A control method and system for the die head motor of a coater based on PID-type iterative learning

Abstract

The present invention discloses a coating machine die motor control method and system based on PID type iterative learning, belonging to the technical field of automatic control. The method includes discretizing the surface density error and the error change rate to form multiple intervals, and setting the initial coefficient according to different intervals; obtaining the data of the coating machine system operating parameters, the surface density weighing system and the die motor in real time, and calculating the current surface density error and error change rate; querying the proportional parameter P corresponding to the interval in which it is located, and inputting it into the proportional controller to obtain the first adjustment amount corresponding to each partition; querying the historical iteration record of the corresponding interval through the PID type iterative controller, and calculating the second adjustment amount corresponding to each partition output by the iterative control part; superimposing the first adjustment amount of each partition with the second adjustment amount to generate the final adjustment amount. The present invention takes into account the simplicity and ease of implementation of traditional proportional control and the adaptive characteristics of iterative control, making the control more flexible, stable and low-cost.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic control, and particularly relates to a control method and system for a die head motor of a coater based on PID-type iterative learning. Background Art

[0002] The coater is used to coat key components such as the positive electrode, negative electrode, and separator to ensure their performance and quality. Among them, for the extrusion coater, the lateral adjustment can precisely control the position of the coating head or nozzle by controlling the movement of the slider with a motor, ensuring that the coating material evenly covers the electrode material.

[0003] In the coater system, classical PID and modified PID algorithms are widely used because of their advantages such as good stability and simple implementation. However, these algorithms have difficulties such as difficult parameter adjustment. Although there are many current non-linear complex control systems, in the lateral closed-loop control of the coater, due to the excessive time-consuming parameter adjustment of the PID system in the motor control of the coater die head, it is difficult to establish an accurate mathematical model, resulting in these algorithms being unable to achieve good results. At the same time, since the coefficients of PID are fixed, it cannot adapt to changes in working conditions. In order to overcome the difficulty of manual debugging of parameters, directly adopting PID-type iterative control has problems such as large storage requirements and slow iterative update. Summary of the Invention

[0004] The purpose of the present invention is to provide a control method and system for a die head motor of a coater based on PID-type iterative learning, which can save storage space and accelerate the iterative speed.

[0005] The technical solutions adopted by the present invention are specifically as follows:

[0006] A control method for a die head motor of a coater based on PID-type iterative learning includes:

[0007] S1. Discretize the surface density error and the error change rate to form multiple intervals, and set the initial coefficients of the proportional controller and PID-type iterative control according to different intervals, and load the configuration parameters saved at the previous shutdown.

[0008] S2. Real-time obtain the operation parameters of the coater system, the data of the surface density weighing system and the die head motor, and after filtering the data, calculate the current surface density error and the error change rate according to the filtered data.

[0009] S3. According to the current surface density error and the error change rate, query the proportional parameter P corresponding to the interval where they are located, and input it into the proportional controller to obtain the first adjustment amount corresponding to each partition.

[0010] S4. Based on the data processed in S2, query the historical iteration records in the corresponding interval through a PID-type iterative controller. Combine the current areal density error, error change rate, and historical iteration records to calculate the corresponding second adjustment amount for each partition output by the iterative control part;

[0011] S5. Superimpose the first adjustment amount and the second adjustment amount for each partition to generate a final adjustment amount and drive the die head motor of the corresponding partition to perform the adjustment;

[0012] S6. Loop through steps S2 - S5.

[0013] In a preferred solution, the steps of setting the coefficients of the proportional controller and the PID-type iterative control in step S1 include:

[0014] According to the production process parameters, perform fuzzy processing on the areal density error and the error change rate, and divide them into multiple discrete intervals;

[0015] Each discrete interval is independently associated with the proportional parameter P and the parameters of the PID-type iterative controller respectively, for realizing differential control in different intervals;

[0016] Set the initial coefficients of the proportional controller and the PID-type iterative control according to the corresponding discrete interval.

