Search

CN-116300566-B - New energy battery post-coating drying multivariable model predictive control system

CN116300566BCN 116300566 BCN116300566 BCN 116300566BCN-116300566-B

Abstract

The invention discloses a new energy battery post-coating drying multivariable model prediction control system, which performs real-time optimization control and rolling update by taking T APC as a calling period, specifically, the control system performs prediction and optimization calculation based on a model according to the opening of a return air valve, the frequency of an exhaust fan and the set value of the frequency of an external circulation fan in the last calling period, a measured value of the concentration of NMP at the current moment, a measured value of the pressure of an oven, a target value of the concentration of NMP and measurable external disturbance change data, outputs an optimal sequence of the opening of the return air valve, the frequency of the exhaust fan and the set value of the frequency of the external circulation fan in the next optimal period as an optimal result, and finally, writes back the output values of all input variables in the optimal sequence at the first moment to a relevant executor of a drying system to complete closed loop control. According to the invention, on the premise of ensuring the stability of the system and the non-standard concentration of the toxic solvent NMP, the optimal control of dynamic optimization based on prediction is realized, the real-time online optimization of the drying process after coating is realized, and the energy consumption is reduced while the cooperative and stable operation of a plurality of sections of ovens is ensured.

Inventors

  • CHU DANLEI
  • JIANG JINGBO

Assignees

  • 厦门奥普拓自控科技有限公司

Dates

Publication Date
20260508
Application Date
20221228

Claims (8)

