Controller operation and maintenance method, device and system

Abstract

The present application discloses a controller operation and maintenance method, device and system, which relates to the field of industrial control technology. The method includes: collecting configuration parameters and operating data of industrial field control subsystems; based on the configuration parameters and operating data, performing performance evaluation on the operating performance of the controller and obtaining performance evaluation results; when the performance evaluation result is that the operating performance of the controller is abnormal, judging whether a model mismatch occurs in the industrial process object based on the configuration parameters and operating data, and obtaining a model mismatch detection result; based on the model mismatch detection result, determining the type of influencing factors that affect the operating performance of the controller, and determining an operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors. The present application can ensure that the industrial controller continues to maintain the operating performance at the initial stage of commissioning, thereby achieving immediate and targeted operation and maintenance of the control loop of complex industrial process objects, and maintaining the stability of the industrial production quality of industrial process objects.

Description

Technical Field

[0001] The present application relates to the field of industrial control technology, and more specifically, to a controller operation and maintenance method, device and system. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] Industrial control software is the crystallization of industrial knowledge and experience accumulated over a long period of time in industrial automation. Industrial controllers built based on industrial control software can not only control the operation of products and equipment in real time, but also optimize the operating status of the system to achieve continuous, stable and efficient production. However, the operating performance of industrial controllers will gradually decline with the aging of industrial process object equipment, changes in working conditions, external disturbances and other factors during operation, which will affect production efficiency, quality and even production safety. Traditional operation and maintenance methods often require manual participation in regular evaluation, testing and optimization and other maintenance measures to ensure the effective operation of industrial controllers.

[0004] Traditional industrial controller operation and maintenance relies on the experience of professional engineers to diagnose controller performance, and it takes a lot of effort to discover abnormal loops. After discovering the abnormal control performance loop, professionals are still required to spend a lot of debugging effort to trim and debug the controller model and parameters, and the effect of operation and maintenance depends on the technical level of professionals. Therefore, traditional operation and maintenance methods often make it impossible for industrial controllers to continue to maintain the operating performance at the initial stage of commissioning. There is an urgent need for an operation and maintenance system with automatic operation capabilities for industrial controller performance monitoring, diagnosis and maintenance. Summary of the invention

[0005] The embodiments of the present application provide a controller operation and maintenance method, device and system to at least solve the problem in the prior art that traditional operation and maintenance methods make it impossible for industrial controllers to continuously maintain their operating performance in the initial stage of operation.

[0006] According to one aspect of the present application, a controller operation and maintenance method is also provided, wherein the controller controls an industrial process object in an industrial field control subsystem, and the operation and maintenance method comprises:

[0007] Collecting configuration parameters and operation data of the industrial field control subsystem, wherein the operation data includes control instruction data of the controller and state data and an initial model of the industrial process object, wherein the configuration parameters are parameters required for operation and maintenance of the operation performance of the controller, and the initial model is a mathematical model that conforms to the initial characteristics of the industrial process object;

[0008] According to the configuration parameters and the operating data, a performance evaluation is performed on the operating performance of the controller and a performance evaluation result is obtained;

[0009] When the performance evaluation result is that the operating performance of the controller is abnormal, determining whether a model mismatch occurs in the industrial process object according to the configuration parameters and the operating data, and obtaining a model mismatch detection result;

[0010] When the performance evaluation result shows that the operating performance of the controller is abnormal, the type of influencing factors affecting the operating performance of the controller is determined according to the model mismatch detection result, and an operation and maintenance strategy for optimizing the operating performance of the controller is determined based on the type of influencing factors.

[0011] In some embodiments, the step of determining the type of influencing factor affecting the operating performance of the controller according to the model mismatch detection result includes:

[0012] When the model mismatch detection result is that the industrial process object has a model mismatch, determining the type of influencing factor is that the real characteristic currently possessed by the industrial process object is different from the initial characteristic;

[0013] Then, the step of determining the operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors includes:

[0014] Acquire a recommended control model determined according to the operating data, wherein the recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object;

[0015] A first recommended controller parameter matching the design index of the controller is determined according to the recommended control model, so as to optimize the operation performance of the controller using a first optimization result composed of the first recommended controller parameter and the recommended control model.

