Intelligent decision-making dynamic cooperative control method for roller cooling system

Abstract

The invention provides an intelligent decision-making dynamic cooperative control method for a roller cooling system, and relates to the field of industrial production intelligent control. The method comprises the following steps that the spraying state of each nozzle is monitored in real time; the minimum cooling liquid injection amount of process cooling in the rolling process is calculated; the sectional cooling flow in the rolling process is calculated, and the sectional cooling flow is not lower than the minimum cooling liquid injection amount; master-slave control is formed by process cooling and subsection cooling, the process cooling is responsible for macroscopic temperature field regulation and control, and local defect compensation is executed by the subsection cooling; when it is detected that the nozzles are blocked or fail, the flow compensation algorithm of the adjacent nozzles is started, and the standby micro nozzles are activated to execute fine compensation flow distribution; and after sectional cooling fails, on the basis of executing fine compensation flow distribution, roller fine cooling is adopted to eliminate defects of composite wave shapes, namely high-order plate shapes. The method can effectively solve the problem of high-order plate shape defects caused by failure of a sectional cooling system, and the stability and precision of plate shape control are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control of industrial production, and in particular to a dynamic collaborative control method for a roller cooling system with intelligent decision-making. Background Art

[0002] Cold-rolled steel strip is increasingly used in industries such as automotive, home appliances, and construction. These industries are placing increasing demands on the quality of cold-rolled steel strip, particularly regarding shape accuracy. Shape quality directly impacts subsequent processing performance and product lifespan. Therefore, improving cold-rolled shape control technology has become a critical issue for steelmakers.

[0003] At present, cold rolling plate shape control technology mainly relies on the segmented cooling system and work roll bending adjustment. However, the segmented cooling system may experience problems such as nozzle clogging or unstable injection volume during actual operation, resulting in poor cooling effect, which in turn affects the plate shape quality. When the segmented cooling system fails, it is usually necessary to rely on the work roll bending for plate shape adjustment, but it is mainly used to adjust symmetrical small plate shape defects, and has limited adjustment capabilities for high-order plate shape defects, and may cause abnormal fluctuations. Problems that will be faced at that time: Failure of the segmented cooling system. Nozzle clogging leads to insufficient coolant supply, which cannot meet the required cooling volume; excessive nozzle injection causes the coolant volume to exceed the standard, affecting the plate shape control effect; working roll bending adjustment limitations. It is mainly used to adjust symmetrical small plate shape defects, and has limited adjustment capabilities for high-order plate shape defects; improper adjustment may cause abnormal fluctuations, affecting the stability of the rolling process. Summary of the Invention

[0004] Purpose of the invention: To propose a dynamic collaborative control method for a roll cooling system with intelligent decision-making, which can maintain the overall cooling effect and reduce production interruptions when local nozzle failure occurs, thereby effectively solving the above-mentioned problems existing in the prior art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic collaborative control method for a roll cooling system with intelligent decision-making, comprising the following steps:

[0006] Add an injection status detection module and a failure prediction module to the cooling system architecture. Add a pressure sensor to the segmented cooling nozzles to monitor the blockage coefficient of each nozzle in real time, which is the flow deviation rate + pressure fluctuation amplitude.

[0007] The LSTM neural network is used to predict nozzle failure, and the input variable is X T :

[0008]

[0009] Where ΔQ nozis the nozzle flow deviation rate; H(p) t is the pressure fluctuation amplitude; is the roller surface temperature gradient; S vib It is the vibration spectrum characteristic of the nozzle.

[0010] Calculate the minimum coolant injection amount for process cooling during rolling. There are two calculation methods:

[0011] 1) Temperature theory method, through the friction heat generated by the friction between the roll and the strip, and the temperature change caused by the deformation work of the strip. According to the change of the strip temperature, the minimum cooling liquid injection amount Q corresponding to different temperature conditions can be determined according to the actual state during the rolling process. min .

[0012] 2) Online control calculation: during the rolling process, the actual rolling force and strip speed are used as the main factors to calculate the roll gap power. The minimum injection amount Q can be obtained through the relationship between the roll gap power and the minimum injection amount. min .

[0013] Calculate the segmented cooling flow rate during the rolling process to ensure that the segmented cooling flow rate is not less than the minimum coolant injection rate. Segmented cooling can effectively control high-order flatness defects in the strip.

[0014] 1) Calculate the residual flatness defects and convert them into the working roll cooling section. Then, mathematically process the residual flatness of each converted section to obtain the residual flatness deviation of each section of the segmented cooling corresponding to the strip. Then, use the PID controller to control the cooling flow of each nozzle.

