Proportional integral derivative parameter adjustment method, device and equipment of industrial control loop and medium

Abstract

This application discloses a method, apparatus, equipment, and medium for adjusting the proportional-integral-derivative (PID) parameters of an industrial control loop, relating to the field of industrial process control technology. The method includes: acquiring closed-loop response data of the industrial control loop, the current PID parameter vector, the artificial control target, the controlled object model information, and the parameter range; determining the qualitative adjustment direction of the target PID parameter based on the response index in the closed-loop response data and the artificial control target; determining the relative adjustment ratio of the target PID parameter based on the qualitative adjustment direction, the current PID parameter vector, the controlled object model information, and the artificial control target; generating candidate parameters based on the current PID parameter vector and the relative adjustment ratio, and generating parameter adjustment instructions based on the candidate parameters to adjust the target PID parameter. This method can obtain accurate PID parameters and reduce the risk of control loop instability.

Description

Technical Field

[0001] This invention relates to the field of industrial process control technology, and in particular to methods, devices, equipment and media for adjusting proportional, integral and derivative parameters of industrial control loops. Background Technology

[0002] Currently, with the continuous improvement of industrial automation and the rapid development of artificial intelligence technology, Proportional-Integral-Derivative (PID) controllers are widely used in industrial process control scenarios such as chemical, energy, pharmaceutical, metallurgy, and advanced manufacturing due to their simple structure, ease of implementation, and robustness. Their parameters directly affect the response speed, overshoot, oscillation level, and steady-state error of the control loop. For example, when tuning PID parameters for industrial control loops, methods such as manual experience-based tuning, classical empirical formulas, or optimization-based automatic tuning are commonly used. Manual tuning relies on experience, is time-consuming, and difficult to reuse. Classical empirical formulas are highly dependent on the object model and operating conditions. Optimization-based tuning methods require process models or extensive online testing, resulting in high deployment costs in industrial settings with strict safety constraints. While using large models to assist PID tuning can introduce natural language target understanding and rule reasoning capabilities, directly outputting PID parameters end-to-end from a single large model can easily lead to problems such as unclear diagnostic criteria, unstable output formats, parameter out-of-bounds errors, and difficulty in verification.

[0003] In summary, the problem that needs to be solved is how to separate control state diagnosis from specific parameter generation and combine it with retrieval-based engineering rules for constrained reasoning in PID parameter tuning. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for adjusting the proportional-integral-derivative (PID) parameters of an industrial control loop, which can obtain accurate PID parameters and reduce the risk of control loop instability. The specific solution is as follows: In a first aspect, this application discloses a method for adjusting the proportional-integral-derivative parameters of an industrial control loop, including: Acquire closed-loop response data of industrial control loop, current proportional-integral-derivative parameter vector, artificial control target, controlled object model information, and parameter range; Response indicators are extracted from the closed-loop response data; these response indicators are used to characterize the dynamic response state of the industrial control loop. The preset state diagnosis knowledge base is retrieved based on the response index and the artificial control target to obtain the target diagnosis rule, and the qualitative adjustment direction of the target proportional-integral-derivative parameters is determined based on the target diagnosis rule. Based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target, the target parameter adjustment rule is determined from the preset parameter adjustment knowledge base, and the relative adjustment ratio of the target proportional-integral-derivative parameter is determined based on the target parameter adjustment rule; Candidate parameters are generated based on the current proportional-integral-derivative (PID) parameter vector and the relative adjustment ratio. When the candidate parameters are within the parameter range, a parameter adjustment instruction for the target PID parameter is output so that the target PID parameter can be adjusted based on the parameter adjustment instruction.

[0005] Optionally, extracting response metrics from the closed-loop response data includes: The overshoot, settling time, integral error, oscillation level, and time delay characteristics are extracted from the closed-loop response data, and the overshoot, settling time, integral error, oscillation level, and time delay characteristics are organized into a dynamic performance index vector of the closed-loop circuit to obtain the response index. Wherein, the overshoot represents the closed-loop peak deviation; the settling time represents the response speed; the integral error represents the steady-state cumulative deviation; the oscillation level represents the severity of the response oscillation; and the time delay characteristic represents the pure time delay characteristic of the controlled object.

[0006] Optionally, the step of retrieving a preset state diagnosis knowledge base based on the response index and the artificial control target to obtain target diagnosis rules, and determining the qualitative adjustment direction of the target proportional-integral-derivative parameters based on the target diagnosis rules, includes: The response indicators and the manual control targets are determined as query conditions, and the preset state diagnosis knowledge base is retrieved through the query conditions to obtain candidate diagnosis rules that match the current working condition. The target diagnostic rule is determined from the candidate diagnostic rules, and the qualitative adjustment direction of the proportional coefficient, integral coefficient and differential coefficient in the target proportional-integral-differential parameters is determined based on the target diagnostic rule.

[0007] Optionally, determining the target parameter adjustment rules from a preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-differential parameter vector, the controlled object model information, and the artificial control target includes: Based on the abnormal scenarios corresponding to the qualitative adjustment direction, and the time delay and gain features in the controlled object model information, a composite query condition is constructed. The preset parameter adjustment knowledge base is retrieved based on the composite query conditions to match the target parameter adjustment rule corresponding to the current working condition from the preset parameter adjustment knowledge base.