[0017] In a preferred solution, the filtering of the data in step S2 includes: abnormal data filtering and mean filtering;

[0018] Among them, the abnormal data filtering adopts upper and lower limit filtering. When the data exceeds the upper limit value or the lower limit value, this abnormal data is filtered out to exclude invalid data. The invalid data includes: data at the moments when the die head of the coater advances or retracts, and other abnormal moments of the coater, as well as abnormal millimeter areal density data;

[0019] The formula for mean filtering is as follows:

[0020] Among them, m is the sliding window size, g is the input value, x is the input parameter, t is the time serial number, and z is the data after mean filtering.

[0021] In a preferred solution, calculating the areal density error and the error change rate according to the filtered data in step S2 includes:

[0022] Use the data after mean filtering and the target value of the current production to calculate to obtain all the current areal density errors;

[0023] Calculate the areal density error and the error change rate;

[0024] The calculation of the areal density error is as follows:

[0025] wherein, is the surface density error of partition r in the t-th pass, is the surface density value of partition r in the t-th pass, is the target value of the surface density specified by the production process;

[0026] The error change rate is calculated as follows:

[0027] wherein, is the surface density error of partition r in the (t - 1)-th pass.

[0028] In a preferred embodiment, the historical iteration record includes the adjustment amounts of the historical outputs of the iterative control parts of each motor, and the corresponding surface density errors and error change rates.

[0029] In a preferred embodiment, obtaining the first adjustment amount corresponding to each partition in step S3 includes:

[0030] Using the intervals corresponding to the surface density error and the error change rate, querying the corresponding proportional parameter P, and then calculating the adjustment amount of the proportional controller for each partition;

[0031] wherein, is the adjustment amount output by the proportional controller, is the proportional parameter of the proportional controller, which is obtained by querying the intervals of the surface density error and the error change rate.

[0032] In a preferred embodiment, the PID-type iterative formula for calculating the second adjustment amount corresponding to each partition of the output of the iterative control part in step S4 is:

[0033] wherein, and , are respectively the three PID coefficient values of the PID-type iterative control, t is the t-th adjustment, is the surface density error of the r-th partition in the t-th adjustment, is the error change rate of the r-th partition in the t-th adjustment, which is obtained by subtracting the previous adjustment error from the current adjustment error by querying the intervals corresponding to the surface density error and the error change rate, is the cumulative adjustment error of the r-th partition, which is obtained by cumulative calculation by querying the historical adjustment errors of the intervals corresponding to the surface density error and the error change rate, is the adjustment amount output by the iterative controller of the r-th partition in the (t - 1)-th pass, is the adjustment amount output by the iterative controller of the r-th partition in the t-th pass.

[0034] In a preferred embodiment, the adjustment amount of the proportional controller part in step S5 plus the adjustment amount of the iterative control part is the final adjustment amount. According to the final adjustment amount, the motor corresponding to each zone executes its corresponding adjustment amount;

[0035] After execution, store the result of this iteration in the data memory for the next iteration.

[0036] The present invention also provides a die head motor control system for a coater based on PID-type iterative learning, including a data acquisition module, a data memory, a control algorithm unit, and an execution module;

[0037] Among them, the data acquisition module is used to obtain the operating parameters of the coater system, the surface density weighing system, and the data of the die head motor in real time, store the data in the data memory, filter the data and use it as the input of the control algorithm unit, and the execution module executes the adjustment according to the output of the control algorithm unit.

[0038] The technical effects achieved by the present invention are:

[0039] By discretizing the surface density error and the error change rate, dividing the continuous numerical values into multiple intervals, the data storage requirements are reduced and the iterative learning speed is accelerated. This design effectively solves the problems of large memory occupation and slow query speed caused by continuous state recording in traditional PID-type iterative control, especially suitable for the characteristics of poor fluidity of high-viscosity materials in the coater, making the influence of the adjustment amount change on the system more controllable.

[0040] The present invention adopts a segmented proportional control strategy, dynamically adjusts the proportional parameter according to different surface density error and error change rate intervals, and enhances the nonlinear fitting ability of the system. Combining the historical error cumulative adjustment mechanism of PID-type iterative control, dynamic compensation is achieved through the PID-type iterative formula, so that the surface density error gradually converges in multiple iterations, resulting in a significant improvement in control accuracy.

