PID self-tuning control method applied to lithium battery industry

Abstract

The invention provides a PID self-tuning control method applied to the lithium battery industry, and the method comprises the steps: obtaining operation parameters of lithium battery production equipment and scene features containing scene type identifiers and material attribute parameters, and carrying out the matching of a PID adjustment model in which adaptive dynamic characteristic parameters are pre-stored; after the deviation between the operation parameters and the scene target value is calculated, the deviation is dynamically corrected based on the material attribute parameters and the association rules; carrying out feature analysis on the corrected deviation by combining the influence weights of the dynamic characteristic parameters and the material attributes on the deviation trend; and then according to an analysis result, calling a parameter adjustment strategy corresponding to the scene to correct a PID adjustment model, and finally outputting PID parameters to adjust the lithium battery production equipment, so as to solve the problems of insufficient control precision and poor scene adaptability caused by the fact that PID parameter setting is not deeply coupled with lithium battery production scene characteristics and material attributes in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of PID deviation correction, and in particular to a PID self-tuning control method applied to the lithium battery industry. BACKGROUND

[0002] As the core energy storage device of the new energy industry, lithium ion batteries have been widely used in electric vehicles, energy storage systems and other fields. The high-precision control of the production process directly determines the performance and safety of the product. In the key processes of lithium battery production such as coating, rolling, liquid injection and sintering, the stability of process parameters such as temperature, pressure, tension and flow is crucial. The PID (Proportional-Integral-Derivative) control, due to its simple structure and fast response, has become the mainstream technology for parameter control. The accuracy of its parameter tuning directly affects the production efficiency and product consistency.

[0003] In the prior art, the PID parameter control of lithium battery production equipment mainly relies on manual tuning based on experience, and the proportional coefficient (Kp), integral time (Ti) and derivative time (Td) are adjusted through trial and error method. It is suitable for simple working conditions but difficult to cope with complex dynamic scenarios. In addition, a general industrial PID self-tuning algorithm can be used to automatically generate parameters based on system step response or critical oscillation characteristics.

[0004] However, the existing technology has obvious limitations: first, it does not consider the differences between multiple scenarios in lithium battery production, and the general PID model is difficult to adapt to the inertia and lag characteristics of the scene, resulting in insufficient control accuracy; second, it ignores the influence of material attribute parameters on process parameters. When the material characteristics fluctuate, the PID controller with fixed parameters cannot dynamically correct the deviation, which easily leads to overshoot or oscillation; third, the correlation between parameter adjustment strategy and scene characteristics and deviation analysis results is weak, making it difficult to adaptively select the optimization direction according to the real-time deviation mode, resulting in low self-tuning efficiency and poor robustness. SUMMARY

[0005] The present application provides a PID self-tuning control method applied to the lithium battery industry to solve the problem of insufficient control accuracy and poor scene adaptability caused by the uncoupling of PID parameter tuning with scene characteristics and material attributes in the prior art.

[0006] In one aspect, the present application provides a PID self-tuning control method applied to the lithium battery industry, comprising: obtaining the running parameters and scene characteristics of the lithium battery production equipment, wherein the scene characteristics include scene type identification, material attribute parameters and scene target values preset based on the scene characteristics; matching the corresponding PID adjustment model according to the scene type identification, wherein the PID adjustment model pre-stores dynamic characteristic parameters, control accuracy threshold, initial parameter configuration and material attribute deviation correlation rules adapted to the scene type identification; calculate the deviation of the running parameter from the scene target value, and dynamically correct the deviation based on the material attribute parameter and the material attribute deviation correlation rule; Based on the dynamic characteristic parameter, the influence weight of the material attribute parameter on the deviation change trend is fused, and the corrected deviation is characterized to obtain a characteristic analysis result; According to the characteristic analysis result, the parameter adjustment strategy corresponding to the scene characteristic is called, and the parameter configuration of the PID adjustment model is corrected based on the parameter adjustment strategy; the parameter adjustment strategy includes an iterative algorithm, a parameter adjustment step and a boundary constraint rule adapted to the material attribute parameter; Based on the parameter configuration, the PID parameter is output to adjust the lithium battery production equipment.

[0007] Optionally, the corresponding PID adjustment model is matched according to the scene type identifier, comprising: Call the initial PID adjustment model corresponding to the scene type identifier from the pre-constructed PID model library; Based on the material attribute parameter, the dynamic characteristic parameter and the initial parameter configuration of the initial PID adjustment model are adaptively adjusted through the pre-set material attribute deviation correlation rule; Verify whether the control accuracy threshold of the adjusted initial PID adjustment model matches the scene target value; If the control accuracy threshold matches the scene target value, a PID adjustment model is generated; If the control accuracy threshold does not match the scene target value, the control accuracy threshold is calibrated based on the scene target value until the parameters of the initial PID adjustment model are completely adapted to the scene characteristics, and the PID adjustment model is generated.

[0008] Optionally, the deviation of the running parameter from the scene target value is calculated, and the deviation is dynamically corrected based on the material attribute parameter and the material attribute deviation correlation rule, comprising: Calculate the initial deviation of the running parameter from the scene target value; Extract the change amount of the material attribute parameter, which is the difference between the material attribute parameter and the material attribute reference value; Call the material attribute deviation correlation rule, and correct the initial deviation based on the weight factor and the change amount through a weighted summation formula to obtain the deviation; If the ratio of the absolute value of the deviation to the control accuracy threshold is greater than a pre-set accuracy threshold, a deviation truncation mechanism is started to constrain the deviation within the pre-set accuracy threshold of the control accuracy threshold.

[0009] Optionally, based on the dynamic characteristic parameter, the influence weight of the material attribute parameter on the deviation change trend is fused to perform feature analysis on the corrected deviation, and a feature analysis result is obtained, including: A dimension of the feature analysis is determined according to the dynamic characteristic parameter, and the dimension is adapted to the response characteristic of the scene; Based on historical influence data of the material attribute parameter, the influence weight of the material attribute parameter on the deviation change trend is calculated by a gradient boosting tree algorithm; The deviation is analyzed in time sequence, the change mode of the deviation is identified in combination with the dimension and the influence weight, and the deviation duration, peak intensity and convergence probability in the change mode are calculated; The dimension, the influence weight, the change mode and the calculated deviation duration, peak intensity and convergence probability are integrated into the feature analysis result.