[0015] 2) In actual rolling production control, segmented cooling is controlled by a PLC. The PLC can only calculate the control variable based on the deviation value at the sampling moment. Therefore, the integral term obtained from the above equation cannot be used directly and needs to be discretized to replace the continuous integration process. Therefore, a PI controller is constructed in the control system to calculate the output value.

[0016] Substituting the obtained output value into the corresponding relationship between the controller output and the segmented coolant injection amount determined by actual production, the segmented coolant injection amount for controlling the high-order residual of the plate shape defect can be obtained.

[0017] A multi-dimensional collaborative cooling architecture is established. Process cooling and segmented cooling form a master-slave control system. The master circuit: Process cooling is responsible for macroscopic temperature field control; the slave circuit: Segmented cooling performs local defect compensation. This ensures that deformation work and frictional heat during rolling do not cause roll and strip temperatures to exceed process cooling requirements, thus ensuring a stable rolling process. Furthermore, segmented cooling control ensures that strip shape accuracy meets quality requirements.

[0018] In order to achieve coordinated optimization control of basic cooling and segmented cooling, it is proposed to calculate additional plate shape defects through cooling amount deviation, and add the calculated plate shape deviation to the actual strip plate shape deviation measured in the plate shape control system to form the total plate shape deviation that the segmented cooling controller needs to control through segmented cooling. An iterative calculation process is formed through the plate shape segmented cooling control system and the process cooling control system until the plate shape control effect required for segmented cooling control is achieved while meeting the minimum cooling amount for process cooling.

[0019] In the actual rolling control system program, it is set that if the segmented cooling amount is less than the minimum cooling amount of process cooling, the system will determine that there is a deviation between the two. The additional plate shape deviation is calculated from the deviation cooling amount and added to the high-order residual of the strip shape defect. The plate shape segmented cooling control is used to eliminate the superimposed plate shape deviation, and finally the total rolling cooling amount that meets the two functional requirements of process cooling and segmented cooling is obtained.

[0020] A dynamic resource redistribution mechanism is established. When a nozzle is detected to be blocked or failed, the flow compensation algorithm of the adjacent nozzle is enabled and the spare micro-nozzle is activated to perform fine compensation flow distribution.

[0021] 1) Flow compensation algorithm calculation: First, the Voronoi diagram is constructed, then the effective coverage area is calculated, and finally the algorithm is implemented for the allocation of spatial weights and the calculation of flow compensation.

[0022] 2) When adjacent nozzles compensate for flow, only the approximate flow requirement is met. Therefore, backup micronozzles are needed for finer compensation. The calculation steps are as follows: First, residual defects are assessed after adjacent nozzle compensation; then, the micronozzle coverage is calculated; and finally, the additional flow required for each micronozzle is calculated based on the magnitude and location of the residual defect.

[0023] Use fine cooling control in flatness control. During the rolling process, quadruple flatness defects are prone to occur. When segmented cooling fails, even if the cooling amount is immediately compensated, this high-order flatness defect will still occur. Adjusting the work roll bending to adjust this high-order flatness defect is difficult to eliminate. Therefore, fine roll cooling is used after cooling compensation to eliminate compound wave shape, i.e., high-order flatness defects.

[0024] The least squares method is used to obtain a mathematical model represented by a fourth-order polynomial, which decomposes the adjustment deviation value into deviation values ​​corresponding to different powers of plate shape defects. In this way, the strip stress adjustment deviation existing in each measuring section is converted into adjustment deviation components corresponding to different plate shape defects.

[0025] In a further embodiment, the nozzle flow deviation rate ΔQ noz The calculation formula is as follows:

[0026]

[0027] Where, is the measured value of the piezoelectric flowmeter; Set the flow value for time t;

[0028] Pressure fluctuation amplitude H(p) t The calculation formula is as follows:

[0029]

[0030] Where p k The probability distribution of 32 equally spaced subintervals of the pressure sensor within 1 second, with a measurement range of 0-5 MPa;

[0031] Roller surface temperature gradient The calculation formula is as follows:

[0032]

[0033] Nozzle vibration spectrum characteristics S vib The calculation formula is as follows:

[0034]

[0035] Where A f is the vibration acceleration spectral density; α is the frequency attenuation factor; ω(f) is the frequency band weight coefficient.