[0008] Optionally, determining the relative adjustment ratio of the target proportional-integral-derivative parameter based on the target parameter adjustment rule includes: Based on the adjustment range of the proportional coefficient bound in the target parameter adjustment rule, determine the first adjustment ratio of the proportional coefficient in the target proportional integral differential parameter; Based on the adjustment range of the integral coefficient bound in the target parameter adjustment rule, determine the second adjustment ratio of the integral coefficient in the target proportional integral differential parameter; Based on the adjustment range of the differential coefficients bound in the target parameter adjustment rules, determine the third adjustment ratio of the differential coefficients in the target proportional integral differential parameter; Wherein, the sign of the first adjustment ratio is consistent with the qualitative adjustment direction of the proportional coefficient; the sign of the second adjustment ratio is consistent with the qualitative adjustment direction of the integral coefficient; and the sign of the third adjustment ratio is consistent with the qualitative adjustment direction of the differential coefficient.

[0009] Optionally, generating candidate parameters based on the current proportional-integral-differential parameter vector and the relative adjustment ratio includes: The first adjustment ratio, the second adjustment ratio, and the third adjustment ratio are combined to form a three-dimensional relative adjustment ratio vector; Candidate parameters are determined by multiplying the three-dimensional relative adjustment scale vector and the current scale integral differential parameter vector element by element.

[0010] Optionally, the step of outputting a parameter adjustment instruction for the target proportional-integral-derivative parameter when the candidate parameter is within the parameter range includes: Determine whether each of the candidate parameters falls between the lower limit and the upper limit of the corresponding parameter value in the parameter range; If the candidate parameter is within the parameter range, then the parameter adjustment instruction for the target proportional-integral-derivative parameter is directly output; If the candidate parameter is not within the parameter range, a clipping operation is performed on the candidate parameter to obtain a revised candidate parameter, and a parameter adjustment instruction is generated based on the revised candidate parameter. The clipping operation includes truncating the portion exceeding the corresponding upper limit to the upper limit and truncating the portion below the corresponding lower limit to the lower limit.

[0011] Secondly, this application discloses a proportional-integral-derivative parameter adjustment device for an industrial control loop, comprising: The data acquisition module is used to acquire closed-loop response data of industrial control loops, current proportional-integral-derivative parameter vectors, artificial control targets, controlled object model information, and parameter ranges. The indicator extraction module is used to extract response indicators from the closed-loop response data; the response indicators are used to characterize the dynamic response state of the industrial control loop. The adjustment direction determination module is used to search the preset state diagnosis knowledge base according to the response index and the artificial control target to obtain the target diagnosis rule, and determine the qualitative adjustment direction of the target proportional integral differential parameter based on the target diagnosis rule; The adjustment ratio determination module is used to determine the target parameter adjustment rule from a preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target, and to determine the relative adjustment ratio of the target proportional-integral-derivative parameter based on the target parameter adjustment rule; The adjustment instruction generation module is used to generate candidate parameters based on the current proportional-integral-derivative parameter vector and the relative adjustment ratio, and output parameter adjustment instructions for the target proportional-integral-derivative parameter when the candidate parameters are within the parameter range, so as to adjust the target proportional-integral-derivative parameter based on the parameter adjustment instructions.

[0012] Thirdly, this application discloses an electronic device, including: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned proportional-integral-derivative parameter adjustment method for the industrial control loop.

[0013] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the aforementioned proportional-integral-derivative parameter adjustment method for an industrial control loop.

[0014] As can be seen, in this application, the closed-loop response data of the industrial control loop, the current proportional-integral-derivative (PID) parameter vector, the artificial control target, the controlled object model information, and the parameter range are obtained; a response index is extracted from the closed-loop response data; the response index is used to characterize the dynamic response state of the industrial control loop; a preset state diagnosis knowledge base is searched according to the response index and the artificial control target to obtain target diagnosis rules, and a qualitative adjustment direction of the target PID parameter is determined based on the target diagnosis rules; a target parameter adjustment rule is determined from a preset parameter adjustment knowledge base according to the qualitative adjustment direction, the current PID parameter vector, the controlled object model information, and the artificial control target, and a relative adjustment ratio of the target PID parameter is determined based on the target parameter adjustment rule; candidate parameters are generated based on the current PID parameter vector and the relative adjustment ratio, and when the candidate parameters are within the parameter range, a parameter adjustment instruction for the target PID parameter is output so that the target PID parameter can be adjusted based on the parameter adjustment instruction.

[0015] In other words, when adjusting the proportional-integral-derivative (PID) parameters, this application first acquires the closed-loop response data, the current parameter vector, the artificial control objective, the controlled object model information, and the parameter range, and extracts the response index. Then, it searches the state diagnosis knowledge base and the parameter adjustment knowledge base to determine the qualitative adjustment direction and the relative adjustment ratio. Finally, it generates candidate parameters based on the parameter vector and the relative adjustment ratio, and outputs an adjustment command when the parameter range is satisfied. By decoupling qualitative diagnosis from quantitative adjustment, the adjustment direction has a diagnostic basis, and the adjustment ratio is constrained by the direction and checked for range deviations. This allows for the output of adjustment commands with diagnostic basis, directional constraints, and range verification. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This application discloses a flowchart of a method for adjusting the proportional-integral-derivative parameters of an industrial control loop. Figure 2 This application discloses a PID parameter tuning response curve before and after tuning. Figure 3 This is a schematic diagram of the proportional-integral-derivative parameter adjustment device for an industrial control loop disclosed in this application. Figure 4This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In current industrial process control, PID controllers are widely used due to their simple structure and robustness. However, their parameter tuning still relies on manual experience, empirical formulas, or optimization algorithms, which suffers from problems such as time consumption, limited applicability, and high on-site deployment costs. While introducing large models to assist parameter tuning can incorporate natural language understanding and rule-based reasoning, directly outputting parameters end-to-end from a single large model can easily lead to ambiguous diagnostic criteria, unstable formats, parameter out-of-bounds errors, and difficulty in verification, thus limiting practical engineering applications. Therefore, this application will specifically introduce a proportional-integral-derivative (PID) parameter tuning method for industrial control loops, which can solve the above problems.