[0010] Optionally, further comprising: The long short-term memory network is used to perform trend prediction on the corrected deviation, and a deviation prediction value is output; In combination with the dynamic characteristic parameter, the deviation prediction value and the control accuracy threshold are calculated to obtain a deviation degree, when the deviation degree is greater than a preset deviation threshold, a pre-adjustment mechanism in the parameter adjustment strategy is triggered to incrementally correct the PID parameter in advance; The deviation prediction value, deviation value and deviation degree are added to the feature analysis result.

[0011] Optionally, according to the feature analysis result, a parameter adjustment strategy corresponding to the scene feature is called, including: Based on the scene type identifier, a basic parameter adjustment strategy matching the scene type identifier is called from a pre-constructed strategy library, the basic parameter adjustment strategy includes an iterative algorithm type, an initial parameter adjustment step and a basic boundary constraint rule adapted to the dynamic characteristic parameter; In combination with the change mode and the influence weight in the feature analysis result, the basic parameter adjustment strategy is corrected to obtain a parameter adjustment strategy.

[0012] Optionally, in combination with the change mode and the influence weight in the feature analysis result, the basic parameter adjustment strategy is corrected to obtain a parameter adjustment strategy, further comprising: A target function matrix is constructed with the joint optimization objectives of minimizing the deviation integral value, minimizing the parameter adjustment amplitude and minimizing the actuator energy consumption; A Pareto optimal solution set of the target function matrix is solved by a non-dominated sorting genetic algorithm; determine a priority of the joint optimization target according to the influence weight; select a correction parameter from the Pareto optimal solution set based on the priority; update a corresponding item in the basic parameter adjustment strategy by using the correction parameter to obtain a parameter adjustment strategy.

[0013] Optionally, the basic parameter adjustment strategy is corrected by combining the change mode and the influence weight in the feature analysis result to obtain a parameter strategy, including: If the change mode is persistent growth and the influence weight exceeds a preset threshold, a global search ratio of an iterative algorithm is increased to obtain the parameter adjustment strategy; If the change mode is fluctuant oscillation, the initial parameter adjustment step is reduced and the basic boundary constraint rule is tightened to obtain the parameter adjustment strategy.

[0014] Optionally, the parameter configuration of the PID regulation model is corrected based on the adjustment strategy, including: Taking the initial parameter configuration as a starting point, a particle swarm algorithm fused with chaotic disturbance is used for parameter optimization in a preset parameter search space to obtain a PID parameter candidate solution; The parameter search space is dynamically shrunk based on a deviation change rate in the feature analysis result; When the parameter variation amount of continuous preset iterations is less than a preset threshold, the parameter optimization is stopped and the PID parameter candidate solution is taken as a candidate correction parameter; The candidate correction parameter is verified for scene adaptability, and a matching degree of a theoretical deviation convergence speed under the candidate correction parameter and a scene dynamic characteristic parameter is calculated; When the matching degree meets a preset condition, the candidate correction parameter is determined as a PID correction parameter, and the parameter configuration of the PID regulation model is updated.

[0015] Optionally, based on the parameter configuration, a PID parameter is output to adjust the lithium battery production equipment, including: The parameter configuration of the PID regulation model is converted into a signal format recognizable by a device actuator to obtain a PID parameter; the PID parameter includes a proportional coefficient, an integral time, and a differential time; Based on the dynamic characteristic parameter, an output frequency of the PID parameter is set; Based on the PID parameter and the output frequency, an actuator is driven to perform a regulation action according to the PID parameter; Running parameters of the actuator are collected; the running parameters include controlled parameters and state parameters; Calculate the deviation value of the controlled parameter and the scene target value; if the actuator state parameter is abnormal and the deviation value is not within the preset range, suspend the PID parameter configuration output and switch to the backup PID parameter configuration corresponding to the scene.

[0016] In another aspect, the present application also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the PID self-tuning control method for the lithium battery industry according to any one of the above aspects when executing the program.

[0017] In another aspect, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the PID self-tuning control method for the lithium battery industry according to any one of the above aspects.

[0018] In another aspect, the present application also provides a computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements the PID self-tuning control method for the lithium battery industry according to any one of the above aspects.

[0019] The PID self-tuning control method for the lithium battery industry provided by the present application comprises the following steps: obtaining the running parameters of a lithium battery production device and a scene feature comprising a scene type identifier and a material attribute parameter, matching a PID regulation model pre-stored with adaptive dynamic characteristic parameters, calculating the deviation of the running parameters and the scene target value, dynamically correcting the deviation based on the material attribute parameter and the associated rules, analyzing the corrected deviation in combination with the influence weight of the dynamic characteristic parameters and the material attribute on the deviation trend, subsequently calling a parameter adjustment strategy corresponding to the scene to correct the PID regulation model according to the analysis result, and finally outputting the PID parameter regulation lithium battery production device, so as to solve the problem that the PID parameter tuning in the prior art is not deeply coupled with the scene characteristics and the material attributes of the lithium battery production, resulting in insufficient control precision and poor scene adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0021] Figure 1 is a flowchart of the PID self-tuning control method for the lithium battery industry provided by the embodiments of the present application; Figure 2 is a structural schematic diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] Figure 1 This is a schematic diagram of the PID self-tuning control method for the lithium battery industry provided in this embodiment of the invention.

[0024] like Figure 1 As shown in the figure, the PID self-tuning control method for the lithium battery industry provided by this invention mainly includes the following steps: 101. Obtain the operating parameters and scenario characteristics of lithium battery production equipment.

[0025] The scene features include scene type identifier, material attribute parameters, and scene target values ​​preset based on scene characteristics.

[0026] Specifically, the operating parameters of the lithium battery production equipment are collected in real time through the equipment's sensors. The operating parameters include controlled process parameters and interference parameters. For example, the coating thickness in the coating process, the rolling force in the rolling process, and the furnace temperature in the sintering process are controlled process parameters, while the ambient humidity, the load current of the equipment motor, and the fluctuation of the material conveying speed are interference parameters.