[0036] In a further embodiment, the temperature theory method is calculated as follows:

[0037]

[0038] Where, ρ is the coolant density; c is the coolant specific heat capacity; dl is the unit arc length;

[0039] ΔT S is the temperature change caused by frictional heat and deformation work:

[0040]

[0041] Where Q F is the heat generated by the deformation work of the strip, Q RB is the heat generated by friction, C WB is the specific heat capacity of the strip, ρ B is the density of the strip, V is the volume of the strip;

[0042] The calculation formula of the online control method is as follows:

[0043] Q min =0.002Pgap

[0044] P gap is the roller gap power:

[0045]

[0046] Where K p is the model coefficient, F act is the actual rolling force, V act is the actual strip speed, and W is the actual strip width.

[0047] In a further embodiment, the residual flatness defect is calculated as follows:

[0048]

[0049] Where dev_c is the high-order residual of the strip shape defect; ref is the target strip shape; mes is the measured strip shape; a j is the j-th plate shape controller control gain; eff j is the unit adjustment coefficient of the j-th shape controller; j=1,…,m is the number of shape controllers.

[0050] In a further embodiment, the process of controlling the cooling flow rate of each nozzle using a PID controller is as follows:

[0051] The obtained residual flatness defects are processed mathematically to obtain the initial output:

[0052]

[0053] Where, e(t) is the deviation of the strip shape; u(t) represents the output of the PI controller related to segmented cooling; K P is the proportional link coefficient; T I is the integration time constant;

[0054] Construct a PI controller to obtain the output value. The output value expression is as follows:

[0055]

[0056] Where u(k) is the output value of the PI controller at the kth sampling moment; K P is the proportional gain coefficient related to the absolute value of the plate shape defect; K I is the integral gain coefficient related to the rolling speed; k is the sampling number.

[0057] In a further embodiment, process cooling and segmented cooling form a master-slave control, and the total cooling amount is expressed as follows:

[0058]

[0059] Where com_c is the additional deviation of the flatness; K com Q is the gain factor for calculating the additional flatness deviation; min is the minimum cooling amount of the process; Q PI Control the cooling amount of each section for plate shape; Cool num The total number of nozzles in the spraying state; Cool num The number of nozzles that is less than the minimum cooling amount of the process.

[0060] In a further embodiment, the flow rate compensation algorithm for adjacent nozzles is as follows:

[0061] Divide the roller surface into n polygonal regions and construct the Voronoi diagram:

[0062]

[0063] Where V i represents the inherent influence area of ​​the i-th nozzle; (x i ,y i ) is the plane coordinate of the nozzle;

[0064] Define the adjacent set N(k) of the failed nozzle k and calculate the expandable coverage area of ​​each adjacent nozzle

[0065]

[0066] Where, Area is the expandable coverage area of ​​the nozzle; d jk is the Euclidean distance between nozzle j and failure point k; λ is the distance attenuation coefficient;

[0067] An improved exponential weighting strategy is used to allocate spatial weights:

[0068]

[0069] Where α is the spatial sensitivity factor; R is the maximum effective compensation radius; d jk is the Euclidean distance between nozzle j and failure point k;

[0070] Calculate basic compensation and dynamic correction Get the final compensation flow in:

[0071]

[0072] Where, is the theoretical required flow rate of the blocked nozzle; γ is the temperature gradient influence coefficient; A0 is the reference coverage area; The original cooling capacity.

[0073] In a further embodiment, the fine filling flow distribution is specifically as follows:

[0074] Calculate the defect energy density E d :

[0075]

[0076] Where, is the amplitude of residual harmonics of plate shape after main compensation; is the kth harmonic allowable threshold; ω k is the harmonic weight coefficient;

[0077] The Gaussian diffusion model is used to define the cooling influence function f of a single micro nozzle. i (x,y), complete the modeling of the nozzle influence area:

[0078]

[0079] Where, is the theoretical flow rate of the micro nozzle; σ is the diffusion radius; (x i ,y i ) is the plane coordinate of the nozzle;

[0080] Calculate basic compensation and dynamic correction Get the final compensation flow in:

[0081]

[0082] Where, is the maximum flow rate of a single micro nozzle; η is the energy conversion coefficient; κ is the temperature curvature sensitivity coefficient; d i is the distance from the micro nozzle to the defect center; D max is the maximum effective range; It is the maximum flow rate of a single micro nozzle.