[0020] See Figure 1 As shown in the figure, this application discloses a method for adjusting the proportional-integral-derivative parameters of an industrial control loop, including: Step S11: Obtain the closed-loop response data of the industrial control loop, the current proportional-integral-derivative parameter vector, the artificial control target, the controlled object model information, and the parameter range.

[0021] In this embodiment, the closed-loop response data can be step response data or on-site operating trend data; the current proportional-integral-derivative parameter vector includes the proportional coefficient, integral coefficient, and derivative coefficient; the manual control target can be an overshoot lower than a set threshold, a settling time lower than a set threshold, or oscillation decay meeting a set condition; the controlled object model information can include the identified transfer function parameters and time delay dominance flag; the parameter range includes the upper and lower limits of the allowed values ​​for the proportional coefficient, integral coefficient, and derivative coefficient. By simultaneously acquiring these five types of information—closed-loop response data, current proportional-integral-derivative parameter vector, manual control target, controlled object model information, and parameter range—a complete and traceable input basis is provided for subsequent state diagnosis and parameter adjustment, avoiding a lack of targeted diagnosis or adjustment due to missing input information.

[0022] Step S12: Extract response indicators from the closed-loop response data; the response indicators are used to characterize the dynamic response state of the industrial control loop.

[0023] In this embodiment, extracting response indicators from the closed-loop response data includes: extracting overshoot, settling time, integral error, oscillation level, and time delay characteristics from the closed-loop response data, and organizing the overshoot, settling time, integral error, oscillation level, and time delay characteristics into a closed-loop dynamic performance indicator vector to obtain the response indicators; wherein, the overshoot represents the closed-loop peak deviation; the settling time represents the response speed; the integral error represents the steady-state cumulative deviation; the oscillation level represents the severity of response oscillation; and the time delay characteristics represent the pure time delay characteristics of the controlled object. In this embodiment, the response indicators are organized into a closed-loop dynamic performance indicator vector x: ; Among them, M p T represents the overshoot, characterizing the closed-loop peak deviation; s J represents the settling time, characterizing the speed of response; e L represents the integral error, characterizing the steady-state cumulative deviation; o Indicates the oscillation level, characterizing the intensity of the oscillation; q d This represents the time delay characteristic, characterizing the impact of pure time delay on loop stability. Specifically, the overshoot M... p The adjustment time T can be calculated based on the relative deviation between the peak value of the closed-loop response and the set value. s The integral error J can be determined based on the time required for the response to enter and remain within a set error band. e The oscillation level L can be obtained by integrating the absolute value or squared error between the set value and the actual response value over the observation time window; o The time delay characteristic q can be obtained based on the number of peaks, the attenuation ratio of adjacent peaks, or the number of zero crossings of the response curve; d The time delay between input change and effective output response can be determined, or by the ratio of pure time delay to time constant in the controlled object identification model. Five indicators together constitute the loop performance evaluation index, corresponding to speed, overshoot, steady-state error, oscillation, and time delay, respectively. This transforms the abstract closed-loop response curve into a quantifiable and comparable indicator vector, providing a unified and objective input for subsequent rule-based diagnosis.

[0024] Step S13: Search the preset state diagnosis knowledge base according to the response index and the artificial control target to obtain the target diagnosis rule, and determine the qualitative adjustment direction of the target proportional-integral-derivative parameters based on the target diagnosis rule.

[0025] In this embodiment, the step of retrieving a preset state diagnosis knowledge base based on the response index and the manual control target to obtain target diagnosis rules, and determining the qualitative adjustment direction of the target proportional-integral-derivative parameters based on the target diagnosis rules, includes: determining the response index and the manual control target as query conditions, and retrieving the preset state diagnosis knowledge base through the query conditions to obtain candidate diagnosis rules matching the current operating condition; determining the target diagnosis rule from the candidate diagnosis rules, and determining the qualitative adjustment direction of the proportional coefficient, integral coefficient, and derivative coefficient in the target proportional-integral-derivative parameters based on the target diagnosis rule. The preset state diagnosis knowledge base is established using typical abnormal scenarios of industrial circuits as an index. Each diagnosis rule is bound to an index threshold range, an abnormal subtype, a diagnosis basis, and the qualitative adjustment direction of the proportional coefficient, integral coefficient, and derivative coefficient. Diagnosis rules can be stored in the form of "rule number, abnormal scenario, abnormal subtype, index threshold condition, diagnosis basis, qualitative adjustment direction of proportional coefficient, qualitative adjustment direction of integral coefficient, and qualitative adjustment direction of derivative coefficient". Abnormal scenarios include slow response, excessive overshoot, and oscillation abnormality. For example, when the overshoot exceeds the target threshold and the oscillation level exceeds the preset level, the abnormal scenario can be marked as excessive overshoot, and the abnormal subtype can be marked as underdamped or excessively strong integral action, with corresponding qualitative adjustment directions provided. Slow response scenarios can correspond to insufficient proportional or integral action; abnormal oscillation scenarios can correspond to multi-peak oscillation or time-delay-induced oscillation. During retrieval, the response index vector, manual control target, and abnormal scenario can be encoded into a query vector. The top few candidate diagnostic rules are obtained by sorting according to similarity, and rule filtering or reordering is performed based on the index threshold conditions. The qualitative adjustment directions include increasing, decreasing, and remaining unchanged, and corresponding qualitative adjustment directions are given for the proportional coefficient, integral coefficient, and derivative coefficient respectively. This state diagnosis process only outputs the qualitative adjustment direction and does not directly output absolute parameters. By first retrieving the state diagnosis knowledge base to obtain engineering diagnostic rules, and then the system performs fusion reasoning based on the retrieved diagnostic rules, since the diagnostic rules provide standardized engineering diagnostic priors for diagnosis, it can avoid the confusion and diagnostic bias caused by relying solely on general knowledge reasoning, thereby ensuring that the output qualitative adjustment direction has clear diagnostic basis and can be verified by engineers.

