METHOD FOR CALCULATING BULK MATERIAL CONVEYING QUANTITIES OR BULK MATERIAL LOADS OF A SWINGING MACHINE
Patent Information
- Application Number
- DE502021005756
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-09-25
- Filing Date
- 2021-09-01
- Publication Date
- 2024-11-28
- Estimated Expiration
- 2041-09-01
AI Technical Summary
Existing methods for determining the bulk material delivery rate of vibrating machines are complex and inefficient due to the need to account for multiple parameters and the difficulty in accurately measuring weight during oscillating movements, which can lead to inaccurate regulation of conveying capacity and potential overloads.
A method using raw measurement data from acceleration, speed, or displacement sensors to create feature data sets, which are then processed by an AI algorithm to develop a regression model that correlates vibration behavior with bulk material loading, allowing for accurate determination of delivery quantity and overload detection.
This approach enables precise calculation of bulk material delivery quantity and detection of overloads by accounting for individual machine characteristics, ensuring optimal utilization and extending machine service life by continuously adapting to changing conditions.
Description
[0001] The invention relates to a method for calculating bulk material conveying quantities or bulk material loadings of a vibrating machine.
[0002] Vibrating machines or vibrating conveyors, such as vibrating screens or vibrating conveyors, typically comprise a movable vibrating body with a screening surface or conveying surface, as well as a fixed support frame against which the movable vibrating body is mounted. Such vibrating machines are used for classifying and transporting bulk materials, for example, from a storage facility or hopper to a location where the bulk material is further processed.
[0003] To drive or move the vibrating body, either magnetic vibrating drives or so-called unbalance exciters are used. Unbalance exciters have rotating unbalances or weights that transfer their acceleration forces to the vibrating body, causing it to vibrate. Unbalance exciters that cause a directed movement of the vibrating body are called directional exciters.
[0004] The regulation and / or control of the conveying capacity or discharge of a vibrating machine is not an easy task to solve due to the many parameters that have to be taken into account, such as drive frequency, vibration amplitude, vibration angle, etc.
[0005] The weight measurement of the conveyed volume provides valuable information for the customer, enabling them to determine and control the conveying capacity of the vibrating machine. This allows the vibrating machine's performance potential to be fully utilized and overloads to be prevented.
[0006] A device and a method for adjusting the quantity of bulk material on a conveyor trough of a vibrating machine is known, for example, from DE 103 01 143 A1.
[0007] Here, a screening machine is mounted on its support frame via four spring elements. Four load beams or load cells are arranged between the spring elements and the support frame, acting as load sensors for determining the load on the conveyor chute. These load cells are used to continuously determine the actual weight of the vibrating machine and the existing load and compare it with a reference value.
[0008] Determining the weight of the conveyed bulk material is difficult due to the vibratory movements, which can be within the resonance range of the vibrating machine. Therefore, measuring the weight using load cells, force sensors with applied strain gauges that generate an electrical signal proportional to the weight based on their deformation, is not readily possible.
[0009] Another method for controlling the discharge of a vibrating trough is known from EP 1 188 695 A1. Here, acceleration sensors measure the vertical acceleration of the vibrating trough and also its drive frequency in order to derive the actual conveying capacity from these values.
[0010] However, this method requires the relationship between vertical acceleration, drive frequency, and the corresponding actual conveying capacity to be determined empirically or theoretically, for example, in the form of a function for each vibrating machine individually. JP 2011 084396 A discloses a method according to the preamble of claim 1, as well as a device according to the preamble of claim 10.
[0011] The object of the present invention is therefore to improve and simplify the determination of the bulk material conveying rate of a vibrating machine.
[0012] This object is achieved by a method according to patent claim 1 and a device according to patent claim 10. Advantageous embodiments of the invention emerge from the subclaims.
[0013] The invention provides a method for determining or calculating the bulk material conveying quantity or bulk material loading of a vibrating machine, in which raw measurement data of the vibrating machine are first recorded using at least one acceleration, speed or displacement sensor.
[0014] A vibrating machine in the sense of the invention is defined as
[0015] Vibrating conveyor machines such as vibrating screens or vibrating conveyors.
[0016] While in vibrating screens, the thickness of the bulk material being conveyed on the screen surface decreases along the conveying direction due to continuous screening, in vibrating trough conveyors, the material thickness remains essentially constant. As a result of changes in material thickness, the position of the center of gravity of a vibrating screen also changes. Despite constant drive conditions, such as a steady impact angle and steady stator frequency of the motor, a shift in the center of gravity changes the distance to the impact axis, which in turn results in a change in the pitching motion. As the material thickness changes and the center of gravity shifts, the amplitude and phase of the longitudinal acceleration also change. Furthermore, the mass of the bulk material load or the absolute load quantity influences, for example, the resulting stroke or maximum stroke of the vibrating machine.Furthermore, the load distribution of the bulk material on the loading or transport surface influences the longitudinal stroke of the vibrating machine. However, the load distribution of the bulk material does not influence the lateral stroke or (horizontal) lateral stroke of the vibrating machine.