[0083] In a further embodiment, when segmented cooling fails, in addition to performing fine flow distribution, fine roller cooling is used to eliminate complex wave-shaped or high-order plate defects. This process is expressed by a polynomial:

[0084] y=A0+A1x+A2x 2 +A3x 3 +A4x 4

[0085] Where A1x is the adjustment deviation component corresponding to the primary flatness defect; A2x 2 A3x is the adjustment deviation component corresponding to the secondary flatness defect; 2, A4x 4 It is the adjustment deviation component corresponding to the third and fourth order flatness defects.

[0086] Beneficial Effects: This invention provides an intelligent decision-making method for dynamic coordinated control of a roll cooling system. By combining process cooling with segmented cooling control and precise roll control, it can reduce plate shape defects caused by segmented cooling system failures. This method enables real-time monitoring and failure warnings, resource coordination, and improved fault tolerance. Even in the event of local nozzle failure, the overall cooling effect can be maintained, minimizing production interruptions. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 It is a flow chart of the dynamic collaborative control method of the roll cooling system with intelligent decision-making of the present invention.

[0088] Figure 2 It is a flow chart of the dynamic resource reallocation mechanism of the present invention.

[0089] Figure 3 This is a comparison chart of working roll temperature after implementation. DETAILED DESCRIPTION

[0090] In the following description, numerous specific details are provided to provide a more thorough understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without one or more of these details. In other instances, certain technical features well known in the art have not been described to avoid confusion with the present invention.

[0091] To reduce plate shape defects caused by failure of segmented cooling systems, this paper proposes a new, intelligent, and coordinated dynamic control method for roll cooling systems. First, it enables real-time monitoring and failure warnings; second, it leverages resource coordination to improve fault tolerance. Even in the event of local nozzle failure, the overall cooling effect can still be maintained, minimizing production interruptions.

[0092] In order to give full play to its role, it is necessary to study the characteristics of the cooling system and formulate a reasonable adjustment strategy.

[0093] Step 1: Add an injection status detection module and a failure prediction module to the cooling system architecture. Add a pressure sensor to the segmented cooling nozzle to monitor the blockage coefficient of each nozzle in real time (i.e., the flow deviation rate + pressure fluctuation amplitude).

[0094] An improved LSTM neural network is used to predict nozzle failure. The input variables are:

[0095]

[0096] Where: ΔQnoz is the nozzle flow deviation rate, H(p) t is the pressure fluctuation amplitude, is the roller surface temperature gradient, S vib It is the vibration spectrum characteristic of the nozzle.

[0097] Flow deviation rate expression:

[0098]

[0099] Where: is the measured value of the piezoelectric flowmeter, Set the flow rate value for time t.

[0100] Pressure fluctuation amplitude expression:

[0101]

[0102] Where: p k The probability distribution of the pressure sensor in 32 equally spaced subintervals within 1 second, measuring range: 0-5MPa.

[0103] Roller surface temperature gradient expression:

[0104]

[0105] Vibration spectrum characteristic expression:

[0106]

[0107] Where: A f is the vibration acceleration spectral density; α = 0.75 is the frequency attenuation factor; ω(f) is the frequency band weight coefficient.

[0108] Sensors must be dustproof and waterproof to avoid interference from water mist and iron oxide scale in the rolling environment. Regularly calibrate sensors to prevent data drift and misjudgment. Avoid over-reliance on a single algorithm; instead, cross-validate it with physical principles. Detect nozzle blockages or leaks in advance to prevent uncontrolled coolant flow. Use predictive models to reduce downtime due to unexpected failures and reduce the frequency of forced intervention in the roll bending system.

[0109] Step 2: Calculate the minimum coolant injection amount for process cooling during rolling:

[0110] 1) Temperature theory method, through the friction heat generated by the friction between the roll and the strip, and the temperature change caused by the deformation work of the strip. According to the change of the strip temperature, the minimum cooling liquid injection amount Q corresponding to different temperature conditions can be determined according to the actual state during the rolling process. min .

[0111] Calculate the minimum cooling injection amount Q of the coolant according to the temperature theory method min expression:

[0112]

[0113] Where: ρ is the coolant density, c is the coolant specific heat capacity, dl is the unit arc length, ΔT S is the temperature change caused by frictional heat and deformation work, Q F is the heat generated by the deformation work of the strip, Q RB is the heat generated by friction, C WB is the specific heat capacity of the strip, ρ B is the density of the strip, and V is the volume of the strip.

[0114] 2) Online control calculation: during the rolling process, the actual rolling force and strip speed are used as the main factors to calculate the roll gap power. The minimum injection amount Q can be obtained through the relationship between the roll gap power and the minimum injection amount. min .