[0041] The present invention combines the simplicity of traditional proportional control and the self-adaptability of iterative control, retains the advantage of fast response of proportional adjustment, and realizes parameter self-tuning through iterative learning. This method avoids the need for complex mathematical modeling, and at the same time uses historical data to optimize the adjustment amount, and finally achieves the comprehensive effects of flexible control, high stability and low hardware cost, meeting the dual requirements of reliability and economy in industrial scenarios. At the same time, the present invention indirectly reflects the error between the adjustment amount and the ideal adjustment amount through the error between the surface density and the target value of the surface density, and uses it to replace the error between the adjustment amount and the ideal adjustment amount, enabling normal iteration under the premise that the ideal adjustment amount is unknown. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is the system flow chart of the present invention; Specific Embodiments

[0043] The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work belong to the scope of protection of the present invention.

[0044] In the description of the present invention, it should be noted that the technical features involved in the different implementation methods of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0045] Please refer to Figure 1 , a control method for the die head motor of a coater based on PID-type iterative learning, including the following steps:

[0046] S1. Discretize the surface density error and the error change rate to form multiple intervals, and set the initial coefficients of the proportional controller and the PID-type iterative control according to different intervals, and load the configuration parameters saved at the previous shutdown (parameters such as the previous adjustment error);

[0047] Among them, during the coater production, the process personnel of the manufacturer will design the production process parameters of this product. Therefore, according to the process parameters, the surface density error and the error change rate are fuzzified, and they are divided into multiple discrete intervals (corresponding to the multiple intervals formed by the above discretization). Each discrete interval is independently associated with the proportional parameter P and the parameters of the PID-type iterative controller, which can achieve differential control in different intervals. For the storage and query of the iterative data used for iterative control, that is, different surface density error and error change rate intervals have independent proportional parameters P and iterative controller parameters. Thus, the initial coefficients of the proportional controller and the PID-type iterative control can be set according to the corresponding intervals, including the proportional coefficient and the three PID coefficient values of the PID-type iterative control;

[0048] S2. Real-time obtain the operation parameters of the coater system, the data of the surface density weighing system and the die head motor, and after filtering the data, calculate the current surface density error and the error change rate according to the filtered data;

[0049] Among them, the filtering of the data includes: abnormal data filtering and mean filtering;

[0050] The abnormal data filtering adopts upper and lower limit filtering. When the data exceeds the upper limit value or the lower limit value, this abnormal data is filtered out to exclude invalid data, including: the data of the die head advancing and retracting the knife of the coater and other abnormal moments of the coater and abnormal millimeter surface density data;

[0051] The mean filtering process is mainly used to smooth data and reduce fluctuations. Its processing formula is as follows:

[0052] Among them, m is the sliding window size, g is the input value, x is the input parameter, t is the time serial number, and z is the data after mean filtering processing.

[0053] The processed data is used to calculate the areal density error;

[0054] Among them, is the areal density error of partition r in the t-th pass, is the areal density value of partition r in the t-th pass, is the target value of the areal density specified by the production process.

[0055] Among them, the error change rate is calculated as follows:

[0056] is the areal density error of partition r in the (t - 1)-th pass.

[0057] After obtaining the areal density error and the error change rate, query the interval where the proportional parameter table is located, so as to obtain the coefficient.

[0058] S3. According to the current areal density error and the error change rate, query the proportional parameter P corresponding to its interval, denoted as , and input it into the proportional controller to obtain the corresponding first adjustment amount for each partition;

[0059] Here, the first adjustment amount is the output corresponding to the proportional controller obtained through the proportional control formula;

[0060] Among them, is the first adjustment amount output by the proportional controller, is the coefficient value of the proportional controller, which is obtained by querying the interval of the surface error and the error change rate.