[0027] Meanwhile, scene feature data is retrieved from the production management system. The scene type identifier is a unique identifier that distinguishes different processes in lithium battery production, such as the character code for coating-positive electrode and rolling-electrode. The material attribute parameters are the key material characteristic data of the current production batch, such as the viscosity of the slurry, solid content, and the areal density and flexibility of the electrode, which can be obtained through offline testing instruments or online testing modules.

[0028] The target value of a scenario is the target range of the controlled parameters preset based on the process standard of that scenario, such as the target value of coating thickness of 50μm±2μm and the target value of sintering temperature of 800℃±5℃. The target value of a scenario is pre-configured by the process document and stored in the system.

[0029] 102. Match the corresponding PID control model based on the scenario type identifier.

[0030] The PID control model pre-stores dynamic characteristic parameters, control accuracy thresholds, initial parameter configurations, and material attribute deviation association rules that are adapted to the scenario type identifier.

[0031] The scene type identifier can distinguish different production processes and can be searched in the pre-stored PID adjustment model library. When a PID adjustment model matching the current scene type identifier is found, the preset dynamic characteristic parameters, control accuracy threshold, initial parameter configuration and material attribute deviation association rule in the PID model are loaded. Specifically, the dynamic characteristic parameters reflect the response characteristics of the controlled object to the control input under different processes, which is the basis for ensuring control stability and rapidity. At the same time, the control accuracy threshold defines the control accuracy level expected to be reached. The initial parameter configuration provides a reasonable starting point for the PID controller, which helps to speed up the convergence of the control process. The material attribute deviation association rule can dynamically adjust the PID parameters based on the internal relationship between the material attribute parameters and the controlled process parameters, so as to cope with the influence of material attribute fluctuations on the production process, thereby further improving the robustness and adaptability of the PID control process.

[0032] Specifically, the corresponding PID adjustment model is matched according to the scene type identifier, including: An initial PID adjustment model corresponding to the scene type identifier is called from the pre-constructed PID model library; Based on the material attribute parameters, the dynamic characteristic parameters and the initial parameter configuration of the initial PID adjustment model are adaptively adjusted through the preset material attribute deviation association rule; It is verified whether the control accuracy threshold of the adjusted initial PID adjustment model matches the scene target value; If the control accuracy threshold matches the scene target value, a PID adjustment model is generated; If the control accuracy threshold does not match the scene target value, the control accuracy threshold is calibrated based on the scene target value until the parameters of the initial PID adjustment model completely adapt to the scene characteristics, and a PID adjustment model is generated.

[0033] First, an initial PID adjustment model uniquely corresponding to the scene type identifier is called from the pre-constructed PID model library, and the PID model library pre-stores the basic control models of each process of lithium battery production, such as coating and rolling.

[0034] Then, based on the material attribute parameters of the current production batch, such as slurry viscosity and pole piece surface density, the dynamic characteristic parameters and the initial parameter configuration in the initial PID adjustment model are adaptively adjusted through the preset material attribute deviation association rule, such as adaptively adjusting the initial values of the inertia coefficient, the lag time, the proportional coefficient and the integral time, so that the initial PID adjustment model preliminarily fits the current material characteristics. The material attribute deviation association rule includes the influence coefficient of different material attributes on the model parameters.

[0035] Then, the control accuracy threshold of the adjusted initial PID adjustment model is verified, that is, whether the maximum deviation range allowed by the initial PID adjustment model matches the scene target value, and the verification is performed by judging whether the upper and lower limits of the control accuracy threshold completely cover the scene target value and do not exceed the preset redundancy range.

[0036] If the control accuracy threshold matches the scene target value, the adjusted initial PID adjustment model is directly determined as the final PID adjustment model. If the control accuracy threshold does not match the scene target value, the control accuracy threshold is calibrated based on the scene target value, and the above parameter adjustment and verification process is repeated until all parameters of the initial PID adjustment model are completely adapted to the scene characteristics, and a usable PID adjustment model is finally generated.

[0037] 103. Calculate the deviation of the running parameter from the scene target value, and dynamically correct the deviation based on the material attribute parameter and the material attribute deviation correlation rule.

[0038] The correction of the deviation can make the PID adjustment model more accurately reflect the influence of the change of the material attribute on the running parameter in the actual production process, thereby improving the control accuracy and stability. Specifically, by introducing the material attribute parameter and the pre-set material attribute deviation correlation rule, the deviation between the running parameter and the scene target value can be adjusted in real time, and the control process is ensured to be more in line with the actual production demand. The dynamic correction mechanism not only improves the self-adaptive ability, but also effectively reduces the control error caused by the fluctuation of the material attribute.

[0039] Specifically, the deviation of the running parameter from the scene target value is calculated, and the deviation is dynamically corrected based on the material attribute parameter and the material attribute deviation correlation rule, including: Calculate the initial deviation of the running parameter from the scene target value. Extract the change amount of the material attribute parameter, which is the difference between the material attribute parameter and the material attribute reference value. Call the material attribute deviation correlation rule, correct the initial deviation based on the weight factor and the change amount by a weighted sum formula, and obtain the deviation. If the ratio of the absolute value of the deviation to the control accuracy threshold is greater than the preset accuracy threshold, start the deviation truncation mechanism to constrain the deviation within the preset accuracy threshold of the control accuracy threshold.

[0040] Specifically, the controlled parameter is first extracted from the running parameter, for example, the controlled parameter is the actual thickness value of the coating process, and the initial deviation of the controlled parameter from the scene target value is calculated. Wherein, the initial deviation = controlled parameter - scene target value.

[0041] After obtaining the initial deviation, a variation of the material attribute parameter is extracted, and the variation is a difference between a measured value of the material attribute parameter and a reference value of the material attribute corresponding to the scene; subsequently, a pre-stored material attribute deviation correlation rule is called, the calling of the pre-stored material attribute deviation correlation rule including a weight factor of an influence of different material attributes on the deviation, for example, the weight factor of the viscosity on the thickness deviation is 0.6, based on the weight factor and the variation, the initial deviation is corrected by a weighted summation formula to obtain a corrected deviation; wherein the weighted summation formula is corrected deviation = initial deviation × [1 + (material attribute variation × weight factor)].

