System and method for operating comminution process in tumbling mill

Through vibration and position sensor detection, internal status information of the rolling mill is generated, which solves the grinding problems of low efficiency and unstable in the prior art, and realizes real-time monitoring and control of the internal status of the rolling mill.

CN118984735APending Publication Date: 2024-11-19SPM INSTR
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202380033455.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-10-09
Filing Date
2023-04-11
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The existing rolling mills are inefficient during the grinding process and are difficult to balance feed rate and consumption, resulting in unstable operating conditions and lack of real-time loading status monitoring information.

Method used

A vibration sensor is used to detect the mechanical vibration signal when the housing rotates, and information indicating the internal state of the rolling mill is generated by the analysis device, and rotation position signal is obtained in combination with the position sensor, and state parameters are extracted to control the crushing process.

Benefits of technology

It improves the efficiency and stability of the grinding process, provides real-time internal state monitoring and control capabilities, and ensures the optimal state of the crushing process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118984735A_ABST
    Figure CN118984735A_ABST
Patent Text Reader

Abstract

A method of operating a tumbling mill, the method comprising: monitoring a toe position of a first material portion within the mill, the toe position being determined in response to a signal measured from a vibration sensor; determining a measurement corresponding to a discharge characteristic of the second material portion discharged from the mill; receiving an input corresponding to a desired discharge characteristic of the discharged material; and determining an operating parameter of the mill based on the determined toe position, the received desired discharge characteristic, and a predetermined correlation between the toe position and the discharge characteristic.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of roller mills and to the monitoring of roller mills. The present invention also relates to a method for generating information relating to the internal state of a roller mill and to the field of controlling roller mills. The present invention also relates to a method for operating a comminution process in a roller mill and to a device for monitoring the internal state of a roller mill. The present invention also relates to a device for controlling the internal state of a roller mill. The present invention also relates to a computer program for monitoring the internal state of a roller mill. The present invention also relates to a computer program for controlling the internal state of a roller mill. Background Art

[0002] In some industries, such as mining, it is necessary to grind large pieces of material to reduce the size of the individual pieces of material received. Roller mills can achieve this grinding of the material.

[0003] A roller mill includes a housing that contains a charge material to be rolled and ground as the housing rotates. US2017 / 0225172 A1 discloses that grinding in a roller mill can be inefficient, particularly where energy is wasted by impacts that do not break up particles, and that autogenous (AG) and semi-autogenous (SAG) mills sometimes operate in an unstable state because it is difficult to balance the feed rate of large particles into the roller mill with the consumption of the charge. According to US2017 / 0225172 A1, in order to control the process, it is necessary to provide real-time information about the current state of the charge in the roller mill. US 2017 / 0225172 A1 discloses the use of rotor dynamics to determine the characteristics of a moving charge within a roller mill.

[0004] US2017 / 0225172 A1 provides a monitoring device for a roller mill. The device includes vibration sensors mounted on the mill's two main bearings and the mill's thrust bearing, generating vibration signals corresponding to the bearings to which the sensors are mounted. These vibration signals are transmitted to an analyzer, which analyzes the signals and displays the mill's operating status numerically or graphically.

[0005] In US2017 / 0225172 A1 Figure 5 In the US 2017 / 0225172 A1, two orbital graphs are disclosed, one generated at time 1 and the other generated at time 2. According to US 2017 / 0225172 A1, by observing the change in the orbital graph from time 1 to time 2, the mill operator will observe that the amplitude of the vibration has been significantly reduced, and the orbital parameters or frequency or phase or progression or other characteristic changes in the orbital graph have also been significantly reduced. According to US 2017 / 0225172 A1, this information will inform the operator that a very significant change has occurred in the overall operation of the mill rotor and the composite material being processed. Summary of the Invention

[0006] In view of the prior art, the problem to be solved is how to generate improved information about the internal state of a roller mill and / or how to obtain an improved method of operating a comminution process in a roller mill.

[0007] This problem is solved by the examples presented in this article. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] For the purpose of simplifying the understanding of the present invention, the present invention will be described by way of example with reference to the accompanying drawings, in which:

[0009] Figure 1A Shown are outlines and schematic side views of a system including a roller mill.

[0010] Figure 1B Another schematic diagram of a system including a tumble mill is shown.

[0011] Figure 1C is a block diagram illustrating a roller mill as a block that receives multiple inputs and generates multiple outputs.

[0012] Figure 2 Shown along Figure 1A Another example of a cross-sectional view taken along line AA of FIG.

[0013] Figure 3 is a schematic block diagram of an example of the analysis device shown in FIG. 1 .

[0014] Figure 4 is a simplified diagram of program memory and its contents.

[0015] Figure 5 is a block diagram illustrating an example of an analysis device.

[0016] Figure 6A is a graphical representation of the signal pair S(i) and P(i) delivered by the A / D converter.

[0017] Figure 6B is a diagram of a sequence of signal pairs S(i) and P(i) delivered by the A / D converter.

[0018] Figure 7 is a block diagram illustrating an example of a portion of a state parameter extractor.

[0019] FIG8 is a simplified diagram of an example of a memory and its contents.

[0020] Figure 9 It shows the operation Figure 7 Flowchart of an example of a method of a state parameter extractor.

[0021] Figure 10 is shown for execution Figure 9 Flowchart of an example of a method of step S#40.

[0022] Figure 11 is a flowchart illustrating another example of a method.

[0023] Figure 12 is shown for execution Figure 9 Flowchart of another example of the method of step S#40.

[0024] Figure 13 is a graph showing a series of temporally consecutive position signals P1 , P2 , P3 , . . . , each position signal P indicating a complete revolution of the monitored housing.

[0025] Figure 14A Another example of a cross-sectional view of a middle portion 98 of a rotating mill housing during operation is shown.

[0026] Figure 15A is a block diagram illustrating an example of a state parameter extractor.

[0027] Figure 15B is a block diagram illustrating another example of a state parameter extractor.

[0028] Figure 16 is a diagram of one example of a visual indication of analysis results.

[0029] Figure 17 is a diagram of another example of a visual indication of analysis results.

[0030] Figure 18 is a diagram of another example of a visual indication of analysis results.

[0031] Figure 19A and Figure 19B is a diagram of yet another example of a visual indication based on the analysis result of the internal state of the roller mill.

[0032] Figure 20 is a block diagram of an example of a compensated decimator.

[0033] Figure 21 It shows the operation Figure 20 Flowchart of an embodiment of a method of compensating a decimator.

[0034] Figure 22A 、 Figure 22B and Figure 22C Shows the operation Figure 20 Flowchart of an embodiment of a method of compensating a decimator.

[0035] Figure 23 Another example of a cross-sectional view of a middle portion of a rotating mill housing during operation is shown.

[0036] Figure 24 A sketch and schematic top view of another system including a tumble mill is shown.

[0037] Figure 25 A sketch and schematic top view of another embodiment of a system including a tumbler mill is shown.

[0038] Figure 26 A sketch and schematic top view of another embodiment of a system including a tumbler mill is shown.

[0039] Figure 27 A sketch and schematic top view of another embodiment of a system including a tumbler mill is shown.

[0040] Figure 28 A sketch and schematic top view of another embodiment of a system including a tumbler mill is shown.

[0041] Figure 29 A sketch and schematic top view of another embodiment of a system including a tumbler mill is shown.

[0042] Figure 30 Another example of a cross-sectional view of a middle portion of a rotating mill housing during operation is shown.

[0043] Figure 31 is a block diagram illustrating another example of a state parameter extractor.

[0044] Figure 40 is a block diagram of a system for monitoring an internal state X of a mill and enabling improved control of the comminution process taking place in the mill.

[0045] Figure 41 A flow chart showing a method of operating a roller mill is shown.

[0046] Figure 42 A flow chart illustrating another method of operating a roller mill is shown.

[0047] Figure 43 A flow chart illustrating yet another method of operating a roller mill is shown.

[0048] Figure 44 A flow chart illustrating yet another method of operating a roller mill is shown.

[0049] Figure 45 A flow chart illustrating yet another method of operating a roller mill is shown.

[0050] Figure 46 A flow chart illustrating yet another method of operating a roller mill is shown. DETAILED DESCRIPTION

[0051] In the following, similar features in different examples will be denoted by the same reference numerals.

[0052] Figure 1A A schematic diagram and side view of a system 5 including a roller mill 10 are shown. For example, the roller mill 10 can be an autogenous (AG) mill. Alternatively, the roller mill 10 can be, for example, a semi-autogenous (SAG) mill. Another example of a roller mill 10 is a ball mill 10.

[0053] Figure 1A A cross-sectional view of section AA is also shown. Section AA is also indicated by reference numeral 15. The roller mill 10 includes a housing 20 having an inner housing surface 22 that forms a chamber 25 for the grinding material. Figure 1A In the cross-sectional view identified by reference numeral 15, the housing 20 is shown as rotating in a counterclockwise direction at a rotational speed f ROT Rotation, like a curved arrow f ROT shown.

[0054] In operation, the tumbling chamber 25 contains a charge 30 of material to be tumbled and ground. The charge material has a material surface 33, which is the boundary between the air in the rotatable housing 20 and the material 30 (see Figure 2 ). The purpose of grinding in a roller mill is to reduce the size of the particles of solid material. This can be achieved, for example, by causing pieces of solid material to fall onto other pieces of solid material. Therefore, the roller mill uses natural forces, namely gravity, to accelerate the impact of charge particles with other particles of the charge. According to some embodiments, the walls of the housing 20 include a strong material, such as steel, in order to withstand the impact of heavy particles, such as large pieces of ore that are roller milled in the chamber 25. According to some embodiments, the walls of the housing 20 include an elastic material in order to reduce wear of the walls. According to some embodiments, the elastic material includes rubber. According to some embodiments, the elastic material includes a polymer, such as polyurethane. According to some embodiments, the inner housing surface 22 includes a surface coating of an elastic material, such as rubber or polyurethane.

[0055] According to some embodiments, the housing 10 is supported on at least two bearings 40 and 50. The housing 20 is rotatable about an axis of rotation 60. In this regard, it should be noted that an axis is an imaginary line about which an object rotates (the axis of rotation). The rotation of the housing is used to lift a portion of the charge, including particles of solid material, so that some of the solid particles can fall back onto another portion of the charge under the influence of gravity. Therefore, it is desirable to set the rotational speed f of the housing 20 to ROT A suitable value is selected so as to obtain a balance between the lifting action and the falling action of the charge 30. Figure 1A15, arrow 62 indicates the direction of the gravitational force g associated with the rotating shell 20 and its charge 30. Thus, the internal conditions of the roller mill 10 are determined in part by the balance between the gravitational force 62 and the centripetal force 65, which acts to press radially from the center (i.e., from the axis of rotation 60) against the portion of the charge 30 that is immobile relative to the interior shell surface 22. In other words, during operation of the roller mill 10, the centripetal force acts to press a portion of the charge 30 against the shell interior surface 22, and this centripetal force is dependent on the rotational speed f of the shell 20. ROT In this regard, it should be noted that the centripetal force acting on a piece of solid material in contact with the interior housing surface 22 depends on the inner diameter of the housing 20. When the force of gravity 62 acting on a particular piece of solid material 68 is greater than the portion 69 of the centripetal force acting on that piece of solid material 68 in a direction opposite to the direction of gravity, that piece of solid material 68 will fall.

[0056] A vibration sensor 70 may be provided to generate a measurement signal S EA .Measurement signal S EA It may depend on mechanical vibrations or shock pulses generated when the housing 20 rotates.

[0057] One example of system 5 is operable when vibration sensor 70 is securely mounted on or at a measuring point on roller mill 10. The measuring point may include a coupling to which sensor 70 may be securely or removably attached. Figure 1A In the example shown, the sensor 70 is mounted on the bearing 40. Alternatively, the sensor 70 may be mounted elsewhere on the roller mill, wherein the sensor 70 is capable of generating a measurement signal S based on mechanical vibrations or shock pulses generated when the housing 20 rotates. EA .

[0058] The roller mill 10 has an input side 80 for receiving pieces of solid material and an output side 90 for delivering output material 95 that has passed through the roller mill 10 .

[0059] The housing 20 may have a generally cylindrical middle portion 98 with the chamber 25 having an inner radius R MIC For example, the inner radius R MIC It can be more than 0.5 meters. Alternatively, the inner radius R MIC The roller mill 10 may alternatively have a chamber middle inner diameter R of more than 8 meters. MIC The middle portion of the housing 20 has a length L from the input side 80 to the output side 90. MIC For example, the length of the middle shell L MIC According to one embodiment, the length L of the intermediate shell is MIC It should be noted that any inner radius R in this exampleMIC Can be used with any of the case lengths L in this example MIC Combined.

[0060] Furthermore, it should be noted that the housing 20 can have a polygonal middle portion 98. An example of such a polygonal housing shape is a housing having at least three housing wall portions that are connected to form the roller mill chamber 25. In this case, it should be noted that for the purposes of this disclosure, a roller mill housing having a middle portion 98 having at least six housing wall portions that are connected to form the chamber 25a can be considered to have a generally cylindrical shape.

[0061] Thus, for purposes of this disclosure, a roller mill housing having a hexagonally shaped mid-section 98 may be considered to have a generally cylindrical shape.

[0062] exist Figure 1A In the example shown, the input side 80 includes a first input 100 for solid material chunks 110. The solid material 110, also referred to as feed material 110, may include rock and ore chunks 115, with the chunks 115 having various sizes. However, the solid material 110 fed into the first input 100 may have been processed such that a maximum solid material particle size exists. In other words, the feed material 110 may include rock and ore chunks 115, with the chunks 115 having a feed particle size distribution. For example, the feed particle size distribution may be such that a specific maximum input solid particle volume V exists. ISPM and / or a specific maximum input solid particle diameter D ISPM Therefore, the maximum solid material particle size in the feed material can be a certain maximum input solid particle volume V ISPM The solid material 110 may, for example, comprise ore chunks 115 having a particle volume of up to ten (10) cubic decimeters, i.e. individual input solid particles 115 having a maximum input solid particle volume V of less than or at most ten (10) cubic decimeters. ISP Alternatively, the maximum solid material particle size may be a specified maximum input solid particle diameter D ISPM .

[0063] Therefore, a single input solid particle 115 has a maximum input solid particle diameter D of less than or at most 250 mm. ISP .

[0064] The feed particles 115 may include useful minerals as well as minerals that are considered less useful. The less useful minerals may be referred to as waste minerals. To separate the useful minerals from the waste minerals, the solid feed material 110 is ground in a roller mill 10. The ground output material 95 delivered from the roller mill 10 may include particles 96 having a diameter of approximately 0.1 mm or less. The particles 96 delivered from the roller mill 10 may be referred to as product particles 96.

[0065] According to some embodiments, the roller mill 10 is operated to perform dry grinding. According to one embodiment, the roller mill 10 is a ball mill that is operated to perform dry grinding. The ball mill 10 includes a plurality of balls 117 for enhancing the grinding of feed particles of solid feed material 110 into ground solid product particles 96. According to one embodiment, the balls 117 of the ball mill include steel balls. According to one embodiment, the roller mill 10 is a ball mill that is used to grind feed particles 115 of a hard substance into a powder 95 known as cement. In this regard, it should be noted that Portland cement (a type of hydraulic cement) is made by heating limestone (i.e., calcium carbonate) and other materials (e.g., clay) during a period known as calcination, which releases carbon dioxide molecules from the calcium carbonate to form calcium oxide or quicklime, which then chemically combines with other materials in the mixture to form calcium silicate and other cementitious compounds. According to one embodiment, the resulting hard matter is then ground into a powder together with a certain amount of gypsum using the ball mill 10 described above for dry grinding to produce cement.

[0066] According to some embodiments, the roller mill 10 operates to grind a solid material 110. An example of a grinding process employing a roller mill 10 operating to grind a solid material 110 is a roller mill 10 in the mining industry. According to some embodiments, the mining industry roller mill 10 operates to grind a solid material 110 that includes a mixture of useful minerals and minerals considered less useful. According to some embodiments, the mining industry roller mill 10 is an autogenous (AG) mill. Alternatively, the mining industry roller mill 10 is a semi-autogenous (SAG) mill. According to some embodiments, the mining industry roller mill 10 is a ball mill 10.

[0067] According to some embodiments, the solid material 110 is an ore having a metal content. The average metal content in the solid material 110 may be, for example, higher than 0.1%. According to some embodiments, the solid material 110 has an average metal content of more than 5% of the desired metal.

[0068] Alternatively, the average metal content in the solid material 110 may be, for example, 50%. According to some embodiments, the solid material 110 has a desired metal content exceeding 40%. According to some embodiments, the solid material 110 has a desired metal content exceeding 40%, and the desired metal is iron. In this case, it should be noted that the desired metal content in the solid material 110 affects the charge density in the roller mill 10. Therefore, according to some embodiments, the charge density in the roller mill 10 may indicate the relationship between the desired metal and the waste mineral in the charge in the roller mill 10.

[0069] According to some embodiments, the grinding process can be facilitated by providing a liquid 120. An example of a grinding process facilitated by providing a liquid 120 is a roller mill used in the mining industry. According to some embodiments, the liquid 120 enters the roller mill 10 at a second input 130 on the input side 80 of the roller mill 10.

[0070] In the rotating shell 20 , an input mass of solid material 110 is mixed with an input liquid 120 to form a charge 30 .

[0071] When the density of the input liquid 120 is different from the density of the input solid material 110, the density of the charge 30 can be controlled by controlling the ratio of the input liquid 120 to the input solid material 110. Thus, when the density of the input liquid 120 is lower than the density of the input solid material 110, the density of the charge 30 can be decreased by increasing the amount of the input liquid 120.

[0072] The input liquid 120 may include water. Water has a density of approximately 997 kilograms per cubic meter. The density of the input solid material mass is typically higher than the density of the input liquid. The input solid material mass typically has a density exceeding 1500 kilograms per cubic meter. The input solid material 110 may include ore containing useful minerals mixed with other minerals.

[0073] An example of a useful mineral is one containing metals, such as aluminum or iron. Aluminum has a density of approximately 2,700 kilograms per cubic meter. Iron has a density of approximately 7,870 kilograms per cubic meter. The "other minerals" mentioned above can include, for example, granite or other rock masses. Granite has a density of approximately 2,700 kilograms per cubic meter.

[0074] Table 1 provides some examples of solid materials and corresponding material properties.

[0075] Table 1

[0076]

[0077]

[0078] In mineralogy, the term toughness describes a mineral's resistance to fracture, beading, cutting, or other forms of deformation.

[0079] A material is considered brittle if it breaks when stressed with only minimal elastic deformation and no significant plastic deformation. Brittle materials absorb relatively little energy before breaking, even high-strength materials.

[0080] A ductile material can be stretched or shaped by striking or applying pressure. A soft material can be pulled or stretched by mechanical forces without breaking.

[0081] Compressive strength, or compressive strength, is the ability of a material or structure to withstand a load that tends to reduce its size. In contrast, tensile strength is the ability of a material or structure to withstand a load that tends to elongate. In other words, compressive strength resists compression (being pushed together), while tensile strength resists tension (being pulled apart).

[0082] It should be noted that gold has a density of 19,320 kilograms per cubic meter, which is much higher than the other solid materials listed above in Table 1. In this regard, it should also be noted that the gold content of ores that include some gold is generally low compared to the content of other solids used as roller mill feed material 110.

[0083] The output side 90 of the roller mill 10 may include a separator for conveying the output material 95 to the output end 200 and for retaining material lumps with a particle size exceeding a limit. The separator may include a screen configured to screen out material lumps with a particle size less than a specified limit for conveying as the output material 95 to the output end 200. The ground output material 95 conveyed from the roller mill 10 may include particles having a diameter less than a certain limit output particle diameter. The limit output particle diameter may be 0.1 mm.

[0084] One measure of the production quality of the tumble mill 10 may be the proportion of output particles having a particle diameter less than 45 μm (where μm means micrometers), or the amount of output particles having a particle diameter less than 45 μm per hour.

[0085] Furthermore, it is desirable to achieve a highly efficient grinding process. One aspect of grinding process efficiency is the amount of material ground per unit time. Therefore, it is desirable to increase or optimize the kilograms per hour of ground solid material with a particle size below a limit. However, this value is typically metric tons per hour of solid material fed to the roller mill 10.

[0086] Another aspect of the efficiency of the grinding process is the amount of ground material per unit of energy, in order to minimize the energy consumption of the grinding process. Therefore, it is desirable to increase or optimize the output expressed in kg / kWh of ground solid material, with the particle size of the ground solid material being less than a limit value. In this context, it should be noted that roller mills may typically have a power consumption exceeding 4 MW. Some roller mills have an average power consumption of 10 MW, and some roller mills may require a peak power consumption of 20 MW. In this context, it should be noted that when a roller mill has an average power consumption of 10 MW, the energy consumption is 10,000 kW-h per hour. Therefore, when the roller mill is operated 24 hours a day for a year, even a small improvement in the energy efficiency of the grinding process (e.g., a one percent (1%) improvement) can save 6 million kW-h of energy per year.

[0087] The efficiency of the grinding process in the roller mill 10 depends on a number of variables that affect the internal conditions of the roller mill 10. One variable that affects the efficiency of the grinding process in the roller mill 10 is the filling level of the roller mill 10. Therefore, it is desirable to control the inflow of input solid material 110 so as to achieve an optimal filling level.

[0088] In order to maximize the amount of output material 95 from the roller mill 10, it is therefore desirable to control the inflow of input material 110 so as to maintain optimal conditions for the roller milling process. The optimal internal conditions for the roller milling process may include a certain filling degree of the housing 20, i.e., a certain charge volume. Therefore, one variable that has an impact on the efficiency of the grinding process in the roller mill 10 is the solid feed rate R s , i.e. the amount of solid material particles fed into the roller mill 10 per unit time.

[0089] Another variable that affects the efficiency of the grinding process in the roller mill 10 is the mineralogical properties of the input solid material particles 110. In this regard, it should be noted that mineralogy is a branch of geology that studies the physical properties of minerals, as well as their chemical and crystal structures. Furthermore, the mineralogical properties of the particles in the charge 30 are not constant over time, as the composition of the solid material 110 (e.g., ore from a mine) typically changes over time. Changes in the mineralogical properties of the particles in the charge 30 can affect the efficiency of the grinding process in the roller mill 10. Therefore, due to changes in the mineralogical properties of the particles in the material charge 30, the efficiency of the grinding process may vary over time. Therefore, if the material feed remains constant, over a certain timeframe, the grinding process may decrease in efficiency, resulting in an increase in the charge volume in the mill 10. Therefore, unless the operator of the roller mill is fully informed of the current charge volume in the mill 10, there is a risk of overloading, which, in the worst case, could lead to a complete halt of the grinding process.

[0090] Another variable that has an impact on the efficiency of the grinding process is the size distribution of the solid material particles 110 fed into the roller mill 10. According to some embodiments, the feeding of the solid material particles 110 is controlled so that a certain proportion of the solid material particles 110 provided at the first input 100 have a single volume greater than one cubic decimeter, thereby improving the efficiency of the grinding process. It has been inferred that controlling the feeding of the solid material particles 110 so that a certain proportion of the solid material particles 110 provided at the first input 100 have a single volume greater than one cubic decimeter increases the efficiency of the grinding process, particularly when the roller mill is an AG mill or a SAG mill.

[0091] The housing 20 is generally opaque, ie, it is not possible to visually inspect the charge in the housing during operation of the roller mill 10. Furthermore, the movement of the tumbled heavy ore during operation of the roller mill 10 prevents placement of a camera or other sensitive detector inside the housing 20.

[0092] One object of this document is to describe a method and system for improving the monitoring of the internal conditions of a roller mill during operation. Another object of this document is to describe a method and system for an improved human-machine interface (HCI) related to the internal conditions of a roller mill during operation. Another object of this document is to describe a method and system for an improved graphical user interface related to the grinding process in a roller mill 10.

[0093] The inventors have realized that during operation of the roller mill 10, there may be mechanical vibrations V IMP , indicating an impact between a protrusion (e.g., a lifter) on the inner surface of the rotating shell 20 and at least one particle in the toe 205 of the material charge 30. The inventors also considered that this mechanical vibration V IMP The current internal state of the roller mill 10 and / or the current state of the grinding process can be indicated. When the protrusions (e.g., lifters) interact with particles in the toe 205 of the material charge 30 in the chamber 25, mechanical vibrations V are generated. IMP The impact force F of the interaction between the rotary elevator and the material charge 30 IMP The impact causes acceleration of at least one particle in the toe 205 of the material charge 30, which results in a mechanical shock vibration V IMP In fact, the impact force F IMP Can cause mechanical shock vibration V IMP , which indicates the current internal state of the roller mill 10 and / or indicates the current state of the grinding process.

[0094] The sensor 70 placed outside the chamber 25 can detect vibrations caused by the interaction of particles of the charge 30 in the chamber 25 during operation of the roller mill 10. Figure 1AThe sensor 70 can generate a measurement signal S according to the mechanical vibration or shock pulse generated when the housing 20 rotates. EA Therefore, the measurement signal S EA The force F between the protrusion (eg, lifter) and at least one particle in the toe 205 of the material charge 30 during operation of the roller mill 10 may depend on and be indicative of the impact force F between the protrusion (eg, lifter) and at least one particle in the toe 205 of the material charge 30. IMP .

[0095] The sensor 70 may be, for example, an accelerometer 70 configured to generate a measurement signal S EA , the amplitude of the measured signal depends on the impact force F IMP The inventors concluded that there may be mechanical vibrations V that indicate the current internal state of the roller mill 10 and / or the current state of the grinding process. IMP , but conventional methods for measuring vibrations and / or for analyzing and / or for visualizing such vibrations may hitherto be insufficient.

[0096] An analysis device 150 is provided for monitoring the tumbling process. The analysis device 150 may also be referred to as a monitoring module 150A.

[0097] The analysis device 150 can be used to analyze the measurement signal S EA Generate information indicating the internal state of the tumbling process. Generate measurement signal S EA The sensor 70 is coupled to the input 140 of the analysis device 150 in order to transmit the measurement signal S EA The analyzing device 150 also has a second input terminal 160 for receiving a position signal E according to the rotational position of the housing 20. P .

[0098] A position sensor 170 is provided to generate a position signal E according to the rotational position of the housing 20. P As described above, the housing 20 is rotatable around the rotation axis 60, so the position sensor 170 can generate a series of housing position signal values ​​P S Position signal E P , used to indicate the instantaneous rotational position of the housing 20. The position mark 180 can be set on the outer surface of the housing 20, so that when the housing 20 rotates around the rotation axis 60, the position mark 180 passes the position sensor 170 once per each rotation of the housing, thereby causing the position sensor 170 to generate a rotation mark signal P S This rotation mark signal P S The analysis device 150 can be configured to determine the rotational position of the housing 20 based on the position signal E. P Generates the speed f indicating the housing 20 ROTWhen the position sensor 170 is an optical device (e.g., a laser transceiver), the position marker 180 may be, for example, an optical device 180 (e.g., a reflector 180) configured to generate a rotation mark signal P when the intensity of the laser reflection changes due to the laser beam irradiating the reflector 180. S Alternatively, when the position sensor 170 is a device 170 configured to detect a changing magnetic field, the position marker 180 may be, for example, a magnetic device 180, such as a strong magnet 180. An example of a device configured to detect a changing magnetic field is a device including an induction coil that generates a current in response to a changing magnetic field. Thus, the device 170 configured to detect a changing magnetic field is configured to generate a rotation marker signal P when passing by the magnetic device 180. S Alternatively, the position sensor 170 may be implemented as an encoder 170 that is mechanically coupled to the rotating mill housing 20 such that the encoder generates, for example, a flag signal P for each rotation of the rotating mill housing 20. S .

[0099] The system 5 may include a control room 220 that allows a mill operator 230 to operate the roller mill 10. The analysis device 150 may be configured to generate information indicating the internal state of the roller mill 10. The analysis device 150 also includes a device human-machine interface (HCI) 210 for enabling user input and user output. The HCI 210 may include a display or screen 210S for providing a visual indication of the analysis results. The displayed analysis results may include information indicating the internal state of the roller mill process, enabling the operator 230 to control the roller mill.

[0100] The roller mill controller 240 is configured to deliver a solid material feed rate set point R SSP , and may also optionally transmit a liquid feed rate set point R LSP According to some embodiments, the set point value R SSP Set by operator 230. According to some embodiments, the set point value R LSP is also set by the operator 230. Thus, the roller mill controller 240 may include a mill user input / output interface 250, enabling the operator to adjust the solid material feed rate R S and / or liquid feed rate R L .

[0101] As described above, the input side 80 of the roller mill includes a first input end 100 for a block of solid material 110, and optionally, the input side 80 may also have a second input end 130 for a liquid 120 (e.g., water) to enter the chamber 25. The solid material 110 may be conveyed to the first input end 100 by a conveyor belt 260. The conveyor belt 260 operates at a conveyor belt speed, and the solid material feed rate R S Solid material 110 is delivered to the first input 100 .

[0102] During operation of the roller mill 10, under certain internal conditions of the roller mill 10, the solid material feed rate R S It can be, for example, 10,000 kg per minute. Similarly, during operation of the roller mill 10, under certain internal conditions of the roller mill 10, the liquid feed rate R L It could be, for example, 1000 kilograms per minute.

[0103] exist Figure 1A and / or Figure 1B The liquid feed rate R is schematically shown by the symbol of the controllable valve 270. L The controllable valve receives the liquid feed rate set point R from the roller mill controller 240. LSP . Similarly, in Figure 1A and / or Figure 1B The solid material feed rate R is schematically shown by the symbol of the controllable valve 280. SF The controllable valve receives the solid material feed rate set point R from the roller mill controller 240. SSP .

[0104] According to some embodiments, the roller mill controller 240 can also set the rotational speed f of the mill housing. ROT Generate setpoint value f ROT_SP . Speed ​​set point value f ROT_SP Also known as U1 SP The speed setpoint value f may be generated in response to user input from operator 230 via user input / output interface 250 ROT_SP , also known as U1 SP ,like Figure 1B shown.

[0105] like Figure 1B As shown, the roller mill controller 240 can also generate a set point value U2 SP and U3 SP , among which U2 SP is the above R SSP , U3 SP is the above R LSP .

[0106] Additionally, the roller mill controller 240 may also generate a ball feed rate set point value U4 SP , R BFSP , used to set the ball feed rate U4, R BF Ball feed rate U4, R BF is the number of grinding balls fed into the ball mill per unit time to enhance the comminution process. Thus, when the mill 10 is a ball mill, i.e., includes grinding balls 1168 (see Figure 36 ) of a roller mill, this set point value may be relevant.

[0107] The balls used in the ball mill can include chrome steel or stainless steel. Alternatively, the ball mill can use balls made of ceramic material. In some examples, the ball mill can use balls including rubber material.

[0108] exist Figure 1A In the example shown, the mill user input / output interface 250 is coupled to the regulator 240, and the HCI 210 is coupled to the analysis device 150 or monitoring module 150A, which is configured to generate information indicative of the internal state of the roller mill 10. Figure 1A As shown coupled only to the monitoring module 150A, the HCI 210 can advantageously be added to the control room 220 without modifying any pre-existing input / output interfaces 250 and regulators 240 used by the mill operator 230 to operate the roller mill 10 .

[0109] The solutions and examples disclosed in this document address an objective of describing methods and systems for improving the monitoring of internal conditions X in a roller mill 10 during operation. Another objective addressed by the solutions and examples disclosed in this document is to describe methods and systems for improving the control of internal conditions X in a roller mill 10 during operation. Furthermore, the solutions and examples disclosed in this document address an objective of describing methods and systems for an improved human-machine interface (HCI) that involves conveying useful information about the internal conditions X of the roller mill during operation. Another objective addressed by this document is to describe methods and systems for an improved graphical user interface related to the grinding process in the roller mill 10.

[0110] Another object addressed by the solutions and examples disclosed in this document is to describe a method and system for improving the control of the output Y of the roller mill 10 during operation. Yet another object addressed by the solutions and examples disclosed in this document is to describe a method and system for an improved human-machine interface (HCI) involving the communication of useful information about the output Y of the roller mill 10 during operation and / or also about corresponding internal states X in the roller mill during operation.