[0016] In some embodiments, the step of determining the type of influencing factor affecting the operating performance of the controller according to the model mismatch detection result includes:

[0017] When the model mismatch detection result is that the industrial process object does not have a model mismatch, determining the type of influencing factor is that the external environment in which the industrial process object is currently operating is different from the initial operating environment;

[0018] Then, the step of determining the operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors includes:

[0019] Second recommended controller parameters matching the design index of the controller are determined according to the initial model, so as to optimize the operation performance of the controller using a second optimization result formed by the second recommended controller parameters.

[0020] In some embodiments, the configuration parameters include a preset control performance change threshold, a reference data time period, and an evaluation data time period, wherein the step of performing a performance evaluation on the operating performance of the controller and obtaining a performance evaluation result according to the configuration parameters and the operating data includes:

[0021] Acquire reference operating data collected within the reference data time period, wherein the reference operating data is the operating data collected when the operating performance of the controller is normal;

[0022] Using the operating data collected within the evaluation data time period as the operating data to be evaluated;

[0023] According to the control performance change threshold, the to-be-evaluated operating data and the reference operating data, the operating performance of the controller is evaluated and a performance evaluation result is obtained.

[0024] In some embodiments, the configuration parameters further include a preset model mismatch threshold and a mismatch detection data time period, and the step of determining whether a model mismatch occurs in the industrial process object according to the configuration parameters and the operating data includes:

[0025] The operating data collected in the mismatch detection data time period is acquired, and whether a model mismatch occurs in the industrial process object is determined according to the model mismatch threshold and the operating data.

[0026] In some of the embodiments, the configuration parameters further include a performance evaluation period, and the performance evaluation period is used to specify a time period for performing a performance evaluation on the operating performance of the controller.

[0027] In some of the embodiments, the configuration parameters further include a mismatch detection period, and the mismatch detection period is used to specify a time period for determining whether a model mismatch occurs in the industrial process object.

[0028] According to another aspect of the present application, a controller operation and maintenance device is further provided, the controller operation and maintenance device is used to operate and maintain a controller that controls an industrial process object in an industrial field control subsystem, the controller operation and maintenance device comprising:

[0029] A first data reading module, a performance evaluation module, a mismatch detection module and a control optimization module;

[0030] The first data reading module is used to collect configuration parameters and operation data of the industrial field control subsystem, the operation data includes control instruction data of the controller and state data and initial model of the industrial process object, the configuration parameters are parameters required for operation and maintenance of the operation performance of the controller, and the initial model is a mathematical model that conforms to the initial characteristics of the industrial process object;

[0031] The performance evaluation module is used to perform a performance evaluation on the operating performance of the controller and obtain a performance evaluation result according to the configuration parameters and the operating data;

[0032] The mismatch detection module is used to, when the performance evaluation result is that the operating performance of the controller is abnormal, determine whether a model mismatch occurs in the industrial process object according to the configuration parameters and the operating data, and obtain a model mismatch detection result;

[0033] The control optimization module is used to determine the type of influencing factors affecting the operating performance of the controller according to the model mismatch detection result, and determine the operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors.

[0034] In some embodiments, a model updating module and a first data writing module are further included, wherein the control optimization module determines the type of influencing factor affecting the operating performance of the controller according to the model mismatch detection result, including:

[0035] When the model mismatch detection result is that the industrial process object has a model mismatch, determining the type of influencing factor is that the real characteristic currently possessed by the industrial process object is different from the initial characteristic;

[0036] Then, the step of the control optimization module determining the operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors includes:

[0037] Acquire a recommended control model determined by the model updating module according to the operating data, wherein the recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object;

[0038] Determine first recommended controller parameters that match the design indicators of the controller based on the recommended control model, and send the first recommended controller parameters to the first data writing module, which feeds back a first optimization result consisting of the first recommended controller parameters and the recommended control model to the controller to optimize the operating performance of the controller using the first optimization result.