[0115] Calculate the roller gap power P based on online control gap , thus obtaining the minimum coolant injection amount Q min :

[0116] Q min =0.002P gap

[0117]

[0118] Where: P gap is the roll gap power, K p is the model coefficient, F act is the actual rolling force, V act is the actual strip speed, and W is the actual strip width.

[0119] This ensures the nozzle operates above the critical flow rate, preventing coolant stagnation and scaling due to low flow rates. The effect of coolant viscosity changes with temperature must be considered; the minimum flow rate must not be lower than the nozzle's critical anti-clogging flow rate.

[0120] Step 3: Calculate the segmented cooling flow rate during the rolling process. Segmented cooling can effectively control high-order strip shape defects.

[0121] 1) The residual flatness defect can be obtained through the formula, and the residual flatness defect is converted to the working roll cooling section. Then, the residual flatness of each section after conversion is mathematically processed to obtain the residual flatness deviation of each section of the segmented cooling corresponding to the strip. Then, the PID controller is used to control the cooling flow of each nozzle.

[0122] The expression for obtaining flatness defects is:

[0123]

[0124] Where: dev_c is the high-order residual of the strip shape defect, ref is the target strip shape, mes is the measured strip shape, a j is the j-th shape controller control gain, eff j is the unit adjustment coefficient of the j-th shape controller, j=1,…,m is the number of shape controllers.

[0125] The obtained residual plate shape defects are processed mathematically to obtain the initial output:

[0126]

[0127] Where: e(t) is the deviation of the strip shape, where e(t) = r(t) - y(t), u(t) represents the PI controller output related to segmented cooling, K P is the proportional link coefficient, T I is the integration time constant.

[0128] 2) In actual rolling production, segmented cooling is controlled by a PLC. The PLC can only calculate the control variable based on the deviation value at the sampling moment. Therefore, the integral term obtained from the above equation cannot be used directly and needs to be discretized to replace the continuous integration process. Therefore, a PI controller is constructed in the control system to calculate the output value.

[0129] Substituting the obtained output value into the corresponding relationship between the controller output and the segmented coolant injection amount determined by actual production, the segmented coolant injection amount for controlling the high-order residual of the plate shape defect can be obtained.

[0130] Construct a PI controller to obtain the output value. The output value expression is:

[0131]

[0132] Where: u(k) is the output value of the PI controller at the kth sampling moment; K P is the proportional gain coefficient related to the absolute value of the plate shape defect; K I is the integral gain coefficient related to the rolling speed; k is the sampling number.

[0133] Through differentiated control of zoned flow, high-order plate shape defects can be directly suppressed, the compensation pressure of the working roll bending roll can be reduced, and its over-limit fluctuation can be avoided.

[0134] Step 4: Establish a multi-dimensional collaborative cooling architecture. Process cooling and segmented cooling form a master-slave control system: the master loop: process cooling is responsible for macroscopic temperature field control; the slave loop: segmented cooling performs local defect compensation. This ensures that deformation work and frictional heat during rolling do not cause roll and strip temperatures to exceed process cooling requirements, thus ensuring a stable rolling process. Furthermore, segmented cooling control ensures that strip shape accuracy meets quality requirements.

[0135] In order to achieve coordinated optimization control of basic cooling and segmented cooling, it is proposed to calculate additional plate shape defects through cooling amount deviation, and add the calculated plate shape deviation to the actual strip plate shape deviation measured in the plate shape control system to form the total plate shape deviation that the segmented cooling controller needs to control through segmented cooling. An iterative calculation process is formed through the plate shape segmented cooling control system and the process cooling control system until the plate shape control effect required for segmented cooling control is achieved while meeting the minimum cooling amount for process cooling.

[0136] In the actual rolling control system program, it is set that if the segmented cooling amount is less than the minimum cooling amount of process cooling, the system will determine that there is a deviation between the two. The additional plate shape deviation is calculated from the deviation cooling amount and added to the high-order residual of the strip shape defect. The plate shape segmented cooling control is used to eliminate the superimposed plate shape deviation, and finally the total rolling cooling amount that meets the two functional requirements of process cooling and segmented cooling is obtained.

[0137] The total cooling amount of rolling required by the two functional requirements of process cooling and segmented cooling is expressed as follows:

[0138]

[0139] Where: com_c is the additional deviation of the flatness, K com To calculate the gain factor of the additional flatness deviation, Q min is the minimum cooling amount of the process, Q PI Control the cooling amount of the plate shape, Cool num is the total number of nozzles in the spraying state, Cool num The number of nozzles that is less than the minimum cooling amount of the process.