[0111] Figure 1B There is shown another schematic diagram of a system 320 including a roller mill 10. Thus, reference numeral 320 relates to a system including a mill 10 having a rotatable shell 20, as described herein. Figure 1B The system 320 may include the above-mentioned Figure 1A and Figure 2 as described and / or as described elsewhere in this document and configured.

[0112] Despite Figure 1A In the example shown, the mill user input / output interface 250 is coupled to the regulator 240, and the HCI 210 is a separate input / output interface coupled to the analysis device 150 or the monitoring module 150A, but Figure 1B The system shown may provide an integrated HCI 210, 250, 210S. Thus, Figure 1B The input / output interface 210 may be configured to enable all of the input and / or output described above in connection with interfaces 210 and 250 .

[0113] Figure 1C is a block diagram showing a roller mill as a block 10B that receives a plurality of inputs U1, ...Uk and generates a plurality of outputs Y1, ...Yn. Figure 1C It should be noted that for analytical purposes, the roller mill 10 can be viewed as a black box 10B having a plurality of input variables, referred to as input parameters U1, U2, U3, ..., Uk, where index k is a positive integer. During operation of the black box roller mill 10B, the black box roller mill 10B has an internal state X and produces a plurality of output variables, also referred to as output parameters Y1, Y2, Y3, ..., Yn, where index n is a positive integer.

[0114] The internal state X of the mill can be described or indicated by a plurality of internal state parameters X1, X2, X3, ..., Xm, where the index m is a positive integer.

[0115] Using the terminology of linear algebra, the input variables U1, U2, U3, ...Uk can be collectively referred to as the input vector U; the internal state parameters X1, X2, X3, ...Xm can be collectively referred to as the internal state vector X; and the output parameters Y1, Y2, Y3, ...Yn can be collectively referred to as the output vector Y.

[0116] At a point in time called r, the internal state X of the mill 10 may be referred to as X(r). The internal state X(r) may be described or indicated by a plurality of parameter values ​​that define different aspects of the internal state X(r) of the mill 10 at time r.

[0117] The internal state X(r) of the black box roller mill 10B depends on the input vector U(r), and the output vector Y(r) depends on the internal state vector X(r). One aspect of the internal state X is the total amount of material 30 in the housing 20, and this amount does not change immediately. Therefore, during operation of the mill 10, the internal state X(r) can be considered as a function of the earlier internal state X(r-1) and the input U(r):

[0118] X(r)=f1(X(r-1), U(r)), wherein X(r-1) indicates the internal state X of the mill 10 at a point in time before a point in time called r.

[0119] Likewise, the output Y of the black box 10B can be viewed as a function of its internal state X:

[0120] Y(r)=f2(X(r)).

[0121] Figure 2 It is along Figure 1A , taken along line AA, shows a more detailed example of the middle portion 98 of the housing 20. The housing 20 has an inner housing surface 22 facing the chamber 25, the inner housing surface 22 including a plurality of protrusions 310. According to some embodiments, at least two protrusions 310 are provided. Figure 2 The example housing 20 shown includes twelve protrusions 310 positioned equidistant from one another on the interior housing surface 22 of the housing 20. The protrusions 310 can be configured to engage and lift the material 30 as the housing rotates about the axis 60. Thus, the protrusions 310 can be referred to as lifters 310. The charge material has a material surface 33, which is a boundary between the air in the rotatable housing 20 and the material 30.

[0122] exist Figure 2 In the clockwise direction at a speed f ROT The housing 20 is shown during rotation. The lift 310 includes a structure, such as an internal structure, texture, bar, protrusion, etc., that protrudes from the housing interior surface 22 toward the center of the housing 20. The lift 310 (also referred to as a protrusion 310) has a leading edge 312 that engages and lifts the material charge 30 as the roller mill 10 rotates about the axis 60, causing the material to fall freely within the interior chamber 25. In one example, the lift 310 includes an elongated rod mounted on the interior housing surface wall 22 so as to at least partially line the interior housing surface 22 of the mill 10. In other examples, the lift 310 is integrally formed with the interior housing surface wall 22 as part of a single, unitary body. According to some embodiments, the leading edges 312 of the protrusion 310 are equidistant. Thus, with reference to Figure 2The example housing 20 shown includes twelve protrusions 310, each of which has a leading edge 312, and the angular distance between any two adjacent leading edges 312 is 30 degrees. In this case, it should be noted that when there are L protrusions 310 on the inner housing surface 22, and the L protrusions 310 are positioned so that the leading edges 312 of the protrusions 310 are equidistant, the angular distance between any two adjacent leading edges 312 is 360 / L degrees. Therefore, when there are L protrusions 310 at angular locations on the inner housing surface 22, and the L protrusions 310 are positioned equidistant from each other, the angular distance between any two adjacent protrusions 310 is 360 / L degrees.

[0123] exist Figure 2 In the example shown, the position sensor 170 is mounted in a fixed manner so as to generate a position signal E having a series of position signal values ​​PS. P , used to indicate the instantaneous rotational position of the housing 20. The position marker device 180 can be disposed on the outer wall surface of the housing 20, such that when the housing 20 rotates about the rotation axis 60, the position marker 180 passes the position sensor 170 once per each rotation of the housing, thereby causing the position sensor 170 to generate a rotation marker signal value PS. The position sensor 170 may include a tachometer 170 that transmits a position signal pulse Ep per rotation, for example.

[0124] The position marking device 180 may include a metal object. For example, the metal object may be a bolt or a metal bracket.

[0125] refer to Figure 1B Solid material particles 115 enter the mill housing 20 through the material feed opening 100 and are broken down during operation of the mill 10 due to collisions with other particles 115, 30 and / or the internal housing surface 22 and / or balls. The breaking generates solid material product particles 96, also referred to as solid material product or product particles 96. The solid material product exits the mill housing 20 through the outlet 200.

[0126] An important aspect of the grinding process is the crushing rate. The crushing rate depends on, for example, the frequency of collisions experienced by the solid feed material particles 115 upon entering the rotating shell 20.

[0127] Another important aspect of the grinding process is the particle size distribution of the solid material product particles 96 generated by the collision. The particle size distribution can also be called the appearance distribution function.

[0128] Another important aspect of the grinding process is the flow rate of the solid material product particles 96 out of the mill housing 20. The transport of particles out of the mill housing 20 may also be referred to as the product discharge rate.

[0129] Therefore, in summary: the solid feed particles 115 are fed at a solid material feed rate R S Enters the chamber 25 of the mill 10. The feed particles 115 have a first particle size distribution, also referred to as a feed size distribution. When the solid feed material is fed into the mill 10, it can be measured. A feed material analyzer 325 can be provided for generating a measurement value indicating at least one feed material characteristic U5, U6. At least one feed material characteristic can include a solid feed material particle size distribution U5 and / or data U6 indicating the weight per unit time of the solid material 110 delivered to the first input end 100 of the mill 10. Therefore, the feed material particle size distribution can be estimated, for example, by measurement. Alternatively, the solid feed material particle size distribution U5 can be predetermined. In some examples, the solid feed material particle size distribution U5 is known due to processing and / or classification before the solid feed material 110 is conveyed to the conveyor belt 260.

[0130] Once received in the grinding chamber 25, the received particles may be collectively referred to as charge material 30. While in the grinding chamber 25, the solid material particles 30 break up, resulting in solid product particles 96 being discharged from the chamber 25 via the output 200. The breakage results in a change in the particle size distribution of the particles.

[0131] Solid material product particles 96 are discharged at a product discharge rate R SDis Out of the grinding chamber 25. The product discharge rate R can be measured SDis , and it can be considered as the output parameter Y1.

[0132] The discharged solid product particles 96 have a second particle size distribution, also referred to as a product particle size distribution. The product particle size distribution can be measured and a value indicative thereof can be provided, for example, as output parameter values ​​Y2, Y3, etc.

[0133] Therefore, feed particles 115 having a first particle size distribution or feed size distribution U4 are fed at a solid material feed rate R S The product particles 96 fed into the mill 10 and having the product particle size distribution Y2 or the second particle size distribution Y2 are discharged at a product discharge rate Y1=R SDis Discharged from the mill 10.

[0134] During the transport of the feed particles 115 from the inlet 100 to the outlet 200, the feed particles 115 are converted into a plurality of smaller product particles 96. This conversion is due to the comminution process that occurs during operation of the roller mill 10.

[0135] It is believed that the product particle size distribution depends on the following:

[0136] - feed particle size distribution, and

[0137] The internal state X of the mill 10 during the period of conveying the received feed particles 115 from the feed input 100 until the output material 95 is discharged from the mill output 200 in the form of solid product particles 96 .

[0138] Therefore, the amount of crushing of the solid feed material particles 115 depends on the above-mentioned crushing rate and the internal state X of the mill 10. The internal state X of the mill 10 determines the time duration T from the delivery of the received feed particles 115 from the feed input 100 until the discharge of the output material 95 from the mill outlet 200. C The average duration of the transport from the feed input 100 to the mill output 200 may be referred to as the duration T C .

[0139] refer to Figure 1B , which illustrates a Cartesian coordinate system with three mutually perpendicular axes, x, y, and z. It should be understood that during operation of the mill 10, the material 30 travels from the mill input side 80 to the mill output side 90 in the positive direction of the x-axis. Thus, while the material 30 is being tumbled in the mill 10, it also gradually travels from the mill input side to the mill output side in a direction parallel to the mill's axis of rotation. However, different individual particles may travel this distance at varying speeds. A large number of small feed particles may be transported from the mill input side to the mill output side and then discharged via the mill output 200 relatively quickly, for example because they are small enough to pass through the outlet grate in a relatively short time, while the largest feed particles require more time in the mill to be ground into product particles small enough to pass through the same outlet grate. Consequently, a certain proportion of large feed particles will remain in the mill longer than a corresponding proportion of small feed particles.

[0140] However, during steady-state operating conditions, the weight flow rate into, through, and out of the mill 10 will be constant or approximately constant. Thus, the transport of material 30 through the mill 10 can be discussed in terms of weight per unit time, for example, measured in kilograms per minute or metric tons per hour.

[0141] In this regard, it should be noted that the average duration of the transport of the feed particles 115 from the feed input 100 to the mill output 200 and the simultaneous conversion of these feed particles 115 into a plurality of smaller product particles 96 depends on the average flow velocity v in the mill chamber 25 along the x-axis. xA Therefore, at least under steady-state conditions of the grinding process, the average duration of the particles being transported from the feed input 100 to the grinding output 200 depends on the length of the grinding chamber 25 in the x-direction and the average flow velocity v xA .

[0142] Figure 3FIG1 is a schematic block diagram of an example of the analysis device 150 shown in FIG1. ​​The analysis device 150 has a function for receiving the analog vibration signal S from the vibration sensor 70. EA The input terminal 140 is connected to the analog-to-digital (A / D) converter 330. The A / D converter 330 samples the data at a specific sampling frequency f S The received analog vibration signal S EA Sampling is performed in order to transmit data with the specific sampling frequency f S The digital measurement data signal S MD , and wherein the amplitude of each sample depends on the amplitude of the analog signal received at the sampling instant. The digital measurement data signal S is transmitted on the digital output terminal 340 coupled to the data processing device 350 MD .

[0143] refer to Figure 3 The data processing device 350 is coupled to a computer-readable medium 360 for storing program code. The computer-readable medium 360 may also be referred to as memory 360. The program memory 360 is preferably a non-volatile memory. The memory 360 may be a read / write memory, i.e., data can be read from the memory and new data can be written to the memory 360. According to one example, the program memory 360 is implemented as flash memory. The program memory 360 may include a first memory segment 370 for storing a first set of program code 380, which is executable to control the basic operations of the analysis device 150. The program memory 360 may also include a second memory segment 390 for storing a second set of program code 394. The second set of program code in the second memory segment 390 may include program code for causing the analysis device 150 to process the detection signal. Signal processing may include processing for generating information indicating the internal state of the roller mill, as discussed elsewhere in this document. Furthermore, signal processing may include controlling the internal state of the roller mill, as discussed elsewhere in this document. Thus, signal processing may include generating data indicative of the internal state of the roller mill, such as in combination with e.g. Figure 5 , Figure 15 and / or Figure 24 The embodiment of the state parameter extractor 450 is disclosed.

[0144] The memory 360 may further include a third memory segment 400 for storing a third set of program codes 410. The program codes 410 in the third memory segment 400 may include program codes for causing the analysis device to perform a selected analysis function. When the analysis function is performed, the analysis device may present corresponding analysis results on the user interface 210, 210S or transmit the analysis results on the port 420.

[0145] The data processing device 350 is also coupled to a read / write memory 430 for data storage. Thus, the analysis device 150 includes a data processor 350 and program code for causing the data processor 350 to perform certain functions, including digital signal processing functions. When a statement is made in this document that the device 150 performs a certain function or a certain method, this statement may mean that a computer program is executed in the data processing device 350 to cause the device 150 to perform the method or function described in this document.

[0146] The processor 350 may be a digital signal processor. The digital signal processor 350 may also be referred to as a DSP. Alternatively, the processor 350 may be a field programmable gate array circuit (FPGA). Thus, the computer program may be executed by the field programmable gate array circuit (FPGA). Alternatively, the processor 350 may include a combination of a processor and an FPGA. Thus, the processor may be configured to control the operation of the FPGA.

[0147] Figure 4 is a simplified illustration of program memory 360 and its contents. The simplified illustration is intended to convey an understanding of the general idea of ​​storing different program functions in memory 360 and is not necessarily a correct technical teaching of the manner in which the programs would be stored in actual memory circuits. A first memory section 370 stores program code for controlling the analysis device 150 to perform basic operations. Although Figure 4 The simplified diagram of FIG shows pseudo code, but it should be understood that the program code can be a machine code or can be processed by the data processing device 350 ( Figure 3 ) consists of program code at any level of execution or interpretation.

[0148] Figure 4 The second memory segment 390 is shown storing a second set of program code 394. When executed on the data processing device 350, the program code 394 in the segment 390 will cause the analysis device 150 to perform a function, such as a digital signal processing function. The function may include the digital measurement data signal S MD Advanced mathematical processing.

[0149] The computer program for controlling the functions of the analysis device 150 can be downloaded from the server computer. This means that the program to be downloaded is transmitted via a communication network. This can be achieved by transmitting the program over the communication network by modulating a carrier wave. Thus, the downloaded program can be loaded into a digital memory, for example, the memory 360 (see Figure 3 and Figure 4 ). Thus, the program 380 and / or the signal processing program 394 and / or the analysis function program 410 can communicate with the user via, for example, port 420 ( Figure 1A and / or Figure 1B and Figure 3) and other communication ports so that they can be loaded into the program memory 360.

[0150] Therefore, the present document also relates to a computer program product, such as program code 380 and / or program code 394 and / or program code 410, which can be loaded into the digital memory of the device. The computer program product comprises software code portions for performing the signal processing methods and / or analysis functions when the product is run on the data processing unit 350 of the device 150. The term "running on the data processing unit" means that the computer program plus the data processing means 350 performs the method described in the present document.

[0151] The phrase "computer program product loadable into the digital memory of an analysis device" means that a computer program can be introduced into the digital memory of an analysis device 150 to implement an analysis device 150 that is programmed to be capable of or adapted to perform the methods described herein. The term "loaded into the digital memory of the device" means that the device, thus programmed, is capable of or adapted to perform the functions and / or methods described herein. The computer program product described above may also be a program 380, 394, 410 loadable onto a computer-readable medium (e.g., a compact disc or DVD). Such a computer-readable medium can be used to transmit the program 380, 394, 410 to a client. Alternatively, as described above, the computer program product may include a carrier wave that is modulated to transmit the computer program 380, 394, 410 via a communications network. Thus, the computer program 380, 394, 410 can be transmitted from a vendor's server to a client having the analysis device 150 via an Internet download.

[0152] Figure 5 is a block diagram showing one example of the analysis device 150. Figure 5 In the example, some functional blocks represent hardware, some functional blocks may represent hardware, or may represent functions implemented by running program codes on the data processing device 350, such as in combination with Figure 3 and Figure 4 discussed.

[0153] Figure 5 The device 150 shows Figure 1A and / or Figure 1B and / or Figure 3 An example of an analysis device 150 is shown in FIG. To simplify understanding, Figure 5 Also shown are some peripheral devices coupled to the device 150. The vibration sensor 70 is coupled to the input 140 of the analysis device 150 to transmit the analog measurement signal S EA (Also called vibration signal S EA ) is transmitted to the analysis device 150.

[0154] Furthermore, a position sensor 170 is coupled to the second input 160. Thus, the position sensor 170 generates a position signal E depending on the rotational position of the housing 20. P is transmitted to the second input 160 of the analysis device 150 .

[0155] The input terminal 140 is connected to an analog-to-digital (A / D) converter 330. The A / D converter 330 operates at a specific sampling frequency f S The received analog vibration signal SEA is sampled so as to transmit the signal having the specific sampling frequency f S The digital measurement data signal S MD , and wherein the amplitude of each sample depends on the amplitude of the analog signal received at the sampling moment. The digital measurement data signal S is transmitted on the digital output terminal 340 MD , the digital output is coupled to the data processing unit 440. The data processing unit 440 includes functional blocks that show the functions performed. In terms of hardware, the data processing unit 440 may include a data processing unit 350, a program memory 360 and a read / write memory 430, as described above in conjunction with Figure 3 and Figure 4 As described. Therefore, Figure 5 The analysis device 150 may include a data processing unit 440 and program codes for causing the analysis device 150 to perform certain functions.

[0156] Digital measurement data signal S MD With position signal E P Therefore, the A / D converter 330 can be configured to process the analog vibration signal S EA While sampling the position signal E P Position signal E P The sampling frequency f can be used S To generate a digital position signal E PD , where the amplitude of each sample P(i) depends on the received analog position signal E at the sampling moment P The amplitude.

[0157] As mentioned above, the simulated position signal E P Can have a marker signal value P S , for example, in the form of an electrical pulse, the marker signal value has an amplitude edge that can be accurately detected and indicates a specific rotational position of the monitored housing 20. Therefore, although the analog position marker signal P S Has an amplitude edge that can be accurately detected, but the digital position signal E PD The switch from a first value, eg, "0" (zero), to a second value, eg, "1" (one), will occur at different times.

[0158] Therefore, the A / D converter 330 can be configured to transmit a series of measurement value pairs S(i) associated with corresponding position signal values ​​P(i). The letter "i" in S(i) and P(i) represents a time point, i.e., a sample number. Therefore, the time series of position signal values ​​P(i) can be analyzed and the digital position signal E(i) can be identified. PD The sample P(i) has switched from a first value (eg, “0” (zero)) to a second value (eg, “1” (one)) to detect the occurrence time of the rotation reference position of the rotating housing.

[0159] Figure 6A is a graphical representation of the signal pair S(i) and P(i) delivered by the A / D converter 330 .

[0160] Figure 6B is a diagrammatic representation of a sequence of signal pairs S(i) and P(i) transmitted by A / D converter 330. The first signal pair comprises a first vibration signal amplitude value S(n) associated with sampling instant "n", which is transmitted simultaneously with a first position signal value P(n), associated with sampling instant "n". This is followed by a second signal pair comprising a second vibration signal amplitude value S(n+1) associated with sampling instant "n+1", which is transmitted simultaneously with a second position signal value P(n+1) associated with sampling instant "n+1", and so on.

[0161] refer to Figure 5 , the signal pair S(i) and P(i) is transmitted to the state parameter extractor 450. The state parameter extractor 450 is configured to generate an amplitude peak value S based on the time series of the measured sample value S(i). P (r). Amplitude peak S P (r) may depend on the impact force F generated when the protrusion 310 on the inner shell surface of the rotating shell interacts with the toe 205 of the material charge 30 IMP (See Figure 2 In this regard, it should be noted that as the housing 20 rotates, the surface 33 of the material 30 will deviate from the horizontal orientation due to a combination of centrifugal force and gravity. Figure 2 As shown, the toe 205 of the charge material 30 is located at the lower edge of the surface 33 .

[0162] The state parameter extractor 450 is further configured to generate a state parameter based on the peak amplitude S P The time duration (T) between the occurrence time of (r) and the occurrence time of the rotation reference position of the rotating shell D ) to generate the time relationship value R T (j), also known as R TAs mentioned above, the time series of the position signal values ​​P(i) can be analyzed and the digital position signal E(i) can be identified. PD The occurrence time of the rotation reference position of the rotating housing is detected by switching from a first value (eg, “0” (zero)) to a second value (eg, “1” (one)).

[0163] Figure 7 is a block diagram showing an example of a part of a state parameter extractor 450. According to an example, the state parameter extractor 450 comprises a memory 460. The state parameter extractor 450 is adapted to receive a sequence of measurement values ​​S(i) and a signal sequence of positions P(i) and a time relationship therebetween, and the state parameter extractor 450 is adapted to provide a time-coupled value sequence S(i), f ROT (i) and P(i). Therefore, a single measurement value S(i) is associated with the corresponding velocity value f ROT (i) Associated, speed value f ROT (i) indicates the rotational speed of the housing 20 at the time the associated single measurement value S(i) is detected. This will be described in detail below with reference to Figures 8-13.

[0164] Figure 8 is a simplified illustration of an example of memory 460 and its contents, and columns #01, #02, #03, #04 and #05 on the left-hand side of the illustration of memory 460 provide illustrative images intended to show the temporal relationship between the detection time of the encoder pulse signal P(i) (see column #02) and the corresponding vibration measurement value S(i) (see column #03).

[0165] As mentioned above, the analog-to-digital converter 330 uses the initial sampling frequency f S For analog electrical measurement signal S EA Sampling is performed to generate a digital measurement data signal S MD . It is also possible to have essentially the same initial time resolution f S The encoder signal P is detected, as shown in column #02 of FIG8.

[0166] Column #01 shows the time course as a series of time slots, each time slot having a duration dt=1 / f Sample ; where f Sample Is the analog electrical measurement signal S EA The initial sampling frequency f for sampling S The sampling frequency has an integer relationship. According to a preferred example, the sampling frequency f Sample is the initial sampling frequency f S According to another example, the sampling frequency f Sample is the first reduced sampling frequency f SR1 , and the initial sampling frequency f S Compared with , it reduces the integer multiple M.

[0167] In column #02 of FIG8 , each positive edge of the encoder signal P is represented by a "1." In this example, positive edges of the encoder signal P are detected in the 3rd, 45th, 78th, and 98th time slots, as shown in column #02. According to another example, negative edges of the position signal are detected, which provides equivalent results to detecting positive edges. According to yet another example, both positive and negative edges of the position signal are detected to provide redundancy by allowing a later selection of whether to use positive or negative edges.

[0168] Column #03 shows a sequence of vibration sample values ​​S(i). Column #05 shows a corresponding sequence of vibration sample values ​​S(j) when integer decimation is performed. Thus, when integer decimation is performed by this stage, it can, for example, be set to provide an integer decimation factor M=10, and as shown in Figure 8, one vibration sample value S(j) (see column #05 in Figure 8) will be provided for every ten samples S(i) (see column #03 in Figure 8). According to one example, very precise position and time information PT associated with the extracted vibration sample values ​​S(j) is maintained by setting the position time (PositionTime) signal in column #04 to a value PT=3 to indicate that a positive edge was detected in time slot #03 (see column #02). Thus, the value of the position time signal after integer decimation indicates the time of detection of the position signal edge P relative to the sample value S(l).

[0169] In the example of Figure 8, the amplitude value of the position time signal at sample i=3 is PT=3, and due to the decimation factor M=10, sample S(l) is transmitted in time slot 10, which means that the edge is detected in M-PT=10-3=7 time slots before the time slot of sample S(l).

[0170] Thus, the device 150 may be operable to process information about the positive edges of the encoder signal P(i) in parallel with the vibration samples S(i) in order to establish the velocity value f by detecting the analog signal. ROT The above signal processing maintains the time relationship between the positive edge of the encoder signal P(i) and the corresponding vibration sample value S(i) and / or the integer-decimated vibration sample value S(j).

[0171] Figure 9 It shows the operation Figure 7 Flowchart of an example of a method of the state parameter extractor 450.

[0172] According to one example, the state parameter extractor 450 analyzes (step S#10) the temporal relationship between three consecutively received position signals to determine whether the monitored rotating housing 20 is in a constant speed phase or an acceleration phase. As described above, this analysis can be performed based on the information in the memory 460 (see FIG. 8 ).

[0173] If the analysis shows that there are the same number of time gaps between the position signals, then the state parameter extractor 450 concludes (at step #20) that the speed is constant, in which case step S#30 is executed.

[0174] In step S#30, the state parameter extractor 450 can calculate the duration between two consecutive position signals by multiplying the duration of the time slot dt=1 / fs by the number of time slots between the two consecutive position signals. When the monitored housing 20 provides a position signal once per full rotation, the rotational speed can be calculated as:

[0175] V=1 / (n diff *dt),

[0176] Among them, n diff = the number of time slots between two consecutive position signals. During the constant speed phase, all sample values ​​S(j) associated with the three analyzed position signals (see column #05 in FIG8 ) can be assigned the same speed value f ROT =V=1 / (n diff *dt), as described above. Thereafter, step S#10 may be performed again for the next three consecutively received position signals. Alternatively, when step S#10 is repeated, the previous third position signal P3 will be used as the first position signal P1 (i.e., P1:=P3) to determine whether the speed is about to change.

[0177] If the analysis (step S#10) shows that the number of time slots between the first and second position signals is different from the number of time slots between the second and third position signals, the state parameter extractor 450 concludes (step S#20) that the monitored rotating housing 20 is in an acceleration phase. Acceleration can be positive, i.e., an increase in rotational speed, or negative, i.e., a decrease in rotational speed, also known as deceleration.

[0178] In the next step S#40, the state parameter extractor 450 operates to establish instantaneous speed values ​​during the acceleration phase and associates each measured data value S(j) with an instantaneous speed value Vp indicating the instantaneous speed at the time when the sensor signal (S) corresponding to the data value S(j) is detected. EA ) value is the rotational speed of the monitored mill housing.

[0179] According to one example, the state parameter extractor 450 operates to establish the instantaneous speed value by linear interpolation. According to another example, the state parameter extractor 450 operates to establish the instantaneous speed value by non-linear interpolation.

[0180] Figure 10 is shown for execution Figure 9 Flowchart of an example of the method of step 40. According to one example, it is assumed that the acceleration has a constant value for the duration between two mutually adjacent position indicators P (see column #02 in FIG8 ). Thus, when

[0181] ● a position indicator P is transmitted once per revolution, and

[0182] ●The gear ratio is 1 / 1, then

[0183] - the angular distance traveled by the rotating housing 20 between two mutually adjacent position indicators P is one (1) revolution, which can also be expressed as 360 degrees, and

[0184] - Duration is T=n diff *dt,

[0185] ■ Among them, n diff is the number of time slots of duration dt between two mutually adjacent position indicators P.

[0186] 8, the first position indicator P is detected in time slot il=#03, and the next position indicator P is detected in time slot i2=#45. Therefore, the duration is n diff1 =i2-i1=45-3=42 time slots.

[0187] Therefore, in step S#60 (see FIG. 8 ), Figure 10 ), the state parameter extractor 450 operates to establish a first number of time slots n between the first two consecutive position signals P1 and P2, that is, between the position signal P(i=3) and the position signal P(i=45). diff1 .

[0188] In step S#70, the state parameter extractor 450 operates to calculate the first rotation speed value VT1.

[0189] The first rotation speed value VTL can be calculated as:

[0190] VT1=1 / (n diff1 *dt),

[0191] Among them, VT1 is the speed expressed in revolutions per second,

[0192] n diff1 = the number of time slots between two consecutive position signals; and

[0193] dt is the duration of a time slot, expressed in seconds.

[0194] Since it is assumed that the acceleration has a constant value during the duration between two position indicators P adjacent to each other, the calculated first velocity value VT1 is assigned to the middle time slot between two consecutive position signals (step S#80).

[0195] Therefore, in this example, where in time slot i P1 = The first position indicator P1 is detected in #03 and in time slot i P2 = The next position indicator P2 is detected in #45; the first intermediate time slot is:

[0196] Time slot i P1-2 =i P1 +(i P2 -i P1 ) / 2=3+(45-3) / 2=3+21)=24.

[0197] Therefore, in step S#80, the first rotation speed value VT1 can be assigned to a time slot (for example, time slot i=24) which represents a time point earlier than the time point of detecting the second position signal edge P (i=45), see Figure 8.

[0198] The retroactive assignment of speed values ​​to time slots representing the points in time between two consecutive position signals advantageously results in a significant reduction in the inaccuracy of the speed values. Although prior art methods of obtaining instantaneous rotational speed values ​​of the roller mill housing 20 may be satisfactory for establishing constant speed values ​​at several mutually different rotational speeds, prior art solutions appear to be unsatisfactory when used to establish speed values ​​of the rotating roller mill housing 20 during an acceleration phase.

[0199] In contrast, the method according to the example disclosed in this document enables speed values ​​to be established with advantageously small inaccuracies even during acceleration phases.

[0200] In the subsequent step S#90, the state parameter extractor 450 operates to establish a second number of time slots n between the next two consecutive position signals. diff2 In the example of FIG8 , this is the number of slots n between slot 45 and slot 78. diff2 , that is, n diff2 =78-45=33.

[0201] In step S#100, the state parameter extractor 450 calculates the second rotation speed value VT2. The second rotation speed value VT2 can be calculated as:

[0202] VT2=Vp61=1 / (ndiff2 *dt),

[0203] Among them, n diff2 = the number of time slots between the next two consecutive position signals P2 and P3. Therefore, in the example of FIG8, n diff2 =33, that is, the number of time slots between time slot 45 and time slot 78.

[0204] Since it can be assumed that the acceleration has a constant value over the duration between two position indicators P adjacent to each other, the calculated second velocity value VT2 is assigned (step S#110) to the middle time slot between two consecutive position signals.

[0205] 8, the calculated second velocity value VT2 is assigned to time slot 61, because 45 + (78 - 45) / 2 = 61.5. Therefore, the velocity at time slot 61 is set to:

[0206] V(61):=VT2.

[0207] Thus, in this example, where one position indicator P is detected in time slot i2=#45 and the next position indicator P is detected in time slot i3=#78; the second intermediate time slot is the integer portion of:

[0208] i P2-3 =i P2 +(i P3 -i P2 ) / 2=45+(78-45) / 2=45+33 / 2=61.5

[0209] Therefore, time slot 61 is the second middle time slot i P2-3 .

[0210] Therefore, in step S#110, the second velocity value VT2 can be advantageously assigned to a time slot (e.g., time slot i=61) that represents a time point earlier than the time point at which the third position signal edge P (i=78) is detected, see Figure 8. This feature enables slightly delayed real-time monitoring of the rotational speed while achieving higher accuracy in detecting the speed.

[0211] In the next step S#120, the first acceleration value of the relevant time period is calculated. The first acceleration value can be calculated as:

[0212] a12=(VT2-VT1) / ((i VT2 -i VT1 )*dt)

[0213] In the example of FIG. 8 , the second speed value VT2 is assigned to the time slot 61, so i VT2=61, the first speed value VT1 is assigned to time slot 24, so i VT1 =24.

[0214] Therefore, in the example of FIG8 , since dt=1 / fs, the acceleration value can be set to:

[0215] a12=fs*(VT2-VT1) / (i VT2 -i VT1 )

[0216] Used for the time period between time slot 24 and time slot 60.

[0217] In the next step S#130, the state parameter extractor 450 operates to associate the established first acceleration value a11 with the time slots in which the established acceleration value a12 is valid. This can be all time slots between the time slot of the first velocity value VT1 and the time slot of the second velocity value VT2. Thus, the established first acceleration value a12 can be associated with every time slot of the duration between the time slot of the first velocity value VT1 and the time slot of the second velocity value VT2. In the example of FIG8 , this is time slots 25 to 60. This is shown in column #07 of FIG8 .

[0218] In the next step S#140, the state parameter extractor 450 operates to establish a velocity value for the measurement value s(j) associated with the duration for which the established acceleration value is valid. Thus, a velocity value is established for each time slot that

[0219] is associated with the measurement s(j), and

[0220] Associated with the established first acceleration value a12.