[0026] Step S14: Based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target, determine the target parameter adjustment rule from the preset parameter adjustment knowledge base, and determine the relative adjustment ratio of the target proportional-integral-derivative parameter based on the target parameter adjustment rule.

[0027] In this embodiment, determining the target parameter adjustment rule from the preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-differential parameter vector, the controlled object model information, and the artificial control target includes: constructing a composite query condition based on the abnormal scenario corresponding to the qualitative adjustment direction, the time delay characteristics and gain characteristics in the controlled object model information; and retrieving the preset parameter adjustment knowledge base based on the composite query condition to match the target parameter adjustment rule corresponding to the current operating condition from the preset parameter adjustment knowledge base. Specifically, the preset parameter adjustment knowledge base is established using a composite index of abnormal scenarios, time delay characteristics of the controlled object, and gain characteristics of the controlled object. It stores the parameter adjustment range under different operating conditions and includes upper and lower limits for proportional, integral, and derivative coefficients. Parameter adjustment rules can be stored in the form of "rule number, abnormal scenario, abnormal subtype, controlled object gain range, time delay dominance flag, recommended relative adjustment range for proportional coefficient, recommended relative adjustment range for integral coefficient, recommended relative adjustment range for derivative coefficient, and upper and lower limit prompts for parameters." The controlled object model information includes the identified transfer function parameters and time delay dominance flag. During retrieval, a query vector is constructed based on the abnormal scenario, abnormal subtype, controlled object model information, and parameter range. Similarity retrieval, precise rule matching, and reordering are performed to constrain the range of relative adjustment ratio values.

[0028] In this embodiment, determining the relative adjustment ratio of the target proportional-integral-derivative (PID) parameter based on the target parameter adjustment rule includes: determining a first adjustment ratio of the proportional coefficient in the target PLD parameter according to the proportional coefficient adjustment range bound in the target parameter adjustment rule; determining a second adjustment ratio of the integral coefficient in the target PLD parameter according to the integral coefficient adjustment range bound in the target parameter adjustment rule; and determining a third adjustment ratio of the derivative coefficient in the target PLD parameter according to the derivative coefficient adjustment range bound in the target parameter adjustment rule. Specifically, the sign of the first adjustment ratio is consistent with the qualitative adjustment direction of the proportional coefficient; the sign of the second adjustment ratio is consistent with the qualitative adjustment direction of the integral coefficient; and the sign of the third adjustment ratio is consistent with the qualitative adjustment direction of the derivative coefficient. Specifically, using the qualitative adjustment direction determined by the state diagnosis as a hard constraint, the output relative adjustment ratio r for the proportional coefficient, integral coefficient, and derivative coefficient is: ; Where r p r i r dThese represent the relative adjustment ratios of the proportional coefficient, integral coefficient, and derivative coefficient, respectively. The inference process for this parameter only outputs the relative adjustment ratios, not directly new absolute parameters. Furthermore, the relative adjustment ratios are consistent with the qualitative adjustment direction to prevent contradictions between qualitative diagnosis and quantitative parameter adjustment ratios: when the qualitative adjustment direction of a parameter is increasing, the corresponding relative adjustment ratio is positive; when it is decreasing, it is negative; and when it remains unchanged, it is zero. By using the relative adjustment ratio as an intermediate output and the qualitative adjustment direction as a hard constraint, and because the relative adjustment ratio does not directly provide absolute parameters, it can adapt to the differences in parameter numerical scales of controlled objects with different dynamic characteristics, and aligns with the practical habit of gradually adjusting parameters proportionally in industrial settings. This avoids contradictions between diagnosis and adjustment and enhances practical adaptability.

[0029] Step S15: Generate candidate parameters based on the current proportional-integral-derivative parameter vector and the relative adjustment ratio, and output a parameter adjustment instruction for the target proportional-integral-derivative parameter when the candidate parameters are within the parameter range, so as to adjust the target proportional-integral-derivative parameter based on the parameter adjustment instruction.