[0042] Finally, it is judged whether a ratio of an absolute value of the corrected deviation to the control accuracy threshold value is greater than a preset accuracy threshold value, if the ratio is greater than the preset accuracy threshold value, a deviation truncation mechanism is started to constrain the corrected deviation in the control accuracy threshold value range, so as to avoid device adjustment oscillation caused by excessive correction.

[0043] For example, the scene target value is a coating thickness of 50 μm, the control accuracy threshold value is ± 2 μm, the preset accuracy threshold value is 150%, that is, ± 3 μm; the material attribute parameter is slurry viscosity, the reference value of the material attribute is 2000 mPa·s, and the weight factor of the viscosity on the thickness deviation in the material attribute deviation correlation rule is 0.0005.

[0044] If the measured value of the coating thickness in the real-time collected running parameter is 52 μm, the initial deviation = 52 μm-50 μm = 2 μm; the current measured value of the slurry viscosity is 2200 mPa·s, and the variation = 2200 mPa·s-2000 mPa·s = 200 mPa·s; after calling the correlation rule, the corrected deviation = 2 μm × [1 + (200 × 0.0005)] = 2 μm × 1.1 = 2.2 μm. Since the ratio of 2.2 μm to the control accuracy threshold value is 110%, which is less than 150%, the truncation mechanism does not need to be started, and the corrected deviation is 2.2 μm.

[0045] If the measured value of the coating thickness is 54 μm, the initial deviation = 4 μm; when the viscosity variation is 200 mPa·s, the corrected deviation = 4 μm × 1.1 = 4.4 μm. At this time, the ratio of 4.4 μm to the control accuracy threshold value is 220%, which is greater than 150%, the truncation mechanism is started to constrain the deviation to 3 μm.

[0046] 104、Based on the dynamic characteristic parameter, the influence weight of the material attribute parameter on the deviation variation trend is fused to perform feature analysis on the corrected deviation, and a feature analysis result is obtained.

[0047] Wherein, the feature analysis on the corrected deviation to obtain the feature analysis result can comprehensively consider the dynamic characteristic parameter and the material attribute parameter to deeply analyze the corrected deviation.

[0048] Specifically, based on the dynamic characteristic parameter, the influence weight of the material attribute parameter on the deviation change trend is fused, the corrected deviation is analyzed, and a feature analysis result is obtained, including: The dimension of the feature analysis is determined according to the dynamic characteristic parameter, and the dimension is adapted to the response characteristic of the scene; Based on the historical influence data of the material attribute parameter, the influence weight of the material attribute parameter on the deviation change trend is calculated by a gradient boosting tree algorithm; The deviation is analyzed in time sequence, the change mode of the deviation is identified in combination with the dimension and the influence weight, and the deviation duration, peak intensity and convergence probability under the change mode are calculated; The dimension, influence weight, change mode and calculated deviation duration, peak intensity and convergence probability are integrated as the feature analysis result.

[0049] Firstly, the dimension of the feature analysis is determined according to the dynamic characteristic parameter, and the dimension needs to be adapted to the response characteristic of the scene, that is, the dimension setting needs to be adapted to the inertia and lag dynamic attribute of the scene; secondly, based on the historical influence data of the material attribute parameter, the influence weight of each material attribute parameter on the deviation change trend is calculated by a gradient boosting tree algorithm, and the weight size reflects the influence degree of different material attributes on the deviation trend in the current scene.

[0050] Then, the corrected deviation is analyzed in time sequence, the change mode of the deviation is identified in combination with the dimension and the influence weight, such as the change mode is persistent increase or decrease, or the change mode is periodic fluctuation, and the deviation duration, peak intensity and convergence probability under the change mode are calculated; finally, the dimension, influence weight, change mode and calculated deviation duration, peak intensity and convergence probability are integrated as the structured feature analysis result.

[0051] Taking the lithium battery rolling process as an example, the dynamic characteristic parameter shows that the lithium battery rolling process has large inertia, and the feature analysis dimension is determined as the deviation accumulation.

[0052] Based on the historical data, the influence weight of the pole piece hardness is 0.5 and the influence weight of the rolling speed is 0.3, which are calculated by a gradient boosting tree algorithm, the corrected rolling thickness deviation is analyzed in time sequence for 15 consecutive sampling periods, the deviation is identified as a ladder type growth mode in combination with the dimension and the influence weight, it is calculated that the ladder type growth mode has lasted for 8 periods, the peak intensity is 4 μm, and the convergence probability is 30%; finally, the integrated feature analysis result is: the dimension is the deviation accumulation, the influence weight is the pole piece hardness 0.5 and the rolling speed 0.3, the change mode is the ladder type growth, the duration is 8 periods, the peak intensity is 4 μm, and the convergence probability is 30%.

[0053] In some embodiments, after obtaining the feature analysis result, the method further includes: The long short-term memory network is used to predict the trend of the corrected deviation, and an output deviation prediction value is obtained. In combination with the dynamic characteristic parameter, a deviation degree of deviation between the deviation prediction value and the control accuracy threshold is calculated, and when the deviation degree is greater than a preset deviation threshold, a pre-adjustment mechanism in the parameter adjustment strategy is triggered to incrementally correct the PID parameter in advance. The deviation prediction value, the deviation value and the deviation degree are added to the feature analysis result.

[0054] The long short-term memory network is used to predict the trend of the corrected deviation, and an output deviation prediction value is obtained.

[0055] Then, in combination with the dynamic characteristic parameter, a deviation degree of deviation between the deviation prediction value and the control accuracy threshold is calculated, and when the deviation degree is greater than a preset deviation threshold, a pre-adjustment mechanism in the parameter adjustment strategy is triggered to incrementally correct the PID parameter in advance. For example, in a large inertia scenario, the cumulative effect of the deviation prediction value over time needs to be considered. The calculation formula of the deviation degree is: ; ; If the calculated deviation degree exceeds the preset deviation threshold, the pre-adjustment mechanism is triggered, the increment correction range of the PID parameter is determined according to the size of the deviation degree and the dynamic characteristic parameter, and the parameter is adjusted in advance to avoid the risk of deviation exceeding the limit.