[0221] During linear acceleration, that is, when the acceleration a is constant, the velocity at any given point in time is given by:

[0222] V(i)=V(i-1)+a*dt,

[0223] in,

[0224] V(i) is the instantaneous velocity at time slot i,

[0225] V(i-1) is the instantaneous velocity at the time slot immediately before time slot i,

[0226] a is the acceleration,

[0227] dt is the duration of the time slot.

[0228] According to one example, the speed can be calculated continuously in this manner for each time slot from time slot 25 to time slot 60, as shown in column #08 in FIG8 . Thus, in this manner, it is possible to establish instantaneous speed values ​​Vp associated with the detected measurement values ​​Se(25), Se(26), Se(27) ... Se(59) and Se(60), which are associated with the acceleration value a12 (see column #08 in FIG8 together with column #03 and column #07 for time slots 25 to 60).

[0229] Thus, in this way it is possible to establish instantaneous velocity values ​​S(j) [see column #05] associated with the detected measurement values ​​S(3), S(4), S(5) and S(6), which are associated with the acceleration value a12.

[0230] According to another example, the instantaneous speed of time slot 30 associated with the first measurement value s(j)=S(3) can be calculated as:

[0231] V(i=30)=Vp30=VT1+a*(30-24)*dt=Vp24+a*6*dt

[0232] The instantaneous speed of time slot 40 associated with the first measurement value s(j)=S(4) can be calculated as:

[0233] V(i=40)=Vp40=VT1+a*(40-24)*dt=Vp40+a*16*dt

[0234] or calculated as:

[0235] V(i=40)=Vp40=V(30)+(40-30)*dt=Vp30+a*10*dt

[0236] The instantaneous speed of the time slot 50 associated with the first measurement value s(j)=S(5) can then be calculated as:

[0237] V(i=50)=Vp50=V(40)+(50-40)*dt=Vp40+a*10*dt

[0238] And the instantaneous speed of time slot 60 associated with the first measurement value s(j)=S(6) can then be calculated as:

[0239] V(i=60)=Vp50+a*10*dt

[0240] As described above, when the measurement sample values ​​S(i) [see column #03 in FIG. 8 ] associated with the established acceleration values ​​have been associated with the instantaneous velocity values, a data array comprising a time series of measurement sample values ​​S(i) can be delivered at the output of the state parameter extractor 450, each value being associated with the velocity value V(i), f ROT (i) Associated.

[0241] Alternatively, if it is desired to reduce the sampling rate, this can be done as follows: as described above, when the measurement sample values ​​S(j) [see column #05 in FIG8 ] associated with the established acceleration values ​​have been associated with the instantaneous velocity values, a data array comprising a time series of measurement sample values ​​S(j) can be delivered at the output of the state parameter extractor 450, each value being associated with a velocity value V(j), f ROT (j) Associated.

[0242] refer to Figure 11 , describes another example of a method. According to this example, the state parameter extractor 450 operates to record (see Figure 11 The position signal (E P ) such that between at least some of the recorded position signal values ​​(P(i)), for example, between a first position signal value P1(i) and a second position signal value P2(i), there is a first time relationship n diff1 According to one example, a second position signal value P2(i) is received and recorded in a time slot (i) that arrives ndiffl time slots after the reception of the first position signal value P1(i) (see Figure 11 Then, the third position signal value P3(i) is received and recorded (see step S#160 in FIG. Figure 11 Step S#170 in ) is in time slot (i), which arrives ndiff2 time slots after receiving the second position signal value P2(i).

[0243] like Figure 11 As shown in step S#180 in FIG. 4 , the state parameter extractor 450 may be operable to calculate the relationship value:

[0244] a12=ndiff1 / ndiff2.

[0245] If the relationship value a12 is equal to unity or approximately equal to unity, the state parameter extractor 450 operates to determine that the velocity is constant and may continue to calculate the velocity according to the constant velocity phase method.

[0246] If the relation value a12 is greater than one, then the relation value indicates a percentage speed increase.

[0247] If the relation value a12 is less than one, then the relation value indicates a percentage speed reduction.

[0248] The relationship value a12 can be used to calculate the speed V2 at the end of the time series based on the speed V1 at the beginning of the time series, for example, as

[0249] V2=a12*V1.

[0250] Figure 12 is shown for execution Figure 9 Flowchart of an example of a method of step S#40. According to one example, it is assumed that the acceleration has a constant value during the duration between two position indicators P adjacent to each other (see column #02 in FIG8 ). Therefore, when

[0251] ●Transmit the position indicator P once per revolution, and

[0252] ●The gear ratio is 1 / 1, then

[0253] - the angular distance between two adjacent position indicators P is one revolution, which can also be expressed as 360 degrees, and

[0254] - duration is T = n*dt,

[0255] ■ where n is the number of time slots of duration dt between the first two position indicators P1 and P2 that are adjacent to each other.

[0256] In step S#200, the first rotation speed value VT1 can be calculated as:

[0257] VT1=1 / (n diff1 *dt),

[0258] Among them, VT1 is the speed expressed in revolutions per second,

[0259] ndiff1 = number of time slots between two consecutive position signals; and

[0260] dt is the duration of the time slot, expressed in seconds. The value of dt can be, for example, the initial sampling frequency f S The reciprocal of .

[0261] Since it is assumed that the acceleration has a constant value during the duration between two position indicators P adjacent to each other, the calculated first velocity value VT1 is assigned to the intermediate position between two consecutive position signals P(i) and P(i+ndiff1). First mid-slot .

[0262] In step S#210, the second velocity value VT2 can be calculated as:

[0263] VT2=1 / (ndiff2*dt),

[0264] Among them, VT2 is the speed expressed in revolutions per second,

[0265] ndiff2 = number of time slots between two consecutive position signals; and

[0266] dt is the duration of the time slot, expressed in seconds. The value of dt can be, for example, the initial sampling frequency f S The reciprocal of .

[0267] Since it is assumed that the acceleration has a constant value during the duration between two position indicators P adjacent to each other, the calculated second velocity value VT2 is assigned to the intermediate position between two consecutive position signals P(i+ndiff1) and P(i+ndiff1+ndiff2). Second mid-slot .

[0268] After that, the speed difference V Delta can be calculated as:

[0269] V Delta =VT2-VT1.

[0270] The speed difference V Delta The value can be divided by the number of time slots between the second intermediate time slot and the first intermediate time slot. The resulting value indicates the velocity difference dV between adjacent time slots. Of course, as mentioned above, this assumes constant acceleration.

[0271] The instantaneous speed value associated with the selected time slot may then be calculated based on the first rotational speed value VT1 and a value indicative of the speed difference between adjacent time slots.

[0272] As described above, when the measurement sample values ​​S(i) associated with the time slot between the first intermediate time slot and the second intermediate time slot have been associated with instantaneous speed values, a data array comprising a time sequence of measurement sample values ​​S(i) is delivered at the output of the state parameter extractor 450, each value being associated with a speed value V(i). The instantaneous speed value V(i) may also be referred to as f ROT (i).

[0273] In summary, according to some examples, the first instantaneous speed value VT1 may be established according to the following factors:

[0274] The angular distance δ-FI between the first position signal P1 and the second position signal P2 p1-p2 , and depends on:

[0275] The corresponding duration δ-T p1-p2 =t P2 -t P1 .

[0276] Thereafter, the second instantaneous velocity value VT2 can be established based on the following factors:

[0277] The angular distance δ-FI between the second position signal P2 and the third position signal P3 P2-P3 , and depends on:

[0278] The corresponding duration δ-T p2-p3 =t P2 -t P1 .

[0279] Thereafter, the instantaneous speed value of the rotating housing 20 may be established by interpolation between the first instantaneous speed value VT1 and the second instantaneous speed value VT2 .

[0280] In other words, according to the example, it is possible to p1-p2 ,δ-FI P2-P3 Two instantaneous speed values ​​VT1 and VT2 are established based on the corresponding time durations between the three consecutive position signals, and thereafter, the instantaneous speed value of the rotating housing 20 can be established by interpolation between the first instantaneous speed value VT1 and the second instantaneous speed value VT2.

[0281] Figure 13 is a diagram showing a series of temporally consecutive position signals P1, P2, P3, ..., each position signal P indicating one complete revolution of the monitored housing 20. Thus, the time values ​​in seconds increase to the right along the horizontal axis.

[0282] The vertical axis indicates rotational speed, graded in revolutions per minute (RPM).

[0283] refer to Figure 13 , shows the effect of the method according to an example. The first instantaneous speed value V(t1)=VT1 can be established according to the following formula:

[0284] The angular distance δ-FI between the first position signal p1 and the second position signal P2 p1-p2 , and depends on:

[0285] The corresponding duration δ-T p1-p2 =t P2 -t P1 By dividing the angular distance δ-FI p1-p2 Divide by the corresponding duration (t P2 -t P1 ) represents the speed V(t1) of the rotating shell 20 at the first intermediate time point t1, also referred to as mtp (intermediate time point), as Figure 13 shown.

[0286] Thereafter, the second instantaneous velocity value V(t2)=VT2 can be established according to the following formula:

[0287] The angular distance δ-FI between the second position signal P2 and the third position signal P3 depends on:

[0288] The corresponding duration δ-T2-3=t P3 -t P2 .

[0289] like Figure 13 As shown, by dividing the angular distance δ-FI by the corresponding duration (t P3 -t P2 ) represents the speed V(t2) of the rotating housing 20 at the second intermediate time point t2 (second mtp).

[0290] Thereafter, the instantaneous speed value of the time value between the first intermediate time point and the second intermediate time point can be established by interpolating between the first instantaneous speed value VT1 and the second instantaneous speed value VT2, as shown in the curve f ROTint shown.

[0291] Mathematically, this can be expressed as the following formula:

[0292] V(t12)=V(t1)+a*(t12–t1).

[0293] Therefore, if the velocity of the housing 20 can be detected at two time points (t1 and t2) and the acceleration a is constant, the instantaneous velocity at any time point can be calculated. Specifically, the housing velocity V(t12) at time T12 (a time point after t1 and before t2) can be calculated using the following formula:

[0294] V(t12)=V(t1)+a*(t12–t1)

[0295] in,

[0296] a is the acceleration, and

[0297] t1 is the first intermediate time point t1 (see Figure 13 ).

[0298] The establishment of the speed value and the reference value as described above can be achieved by executing the corresponding method steps. Figure 20 、 Figure 21 22, and this may be accomplished by a computer program 94 stored in the memory 60, as described above. The computer program may be executed by the DSP 50. Alternatively, the computer program may be executed by a field programmable gate array circuit (FPGA).

[0299] When the processor 350 executes the corresponding program codes 380, 394, 410, the speed value f ROT The establishment of (i) can be performed by the analysis device 150, as described above in conjunction with Figure 4 As discussed. The data processor 350 may include a central processing unit 350 for controlling the operation of the analysis device 14. Alternatively, the processor 50 may include a digital signal processor (DSP) 350. According to another example, the processor 350 includes a field programmable gate array circuit (FPGA). The operation of the field programmable gate array circuit (FPGA) can be controlled by the central processing unit 350, which may include a digital signal processor (DSP) 350.

[0300] Identification of data related to the toe of a charge in a roller mill

[0301] As described above, the roller mill housing 20 has an inner housing surface 22 facing the chamber 25, and the inner housing surface 22 includes a plurality of protrusions 310, also known as lifters, which can be configured to engage and lift the material 30 when the housing rotates about the axis 60 (see, for example, FIG. Figure 2 ). The number of protrusions 310 provided on the inner housing surface 22 facing the chamber 25 is herein denoted by the variable L. Figure 2 The case where there are twelve protrusions 310 is shown, i.e., L=12, but the number L of protrusions 310 may be higher or lower. According to some embodiments, the number L of protrusions 310 may be at least one, i.e., the number L of protrusions 310 may be L=1. According to some embodiments, the number L of protrusions 310 may be any number greater than L=1. According to some embodiments, the number L of protrusions 310 may be any value within the range of L=2 to L=60. According to some embodiments, the number L of protrusions 310 may be any value within the range of L=2 to L=35.

[0302] The number L of protrusions 310 is an important factor in analyzing the vibrations caused by the rotation of the mill housing 20. The inventors have recognized that the interaction of the protrusions 310 with the toe of the charge forces the material of the charge to accelerate in the direction of motion of the protrusions 310, thereby causing the mechanical vibration V IMP The inventors have also recognized that this mechanical vibration V caused by the interaction of the protrusion 310 with the toe of the charge IMP will be repeated, i.e. there is a repetition frequency f R . refer to Figure 2 , it should be noted that it shows the rotating mill housing 20 when the protrusion 310C impacts the toe 205 of the material charge 30. The impact of the protrusion 310C on the bulk of the material in the toe 205 causes the bulk of the material in the toe to move in the direction A of motion of the protrusion 310C. ACC The acceleration causes a force F on the leading edge surface of the protrusion 310C.IMP By the way, this impact force F IMP can be estimated to be an order of magnitude:

[0303] F IMP =m 205 *a 205

[0304] in,

[0305] m 205 is the weight of the accelerating toe,

[0306] a 205 is the magnitude of the acceleration at the toe.

[0307] Therefore, the measurement signal S MD (See, for example, Figure 5 ) may include at least one vibration signal characteristic S of the vibration movement of the roller mill housing 20 depending on the rotational movement FIMP , wherein the vibration signal characteristic S FIMP With repetition frequency f R The repetition frequency depends on the rotational speed f of the rotating roller mill housing 20. ROT .

[0308] In addition, the vibration signal characteristic S FIMP The magnitude of the peak amplitude seems to depend on the impact force F IMP size.

[0309] Therefore, the inventors concluded that the vibration signal feature S FIMP The energy or magnitude measurement appears to indicate the impact force F IMP size.

[0310] The vibration signal characteristic S of the vibration movement of the mill housing 20 depending on the rotational movement FIMP The presence of can thus provide an indication of the toe 205 of the charge of the monitored roller mill housing 20. In fact, the vibration signal characteristic S of the vibration motion of the mill housing 20 depends on the rotational motion FIMP An indication of the position of the toe 205 of the charge of the roller mill housing 20 being monitored may be provided, the position being indicated relative to a reference position value.

[0311] The inventors concluded that the mechanical vibration V caused by the interaction of the protrusion 310 with the charge toe IMP The repetition frequency f R Depending on the number L of protrusions 310 provided on the inner housing surface 22 and the rotational speed f of the housing 20 ROT .

[0312] When the roller mill housing 20 is monitored ConstantWhen the speed of rotation is R The discussion can be based on repetitions per unit time or on repetitions per revolution of the monitored housing, without distinguishing between the two. variable Things are more complicated, as discussed elsewhere in this disclosure, for example, in conjunction with Figure 20 、 Figure 21 、 Figure 22A 、 Figure 22B and Figure 22C In fact, it seems that even very small changes in the rotational speed of the mill housing can have a significant adverse effect on the quality of the detected signal, as far as the ambiguity of the detected vibration signal is concerned. ROT Very precise detection is crucial.

[0313] Furthermore, the inventors have recognized that not only the mechanical vibration V IMP The amplitude of the mechanical vibration V IMP The occurrence time of the measurement signal S can indicate the data related to the charging toe 205 in the roller mill. MD (See, for example, Figure 5 ) may include at least one vibration signal amplitude component S of the vibration motion of the roller mill housing 20 depending on the rotational motion FIMP ;

[0314] Among them, the vibration signal amplitude component S FIMP With repetition frequency f R , the repetition frequency:

[0315] The rotational speed f of the roller mill housing 20 depends on the rotational movement ROT , and also

[0316] depends on the number L of protrusions 310 provided on the inner housing surface 22 of the mill housing 20; and

[0317] Among them, there is a temporal relationship between:

[0318] Repetitive vibration signal amplitude component S FIMP The emergence of

[0319] With a second repetition frequency f P The frequency of the position signal P(i) depends on the rotational speed f of the rotating roller mill housing 20. ROT .

[0320] Regarding the constant speed, the inventors concluded that if the speed f ROT is constant, then the digital measurement signal S comprising a time series of vibration sample values ​​S(i) MDWith repetition frequency f R , the repetition frequency depends on the number L of protrusions 310 provided on the inner housing surface 22 .

[0321] The state parameter extractor 450 may optionally include a device coupled to receive the digital measurement signal S MD or depending on the digital measurement signal S MD (See Figure 15A and / or Figure 15B ) is a fast Fourier transform (FFT) of the signal. Regarding the analysis of the roller mill with the rotating shell 20, the analysis is performed at a frequency higher than the rotation frequency f of the rotating shell 20. ROT The signal frequency may be of interest. In this case, the rotational frequency f of the housing 20 is ROT can be called "1st order". If the signal of interest occurs ten times per revolution of the housing, the frequency can be called 10th order, ie the repetition frequency f R (measured in Hz) divided by the speed f ROT (measured in revolutions per second rps) equal to 10Hz / rps, that is, Oi=f R / f ROT =10th order.

[0322] The maximum order is called O MAX , and the total number of frequency bins in the FFT is used as B n , the inventors concluded that, according to one example, the following formula applies:

[0323] Oi*B n =N R *O MAX .

[0324] In contrast, N R =Oi*B n / O MAX ,in,

[0325] O MAX is the maximum order; and

[0326] B n is the number of frequency bins in the spectrum produced by the FFT, and

[0327] Oi is the number L of protrusions 310 in the roller mill housing 20 that are monitored.

[0328] The above variables should be set to MAX 、B n and Oi, so that the variable N R In conjunction with the above example, it should be noted that the FFT analyzer is configured to receive a reference signal, ie, a position marker signal value PS or E, once per rotation of the rotating housing 20. PIf combined with Figure 2 As described above, a position marking device 180 may be provided so that when the housing 20 rotates around the rotation axis 60, the position mark 180 passes the position sensor 170 once per rotation of the housing 20, thereby causing the position sensor 170 to generate a rotation mark signal value PS, E P .

[0329] Incidentally, referring to the above example of FFT analyzer settings, the integer N obtained R The number of revolutions of the monitored housing 20 can be indicated, during which the digital signal S is analyzed. MD According to one example, the above variable O MAX 、B n and Oi can be set via the human-machine interface HCI 210, 210S (see, for example, FIG. 1 and / or Figure 5 and / or Figure 15A and / or Figure 15B ).

[0330] Consider when the digital measurement signal S MD When the FFT analyzer is used: In this case, when the FFT analyzer is set for 10 protrusions, that is, L=10, and B n = 160 frequency bins, and user focus analysis up to O MAX =100th order frequency, then N R The value becomes N R =Oi*B n / O MAX =10*160 / 100=16.

[0331] Therefore, when B is needed n = 160 frequency bins, 16 shell rotations (N R =16) during measurement, the number of protrusions is L=10; and the user focuses on analysis up to O MAX =100th order frequency. Combined with the settings of the FFT analyzer, the order value O MAX Can indicate digital measurement signal S MD The highest frequency to be analyzed in .

[0332] According to some embodiments, when the FFT analyzer is configured to receive the reference signal, i.e., the position marker signal value PS, once per rotation of the rotating housing 20 , the settings of the FFT analyzer should meet the following criteria:

[0333] The integer value Oi is set equal to L, the number of protrusions 310 in the housing 20, and

[0334] Select the variable O MAX and B n, so that the mathematical expression Oi*B n / O MAX In other words: when the integer value Oi is set equal to L, then the variable O can be set MAX and B n should be set to an integer value so that the variable N R is a positive integer,

[0335] Among them, N R =Oi*B n / O MAX .

[0336] According to an example, a value B can be selected from a set of values ​​by n To set the number of bins B n Frequency resolution B n The optional value group can include:

[0337] B n =200

[0338] B n =400

[0339] B n =800

[0340] B n =1600

[0341] B n =3200

[0342] An example of a constant velocity phase

[0343] If combined Figure 9 As described in step S#30 in FIG. 1 , the state parameter extractor 450 can identify the constant speed stage, ie, the constant speed f of the housing 20. ROT status.

[0344] Figure 14A and Figure 14B Another example of a cross-sectional view of the middle portion 98 of the rotating mill housing 20 during operation is shown. This view may be taken, for example, along Figure 1A According to the line AA. Figure 14A In the example of FIG. 5 , the roller mill housing 20 has six protrusions 310 configured to engage the material charge 30 when the housing is rotated about the axis 60 , ie, the number L=6.

[0345] The inner diameter of the housing 20 may be, for example, 600 cm, and the rotational speed may be constant, for example, 13.6 revolutions per minute. For the purposes of this example, the sampling frequency is such that there are n=7680 samples per revolution, and the rotational speed of the housing 20 is f ROT .

[0346] As described above, the housing 20 can rotate about the rotation axis 60, so the position sensor 170 can generate a position signal Ep for indicating the instantaneous rotation position of the housing 20. The position mark 180 can be provided on the outer surface of the housing 20, so that when the housing 20 rotates about the rotation axis 60, the position mark 180 passes the position sensor 170 once for each rotation of the housing, so that the position signal Ep presents a position mark signal value P. S Each such position marker signal value P S Indicates the rest position, i.e. the position of the stationary stator.

[0347] Figure 14A The rotational position of the rotating housing 20 is shown, wherein the position mark 180 is located at the same rotational position as the static position sensor 170, and the protrusion 310A has passed through the toe 205. The protrusion 310A is followed by an adjacent protrusion 310B.

[0348] Figure 14B Another rotation position of the rotating housing 20 is shown. Figure 14A The position shown is a little later. Figure 14B In the example, the adjacent protrusion 310B is located at the position where it strikes the toe 205. The impact time causes the vibration V IMP , which results in a signal characteristic event in the vibration signal. Therefore, Figure 14B The rotational position shown is the event position 205E of the rotating housing 20. The event position 205E is the rotational position of the rotating housing 20 when the protrusion hits the toe 205. Therefore, the event position 205E indicates the toe position 205. Figure 14B As shown, the position 205E of the toe 205 can be expressed as a percentage of the distance between two adjacent static positions P3 and P4.

[0349] When there is a position mark signal value P per revolution S And the speed f ROT When constant or approximately constant, there will be a constant or approximately constant number of vibration sample values ​​S(i) for each revolution of the mill housing 20. For the purposes of this example, the position signal P(0) indicates vibration sample i=0, as shown in Table 2 (see below). For the purposes of this example, the position signal P(0) relative to the housing 20 may not be important as long as the repetition frequency f P The rotational speed f of the roller housing 20 depends on the rotational movement ROT Therefore, if the housing 20 rotates once, the position signal E P There is a pulse Ps, and the digital position signal will have a position signal value P(i)=1 every time it rotates one circle, and the rest of the position signal values ​​are zero.

[0350] Table 2

[0351]

[0352]

[0353] Therefore, at a constant speed f ROT In this case, there may be n time slots per revolution, as shown in Table 2, where n can be a positive integer. In the example of Table 2, n=7680.

[0354] There is a position signal P per revolution S , we know that the position signal will be repeated every n time slots, because the speed f ROT Therefore, multiple virtual position signals P can be generated by calculation. C In one example, consider generating a virtual position signal P C A virtual position signal P is provided for each protrusion 310. C Can be used to establish temporal relationships between:

[0355] Repetitive vibration signal amplitude component S FIMP The emergence of

[0356] The presence of a position signal P(i) with a second repetition frequency f P The repetition frequency depends on the rotational speed f of the rotating roller mill housing 20. ROT .

[0357] There are L equally spaced protrusions 310 in the mill housing, and each revolution has a position signal P S and constant speed f ROT , a virtual position signal P can be generated for each protrusion C , so the position signal P S 、P C The total number of uniform distribution. Each such position mark signal value P S and P C Indicates the rest position, i.e. the position of the stationary stator, e.g. Figure 14A and Figure 14B "P S ” and “P C ” as shown.

[0358] Therefore, as shown in Table 3, when n time slots are provided per revolution, the position signal P S or P C Will appear at every n / L sample value position. In Table 3, n = 7680, L = 6, so a position signal P is provided every 1280 samples. C , the calculated position signal is represented as 1C.

[0359] As shown in the example of FIG. 14 , the position marker signal value P S and P C L fixed positions P1 , P2 , P3 , P4 , P5 and PL are indicated, where L=6 since there are 6 protrusions 310 in the housing 20 shown.

[0360] It can be assumed that the position of the toe 205 of the mill is substantially constant during a single rotation of the mill housing 20. In other words, the position of the toe 205 is substantially fixed.

[0361] Since the amplitude component S of the vibration signal FIMP , S P is generated by the interaction of the protrusion and the toe of the charge (see Figure 14B ), so each protrusion 310 will have a vibration signal amplitude component S FIMP , the frequency of Sp is repeated. Therefore, it can be assumed that:

[0362] Repetitive vibration signal amplitude component S FIMP 、S p The emergence of

[0363] The temporal relationship between the occurrences of the position signals P, PC is approximately constant for each of the L data blocks, in this example L=6.

[0364] Table 3 shows the principle of the time course of the position signal value P(i), and the calculated position signal value P(i) is represented as "1C".

[0365] Table 3

[0366]

[0367]

[0368] Table 4

[0369]

[0370]

[0371] Table 5

[0372]

[0373]

[0374]

[0375] As described above, the housing 20 can rotate about the rotation axis 60, so the position sensor 170 installed in a fixed manner can generate a position signal Ep having a series of housing position signal values ​​PS , used to indicate the instantaneous rotation position of the housing 20. Figure 23 As shown, the position mark 180 can be provided on the outer surface of the housing 20 so that when the housing 20 rotates about the rotation axis 60, the position mark 180 passes through the position sensor 170 during one rotation of the housing 20, thereby causing the position sensor 170 to generate a rotation mark signal value P S .

[0376] As described above, the position sensor 170 can generate a position signal Ep having a series of housing position signal values ​​P S , used to indicate the instantaneous rotation position of the housing 20 when the housing 20 rotates. Referring to Tables 2-4 in this document, such a mark signal value P S Shown as "1" in column #2 of Table 2-4.

[0377] When the rotating housing is provided with a position marking device 180, a marking signal value P is provided once per revolution. S In Table 2-4, the signal value P is marked S In column #2 it is shown as "1". There are L equally spaced protrusions 310 in the mill housing, and one position signal P per revolution and a constant speed f ROT , it is possible to generate a virtual position signal P for each protrusion C , so that the position signals P, P C The total number of is evenly distributed as described above. Therefore, as shown in Table 3, when n time slots are provided per revolution, the position signal P or P C Will appear at every n / L sample value position. In Table 3, n = 7680, and L = 6, so a position signal P is provided every 1280 samples. C , the calculated position signal indication is 1C.

[0378] It is believed that when the marking signal value P is provided once per revolution S (Indicated as "1" in column #2 of Table 2-4) and generate virtual position signal values ​​P in a uniformly distributed manner C The mutually equidistant positions of the protrusions 310 are important so that when n time slots are provided per revolution in the sequence of housing position signal values ​​for indicating the instantaneous rotational position of the housing 20, the position signal P or P C It will appear at every n / L sample value position, as shown in Table 3. In Table 3, the actual detected rotation mark signal value P S is reflected as "1" (see column #2 in Table 3, time slot "0" and time slot "7680"), and the virtual position signal value P C Reflected as "1C" (see column #2 in Table 3, time slot "0" and time slot "7680").

[0379] It is believed that this is important for some embodiments of the present disclosure because the position mark 180 results in the generation of a position reference signal value, and the protrusion 310, when engaging material in the rotating mill charge, results in the generation of a signal event, e.g., an amplitude peak in the vibration signal (see, e.g., reference S in FIG. 1 and FIG. 15 ). EA 、S MD , Se(i), S(j), S(q)). In addition, the duration between the occurrence of the position reference signal value and the occurrence of the signal event in the vibration signal, as discussed elsewhere in this disclosure, caused by the protrusion 310 engaging material in the rotating mill housing charge, can be indicative of the internal state of the mill in operation.

[0380] Table 4 is a diagram of the first block (i.e., block I) with n / L=7680 / 6=1280 consecutive time slots. It should be understood that if there is a constant speed phase during the duration of a complete rotation of the housing 20 (see Figure 9 ), then each of blocks I to VI (see Table 3) will have the same appearance as block I shown in Table 4.

[0381] According to an embodiment of the present disclosure, referring to column #03 in Table 4, the vibration sample value S(i) is analyzed to detect the vibration signal feature S FIMP . Vibration signal characteristics S FIMP It can be expressed as the peak amplitude sample value S P According to one example, referring to column #03 in Table 4, the vibration sample value S(i) is analyzed by a peak detector to detect the peak sample value S P Referring to Table 5, the peak analysis results in the detection of the highest vibration sample amplitude value S(i). In the example shown, the vibration sample amplitude value S(i=760) is detected as the one holding the highest peak value S P .

[0382] Peak S has been detected P In time slot 760, the repetitive vibration signal amplitude component S P The temporal relationship between the occurrence of position signal P(i) and the occurrence of position signal P(i). In Table 5, the time slots in which position signal P(i) is transmitted are represented as 0% and 100%, respectively, and all time slots in between are labeled with their corresponding positions, as shown in column #02 in Table 5. As shown in the example in column #02 of Table 5, the time position of time slot number i = 760 is 59% of the time distance between time slot i = 0 and time slot i = 1280. In other words, 760 / 1280 = 0.59 = 59%.

[0383] Therefore, the inventors concluded that:

[0384] Repetitive vibration signal amplitude component SFIMP The temporal relationship between the occurrence of and the occurrence of the position signal P(i) may be used as an indication of the relative physical position of the toe 205 of the charge between two consecutive protrusions 310 in the rotating housing 20 .

[0385] Therefore, the distance between two adjacent leading edges expressed as a percentage can be obtained by the following formula (see Figure 2 312C and 312D in combination with the position of the toe 205 of Table 5):

[0386] For the first reference signal from sample number N0=0 to sample number N B = the total number of samples where the second reference signal appears in 1280 (N B -N0=N B -0=N B =1280) for counting, and

[0387] For the period from the first reference signal at N0=0 to the first reference signal at sample number N P The peak amplitude value S at P The number of samples that appear (N P -N0=N P -0=N P ) to count, and

[0388] Based on the other number N P and the total number N B Generate the first time relationship (R T (r); T D ; FI(r)). This can be summarized as:

[0389] R T (r)=R T (760)=(N P -N0) / (N B -N0)=(760-0) / (1280-0)=0.59=59%

[0390] Therefore, the relative toe position can be generated by:

[0391] The total number of samples (N) from the first reference signal to the second reference signal B ) to count, and

[0392] From the first reference signal to the sample number N P The peak amplitude value S at P The number of samples that appear (N P ) to count, and

[0393] Based on the sample number N PWith the total number of samples (ie N B ) generates the first time relationship (R T (r); T D ; FI(r)).

[0394] Referring to FIG14 , it should be noted that at the time point shown, the position marker 180 is depicted as just passing the position of the position sensor 170. Therefore, the time point shown may be the time point indicated by the time slot 1280, i.e., when the position signal P (i=1280) is generated. Since the housing rotates in a clockwise direction, the most recent sample peak S P is generated by the impact of the protrusion 310A and the toe 205 (see Figure 14A and Figure 14B and Table 5). Therefore, it is detected as maintaining the highest peak S P The vibration sample amplitude value S (i = 760) appears before the position signal P (i = 1280) appears at time T SP =dt*(1280-760).

[0395] Because S = v*t, where S = distance, v = constant velocity, and t is time, the time relationship can be directly converted to distance. Therefore, column #02 of Table 5 can be considered to indicate the physical position of toe 205 at 59% of the distance between protrusion 310A and protrusion 310B (see Figure 14 in conjunction with column #02 of Table 5).

[0396] According to another example, referring to Table 6, the repetitive vibration signal amplitude component S P The temporal relationship between the occurrence of and the occurrence of the position signal P(i) can be considered as a phase deviation expressed in degrees.