[0030] In this embodiment, generating candidate parameters based on the current proportional-integral-differential parameter vector and the relative adjustment ratio includes: forming a three-dimensional relative adjustment ratio vector by combining the first adjustment ratio, the second adjustment ratio, and the third adjustment ratio; and determining candidate parameters by multiplying the three-dimensional relative adjustment ratio vector with the current proportional-integral-differential parameter vector element by element.

[0031] In this embodiment, the system generates candidate PID parameters based on the current PID parameter vector and the relative adjustment ratio: ; Where K0 represents the current PID parameter vector, K C represents the candidate PID parameter vector, r represents the three-dimensional relative adjustment ratio vector, and ⊙ represents element-wise multiplication.

[0032] In this embodiment, the step of outputting a parameter adjustment instruction for the target proportional-integral-derivative (PID) parameter when the candidate parameter is within the parameter range includes: determining whether each parameter in the candidate parameters falls between the lower limit and the upper limit of the corresponding parameter value in the parameter range; if the candidate parameter is within the parameter range, directly outputting a parameter adjustment instruction for the target PID parameter; if the candidate parameter is not within the parameter range, performing a clipping operation on the candidate parameter to obtain a revised candidate parameter, and generating the parameter adjustment instruction based on the revised candidate parameter; the clipping operation includes truncating the portion exceeding the corresponding upper limit to the upper limit, and truncating the portion below the corresponding lower limit to the lower limit. That is, determining whether each parameter in the candidate parameters falls between the lower limit and the upper limit of the corresponding parameter value in the parameter range. If the candidate parameter meets the parameter range, a parameter adjustment instruction is output, including the candidate parameter, state diagnosis conclusion, qualitative adjustment direction, and parameter adjustment description. If the candidate parameter exceeds the corresponding upper and lower limits, one of the following two processing methods is executed: First, a limiting process is used to trim the out-of-bounds parameter to the corresponding upper and lower limits to obtain a corrected candidate parameter, and a parameter adjustment instruction is output based on the corrected candidate parameter. Second, the parameter over-limit information is sent back to the parameter inference process, which regenerates the relative adjustment ratio under the hard constraint of the qualitative adjustment direction until the candidate parameter meets the parameter range constraint, and then the corresponding parameter adjustment instruction is output. Since a value boundary verification process is added before all candidate parameters are issued, and two fault-tolerant handling schemes, limiting trimming and regeneration, are provided for out-of-bounds parameters, the direct deployment of out-of-range parameters can be avoided at the algorithm link level, thereby significantly reducing the risk of control loop instability.

[0033] As can be seen, in this embodiment, the closed-loop response data of the industrial control loop, the current proportional-integral-derivative (PID) parameter vector, the artificial control target, the controlled object model information, and the parameter range are acquired; a response index is extracted from the closed-loop response data; the response index is used to characterize the dynamic response state of the industrial control loop; a preset state diagnosis knowledge base is searched according to the response index and the artificial control target to obtain target diagnosis rules, and a qualitative adjustment direction of the target PID parameter is determined based on the target diagnosis rules; a target parameter adjustment rule is determined from the preset parameter adjustment knowledge base according to the qualitative adjustment direction, the current PID parameter vector, the controlled object model information, and the artificial control target, and a relative adjustment ratio of the target PID parameter is determined based on the target parameter adjustment rule; candidate parameters are generated based on the current PID parameter vector and the relative adjustment ratio, and a parameter adjustment instruction for the target PID parameter is output when the candidate parameter is within the parameter range, so that the target PID parameter can be adjusted based on the parameter adjustment instruction. In other words, when adjusting the proportional-integral-derivative (PID) parameters, this application first acquires the closed-loop response data, the current parameter vector, the artificial control objective, the controlled object model information, and the parameter range, and extracts the response index. Then, it searches the state diagnosis knowledge base and the parameter adjustment knowledge base to determine the qualitative adjustment direction and the relative adjustment ratio. Finally, it generates candidate parameters based on the parameter vector and the relative adjustment ratio, and outputs an adjustment command when the parameter range is satisfied. By decoupling qualitative diagnosis from quantitative adjustment, the adjustment direction has a diagnostic basis, and the adjustment ratio is constrained by the direction and checked for range deviations. This allows for the output of adjustment commands with diagnostic basis, directional constraints, and range verification.

[0034] In summary, this application employs a two-layer structure—state diagnosis retrieval enhancement generation and parameter inference retrieval enhancement generation—to separate the handling of "judging loop problems" and "providing adjustment ratios." This effectively compensates for the shortcomings of traditional end-to-end large model tuning methods, such as lack of interpretability and difficulty in manual verification. Furthermore, the state diagnosis knowledge base constructs a retrieval index based on typical adverse industrial control conditions, such as slow response, excessive overshoot, and abnormal oscillations, providing standardized engineering diagnostic prior rules for large model inference. This effectively avoids the problems of condition identification confusion and diagnostic bias caused by models relying solely on general textual knowledge for inference. The parameter inference layer uses relative adjustment ratios as intermediate outputs, combined with parameter iteration calculation formulas to solve for candidate PID parameters. This adapts to the differences in parameter numerical scales of controlled objects with different dynamic characteristics and aligns with the practical habit of gradually adjusting parameters in industrial settings, resulting in stronger adaptability. A value boundary verification process is added before all candidate PID parameters are issued, providing two fault-tolerant handling schemes for parameters exceeding limits: amplitude limiting and regeneration. This avoids the direct deployment of parameters exceeding the range at the algorithmic level, significantly reducing the risk of control loop instability.