[0056] The obtained future deviation prediction value, the current corrected deviation value and the calculated deviation degree are integrated into the existing feature analysis result, so that the feature analysis result not only includes the features of the historical deviation, but also includes the prediction information of the future deviation and the deviation risk index, thereby providing a more comprehensive decision basis for the subsequent calling of the parameter adjustment strategy and improving the foresight and pertinence of the strategy.

[0057] For example, assuming that the control accuracy threshold of the coating process is ±2μm, the dynamic characteristic parameter is small inertia, and the preset deviation threshold is 30%. The obtained feature analysis result includes: the current corrected deviation is 1.5μm (period 20), the change mode is slow growth, and the duration is 5 periods.

[0058] The long short-term memory network is used to predict the deviation, and the corrected deviation time series data of the 1st-20th periods, such as 0.8 μm, 1.0 μm, 1.2 μm, …, 1.5 μm, are input into the trained long short-term memory network model. The long short-term memory network model learns that the deviation increases at a rate of 0.1 μm / period, and outputs the deviation prediction value of the future 10 periods, wherein the prediction value of the 25th period is 2.0 μm, and the prediction value of the 30th period is 2.5 μm.

[0059] In combination with the small inertia characteristic, the long-term cumulative effect does not need to be considered, the deviation degree of the prediction value of the 30th period from the upper limit of the control precision threshold is (2.5-2) / 2*100% = 25%, which is less than 30%, and the pre-adjustment is not triggered; the deviation degree of the prediction value of the 35th period is (3.0-2) / 2*100% = 50%, which is greater than 30%, and the pre-adjustment mechanism is triggered. According to the deviation degree 50% and the small inertia characteristic, the correction range of the proportional coefficient increment is determined to be 15% of the base value, and the parameter is adjusted in advance to suppress the growth of the deviation.

[0060] Finally, the deviation prediction value, the current corrected deviation, and the deviation degree are added to the original feature analysis result to form complete analysis data containing historical features and future predictions, which provides a forward-looking basis for subsequent parameter adjustment strategies.

[0061] 105、According to the feature analysis result, the parameter adjustment strategy corresponding to the scene feature is called, and the parameter configuration of the PID regulation model is corrected based on the parameter adjustment strategy. The parameter adjustment strategy includes an iterative algorithm adapted to the material attribute parameter, a parameter adjustment step, and a boundary constraint rule.

[0062] Specifically, according to the feature analysis result, the parameter adjustment strategy corresponding to the scene feature is called, including: Based on the scene type identifier, the base parameter adjustment strategy matching the scene type identifier is called from the pre-constructed strategy library.

[0063] The base parameter adjustment strategy includes an iterative algorithm type adapted to the dynamic characteristic parameter, an initial parameter adjustment step, and a base boundary constraint rule. The iterative algorithm includes gradient descent method, genetic algorithm, and particle swarm optimization algorithm. The iterative algorithm can automatically select the optimal iterative path for parameter adjustment according to the different characteristics of the material attribute parameter. The parameter adjustment step is dynamically determined according to the historical adjustment effect and the deviation prediction value, to ensure that the adjustment process is neither too aggressive nor too conservative. The boundary constraint rule is used to limit the range of parameter adjustment to prevent system instability or exceed the physical limit of the equipment due to excessive adjustment. By comprehensively using the parameter adjustment strategy, the PID regulation model parameters can be accurately corrected, thereby improving the stability and response speed of the PID regulation model.

[0064] The basic parameter adjustment strategy is corrected according to the change mode and the influence weight in the characteristic analysis result, and a parameter adjustment strategy is obtained.

[0065] Specifically, the basic parameter adjustment strategy is corrected according to the change mode and the influence weight in the characteristic analysis result, and a parameter strategy is obtained, including: If the change mode is persistent growth and the influence weight exceeds a preset threshold, the global search proportion of the iterative algorithm is increased to obtain the parameter adjustment strategy. If the change mode is fluctuant oscillation, the initial parameter adjustment step is reduced and the basic boundary constraint rule is tightened to obtain the parameter adjustment strategy.

[0066] If the change mode is persistent growth, that is, the corrected deviation shows a stable increasing trend over time, and no downward inflection point appears in continuous multiple sampling periods, and the influence weight exceeds the preset threshold, it indicates that the fluctuation of the material attribute is the main driving factor of the deviation growth. At this time, the global search proportion of the iterative algorithm needs to be increased to quickly find the optimal parameter combination that can suppress the growth of the deviation by expanding the parameter search range, and then the parameter adjustment strategy suitable for the persistent growth scenario is obtained.

[0067] If the change mode is fluctuant oscillation, that is, the deviation repeatedly fluctuates around the control accuracy threshold, the fluctuation frequency exceeds the preset period and the amplitude does not significantly decay, indicating that the current parameter adjustment rhythm does not match the scene response characteristics, and is easy to cause overshoot. At this time, the initial parameter adjustment step needs to be reduced, such as reducing from 0.2 to 0.1, to reduce the amplitude of single parameter change, and the basic boundary constraint rule is tightened to narrow the parameter fluctuation range from ±5% to ±3%, to avoid parameter adjustment from causing the oscillation to intensify. Finally, the parameter adjustment strategy suitable for the fluctuant oscillation scenario is obtained.

[0068] In addition, the basic parameter adjustment strategy is corrected according to the change mode and the influence weight in the characteristic analysis result, and a parameter adjustment strategy is obtained, further including: A target function matrix is constructed with the minimization of the deviation integral value, the minimization of the parameter adjustment amplitude and the minimization of the actuator energy consumption as the joint optimization objectives.

[0069] The deviation integral value is the integral of the absolute value of the corrected deviation within a preset time, and the calculation formula is: ; Wherein, is the deviation integral value, is the integral time, is the absolute value of the difference between the set value and the actual output value, and the deviation integral value reflects the cumulative influence of the deviation; The parameter adjustment amplitude is the sum of the absolute values of the PID parameter changes, and the calculation formula is: ; wherein, is a parameter adjustment amplitude, is a parameter value, is a number of iterations, and the parameter adjustment amplitude reflects stability of parameter adjustment.