[0017] Furthermore, there is a connection between the load and the rotational speed of the rotor or the unbalance exciter, since an increased inertial mass counteracts the centrifugal forces of the exciter and therefore leads to an increased moment of inertia on the drive shaft.
[0018] The applicant has therefore determined that no fixed algorithm can be used to determine the bulk material flow rate or bulk material loading of a vibrating machine. Rather, a learning AI algorithm is required to take into account the individual characteristics of a vibrating machine that influence its vibration behavior. Among other things, the mass of the vibrating machine, its geometry, the screen lining, and / or specific bulk material characteristics are taken into account.
[0019] In order to account for the individual vibration behavior of a vibrating machine, the method according to the invention records the raw measurement data: acceleration, velocity, or displacement for at least two different loading states of the vibrating machine, depending on the selected sensor. Of interest, for example, may be a measurement at 0% loading, when there is no bulk material on the vibrating machine, and a measurement at 100% loading, when the nominal load of the vibrating machine is reached. This raw measurement data is processed by an electronic evaluation unit in a single calculation into at least one characteristic from the list: amplitude, frequency, and phase.
[0020] Subsequently, the features obtained in this way are used to create so-called feature data sets. These feature data sets can consist of just one or several features from the list: amplitude, frequency, and / or phase. These feature data sets, or vibration signals, are saved and thus available for subsequent analysis.
[0021] It is important here that the feature data sets resulting from a vibrating machine A are only used for evaluations concerning the same vibrating machine A.
[0022] Investigations by the applicant have shown that the raw measurement data, or accordingly, the feature data sets or vibration signals, interact with the loading quantity of the vibrating machine. Therefore, the method according to the invention extracts or filters out from the feature data sets or vibration signals those indicators that exhibit a high correlation with a reference signal of the bulk material conveying rate or bulk material loading.
[0023] Based on the stored feature data sets, each of which demonstrates the correlation between load and vibration behavior, a classification model or regression model is then created and / or trained. The classification model or regression model assumes the indicators obtained from the feature data sets as input variables and a corresponding load specification, loading information—for example, 0% and 100% load—or a reference signal as the response variable.
[0024] If only a load information of 0-100% is used as the reaction variable, the regression model can at least estimate / calculate an absolute load in %. This information is also valuable for detecting overload or underload in a vibrating machine.
[0025] According to the invention, a reference signal or reference load signal can serve as the reaction variable. This reference signal can, for example, be a force measurement signal or a motor current signal resulting from an upstream, alternative, or indirect measurement process of the bulk material flow rate or bulk material load. For example, the reference load signal can be obtained by previously weighing the bulk material using a hopper scale or a gravimetric dosing device.
[0026] Alternatively, a feature of the acceleration, speed, or displacement sensor itself can be used as a reference signal.
[0027] By means of the method according to the invention, a multivariate classification model or regression model is generated which takes into account the indicators of the feature data sets in linear and / or non-linear form and, if necessary, including coefficients.
[0028] Based on the regression model, the prediction of another measured value is possible due to a large proportion of common variance between two variables and the knowledge of a measured value or reference value.
[0029] The classification model or regression model thus obtained can then be implemented in an electronic evaluation device in such a way that the actual load quantity of the vibrating machine can be determined and / or displayed on the basis of currently measured raw measurement data or vibration data.
[0030] One embodiment of the method provides for the step "Creating the classification or regression model" to be repeated after a period of time Δt, after the occurrence of wear on the vibratory conveyor, after maintenance measures, and / or after other system changes such as changes in the load, machine components, drive characteristics, or material properties. This ensures that the created regression model can be continuously adapted to changing boundary conditions. A change in the vibration behavior of a vibratory machine or vibratory conveyor resulting from wear or other changed conditions is often not immediately detectable.By repeating the measurements and training the created regression model - for example, by adjusting the input variables - it can be ensured that the determination of the bulk material flow rate or bulk material loading is carried out correctly despite system changes.
[0031] To create the regression model, an AI algorithm can be used that is weakly learning and thus adaptive, allowing it to adapt to the individual machine characteristics. The regression model itself, in the form of a mathematical equation, represents a working algorithm.
[0032] An advantageous embodiment of the method according to the invention also provides that only input variables that have a high correlation with the reference signal or reference load signal of the bulk material flow rate or bulk material load are used to create the regression model. A correlation calculation is used to determine which input variables are suitable for the regression model.
[0033] An embodiment of the method according to the invention provides that the regression model uses variables in the form C1*X1+C2*X2+C3*X2^2...CN*Xn^n, wherein the input variables X are taken into account based on the feature data sets as linear or non-linear factors and / or using coefficients C1, C2,..;CN ≥0.
[0034] To validate the method according to the invention, the regression model can advantageously be tested using feature data sets that were not used in the creation of the model. Historical data sets or feature data sets are thus divided into training and validation data sets.
[0035] To determine and display the bulk material conveyance rate or bulk material load of a vibratory conveyor, the invention also provides a device suitable for recording raw measurement data from the vibratory conveyor using at least one acceleration, speed, or displacement sensor. Furthermore, this device provides an electronic evaluation unit with which the raw measurement data is processed into at least one feature consisting of a direction-related vibration measurement variable from the list: amplitude, frequency, and phase.