[0039] In some embodiments, a first data writing module is further included, and the control optimization module determines the type of influencing factor affecting the operating performance of the controller according to the model mismatch detection result, including:

[0040] When the model mismatch detection result is that the industrial process object does not have a model mismatch, determining the type of influencing factor is that the external environment in which the industrial process object is currently operating is different from the initial operating environment;

[0041] Then, the step of the control optimization module determining the operation and maintenance strategy for optimizing the operating performance of the controller based on the type of influencing factors includes:

[0042] Determine second recommended controller parameters that match the design indicators of the controller based on the initial model, and send the second recommended controller parameters to the first data writing module, which feeds back a second optimization result consisting of the second recommended controller parameters to the controller to optimize the operating performance of the controller using the second optimization result.

[0043] According to another aspect of the present application, a controller operation and maintenance system is also provided, including:

[0044] Controller operation and maintenance device;

[0045] An industrial field control subsystem, the industrial field control subsystem comprising a controller to be operated and maintained and an industrial process object controlled by the controller;

[0046] The controller operation and maintenance device is used to operate and maintain the controller.

[0047] In some of the embodiments, a data acquisition and storage subsystem is also included, and the data acquisition and storage subsystem is used to acquire and store the operating data generated by the industrial field control subsystem and the data generated when the controller operation and maintenance device operates and maintains the controller.

[0048] The embodiment of the present application collects the operating data of the industrial field control subsystem where the controller is located and the configuration parameters required for operation and maintenance, and determines the cause of the controller's control performance degradation based on the configuration parameters and operating data, thereby improving the performance of the control loop according to the type of factors affecting the performance degradation, and automatically monitoring, diagnosing and maintaining the operating performance of the industrial controller. Ensure that the industrial controller continues to maintain the operating performance at the initial stage of commissioning, achieve instant and targeted operation and maintenance of the control loop of complex industrial process objects, and maintain the stability of the industrial production quality of industrial process objects. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0050] Figure 1 is a flow chart of a controller operation and maintenance method according to an embodiment of the present application;

[0051] Figure 2 is a schematic diagram of a controller operation and maintenance device according to an embodiment of the present application;

[0052] Figure 3 is a schematic diagram of a controller operation and maintenance system according to an embodiment of the present application;

[0053] Figure 4 is a block diagram of the operation and maintenance principle of a controller operation and maintenance system according to an embodiment of the present application;

[0054] Figure 5 yes Figure 4 A schematic diagram of the operation and maintenance process of the controller operation and maintenance system shown.

[0055] The above drawings include the following reference numerals:

[0056] 10. Controller operation and maintenance device; 11. First data reading module; 12. Mismatch detection module; 13. Performance evaluation module; 14. Control optimization module; 15. Model updating module; 16. First data writing module; 17. Monitoring interface module; 20. Data acquisition and storage subsystem; 21. Second data writing module; 22. Data cloud storage module; 23. Second data reading module; 30. Industrial field control subsystem; 31. Third data reading module; 32. Controller; 33. Industrial process object; 34. Third data writing module; 35. Set value input module. DETAILED DESCRIPTION

[0057] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0058] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0059] In the first embodiment of the present invention, a controller 32 operation and maintenance method is provided. The controller 32 controls the industrial process object 33 in the industrial field control subsystem 30. The industrial field control subsystem 30 and the controller operation and maintenance device 10 for operating and maintaining the controller 32 constitute a controller 32 operation and maintenance system. Figure 1 When the controller operation and maintenance device 10 uses the operation and maintenance method provided by the embodiment of the present invention to operate and maintain the controller 32, the operation and maintenance method includes the following steps:

[0060] Step S11: Collect configuration parameters and operating data of the industrial field control subsystem 30. The operating data includes control instruction data of the controller 32 and status data and initial model of the industrial process object 33. The configuration parameters are parameters required for operation and maintenance of the operating performance of the controller 32. The initial model is a mathematical model that conforms to the initial characteristics of the industrial process object 33.