[0070] The actuator energy consumption is an integral of the real-time power of the actuator, and the calculation formula is: ; wherein, is the actuator energy consumption, reflecting the energy consumption cost of the regulation process; is the real-time power. The three objective functions are integrated in matrix form to form a mathematical expression of joint optimization.

[0071] The Pareto optimal solution set of the objective function matrix is solved by the non-dominated sorting genetic algorithm; wherein, the non-dominated sorting genetic algorithm (NSGA-II) distinguishes the superiority and inferiority grade and the crowding degree calculation of the solution by non-dominated sorting of the population, and iteratively searches the solution of simultaneously optimizing the three objective functions. The final Pareto optimal solution set is a set of non-dominated solutions, that is, there is no solution in the set that is superior to another solution in all objectives, each solution is optimal in different objectives, forming a balanced scheme between the objectives.

[0072] The priority of the joint optimization target is determined according to the influence weight.

[0073] wherein, the priority of the joint optimization target is determined according to the influence weight. The influence weight reflects the influence degree of the material attribute parameter on the deviation change trend, such as the influence weight of a certain material attribute is 0.7, indicating that its influence on the deviation is dominant. If the influence weight is high, it means that the fluctuation of the material attribute is the main cause of the deviation, at this time, the deviation integral value is preferentially minimized to quickly suppress the deviation; if the influence weight is low, the parameter adjustment amplitude minimization or the actuator energy consumption minimization can be preferentially considered to balance the regulation stability and cost.

[0074] The correction parameter is selected from the Pareto optimal solution set based on the priority.

[0075] wherein, the solution with the smallest is filtered out from the Pareto optimal solution set; if there are multiple same solutions, further filtering is performed according to the secondary priority, and finally the unique solution is determined as the correction parameter. The correction parameter specifically includes the global search ratio of the iteration algorithm, the parameter adjustment step coefficient, and the boundary constraint threshold, which are used to adjust the basic parameter adjustment strategy.

[0076] The corresponding item in the basic parameter adjustment strategy is updated by using the correction parameter to obtain the parameter adjustment strategy.

[0077] Wherein, the selected correction parameter replaces the corresponding initial value in the basic strategy, such as updating the parameter adjustment step of the basic strategy from 0.2 to the correction parameter 0.15, and updating the global search ratio from 30% to 45%, so that the adjustment strategy adapts to the change mode and influence weight of the current scene, forming a final executable parameter adjustment strategy.

[0078] For example, taking the lithium roller pressing process as an example, the scene feature is volatility shock, the influence weight of the material attribute is 0.7, and the basic parameter adjustment strategy includes an initial step of 0.2, a global search ratio of 30%, and a boundary constraint threshold of ±5.

[0079] The NSGA-II is used to obtain a Pareto optimal solution set, including three non-dominated solutions, solution A (J1=8, J2=1.2, J3=200), solution B (J1=10, J2=0.8, J3=150), and solution C (J1=6, J2=1.5, J3=250).

[0080] Since the influence weight of the pole piece hardness is 0.7>0.6, the priority is determined as J1(minimize deviation integral value)>J2(parameter adjustment amplitude)>J3(energy consumption).

[0081] Since J1 of solution C is 6, the minimum, the correction parameter corresponding to solution C is selected, wherein the correction parameter of solution C is a step length coefficient of 0.15, a global search ratio of 45%, and a boundary constraint threshold of ±4.

[0082] The initial step of 0.2 is changed to 0.15, the global search ratio of 30% is changed to 45%, and the boundary constraint threshold of ±5 is changed to ±4, to obtain a parameter adjustment strategy that adapts to the volatility shock mode and the high weight material attribute.

[0083] In addition, based on the parameter configuration of the adjustment strategy to correct the PID regulation model, including: Taking the initial parameter configuration as a starting point, a particle swarm algorithm with chaos disturbance fusion is used to optimize the parameters in a preset parameter search space to obtain a PID parameter candidate solution.

[0084] The parameter search space is dynamically contracted based on the deviation change rate in the feature analysis result.

[0085] When the parameter variation of continuous preset iterations is less than a preset threshold, the parameter optimization is stopped and the PID parameter candidate solution is taken as a candidate correction parameter.

[0086] The scene adaptability of the candidate correction parameter is verified, and the matching degree of the theoretical deviation convergence speed under the action of the candidate correction parameter and the scene dynamic characteristic parameter is calculated.

[0087] When the matching degree meets the preset condition, the candidate correction parameter is determined as the PID correction parameter, and the parameter configuration of the PID adjustment model is updated.

[0088] The PID parameter candidate solution is obtained by using a particle swarm algorithm with chaos disturbance for parameter optimization. The initial parameter configuration is the benchmark value of the initial proportional coefficient (Kp), integral time (Ti), and differential time (Td) of the PID adjustment model, such as Kp=10, Ti=5s, and Td=1s.

[0089] The preset parameter search space is the allowed value range of each parameter, such as Kp∈[5,20], Ti∈[2s,10s], and Td∈[0.5s,3s].

[0090] In the particle swarm algorithm with chaos disturbance, the particle swarm algorithm searches for parameters by updating the particle position; the chaos disturbance generates a random sequence through Logistic mapping, such as x(n+1)=4x(n)(1-x(n)), and disturbs the particle position that falls into a local optimum in the iteration process to avoid optimization stagnation, and finally obtains a set of temporarily selected PID parameters (Kp, Ti, and Td) as the candidate solution.

[0091] The deviation change rate is the change amount of the corrected deviation per unit time, such as A positive value indicates that the deviation increases, and a negative value indicates that the deviation decreases. If the absolute value of the deviation change rate is large, the parameter needs to be adjusted significantly, at which time the search space is contracted by 20%, such as the Kp range from [5,20] to [8,17], focusing on a better area; if the absolute value of the deviation change rate is small, the parameter is close to the optimum, and the search space is contracted by 50%, such as the Kp range to [9,11], improving the optimization accuracy.