  1. 1. The new energy battery post-coating drying multivariable model prediction control system is applied to a battery coating drying system, and comprises m sections of ovens, n sections of matched total air supply pipes and total air return pipes, wherein each section of oven comprises an electric air return valve, an internal circulation fan, an electric heating bag and an exhaust fan, and is characterized in that the air return quantity is controlled by adjusting the frequency of the m internal circulation fans and the opening degree of the m air return valves, the temperature in the ovens is controlled by adjusting the power of the m electric heating bags, the change of the pressure in the ovens and the change of the NMP concentration are compensated by adjusting the frequency of the m exhaust fans, the NMP concentration and the total NMP concentration of each section of oven are controlled by adjusting the frequency of the n external circulation fans, and n is smaller than m; The control system carries out real-time optimization control and rolling update by taking T APC as a calling period, specifically, the control system carries out model-based prediction and optimization treatment according to the opening of a return air valve, the frequency of an exhaust fan and the set value of the frequency of an external circulation fan in the last calling period, the measured value of the concentration of NMP at the current moment, the measured value of the pressure of an oven, the measured value of the concentration of NMP at the current moment and the measured value of the change of external disturbance data, optimally outputs an optimal sequence of the opening of the return air valve, the frequency of the exhaust fan and the set value of the frequency of the external circulation fan in the optimal sequence, and finally, writes back the output values of all input variables in the optimal sequence at the first moment to a relevant executor of the drying system to complete closed-loop control; the control system defines an input variable U and an output variable Y, and the multi-input multi-output mathematical relationship of the system is expressed as follows: (1) wherein G is a model matrix of (2m+1) x (3m+n), each element in the matrix is a transfer function of 1 x 1, U is a column vector of 3m+n rows, the model matrix comprises m return valve openings U 1 、u 2 、……、u m , m internal circulation fan frequencies U m+1 、u m+2 、……、u 2m , m exhaust fan frequencies U 2m+1 、u 2m+2 、……、u 3m and n external circulation fan frequencies U 3m+1 、u 3m+2 、……、u 3m+n , Y is a column vector of 2m+1 rows, the model matrix comprises m sections of oven NMP concentration Y 1 、y 2 、……y m and m sections of oven cavity pressure 、 、......、 And 1 total NMP concentration of the drying system According to the above description, the multivariable control problem is specifically expressed as: (2) in the formula, Representing a mathematical relation expression between the jth control variable and the ith controlled variable, namely a corresponding continuous time transfer function model; the control system converts the continuous time transfer function model into a discrete time system mathematical model, (3) The control system defines a target optimization function J 1 in the form of a constrained quadratic programming that includes predicted future time input and output variables as follows: (4) (5) (6) (7) (8) Wherein the symbols are Represents a future k + period, referenced at a kth control period, wherein, Called control step size, representing the input sequence Is a time span of (2); Called prediction step length, representing the predicted output trajectory Is a time span of (2); 、 A predicted sequence representing the future output variable at time { k, k+1, k+H y } and the input variable at time { k, k+1, k+Hu } from the beginning of the current k period; upper and lower bounds for the output variables; upper and lower bounds for the input variables; is the maximum rate of change, wherein, Representation of weighted Q 1 、Q 2 、Q 3 respectively corresponds to the weight matrix of each of three summation items in the objective function J 1 , Q 1 is a H y row H y column value matrix, Q 2 and Q 3 are H u row H u column value matrices, The target value of the output variable is manually set or set by checking the formula according to different batches by a user; The control system optimizes based on the model G (z) and the objective function J 1 , and completes the optimization control of each APC period by solving an input sequence with the minimum objective function value meeting constraint conditions through quadratic programming, wherein the first term on the right side of the equation of the formula (4) expresses the minimization of errors between a predicted track and a control target, the second term expresses the minimization of errors between a control variable U and a preset target thereof, and the third term expresses the minimization of a change index of the control variable U each time.
  2. 2. The predictive control system for a drying multivariable model after coating of a new energy battery according to claim 1, wherein the control system comprises the following specific implementation steps: First, based on the objective function and constraint conditions, the current input/output variable takes on value, and the adjustable weight Qi, i=1, 2,3, outputs the target value Solving the quadratic programming problem to obtain the optimal input sequence meeting the constraint condition ; Second, the optimal input sequence is selected Is the first element of (2) The rest of the input sequence elements are sent into a controller of the drying system to form on-line closed-loop control Discarding; Third, the next sampling time, the system will obtain new output value Then using k+1 time to update , ) Repeating the above two steps to obtain the optimal system input of the next step ; The above process is repeated to obtain the optimal input set value after each APC period rolling optimization The method is applied to a control system to finish the application of the multivariable model predictive controller.
  3. 3. The predictive control system for a drying multivariable model after coating of a new energy battery according to claim 1, wherein the control system performs dimension reduction treatment on input variables of the multivariable model, namely, separates the frequency U m+1 、u m+2 、……、u 2m of an internal circulation fan from U and does not participate in the optimization of APC; The frequency of the internal circulation fan is selected to be not involved in the optimization control, the frequency of the internal circulation fan is set to be an initialization given frequency value only after the batch is started and stopped, the initialization setting frequency value can be manually adjusted by an operator, a stable state can be achieved, the frequency can be secondarily adjusted to a new fixed value, or a condition judgment logic is programmed to solidify manual operation into a logic program, but the frequency does not participate in APC control.
  4. 4. The method for predicting and controlling the post-coating drying multivariable model of the new energy battery according to claim 1 or 3, wherein the control system performs dimension reduction on the frequencies of the external circulation fans in the input variables of the multivariable model matrix, and adopts a linkage design for a plurality of independent external circulation fan frequencies, namely selecting the ith external circulation fan frequency As a reference, other external circulation fans are regulated in frequency following manner as follows: (10) Where, when j=i, aj=1, bj=0, the corresponding original value is not corrected.
  5. 5. The system for predictive control of a post-coating oven-drying multivariable model of claim 1 or 3, wherein said control system compares static gain values of transfer functions of each row of the multivariable model matrix, and uses "0" instead of a model having an influence less than a set threshold when the model gain of the ith return air valve opening to NMP concentration and pressure of the jth (j not pi) oven is one fifth or less than that of the present oven (I=1, 2,..m, j +.i, j +.i+1; or i=m+1, m+2,..2, j +.i-m, j +.i-m+1; likewise, when the model gain of the ith exhaust fan for the jth (j +.i) oven NMP concentration and pressure is one fifth or less of the present oven, the "0" is replaced by (i=m+1,m+2,...,2m,j≠i-m,j≠i-m±1)。
  6. 6. The predictive control system for a drying multivariable model after coating a new energy battery according to claim 1 or 3, wherein the control system performs dimension reduction processing on the input variables of the multivariable model, wherein the control system links the linear relation function mapping of the outlet exhaust fan frequency and the inlet return air valve opening design of the same section in the input variables of the multivariable model, so as to combine the control designs of the two variables into the control design of a single variable, and the adjustment mode is as follows: (13) in the formula, And Respectively represents the frequency of the exhaust fan and the opening degree of the return air valve of the same section, And The following adjusting parameters of the ith section of oven exhaust fan are set manually according to the air quantity balance requirement.
  7. 7. The predictive control system for a drying multivariable model after coating a new energy battery according to claim 4, wherein the control system performs dimension reduction treatment on the input variables of the multivariable model, wherein the control system links the linear relation function mapping of the design of the frequency of the outlet exhaust fan and the opening of the inlet return air valve of the same section in the input variables of the multivariable model so as to combine the designs of the two variables into the design of a single variable, and the adjustment mode is as follows: (13) in the formula, And Respectively represents the frequency of the exhaust fan and the opening degree of the return air valve of the same section, And The following adjusting parameters of the ith section of oven exhaust fan are set manually according to the air quantity balance requirement.
  8. 8. The predictive control system for a drying multivariable model after coating a new energy battery according to claim 5, wherein the control system performs dimension reduction treatment on the input variables of the multivariable model, wherein the control system links the linear relation function mapping of the design of the frequency of the outlet exhaust fan and the opening of the inlet return air valve of the same section in the input variables of the multivariable model so as to combine the designs of the two variables into the design of a single variable, and the adjustment mode is as follows: (13) in the formula, And Respectively represents the frequency of the exhaust fan and the opening degree of the return air valve of the same section, And The following adjusting parameters of the ith section of oven exhaust fan are set manually according to the air quantity balance requirement.