[0061] S4. According to the data processed in S2, query the historical iteration records of the corresponding interval through the PID-type iterative controller, and combine the current areal density error, the error change rate and the historical iteration records to calculate the corresponding second adjustment amount for each partition output by the iterative control part;

[0062] In this step, query the iterative historical records in the same state according to the areal density error and the error change rate, such as the adjustment amounts output by the iterative control parts of each motor and the corresponding areal density errors and error change rates, and then calculate the corresponding second adjustment amount for each partition output by the iterative control part;

[0063] Meanwhile, since the areal density error and the error change rate are continuous values, if the states are recorded continuously, there will be a problem that the state space is too large, the iteration speed is slow, and the memory requirement is high. Therefore, in this application, the areal density error and the error change rate are discretized into multiple intervals, which can save storage space and speed up the iteration. At the same time, since the production material of the coater is a high-viscosity liquid with poor fluidity, increasing the adjustment amount by 1 results in only a very small change in the flow state, which is also the main reason for discretizing the iteration space. Moreover, the non-linear ability of the iterative control is low. To increase the non-linear ability of the control, the basic proportional control is segmented, that is, different intervals of the areal density error and the error change rate correspond to different basic proportional controls. This can increase the fitting ability of the control system and make the control more accurate.

[0064] Specifically, query the corresponding iteration parameters in the data memory according to the intervals where the current areal density error and the error change rate are located, including the output of the previous iteration term and the historical adjustment error.

[0065] Moreover, due to the complex state of the coater die head, it is difficult to establish an effective system mathematical model. Therefore, the ideal adjustment amount for motor control in different states is unknown.

[0066] In theory, the difference between the adjustment amount output by the controller and the ideal adjustment amount is used as the error of iterative learning in iterative control. Since the ideal adjustment amount cannot be obtained, and the error between the adjusted areal density and the target value of the areal density can indirectly reflect the error between the adjustment amount and the ideal adjustment amount, therefore, it is used to replace the error between the adjustment amount and the ideal adjustment amount as the error of iterative learning. The PID-type iterative formula is:

[0067] where 、 , are the three PID coefficient values of the PID-type iterative control respectively, t is the t-th adjustment, is the areal density error of the t-th adjustment in the r-th partition, is the error change rate of the t-th adjustment in the r-th partition, which is obtained by subtracting the previous adjustment error from the current adjustment error corresponding to the intervals of the areal density error and the error change rate, is the cumulative adjustment error in the r-th partition, which is obtained by cumulative calculation by querying the historical adjustment error corresponding to the intervals of the areal density error and the error change rate, is the adjustment amount output by the iterative controller of the (t - 1)-th adjustment in the r-th partition, is the adjustment amount output by the iterative controller of the t-th adjustment in the r-th partition.

[0068] S5. Superimpose the first adjustment amount and the second adjustment amount of each zone to generate a final adjustment amount and drive the die head motor of the corresponding zone to perform the adjustment, which is obtained through the following formula;

[0069] where, is the total adjustment amount for the t-th adjustment of the r-th zone.

[0070] Since the role of the proportional controller therein is an initial adjustment amount, and the adjustment amount output by the iterative controller is an iterative change amount. Therefore, the adjustment amount executed by the final execution module is the sum of the adjustment amounts of the proportional controller and the iterative controller.

[0071] S6. Loop and execute steps S2 - S5.

[0072] The following combines a calculation example to illustrate the actual calculation process of the algorithm of the present invention. There are 25 zones on the production line in this example. The data of one of the zones is selected for display and analysis. These 3 adjustments are not continuous in time, and there will be adjustments in other intervals in between.

[0073] Table 1: Cumulative adjustment times table within the same surface density error and surface density error change rate intervals.

[0074]

[0075] According to the process parameters, perform fuzzification, which discretizes the surface density error and the error change rate into 7 sets respectively, corresponding to 36 intervals. Different intervals correspond to different proportional control parameters and iterative control storage spaces.

[0076] Parameters of the initial PID-type iterative controller 、 , are respectively set to 1.2, 0.01, 0.2.

[0077] Table 2: Proportional parameter table of the proportional controller.

[0078]

[0079] When the algorithm outputs for the first time, the surface density error is -1.1, and the surface density error change rate is 1. As shown in Table 2, query the proportional parameter table of the proportional controller to obtain = 6.

[0080] The adjustment amount output by the proportional controller = 6*(-1.1) = -6.6.

[0081] Since it is the first adjustment and t = 0, that is, the zero-th adjustment, there is no previous adjustment information, so The surface density error of the t-th adjustment of the r-th partition is , The error change rate of the t-th adjustment of the r-th partition is , The cumulative adjustment error of the r-th partition is , so the , , and of the iterative controller are all 0. Therefore, the adjustment amount = 0.