[0397] Table 6

[0398]

[0399]

[0400] In fact, by using the position signal as a digital measurement signal S MD, S(i), S(j), and adjusting the settings of a fast Fourier converter in some manner, the fast Fourier converter can be used to extract the amplitude peaks as well as the phase values, as described below. Therefore, when the total distance between the protrusion 310A and the protrusion 310B is considered to be 360 ​​degrees, column #02 of Table 6 can be considered to indicate the physical position of the toe 205 at a position of 213.75 degrees of the distance between the protrusion 310A and the protrusion 310B (see Figure 14 in conjunction with column #02 of Table 6). When expressed as a portion of the distance between two adjacent protrusions 310, the physical position of the toe 205 can be referred to as the relative position of the toe 205. In other words, the present disclosure provides a way to identify the relative toe position of the toe 205 of the charge in the roller mill. Therefore, the present disclosure provides a way to generate information indicating the position of the toe 205 when expressed as a portion of the distance between two adjacent protrusions 310 in the rotating shell 20. Reference Figure 15A and / or Figure 15B and Figure 16 , the relative toe position can be expressed as a phase angle FI(r), as shown below in conjunction with Figures 15 and Figure 16 According to embodiments of the disclosure, relative toe position may be expressed as a percentage (see column #02 of Table 5 above).

[0401] Furthermore, according to embodiments of the disclosure, the relative toe position can be expressed as a duration, or a fraction of a duration. As described above, in conjunction with Table 5, since S = v*t, where S = distance, v = velocity of the protrusion, and t is time, the time relationship can be directly converted to distance. In this case, it should be noted that the velocity v of the protrusion depends on the angular velocity f of the housing 20. ROT and the radius R of the housing 20 MIC (See Figure 14).

[0402] Figure 15A is a block diagram illustrating an example of the state parameter extractor 450 . Figure 15A The state parameter extractor 450 includes receiving a quantity vibration signal S MD , S(i) and the digital position signal (Pi). The housing speed detector 500 may also be referred to as a housing speed value generator 500. The housing speed detector 500 may generate a housing speed value based on the received digital vibration signal S MD , S(i) and the digital position signal (Pi) generate three signals S(j), P(j) and f ROT (j). This can be done, for example, as described above with respect to Figures 7 to 13 In this respect, it should be noted that the three signals S(j), P(j) and f(j) can be transmitted simultaneously. ROT(j), i.e., these signals are all associated with the same time slot j. In other words, the three signals S(j), P(j), and f(j) can be provided in a synchronized manner. ROT (j). Providing signals such as S(j), P(j) and ROT(j) in a synchronized manner advantageously provides accurate information about the temporal relationship between the signal values ​​of the respective signals. Thus, for example, the velocity value f delivered by the shell velocity value generator 500 ROT (j) indicates the instantaneous rotation speed of the housing 20 when the amplitude value S(j) is detected.

[0403] It should be noted that the signals S(j) and P(j) transmitted by the shell velocity value generator 500 are delayed relative to the signals S(i) and (Pi) received by the shell velocity value generator 500. It should also be noted that the signals S(j) and P(j) are equally delayed relative to the signals S(i) and (Pi), thereby preserving the temporal relationship between the two. In other words, the signals S(j) and P(j) are synchronously delayed.

[0404] The housing speed detector 500 may deliver a signal indicating whether the rotational speed remains constant for a sufficiently long time, in which case the signals S(j) and P(j) may be delivered to the fast Fourier converter 510 .

[0405] As described above, the variables Y, Z, and L should be set so that the variable X is a positive integer. According to one example, the variables Y, Z, and L can be set by the human-machine interface HCI 210, 210S (see, for example, FIG. 1 and / or Figure 5 and / or Figure 15A and / or Figure 15B As described above, the resulting integer X can indicate the number of revolutions of the monitored roller mill housing 20 during which the digital signals S(j) and P(j) are analyzed by the FFT 510. Thus, based on the settings of the variables Y, Z, and L, the FFT 510 can generate a value X that indicates the analysis duration of the measurement session, and after the measurement session, the FFT 510 delivers a set of state values ​​Sp(r) and FI(r).

[0406] The concept "r" in the state values ​​Sp(r) and FI(r) represents a point in time. It should be noted that there may be a time delay between receiving the first pair of input signals S(j), P(j) at the input of FFT 510 and transmitting the pair of state values ​​Sp(r) and FI(r) from FFT 510. The pair of state values ​​Sp(r) and FI(r) can be based on a time sequence of the input signal pair S(j), P(j). The duration of the time sequence of the input signal pair S(j), P(j) should include at least two consecutive position signal values ​​P(j) = 1 and the corresponding input signal pair.

[0407] As described below, the state values ​​Sp(r) and FI(r) can also be referred to as C L and ФL. As above about Figure 2 As described above, the vibration signal S EA 、S MD , S(j), S(r) will present a signal characteristic S indicating the impact of the protrusion with the toe 205 FIMP , and when there are L protrusions 310 in the housing 20 (see FIG. 1 in conjunction with FIG. 14 and Figure 15A and / or Figure 15B ), each time the housing 20 rotates one circle, the signal characteristic S FIMP Will be repeated L times.

[0408] To convey an intuitive understanding of this signal processing, it may be helpful to consider the superposition principle and repetitive signals such as sinusoids. Sinusoids can exhibit both amplitude and phase values. Simply put, the superposition principle, also known as the superposition property, states that for all linear systems, the net response caused by two or more stimuli at a given location and time is the sum of the responses caused by each stimulus individually.

[0409] Sound waves are one such stimulus. Similarly, a vibration signal (eg, including a signal characteristic S indicating the impact of the protrusion with the toe 205) FIMP The vibration signal S EA 、S MD , S(j), S(r)) is one of such stimuli. In fact, including the signal feature S FIMP The vibration signal S EA 、S MD , S(j), S(r) can be considered as the sum of sinusoidal signals, each of which exhibits an amplitude value and a phase value. In this regard, reference is made to the Fourier series (see Equation 1 below):

[0410] n=∞

[0411] F(t) = ∑Cn sin(nωt +Ф n ) (Formula 1)

[0412] n=0

[0413] in,

[0414] n=0 the average value of the signal over a period of time (can be zero, but does not have to be zero),

[0415] n=1 corresponds to the fundamental frequency of the signal F(t),

[0416] n=2 corresponds to the first harmonic part of the signal F(t),

[0417] ω=angular frequency, that is (2*π*f ROT ),

[0418] f ROT = housing speed in cycles per second,

[0419] t = time,

[0420] φ n = the phase angle of the nth partial, and

[0421] C n = the amplitude of the nth partial tone.

[0422] From the Fourier series above, it can be concluded that the time signal can be considered as the superposition of multiple sinusoidal signals.

[0423] A harmonic is any frequency greater than the fundamental frequency of a signal.

[0424] In the above example, it should be noted that the fundamental frequency will be f ROT , that is, the shell rotation speed, because the FFT 510 receives the marker signal value P(j)=1 only once every time the shell 20 rotates one circle (see, for example, FIG. 14 ).

[0425] Using the Fourier analysis model, the fundamental frequency and its overtones are collectively referred to as partials. Harmonics, or more precisely, harmonic partials, are partials whose frequencies are integer multiples of the fundamental frequency (including the fundamental frequency, which is itself 1).

[0426] refer to Figure 15A and / or Figure 15B As with Equation 1 above, FFT 510 can deliver an amplitude value C with n=L n (r), namely C L (r) = Sp(r) FFT 510 can also transmit the phase angle of the partial tone (n = L), that is, ΦL(r) = FI(r).

[0427] Consider now an example where the mill housing has ten (10) protrusions 310 when it rotates at a speed of 10 revolutions per minute (rpm). A speed of 10 rpm means one rotation every 6 seconds, i.e., f ROT = 0.1667 rpm. There are ten protrusions (i.e. L = 10) and f ROT = 0.1667 rpm, so that the repetition frequency f of the signal associated with the protrusion 310 is R is 1.667Hz, because the repetition frequency f R is the 10th order frequency.

[0428] Position signals P(j), P(q) (see Figure 15A and / or Figure 15B) can be used as a reference signal for the digital measurement signals S(j) and S(r). According to some embodiments, when the FFT analyzer is configured to receive the reference signal, i.e., the position signals P(j) and P(q), once per rotation of the rotating housing 20, the settings of the FFT analyzer should meet the following criteria:

[0429] The integer value Oi is set equal to L, the number of protrusions 310 in the housing 20, and

[0430] Select the variable that can be set. MAX and B n , so that the mathematical expression Oi*B n / O MAX In other words: when the integer value Oi is set equal to L, the variable O can be set MAX and B n should be set to an integer value so that the variable N R is a positive integer,

[0431] Among them, N R =Oi*B n / O MAX ,

[0432] O MAX is the maximum order; and

[0433] B n is the number of bins in the spectrum produced by the FFT, and

[0434] Oi is the frequency of interest, expressed as an integer of order, and where f ROT is the frequency of order 1, i.e. the fundamental frequency. In other words, the speed f of the housing 20 is ROT is the fundamental frequency, and L is the number of protrusions 310 in the housing 20 .

[0435] Using the above settings, the integer value Oi is set equal to L, and referring to the above Figure 15A and / or Figure 15B Using Equation 1, FFT 510 can deliver an amplitude value C with n=L n , that is, C L =Sp(r). FFT 510 can also transmit the phase angle of part (n=L), that is, Φ L =FI(r).

[0436] Therefore, according to an embodiment of the present disclosure, when the rotating housing 20 rotates once, the FFT 510 receives the position reference signal P(j), P(q) once, and the FFT analyzer can be configured to generate a peak amplitude value C of the signal. L , the repetition frequency of the signal f Ris the L-order frequency, where L is the number of equally spaced protrusions 310 in the rotating housing 20 .

[0437] Referring to the discussion above regarding Equation 1 in this disclosure, the repetition frequency f R The amplitude of the signal with the L-order frequency can be called C n , where n = L, i.e. C L . Refer to formula 1 and Figure 15A and / or Figure 15B , can transmit amplitude value C L , as the peak amplitude value, in Figure 15A and / or Figure 15B It is represented as Sp(r).

[0438] Referring again to Formula 1 above, in the present disclosure, the repetition frequency f may be transmitted R The phase angle value Ф of the signal with L-order frequency L , as a time indication value, which indicates the impact force F IMP The duration T between the appearance of the rotation reference position of the rotating housing and the appearance of the rotation reference position of the rotating housing D1 .

[0439] Therefore, according to an embodiment of the present disclosure, when the FFT 510 receives the position reference signals P(j), P(q) once every time the rotating housing 20 rotates once, the FFT analyzer can be configured to generate a repetition frequency f R The phase angle value Φ of the signal with L-order frequency L , where L is the number of equally spaced protrusions 310 in the rotating housing 20 .

[0440] Therefore, using the above settings, the integer value Oi is set equal to L, and referring to the above Figure 15A and / or Figure 15B Using Equation 1, FFT 510 can generate a phase angle value Φ L .

[0441] Combine Figure 1A refer to Figure 15A and / or Figure 15B , state value Sp(r)=C L and FI(r)=Ф L The displayed analysis results may include information indicating the internal status of the tumbling process, for enabling the operator 230 to control the tumbling machine.

[0442] Figure 15B is a block diagram illustrating an example of the state parameter extractor 450 . Figure 15BThe example state parameter extractor 450 in FIG. 5 includes a shell velocity detector 500, a velocity variation compensation extractor 470, a time-synchronized averager 471, and a fast Fourier transformer 510FFT. In this disclosure, the abbreviation TSA may be used for the time-synchronized averager. The example state parameter extractor 450 may be Figure 15A The state parameter extractor 450 described in

[15] is further configured to add a time synchronized averager TSA471. TSA471 is configured to receive the vibration signal value S(q) and the corresponding position signal value P(q) transmitted by the speed variation compensation extractor 470. TSA471 is configured to receive the vibration signal value S(q) and the corresponding position signal value P(q) corresponding to a plurality of rotations or cycles, and to generate an average vibration signal value S(t), wherein the average vibration signal value S(t) is based on the number N of vibration values ​​detected at the same rotational position of the housing 20. TSA .

[0443] As discussed elsewhere in this disclosure, the compensated decimator 470 is configured to generate a decimated digital vibration signal S MDR , so that even when the rotation speed changes, the number N of vibration sample values ​​per revolution of the rotating shell V Also maintained at a constant value or substantially constant value. Therefore, when the compensating extractor 470 delivers N per revolution of the rotating shell V vibration sample values, then every Nth V The vibration sample values ​​are associated with the same rotational position.

[0444] Therefore, when N V = 100, the speed variation compensation extractor 470 outputs one hundred (100) vibration signal values ​​S(q) per revolution. V = 100 and the TSA 471 is configured to generate a single average vibration signal value S TSA (t), as N TSA When the average value of the vibration signal value S(q) is obtained, the output average value S TSA (t) can be generated as:

[0445] S TSA (t)=(S(q)+S(q+100)+S(q+200)) / N TSA .

[0446] For example, if N TSA =3, then TSA 471 transmits S TSA (t), where each value is based on three vibration signal values ​​S(q). Therefore, it should be understood that the signal values ​​S(q), S(q+100), and S(q+200) all represent the qth position. Therefore, the TSA 471 transmits the average signal value S TSA(t), which has the same number N of outputs per revolution as provided by the speed variation compensation extractor 470 V For example, if the speed variation compensation extractor 470 outputs N per revolution V = 100 (one hundred) vibration signal values ​​S(q), then TSA 471 transmits N per revolution V = 100 (one hundred) average vibration signal values ​​S(t).

[0447] The combination of the housing speed detector 500, the speed variation compensation extractor 470, and the time-synchronized averager 471 allows the output from the TSA 471 to have a vibration value averaged over several revolutions, which advantageously reduces noise. It should be noted that the TSA is configured to generate an average vibration value such that the average vibration value represents an average of multiple vibration values ​​detected at the same rotational position of the housing 20.

[0448] like Figure 15B As shown, the output signal P of TSA 471 TSA and S TSA It can be provided to FFT 510.

[0449] In some examples, the output P of TSA 471 TSA S TSA Provided to HCI 210.

[0450] In some examples, the HCI 210 is configured to set the number of revolutions or cycles that the TSA 471 is configured to average.

[0451] Information about the current internal state X of the mill 10 may be conveyed by one or more internal state values ​​in such a way that it makes intuitive sense to an operator 230 of the mill 10 .

[0452] Figure 16 is a diagram of an example of a visual indication of analysis results.

[0453] According to one example, the visual indication of the analysis results may include providing a polar coordinate system 520. A polar coordinate system is a two-dimensional coordinate system in which each point on a plane is determined by a distance from a reference point 530 and an angle from a reference direction 540. Reference point 530 (analogous to the origin of a Cartesian coordinate system) is referred to as a pole 530, and a ray from the pole in the reference direction is a polar axis. The distance to the pole is referred to as a radial coordinate, radial distance, or simply radius, and the angle is referred to as an angular coordinate, polar angle, or azimuth.

[0454] According to one example, the amplitude value Sp(r) is used as the radius, and the time relationship values ​​FI(r), Φ(r), T D Used as angular coordinate.

[0455] In this manner, by providing the internal status indicator object 550 on the display 210S, the internal status of the monitored roller mill ( Figure 16 Combine Figure 1A and / or Figure 1B ). Figure 16 Combine Figure 1A and / or Figure 1B and Figure 14 may help to understand the following examples.

[0456] Thus, one example relates to an electronic roller mill monitoring system 150, 210S for generating and displaying information related to a grinding process in a roller mill 10 having a housing 20 that rotates at a rotational speed f ROT Rotates about axis 60 for grinding material charge 30 by tumbling the material charge in a rotating housing. Example monitoring system 150 includes:

[0457] a computer-implemented method of representing on a screen display 210S the internal state of the grinding process in the roller mill,

[0458] The method includes:

[0459] The screen display 210S displays:

[0460] Polar coordinate system 520, the polar coordinate system 520 having

[0461] Reference point (O, 530), and

[0462] Reference directions (0°, 360°, 540°); and

[0463] The first internal state indicator object (550, S P1 , T D1 ), which indicates the internal state of the grinding process, has a first radius (Sp(r), S P1 ), and has a first polar angle (FI(r), Φ(r), T relative to the reference direction (0°, 360°, 540°) D , T D1 ),

[0464] The first radius (Sp(r), S P1 ) indicates the impact force (F) generated when the protrusion (310) on the inner shell surface of the rotating shell interacts with the toe portion 205 of the charge material (30) IMP ),and

[0465] The first polar angle (F1(r), Φ(r), T D , T D1) indicates the position of the toe 205 between the two protrusions 310 in the rotating housing 20.

[0466] As described above, the state parameter extractor 450 can be configured to generate a continuous pair of state values ​​Sp(r) and FI(r). The state parameter extractor 450 can also generate time derivative values ​​of the state values ​​Sp(r) and FI(r), respectively. This can be done, for example, by subtracting the most recent previous state value Sp(r-1) from the most recent state value Sp(r) divided by the duration between the two values. Similarly, the numerical derivative of the internal state value FI can be obtained. Therefore, derivative values ​​dSp(r) and dFI(r) can be generated. The derivative values ​​dSp(r) and dFI(r) can be used to indicate the first internal state indicator object (550, S P1 , T D1 ) movement.

[0467] Figure 17 and Figure 18 is another example of a visual indication of analysis results. Figure 17 and Figure 18 , the derivative value can be used to display an arrow 560 on the screen display 210S, the arrow 560 originating from the first internal state indicator object (550, S P1 , T D1 ) and has an extension that depends on the magnitude of the derivative value. In other words, the absence of arrow 560 means that the internal state is stable and has not changed for a period of time. Figure 18 The arrow in 560 is Figure 17 The arrow 560 is long, indicating Figure 18 The internal state of the mill shown in Figure 17 The internal state of the mill shown in changes more rapidly.

[0468] Figure 19A and Figure 19B Another example of a visual indication based on the analysis result of the internal state of the roller mill 10 is shown. A recent internal state indicator object 550(r) indicates the current internal state of the mill 10. Another internal state indicator object 550(r-1) indicates the most recent previous internal state of the mill 10.

[0469] The internal state indicator object 550 (1) shown as a small hollow circle indicates the internal state of the mill 10, which is almost empty. It should be noted that when the roller mill is started from the no-load state, the initial internal state indicator object appears at the initial polar angle φ (1), which represents the first detected toe position of the mill. Figure 19A and Figure 19B, starting from the small hollow circle 550(1), the first thirty-one (31) detected toe positions are represented as hollow circles. Based on experimental measurements, it seems that the initial polar angle Φ(1) can be used as a reference toe position value. Therefore, the initial polar angle Φ(1) can be called the reference toe position value Φ TR . For its internal state is Figure 19A and Figure 19B The particular roller mill represented by the display 210S shown, the reference toe position corresponds to an angle value φ of approximately 47 degrees. TR ,like Figure 19A and Figure 19B shown.

[0470] The first thirty-one (31) detected toe positions are represented by hollow circles, and the subsequent sequence of toe positions are represented by shaded circles, one of which is in Figure 19A It is represented as 550(p). Figure 19A The shaded circles in FIG. 3 indicate that the filling degree of the mill housing 20 is higher than the filling degree indicated by the open circles. Figure 19A The solid black circles in φ indicate that the filling degree of the mill shell 20 is higher than the filling degree shown by the shaded circles. Therefore, it should be noted that the initial lowest detected filling degree appears to be represented by a relatively small radius, i.e. a low peak amplitude value Sp at the initial polar angle φ(1).

[0471] refer to Figure 19A , the gradually increasing detected toe position FI(r) and the correspondingly increasing filling degree of the mill housing 20 present an image of a spiral arm rotating outward in the counterclockwise direction starting from the first internal state indicator object 550(1), as shown Figure 19A As shown by the curved arrow 560A in FIG.

[0472] In this way, the current internal state of the roller mill 20 can be represented and visualized so that it is intuitively meaningful to the operator 230 of the mill system 5. It should be noted that although Figure 17 As shown, the display of a single internal state indicator object 550 indicates the current internal state or the most recently detected internal state of the mill 10, but as shown in FIG. Figure 19A As shown, the display of the time progression of the internal state indicator object ranging from the initial state 550 (1) via intermediate states (e.g., 550 (p) and 550 (r-1) to 550 (r)) represents the current internal state 550 (r) and the history of several earlier internal states 550 (p), 550 (p+1), 550 (r-1) of the mill 10.

[0473] In other words, the increasing polar angle F1(r) and the increasing radius S p(r) in combination, presents an image of a spiral arm rotating outward from the first internal state indicator object 550(1), as shown in FIG. Figure 19A The "angular length" X6(r) of the spiral arm from the initial polar angle Φ(1) of the first internal state indicator object 550(1) to the current or most recently detected toe position FI(r) appears to indicate the absolute position X6(r) of the toe 205 (e.g., see Figure 2 and Figure 14). In this regard, it should be noted that Figure 19A 360 degrees in the polar coordinate system 520 corresponds to two adjacent protrusions (e.g., Figure 2 100% of the distance between the leading edges of 312C and 312D).

[0474] Example of a speed change phase state parameter extractor

[0475] As mentioned above, if the roller mill housing 20 is variable Speed ​​f ROT The analysis of the measured data is more complicated if the mill housing rotates. In fact, it seems that even very small changes in the rotational speed of the mill housing can have a significant negative impact on the quality of the detected signal in terms of the tailing effect. Therefore, the rotational speed f of the mill housing 20 is ROT Very precise detection of the velocity seems to be crucial, and precise compensation for any speed changes also seems to be crucial.

[0476] refer to Figure 15A and / or Figure 15B , the housing speed detector 500 may transmit a signal indicating when the rotational speed changes, such as in conjunction with Figure 9 As discussed. Again refer to Figure 15A and / or Figure 15B , signals S(j) and P(j) and speed value f ROT (j) may be transmitted to a speed variation compensation decimator 470. The speed variation compensation decimator 470 may also be referred to as a fractional decimator. The decimator 470 is configured to calculate the fractional decimator based on the received speed value f. ROT (j) Extracting the digital measurement signal S MD According to an example, the decimator 470 is configured to decimate the digital measurement signal S by a variable decimation factor D. MD , based on the variable speed value f during the measurement session ROT (j) Adjusting the variable decimation factor D. Thus, the compensation decimator 470 is configured to generate a decimated quantity vibration signal S MDR , so that when the rotation speed changes, the number N of vibration sample values ​​of the rotating shell per rotation V According to some embodiments, when the number of sample values ​​per rotation N isV When the variation is less than 5%, the number of sample values ​​per rotation of the rotating shell is considered to be a substantially constant value. V The number of sample values ​​per rotation of the rotating shell is considered to be a substantially constant value when the variation is less than 1%. According to a most preferred embodiment, the number of sample values ​​per rotation of the rotating shell is considered to be a substantially constant value when the variation is less than 0.2%.

[0477] therefore, Figure 15A Examples and Figure 15B The embodiment includes a fractional decimator 470 for decimating by a factor D=N / U D The sampling rate is decimated, where U D and N are both positive integers. Thus, the fractional decimator 470 advantageously decimates the sampling rate. Thus, the speed variation compensation decimator 470 can operate to obtain the sample rate by the fraction D=N / U. D To extract the signals S(j) and P(j) and f ROT (j). According to one embodiment, U D The values ​​of U and N can be selected in the range from 2 to 2000. According to one embodiment, U D The values ​​of U and N can be selected in the range from 500 to 1500. According to another embodiment, U D and N can be chosen in the range from 900 to 1100. In this article it should be noted that the term "fraction" has the following background: A fraction (from the Latin fractus, "to break") represents a part of a whole, or more generally, any number of equal parts. In a positive common fraction, both the numerator and the denominator are natural numbers. The numerator represents a number of equal parts, and the denominator represents how many parts make up a unit or a whole. Common fractions are quantities that represent rational numbers. The same quantity can also be expressed as a decimal, a percentage, or a negative exponent. For example, 0.01, 1%, and 10 -2 Both are equal to the fraction 1 / 100. Therefore, the fraction D = N / U D can be thought of as an inverse fraction.

[0478] Therefore, the resulting signal S delivered by the fractional decimator 470 is MDR With sampling rate:

[0479] f SR =f S / D=f S *U D / N

[0480] Among them, f S is the signal S received by the fractional decimator 470 REDThe sampling rate.

[0481] Score value U D / N depends on the speed control signal received at input port 490. The speed control signal may be an indication of the rotational speed f of the rotating housing. ROT signal.

[0482] The variable decimator value D of the decimator can be set to D=f S / f SR , where f S is the initial sampling rate of the A / D converter, f SR is the vibration signal S that indicates the number of samples to be extracted MDR For example, when there are twelve (12) protrusions to be monitored in the mill housing, the set point value f SR It can be set to 768 samples per revolution, that is, the number of samples per revolution is set to the number of extracted vibration signals S MDR The compensation decimator 470 is configured to oscillate the signal S by the decimated amount. MDR The position signal P(q) is generated at regular intervals that depend on the set point value f SR For example, when f SR When set to 768 samples per revolution, the position signal P(q) can be transmitted once every 768 samples of the extracted vibration signal S(q).

[0483] Therefore, the sampling frequency f of the output data value R(q) SR (also known as f SR2 ) than the input sampling frequency f S =lower by a factor D. The factor D can be set to any number greater than 1 and can be a fraction, as discussed elsewhere in this disclosure. According to a preferred embodiment, the factor D can be set to a value between 1.0 and 20.0. In a preferred embodiment, the factor D is a fraction that can be set to a value between about 1.3 and about 3.0. The integer U can be D and N are set to appropriate values ​​to obtain the factor D.

[0484] The factor D is equal to N divided by U D :

[0485] D=N / U D .

[0486] According to one embodiment, the integer U D and N can be set to large integers so that the factor D = N / U D Able to follow speed changes with minimal error. Select variable U DIt is advantageous for N to be an integer greater than 1000 to have high accuracy when adjusting the output sampling frequency to track the change in the rotation speed of the housing 20. Therefore, for example, N is set to 500 and U is set to D If set to 1001, then D = 2.002.

[0487] The variable D is set to a suitable value at the start of the measurement and this value is associated with the specific rotational speed of the rotating component to be monitored. Thereafter, during the measurement session, the fractional value D is automatically adjusted in response to the rotational speed of the rotating component to be monitored so that the output signal S MDR A substantially constant number of sample values ​​is provided per revolution of the rotating housing.

[0488] Figure 20 is a block diagram of one example of a compensated decimator 470. This example of a compensated decimator is designated 470B.

[0489] The compensation extractor 470B may include a memory 604 adapted to receive and store data values ​​S(j) and corresponding rotational speeds f of the rotating mill housing being monitored. ROT Therefore, the memory 604 can store each data value S(j) so that it corresponds to the sensor signal S corresponding to the data value S(j). EA The monitored mill housing speed f ROT (j) is associated with the value. Refer to the above Figure 7-13 Describes the corresponding speed value f ROT (j) Provision of the associated data value S(j).

[0490] The compensation decimator 470B receives a signal having a sampling frequency f SR1 The signal S MD , as a sequence of data values ​​S(j) and transmits at its output 590 a data signal with a reduced sampling frequency f SR The output signal S MDR , as a sequence of another number of data values ​​R(q).

[0491] The compensation extractor 470B may include a memory 604 adapted to receive and store data values ​​S(j) and corresponding rotational speeds f of the rotating mill housing being monitored. ROT The memory 604 may store data values ​​S(j) in blocks such that each block is associated with a value indicating the associated rotational speed of the mill housing being monitored, as described below in conjunction with Figure 21 As stated.

[0492] The compensation extractor 470B may further include a compensation extraction variable generator 606 adapted to generate a compensation value D. The compensation value D may be a floating point number. Thus, in response to the received speed value f ROT, the compensation number can be controlled as a floating point value so that the floating point value indicates the speed value f with a certain inaccuracy ROT As mentioned above, when implemented by a suitably programmed DSP, the inaccuracy of floating point values ​​may depend on the DSP's ability to generate floating point values.

[0493] Furthermore, the compensation decimator 470B may further include an FIR filter 608. In this regard, the acronym FIR stands for Finite Impulse Response. The FIR filter 608 is a low-pass FIR filter having a specific low-pass cutoff frequency, adapted to reduce the power consumption by a factor D. MAX Extract. Factor D MAX Can be set to a suitable value, for example, 20.000. In addition, the compensation decimator 470B may further include a filter parameter generator 610.

[0494] Reference below Figure 21 and 22 describe the operation of the compensation extractor 470B.

[0495] Figure 21 It shows the operation Figure 20 Flowchart of an embodiment of a method of compensating sampler 470B.

[0496] In a first step S2000, the rotational speed f of the mill housing to be monitored is ROT is recorded in the memory 604 ( Figure 20 and Figure 21 ), and this can be done at approximately the same time as the start of the vibration measurement. According to another example, the rotational speed of the mill housing to be monitored is measured over a period of time. The maximum detection speed f ROTmax and minimum detection speed f ROTmin It can be recorded in, for example, the memory 604 ( Figure 20 and Figure 21 ).

[0497] In step S2010 , the recorded speed values ​​are analyzed to determine whether the rotation speed has changed.

[0498] In step S2020, the user interface 210, 210S displays the recorded speed value f RO Or speed value f ROTmin 、f ROTmax , and requests the user to input the desired order value Oi. As mentioned above, the mill housing rotation frequency f ROT This is often referred to as "1st order." An interesting signal might occur ten times per revolution of the mill housing (10th order). Furthermore, it might be interesting to analyze the overtones of some signals, so it might be interesting to measure signals up to 100th order, 500th order, or even higher. Thus, the user can input the order Oi using the user interface 210, 210S.

[0499] In step S2030, determine the appropriate output sampling rate f SR In the present disclosure, the output sampling rate f SR Can also be called f SR2 According to one embodiment, the output sampling rate f SR is set to f SR =C*Oi*f ROTmin ,

[0500] in,

[0501] C is a constant with a value greater than 2.0,

[0502] Oi is a quantity that indicates the relationship between the rotational speed of the mill housing being monitored and the repetition frequency of the signal to be analyzed.

[0503] f ROTmin is the lowest rotational speed of the monitored mill housing to be expected during the upcoming measurement session. According to one embodiment, as described above, the value f ROTmin It is the lowest rotation speed detected in step S2020.

[0504] Considering the sampling theorem, the constant C may be selected to be 2.00 (two) or higher. According to an embodiment of the present disclosure, the constant C may be preset to a value between 2.40 and 2.70.

[0505] According to one embodiment, the factor C is advantageously chosen such that 100*C / 2 represents an integer. According to one embodiment, the factor C may be set to 2.56. C is chosen to be 2.56 such that 100*C=256=2 raised to the eighth power.

[0506] In step S2050 , the compensation extraction variable value D is determined. When the rotational speed of the monitored mill housing changes, the compensation extraction variable value D will change according to the instantaneously detected speed value.

[0507] According to one embodiment, the maximum compensation decimation variable value D MAX Set to D MAX =f ROTmax / f ROTmin The value of the minimum compensation extraction variable value D MIN is set to 1.0. After that, the actual speed value f ROT Perform instantaneous real-time measurements and set the instantaneous compensation value D accordingly.

[0508] f ROT is a value indicating the measured rotational speed of the rotating mill housing to be monitored.

[0509] In step S2060 , the actual measurement is started and the desired total duration of the measurement can be determined. The total duration of the measurement can be determined based on the desired number of revolutions X of the monitored mill housing.

[0510] When the measurement starts, the digital signal S MD is passed to the input 480 of the compensation decimator. In the following, the signal S is discussed in terms of a signal having sample values ​​S(j). MD , where j is an integer.

[0511] In step S2070, the data value S(j) is recorded in the memory 604, and each vibration data value S(j) is compared with the rotation speed value f ROT (j) Associated.

[0512] In a subsequent step S2080, the recorded rotational speed values ​​are analyzed and the recorded data values ​​S(j) are divided into data blocks according to the rotational speed values. In this manner, multiple blocks of data value blocks S(j) are generated, each of which is associated with a rotational speed value. The rotational speed value indicates the rotational speed of the monitored mill housing at the time that particular block of data value S(j) was recorded. The individual data blocks can have different sizes, i.e., each data block can store a different number of data values ​​S(j).

[0513] For example, if the monitored rotating mill housing is first moved at a first speed f during a first time period ROT1 rotates, and then, during a second, shorter period of time, changes speed to a second speed f ROT2 Rotation, the recorded data value S(j) can be divided into two data blocks, the first block of data values ​​and the first speed value f ROT1 The second data block value is associated with the second speed value f ROT2 In this case, the second data block will contain fewer data values ​​than the first data block because the second time period is shorter.

[0514] According to one embodiment, when all recorded data values ​​S(j) have been divided into blocks, and all blocks have been associated with rotational speed values, the method proceeds to step S2090 .