Description

New energy battery post-coating drying multivariable model predictive control system Technical Field The invention belongs to the field of drying control after battery coating, and particularly relates to a new energy battery drying multivariable model predictive control system after coating. Background The drying process after the battery production line coating is an important link in the production of new energy batteries, and is a process of drying a wet film after the battery coating, and the drying operation is carried out on the coated coating in a drying chamber by using circulated high-temperature hot air. Since the NMP solvent is used for the battery cathode coating, the solvent is vaporized during the drying process to generate a weak toxic gas, and when the gas concentration reaches a certain strength, the risk of explosion exists, so that the NMP gas concentration needs to be strictly controlled within a safety threshold in the battery drying process. The battery coating and drying system adopts a cascade design of a plurality of ovens, the coated pole piece uniformly passes through a plurality of ovens at a set vehicle speed, the temperatures and NMP concentrations in different ovens are different, and the quality of the pole piece leaving the drying procedure and the drying process of the plurality of ovens have strong coupling relations, so that the battery coating and drying system is a typical multivariable, nonlinear and strong coupling complex control system with a certain time delay. The battery coating and drying system needs to ensure that the temperature and pressure in each oven can meet the control target of drying stably, and simultaneously avoid the concentration of NMP exceeding the standard, so the design of the control system is a typical multi-target control problem. In the existing actual operation process, a plurality of independent single-loop control designs are adopted under most conditions, and part of loops are fixed on a set value for a long time, so that real-time adjustment is difficult to be carried out according to the temperature, pressure and NMP concentration changes in an actual oven, and therefore, in order to prevent the NMP from exceeding the standard, the NMP is always in overshoot control for a long time, and the NMP concentration is reduced, and meanwhile, the waste of energy consumption is brought. In addition, battery pole pieces of different specifications correspond to different gains and time delays of dynamic response of a battery coating and drying system, NMP gas evaporation capacity generated in an oven is also different, further temperature and pressure changes in the oven are also different, optimal working points of different pole pieces are different, in the process of switching products and standby to production switching, a real-time controller is absent, the existing manual operation is often used for ensuring that NMP concentration reaches the standard, NMP concentration is greatly reduced by greatly increasing circulating air quantity in the switching process, and excessive low-temperature return air is heated before entering a drying chamber in a high-ventilation and quick-circulation mode, so that more electric energy is consumed, and a large amount of energy is wasted. In a word, the existing battery coating and drying system lacks effective real-time adjustment control means, the drying quality of the battery is unstable, and meanwhile energy consumption is wasted. Disclosure of Invention Aiming at the problems existing in the prior art, the invention aims to provide a new energy battery post-coating drying multivariable model predictive control system which can realize the online real-time optimization of the drying process while guaranteeing the stability of the drying system. In order to achieve the above purpose, the invention adopts the following technical scheme: The new energy battery post-coating drying multivariable model prediction control system comprises m sections of ovens, n sections of total air supply pipes and total air return pipes, wherein each section of oven comprises an electric air return valve, an internal circulation fan, an electric heating bag and an exhaust fan, the air return quantity is controlled by adjusting the frequency of the m internal circulation fans and the opening of the m air return valves, the temperature in the ovens is controlled by adjusting the power of the m electric heating bags, the change of the pressure in the ovens and the change of the NMP concentration are compensated by adjusting the frequency of the m exhaust fans, the NMP concentration and the total NMP concentration of each section of ovens are controlled by adjusting the frequency of the n external circulation fans, and n is smaller than m; The control system performs real-time optimization control and rolling update by taking T APC as a calling period, specifically, the control system performs model-based prediction and opt