[0061] In the embodiment of the present invention, the initial characteristics of the industrial process object 33 are the characteristics before the operating performance of the controller 32 changes. Since the operation and maintenance process is a control process that continuously cycles to maintain the operating performance of the controller 32, the initial model of the industrial process object 33 can be a nominal model that meets the original characteristics of the industrial process object 33 determined when the controller 32 is designed in some of the operation and maintenance rounds. Before operation and maintenance, the user configures the nominal model in the configuration parameters, or it can be a recommended control model that meets the actual characteristics of the industrial process object 33 during the last operation and maintenance determined based on the operation data after the last operation and maintenance. The control instruction data of the controller 32, such as the speed, valve opening, feed amount, etc., are instruction data for controlling the industrial process object 33, and the status data is determined according to the controlled industrial process object 33, such as temperature, flow, weight, density, etc.

[0062] Step S12: Based on the configuration parameters and operating data, the operating performance of the controller 32 is evaluated and a performance evaluation result is obtained. The performance evaluation result is used to determine whether the control performance of the controller 32 has deteriorated. If not, the industrial field control subsystem where the controller 32 is located continues to operate. If it is determined that the performance has deteriorated, a corresponding operation and maintenance mechanism is implemented for the controller 32, which can avoid unnecessary monitoring and diagnosis processes, improve operation and maintenance efficiency, and ensure the effectiveness and timeliness of operation and maintenance of the controller 32.

[0063] Step S13: When the performance evaluation result shows that the operating performance of the controller 32 is abnormal, it is determined whether a model mismatch occurs in the industrial process object 33 according to the configuration parameters and the operating data, and a model mismatch detection result is obtained.

[0064] In the embodiment of the present invention, when judging whether the industrial process object 33 has model mismatch according to the configuration parameters and the operation data, the model mismatch detection algorithm adopted includes but is not limited to the correlation analysis method, the system identification method, etc. The configuration parameters used for model mismatch detection include the pre-set model mismatch threshold and the mismatch detection data time period. When step S13 judges whether the industrial process object 33 has model mismatch according to the configuration parameters and the operation data, the operation data collected in the mismatch detection data time period can be obtained, and the industrial process object 33 can be judged whether the model mismatch has occurred according to the model mismatch threshold and the selected operation data. For example, the data within the period from 12:00 of a certain year / month / day to 15:00 of a certain year / month / day is selected for mismatch detection. The configuration parameters also include the mismatch detection cycle, which is used to specify the time period for judging whether the industrial process object 33 has model mismatch, and determine how often the mismatch detection process is run, so as to automatically judge the operation performance of the controller 32 based on the data within a specific time period, and realize regular automatic monitoring and diagnosis of the operation performance of the controller 32.

[0065] Step S14: According to the model mismatch detection result, determine the type of influencing factors affecting the operating performance of the controller 32, and determine the operation and maintenance strategy for optimizing the operating performance of the controller 32 based on the type of influencing factors, so as to optimize the operating performance of the controller 32 according to different influencing factors to improve the performance of the control loop.

[0066] When the embodiment of the present invention performs performance evaluation on the operating performance of the controller 32 according to the configuration parameters and operating data, the performance evaluation algorithm adopted includes but is not limited to the minimum variance control, the evaluation algorithm based on the linear quadratic Gaussian and other benchmarks, the evaluation algorithm based on abnormal detection and the evaluation algorithm based on data statistical indicators. The configuration parameters collected during the performance evaluation include the preset control performance change threshold, the benchmark data time period and the evaluation data time period, wherein step S12 performs performance evaluation on the operating performance of the controller 32 according to the configuration parameters and the operating data and obtains the performance evaluation result, specifically including:

[0067] Obtain the benchmark operating data collected during the benchmark data time period, wherein the benchmark operating data is the operating data collected when the operating performance of the controller 32 is not abnormal. The operating data collected during the evaluation data time period is used as the operating data to be evaluated. According to the control performance change threshold, the operating data to be evaluated and the benchmark operating data, the operating performance of the controller 32 is evaluated and the performance evaluation result is obtained. The benchmark operating data collected during the benchmark data time period is the data when the operating performance of the controller 32 is good. The operating data to be evaluated collected during the evaluation data time period, such as the operating data during the period from 11:00 in a certain year / month / day to 14:00 in a certain year / month / day, is used for performance evaluation. The configuration parameters also include a performance evaluation cycle, which is used to specify the time period for the performance evaluation of the operating performance of the controller 32, thereby determining how often the performance evaluation algorithm is run to evaluate the operating performance of the controller 32, such as running the performance evaluation algorithm once a day or a week to evaluate the operating performance of the controller 32. The operation and maintenance objectives of the controller 32 can be achieved under the operation and maintenance modes such as cost saving, instant access, and uninterrupted operation and maintenance, so as to continuously maintain the stability of industrial production quality.

[0068] When the performance of the controller 32 is evaluated, the embodiment of the present invention takes the evaluation algorithm based on data statistical indicators as an example to introduce a specific implementation process of the performance evaluation of the controller 32. The basic idea of ​​the evaluation algorithm based on data statistical indicators is to compare the changes in the operating performance of the controller 32 in the evaluation data time period relative to the benchmark data time period by calculating the benchmark operating data in the benchmark data time period and the statistical indicators related to the operating data to be evaluated in the evaluation data time period, and then point out the performance improvement space of the industrial field control subsystem 30 under various changes, wherein the statistical indicators include but are not limited to the control error standard deviation, online rate, covariance, saturation rate, in-range rate, number of changes, minimum variance performance indicators, etc. Taking the evaluation and monitoring method based on covariance as an example, when the embodiment of the present invention evaluates the operating performance of the controller 32 according to the control performance change threshold, the operating data to be evaluated and the benchmark operating data, the main implementation steps are as follows:

[0069] 11) Select the output data with better performance in the benchmark data time period as the benchmark data, obtain the operation data in the monitoring phase (i.e., the evaluation data time period), and the user sets the control performance change threshold in the configuration parameters. The control performance change threshold includes a performance deterioration threshold and a performance improvement threshold;

[0070] 12) Calculate the covariance matrix of the output variables of the industrial process object 33 in the industrial field control subsystem 30 in the reference data time period and the covariance matrix of the evaluation data time period, perform generalized eigenvalue decomposition on the covariance matrix of the reference data time period and the covariance matrix of the evaluation data time period, obtain the matrix eigenvalues ​​of each covariance matrix, and calculate the overall performance index of the monitoring stage;

[0071] 13) Perform generalized eigenvalue statistical inference to obtain the confidence interval of the matrix eigenvalues ​​of each covariance matrix in the corresponding characteristic direction, compare it with the performance deterioration threshold and performance improvement threshold set by the user, and judge the change of performance in the corresponding characteristic direction;

[0072] 14) By vector projection, determine the main output variables that cause the characteristic direction of deterioration or improvement of the performance of the controller 32.

[0073] When evaluating the performance of the controller 32, the embodiment of the present invention takes the evaluation algorithm based on the minimum variance control benchmark as an example to introduce the specific implementation process of the performance evaluation of the controller 32. The basic idea of ​​the minimum variance control evaluation algorithm is to analyze the process history data and the disturbance model to find the ideal minimum variance of the process output of the current control system, and use this as the evaluation benchmark to point out the performance improvement space of the control system under various working conditions. The main steps are as follows: based on the historical closed-loop output data, build a time series model, perform step response expansion, solve the minimum variance control output item, based on the real-time closed-loop output data, solve the control error variance, calculate the minimum variance performance index, and when the minimum variance performance index is significantly less than 1, it means that the performance of the controller 32 may be abnormal and there is room for improvement.