[0515] In step S2090, the first block of data values ​​S(j) is selected and the corresponding speed value f is determined. ROT The compensation decimation value D is associated with the first block data value S(j). According to one embodiment, when all blocks have been associated with corresponding compensation decimation values ​​D, the method proceeds to step S2100. Therefore, the value of the compensation decimation value D is determined according to the speed f. ROT Make adjustments.

[0516] In step S2100, a block of data values ​​S(j) and associated compensated decimated values ​​D are selected, as described above in step S2090.

[0517] In step S2110 , a block of output values ​​R is generated in response to the selected block of input values ​​S and the associated compensated decimated values ​​D. This may be done as described with reference to FIG.

[0518] In step S2120, it is checked whether there are any remaining input data values ​​to be processed. If there is another block of input data values ​​to be processed, step S2100 is repeated. If there are no remaining blocks of input data values ​​to be processed, the measurement session is completed.

[0519] Figure 22A 、 Figure 22B and Figure 22C Shows the operation Figure 20 Flowchart of an embodiment of a method of compensating sampler 470B.

[0520] In step S2200, an input data value block S(j) and an associated specific compensation decimation value D are received. According to one embodiment, the received data is as above Figure 21 The input data values ​​S(j) in the received input data value block S are all associated with a specific compensation decimation value D.

[0521] In steps S2210 to S2390, the FIR filter 608 (see Figure 20 ) is applied to the specific compensation decimation value D received in step S2200 and generates a set of corresponding output signal values ​​R(q). This will be described in more detail below.

[0522] In step S2210, a filter setting suitable for a specific compensation decimation value D is selected. Figure 20 As mentioned, the FIR filter 608 is a low-pass FIR filter having a filter size suitable for filtering with a factor D. MAX A low-pass cutoff frequency at which decimation occurs. Factor D MAX Can be set to a suitable value, for example, 20.

[0523] Filter ratio F R is set to depend on the factor D MAX and the value of the specific compensation decimation value D received in step S2200. Step S2210 may be performed by the filter parameter generator 610 ( Figure 20 ) to execute.

[0524] In step S2220, a starting position value x is selected in the received input data block s(j). It should be noted that the starting position value x does not have to be an integer. The FIR filter 608 has a length FLENGTH , and then according to the filter length F LENGTH and filter ratio F R To select the starting position value x. The filter ratio FR is as set in step S2210 above. According to one embodiment, the starting position value x can be set to x:=F LENGTH / F R .

[0525] In step S2230, a filtered sum value SUM is prepared and set to an initial value, for example, SUM:=0.0.

[0526] In step S2240, a position j in the received input data that is adjacent to and preceding the position x is selected. The position j may be selected as an integer portion of x.

[0527] In step S2250, a position Fpos in the FIR filter is selected that corresponds to the selected position j in the received input data. The position Fpos may be a compensation amount. Relative to the middle position of the filter, the filter position Fpos may be determined as:

[0528] Fpos=[(xj)*F R ]

[0529] Among them, F R is the filtering ratio.

[0530] In step S2260, it is checked whether the determined filter position value Fpos is outside the allowed limit, i.e., points to a position outside the filter. If this is the case, the following step S2300 is performed. Otherwise, the following step S2270 is performed.

[0531] In step S2270, the filter values ​​are calculated by interpolation. It should be noted that adjacent filter coefficient values ​​in an FIR low-pass filter usually have similar values. Therefore, interpolation will be advantageously accurate. First, the integer position value IFpos is calculated:

[0532] IFpos:=integer part of Fpos

[0533] The filter value Fval at position Fpos will be:

[0534] Fval=A(IFpos)+[A(IFpos+1)-A(IFpos)]*[Fpos-IFpos]

[0535] where A(IFpos) and A(IFpos+1) are the values ​​in the reference filter and the filter position Fpos is the position between these values.

[0536] In step S2280, in response to the signal position j, an update of the filtered sum value SUM is calculated:

[0537] SUM:=SUM+Fval*S(j)

[0538] In step S2290, move to another signal position:

[0539] Set j:=j-1

[0540] Thereafter, go to step S2250.

[0541] In step 2300, a position j is selected in the received input data that is adjacent to and after position x. This position j can be selected as the integer part of x plus 1 (one), that is, j:=1+integer part of x.

[0542] In step S2310, a position is selected in the FIR filter corresponding to the selected position j in the received input data. The position Fpos may be a compensation amount. Relative to the middle position of the filter, the filter position Fpos may be determined as:

[0543] Fpos=[(jx)*F R ]

[0544] Among them, F R is the filtering ratio.

[0545] In step S2320, it is checked whether the determined filter position value Fpos is outside the allowed limit, i.e., points to a position outside the filter. If this is the case, the following step S2360 is performed. Otherwise, the step S2330 is performed.

[0546] In step S2330, the filter values ​​are calculated by interpolation. It should be noted that adjacent filter coefficient values ​​in an FIR low-pass filter usually have similar values. Therefore, interpolation will be advantageously accurate. First, the integer position value IFpos is calculated:

[0547] IFpos:=integer part of Fpos

[0548] The filtered value of position Fpos is:

[0549] Fval(Fpos)=A(IFpos)+[A(IFpos+1)-A(IFpos)]*[Fpos-IFpos]

[0550] where A(IFpos) and A(IFpos+1) are the values ​​in the reference filter and the filter position Fpos is the position between these values.

[0551] In step S2340, in response to the signal position j, an update of the filtered sum value SUM is calculated:

[0552] SUM:=SUM+Fval*S(j)

[0553] In step S2350, move to another signal position:

[0554] Set j:=j+1

[0555] Thereafter, go to step S2310.

[0556] In step S2360, the output data value R(j) is transmitted. The output data value R(j) may be transmitted to the memory such that consecutive output data values ​​are stored in consecutive memory locations. The value of the output data value R(j) is:

[0557] R(j):=SUM.

[0558] In step S2370, the position value x is updated:

[0559] x:=x+D.

[0560] In step S2380, update the position value j

[0561] j:=j+1.

[0562] In step S2390, it is checked whether the expected number of output data values ​​has been generated. If the expected number of output data values ​​has not been generated, then go to step S2230. If the expected number of output data values ​​has been generated, then go to step S2390. Figure 21 Step S2120 in the described method.

[0563] In practice, step S2390 is designed to ensure that an output signal value R(q) corresponding to the input data value block S received in step S2200 is generated, and when an output signal value R corresponding to the input data value S has been generated, the execution of Figure 21 Step S2120 in .

[0564] The method described with reference to FIG. 22 may be implemented as a computer program subroutine, and steps S2100 and S2110 may be implemented as a main program.

[0565] Figure 23 Another example of a cross-sectional view of the middle portion 98 of the rotating mill housing 20 during operation is shown. This view may be taken, for example, along Figure 1A According to the line AA. Figure 23In the example of FIG. 5 , the roller mill housing 20 has six protrusions 310 configured to engage the material charge 30 when the housing rotates about the axis 60 , ie, the number L=6. For clarity, Figure 23 The protrusions in the example are labeled 3101, 3102, 3103, 3104, 3105 and 3106 respectively.

[0566] A position sensor 170 is provided to generate a position signal E according to the rotational position of the housing 20. P As described above, the housing 20 can rotate about the rotation axis 60, so the position sensor 170 installed in a fixed manner can generate a position signal E P , the position signal has a series of shell position signal values ​​P S , used to indicate the instantaneous rotation position of the housing 20. Figure 23 As shown, a plurality of position marks 180 may be provided on the outer surface of the housing 20, so that when the housing 20 rotates about the rotation axis 60, several position marks 180 pass by the position sensor 170 in one rotation of the housing 20, and each mark 180 thereby causes the position sensor 170 to generate a rotation mark signal value Ps. According to one embodiment, L position marks 180 are provided on the housing 20, so that when the housing 20 rotates about the rotation axis 60, the position marks 1801...180 L Continuously passing through the position sensor 170, so that the position sensor 170 generates L rotation mark signal values ​​Ps in one rotation of the housing 20. Figure 23 In the illustrated embodiment, there are six protrusions 310 , ie, L=6, and six position marks 1801 , 1802 , 1803 , 1804 , 1805 and 1806 .

[0567] It is believed to be important that the arrangement of the position markings 180 in terms of angular position mirrors the arrangement of the protrusions 310 on the inner surface 22 of the housing 20 in terms of angular position.

[0568] exist Figure 23 In the embodiment of FIG. 1 , L position marks 180 are positioned on the periphery of the housing 20 in an equidistant manner, so that the position sensor 170 generates a mark signal Ps every 360 / L degrees during the rotation of the housing 20. In this case, it should be noted that Figure 23 In the embodiment, the L-shaped protrusions 3101, 3102, 3103, 3104, 3105 and 310 LThe protrusions 310 are positioned on the inner surface 22 of the housing 20 in an equidistant manner. It is believed that the equidistant positions of the protrusions 310 and the equidistant positions of the position marks 180 are important for some embodiments of the present disclosure. This is believed to be important for some embodiments of the present disclosure because the position marks 180 cause the generation of a position reference signal value and the protrusions 310 cause the generation of a signal event, such as an amplitude peak, in the vibration signal when engaging material in the rotating mill charge (see reference S). EA 、S MD , Se(i), S(j), S(q), e.g., in FIG. 1 and FIG. 15 ). Additionally, the duration between the occurrence of the position reference signal value and the occurrence of the signal event in the vibration signal can be indicative of an internal state of the mill in operation, as discussed elsewhere in this disclosure, which duration is caused by the protrusion 310 engaging material in the charge of the rotating mill housing. For example, the duration between the occurrence of the position reference signal value and the occurrence of the signal event in the vibration signal can be indicative of an internal state, e.g., the position of the toe 205, which duration is caused by the protrusion 310 engaging material in the charge of the rotating mill housing.

[0569] However, the actual placement of the position mark 180 relative to the position of the protrusion 310 is considered to be less important. Figure 23 The position marks 180 are shown as being positioned at the same angular position as the protrusions 310, but it should be noted that the position marks 180 may also be displaced in terms of angular position. However, if the position marks 180 are displaced in terms of angular position, it is believed that it is important that all position marks 180 are displaced equally to maintain the mutually equidistant positions of the position marks 180. More specifically, it is believed that it is important that the arrangement of the position marks 180 in terms of angular position mirrors the arrangement of the protrusions 310 on the inner surface 22 of the housing 20 in terms of angular position.

[0570] As mentioned above, combined Figure 19A and Figure 19B It has been observed that when the roller mill is started from an empty state, the initial internal state indicator object appears at an initial polar angle Φ(1), which represents the first detected toe position 205 of the mill. Based on experimental measurements, it seems that the initial polar angle Φ(1) can be used as a reference toe position value. Therefore, the initial polar angle Φ(1) can be referred to as the reference toe position value Φ TR . For its internal state is Figure 19A and Figure 19B The particular roller mill represented by the display 210S shown, the reference toe position corresponds to an angle value φ of approximately 47 degrees. TR ,like Figure 19A and Figure 19B Reference Figure 214, it is believed that if the position marker 180 is physically moved to a different position in terms of angular position, the reference toe position value Φ TR The angle value of will change to a numerically different angle value.

[0571] like Figure 23 The illustrated arrangement of the rotating mill housing 20 may be used in conjunction with the state parameter extractor 450 exemplified in the present disclosure. Figure 15A and Figure 15B ,like Figure 23 As shown, the arrangement of the rotating mill housing 20 can be used to generate a marker signal P(i), which is transmitted to the housing speed value generator 500. Therefore, during the rotation of the housing 20, the housing speed value generator 500 will receive a marker signal P(i) having a position indicator signal value every 360 / L degrees. Therefore, when the rotation speed f ROT When constant, during the rotation of the housing 20, the fast Fourier transformer 510 receives the flag signal value P(j)=1 from the speed value generator 500 every 360 / L degrees. Alternatively, when the rotation speed f ROT When changing, during the rotation of the housing 20, the fast Fourier transformer 510 will receive a flag signal value P(q)=1 from the extractor 470, 470B every 360 / L degrees.

[0572] Furthermore, when the speed value generator 500 receives a marker signal P(i) having a position indication signal value (e.g., P(i)=1) every 360 / L degrees during the rotation of the housing 20, the speed value generator 500 will be able to generate an even more accurate speed value f ROT (j).

[0573] As for the appropriate setting of the FFT 510 when a signal with a marker value P(j)=1 is received every 360 / L degrees during the rotation of the housing 20, this means that the fundamental frequency will be the repetition frequency f R .

[0574] As above about Figure 2 As described above, the vibration signal S EA 、S MD , S(j), S(q) will show signal characteristics S FIMP , which indicates the impact of the protrusion with the toe 205, and when there are L protrusions 310 in the housing 20 (see Figure 23 Combined with the following formula 2), the signal characteristic S FIMP Will be repeated L times.

[0575] Referring again to the Fourier series (see Equation 2 below):

[0576]

[0577] in,

[0578] n=0 the average value of the signal over a period of time (can be zero, but does not have to be zero),

[0579] n=1 corresponds to the fundamental frequency of the signal F(t),

[0580] n=2 corresponds to the first harmonic part of the signal F(t),

[0581] ω = angular frequency of interest, i.e. (2*π*f R ),

[0582] f R = the frequency of interest, expressed in cycles per second,

[0583] t = time,

[0584] Ф n = the phase angle of the nth partial tone,

[0585] C n = the amplitude of the nth partial tone.

[0586] In this embodiment, it should be noted that when the FFT 510 receives a marker signal value P(j)=1 every 360 / L degrees during rotation of the housing 20 , the fundamental frequency will be one per protrusion 310 .

[0587] As mentioned above, the setting of FFT 510 should take into account the reference signal. As mentioned above, the position signals P(j), P(q) (see Figure 15A and / or Figure 15B ) can be used as a reference signal for the digital measurement signals S(j) and S(q).

[0588] According to some embodiments, when the FFT analyzer is configured to receive reference signals, i.e., position signals P(j), P(q), every 360 / L degrees during the rotation of the housing 20, and L is the number of protrusions 310 in the housing 20, the settings of the FFT analyzer should meet the following criteria:

[0589] The integer value Oi is set to one, i.e. equal to 1, and

[0590] The settable variables Y and Z are selected so that the mathematical expression Oi*Z / Y becomes a positive integer. In other words: when the integer value Oi is set equal to 1, then the settable variables Y and Z should be set to integer values ​​so that the variable X is a positive integer.

[0591] Where X = Oi*Z / Y

[0592] Using the above settings, the integer value Oi is set equal to 1, and referring to the above Figure 15A and / or Figure 15B Using Equation 2, FFT 510 can deliver the magnitude value C for n=1. n , that is, C1 = Sp(r). FFT 510 can also transmit the phase angle of the fundamental frequency (n=1), that is, Φ1 = FI(r).

[0593] Combine Figure 1A and / or Figure 1B and the above formula 2 reference Figure 15A and / or Figure 15B , the state values ​​Sp(r)=C1 and FI(r)=Φ1 can be transmitted to the human machine interface (HCI) 210 for providing a visual indication of the analysis results. As described above, the displayed analysis results can include information indicating the internal state of the tumbling process for enabling the operator 230 to control the tumbling machine.

[0594] refer to Figure 16 、 Figure 17 、 Figure 18 、 Figure 19A and Figure 19B , example illustrations of visual indications of analysis results are available for the configuration of a rotating mill housing 20, as shown Figure 23 As shown, the FFT 510 will receive marker signals P(i), P(j), P(q) having position indication signal values ​​every 360 / L degrees, where L is the number of protrusions 310 in the housing 20.

[0595] While the above discussion of the configuration of the FFT 510 involves Fourier series and Equations 1 and 2 for the purpose of conveying an intuitive understanding of the context of the configuration of the FFT transformer 510, it should be noted that the use of digital signal processing may involve a discrete Fourier transform (see Equation 3 below):

[0596] Formula 3:

[0597]

[0598] Thus, according to embodiments of the present disclosure, the aforementioned discrete Fourier transform (DFT) may be included in signal processing for generating data indicative of the internal state of the roller mill, such as discussed in conjunction with the embodiment of the state parameter extractor 450. In this regard, reference is made to, for example, Figure 3 、 Figure 4 、 Figure 5 , Figure 15 and / or Figure 24 In view of the above discussion of the subject of FFT and Fourier series, the Discrete Fourier Transform will not be discussed in further detail as it is well known to skilled readers of this disclosure.

[0599] although Figure 23It is shown that a plurality of position marks 180 may be provided on the outer surface of the housing 20, each mark 180 thereby causing the position sensor 170 to generate a rotational mark signal value Ps, but it should be noted that such a position signal may alternatively be generated by an encoder 170 mechanically coupled to the rotating mill housing 20. Thus, the position sensor 170 may be implemented by an encoder 170 that is mechanically coupled to the rotating mill housing 20 such that during rotation of the mill housing 20, the encoder generates, for example, one mark signal Ps on each protrusion 310 in the rotating housing 20.

[0600] In summary, regarding the appropriate settings of the FFT 510 and the above formulas 1 and 2, it should be noted that the phase angle of the nth partial tone (ie, Φ n ) can indicate the relative position of the toe 205. In particular, the phase angle of the nth partial (ie, Φ n ) can indicate the position of the toe 205, expressed as a fraction of the distance between two adjacent protrusions 310 in the rotating housing 20. Referring to Table 6 above and FIG. 14 , the total distance between two adjacent protrusions can be considered to be 360 ​​degrees, and the phase angle value of the nth partial tone (i.e., Φ n ) divided by 360 degrees can indicate a percentage of the total distance between two adjacent protrusions. This can be seen, for example, by comparing column #2 in Table 5 and Table 6 above. As described above, Φ n = Phase angle of the nth partial, C n = amplitude of the nth partial tone. As described above, taking into account the number L of protrusions in the rotating housing 20 and the number of reference signals generated and the resulting order Oi of the signal of interest, the FFT 510 can be set to transmit the phase angle Φ of the nth partial tone. n and the amplitude C of the nth partial n , so that the phase angle of the nth partial tone (ie Ф n ) may indicate the relative position of the toe 205. Additionally, as described above, the FFT 510 may be configured such that the variable X is a positive integer, where

[0601] X=Oi*Z / Y

[0602] And among them,

[0603] Oi is set to an integer value,

[0604] Y is set to an integer value,

[0605] Z is set to an integer value.

[0606] Figure 24A schematic top view of another system 700 including a roller mill 10 is shown. For example, the roller mill 10 can be an autogenous (AG) mill. Alternatively, the roller mill 10 can be a semi-autogenous (SAG) mill. Another example roller mill 10 is a ball mill 10. The roller mill 10 includes a housing 20 having an interior housing surface 22 that forms a chamber 25 for grinding material. Figure 24 The roller mill system 700 can be configured as described in any other embodiment described in this specification, for example, with respect to FIG. 1- Figure 31 However, despite Figure 1A and / or Figure 1B The roller mill system is described as having a vibration sensor 70 on the input side of the mill, but it should be noted that Figure 24 The tumble mill system 700 can be configured with:

[0607] For generating a first measurement signal S EAIN The first vibration sensor 70 IN ,as well as

[0608] For generating the second measurement signal S EAOUT The second vibration sensor 70 OUT .

[0609] By the first vibration sensor 70 IN The first measurement signal S generated EAIN The signal processing of the signal S may be as described in any other embodiment of the present disclosure. EA As described, for example, with respect to FIG1 above Figure 31 Similarly, the second vibration sensor 70 OUT The generated second measurement signal S EAOUT The signal processing of the signal S may be as described in any other embodiment of the present disclosure. EA As described, for example, with respect to FIG1 above Figure 31 Therefore, compared with the above embodiment, the difference is that, in the system 700, based on the first measurement signal S EAIN provides data indicative of the internal state of the input side of the roller mill and based on the second measurement signal S EAOUT Provides data indicating the internal state of the output side of the roller mill. Thus, regarding the provision of position signals or reference signals, Figure 24 The roller mill system 700 can be configured as described in any embodiment of the present disclosure.

[0610] Figure 24The illustrated analysis device 150 may include a first state parameter extractor 4501 and a second state parameter extractor 4502. The state parameter extractors 4501 and 4502 may operate as described in any other described embodiments, e.g., with reference to Figure 5 and / or FIG. 15 and / or as regards Figure 30-Figure 31 Therefore, the first state parameter extractor 4501 can be configured to generate the parameter S P1 (r), R T1 (r), f ROT (r), dS P1 (r) and dR T1 (r).

[0611] Similarly, the second state parameter extractor 4502 may be configured to generate the parameter S P2 (r), R T2 (r), f ROT (r), dS P2 (r) and dR T2 (r). However, the rotational speed of the housing f ROT (r) is of course the same, so if a state parameter extractor delivers the speed value f ROT (r) is enough.

[0612] refer to Figure 24 , shows a Cartesian coordinate system having three mutually perpendicular axes x, y and z. It will be appreciated that during operation of the mill 10, the material 30 travels in the positive direction of the x-axis from the input side 80 to the output side 90 of the mill.

[0613] Figure 24 The roller mill system 700 advantageously provides a parameter indicating the internal state of the input side of the roller mill: S P1 (r), R T1 (r), dS P1 (r) and dR T1 (r), and a parameter indicating the internal state of the output side of the roller mill: S P2 (r), R T2 (r), dS P2 (r) and dR T2 (r).

[0614] Comparison of input-side parameters with corresponding output-side parameters can advantageously add another dimension to the understanding of the internal state of the mill 10. For example, R T2 (r) and R T1 The relationship between (r) shows that:

[0615] - Is the toe position the same on the input and output sides, or

[0616] - Is the toe position higher on the input side when R T1 (r)>R T2 (r) when instructed; or

[0617] - Is the toe position higher on the output side when R T2 (r)>R T1 (r) indicates when

[0618] A higher toe position on the output side may indicate an incipient abnormality. For example, when the outflow of output material 95 decreases, perhaps due to a blockage, while the inflow of solid material 110 continues at an unreduced rate, the risk of overloading will increase, which may lead to a reduction in the efficiency of the grinding process in the roller mill. Therefore, Figure 24 The roller mill system 700 can advantageously provide early indication of incipient anomalies. Thus, based on a comparison of input-side parameters with corresponding output-side parameters, the roller mill system 700 can adjust control parameters to avoid anomalies, such as mill overload.

[0619] refer to Figure 24 It should be noted that the vibration sensor 70 OUT connected to the non-rotating portion of the body of the mill structure 10 and the vibration sensor 70 OUT Positioned to detect vibrations primarily in the horizontal direction Y (see a Cartesian coordinate system with three mutually perpendicular axes x, y and z, where Y is the horizontal direction). Likewise, the vibration sensor 70 IN connected to the non-rotating portion of the body of the mill structure 10 and the vibration sensor 70 IN Positioned to detect vibrations primarily in the horizontal direction Y. Experimental measurements appear to indicate that improved vibration signal quality is obtained when the vibration sensor is configured to detect vibrations primarily in the horizontal direction Y, compared to the vibration signal quality obtained when the vibration sensor is configured to detect vibrations primarily in the vertical direction Z. As described above, for example, in combination Figure 2 , the protrusion 310 interacts with the toe 205 of the charge, forcing the material in the toe of the charge to accelerate in the direction of motion of the protrusion 310, as shown Figure 2 As shown, the mechanical vibration V IMP The impact of the protrusion 310C on the mass of material in the toe 205 causes the mass of material in the toe to be moved at the time of the motion of the protrusion 310C. ACC The acceleration in the direction of the protrusion 310C causes a force F on the leading edge surface of the protrusion 310C. IMP Since the weight of the solid material in the charge 30 of the mill is in metric tons, the impact force F IMPHowever, since the mill structure will typically rest on a very hard floor surface which tends to attenuate vibrations in the vertical direction, it would appear that vibration detection in the horizontal direction Y provides improved vibration signal quality.

[0620] Figure 25 A schematic top view of yet another embodiment of a system 720 including a roller mill 10 is shown.

[0621] Figure 25 The roller mill system 720 can be combined with Figure 24 However, despite Figure 24 The roller mill system 700 is depicted as having a vibration sensor 70 connected to a non-rotating portion of the body of the grinding structure 10. OUT and a vibration sensor 70 connected to another non-rotating portion of the body of the grinding structure 10 IN ,but Figure 25 The roller mill system 720 is different in that it provides a vibration sensor 70 connected to the rotating shell 20 of the grinding structure 10. 20 .like Figure 25 As shown, a vibration sensor 70 is directly provided on the rotating housing 20. 20 will generate high amplitude, especially when the vibration sensor 70 20 Located on the exterior of the housing, directly on the side of the housing wall opposite the protrusion 310 .

[0622] Figure 25 The tumbler system 720 may optionally include:

[0623] For generating a first measurement signal S EAIN The first vibration sensor 70 20IN ,as well as

[0624] For generating the second measurement signal S EAOUT The second vibration sensor 70 20OUT .like Figure 25 As shown, the first vibration sensor 70 20IN The measuring point position 310 may be closer to the input side 80 than the output side 90. IN The second vibration sensor 70 is firmly connected to the outer surface of the housing 20. 20OUT The measuring point position 310 may be closer to the output side 90 than the input side 80. OUT The outer surface of the housing 20 is firmly connected to the housing 20.

[0625] First vibration sensor 70 20IN and the second vibration sensor 70 20OUTThe sensors 70 on the outer surface of the housing 20 may be configured to communicate wirelessly with the device 150, for example via transceiver units 740 and 750, respectively. 20 , 70 20IN , 70 20OUT The power supply may be via a battery, or alternatively, via an induction device (not shown) attached to the outer surface of the rotating housing 20, which operates as a generator by interacting with one or more fixed permanent magnets. In this manner, as the housing 20 rotates, the induction device will repeatedly pass through the magnetic field of the fixed one or more permanent magnets, thereby inducing a magnetic field that can be used as a sensor 70. 20 , 70 20IN , 70 20OUT The current of the power supply.

[0626] Figure 25 The roller mill system 720 can also advantageously provide a parameter S indicating the internal state of the roller mill input side. P1 (r), R T1 (r), dS P1 (r) and dR T1 (r) and the parameter S indicating the internal state of the output side of the roller mill P2 (r), R T2 (r), dS P2 (r) and dR T2 (r). Therefore, it is straightforward and unambiguous for the skilled reader of this disclosure to infer that Figure 25 The roller mill system 720 can advantageously be operated in a manner substantially similar to Figure 24 The roller mill system 700 provides for early indication of incipient abnormalities. In particular, Figure 25 The roller mill system 720 can advantageously implement a comparison of input side parameters with corresponding output side parameters in the manner described above with respect to the roller mill system 700. Figure 25 The roller mill system 720 can also advantageously enable adjustment of control parameters to avoid abnormalities such as mill overload.

[0627] Figure 26 A schematic top view and a diagrammatic top view of another embodiment of a system 730 including a roller mill 10 are shown. For example, the roller mill 10 can be an autogenous (AG) mill. Alternatively, the roller mill 10 can be, for example, a semi-autogenous (SAG) mill. Another example of a roller mill 10 is a ball mill 10. The roller mill 10 includes a housing 20 having an interior housing surface 22 that forms a chamber 25 for grinding material. Figure 26 The roller mill system 730 may include the components described in any other embodiments described in the present disclosure and be configured, for example, with respect to FIG. Figure 25 and / or Figure 30-Figure 31described. In particular, Figure 26 The device 150 shown in FIG. 1 may be configured as described in any other embodiment described in this disclosure, for example, with respect to FIG. Figure 25 and / or Figure 30-Figure 31 described.

[0628] However, in Figure 26 In the embodiment of the system 730 shown, the device 150 includes a monitoring module 150A and a control module 150B. Although the figures illustrate the device 150 as two blocks, it should be understood that the device 150 can also be provided as a single entity 150 including the monitoring module 150A and the control module 150B, as shown by the unified reference numeral 150.

[0629] The system 730 is configured to control the internal state of the roller mill 10, which has a certain rotation speed f ROT The housing 20 rotates about an axis 60 for grinding a charge of material 30 by tumbling the material within the rotating housing.

[0630] The housing 20 has an interior housing surface 22 that includes a first number L of protrusions 310 configured to engage material when the housing 20 rotates about the axis 60. The system 730 may include devices 170, 180 for generating a position signal. The devices 170, 180 may include a position sensor 170 and a marker 180, as described elsewhere in this disclosure. The position signal is E P , P(i), P(j), P(q), indicating the rotational position of the rotating shell 20, and the position signal includes a time series of position signal sample values ​​P(i), P(j), P(q).

[0631] Provide sensors 70, 70 IN , 70 OUT , 330, and is configured to generate a mechanical vibration V according to the rotation of the housing IMP To generate the vibration signal S EA 、S MD , Se(i), S(j), S(q). Vibration signal S EA , Se(i), S(j), S(q) may include a time series of vibration sample values ​​Se(i), S(j), S(q).

[0632] The device 150 of the system 730 may include a monitoring module 150A and a control module 150B. The monitoring module 150A includes a state parameter extractor 450, 4501, 4502, 450C, which is configured to detect a first occurrence of a first reference position signal value in the time series of position signal sample values ​​P(i), P(j), P(q) (see Tables 2, 3, and 4 above, where column #2 shows position signals with values ​​1, 1C).

[0633] The state parameter extractor 450 may be configured to detect a second occurrence of the second reference position signal value 1, 1C, 100% in the time series of the position signal sample values ​​P(i), P(j), P(q). The state parameter extractor 450 may also be configured to detect an event feature S in the time series of the vibration sample values ​​Se(i), S(j), S(q). P (r), Sp occurs. This event may be caused by the protrusion 310 impacting the toe 205 of the charge 30, causing an impact vibration that may cause a vibration signal characteristic, as discussed elsewhere in this disclosure. The state parameter extractor 450 may be configured to generate an indication of a first temporal relationship R between T (r), T D , FI(r), X1(r) data:

[0634] Event characteristics appear, and

[0635] First appearance and second appearance.

[0636] As described above, the system 730 includes a control module 150B configured to receive data indicative of the internal state of the mill 10 from the mill monitoring modules 150, 150A. The data indicative of the internal state may include any information generated or transmitted by the state parameter extractor 450, as described in this disclosure with respect to FIG. Figure 31 Any of the ones described in . Figure 26 , the control module 150B includes a regulator 755 for controlling the angular toe position FI(r), A based on the following TOE (See Figure 26 Combine Figure 2 ):

[0637] Toe position reference value FI REF (r)(See Figure 26 ),

[0638] The first time relationship R T (r), T D , FI(r), X1(r) (see Figure 1- Figure 31 ),as well as

[0639] Toe position error value FI ERR (r)(See Figure 26 ).

[0640] Toe position error value (FI ERR (r)) depends on the toe position reference value FI REF (r) and the first time relationship R T (r), T D , FI(r)(see Figure 3-Figure 26 ). Toe position reference value FI REF (r) can be entered manually ( Figure 26 ) is generated, but can be combined as above, for example Figure 1A and / or Figure 1B Complete as discussed.

[0641] like Figure 26 As shown, the toe position error value (FI ERR (r)) may depend on the toe position reference value FI REF (r) and the first time relation R T (r), T D , FI(r), and the difference between X1(r).

[0642] The regulator 755 may be configured to adjust the toe position reference value FI according to the REF (r) to control the solid material feed rate set point R SSP . Combined Figure 1A The solid material feed rate R S Depends on the solid material feed rate set point R SSP (See Figure 26 ). If combined Figure 1A As mentioned above, the solid material feeding rate R S is the amount of solid material fed to the input end 100 of the roller mill 10 per unit time.

[0643] The regulator can also be configured to adjust the toe position reference value FI REF (r) to control the liquid feed rate set point R LSP Liquid feed rate R L The liquid feed rate set point R LSP If combined with Figure 1A As mentioned above, the liquid feed rate R L It may be the amount of liquid fed to the input 130 of the roller mill 10 per unit time.

[0644] The event signature may indicate the impact force F generated when the protrusion 310 on the inner shell surface 22 of the rotating shell 20 interacts with the toe 205 of the charge material 30.IMP .

[0645] The state parameter extractor 450 may be configured to generate the first time relationship R T (r), T D , FI(r), X1(r), as the phase angle FI(r).