[0082] The total adjustment amount of this output = + = -6.6 + 0 = -6.6. After adjustment, the data collector detects that the error between the surface density after this adjustment and the target surface density is -0.5, that is, the adjustment error = -0.5.

[0083] When the algorithm is output for the second time, the surface density error is -1.1, and the surface density error change rate is 1. As shown in Table 2, querying the proportional parameter table of the proportional controller gives = 6. Querying the data memory, the control error of the previous trip is obtained as = -0.5 and = 0.

[0084] The adjustment amount output by the proportional controller this time = 6 * (-1.1) = -6.6.

[0085] The adjustment amount of the iterative controller = + * + * + * = 0 + 1.2 * (-0.5) + 0.01 * (-0.5) + 0.2 * (-0.5) = -0.705.

[0086] The total adjustment amount of this output = + = -0.705 - 6.6 = -7.305. After adjustment, the data collector detects that the error between the surface density after this adjustment and the target surface density is 0.1, that is, = 0.1.

[0087] When the algorithm output is executed for the third time, the surface density error is -1.1, and the change rate of the surface density error is 1. As shown in Table 2, by querying the proportional parameter table of the proportional controller, = 6.

[0088] The adjustment amount output by the proportional controller = 6 * (-1.1) = -6.6.

[0089] The adjustment amount of the iterative controller = + * + * + * = -0.705 + 1.2 * 0.1 + 0.01 * (-0.5 + 0.1) + 0.2 * (0.1 - (-0.5)) = -0.469.

[0090] The total adjustment amount of this output = + = -0.469 - 6.6 = -7.069. After the adjustment, the data collector detects that the error between the surface density after this adjustment and the target surface density is 0, that is, = 0.

[0091] In the present invention, since the parameter adjustment of the PID system in the motor control of the coater die takes too much time, and at the same time, since the coefficients of the PID are fixed, it cannot adapt to the changes in the working conditions. In order to overcome the difficulty of manual debugging of the parameters, the PID-type iterative control is directly adopted, which has problems such as large storage requirements and slow iterative update. Therefore, the present invention discretizes the surface density error and the error change rate, which can save storage space and speed up the iterative speed. At the same time, since the production material of the coater is a high-viscosity liquid with poor fluidity, increasing one adjustment amount results in very little change in the flow state, which is also the main reason for discretizing the iterative space. And in order to avoid the low non-linear ability of the iterative control caused by the above operations and increase the non-linear ability of the control, the present application segments the basic proportional control, that is, different intervals of the surface density error and the error change rate correspond to different basic proportional controls. This can increase the fitting ability of the control system and make the control more accurate.

[0092] The present invention also claims protection for a coater die motor control system based on PID-type iterative learning, including: a data acquisition module, a data memory, a control algorithm unit, and an execution module;

[0093] Among them, the data acquisition module is used to obtain the operation parameters of the coater system, the data of the surface density weighing system and the die head motor in real time, store the data in the data memory, filter the data and use it as the input of the control algorithm unit, and the execution module executes the adjustment according to the output of the control algorithm unit.

[0094] Obviously, the above embodiments are only examples for clear illustration and not limitations on the implementation methods. For those of ordinary skill in the art, other different forms of changes or variations can be made based on the above description. It is not necessary and impossible to enumerate all implementation methods here. And the obvious changes or variations derived therefrom are still within the protection scope of the present invention.