[0036] Furthermore, the evaluation device generates feature sets consisting of at least one feature, and then uses these feature data sets to create a regression model. The device also has a display that indicates the bulk material load or bulk material conveyance rate of the vibratory conveyor based on the created model.
[0037] As described above, determining the bulk material loading or bulk material conveying rate of a vibrating machine or vibrating conveyor machine is a challenge, since both the bulk material quantity on the vibrating machine and the characteristics of the vibrating machine can change continuously.
[0038] Nevertheless, the bulk material load or bulk material flow rate represents valuable information. In addition to serving as a performance indicator for the vibrating machine, it also allows conclusions to be drawn about potential machine overload, which affects the service life of the vibrating machine. Conversely, a persistently low bulk material flow rate can indicate inefficient plant utilization. The method according to the invention therefore offers the advantage over previously known methods that the actual vibration behavior of the vibrating machine is taken into account. This allows, on the one hand, the underlying regression model to be continuously adjusted, but also—if adjustments to the regression model or algorithm become necessary—to detect malfunctions in the vibration behavior.
[0039] The method according to the invention is explained in more detail below using a process diagram, wherein further features and advantages of the invention are disclosed.
[0040] It shows Fig. 1 a schematic representation of the processes of the method according to the invention.
[0041] In Fig. 1The method according to the invention for calculating the bulk material conveying rate of a vibrating machine 1 in the form of a vibrating screen is shown schematically. At least one sensor 12 is attached to the vibrating machine 1 and is in data communication with a computing unit of an evaluation device 2. The data connection, which is shown in dashed lines in the figure, can be established via a radio connection, a wired connection, a permanent connection, or a temporary connection. The measurement data supplied by the sensor 12 are processed and stored in the computing unit into feature data sets 13. A regression model 6 is formed from the feature data sets 13, which serve as input variables, and reference signals 7, which originate from an upstream or separate measurement process for the bulk material loading. The regression model 6 based on the feature data sets 13 is validated and trained using feature data sets 9, which were not used to create the model.
[0042] The validated regression model 8 is then transferred to a software 10 and transferred to the evaluation device 2 in order to display the calculation of the bulk material loading.
Claims
1. A method for calculating the bulk material feed rate or the bulk material loading of a vibratory conveyor machine (1), wherein: a) raw measurement data of the vibratory conveyor machine (1) are recorded with at least one acceleration, velocity or displacement sensor (12) at at least two points in time with different loading states, b) raw measurement data are processed into at least one vibration data feature from the list: Amplitude, frequency, phase, with the following steps: c) create and store feature data sets (13) consisting of at least one vibration data feature, d) create a regression model (6) using the stored feature datasets (13), characterized by the step: e) determination and display of the current actual load of the vibratory conveyor machine (1) based on the created regression model (6) and at least one current feature data set (13).
2. The method according to claim 1, characterized in that the step d) of creating the regression model (6) is repeated after a period of time Δt, after the occurrence of wear on the vibratory conveyor machine (1), after maintenance measures and / or after other system changes, such as change of the loading, the machine components, the drive properties or of material properties.
3. The method according to any of claims 1 or 2, characterized in that the step a) of acquisition of the raw measurement data is carried out at least at 0% bulk material loading and 100% nominal load bulk material loading.
4. The method according to any one of the preceding claims, characterized in that, for training of the regression model (6), the model-based bulk material loading values are matched with a reference signal or reference load signal of the bulk material conveyor quantity or bulk material loading.
5. The method according to claim 4, characterized in that the reference load signal or reference signal is a force measurement signal or a motor current signal resulting from an upstream, alternative or indirect measurement process of the bulk material feed rate or bulk material load.
6. The method according to any one of claims 4 or 5, characterized in that for training of the regression model (6) only forecasting variables are used whose model-based bulk material loading value has a high correlation to the reference load signal of the bulk material conveyor quantity or bulk material loading.
7. The method according to any one of the preceding claims, characterized in that a multivariate regression method is used in the training phase of the regression model (6).
8. The method according to claim 7, characterized in that the regression model is created in the form C1*X1+C2*X2+C3*X2^2.... CN*Xn^n = bulk loading, wherein the forecasting variables X are considered as linear or nonlinear factors and / or by using coefficients C1, C2,...;CN ≥0.
9. The method according to any of the preceding claims, characterized in that the regression model (6) is validated using feature data sets (13) that have not been used for training the regression model.
10. A device for the determination of a bulk material conveyor quantity or a bulk material load of a vibratory conveyor machine (1), comprising - at least one acceleration sensor, velocity sensor or route sensor (12) arranged to acquire raw measurement data of the vibratory conveyor machine (1), - an electronic evaluation unit (2) for - processing the raw measurement data into at least one feature consisting of a directional vibration measurand from the list: Amplitude, frequency, and phase - creating feature sets (13) consisting of at least the one feature, - creating a regression model using the stored feature datasets (13), characterized in that the device comprises a screen or a display showing a model-based bulk material load value or model-based bulk material feed rate of the vibratory conveyor machine (1).