[0092] When the parameter change amount of continuous preset iterations is less than the preset threshold, the parameter optimization is stopped, and the PID parameter candidate solution is used as the candidate correction parameter. The preset number of iterations is set to 3-5 times, and the preset threshold is 1% of the initial parameter configuration, such as the Kp initial value 10 and the threshold 0.1. If the change amount of Kp, Ti, and Td in continuous 3 iterations is less than the corresponding threshold, it indicates that the parameter has tended to be stable, the optimization is stopped, and the current PID parameter candidate solution is the candidate correction parameter.

[0093] When the candidate correction parameter is verified for scene adaptability, the matching degree between the theoretical deviation convergence speed under the action of the candidate correction parameter and the scene dynamic characteristic parameter is calculated. For example, the control effect of the candidate correction parameter in the current scene is simulated to obtain the theoretical deviation convergence speed, such as the time for the deviation to decrease from 3μm to 0μm.

[0094] When the matching degree meets the preset condition, the candidate correction parameter is determined as the PID correction parameter, and the parameter configuration of the PID adjustment model is updated. For example, the preset condition is set as matching degree ≥ 80%, if the calculated matching degree is 85%, the candidate correction parameter is valid, the candidate correction parameter replaces the parameter configuration of the original PID adjustment model, and the PID adjustment model parameter correction is completed.

[0095] 106、Based on the parameter configuration, output the PID parameter to adjust the lithium battery production equipment.

[0096] Among them, the parameter configuration is the PID correction parameter obtained through the optimization process, including the specific values of the proportional coefficient Kp, the integral time Ti and the derivative time Td.

[0097] Specifically, based on the parameter configuration, the PID parameter is output to adjust the lithium battery production equipment, including: The parameter configuration of the PID adjustment model is converted into a signal format that can be recognized by the device actuator, and the PID parameter is obtained.

[0098] Among them, the PID parameter contains the proportional coefficient, the integral time and the derivative time. The parameter configuration of the PID adjustment model needs to be converted into a signal format that can be recognized by the device actuator according to the signal protocol of the actuator type. For example, for an analog actuator, it is converted into a 4-20mA current signal, and for a digital actuator, it is converted into a digital instruction of Modbus-RTU protocol, so that the actuator can directly parse the parameter meaning.

[0099] Based on the dynamic characteristic parameter, the output frequency of the PID parameter is set.

[0100] Among them, the dynamic characteristic parameter reflects the response speed of the scene, such as slow response in large inertia scene and fast response in small inertia scene, and the output frequency needs to be adapted to the scene. For example, for the rolling process, the rolling process has large inertia, the lag time in the dynamic characteristic parameter is > 5s, and the output frequency is set to 1Hz to avoid frequent adjustment and cause device vibration. For the coating process, the coating process has small inertia, the lag time is < 1s, and the output frequency is set to 10Hz to ensure fast response to deviation changes.

[0101] Based on the PID parameter and the output frequency, the actuator executes the adjustment action according to the PID parameter.

[0102] Among them, the actuator adjusts the physical quantity in real time according to the received PID parameter and output frequency. For example, the servo motor of the coating machine adjusts the speed according to the Kp value, accumulates the correction speed deviation according to the Ti value, and predicts the speed trend according to the Td value, and updates the adjustment amount every time, so that the controlled parameter converges to the scene target value.

[0103] The running parameters of the actuator are collected.

[0104] Wherein, the operating parameters include controlled parameters and state parameters; the controlled parameters are collected in real time by device sensors, that is, process parameters adjusted by actuators, such as coating thickness and roll pressure; the state parameters are collected by feedback modules of actuators, such as motor operating current, valve opening, and response delay time; the collection frequency is consistent with the output frequency of PID parameters, so as to ensure data synchronization.

[0105] The deviation value of the controlled parameter and the scene target value is calculated; if the state parameter of the actuator is abnormal and the deviation value is not within the preset range, the output of the PID parameter configuration is suspended and switched to the standby PID parameter configuration corresponding to the scene.

[0106] Wherein, the deviation value is the difference between the measured value of the controlled parameter and the scene target value; the state parameter is abnormal, for example, the current exceeds the rated range, and the response delay is greater than 200 ms; when both conditions are met, the parameter output is immediately suspended, the pre-stored standby parameter is called, and the standby parameter is the historical stable running PID parameter, so as to avoid device out of control or product defect expansion, and record abnormal information for subsequent optimization.

[0107] Figure 2 is a structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0108] As shown in Figure 2 , the electronic device can include a processor 210, a communications interface 220, a memory 230, and a communications bus 240, wherein the processor 210, the communications interface 220, and the memory 230 complete mutual communication through the communications bus 240. The processor 210 can call logical instructions in the memory 230 to execute the PID self-tuning control method applied to the lithium battery industry.

[0109] In addition, the logical instructions in the memory 230 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0110] In another aspect, the present application also provides a computer program product comprising a computer program, which can be stored on a non-transitory computer readable storage medium, and the computer program, when executed by a processor, enables a computer to perform the PID self-tuning control method for the lithium battery industry provided by the above methods.

[0111] In yet another aspect, the present application also provides a non-transitory computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the PID self-tuning control method for the lithium battery industry provided by the above methods.

[0112] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0113] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary general hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0114] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A PID self-tuning control method applied to the lithium battery industry, characterized in that, The method comprises the following steps: acquiring the running parameters of a lithium battery production device and scene characteristics, wherein the scene characteristics comprise a scene type identifier, material attribute parameters, and a scene target value preset based on the scene characteristics; matching a corresponding PID adjustment model according to the scene type identifier, wherein the PID adjustment model pre-stores dynamic characteristic parameters, a control accuracy threshold, initial parameter configurations, and material attribute deviation correlation rules that are adapted to the scene type identifier; calculating the deviation of the running parameters from the scene target value, and dynamically correcting the deviation based on the material attribute parameters and the material attribute deviation correlation rules; based on the dynamic characteristic parameters, fusing the influence weight of the material attribute parameters on the deviation trend, and performing feature analysis on the corrected deviation to obtain a feature analysis result; based on the feature analysis result, calling a parameter adjustment strategy corresponding to the scene characteristics, and correcting the parameter configurations of the PID adjustment model based on the parameter adjustment strategy; the parameter adjustment strategy comprises an iteration algorithm, a parameter adjustment step, and boundary constraint rules that are adapted to the material attribute parameters; based on the parameter configurations, outputting PID parameters to adjust the lithium battery production device.