Claims

1. A coating machine die motor control method based on PID iterative learning, characterized in that: include: S1. Discretize the surface density error and error change rate to form multiple intervals, set the initial coefficients of the proportional controller and PID iterative control according to different intervals, and load the configuration parameters saved at the last shutdown; S2, real-time acquisition of the operating parameters of the coating machine system, the surface density weight measurement system and the die head motor data, and after filtering the data, the current surface density error and error change rate are calculated according to the filtered data; S3, according to the current surface density error and error change rate, query the proportional parameter P corresponding to the interval in which it is located, and input it into the proportional controller to obtain the first adjustment amount corresponding to each partition; S4, according to the data processed by S2, query the historical iteration record of the corresponding interval through the PID type iterative controller, combine the current surface density error, error change rate and historical iteration record, calculate the second adjustment amount corresponding to each partition output by the iterative control part, and the historical iteration record includes the adjustment amount of the historical output of the iterative control part of each motor and the corresponding surface density error and error change rate; S5, superimposing the first adjustment amount and the second adjustment amount of each partition to generate a final adjustment amount and drive the die head motor of the corresponding partition to perform adjustment; S6, loop through steps S2-S5; The step S3 of obtaining the first adjustment amount corresponding to each partition includes: Using the intervals corresponding to the surface density error and the error change rate, the corresponding proportional parameter P is queried, and then the adjustment amount of the proportional controller of each partition is calculated; The calculation of the adjustment amount is as follows: in, is the regulation amount output by the proportional controller, is the proportional parameter of the proportional controller, is the surface density error of partition r in the tth pass, which is obtained by interval query of surface density error and error change rate; The PID-type iterative formula for calculating the second adjustment amount corresponding to each partition output by the iterative control part in step S4 is: in, , , are the three PID coefficient values ​​of PID iterative control, t is the tth adjustment, is the surface density error of the t-th adjustment of the r-th partition, is the error change rate of the tth adjustment of the rth partition, which is obtained by subtracting the last adjustment error and the current adjustment error in the interval corresponding to the surface density error and the error change rate. is the cumulative adjustment error of the rth partition, which is calculated by querying the historical adjustment error of the corresponding interval of the surface density error and the error change rate. is the adjustment value output by the iterative controller of the t-1th iteration of the rth partition, It is the adjustment value output by the iterative controller of the tth pass in the rth partition.

2. A coating machine die motor control method based on PID iterative learning according to claim 1, characterized in that: The step of setting the initial coefficients of the proportional controller and the PID type iterative control in step S1 includes: According to the production process parameters, the surface density error and error change rate are fuzzy processed and divided into multiple discrete intervals; Each discrete interval is independently associated with a proportional parameter P and a parameter of a PID-type iterative controller to achieve differentiated control between discrete intervals; The initial coefficients of the proportional controller and PID-type iterative control are set according to the corresponding discrete intervals.

3. A coating machine die motor control method based on PID iterative learning according to claim 1, characterized in that: The filtering of data in step S2 includes: abnormal data filtering and mean filtering; Among them, the abnormal data filtering adopts upper and lower limit filtering. When the data exceeds the upper or lower limit, the abnormal data is filtered out to exclude invalid data. Invalid data includes: coating machine die head advance and retreat, other coating machine abnormal time data and abnormal millimeter surface density data; The mean filter processing formula is as follows: Among them, m is the sliding window size, g is the input value, x is the input parameter, t is the time series number, and z is the data after mean filtering.

4. A coating machine die motor control method based on PID iterative learning according to claim 3, characterized in that: The step S2 calculates the surface density error and the error change rate according to the filtered data, including: The data after mean filtering is calculated with the target value of this production to obtain all current surface density errors; Calculate the surface density error and error change rate; The surface density error is calculated as follows: in, is the surface density error of partition r in the tth pass, is the surface density value of partition r in the tth pass, The target value of the area density specified for the production process; The error rate of change calculation is as follows: in, is the surface density error of partition r in the t-1th pass.

5. The coating machine die motor control method based on PID iterative learning according to claim 1, characterized in that: The first adjustment amount and the second adjustment amount in step S5 are superimposed to form a final adjustment amount. According to the final adjustment amount, the motor corresponding to each partition executes its corresponding adjustment amount. After execution, the result of this iteration is stored in the data storage for the next iteration.

6. A coating machine die motor control system based on PID iterative learning, characterized in that: Used to implement the coating machine die motor control method based on PID iterative learning as described in any one of claims 1 to 5, the system includes a data acquisition module, a data storage device, a control algorithm unit and an execution module; Among them, the data acquisition module is used to obtain the data of the coating machine system operating parameters, the surface density weight measurement system and the die head motor in real time, and store the data in the data storage device. After filtering the data, it is used as the input of the control algorithm unit, and the execution module performs adjustment according to the output of the control algorithm unit.

Images