[0035] In a first specific embodiment, the temperature control loop at the top of the petrochemical distillation column experiences frequent fluctuations in feed composition and steam pressure. This application first acquires temperature step response data and current PID parameters. Response indicators such as overshoot and settling time are extracted from the closed-loop response data. A preset state diagnosis knowledge base is retrieved based on the manually controlled target to diagnose insufficient proportional gain and provide a qualitative adjustment direction for increasing the proportional coefficient. Then, based on this qualitative direction and the time delay and gain characteristics in the controlled object model, a composite query condition is constructed. A preset parameter adjustment knowledge base is retrieved to match the target parameter adjustment rule. The relative adjustment ratio of the proportional coefficient is determined according to the adjustment range bound by this rule, while the integral and derivative coefficients remain unchanged. This three-dimensional relative adjustment ratio vector is multiplied element-wise with the current PID parameter vector to generate candidate parameters. After parameter range verification and necessary amplitude limiting and clipping, a parameter adjustment command is output and sent to the distributed control system, effectively suppressing temperature oscillations.

[0036] In another specific embodiment, when faced with false liquid levels caused by feedwater flow disturbances, the three-impulse control loop for the boiler drum liquid level in a thermal power plant extracts the oscillation level and time delay characteristics of the liquid level response curve to diagnose excessive differential action, providing a qualitative direction for reducing the differential coefficient. Then, combined with the gain characteristics of the controlled object, the relative adjustment ratio of the differential coefficient is determined, while the proportional and integral coefficients remain unchanged. After candidate parameters are generated, they undergo boundary value verification. If they exceed the allowable range, amplitude limiting and clipping are performed. Finally, the adjustment command is sent to the distributed control system, significantly improving the stability of the liquid level response.

[0037] In another specific embodiment, in the temperature-pressure cascade control loop of the fine chemical batch reactor, considering the characteristics of intense exothermic reaction and significant time delay of the object, this application takes short adjustment time and low overshoot as the artificial control target. It successively diagnoses excessive integral action and improper proportional gain, and determines the relative adjustment ratio of integral coefficient and proportional coefficient in turn. After generating candidate parameters by multiplying them element by element and checking the parameter range, the parameters are updated online to ensure that the temperature accurately tracks the set curve throughout the entire reaction cycle.

[0038] In a verification operation based on measured closed-loop response data, a first-order hysteresis object is selected, whose transfer function can be expressed as G(s) = 0.8173e {-1.228s} / (3.623s+1) Where 0.8173 is the process gain, 3.623s is the time constant, and 1.228s is the pure time delay. The initial PID parameters are K. p =2、K i =2、K d =0.1, the manual control target is overshoot less than 10%. The state diagnosis layer outputs the abnormal scenario as excessive overshoot and the abnormal subtype as excessive integral action, and gives K. pUnchanged, K i Decrease, K d The qualitative parameter tuning direction remains unchanged; under the constraint of the above qualitative parameter tuning direction, the parameter inference layer outputs a relative adjustment ratio r = (0, -0.75, 0). Based on the current PID parameter vector K0 = (2, 2, 0.1) and the formula K... c =K0⊙(1+r) generates candidate PID parameters K c =(2, 0.5, 0.1). After parameter range checking, the measured overshoot decreased from 71.06% to 8.79%, and the settling time decreased from 20.16 s to 5.43 s. The response curves before and after tuning are shown below. Figure 2 As shown, by Figure 2 It is evident that after tuning, the peak overshoot and oscillation amplitude of the closed-loop response are reduced, and the response enters the vicinity of the set value more quickly, indicating that the relative adjustment ratio of the parameter inference layer output can improve the closed-loop dynamic quality of the identified object.

[0039] refer to Figure 3 As shown in the figure, this application also discloses a proportional-integral-derivative parameter adjustment device for an industrial control loop, comprising: The data acquisition module 11 is used to acquire the closed-loop response data of the industrial control loop, the current proportional-integral-derivative parameter vector, the artificial control target, the controlled object model information, and the parameter range; The indicator extraction module 12 is used to extract response indicators from the closed-loop response data; the response indicators are used to characterize the dynamic response state of the industrial control loop. The adjustment direction determination module 13 is used to search the preset state diagnosis knowledge base according to the response index and the artificial control target to obtain the target diagnosis rule, and determine the qualitative adjustment direction of the target proportional integral differential parameter based on the target diagnosis rule; The adjustment ratio determination module 14 is used to determine the target parameter adjustment rule from the preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information and the artificial control target, and to determine the relative adjustment ratio of the target proportional-integral-derivative parameter based on the target parameter adjustment rule; The adjustment instruction generation module 15 is used to generate candidate parameters based on the current proportional-integral-derivative parameter vector and the relative adjustment ratio, and output parameter adjustment instructions for the target proportional-integral-derivative parameter when the candidate parameters are within the parameter range, so as to adjust the target proportional-integral-derivative parameter based on the parameter adjustment instructions.

[0040] As can be seen, in this embodiment, when adjusting the proportional-integral-derivative (PID) parameters, the closed-loop response data, current parameter vector, artificial control target, controlled object model information, and parameter range are first acquired, and response indicators are extracted. Then, the state diagnosis knowledge base and parameter adjustment knowledge base are retrieved to determine the qualitative adjustment direction and relative adjustment ratio. Finally, candidate parameters are generated based on the parameter vector and relative adjustment ratio, and an adjustment command is output when the parameter range is satisfied. By decoupling qualitative diagnosis from quantitative adjustment, the adjustment direction has a diagnostic basis, and the adjustment ratio is constrained by the direction and checked for range to avoid exceeding the limits. In this way, adjustment commands with diagnostic basis, directional constraints, and range verification can be output.