[0646] First time relationship R T (r), T D , FI(r), X1(r) indicate the toe position 205, A TOE (r)(See Figure 2 Combine Figure 26 ). First time relationship R T (r), T D , FI(r), X1(r) may indicate the ratio of the distance between two adjacent protrusions 310 in the mill housing.

[0647] Alternatively, the relation value X1 (r) may indicate a relative position of the toe 205 , ie, a position of the toe 205 relative to two predetermined stator positions that are separated from each other in a manner corresponding to the positions of two adjacent protrusions 310 .

[0648] Furthermore, the relational value X1(r) may indicate the absolute toe position.

[0649] The absolute toe position, also referred to as Sp(r), may be generated based on the combination of the relation value X1(r) and the second internal state parameter X2(r). In particular, the absolute toe position may be generated based on the time series of the combination of the first internal state parameter X1(r) and the second internal state parameter X2(r), for example, in combination with Figure 16 、 Figure 17 、 Figure 18 、 Figure 19A and Figure 19B As discussed. Figure 19A As shown, the initial data set 550(r) = 550(1) corresponding to X1(r = 1) and X2(r = 1) represents an empty or nearly empty condition of the mill. As the filling degree of the mill increases, the absolute toe position increases in the direction of arrow 560A. Figure 19A and Figure 19B From the initial angle FI TR The "angle length" to the current indicator 550(r) indicates the absolute toe position. Figure 2 refer to Figure 19B , Figure 2 A toe angle of 0 degrees at the mid-vertical line 960 may correspond to Figure 19B The initial angle FI in TR As r increases, the value of X2(r) increases. Figure 19B, and the weight of the indicator toe 205 increases, as discussed elsewhere in this document. When the value of the parameter X1(r) increases to a value equal to 360 degrees, the indicator toe position 205 is at an angle FI from the initial angle FI. TR The angle distance is 360 / L.

[0650] When the value of the parameter X1(r) increases to a value exceeding 360 degrees, it indicates that the absolute toe position 205 is located at a distance from the initial angle FI TR The angle distance is 360 / L+X1(r).

[0651] The state parameter extractor 450 may be configured to generate the event signature as an amplitude value S P (r), Sp, C L (r), C1(r), X2(r).

[0652] The state parameter extractor 450 may include a device configured to generate the first time relationship R T (r), T D , FI(r), X1(r) Fourier transformer 510 (see Figure 15A and / or Figure 15B ).

[0653] As discussed in conjunction with Table 5, the state parameter extractor 450 may be configured to extract the total number of samples N from the first occurrence to the second occurrence. B Furthermore, the state parameter extractor 450 may be configured to count another number of samples N from the first occurrence to the occurrence of the event. P The state parameter extractor 450 may be configured to generate the first time relationship R based on the other number and the total number. T (r), T D , FI(r), X1(r).

[0654] The state parameter extractor 450 may be configured to extract the total number of samples N from the first occurrence to the second occurrence. B The state parameter extractor 450 may be configured to count another number of samples N from the first occurrence to the occurrence of the event. P Furthermore, the state parameter extractor 450 may be configured to generate the first time relationship R based on the relationship between the other number and the total number. T (r), T D , FI(r), wherein the relationship between the further quantity and the total number may be indicative of a toe position 205 .

[0655] According to one embodiment, the regulator 755 operates to control the input side toe position (FI IN (r), A TOE_IN ):

[0656] Toe position reference value (FI REFIN (r)),

[0657] The first time relationship (FI IN (r)), and

[0658] Toe position error value (FI ERRI (r)), where

[0659] The toe position error value (FI ERRI (r)) depends on the following:

[0660] The toe position reference value (FI REFIN (r)), and

[0661] The first time relationship (R T (r); T D ;FI IN (r)).

[0662] According to one embodiment, the regulator 755 operates to adjust the toe position reference value FI REFIN (r) Control the solid material feed rate set point R SSP .

[0663] According to one embodiment, the regulator 755 operates to control the solid material feed rate set point R based on SSP :

[0664] Toe position reference value (FI REFIN (r)),

[0665] The first time relationship (FI IN (r)), and

[0666] Toe position error value (FI ERRI (r)), where

[0667] The toe position error value (FI ERRI (r)) depends on the following:

[0668] The toe position reference value (FI REFIN (r)), and

[0669] The first time relationship (FI IN (r)).

[0670] although Figure 26 、 Figure 28 and Figure 29 Two feedback signals are shown, namely a first time relationship FI related to the input side of the mill 10 IN (r) and a first time relationship FI associated with the output side of the mill 10 OUT (r), but it will be appreciated that the system may be operated with a single feedback signal. Thus, for example, the regulator may be provided with a single input, for example for receiving a first time relation FI relating to the state of the input side of the mill IN (r).

[0671] According to one embodiment, the regulator 755 operates to control the solid material feed rate set point R based on SSP :

[0672] Toe position reference value (FI REFOUT (r)),

[0673] The first time relationship (FI OUT (r)), and

[0674] Toe position error value (FI ERRO (r)), where

[0675] The toe position error value (FI ERRO (r)) depends on the following:

[0676] The toe position reference value (FI REFOUT (r)), and

[0677] The first time relationship (FI OUT (r)).

[0678] The regulator 755 may be configured to include a proportional-integral-derivative controller (PID controller). Alternatively, the regulator 755 may be configured to include a proportional-integral controller (PI controller). Alternatively, the regulator 755 may be configured to include a proportional controller (P controller).

[0679] Alternatively, the regulator 755 can be configured to include a Kalman filter, also known as a linear quadratic estimator (LQE). A Kalman filter is an algorithm that uses a series of measurements observed over time, including statistical noise and other inaccuracies, and produces estimates of unknown variables by estimating the joint probability distribution over the variables at each time period. These estimates tend to be more accurate than estimates based on a single measurement alone.

[0680] Figure 27A schematic block diagram of a distributed process monitoring system 770 is shown. Reference numeral 780 relates to a client location having a mill 10 with a rotatable housing 20, as discussed above with respect to the previous figures in this document. Client location 780, which may also be referred to as client part or mill location 780, may be, for example, a mining company's site, or the site of an ore mill, or a manufacturing mill, such as for cement production.

[0681] When one sensor 70 or several sensors 70, 70 IN , 70 OUT The distributed process monitoring system 770 is operable when attached to or at a measuring point associated with the housing 20. As described above, such a measuring point may be, for example, at the bearings 40, 50 (see Figure 26 and Figure 27 ) or at the measuring point position 310 IN 、310 OUT (See Figure 25 ).

[0682] Measurement signal S EA 、S EAIN 、S EAOUT and E P (See, for example, FIG. 1 , Figure 27 、 Figure 26 、 Figure 25 ) can be coupled to an input port of the mill position communication device 790. The mill position communication device 790 may include a device for measuring the signal S EA 、S EAIN 、S EAOUT and E P The A / D converter 975 may operate as disclosed elsewhere in this document with respect to the A / D converter 330, for example, in combination with Figure 3 and Figure 5 The mill position communication device 790 has a communication port 800 for bidirectional data exchange. The communication port 800 can be connected to a communication network 810, for example, via a data interface 820, for implementing the measurement signal S EA 、S EAIN 、S EAOUT and E P The communication network 810 may be the World Wide Web, also known as the Internet. The communication network 810 may also include a public switched telephone network.

[0683] A server computer 830 is connected to the communication network 810. The server 830 may include a database 840, a user input / output interface 850, data processing hardware 852, and a communication port 855. The server computer 830 is located at a server location 860, which is geographically separated from the mill location 780. The server location 860 may be in a first city, such as Stockholm, Sweden, while the mill location 780 may be in a rural area near the mill and / or in another country, such as Norway, Australia, or the United States. Alternatively, the server location 860 may be in a first part of a country, while the mill location 780 may be in another part of the same country. The server location 860 may also be referred to as a supplier component 860 or a supplier location 860.

[0684] According to one example, the central control location 870 includes a monitoring computer 880 with data processing hardware and software for monitoring and / or controlling the internal conditions of the mill 10 at the remote mill location 780. The monitoring computer 880 may also be referred to as a control computer 880. The control computer 880 may include a database 890, a user input / output interface 900, data processing hardware 910, and communication ports 920, 920A, or several communication ports 920, 920A, 920B. The central control location 870 may be geographically separated from the mill location 780. The central control location 870 may be located in a first city, such as Stockholm, Sweden, while the mill location 780 may be located in a rural area near the mill and / or in another country, such as Norway, Australia, or the United States. Alternatively, the central control location 870 may be located in a first part of a country, while the mill location 780 may be located in another part of the same country. The control computer 880 may be coupled to communicate with the mill location communication device 790 via the communication ports 920, 920A. Thus, the control computer 880 can receive the measurement signal S from the mill position communication device 790 via the communication network 810. EA 、S EAIN 、S EAOUT and E P (See, for example, FIG. 1 , Figure 27 、 Figure 26 、 Figure 25 ).

[0685] The system 770 may be configured to receive the measurement signal S in real time or substantially real time. EA 、S EAIN 、S EAOUT and E P , or the mill 10 can be monitored and / or controlled in real time from location 870. Furthermore, the control computer 880 can include a monitoring module 150, 150A as disclosed in any example of this document, for example, as disclosed in connection with any of the above Figures 1-26.

[0686] The supplier company may occupy the server location 860. The supplier company may sell and deliver the devices 150 and / or monitoring modules 150A and / or software for such devices 150 and / or monitoring modules 150A. Thus, the supplier company may sell and deliver software for the control computer 880 at the central control location 870. Such software 370, 390, 400 may be combined, for example, with Figure 4 Such software 370, 390, 400 may be delivered by transmission over the communications network 810. Alternatively, such software 370, 390, 400 may be delivered as a computer-readable medium 360 for storing program code. Thus, a computer program 370, 390, 400 may be provided as an article of manufacture comprising a computer storage medium having encoded therein a computer program.

[0687] According to an exemplary embodiment of the system 770, the monitoring computer 880 may substantially continuously receive the measurement signal S from the mill position communication device 790, for example via the communication network 810. EA 、S EAIN 、S EAOUT and E P (See, for example, FIG. 1 , Figure 27 、 Figure 26 、 Figure 25 ) to enable continuous or substantially continuous monitoring of the internal conditions of the mill 10. The user input / output interface 900 at the central control location 870 may include a screen 900S for displaying images and data, as discussed in conjunction with the HCI 210 elsewhere in this document. Thus, the user input / output interface 900 may include a display or screen 900S, 210S for providing a visual indication of the results of the analysis. The displayed analysis results may include information indicating the internal conditions of the tumbling process, enabling an operator 930 at the central control location 870 to control the tumbling mill 10.

[0688] In addition, the monitoring computer 880 at the central control location 870 can be configured to transmit information indicating the internal status of the tumbling process to the HCI 210 via the communication ports 920, 920B and via the communication network 810. In this manner, the monitoring computer 880 at the central control location 870 can be configured to enable the operator 230 at the client location 780 to control the tumbling machine. The local operator 230 at the client location 780 can be placed in the control room 220 (see Figure 1A and / or Figure 1B and / or Figure 27Thus, the client location 780, 220 may include a second mill location communication device 790B. The second mill location communication device 790B has a communication port 800B for bidirectional data exchange, and the communication port 800B may be connected to the communication network 810, for example, via a data interface 820B.

[0689] Although two location communication devices 790, 790B have been described for clarity, alternatively a single mill location communication device 790, 790B and / or a single communication port 800, 800B may be provided for bidirectional data exchange. Thus, items 790 and 790B may be integrated into a single unit at the mill location 780, and similarly, items 820 and 820B may be integrated into a single unit at the mill location 780.

[0690] Figure 28 Shown is a schematic block diagram of yet another embodiment of a distributed process monitoring system 940. Reference numeral 780 relates to the mill location of a mill 10 having a rotatable housing 20 as discussed above with respect to the previous figures in this document. Figure 28 The distributed process monitoring system 940 may include components and be configured as described in any other embodiment described in this disclosure, for example, with respect to FIG. Figure 31 Specifically, Figure 28 The monitoring device 150 shown, also referred to as monitoring module 150A, can be configured as described in any other embodiment described in this disclosure, for example, with respect to FIG. Figure 31 Specifically, Figure 28 The process monitoring system 940 shown may be configured to include a monitoring module 150A, such as in conjunction with Figure 27 disclosed, but located in a central control location 870.

[0691] In addition, Figure 28 In the process monitoring system 940 shown, the mill location 780 includes a control module 150B, as described above in conjunction with Figure 26 described.

[0692] Thus, the internal state of the mill 10 can be automatically controlled by the control module 150B located at or near the mill location 780, and the monitoring computer 880 at the central control location 870 can be configured to transmit information indicating the internal state of the rolling mill process to the HCI 900, 900S to enable the operator 930 at the central control location 870 to monitor the internal state of the rolling mill 10.

[0693] Measurement signal S EA 、S EAIN 、S EAOUT and EP (See, for example, FIG. 1 , Figure 27 、 Figure 26 、 Figure 25 ) can be coupled to an input port of the mill position communication device 790. The mill position communication device 790 may include a device for measuring the signal S EA 、S EAIN 、S EAOUT and E P The A / D converter 975 may operate as disclosed elsewhere in this document with respect to the A / D converter 330, for example, in combination with Figure 3 and Figure 5 The mill position communication device 790 has a communication port 800 for bidirectional data exchange. The communication port 800 can be connected to a communication network 810, for example, via a data interface 820. The communication port 800 can be connected to a communication network 810, for example, via a data interface 820, for implementing a communication corresponding to the measurement signal S EA 、S EAIN 、S EAOUT and E P transmission of digital data.

[0694] Furthermore, the client location 780 may include a second mill location communication device 790B having a communication port 800B for bidirectional data exchange, and the communication port 800B may be connected to a communication network 810, for example via a data interface 820B, to enable the control module 150B to receive data indicating the internal state of the mill 10.

[0695] like Figure 28 As shown, data indicative of the internal state of the mill 10 may be generated by the monitoring module 150A located at a central location 870 .

[0696] Although for the sake of clarity, Figure 28 Two location communication devices 790, 790B are described, but alternatively, a single mill location communication device 790, 790B and / or a single communication port 800, 800B may be provided for bidirectional data exchange. Thus, items 790 and 790B may be integrated into one unit at mill location 780, and similarly, items 820 and 820B may be integrated into one unit at mill location 780.

[0697] Figure 29 Shown is a schematic block diagram of yet another embodiment of a distributed process control system 950. Again, reference numeral 780 relates to the mill location of a mill 10 having a rotatable housing 20 as discussed above with respect to previous figures in this document. Figure 29The distributed process monitoring system 950 may be componentized and configured as described in any other embodiment described in this disclosure, for example, with respect to FIG. Figure 31 Specifically, Figure 28 and Figure 29 The monitoring device 150 shown, also referred to as monitoring module 150A, may be configured as described in any other embodiment described in this disclosure, for example, as described with respect to FIG. Figure 31 In addition, Figure 29 The process monitoring system 950 shown may be configured to include the above-mentioned Figure 26 The control module 150B described and as combined Figure 27 The disclosed monitoring module 150A.

[0698] exist Figure 29 In the example of FIG, the monitoring module 150A and the control module 150B are disposed at a control location 870. The control location 870 may be remote from the mill location 780. Data communication between the control location 870 and the mill location 780 may be provided via data ports 820 and 920 and the communication network 810, as discussed above in conjunction with the previous figures.

[0699] Figure 30 Another example of a cross-sectional view of the middle portion 98 of the rotating mill housing 20 during operation is shown. This view may be taken, for example, along Figure 1A According to the line AA. Figure 30 In the example of FIG. 5 , the roller mill housing 20 has four protrusions 310 configured to engage the charge material 30 when the housing rotates about the axis 60 , ie, the number L=4. For clarity, Figure 23 The protrusions in the example are referred to as 3101, 3102, 3103 and 3104 respectively.

[0700] A position sensor 170 is provided to generate a position signal Ep according to the rotational position of the housing 20. Figure 30 As shown, the position sensor 170 is positioned along a vertical line 960 of the rotation axis 60, and the housing 20 is rotatable around the rotation axis. In addition, the position mark 180 is provided on the outer surface of the housing 20 so that when the protrusion 310 passes through the vertical line 960, the position sensor 170 generates a mark signal P S When the roller mill housing 20 has four protrusions 310, as shown in FIG. Figure 30As shown, and when a single static position sensor 170 is provided, a method according to one example may receive a first static position signal Ps1, P1 in response to a first channel of the position marker 180, and the method may receive a second static position signal Ps2 in response to a second channel of the position marker 180, and then the method may generate a virtual static position signal P in the recorded time series of position signal sample values. C When the protrusions are evenly distributed on the inner circumference of the housing 20, the generated virtual static position signal P is inserted C , so as to be uniformly distributed in the recorded time series of position signal sample values. This advantageously results in a recorded time series of position signal sample values ​​which indicates uniformly distributed L static positions, such as Figure 30 As shown. Figure 30 In the example of FIG, there are four protrusions, ie L=4, so the recorded time series of position signal sample values ​​indicates four static positions P1, P2, P3 and P4=PL. L static position signals P S 、P C It can be used as a reference position signal thereafter. The event signal feature will also appear L times per revolution in the time series of the vibration signal sample values ​​and can be compared to a static position signal P S 、P C Or two static position signals P S 、P C To analyze the occurrence of event signal characteristics.

[0701] As mentioned above, combined Figure 19A and Figure 19B It has been observed that when the roller mill is started from a no-load state, the initial internal state indicator object appears at an initial polar angle Φ(1) which represents the first detected toe position 205 of the mill. Based on experimental measurements, it appears that the initial polar angle Φ(1) can be used as a reference toe position value. Therefore, the initial polar angle Φ(1) can be referred to as the reference toe position value Φ TR . For its internal state is Figure 19A and Figure 19B The particular roller mill represented by the display 210S shown, the reference toe position corresponds to an angle value φ of approximately 47 degrees. TR ,like Figure 19A and Figure 19B Reference Figure 2 14, it is believed that if the position marker 180 is physically moved to a different position in terms of angular position, the reference toe position value Φ TR The angle value of will change to a numerically different angle value.

[0702] refer to Figure 30It is believed that the position mark 180 is provided on the outer surface of the housing 20 so that when the protrusion 310 passes the vertical line 960, the position sensor 170 generates a rotation mark signal value P S , this signal value will result in a reference toe position value Ф TR has a very small value or zero value, because the initial polar angle Φ(1) = Φ TR Indicates the toe position first detected when starting the roller mill from a no-load condition.

[0703] In this regard, when the toe is at the lowest portion of the housing 20, at a position indicated by a vertical line 960 from the rotational axis 60, a zero degree absolute toe position X6 may be represented (see FIG. Figure 30 and Figure 2 ).

[0704] refer to Figure 30 、 Figure 2 and Figure 14A and Figure 14B It is believed that if the position marker 180 is physically moved to a different position in terms of angular position, the reference toe position value Φ TR The angle value of will change to a numerically different angle value.

[0705] Figure 31 is a block diagram illustrating another example of a state parameter extractor 450 (referred to as state parameter extractor 450C). As described below, state parameter extractor 450C may include a vibration event feature detector, a position signal value detector, and a relationship generator. As described below, the vibration event feature detector may be implemented by a peak detector.

[0706] According to aspects of the solution disclosed in this document, reference position signal values ​​Ep, I, 1C are generated at L predetermined rotational positions of the rotatable housing 20, the L predetermined rotational positions following a pattern reflecting the angular positions of the L protrusions 310 in the housing 20. Providing such reference position signal values ​​Ep, I, 1C, and providing vibration event signature detection in the manner disclosed herein, enables data indicative of the position of the toe 205 to be generated in an advantageously accurate manner.

[0707] Although the example has been described with the protrusions 310 positioned in an equidistant pattern, i.e., evenly distributed in the housing 20, this solution is also applicable to other patterns of angular positions of the L protrusions 310 in the housing 20. When other patterns of angular positions of the L protrusions 310 in the housing are used, it is important that the reference position signal values ​​Ep, I, 1C are generated at L predetermined rotational positions of the rotatable housing 20, the L predetermined rotational positions following a pattern that reflects the angular positions of the L protrusions 310 in the housing 20.

[0708] refer to Figure 5, the A / D converter 330 may be configured to transmit a series of vibration measurement value pairs S(i) associated with corresponding position signal values ​​P(i) to the state parameter extractor 450 .

[0709] Figure 31 The state parameter extractor 450C is adapted to receive the measurement value sequence S(i) and the position signal sequence P(i) as well as the time relationship between them.

[0710] Thus, a single measurement value S(i) is associated with a corresponding position value P(i). This signal pair S(i) and P(i) is transferred to the memory 970. Figure 31 , the state parameter extractor 450C includes a memory 970.

[0711] Memory 970 is operable to receive data in the form of signal pairs S(i) and P(i) to enable analysis of the temporal relationship between occurrences of events in the received signals. Columns #2 and #3 in Table 3 provide an example of data collected in memory 970 during one complete rotation of the housing when position signals 1, 1C are provided six times per rotation, since there are L=6 protrusions 310 in housing 20. Tables 4 and 5 provide more detailed information regarding example signal values ​​in the first 1280 time slots of Table 3.

[0712] The position signals 1, 1C may be generated by a physical marking device 180, and / or some position signals 1C may be virtual position signals. The time series of position signal sample values ​​P(i), P(j), P(q)) should be provided in an occurrence pattern that reflects the angular position of the protrusion 310 in the housing 20.

[0713] For example, when there are six (L = 6) equally spaced protrusions 310 in the housing 20, the angular distance between any two adjacent protrusions 310 is 60 degrees. This is because 360 ​​degrees is a full circle, and when L = 6, the angular distance between any two adjacent protrusions is 360 / L = 360 / 6 = 60. Therefore, the corresponding time series of position signal sample values ​​P(i) representing a full circle of the housing 20 should include six (L = 6) position signal values ​​1, 1C with corresponding occurrence patterns, as shown in Table 3.

[0714] The state parameter extractor 450C further includes a position signal value detector 980 and a vibration event feature detector 990. The vibration event feature detector 990 may be configured to detect vibration signal events, such as amplitude peaks in the received sequence of measurement values ​​S(i).

[0715] The output of the position signal value detector 980 is coupled to the start / stop input 995 of the reference signal time counter 1010 and the start input 1015 of the event signature time counter 1020. The output of the position signal value detector 980 can also be coupled to the start / stop input 1023 of the vibration event signature detector 990 to indicate the start and stop of the duration to be analyzed. When the position signal value 1, 1C is detected, the detector 990 sends it at its output.

[0716] The vibration event signature detector 990 is configured to analyze all sample values ​​S(i) between two consecutive position signal values ​​1, 1C to detect the highest peak amplitude value Sp therein. The vibration event signature detector 990 has a first output 1021 coupled to a stop input 1025 of an event signature time counter 1020.

[0717] The reference signal time counter 1010 is configured to count the duration between two consecutive position signal values ​​1, 1C, thereby generating a first reference duration value T at the output terminal 1030. REF1 This can be achieved, for example, by means of a reference signal time counter 1010, which is a clock timer that counts the duration between two consecutive position signal values ​​1, 1C. Figure 14B , the first reference duration value T REF1 In this way, the time duration between the static position signal P4 and the static position signal P5 can be indicated.

[0718] Alternatively, the reference signal time counter 1010 may count the number of time slots between two consecutive position signal values ​​1, 1C (see column #01 in Table 3).

[0719] The event signature time counter 1020 is configured to count the duration from the occurrence of the position signal value 1, 1C to the occurrence of a vibration signal event (such as an amplitude peak). This can be achieved by:

[0720] When information is received at the start input 1015 that the position signal value detector 980 has detected the occurrence of the position signal value 1, 1C, the event signature time counter 1020 starts counting.

[0721] - When receiving information on the stop input 1025 that the vibration event signature detector 990 detects a vibration signal event, such as an amplitude peak in the received sequence of measurement values ​​S(i), the event signature time counter 1020 stops counting.

[0722] In this manner, the event feature time counter 1020 can be configured to count the time duration from the occurrence of the position signal value 1, 1C to the occurrence of the amplitude peak. The time duration from the occurrence of the position signal value 1, 1C to the occurrence of the amplitude peak is referred to herein as the second reference duration value T REF2 The second reference duration value T REF2 Can be transmitted on output 1040. Reference Figure 14B , the second reference duration value T REF2 In this way, the time duration between the occurrence of the static position signal P4 and the occurrence of the amplitude peak can be indicated.

[0723] refer to Figure 31 The output terminal 1040 is coupled to the input terminal of the relationship generator 1050 so as to provide the second reference duration value T to the relationship generator 1050. REF2 .

[0724] The relationship generator 1050 is further coupled to receive the first reference duration value T from the output terminal 1030 of the reference signal time counter 1010. REF1 The relationship generator 1050 is configured to generate a second reference duration value T based on the received second reference duration value T REF2 and the received first reference duration value T REF1 Generate the relationship value X1. The relationship value X1 can also be called R T (r), T D , FI(r). The housing 20 may generate L relationship values ​​X1 per one rotation. Furthermore, the L relationship values ​​X1 generated from a single rotation of the housing may be averaged to generate one value X1(r) per rotation of the housing 20. In this manner, the state parameter extractor 450C may be configured to transmit an updated value X1(r) per rotation.

[0725] For the sake of clarity, an example of a relationship value X1 is generated as follows: Figure 31 Refer to column #03 in Table 4: The vibration sample values ​​S(i) are analyzed by the vibration event feature detector 990 to detect the vibration signal feature S FIMP .

[0726] Vibration signal characteristics S FIMP It can be expressed as a peak amplitude sample value Sp. Referring to Table 5, the peak analysis results in the detection of the highest vibration sample amplitude value S(i). In the example shown, the vibration sample amplitude value S(i=760) is detected as holding the highest peak value Sp.

[0727] Having detected that the peak Sp is located in the time slot 760, a time relationship value X1 can be established.

[0728] In Table 5, in the time series of position signal sample values ​​P(i), time slots carrying position signal values ​​1, 1C are indicated as 0% and 100%, respectively.

[0729] As shown in the example in column #02 of Table 5, the time position of slot number i = 760 is the 59% position of the time distance between slot i = 0 and slot i = 1280. In other words, 760 / 1280 = 0.59 = 59%.

[0730] Therefore, the position of the toe 205 is expressed as a percentage of the distance between two adjacent static positions PC (see Table 5 for details). Figure 14B The static positions P4 and P5 in the equation can be obtained by:

[0731] For the first reference signal from sample number N0=0 to sample number N B = the total number of samples where the second reference signal appears in 1280 (N B –N0=N B –0=N B =1280) for counting, and

[0732] From the first reference signal appearing at N0=0 to the peak amplitude value Sp at sample number N P The number of samples that appear at the other location (N P –N0=N P –0=N P ) to count, and

[0733] Based on the other number N P and the total number N B , generate the first time relationship (X1, R T (r); T D ; FI(r)). This can be summarized as:

[0734] X1(r)=R T (r)=R T (760)=(N P –N0) / (N B –N0)=(760-0) / (1280-0)=0.59=59%.

[0735] Therefore, the relative toe position X1, R T It can be generated by:

[0736] The total number of samples (N) from the first reference signal to the second reference signal B ) to count, and

[0737] From the first reference signal to the sample number N PThe number of samples (N) where the peak amplitude value Sp appears P ) to count, and

[0738] Based on the sample number N P With the total number of samples (ie N B ) to generate the first time relationship (X1; R T (r); T D ; FI(r)).

[0739] The relationship generator 1050 may generate updates of the relationship value X1 using a transmission frequency that depends on the rotation speed of the housing 20 .

[0740] As described above, the state parameter extractor 450C can be configured to transmit an updated value X1(r) once per revolution. In this manner, the transmitted updated value X1(r) can be based on L values ​​generated during one revolution. The latest update numbered r of the first internal state parameter X1(r) can be transmitted on the first state parameter extractor output 1060.

[0741] refer to Figure 31 , the vibration event signature detector 990 can be configured to detect a peak amplitude sample value Sp. The vibration event signature detector 990 has an output 1070 for transmitting the detected vibration signal amplitude peak value Sp. The detected vibration signal amplitude peak value Sp can be transmitted from the output 1070 of the vibration signal peak amplitude detector 990 to the output 1080 of the state parameter extractor 450C. The output 1080 constitutes a second state parameter extractor output for transmitting a second internal state parameter X2(r), also referred to as Sp(r). The second internal state parameter X2(r) is transmitted at the same transmission frequency as the first internal state parameter X1(r).

[0742] Furthermore, the first internal state parameter X1(r) and the second internal state parameter X2(r) are preferably transmitted simultaneously as a set of internal state parameter data (X1(r); X2(r)). In the symbol X1(r), "r" is the number of samples indicating the time slot, i.e., increasing values ​​of "r" indicate the time progression, in the same manner as the number "i" in column #01 in Table 3.

[0743] As mentioned elsewhere in this document, the vibration signal characteristic S FIMP The peak amplitude sample value Sp seems to depend on the impact force F IMP The impact force F of the interaction between the rotational motion protrusion 310 and the material charge 30 IMP The impact causes acceleration of at least one particle in the toe 205 of the material charge 30, which results in a mechanical shock vibration V IMP . refer to Figure 2It should be noted that the impact of the protrusion 310C on the bulk of the material in the toe 205 causes the bulk of the toe material to be ACC Accelerate in direction. Direction A ACC is the direction in which the protrusion 310C moves. This acceleration causes a force F against the leading edge surface 312C of the protrusion 310C. IMP The impact force F IMP It can be estimated to be an order of magnitude:

[0744] F IMP =m 205 *a 205

[0745] in,

[0746] m 205 is the weight of the accelerated part of the toe,

[0747] a 205 is the magnitude of the acceleration of that part of the toe.

[0748] In view of the above, the inventors have concluded that the magnitude of the detected peak amplitude sample value Sp can advantageously indicate the charge density in the roller mill 10 .

[0749] In this context, it should be noted that the content of the desired metal in the solid material 110 of the charge material 30 affects the charge density in the roller mill 10, as discussed in conjunction with Table 1 of this document. Therefore, the charge density in the roller mill 10 can indicate the relationship between the desired metal and the waste mineral in the charge in the roller mill 10.

[0750] Therefore, the inventors have concluded that the magnitude of the detected peak amplitude sample values ​​Sp can advantageously indicate the relationship between the desired metal and the waste mineral in the charge to the roller mill 10 .

[0751] Furthermore, the inventors have concluded that the magnitude Sp of the detected peak amplitude sample values, combined with data indicative of the toe position, ie, the aforementioned relationship value X1 , can advantageously indicate the degree of filling of the roller mill 10 .

[0752] In this regard, it should be noted that the filling level of the roller mill 10 has an impact on the efficiency of the grinding process. In order to maximize the amount of output material 95 from the roller mill 10, it is therefore desirable to control the inflow of input material 110 so as to maintain optimal conditions for the roller milling process, including an optimal filling level. The optimal internal conditions for the roller milling process may include a certain filling level of the housing 20, i.e., a certain charge volume.

[0753] Therefore, the solid material feed rate set point R can be controlled according to the combination of the relationship value X1(r) and the magnitude of the detected peak amplitude sample value X2(r). SSP .

[0754] Furthermore, the inventors have concluded that the magnitude of the detected peak amplitude sample value Sp(r) = X2(r) combined with the data indicating the toe position, i.e., the relationship value X1(r) discussed above, can advantageously indicate the absolute toe position value X6(r) of the roller mill 10.

[0755] X1(r) X2(r) X6(r-1) => X6(r)

[0756] exist Figure 19A and Figure 19B As can be seen, the increasing polar angle X1(r) = FI(r) combined with the increasing radius value X2(r) = S P (r) presents an image of a spiral arm rotating outward from the first internal state indicator object 550(1), as shown in FIG. Figure 19A The "angular length" X6(r) of the spiral arm from the initial polar angle φ(1) of the first internal state indicator object 550(1) to the current or most recently detected toe position FI(r) appears to indicate the absolute position X6(r) of the toe 205 (see, for example, Figure 2 and Figure 14). In this regard, it should be noted that Figure 19A 360 degrees in the polar coordinate system 520 corresponds to a relative toe position value X1 indicating 100% of the distance between two adjacent static reference positions, as shown in FIG. Figure 14B discussed.