[0140] Avoid excessive roll bending system response when cooling fails, reducing fluctuation risks. Improve overall control efficiency through division of labor and collaboration. A priority arbitration mechanism should be established. Regularly check communication links to prevent signal interruptions that could lead to control disruptions.

[0141] Step 5: Establish a dynamic resource redistribution mechanism. When a nozzle is detected to be blocked or failed, the flow compensation algorithm of the adjacent nozzle is activated, and the spare micro nozzle is activated for fine filling.

[0142] 1) Flow compensation algorithm calculation: First, the Voronoi diagram is constructed, then the effective coverage area is calculated, and finally the algorithm is implemented for the allocation of spatial weights and the calculation of flow compensation.

[0143] Flow compensation algorithm calculation:

[0144] (1) Construction of Voronoi diagram

[0145] Parameter input:

[0146] The plane coordinates of all nozzles (x i ,y i )(i=1,2,…,n); strip width W and rolling direction length L.

[0147] Calculation process:

[0148] Fortune algorithm is used to generate Voronoi diagram and divide the roller surface into n polygonal areas:

[0149]

[0150] Where: V i represents the inherent influence area of ​​the i-th nozzle.

[0151] (2) Calculation of effective coverage area

[0152] Define the adjacent set of failed nozzle k:

[0153]

[0154] Calculate the expandable coverage area of ​​each adjacent nozzle:

[0155]

[0156] Where: Area is the expandable coverage area of ​​the nozzle, d jk is the Euclidean distance between nozzle j and failure point k; λ = 25 mm is the distance attenuation coefficient.

[0157] (3) Spatial weight allocation

[0158] Adopting an improved index weighted strategy:

[0159]

[0160] Where: α = 0.15 is the spatial sensitivity factor, R = 50 mm is the maximum effective compensation radius, d jk is the Euclidean distance between nozzle j and failure point k.

[0161] (4) Calculation of flow compensation

[0162] Basic compensation amount:

[0163]

[0164] Where: is the basic compensation amount, is the theoretical required flow rate for a blocked nozzle.

[0165] Dynamic correction amount:

[0166]

[0167] Where: is the dynamic correction amount, γ=0.03 ​​is the temperature gradient influence coefficient, A0=100mm 2 The base coverage area.

[0168] Final compensation flow:

[0169]

[0170] Where: For the final compensation flow, is the original cooling capacity, is the basic compensation amount, is the dynamic correction amount.

[0171] 2) When adjacent nozzles compensate for flow, only the approximate flow requirement is met. Therefore, backup micronozzles are needed for finer compensation. The calculation steps are as follows: First, residual defects are assessed after adjacent nozzle compensation; then, the micronozzle coverage is calculated; and finally, the additional flow required for each micronozzle is calculated based on the magnitude and location of the residual defect.

[0172] Fine compensation calculation method:

[0173] (1) Quantification of residual defects

[0174] Input parameters:

[0175] Plate shape residual harmonic amplitude after main compensation: Defect area coordinates: (x d ,y d ).

[0176] Defect energy density calculation:

[0177]

[0178] Where: is the amplitude of the plate shape residual harmonics after main compensation, is the kth harmonic threshold, ω k is the harmonic weight coefficient.

[0179] (2) Micro-nozzle matching mapping

[0180] Modeling of nozzle influence area:

[0181] The Gaussian diffusion model is used to define the cooling influence function of a single micro nozzle:

[0182]

[0183] Where: is the theoretical flow rate of the micro nozzle, and σ = 2.5 mm is the diffusion radius.

[0184] (3) Fine-filling flow distribution

[0185] Basic supplement amount:

[0186]

[0187] Where: As the basic supplement amount, is the maximum flow rate of a single micro nozzle, and η=0.18 is the energy conversion coefficient.

[0188] Dynamic correction items:

[0189]

[0190] Where: is the dynamic correction value, κ=0.05 is the temperature curvature sensitivity coefficient, d i is the distance from the micro nozzle to the defect center, D max =10mm is the maximum effective distance.

[0191] Final fine filling flow:

[0192]

[0193] Where: is the final fine filling flow, As the basic supplement, is the dynamic correction amount, It is the maximum flow rate of a single micro nozzle.