2. The PID self-tuning control method for lithium battery industry according to claim 1, characterized in that, The method comprises the following steps: calling an initial PID adjustment model corresponding to the scene type identifier from a pre-constructed PID model library; based on the material attribute parameters, adaptively adjusting the dynamic characteristic parameters and initial parameter configurations of the initial PID adjustment model through a pre-set material attribute deviation correlation rule; verifying whether the control accuracy threshold of the adjusted initial PID adjustment model matches the scene target value; if the control accuracy threshold matches the scene target value, generating a PID adjustment model; if the control accuracy threshold does not match the scene target value, calibrating the control accuracy threshold based on the scene target value until the parameters of the initial PID adjustment model are completely adapted to the scene characteristics, and generating the PID adjustment model.

3. The PID self-tuning control method for lithium battery industry of claim 1, wherein, The method comprises the following steps: calculating the initial deviation of the running parameters from the scene target value; extracting the change amount of the material attribute parameters, wherein the change amount is the difference between the material attribute parameters and the material attribute reference value; calling the material attribute deviation correlation rule, correcting the initial deviation based on a weight factor and the change amount through a weighted summation formula to obtain a deviation; if the ratio of the absolute value of the deviation to the control accuracy threshold is greater than a pre-set accuracy threshold, starting a deviation truncation mechanism to constrain the deviation within the pre-set accuracy threshold of the control accuracy threshold.

4. The PID self-tuning control method for lithium battery industry of claim 1, wherein, The method comprises the following steps: Determine the dimension of the feature analysis according to the dynamic characteristic parameter, the dimension is adapted to the response characteristic of the scene; Based on the historical influence data of the material attribute parameter, the influence weight of the material attribute parameter on the deviation change trend is calculated by the gradient boosting tree algorithm; Perform time series analysis on the deviation, combine the dimension and the influence weight, identify the change mode of the deviation, and calculate the deviation duration, peak intensity and convergence probability under the change mode; Integrate the dimension, influence weight, change mode and calculated deviation duration, peak intensity and convergence probability into a feature analysis result.

5. The PID self-tuning control method for lithium battery industry according to claim 4, characterized in that, Also includes: Adopting long short-term memory network to perform trend prediction on the corrected deviation, and outputting a deviation prediction value; Combined with the dynamic characteristic parameter, calculate the deviation of the deviation prediction value and the control accuracy threshold, when the deviation is greater than the preset deviation threshold, trigger the pre-adjustment mechanism in the parameter adjustment strategy, and incrementally correct the PID parameter in advance; Add the deviation prediction value, deviation value and deviation to the feature analysis result.

6. The PID self-tuning control method for lithium battery industry according to claim 1, characterized in that, According to the feature analysis result, call the parameter adjustment strategy corresponding to the scene feature, including: Based on the scene type identifier, call the basic parameter adjustment strategy matching the scene type identifier from the pre-constructed strategy library, the basic parameter adjustment strategy includes the iteration algorithm type, the initial parameter adjustment step and the basic boundary constraint rule adapted to the dynamic characteristic parameter; Combined with the change mode and the influence weight in the feature analysis result, the basic parameter adjustment strategy is modified to obtain a parameter adjustment strategy.

7. The PID self-tuning control method for lithium battery industry of claim 6, wherein, Combined with the change mode and the influence weight in the feature analysis result, the basic parameter adjustment strategy is modified to obtain a parameter strategy, including: If the change mode is persistent growth and the influence weight exceeds a preset threshold, increase the global search proportion of the iteration algorithm to obtain the parameter adjustment strategy; If the change mode is fluctuant oscillation, reduce the initial parameter adjustment step and tighten the basic boundary constraint rule to obtain the parameter adjustment strategy.

8. The PID self-tuning control method for lithium battery industry of claim 6, wherein, Combined with the change mode and the influence weight in the feature analysis result, the basic parameter adjustment strategy is modified to obtain a parameter adjustment strategy, further including: Minimizing the deviation integral value, minimizing the parameter adjustment amplitude and minimizing the actuator energy consumption as the joint optimization objective to construct a target function matrix; Solve the Pareto optimal solution set of the target function matrix by non-dominated sorting genetic algorithm; Determine the priority of the joint optimization objective according to the influence weight; Based on the priority, select a modified parameter from the Pareto optimal solution set; Update the corresponding item in the basic parameter adjustment strategy by using the modified parameter to obtain a parameter adjustment strategy.

9. The PID self-tuning control method for lithium battery industry according to claim 1, characterized in that, The parameter configuration of the PID regulation model is corrected based on the parameter adjustment strategy, including: Taking the initial parameter configuration as the starting point, performing parameter optimization in the preset parameter search space by using the particle swarm algorithm fused with chaotic disturbance to obtain a PID parameter candidate solution; dynamically shrink the parameter search space based on a bias change rate in the characteristic analysis result; when the parameter change amount of a preset number of consecutive iterations is less than a preset threshold, stopping the parameter optimization and taking the PID parameter candidate solution as a candidate correction parameter; performing scene adaptability verification on the candidate correction parameter, and calculating a matching degree between a theoretical bias convergence speed under the candidate correction parameter and the scene dynamic characteristic parameter; when the matching degree meets a preset condition, determining the candidate correction parameter as a PID correction parameter, and updating the parameter configuration of the PID adjustment model.

10. The PID self-tuning control method for lithium battery industry of claim 1, wherein, the PID parameter based on the parameter configuration, including: converting the parameter configuration of the PID adjustment model into a signal format recognizable by a device actuator to obtain a PID parameter; the PID parameter includes a proportional coefficient, an integral time, and a differential time; based on the dynamic characteristic parameter, setting an output frequency of the PID parameter; based on the PID parameter and the output frequency, driving the actuator to perform an adjustment action according to the PID parameter; collecting running parameters of the actuator; the running parameters include a controlled parameter and a state parameter; calculating a bias value of the controlled parameter and the scene target value; if the actuator state parameter is abnormal and the bias value is not within a preset range, suspending the PID parameter configuration output and switching to a standby PID parameter configuration corresponding to the scene.

Images