[0757] In particular, please note that Figure 19A and Figure 19B , the change of the absolute toe position X6 has a certain slowness.

[0758] Figure 32 is a block diagram of a system 5, 320, 770 including a roller mill illustrated as a block 10 that receives a plurality of inputs U1, ...Uk and generates a plurality of outputs Y1, ...Yn. Figure 32 and Figure 1C It should be noted that for the purposes of analysis, the roller mill 10 can be considered as a black box 10B having a plurality of input variables, which are referred to as input parameters U1, U2, U3, ..., Uk, where index k is a positive integer. During operation of the roller mill 10, 10B, the roller mill has an internal state X, and for the purposes of analysis, the roller mill 10 can be considered as a black box 10B having a plurality of output variables, which are also referred to as output parameters Y1, Y2, Y3, ..., Yn, where index n is a positive integer.

[0759] The internal state X of the mill can be described or indicated by a plurality of internal state parameters X1, X2, X3, ..., Xm, where the index m is a positive integer.

[0760] Using the terminology of linear algebra, the input variables U1, U2, U3, ...Uk can be collectively referred to as the input vector U. Therefore, the dimension of the input vector U is k:

[0761] Input vector U: Dim(U)=k

[0762] Similarly, the internal state parameters X1, X2, X3, ...Xm can be collectively referred to as the internal state vector X.

[0763] The dimension of the internal state vector X is m:

[0764] Internal state vector X: Dim(X) = m

[0765] The output parameters Y1, Y2, Y3, ...Yn can be collectively referred to as the output vector Y:

[0766] The dimension of the output vector Y is n:

[0767] Output vector Y: Dim(Y)=n

[0768] At a time point called r, the internal state X of the mill 10 can be referred to as X(r). This internal state X(r) can be described or indicated by a plurality of internal state parameters X1, X2, X3, ..., Xm, as described above. These internal state parameters define different aspects of the internal state X(r) of the mill 10 at time r.

[0769] The internal state X(r) of the roller mill 10 depends on the input vector U(r). One aspect of the internal state X is the total amount of material 30 in the housing 20, and this amount does not change immediately. Therefore, during operation of the mill 10, the internal state X(r) can be considered as a function of the earlier internal state X(r-1) and the input U(r):

[0770] X(r) = f1(X(r-1), U(r) ), (Formula 4)

[0771] Here, X(r-1) indicates the internal state X of the mill 10 at a point in time before a point in time called r.

[0772] The output Y of the roller mill 10 can be viewed as a function of the internal state X. Thus, using the terminology of linear algebra, the output vector Y(r) depends on the internal state vector X(r):

[0773] Y(r) = f2(X(r)) (Formula 5)

[0774] One aspect of this document aims to solve the problem of how to keep the internal comminution process of the mill 10 at a suitable operating point. Therefore, during the operation of the mill 10, it may be desirable to offset deviations from such a suitable operating point. This problem can be solved by providing a linearized model of the comminution process at the operating point. When the above functions f1 and f2 are considered at operating points near the suitable operating point, respectively, the functions can be linear. Therefore, at the selected operating point, the internal state X(r) can be regarded as a function of the earlier internal state X(r-1) and the input U(r) according to the linear model, which can be written as follows:

[0775] X(r) = A *X(r-1)+B*U(r) (Formula 6)

[0776] Where A and B are coefficient matrices.

[0777] In this regard, it should be noted that in linear algebra, a coefficient matrix is ​​a matrix comprising the coefficients of the variables in a set of linear equations. As known to technical readers of this document, coefficient matrices are used to solve systems of linear equations.

[0778] In this regard, it should be noted that the coefficients in matrices A and B, respectively, may be constants.

[0779] Similarly, at a chosen operating point, the output vector Y(r) depends on the internal state vector X(r) according to a linear model that can be written as follows:

[0780] Y(r)=C*X(r) (Formula 7)

[0781] Where C is the coefficient matrix.

[0782] However, Equation 7 does not imply that a change in state X necessarily translates immediately into a change in state Y, as there may sometimes be a delay between the occurrence of a change in internal state X and the occurrence of a corresponding change in state Y(r) of product materials 95, 96. However, when operating in a steady state, a causal relationship appears to exist between the internal state X of the comminution process occurring in mill 10 at time r and the state Y(r) of product materials 95, 96 at the same time r. Therefore, Equation 7 is valid, at least when roller mill 10 is operating in a steady state.

[0783] Referring to Formula 7, the coefficients in the matrix C can be constants. The constant values ​​of the coefficients in the matrix C can be obtained at the selected operating point X. OP Set the derivative C = dY / dX.

[0784] refer to Figure 32The system includes a monitoring module 150A for generating an m-dimensional internal state vector X, where m is a positive integer. In one example, Dim(X) is at least 2. The values ​​in the internal state vector X can be Figures 1A to 31 generated in any of the disclosed ways.

[0785] The monitoring module 150A may be adapted to transmit 1122 information describing the internal state X of the mill, e.g., via the user interface 210, during operation of the mill 10, as indicated by arrow 1122. Thus, one or several values ​​of the internal state vector X may be transmitted to the operator 230 via the user interface 210. This advantageously simplifies the operator 230 of the mill 10 making appropriate adjustments 1124 to the setpoint value (index SP) to affect the input vector U. Thus, by adjusting, for example, the speed setpoint value U1 SP (See Figure 32 Combine Figure 1B ), the operator 230 can adjust the speed f ROT , U1.

[0786] In this way, the operator adjusts the relevant set point value U SP The corresponding input variables U1, U2, U3, ...Uk can be adjusted.

[0787] Set point value U1 SP , U2 SP , U3 SP , ...Uk can be collectively referred to as the set point vector U SP Therefore, set the point vector U SP The dimension of is k:

[0788] Set point vector U SP :Dim(U SP )=k.

[0789] Figure 32 The system 5, 320, 770 may include the monitoring module 150A described in any other embodiment described in this disclosure, for example, with respect to FIG. Figure 31 Any one of .

[0790] Figure 33 is a block diagram of another system 730, 940, 950 including a roller mill illustrated as block 10 that receives a plurality of inputs U1, ...Uk and generates a plurality of outputs Y1, ...Yn.

[0791] Figure 33 The system 940 may include a monitoring module 150A as described in any other embodiment described in the present disclosure, for example, with respect to FIG. Figure 31 In addition, Figure 33The system 940 may include a control module 150B as described in any other embodiment described in this disclosure, for example, Figure 28 .

[0792] Figure 33 The monitoring module 150A may be adapted to transmit information describing the internal state X of the mill during operation of the mill 10, for example via the user interface 210. Thus, one or several values ​​of the internal state vector X may be transmitted 1122 to the operator 230 via the user interface 210, as indicated by arrow 1122. This advantageously simplifies the operator 230 of the mill 10 in controlling the mill setpoint value U and / or the internal state reference value X. REF (index REF) makes appropriate adjustments 1126 to affect the internal state X of the mill during operation of the mill 10. Arrow 1126 indicates the desired internal state X. REF Related user input. Internal state reference parameter X1 REF , X2 REF , X3 REF ,...,Xm REF It can be collectively referred to as the internal state reference vector X REF .

[0793] Internal state reference vector X REF The dimension of is m:

[0794] Internal state reference vector X REF :Dim(X REF )=m

[0795] In this way, the operator 230 adjusts the mill set point value U and / or the associated internal state reference parameter value X1 REF , X2 REF , X3 REF ,...,Xm REF The internal state X of the mill may be influenced during operation of the mill 10. Thus, in response to user input, the user interface 210 may be configured to generate an internal state reference vector X REF The value of .

[0796] like Figure 33 As shown, the internal state reference vector X REF is transmitted to the reference input of the control module 150B. Figure 26 refer to Figure 33 The control module 150B is a multivariable control module, and the control module 150B also receives the above-mentioned internal state vector X from the monitoring module 150A.

[0797] In this regard, the internal state vector X may indicate the current state of the grinding process in the mill 10, and the internal state reference vector XREF Indicates the desired state of the grinding process.

[0798] The multivariable control module 150B may be adapted to receive an internal state reference vector X based on the received internal state reference vector X. REF and the received internal state vector X, generating an internal state error vector X ERR .

[0799] Internal state error vector X ERR Including internal state error value X1 ERR , X2 ERR , X3 ERR ,...,Xm ERR

[0800] Internal state error vector X ERR The dimension of is m:

[0801] Internal state error vector X ERR :Dim(X ERR )=m

[0802] The error vector is transmitted to the regulator 755, 755C. Figure 33 The regulator 755, 755C is suitable for generating a set point vector U SP Therefore, the set point vector U SP Including the above set point values ​​for controlling or regulating the corresponding input variables U1, U2, U3, ... Uk (see Figure 33 and Figure 34 ).

[0803] Therefore, about Figure 33 The described system advantageously simplifies the operation of the operator 230 of the mill 10 by transmitting 1122 information indicative of the internal state X of the mill during operation, while also allowing the operator to provide 1126 information describing the desired internal state, for example, in the form of the internal state reference vector X described above. REF The reference value is in the form of .

[0804] The regulator 755, 755C may be a multivariable regulator configured to include a multivariable proportional-integral-derivative controller (PID controller). Alternatively, the regulator 755, 755C may be configured to include a multivariable proportional-integral controller (PI controller). Alternatively, the regulator 755, 755C may be configured to include a multivariable proportional controller (P controller).

[0805] Alternatively, the regulator 755, 755C can be configured to include a Kalman filter, also known as a linear quadratic estimator (LQE). A Kalman filter is an algorithm that uses a series of measurements observed over time, including statistical noise and other inaccuracies, and produces estimates of unknown variables by estimating the joint probability distribution over the variables at each time period. These estimates tend to be more accurate than estimates based on a single measurement alone.

[0806] Figure 34 There is shown another schematic diagram of a system 1130 including a roller mill 10. Thus, reference numeral 1130 relates to a system including a mill 10 having a rotatable shell 20, as described herein. Figure 34 The system 1130 may include multiple components and may be as described above with respect to Figure 1A and / or Figure 1B described and / or configured as described in any other examples described in this disclosure, for example, with respect to FIG. Figure 33 .

[0807] The monitoring module 150A may include a state parameter extractor function as described elsewhere in this document for generating internal state parameter values ​​X1, X2, X3, ..., Xm. It should be noted that the internal state X of the mill 10 at a time point referred to as r may be referred to as X(r). This internal state X(r) may be described or indicated by a plurality of parameter values ​​that define different aspects of the internal state X(r) of the mill 10 at time r. Thus, the values ​​of the internal state parameters X1, X2, X3, ..., Xm at time r may be collectively referred to as the internal state vector X(r).

[0808] Figure 34 The system shown may provide an integrated HCI 210, 250, 210S. Thus, Figure 34 The input / output interface 210 can be configured to enable all of the above-mentioned inputs and / or outputs. Figure 34 The input / output interface 210 can be configured to provide 1132 information related to the state of the output material. The state of the output material can be described by output parameters Y1, Y2, Y3, ...Yn, collectively referred to as the output vector Y. As described above, the dimension of the output vector Y is n:

[0809] Output vector Y: Dim(Y)=n

[0810] Vector Y may also be referred to as the discharged material state vector Y.

[0811] Figure 34The system 1130 includes a regulator 1190. The regulator 1190 can be configured to enable all of the functionality described with reference to the regulator 240 described elsewhere in this document. Alternatively, the regulator 1190 can be configured to enable all of the functionality described with reference to the regulator 755 described elsewhere in this document. In addition to the functionality described in regulator 240 and / or regulator 755, the regulator 1190 can also be configured to perform additional functions, such as transmitting and / or receiving information related to the output materials 95, 96, for example, in the form of output parameters Y1, Y2, Y3, ... Yn. Therefore, the regulator 1190 can also be indicated by the reference numerals 240C and / or 755C.

[0812] Thus, the regulator 1190 may be configured to transmit information related to the output material 95, 96 to the mill operator 230, as indicated by arrow 1132. Additionally, the regulator 1190 may be configured to receive information related to the output material 95, 96 from the mill operator 230, as indicated by arrow 1196.

[0813] Figure 35 It can be done by Figure 34 Schematic diagram of the overall information transmitted by the input / output interface 210. Figure 34 and Figure 35 , it should be noted that Figure 34 The regulator 1190 and 755C are coupled via the coupling 1100 for data exchange with the input / output interface 210. The information to be transmitted via the coupling 1100 includes the internal state reference vector X REF reference value.

[0814] refer to Figure 34 System 1130 includes a product analyzer 1140 configured to analyze at least a portion of the product particles 96. Analyzer 1140 is configured to generate at least one product measurement value Y1, Y2, Y3, ..., Yn based on the product particle analysis. Product analyzer 1140 may, for example, include an Outotec PSI 300i analyzer. Alternatively, product analyzer 1140 may include a PSI 500i particle size analyzer, available from Metso Outotec. Furthermore, product analyzer 1140 may include a device configured to measure or estimate the weight per unit time of material 95 discharged from the mill, such as a scale. However, in some examples, product analyzer 1140 may be configured to analyze or measure a portion of product particles 96, rather than all product particles 96.

[0815] In practice, at least one output material measurement value Y1, Y2, Y3, ... Yn may indicate an output material state Y, which is an instantaneous state of the output material 95. When the analyzer 1140 provides two or more output material measurements, these values ​​may be provided in the form of an output vector Y as described above.

[0816] The at least one product measurement may for example comprise an indication of the product discharge rate R SDis The product discharge rate R SDis constitutes data indicating the weight per unit time of solid material discharged from the mill after grinding. Therefore, the product discharge rate R SDis is a measure of the weight per unit time of the material 95 discharged from the mill. The product discharge rate R SDis It can also be called output parameter Y1.

[0817] However, the inventors concluded that the product discharge rate R SDis The measurement of Y1 does not necessarily require analysis of the material 95 discharged from the mill. In fact, it indicates the product discharge rate R SDis Data for the toe position (i.e., Y1) can be obtained by analyzing the solid feed material 110. The ability to control the toe position advantageously improves the ability to maintain the mill in steady-state operating conditions even when the properties of the solid feed material change. During steady-state operation, the outflow from the mill is equal to the inflow to the mill, and therefore, the weight per unit time of the solid discharge material 95 is equal to the weight per unit time of the solid feed material 110 during steady-state operation. If there is no steady-state operation, the toe position will change, and as discussed elsewhere in this disclosure, measures can then be taken to control the toe position, for example, to return to steady-state operation.

[0818] In order to analyze the weight per unit time of the ground material (ie the product particles), the weight of the feed 110 can therefore be measured (see, for example, Figure 1A ).

[0819] As described elsewhere in this disclosure, the solid feed material 110, also referred to as feed material 110, may include rock and ore pieces 115, the pieces 115 having various sizes. The solid material 110 may be delivered to the first input end 100 of the mill 10 by a conveyor belt 260. The conveyor belt 260 operates at a controllable conveyor belt speed to convey the solid feed 110 at a solid feed rate R SThe conveyor belt 260 is fed to the first input end 100. According to one example, the conveyor belt 260 includes a scale configured to measure the weight of the solid material 110 on the conveyor belt 260 with a controllable conveyor belt speed. Therefore, the combination 325 of the conveyor belt and the scale provides information about the weight per unit time of the solid material 110 fed to the first input end 100 of the mill 10. In other words, based on the weight information from the scale and the information about the conveyor belt speed, data U6 indicating the weight per unit time of the solid material 110 fed to the first input end 100 of the mill 10 can be obtained.

[0820] refer to Figure 34 The combination of a conveyor belt and a scale is schematically shown as a feed material analyzer 325. The feed material analyzer 325 can be provided to generate measurements indicative of at least one feed material characteristic U5, U6. The at least one feed material characteristic can include a solid feed material size distribution U5 and / or data U6 indicative of a weight per unit time of the solid material 110 delivered to the first input end 100 of the mill 10.

[0821] For example, during steady state, when the conveyor belt and scale combination is configured to weigh the solid material 110 over a 20-meter section of the conveyor belt 260, the conveyor belt 260 moves at a speed of 10 meters per second, and the scale indicates 100 kg, this means that the flow rate is approximately 100 kg over a duration of 2 seconds. Therefore, the weight per unit time of the solid feed material 110 can be approximately 180 metric tons per hour.

[0822] The product discharge rate R, also known as the output parameter Y1 SDis The product discharge rate R may be estimated based on data U6 indicating the weight per unit time of the solid material 110 fed to the first input 100 of the mill 10. According to one example, the product discharge rate R SDis (ie Y1) is estimated to be equal to data U6 indicating the weight per unit time of solid material 110 delivered to the first input 100 of the mill 10. As described above, during steady state, the product discharge rate R from the mill SDis (ie Y1) is equal to the solid material feed rate U6 to the mill, and therefore the product discharge rate R SDis (ie, Y1) can be estimated using the solid material feed rate U6 during steady state operating conditions.

[0823] The instantaneous state of the output material 95, i.e., the discharged material state Y, can be identified by measuring at least one output material measurement value Y1, Y2, Y3, ... Yn. In practice, it may be desirable to generate more than one output material measurement value in order to obtain information indicative of the discharged material state (Y).

[0824] The at least one output material measurement value may be one or more selected from the group consisting of:

[0825] a value Y1 indicating the weight per unit time of the output material 95;

[0826] a value Y1 indicating the weight per unit time of the product pellets 96;

[0827] a value Y2 indicating the median particle size;

[0828] a value Y3 indicative of the weight per unit time of said product particles 96 having a particle size below a predetermined product particle size limit;

[0829] a value Y4 indicating a proportion or percentage share of said product particles having a particle size between a lower product particle size limit Y8 and an upper product particle size limit Y9;

[0830] a value Y5 indicating the count of said product particles having a particle size between a lower product particle size limit and an upper product particle size limit, that is, the number of product particles;

[0831] a value Y6 indicative of the product particle size distribution, such as standard deviation; and

[0832] A value Y7 indicating the particle size of the product particles.

[0833] The product particle size Y7 can be at least one selected from the following groups:

[0834] -Median particle size of the product particles;

[0835] - Average particle size of the product;

[0836] - product particle median diameter value; and

[0837] -The average diameter of the product particles.

[0838] The product particle size limit may be at least one selected from the following group:

[0839] - Product particle diameter value; and

[0840] -The maximum width of the product particles.

[0841] The value Y6 indicating the particle size distribution of the product particles may be at least one selected from the following group:

[0842] - Standard deviation value;

[0843] - Variance value;

[0844] - the range between the highest and lowest particle sizes;

[0845] -Interquartile range.

[0846] The range between the minimum and maximum particle size values ​​may be between:

[0847] 30 microns and 20 mm;

[0848] 150 microns and 300 microns;

[0849] 200 microns and 220 microns; and / or

[0850] 0 mm and 40 mm.

[0851] Thus, product analyzer 1140 can be configured to analyze at least a portion of product particles 96 to generate at least one product measurement value Y1, Y2, Y3, ..., Yn based on the product particle analysis. The at least one product measurement value Y1, Y2, Y3, ..., Yn can be provided with information indicating a point in time when the at least one product measurement value Y1, Y2, Y3, ..., Yn was generated.

[0852] Furthermore, at a time point designated as w, the state of the discharged material Y may be referred to as Y(w). This state of the discharged material Y(w) may be described or indicated by a plurality of parameter values ​​Y1(w), Y2(w), Y3(w), ... Yn(w), which define different aspects of the material 95, 96 discharged from the mill 10 at time w. Thus, the discharged material parameter values ​​Y1, Y2, Y3, ... Yn at time w may be collectively referred to as a discharged material state vector Y(w), also referred to as an output vector Y(w).

[0853] As described above, there is a causal relationship between a certain internal state X(r) and a certain output Y(r), so the output Y of the roller mill 10 can be considered as a function of the internal state X.

[0854] refer to Figure 34 , the output vector Y can be transmitted to the first input terminal of the correlator 150C1. In addition, the internal state vector X can be transmitted by the module 150A to the second input terminal of the correlator 150C1. The correlator 150C1 is configured to identify the correspondence between the internal state X and the corresponding output Y.

[0855] However, in order to perform the correlation, it is desirable to ensure that the measured values ​​of the output Y(w) refer to at least approximately the same point in time as the internal state X(r). In other words, the values ​​in the internal state vector X(r) may need to be synchronized with the values ​​in the corresponding output vector Y(w). Figure 34, the output vector Y(w) may be passed to a first input of an optional synchronizer 1150. The synchronizer 1150 is optional because it may not be needed, for example, when the internal state vector X(r) and the corresponding output vector Y(w) are generated in a synchronized manner such that

[0856] - time point w is the same time point as time point r, or

[0857] - so that the time point w is at least approximately the same point as the time point r.

[0858] like Figure 34 As shown, time synchronization vectors X(t) and Y(t) are received by the association data generator 1160.

[0859] The associated data generator 1160 generates an associated data set 1170. According to one example, the associated data generator 1160 generates the associated data set by performing the following associations:

[0860] At least one state parameter value received, e.g., X1(t), and

[0861] At least one corresponding product measurement value is received, for example, Y2(t).

[0862] The associated data generator 1160 may receive a plurality of time-stamped internal state vectors X(r) and a plurality of time-stamped corresponding output vectors Y(w). The received information vectors may be received in a time-interleaved manner, such as X(10), Y(12), X(14), Y(16), X(18), Y(20), X(22), Y(24), wherein the synchronizer 1150 receives the vector X in a time period between the reception of two consecutive vectors Y. For example, this is the case when the vector X(18) is time-stamped in a time period between t=20 and t=16, and the Y vectors Y(16) and Y(20) are time-stamped at time points between t=16 and t=20, respectively. When the mill 10 is operated under steady-state conditions, i.e., when all values ​​in the vectors X and Y are stable over time, the synchronizer 1150 may generate the vector pair X and Y by adjusting the time stamps so that the generated vector pair X and Y have the same time stamp. The same timestamp can be, for example, an intermediate timestamp. For example, upon receiving the above-mentioned vectors X(18) and Y(20), the synchronizer 1150 can arrange them into a vector pair marked with the intermediate time t=19. Therefore, in response to receiving the vectors X(t) and Y(t+2), the synchronizer 1150 can generate a vector pair X(t+1) and Y(t+1) to transmit to the associated data generator 1160.

[0863] In addition, the transmission frequency of the X vector and the Y vector may be different. This problem can be solved, for example, by configuring the synchronizer 1150 to transmit the following to the associated data generator 1160:

[0864] The received pairs of vectors X and Y are such that each timestamp vector Y is associated with the vector X having the closest earlier timestamp. Therefore, the synchronizer 1150 may have to discard or reject some vectors.

[0865] Therefore, for example, when the transmission frequency of the X vector is lower than the transmission frequency of the Y vector, the synchronizer 1150 may receive the following vectors:

[0866] vector X(34),

[0867] vector Y(36),

[0868] vector X(37),

[0869] vector Y(38),

[0870] vector X(40),

[0871] vector Y(40),

[0872] Vector Y(42)

[0873] vector X(43),

[0874] vector Y(44),

[0875] The synchronizer 1150 may then transmit vector X and Y pairs 1165 to the association data generator 1160 such that each timestamped vector Y is associated with the vector X with the closest earlier timestamp. In the above example, the synchronizer 1150 may transmit the following pairs:

[0876] Vector X(34), vector Y(36),

[0877] Vector X(37), vector Y(38),

[0878] Vector X(40), Vector Y(40),

[0879] Vector X (43), vector Y (44), and as the same sequence vector Y (42), can be discarded.

[0880] Table 7 below is an example of consecutive pairs 1165 of vectors X and Y arranged in time order.

[0881] Table 7: Successive pairs of vectors X and Y arranged in time order 1165

[0882]

[0883]

[0884] The example of a continuous pair 1165 of vectors X and Y shown in Table 7 includes information indicative of the toe position X1 and information indicative of the corresponding output parameter Y2. The output parameter Y2 indicates the median particle size of the particles produced by the tumble mill.

[0885] The correlation data generator 1160 may be configured to perform correlation based on the received vector X and Y pairs 1165. According to one example, the correlation data generator 1160 may be configured to perform regression analysis based on a large number of received vector X and Y pairs 1165.

[0886] Regression analysis uses one or more statistical procedures to estimate the relationship between a dependent variable (i.e., the values ​​in vector Y) and one or more independent variables (i.e., the values ​​in vector x).

[0887] Figure 36 The housing 20 of the ball mill 10 rotates at a speed f ROT Therefore, Figure 36 Can be combined with the above Figure 1A 、 Figure 2 、 Figure 14A 、 Figure 14B 、 Figure 23 and Figure 30 Compared with the cross-sectional view. Figure 36 The roller mill 10 is operated as a ball mill, and thus the charge material 30 includes a plurality of grinding balls 1168 and solid feed materials 110, 115. In some examples, the charge material 30 may also include a liquid feed material, such as water.

[0888] Figure 37A is an example of an output graph of a particle size analyzer included in the product analyzer 1140. As described above, in conjunction with Figure 34 The system 1130 includes a product analyzer 1140 configured to analyze at least a portion of the product particles 96. The analyzer 1140 is configured to generate at least one product measurement value Y1, Y2, Y3, ... Yn based on the product particle analysis. Figure 37A As shown in the graph of , the particle size analyzer can generate data values ​​indicating the particle size distribution. Based on the data values ​​indicating the particle size distribution, a data value Y2 indicating the median particle size can be obtained.

[0889] exist Figure 37A In the example graph, it can be seen that in the range of 230-280 microns, the peak in the graph represents the highest particle count. Figure 37AAlso shown is the desired particle size range, i.e., a target range of 200 to 220 microns. As described above, the product analyzer 1140 may, for example, comprise an Outotec PSI 300i analyzer. Alternatively, the product analyzer 1140 may comprise a PSI 500i particle size analyzer available from Metso Outotec.

[0890] Figure 37B is for a constant or approximately constant speed U1 = f ROT =18rpm operation Figure 36 The vectors x1 and y2 of the ball mill are shown as a large number of continuous pairs 1165. Figure 37B In FIG, the density of black dots, each of which represents a pair of values ​​X1-Y1, is very high when X1 = 60 degrees. The high density of black dots at values ​​near X1 = 60 degrees indicates that the mill is often operated at or near the operating point of X1 at 60 degrees.

[0891] It can be seen that X1 = 60 corresponds to a median particle size Y1 of about 260 microns. The density of black dots is lower at about X1 = 40 to 50 degrees, indicating that the mill is generally less operated at an operating point where X1 is between 40 and 50 degrees.

[0892] Combine Figure 36 refer to Figure 1A , the radius of the ball mill is R MIC = 1930 mm, and has 28 protrusions 310 ( Figure 36 During operation of the ball mill 10 , the grinding ball feed rate U4 can be set or selected such that the amount of grinding balls remains constant or approximately constant.

[0893] refer to Figure 34 , the associated data generator 1160 can be configured based on Figure 37B The pairs 1165 of received values ​​X1 and Y2 are shown to perform regression analysis, also known as correlation analysis.

[0894] Regression analysis can employ linear regression, for example. When applied to a single dependent variable Y2 and a single independent variable (such as X1 or X6), linear regression analysis operates according to specific mathematical criteria to identify the linear relationship that best fits the data, i.e., line 1180. For example, the ordinary least squares method calculates a unique line 1180 that minimizes the sum of the squared differences between the true data and the line. Thus, as described above, Figure 37B Line 1180 in FIG. 1 is a graphical representation of the results of a linear regression based on the received value X1 - Y1 pairs.

[0895] Therefore, the correlation data set 1170 generated by the correlator 150C1 may include a data table or a linear equation. The linear equation of a line in a two-dimensional space (such as the two-dimensional space X1-Y2) can be established based on the points through which the line passes. For example, when a line passes through two points (X1 A , Y2 A ) and (X1 B , Y2 B ), the symmetric equation of the line is given by Equation 8:

[0896] (X1 - X1 A ) / (X1 B - X1A A ) = (Y2 – Y2 a ) / (Y2 A – Y2 B ) (Equation 8)

[0897] Where

[0898] X1 A and Y2 A are numerical values defining a point in the two-dimensional space X1-Y2, and

[0899] X1 B and Y2 B are numerical values defining another point in the two-dimensional space X1-Y2.

[0900] Therefore, Equation 8 defines the linear relationship between the internal state parameter X1 and the output parameter Y2 at the operating point of the rolling mill (see Figure 37C ).

[0901] Figure 37C is the curve graph of the generated linear regression result. Therefore, the curve in the figure shows the line identified as line 1180, which shows the relationship between X1 and Y2, at least when the ball mill operates at a constant or substantially constant rotational speed f ROT = 18 rpm at an operating point of 0 degrees < X1 < 80 degrees. Therefore, when linear regression analysis is applied to a single dependent variable Y2 and a single independent variable, the correlation data generator 1160 can transmit the correlation data 1170 indicating the linear relationship 1180.

[0902] Refer to Figure 34 , the correlation data set 1170 generated by the correlator 150C1 can be transmitted to the internal state reference value generator 150c2.

[0903] The internal state reference value generator 150c2 can be configured to use the received correlation data 1170 to convert the expected value Y REF into the corresponding internal state reference value X REFTable 8 is used to convert the expected value Y2 REF Converted into the corresponding internal state reference value X1 REF In fact, Table 8 is an example data set corresponding to the information in Table 7 above.

[0904] Table 8: Related data set 1170 in the form of a related table for the desired product characteristic Y2 REF Converted to internal state parameter reference value X1 REF

[0905]

[0906]

[0907] An example correlation data table 1170 (an example of which is shown in Table 8) indicates a correlation between an internal state parameter value X1 indicating toe position and an output parameter Y2 indicating the median particle size of particles produced by the tumble mill.

[0908] Based on the above equation 8 and any two points (X1, Y2) or (X1 REF , Y2 REF ), we can generate a linear equation that defines the relationship. Express the dependent variable Y2 as a function of the independent variable X1, Figure 37C Line 1180 in can be described by rewriting Equation 8 as Equation 9:

[0909] Y2=(Y2 B -Y2 A ) / (X1 B -X1 A )*X1-(Y2 B -Y2 A A) / (X1 B -X1 A )*X1 A +Y2 A

[0910] (Equation 9)

[0911] refer to Figure 37C , the linear relationship example 1180 shown includes two points:

[0912] (X1 A , Y2 A ) = (40, 200), and

[0913] (X1 B , Y2 B )=(47,220)

[0914] Therefore, by plugging these example values ​​into Equation 9 and expressing the dependent variable Y2 as a function of the independent variable X1, Figure 37C The line 1180 in can be described by the following equation:

[0915] Y2=k1*X1+m1=2.86X1+85.71

[0916] As above combined Figure 34 As described above, the associated data generator 1160 generates an associated data set 1170. The associated data set 1170 can be delivered in the form of a linear equation 1170 for converting the expected value Y REF Converted to the corresponding internal state reference value X REF . Refer again Figure 37C And the linear relationship example 1180 shown including these two points:

[0917] (X1 A , Y2 A ) = (40, 200), and

[0918] (X1 B , Y2 B )=(47,220)

[0919] It should be understood that the above equation 8 can be rewritten as equation 10 (below) so that the desired internal state value (ie, the internal state reference value X1) REF ) is expressed as the expected value Y2 REF Function:

[0920] X1 REF =k2*Y1 REF +m2 (Equation 10)

[0921] in,

[0922] k2=(X1 B -X1 A ) / (Y2 B -Y2 A )

[0923] m2=X1 A -(X1 B -X1 A ) / (Y2 B -Y2 A )*Y2 A .

[0924] Therefore, reference Figure 34 , the associated data generator 1160 can generate an associated data set 1170 in the form of a linear equation 1170 for converting the expected value Y REF Converted to the corresponding internal state reference value X REF, and the linear equation 1170 can be, for example, as shown in Equation 10 above.

[0925] Figure 37D Part I of FIG is an example of a graph showing the density of toe position values ​​X6 over a period of mill operation. Figure 37D It can be inferred from the graph in Part I of that during this time period, the mill most often operates with toe position values ​​X6 in the range of 550 to 600 degrees. In this regard, a toe position value X6 of zero degrees may indicate an empty mill, and a toe position value X6 of 720 degrees (= 2*360) may indicate an empty mill. Figure 2 Toe position A shown TOE As described above, the distance between two adjacent protrusions 310 can be expressed as a phase value of 360 degrees, and thus Figure 2 Toe position A shown TOE Corresponding to approximately 720 degrees.