[0041] In some specific embodiments, the indicator extraction module 12 may specifically include: The response index acquisition unit is used to extract overshoot, settling time, integral error, oscillation level, and time delay characteristics from the closed-loop response data, and organize the overshoot, settling time, integral error, oscillation level, and time delay characteristics into a dynamic performance index vector of the closed-loop circuit to obtain the response index; wherein, the overshoot represents the closed-loop peak deviation; the settling time represents the response speed; the integral error represents the steady-state cumulative deviation; the oscillation level represents the severity of response oscillation; and the time delay characteristics represent the pure time delay characteristics of the controlled object.

[0042] In some specific embodiments, the adjustment direction determination module 13 may specifically include: The rule matching unit is used to determine the response indicators and the manual control targets as query conditions, and to retrieve the preset state diagnosis knowledge base through the query conditions to obtain candidate diagnostic rules that match the current working condition. The qualitative adjustment direction determination unit is used to determine the target diagnostic rule from the candidate diagnostic rules, and determine the qualitative adjustment direction of the proportional coefficient, integral coefficient and differential coefficient in the target proportional-integral-differential parameters based on the target diagnostic rule.

[0043] In some specific embodiments, the adjustment ratio determination module 14 may specifically include: The query condition construction unit is used to construct composite query conditions based on the abnormal scenario corresponding to the qualitative adjustment direction, the time delay characteristics and gain characteristics in the controlled object model information; The parameter adjustment rule matching unit is used to retrieve the preset parameter adjustment knowledge base based on the composite query conditions, so as to match the target parameter adjustment rule corresponding to the current working condition from the preset parameter adjustment knowledge base.

[0044] In some specific embodiments, the adjustment ratio determination module 14 may specifically include: The first adjustment ratio unit is used to determine the first adjustment ratio of the proportional coefficient in the target proportional integral differential parameter according to the adjustment range of the proportional coefficient bound in the target parameter adjustment rule; The second adjustment ratio unit is used to determine the second adjustment ratio of the integral coefficient in the target proportional integral differential parameter according to the adjustment range of the integral coefficient bound in the target parameter adjustment rule. The third adjustment ratio unit is used to determine the third adjustment ratio of the differential coefficient in the target proportional integral differential parameter according to the differential coefficient adjustment range bound in the target parameter adjustment rule.

[0045] In some specific embodiments, the adjustment instruction generation module 15 may specifically include: A proportional vector determination unit is used to form a three-dimensional relative adjustment ratio vector by combining the first adjustment ratio, the second adjustment ratio, and the third adjustment ratio. The candidate parameter determination unit is used to determine candidate parameters by multiplying the three-dimensional relative adjustment scale vector and the current scale integral differential parameter vector element by element.

[0046] In some specific embodiments, the adjustment instruction generation module 15 may specifically include: The range determination unit is used to determine whether each parameter in the candidate parameters falls between the lower limit and the upper limit of the corresponding parameter in the parameter range; The first parameter adjustment instruction generation unit is used to directly output a parameter adjustment instruction for the target proportional-integral-derivative parameter if the candidate parameter is within the parameter range. The second parameter adjustment instruction generation unit is used to perform a limiting and pruning operation on the candidate parameter if the candidate parameter is not within the parameter range, so as to obtain a revised candidate parameter, and generate the parameter adjustment instruction based on the revised candidate parameter; the limiting and pruning operation includes truncating the part exceeding the corresponding upper limit value to the upper limit value, and truncating the part below the corresponding lower limit value to the lower limit value.

[0047] Furthermore, embodiments of this application also disclose an electronic device, Figure 4 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0048] Figure 4This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the proportional-integral-derivative parameter adjustment method for the industrial control loop disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0049] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0050] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0051] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the proportional-integral-derivative parameter adjustment method of the industrial control loop executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0052] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned proportional-integral-derivative parameter adjustment method for the industrial control loop. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0053] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0054] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0056] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0057] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for adjusting the proportional-integral-derivative parameters of an industrial control loop, characterized in that, include: Acquire closed-loop response data of industrial control loop, current proportional-integral-derivative parameter vector, artificial control target, controlled object model information, and parameter range; Response indicators are extracted from the closed-loop response data; these response indicators are used to characterize the dynamic response state of the industrial control loop. The preset state diagnosis knowledge base is retrieved based on the response index and the artificial control target to obtain the target diagnosis rule, and the qualitative adjustment direction of the target proportional-integral-derivative parameters is determined based on the target diagnosis rule. Based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target, the target parameter adjustment rule is determined from the preset parameter adjustment knowledge base, and the relative adjustment ratio of the target proportional-integral-derivative parameter is determined based on the target parameter adjustment rule; Candidate parameters are generated based on the current proportional-integral-derivative (PID) parameter vector and the relative adjustment ratio. When the candidate parameters are within the parameter range, a parameter adjustment instruction for the target PID parameter is output so that the target PID parameter can be adjusted based on the parameter adjustment instruction.

2. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to claim 1, characterized in that, Extracting response metrics from the closed-loop response data includes: The overshoot, settling time, integral error, oscillation level, and time delay characteristics are extracted from the closed-loop response data, and the overshoot, settling time, integral error, oscillation level, and time delay characteristics are organized into a dynamic performance index vector of the closed-loop circuit to obtain the response index. Wherein, the overshoot represents the closed-loop peak deviation; the settling time represents the response speed; the integral error represents the steady-state cumulative deviation; the oscillation level represents the severity of the response oscillation; and the time delay characteristic represents the pure time delay characteristic of the controlled object.

3. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to claim 1, characterized in that, The step of retrieving a preset state diagnosis knowledge base based on the response index and the artificial control target to obtain target diagnosis rules, and determining the qualitative adjustment direction of the target proportional-integral-derivative parameters based on the target diagnosis rules, includes: The response indicators and the manual control targets are determined as query conditions, and the preset state diagnosis knowledge base is retrieved through the query conditions to obtain candidate diagnosis rules that match the current working condition. The target diagnostic rule is determined from the candidate diagnostic rules, and the qualitative adjustment direction of the proportional coefficient, integral coefficient and differential coefficient in the target proportional-integral-differential parameters is determined based on the target diagnostic rule.

4. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to claim 1, characterized in that, The step of determining the target parameter adjustment rules from a preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target includes: Based on the abnormal scenarios corresponding to the qualitative adjustment direction, and the time delay and gain features in the controlled object model information, a composite query condition is constructed. The preset parameter adjustment knowledge base is retrieved based on the composite query conditions to match the target parameter adjustment rule corresponding to the current working condition from the preset parameter adjustment knowledge base.

5. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to claim 1, characterized in that, The step of determining the relative adjustment ratio of the target proportional integral derivative parameter based on the target parameter adjustment rule includes: Based on the adjustment range of the proportional coefficient bound in the target parameter adjustment rule, determine the first adjustment ratio of the proportional coefficient in the target proportional integral differential parameter; Based on the adjustment range of the integral coefficient bound in the target parameter adjustment rule, determine the second adjustment ratio of the integral coefficient in the target proportional integral differential parameter; Based on the adjustment range of the differential coefficients bound in the target parameter adjustment rules, determine the third adjustment ratio of the differential coefficients in the target proportional integral differential parameter; Wherein, the sign of the first adjustment ratio is consistent with the qualitative adjustment direction of the proportional coefficient; the sign of the second adjustment ratio is consistent with the qualitative adjustment direction of the integral coefficient; and the sign of the third adjustment ratio is consistent with the qualitative adjustment direction of the differential coefficient.

6. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to claim 5, characterized in that, The generation of candidate parameters based on the current proportional-integral-differential parameter vector and the relative adjustment ratio includes: The first adjustment ratio, the second adjustment ratio, and the third adjustment ratio are combined to form a three-dimensional relative adjustment ratio vector; Candidate parameters are determined by multiplying the three-dimensional relative adjustment scale vector and the current scale integral differential parameter vector element by element.

7. The method for adjusting the proportional-integral-derivative parameters of an industrial control loop according to any one of claims 1 to 6, characterized in that, The step of outputting a parameter adjustment instruction for the target proportional-integral-derivative parameter when the candidate parameter is within the parameter range includes: Determine whether each of the candidate parameters falls between the lower limit and the upper limit of the corresponding parameter value in the parameter range; If the candidate parameter is within the parameter range, then the parameter adjustment instruction for the target proportional-integral-derivative parameter is directly output; If the candidate parameter is not within the parameter range, a clipping operation is performed on the candidate parameter to obtain a revised candidate parameter, and a parameter adjustment instruction is generated based on the revised candidate parameter. The clipping operation includes truncating the portion exceeding the corresponding upper limit to the upper limit and truncating the portion below the corresponding lower limit to the lower limit.

8. A proportional-integral-derivative parameter adjustment device for an industrial control loop, characterized in that, include: The data acquisition module is used to acquire closed-loop response data of industrial control loops, current proportional-integral-derivative parameter vectors, artificial control targets, controlled object model information, and parameter ranges. The indicator extraction module is used to extract response indicators from the closed-loop response data; the response indicators are used to characterize the dynamic response state of the industrial control loop. The adjustment direction determination module is used to search the preset state diagnosis knowledge base according to the response index and the artificial control target to obtain the target diagnosis rule, and determine the qualitative adjustment direction of the target proportional integral differential parameter based on the target diagnosis rule; The adjustment ratio determination module is used to determine the target parameter adjustment rule from a preset parameter adjustment knowledge base based on the qualitative adjustment direction, the current proportional-integral-derivative parameter vector, the controlled object model information, and the artificial control target, and to determine the relative adjustment ratio of the target proportional-integral-derivative parameter based on the target parameter adjustment rule; The adjustment instruction generation module is used to generate candidate parameters based on the current proportional-integral-derivative parameter vector and the relative adjustment ratio, and output parameter adjustment instructions for the target proportional-integral-derivative parameter when the candidate parameters are within the parameter range, so as to adjust the target proportional-integral-derivative parameter based on the parameter adjustment instructions.

9. An electronic device, characterized in that, include: Memory is used to store computer programs; A processor for executing the computer program to implement the proportional-integral-derivative parameter adjustment method for an industrial control loop as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the proportional-integral-derivative parameter adjustment method for an industrial control loop as described in any one of claims 1 to 7.

Images