[0194] Even if a local nozzle fails, the overall cooling effect can still be maintained, minimizing production interruptions. Avoid single-point failures that could cause entire lines to shut down. Avoid over-reliance on a single compensation nozzle by limiting its maximum flow rate. Establish a "cold zone blacklist" to deactivate nozzles that fail more than three times in a row and issue an alarm. Consider the rolling direction when compensating.

[0195] Step 6: Use fine cooling control in flatness control. During the rolling process, quadruple flatness defects are prone to occur. When segmented cooling fails, even if the cooling amount is immediately compensated, this high-order flatness defect will still appear. Adjustment of the work roll bending to this high-order flatness defect is difficult to eliminate. Therefore, after cooling compensation, fine roll cooling is used to eliminate compound wave shape, i.e., high-order flatness defects.

[0196] Using the least squares method, a mathematical model represented by a fourth-order polynomial was obtained, which decomposed the adjustment deviation value into deviation values ​​corresponding to plate defects of different powers. In this way, the strip stress adjustment deviation existing in each measuring section was converted into adjustment deviation components corresponding to different plate defects.

[0197] Polynomial representation:

[0198] y=A0+A1x+A2x 2 +A3x 3 +A4x 4

[0199] Where: A1x is the adjustment deviation component corresponding to the primary flatness defect; A2x 2 A3x is the adjustment deviation component corresponding to the secondary flatness defect; 3 , A4x 4 It is the adjustment deviation component corresponding to the third and fourth order flatness defects.

[0200] The intelligent decision-making dynamic cooperative control method of the roll cooling system proposed in this invention was tested in a 1450mm cold rolling strip shape control system. Figure 3 As shown, curve a is the working roll temperature of the dynamic collaborative control method of the roll cooling system using intelligent decision-making, and curve b is the working roll temperature of the dynamic collaborative control method of the roll cooling system without intelligent decision-making. By comparing and analyzing the two sets of data, it can be seen that: with the intelligent collaborative cooling control system using dynamic compensation, the temperature distribution of the working roll is more uniform, and the plate quality of the strip is improved.

[0201] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0202] The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, a computer, a server or a data center to another website, a computer, a server or a data center by a wired (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0203] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0204] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes may be made to it in form and detail without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A dynamic collaborative control method for a roll cooling system with intelligent decision-making, characterized in that: The steps include: Add pressure sensors in the segmented cooling nozzles to monitor the blockage coefficient of each nozzle in real time; Calculate the minimum coolant injection amount for process cooling during rolling; Calculating the segmented cooling flow rate during the rolling process, wherein the segmented cooling flow rate is not less than the minimum coolant injection amount; Process cooling and segmented cooling form a master-slave control system. Process cooling is responsible for macro temperature field control, while segmented cooling performs local defect compensation. When a nozzle is detected to be clogged or failed, the flow compensation algorithm of the adjacent nozzle is enabled and the spare micro nozzle is activated to perform fine compensation flow distribution; When segmented cooling fails, on the basis of executing fine filling flow distribution, fine cooling of the roll is used to eliminate compound wave shape, that is, high-order plate shape defects.

2. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 1, characterized in that: The LSTM neural network is used to predict nozzle failure, and the input variable is X T : Where ΔQ noz is the nozzle flow deviation rate; H(p) t is the pressure fluctuation amplitude; is the roller surface temperature gradient; S vib It is the vibration spectrum characteristic of the nozzle.

3. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 2, characterized in that: The nozzle flow deviation rate ΔQ noz The calculation formula is as follows: Where, is the measured value of the piezoelectric flowmeter; Set the flow value for time t; The pressure fluctuation amplitude H(p) t The calculation formula is as follows: Where p k The probability distribution of 32 equally spaced subintervals of the pressure sensor within 1 second, with a measurement range of 0-5 MPa; The roller surface temperature gradient The calculation formula is as follows: The nozzle vibration spectrum characteristic S vib The calculation formula is as follows: Where A f is the vibration acceleration spectral density; α is the frequency attenuation factor; ω(f) is the frequency band weight coefficient.

4. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 1, characterized in that: The minimum coolant injection amount is calculated according to a temperature theory method or an online control method; The temperature theory method calculation formula is as follows: Where, ρ is the coolant density; c is the coolant specific heat capacity; dl is the unit arc length; ΔT S is the temperature change caused by frictional heat and deformation work: Where Q F is the heat generated by the deformation work of the strip, Q RB is the heat generated by friction, C WB is the specific heat capacity of the strip, ρ B is the density of the strip, V is the volume of the strip; The online control method is calculated as follows: Q min =0.002P gap P gap is the roller gap power: Where K p is the model coefficient, F act is the actual rolling force, V act is the actual strip speed, and W is the actual strip width.

5. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 1, characterized in that: The calculation of the segmented cooling flow rate during the rolling process specifically includes: Calculate the residual flatness defects and transfer them to the work roll cooling section; mathematically process the residual flatness of each section after conversion to obtain the residual flatness deviation of each section of the strip cooling process; and use a PID controller to control the cooling flow rate of each nozzle. The calculation formula of the residual flatness defect is as follows: Where dev_c is the high-order residual of the strip shape defect; ref is the target strip shape; mes is the measured strip shape; a j is the j-th plate shape controller control gain; eff j is the unit adjustment coefficient of the j-th shape controller; j=1,…,m is the number of shape controllers.

6. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 5, characterized in that: The PID controller is used to control the cooling flow of each nozzle, specifically including: The obtained residual flatness defects are processed mathematically to obtain the initial output: Where, e(t) is the deviation of the strip shape; u(t) represents the output of the PI controller related to segmented cooling; K P is the proportional link coefficient; T I is the integration time constant; Construct a PI controller to obtain the output value. The output value expression is as follows: Where u(k) is the output value of the PI controller at the kth sampling moment; K P is the proportional gain coefficient related to the absolute value of the plate shape defect; K I is the integral gain coefficient related to the rolling speed; k is the sampling number.

7. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 1, characterized in that: Process cooling and segmented cooling form a master-slave control, and the total cooling amount is expressed as follows: Where com_c is the additional deviation of the flatness; K com Q is the gain factor for calculating the additional flatness deviation; min is the minimum cooling amount of the process; Q PI Control the cooling amount of each section for plate shape; Cool num The total number of nozzles in the spraying state; Cool num The number of nozzles that is less than the minimum cooling amount of the process.

8. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 1, characterized in that: The flow compensation algorithm for enabling adjacent nozzles specifically includes: Divide the roller surface into n polygonal regions and construct the Voronoi diagram: Where V i represents the inherent influence area of ​​the i-th nozzle; (x i ,y i ) is the plane coordinate of the nozzle; Define the adjacent set N(k) of the failed nozzle k and calculate the expandable coverage area of ​​each adjacent nozzle Where, Area is the expandable coverage area of ​​the nozzle; d jk is the Euclidean distance between nozzle j and failure point k; λ is the distance attenuation coefficient; An improved exponential weighting strategy is used to allocate spatial weights: Where α is the spatial sensitivity factor; R is the maximum effective compensation radius; d jk is the Euclidean distance between nozzle j and failure point k; Calculate basic compensation and dynamic correction Get the final compensation flow in: Where, is the theoretical required flow rate of the blocked nozzle; γ is the temperature gradient influence coefficient; A0 is the reference coverage area; The original cooling capacity.

9. The method for dynamic coordinated control of a roll cooling system with intelligent decision-making according to claim 8, characterized in that: The activation of the standby micro-nozzle to perform fine filling flow distribution specifically includes: Calculate the defect energy density E d : Where, is the amplitude of residual harmonics of plate shape after main compensation; is the kth harmonic allowable threshold; ω k is the harmonic weight coefficient; The Gaussian diffusion model is used to define the cooling influence function f of a single micro nozzle. i (x,y), complete the modeling of the nozzle influence area: Where, is the theoretical flow rate of the micro nozzle; σ is the diffusion radius; (x i ,y i ) is the plane coordinate of the nozzle; Calculate basic compensation and dynamic correction Get the final compensation flow in: Where, is the maximum flow rate of a single micro nozzle; η is the energy conversion coefficient; κ is the temperature curvature sensitivity coefficient; d i is the distance from the micro nozzle to the defect center; D max is the maximum effective range; It is the maximum flow rate of a single micro nozzle.

10. The intelligent decision-making dynamic collaborative control method for a roll cooling system according to claim 1, characterized in that: When segmented cooling fails, on the basis of performing fine filling flow distribution, fine roller cooling is used to eliminate compound wave shape, that is, high-order plate shape defects. This process can be expressed by a polynomial: y=A0+A1x+A2x 2 +A3x 3 +A4x 4 Where A1x is the adjustment deviation component corresponding to the primary flatness defect; A2x 2 A3x is the adjustment deviation component corresponding to the secondary flatness defect; 3 , A4x 4 It is the adjustment deviation component corresponding to the third and fourth order flatness defects.

Images