[0926] The actual value of parameter X6 will depend on the number of protrusions 310 (also referred to as lifters 310 ).

[0927] Figure 37D Part II shows the dependence of the toe position value X6 (ie A TOE ) discharge rate R SDis (ie Y1) products (see Figure 2 Gray dots 1183 represent the product discharge rate R depending on the toe position value X6. SDis (i.e. the maximum value detected by Y1).

[0928] The curved black dashed line 1185 represents the product discharge rate R in dependence on the toe position value X6. SDis The average value of (i.e. Y1) is 1183. Figure 37D It can be inferred from the graph 1185 in part II of FIG. 1 that during the measurement period for this particular mill, for toe position values ​​X6 ranging from about 550 degrees to about 600 degrees, the product discharge rate R SDis (ie, Y1) shows a maximum value Y1max. Line 1188 represents a toe position value X6 at or near the toe position value X6, resulting in a maximum product discharge rate value R of approximately 850 tons per hour. SDis (i.e. Y1).

[0929] Gray dots 1187 represent all detected product discharge rate values ​​R that depend on the toe position value X6. SDis The gray dots 1187 are dense enough to form a line indicating the product discharge rate value R depending on the toe position value X6. SDis (i.e. the average value of Y1).

[0930] from Figure 37D It can be inferred from the gray point 1187 in Part II that for toe position values ​​X6 below about 500 degrees, the product discharge rate R SDis (i.e. Y1) decreases rapidly. Therefore, based on this correlation, the product discharge rate limit value Y1 can be identified limit , and the corresponding toe position limit value X6 can also be identified limit .

[0931] In general, Figure 37D Part II is the toe position value X6 and the product discharge rate R SDis It can therefore be inferred that the correlation disclosed in the present disclosure advantageously makes it possible to control the mill so as to optimize the product discharge rate value R SDis (ie Y1) toe position value X6 operation.

[0932] Used for Figure 37D The data is based on the product discharge rate R SDis (ie, Y1), which is estimated based on data U6 representing the weight per unit time of the solid material 110 delivered to the first input end 100 of the mill 10, as described above in conjunction with Figure 34 Discussed in detail.

[0933] More complex cases of multivariable monitoring systems

[0934] Figure 37B and Figure 37C Serves as an illustration of the functionality of the correlation data generator 1160 in the relatively simple case of applying regression analysis to a single dependent variable Y2 and a single independent variable X1. In fact, ball mills inherently exhibit less toe position variation than non-ball mill roller mills, particularly when the ball mill is also operated at a constant or approximately constant rotational speed U1, as described above.

[0935] However, another objective addressed by the solutions and examples disclosed in this document is to describe methods and systems for improving the monitoring and / or control of internal conditions X in a roller mill 10 during operation. When the roller mill 10 operates at a variable rotational speed X5 = U1 and also exhibits variations in toe position X1, the aforementioned regression analysis applied to a single dependent variable Y2 and a single independent variable X1 may not be sufficient. However, to address this issue, the correlation data generator 1160 may apply a regression analysis to a plurality of data pairs 1165, including:

[0936] The received internal state vector X(t) of dimension m, and

[0937] The corresponding output vector Y(t) of dimension n received is

[0938] Wherein, m and n are positive integers.

[0939] Thus, when m state parameter values ​​X1, X2, X3, ..., Xm are to be correlated with n product measurement values ​​Y1, Y2, Y3, ..., Yn, the correlation data generator 1160 may be configured to generate a set of correlation data 1170 by performing the following correlation:

[0940] The received internal state vector X(t),

[0941] as well as

[0942] The corresponding output vector Y(t) received,

[0943] in,

[0944] X(t) is an m*1 vector, where m is a positive integer, and

[0945] Y(t) is an n*1 vector, where n is a positive integer.

[0946] Thus, in this case, the correlation data generator 1160 can be configured to perform a regression analysis to identify a more complex linear combination (i.e., more complex than a line in two-dimensional space) that most closely fits the data according to a particular mathematical criterion. For example, the correlation data generator 1160 can perform an ordinary least squares method applied to a plurality of received vectors X(t) in m dimensions and a plurality of received corresponding output vectors Y(t) in n dimensions to calculate a unique hyperplane that minimizes the sum of the squared differences between the received data and the hyperplane.

[0947] Figure 38 is an example diagram of plane 1180B in the three-dimensional space X5, X1, Y2. Therefore, Figure 38 is an example diagram of a plane calculated based on the ordinary least squares method, applied to two received vectors X5(t) and X1(t) and one received corresponding output vector Y2(t). The plane 1180B appears to show that the particle size y2 depends on the rotation speed of the housing X5(f ROT ) and the position X1, X6 of the toe 205. In this respect, it should be noted that if the particle size depends on the relative toe position X1, it also depends on the absolute toe position X6.

[0948] Therefore, when receiving an m-dimensional vector X(t) and a plurality of corresponding n-dimensional output vectors Y(t), the association data generator 1160 is configured to generate a multidimensional association data set 1170. According to one example, the multidimensional association data set 1170 may be transmitted as data 1170 indicating the aforementioned hyperplane. Alternatively, the multidimensional association data set 1170 may be transmitted as data 1170 indicating a coefficient matrix C, as discussed above with respect to Formula 7.

[0949] According to one example, when generating the correlation data set 1170 , the correlation data generator 1160 may be configured to include Kalman filtering, also known as linear quadratic estimation (LQE).

[0950] This solution advantageously makes it possible to identify and / or determine a causal relationship between an internal state X of the comminution process and at least one output material measurement value Y.

[0951] Furthermore, the solution advantageously enables the identification and / or determination of a causal relationship between the internal state X of the comminution process and the discharge material state Y. The discharge material state Y may also be referred to as product material state Y.

[0952] This solution is general because it allows to define the desired discharge material state Y REF , and allows testing alternative internal states of the comminution process, also known as operating points X OP , in order to search and identify the internal state X of the crushing process BEP , the comminution process results in or produces the desired discharge material state Y REF , or to cause or produce a discharge material state Y as close as possible to the desired REF The discharged material state Y. This internal state can be called the optimal operating point BEP. The parameter values ​​at the BEP can be collectively referred to as the internal state BEP vector X BEP .

[0953] Furthermore, the detected instantaneous pulverization process internal state X(r) is recorded in association with the corresponding instantaneous discharged material state Y(r), generating association data indicating the association between:

[0954] The internal state of the instantaneous crushing process X(r), and

[0955] The corresponding instantaneous discharged material state Y(r).

[0956] By repeatedly recording a plurality of different detected instantaneous pulverization process internal states X(r) and associating them with the instantaneous discharged material states Y(r) caused by the corresponding instantaneous pulverization process internal states X(r), where r is a numerical variable indicating a plurality of different time points, an associated data set can be generated. This associated data set indicates the association between:

[0957] Multiple instantaneous crushing process internal states X(r), and

[0958] A plurality of corresponding instantaneous discharged material states Y(r).

[0959] The ball mill operating characteristic curve or BMOC curve of a ball mill is a graph showing the median particle size (Y2) of the product particles produced by the ball mill as the toe position value is varied, as shown in Figures 37 and Figure 38 shown.

[0960] The BMOC curve is created by plotting the toe position values ​​versus the median particle size (Y2) of the product particles at different toe positions.

[0961] Roller mill operating point or X OP or TOP is a specific point within the operating characteristics of the roller mill. It has been found that when the toe position value is around a specific roller mill operating point (X OP When the toe position value of the mill is varied within a certain range of values, there is a linear relationship between the toe position value and the product particle size distribution (Y). In the context of this document, the term mill operating area (MOA) can be used to describe this certain range of toe position values.

[0962] The Mill Operating Characteristic or MOC curve of a roller mill is a graph showing the product particle size distribution (Y) of the product particles generated by the roller mill when at least one of the state parameter values ​​(X1, X2, X3, X4, X5, X6) is changed. Thus, for example, when the rotational speed of the housing (f ROT ) is held constant, an MOC curve is created by plotting the measured values ​​of the product particle size distribution (Y) against the toe position value.

[0963] It has been found that there is a linear relationship between:

[0964] The median particle size (Y2) of the product particles produced by the ball mill, and

[0965] At least when the toe position value X1 varies within a certain range of toe position values, the toe position value.

[0966] Reference again Figure 34 , the internal state reference value generator 150c2 may be configured to use the received associated data 1170 to convert the expected value YREF Converted to the corresponding internal state reference value X REF .

[0967] In this regard, it is noted that a linear relationship may also exist between:

[0968] On the one hand, there is a single dependent variable, such as the median size Y2 of the product particles generated by the ball mill, and on the other hand, there are two independent variables, such as the toe position value, at least when the toe position value X1 varies within a specific range of toe position values, and the rotation speed X5 (f ROT ).

[0969] In this case, the linear relationship 1170 can be represented by a plane equation in three-dimensional space. Figure 34 , the associated data generator 1160 can generate the associated data set 1170 in the form of a plane equation in a three-dimensional space. The plane equation in a three-dimensional space (for example, the space (X5, X2, Y2)) can be expressed as Equation 11:

[0970] a*X5+b*X2+c*Y2=0 (Equation 11)

[0971] The plane equation written in the form of Equation 11 is called a linear equation in three-dimensional space.

[0972] Using connected data to operate the mill

[0973] refer to Figure 34 , the task of the operator 230 in the control room 220 is to operate the roller mill 10. The operator can use the regulator 1190 to operate the mill 10. The regulator 1190 is coupled to the user interface 210, 210B, also known as the human machine interface (HCI) 210B, such as Figure 34 shown.

[0974] Figure 34 The illustrated example control room 220 includes an internal state control system 1200 including an internal state reference value generator 150c2 and user interfaces 210, 210B and a regulator 755C or a regulator 240C.

[0975] The internal state control system 1200 may be configured to perform the following steps:

[0976] (Step S3000:) The user interface 210 transmits a request to the operator to provide an indication of a desired discharge material state Y REF As described above, the user input information indicates the desired discharge material state Y REFThe user input may indicate at least one desired output material measurement value, such as Y1 and / or Y2. For example, the user input may indicate a desired product particle median size Y2. REF , and / or desired product particle size distribution Y3 REF 、Y4 REF , or the desired amount of material output per time unit Y1 REF .

[0977] This request S3000 may be generated by software included in the regulator 755C, or software included in the regulator 240C, or software included in the internal state reference value generator 150c2.

[0978] The internal state control system 1200 may also be configured to:

[0979] (Step S3005:) For example, a request indicating a desired discharge material state Y is received via the user interface 210. REF and / or desired product median particle size Y2 REF and / or data of the desired product particle size distribution Y2, Y3, Y4.

[0980] Furthermore, the internal state control system 1200 may be configured to perform a method comprising the following steps:

[0981] S3010: Generate a toe position reference value (X1 based on the following content) REF ;FI REF ):

[0982] The data indicates the desired discharge material state Y REF and / or the desired product median particle size (Y2 REF ) and / or desired product particle size distribution Y2 REF 、Y3 REF 、Y4 REF ,as well as

[0983] associated data set (1170); the associated data set (1170)

[0984] Indicates a causal relationship between:

[0985] Specific toe position value (X1(r), FI(r), T D , R T (r);X6,A TOE (r)), and

[0986] The corresponding median particle size of the specific product (Y2),

[0987] At the housing speed (U1, f ROT);

[0988] and / or indicate a causal relationship between:

[0989] Specific internal state X REF ,as well as

[0990] Corresponding specific discharge material state Y REF .

[0991] Corresponding specific discharge material state Y REF Specific product particle size distributions (Y2, Y3, Y4) may be included.

[0992] Step S3010 may include transmitting the received data from the user interface 210 to the internal state reference value generator 150c2 (see Figure 34 and / or Figure 35 and / or Figure 39 ).

[0993] As described above, the internal state reference value generator 150c2 is configured to generate a reference value corresponding to the desired discharge material state Y REF The relevant data is converted into an indication of the corresponding desired internal state X REF The data and / or indication corresponding to the desired toe position reference value X1 REF (r), FI REF (r) data.

[0994] Combine Figure 35 refer to Figure 34 , the internal state control system 1200 may also be configured as:

[0995] S3020: Make the user interface (210, 210S, 240, 250) transmit the desired internal state X corresponding to the instruction REF The information and / or indication of the corresponding desired toe position reference value (X1 REF (r), FI REF (r)) data, and

[0996] S3020: Cause the user interface (210, 210S, 240, 250) to transmit an indication of the actual toe position value (X1(r), FI(r), T) received from the monitoring module 150A, for example. D , R T (r);X6,A TOE (r)) information,

[0997] S3020: Receive the solid material feed rate (U2, R) via the user interface (210, 210S, 240, 250) S ) related to the first user input;

[0998] S3020: Generate solid material feed rate set point value (U2 SP , R SSP ), thereby affecting the internal state (X) to control or influence the desired discharge material state Y REF , product particle median size (Y2); where,

[0999] The generated solid material feed rate set point value (U2 SP , R SSP ) based on the received first user input.

[1000] According to one example, the solid material feed rate setpoint value U2 generated based on the received first user input SP The material in the rotating shell (20) is tumbled in the affected internal state (X) to generate product particles having a product particle median particle size (Y2) corresponding to the affected internal state (X) of the comminution process.

[1001] System for monitoring and providing operators with improved information content about the comminution process

[1002] Figure 39 is a block diagram of a system 1130 for monitoring an internal state X of a mill 10 and for providing an operator 230 of the mill 10 with improved information content.

[1003] System 1130 includes roller mill 10, as described above in conjunction with Figures 34 to 38 Discussed. Figure 39 , system 1130 is shown as a block diagram including a rolling mill, which is shown as a block 10 that receives multiple inputs U1, ...Uk and generates multiple outputs Y1, ...Yn. Thus, in terms of signal processing and analysis, the mill 10 receives an input vector U and generates an output vector Y in the manner discussed elsewhere in this document. Figure 39 The system 1130 may include multiple components and may be as described above with respect to Figure 1A and / or Figure 1B described and / or configured as described in any other examples described in this disclosure, for example, with respect to FIG. Figure 38 .

[1004] like Figure 39 As shown, the system 1130 includes a monitoring module 150A and / or a correlation module 150C. As described above, the correlation module 150C is operable to generate a correlation data set 1170 during operation of the mill 10 and / or the correlation module 150C is operable to associate the data set 1170 with the desired discharge material state Y. REF The relevant data is converted into an indication of the corresponding desired internal state X REFThe conversion step is based on the associated data set 1170 related to the operating mill 10.

[1005] Figure 39 The illustrated system 1130 includes an internal state control system 1200 , which includes an internal state reference value generator 150 c 2 , user interfaces 210 , 210B, and a regulator 240C.

[1006] The internal state control system 1200 may be configured to perform the following steps:

[1007] (Step S3000:) The user interface 210 transmits a request to the operator to provide an indication of the desired discharge material state Y REF As described above, the user input information indicates the desired discharge material state Y REF The user input may indicate at least one desired output material measurement value, such as Y1 and / or Y2. For example, the user input may indicate a desired product particle median size Y2. REF , and / or desired product particle size distribution Y3 REF 、Y4 REF , or the desired amount of material output per time unit Y1 REF .

[1008] This request S3000 may be generated by software included in the regulator 240C.

[1009] The internal state control system 1200 may also be configured to:

[1010] (Step S3005:) For example, a request indicating a desired discharge material state Y is received via the user interface 210. REF and / or desired product median particle size Y2 REF and / or data of the desired product particle size distribution Y2, Y3, Y4.

[1011] Furthermore, the internal state control system 1200 may be configured to perform a method comprising the following steps:

[1012] S3010: Generate the corresponding expected internal state X REF (also called the internal state reference vector X REF ), the expected internal state X REF Can include toe position reference value (X1 REF ;FI REF ). Internal state reference vector X REF Can be based on:

[1013] Indicates the desired ...

Claims

1. A system (150, 450, 1140, 150A, 150C) for monitoring and / or operating a comminution process in a tumble mill (10), the tumble mill comprising a rotatable shell (20) having an inner shell surface (22), the inner shell surface having a first number (L) of protrusions (310), the protrusions being configured to engage material as the shell (20) rotates, for grinding a charge of material (30) by tumbling the material in the rotating shell to generate product particles (95, 96) at a mill output (200), whereby when the protrusions (310) engage the toes (205) of the material (30), a rotational speed (U1, f1) of the rotatable shell (20) is caused. ROT )’s first repetition frequency (f R ) vibration (V IMP ); The system comprises: A sensor (10) for generating a signal indicating the vibration (V IMP ) signal (S FIMP ; S EA ,S MD ,Se(i),S(j),S(q)); A position signal generator for generating a position signal (E) indicating the rotational position of the rotatable housing (20) P , P(i), P(j), P(q)); One or more hardware processors configured to: generating at least one state parameter value (X1(r), FI(r)) indicative of an internal state (X) of the comminution process based on the vibration signal and the position signal; X2(r);X3(r);X4(r);X5,f ROT ;X6,A TOE (r);X7(r)); Solid material feed rate set point value (U2 SP , R SSP ), for setting a solid material feed rate (U2, R) for feeding the solid material (115) to the solid material feed input end (100) of the mill S ), the solid material feed rate (U2, R S ) is the amount of solid material feed particles (115) per unit time, and the solid material feed rate set point value (U2 SP , R SSP ) thereby affecting the internal state (X); an analyzer (1140) configured to analyze at least a portion of the product particles (96), and / or an analyzer (1140) configured to analyze at least a portion of the solid feed material (115); The analyzer (1140) is configured to generate at least one product measurement value (Y1, Y2, Y3, ... Yn) based on the analysis, the at least one product measurement value (Y1, Y2, Y3, ... Yn) indicating a discharged material state (Y(r), Y1, Y2, Y3, ... Yn); The system further comprises: For receiving an indication of the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,......Yn REF ) data input; A1) a user interface (210, 210S, 240, 250), configured to transmit an indication of at least one internal state parameter reference value (X1 REF (r), FI REF (r);X6 REF , A TOEREF (r)) information, the internal state parameter reference value (X1 REF (r), FI REF (r)) indicates the discharge material state (Y) for causing the desired REF (r))'s desired internal state (X REF );as well as A2) The user interface (210, 210S, 240, 250) is further configured to transmit an indication of the state parameter value (X1(r), FI(r); X6, A TOE (r)) information; or B) a user interface (210, 210S, 240, 250), configured to transmit an indication of at least one internal state parameter error value (X1 ERR (r), FI ERR (r);X6 ERR , A TOE_ERR (r)) information, the internal state parameter reference value (X1 ERR (r), FI ERR (r);X6 ERR , A TOE_ERR (r)) indicates the discharge material state (Y) for causing the desired REF (r))'s desired internal state (X REF ) and the state parameter value (X1(r), FI(r); X6, A TOE (r)) differences; as well as C1) The user interface (210, 210S, 240, 250) is configured to receive information related to the solid material feed rate (U2, R S ) is used to generate the solid material feed rate set point value (U2 SP , R SSP ), thereby affecting the internal state (X(r)) for controlling or influencing the discharged material state (Y(r)); or C2) The user interface (210, 210S, 240, 250) is configured to receive information related to the desired internal state (X REF ) related user input for use based on the desired internal state (X REF ) automatically generates the solid material feed rate set point value (U2 SP , R SSP ) in order to influence the internal state (X(r)) for controlling or influencing the discharged material state (Y(r)).

2. The system according to claim 1, wherein: The at least one state parameter value includes a value (X1(r), FI(r); X6, A) indicating the position of the toe (205); TOE (r)).

3. The system according to claim 1, further comprising: The one or more hardware processors are further configured to: The solid material feed rate set point value (U2) is generated based on the received first user input SP , R SSP ).

4. A system according to any preceding claim, wherein: The internal state (X) of the comminution process is indicated by a plurality of internal state parameters (X1, X2, X3, ..., Xm), wherein the internal state parameters (X1, X2, X3, ..., Xm) describe an aspect of the comminution process.

5. A system according to any preceding claim, wherein: The one or more hardware processors are configured to generate an internal state vector (X), which includes an integer m internal state parameters (X1, X2, X3, ..., Xm), wherein the internal state parameters (X1, X2, X3, ..., Xm) describe an aspect of the comminution process.

6. A system according to any preceding claim, wherein: The discharge material state (Y(r), Y1, Y2, Y3, ... Yn) is indicated by a plurality of discharge material state parameters (Y1, Y2, Y3, ... Yn), wherein a single discharge material state parameter (Y1, Y2, Y3, ... Yn) describes one aspect of the product particles (95, 96) generated at the mill output end (200).

7. A system according to any preceding claim, wherein: The one or more hardware processors are configured to generate a discharge material state vector (Y), the discharge material state vector (Y) comprising an integer n discharge material state parameters (Y1, Y2, Y3, ... Yn), wherein a single discharge material state parameter (Y1, Y2, Y3, ... Yn) describes an aspect of the product particles (95, 96) generated at the mill output end (200).

8. A system according to any preceding claim, wherein: The controller (150B) is configured to generate a set point vector (U SP ), the set point vector (U SP ) includes multiple set point parameter values ​​(U1 SP , U2 SP , U3 SP , ...Uk), where the single set point parameter value (U1 SP , U2 SP , U3 SP , ...Uk) describes a set point value for controlling one aspect of the operation of the mill, the solid material feed rate set point value (U2 SP , R SSP ) is the set point parameter value (U1 SP , U2 SP , U3 SP , ...Uk).

9. A system according to any preceding claim, wherein: The one or more hardware processors are configured to determine a desired internal state (X) of the shredding process based on the desired internal state (X) of the shredding process. REF ) generates the solid material feed rate set point value (U2 SP , R SSP ),in, The desired internal state (X) of the pulverization process REF ) based on the following: The desired discharge material state (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ),as well as Correlation data (1170, 1180) indicating a causal relationship between the desired discharged material state and a corresponding internal state (X) of the comminution process.

10. A system according to any preceding claim, wherein: The one or more hardware processors are configured to determine a desired internal state (X) of the shredding process based on the desired internal state (X) of the shredding process. REF ) generates the set point vector (U SP ),in, The desired internal state (X) of the pulverization process REF ) based on the following: The desired discharge material state (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ),as well as Correlation data (1170, 1180) indicating a causal relationship between the desired discharged material state and a corresponding internal state (X) of the comminution process.

11. A system according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes a value (Y1) indicating a desired weight per unit time of the product particles (95, 96) REF ).

12. A system according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes data indicating the expected median particle size (Y2 REF ) value (Y1 REF ).

13. A system according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes the desired discharge material state vector (Y REF (r)); The desired discharged material state vector (Y REF (r)) include: Data (Y1) indicating the desired weight per unit time of the product particles (95, 96) REF ), and / or Indicates the expected median particle size (Y2 REF ) data (Y2 REF ).

14. A method for monitoring and / or operating a comminution process in a tumble mill (10), the tumble mill comprising a rotatable shell (20) having an inner shell surface (22), the inner shell surface having a first number (L) of protrusions (310), the protrusions being configured to engage material as the shell (20) rotates, for grinding a charge of material (30) by tumbling the material in the rotating shell to generate product particles (95, 96), so that when the protrusion (310) engages with the toe (205) of the material (30), a rotation speed (U1, f) depending on the rotation speed of the rotatable shell (20) is caused ROT )’s first repetition frequency (f R ) vibration (V IMP ); The method comprises: A sensor (10) generates a signal indicating the vibration (V IMP ) signal (S FIMP ; S EA ,S MD ,Se(i),S(j),S(q)); A position signal generator including a position sensor is used to generate a position signal (E) indicating the rotational position of the rotatable housing (20). P , P(i), P(j), P(q)); At least one state parameter value (X1(r), FI(r); X2(r); X3(r); X4(r); X5, f) indicating an internal state (X) of the pulverizing process is generated by one or more hardware processors based on the vibration signal and the position signal. ROT ;X6,A TOE (r); X7(r)); The solid material feed rate set point value (U2 SP , R SSP ), for setting a solid material feed rate (U2, R) for feeding the solid material (115) to the solid material feed input end (100) of the mill S ), the solid material feed rate (U2, R S ) is the amount of solid material feed particles (115) per unit time, and the solid material feed rate set point value (U2 SP , R SSP ) thereby affecting the internal state (X); analyzing at least a portion of the product particles (96) to generate at least one product measurement value (Y1, Y2, Y3, . . . Yn) based on the analysis; and / or analyzing at least a portion of the solid feed material (115) to generate at least one product measurement value (Y1, Y2, Y3, . . . Yn) based on the analysis; The at least one product measurement value (Y1, Y2, Y3, ... Yn) is indicative of a discharged material state (Y(r), Y1, Y2, Y3, ... Yn); The method further comprises: A) receiving an indication of the desired discharge material status (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) data; and A1) transmitting, via a user interface (210, 210S, 240, 250), a reference value (X1) indicating at least one internal state parameter REF (r), FI REF (r);X6 REF , A TOEREF (r)) information; the internal state parameter reference value (X1 REF (r), FI REF (r)) indicates the discharge material state (Y) for causing the desired REF (r))'s desired internal state (X REF );as well as A2) transmits the state parameter value (X1(r), FI(r); X6, A2) via the user interface (210, 210S, 240, 250) TOE (r)) information; or B) transmitting, via a user interface (210, 210S, 240, 250), an error value (X1) indicating at least one internal state parameter ERR (r), FI ERR (r);X6 ERR , A TOEERR (r)) information, the internal state parameter reference value (X1 ERR (r), FI ERR (r);X6 ERR , A TOEERR (r)) indicates the discharge material state (Y) for causing the desired REF (r))'s desired internal state (X REF ) and the state parameter value (X1(r), FI(r); X6, the difference between ATOE(r); and C1) receives the solid material feed rate (U2, R) via the user interface (210, 210S, 240, 250) S ) is used to generate the solid material feed rate set point value (U2 SP , R SSP ), thereby affecting the internal state (X(r)) for controlling or influencing the discharged material state (Y(r)); or C2) receiving, via the user interface (210, 210S, 240, 250), the desired internal state (X REF ) related user input for use based on the desired internal state (X REF ) automatically generates the solid material feed rate set point value (U2 SP , R SSP ) in order to influence the internal state (X(r)) for controlling or influencing the discharged material state (Y(r)).

15. The method according to claim 14, wherein: The at least one state parameter value includes a value (X1(r), FI(r); X6, A) indicating the position of the toe (205); TOE (r)).

16. The method according to claim 14 or 15, further comprising: Based on the indication of the desired discharge material state (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) generates at least one internal state parameter reference value (X1 REF (r); X2 REF (r); X5 REF (r);X6 REF , A TOEREF (r)), the internal state parameter reference value (X1 REF (r);FI REF ) indicates the desired discharge material state (Y REF (r))'s desired internal state (X REF );and The solid material feed rate set point value (U2) is generated based on the following items: SP , R SSP ): The at least one internal state parameter reference value (X1 REF (r), FI REF (r); X6 REF , A TOEREF (r)); and The at least one state parameter value (X1(r), FI(r); X2(r); X3(r); x4(r);X5,f ROT ;X6,A TOE (r);X7(r))。 17. A method according to any preceding claim, wherein: The internal state (X) of the comminution process is indicated by a plurality of internal state parameters (X1, X2, X3, ..., Xm), wherein the internal state parameters (X1, X2, X3, ..., Xm) describe an aspect of the comminution process.

18. The method according to any preceding claim, further comprising: An internal state vector (X) is generated, the internal state vector (X) comprising an integer number m of internal state parameters (X1, X2, X3, ..., Xm), wherein the internal state parameters (X1, X2, X3, ..., Xm) describe an aspect of the comminution process.

19. A method according to any preceding claim, wherein: The discharge material state (Y(r), Y1, Y2, Y3, ... Yn) is indicated by a plurality of discharge material state parameters (Y1, Y2, Y3, ... Yn), wherein a single discharge material state parameter (Y1, Y2, Y3, ... Yn) describes one aspect of the product particles (95, 96) generated at the mill output end (200).

20. A system according to any preceding claim, further comprising: A discharge material state vector (Y) is generated, the discharge material state vector (Y) comprising an integer n discharge material state parameters (Y1, Y2, Y3, ... Yn), wherein a single discharge material state parameter (Y1, Y2, Y3, ... Yn) describes an aspect of the product particles (95, 96) generated at the mill output end (200).

21. The method according to any preceding claim, further comprising: Generate the set point vector (U SP ), the set point vector (U SP ) includes multiple set point parameter values ​​(U1 SP , U2 SP , U3 SP , ...Uk), where the single set point parameter value (U1 SP , U2 SP , U3 SP , ...Uk) describes a set point value for controlling one aspect of the operation of the mill, the solid material feed rate set point value (U2 SP , R SSP ) is the set point parameter value (U1 SP , U2 SP , U3 SP , ...Uk).

22. The method according to any preceding claim, further comprising: Based on the desired internal state (X) of the pulverization process REF ) generates the solid material feed rate set point value (U2 SP , R SSP ),in, The desired internal state (X) of the pulverization process REF ) based on the following: The desired discharge material state (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ),as well as Correlation data (1170, 1180) indicating a causal relationship between the desired discharged material state and a corresponding internal state (X) of the comminution process.

23. A method according to any preceding claim, wherein: Based on the desired internal state (X) of the pulverization process REF ) generates the set point vector (U SP ),in, The desired internal state (X) of the pulverization process REF ) based on the following: The desired discharge material state (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ),as well as Correlation data (1170, 1180) indicating a causal relationship between the desired discharged material state and a corresponding internal state (X) of the comminution process.

24. A method according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes a value (Y1) indicating a desired weight per unit time of the product particles (95, 96) REF ).

25. A method according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes data indicating the expected median particle size (Y2 REF ) value (Y1 REF ).

26. A method according to any preceding claim, wherein: Indicates the desired state of discharged material (Y REF (r), Y1 REF , Y2 REF , Y3 REF ,...Yn REF ) includes the desired discharge material state vector (Y REF (r)); The desired discharged material state vector (YREF(r)) includes: Data (Y1) indicating the desired weight per unit time of the product particles (95, 96) REF ), and / or Indicates the expected median particle size (Y2 REF ) data (Y2 REF ).

27. A method according to any preceding method claim, wherein: The vibration signal indicates a vibration signal having a first repetition frequency (f R ) of the vibration (V IMP ), the first repetition frequency (f R ) is the repetitive event of the shell (S P ) the number of occurrences of the rotation, wherein the first repetition frequency ((f R ) depends on the housing rotation speed (f ROT ), and among them, The reference signal indicates the reference position repetition frequency (f R ), the reference position repetition frequency (f R , X4) is equal to the pulse repetition frequency (f R ).

28. The method according to any preceding method claim, further comprising: Detecting a first reference position signal value (1; 1C, 0%) at the first occurrence in the time series of position signal sample values ​​(P(i), P(j), P(q)); Detecting a second reference position signal value (1; 1C; 100%) the second occurrence in the time series of said position signal sample values ​​(P(i), P(j), P(q)); Detection event signature (S P (r); Sp) the third occurrence in the time series of measured sample values ​​(Se(i), S(j), S(q)); Based on the position signal sample value (P(i); p(j); P(q)), generating (1010) a time series indicating that the first occurrence (1; 1C, 0%, 0°, P C ) and the second occurrence (1; 1C, P C , 100%, 360°) between the reference duration (T REF1 ) data; and A first duration (T) is generated (1020) based on REF2 ): time information (i, dt; j, t) associated with the third occurrence, and Time information (i, dt; j, t); The first duration indicates The first appears(1; 1C, 0%, 0°, P C )and The third Appearance (S P (r); Sp), or The first duration indicates The third Appearance (S P (r);Sp) and The second Appearance (1; 1C, P C , 100%, 360°); as well as Generate (1050) an indication of the first duration (T REF2 ) and the reference duration (T REF1 ) between the first time relationship (R T (r); T D ; FI(r), φ(r)) data.

Citation Information

Patent Citations

  • Energy-saving control method for ball-milling machine

    CN108452895A

  • Discharge from grinding mills

    CN1798613A

  • A Mill Process System

    SE2100048A1

  • Method of controlling the process of grinding materials in the mill

    SU1389847A1

  • System for controlling an internal state of a tumbling mill

    WO2021206600A1