[0074] Wherein, the step S13 of determining the type of influencing factors affecting the operating performance of the controller 32 according to the model mismatch detection result includes:

[0075] When the model mismatch detection result is that the industrial process object 33 has a model mismatch, the type of influencing factor is determined to be that the actual characteristics currently possessed by the industrial process object 33 are different from the initial characteristics. That is, if a mismatch occurs in the model detection, then the initial characteristics of the industrial process object 33 itself, such as the structure and parameters, have changed significantly compared to the initial stage of commissioning or the last operation and maintenance. For example, during the first operation and maintenance, the initial characteristics may be the original characteristics corresponding to the nominal model of the industrial process object 33 determined when the controller 32 was designed, that is, the original characteristics determined when the industrial process object 33 was designed have changed significantly. Then, step S13 determines the operation and maintenance strategy for optimizing the operating performance of the controller 32 based on the type of influencing factors, including:

[0076] The recommended control model determined according to the operation data is obtained. The recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object 33. The embodiment of the present invention adopts a model online identification algorithm to determine the recommended control model according to the operation data, wherein the model online identification algorithm includes but is not limited to the least square method, the prediction error algorithm, the subspace method, etc. Then, the first recommended controller parameters matching the design index of the controller 32 are determined according to the recommended control model, so as to optimize the operation performance of the controller 32 using the first recommended controller parameters and the first optimization result composed of the recommended control model, so as to ensure that the controller 32 continues to maintain the operation performance at the initial stage of commissioning when the structure, parameters and other characteristics of the industrial process object 33 change, wherein the recommended control model includes a finite step response model, a transfer function model, and a state space equation model. The first recommended controller parameters include the proportional coefficient Kp, the integral coefficient Ti, the differential coefficient Td of the PID controller, the feedback gain of the linear quadratic controller LQR, the smoothing factor, the equivalent deviation, the closed-loop reference time, the model feedback correction parameter, etc. of the model predictive controller MPC, wherein the PID controller is the abbreviation of the proportional-integral-differential controller.

[0077] Secondly, step S13 includes the following steps: determining the type of influencing factors affecting the operating performance of the controller 32 according to the model mismatch detection result:

[0078] When the model mismatch detection result is that the industrial process object 33 does not have a model mismatch, the type of influencing factor is determined to be that the external environment in which the industrial process object 33 is currently operating is different from the initial operating environment. In other words, there is an external disturbance in the operating environment of the industrial process object 33. Then step S13 determines the operation and maintenance strategy for optimizing the operating performance of the controller 32 based on the type of influencing factor, including:

[0079] According to the initial model, second recommended controller parameters matching the design index of the controller 32 are determined, and the operating performance of the controller 32 is optimized using the second optimization result formed by the second recommended controller parameters, so as to ensure that the controller 32 continues to maintain the operating performance at the initial stage of commissioning when the operating environment of the industrial process object 33 changes externally. The second recommended controller parameters include the proportional coefficient Kp, integral coefficient Ti, differential coefficient Td of the PID controller, the feedback gain of the linear quadratic controller LQR, the smoothing factor, equivalent deviation, closed-loop reference time, model feedback correction parameters, etc. of the model predictive controller MPC.

[0080] The controller 32 operation and maintenance method proposed in the embodiment of the present invention includes controller 32 performance evaluation, model mismatch detection of controlled industrial procedure objects, model online identification and control parameter optimization, etc., forming a complete and fully autonomous closed-loop operation and maintenance process. By collecting the operation data of the industrial field control subsystem 30 where the controller 32 is located and the configuration parameters required for operation and maintenance, the cause of the decline in the control performance of the controller 32 is determined according to the configuration parameters, the initial model of the industrial process object 33 and the control instruction data of the controller 32, the performance of the control loop is improved in a targeted manner, and the operation performance of the industrial controller 32 is automatically monitored, diagnosed and maintained. Ensure that the industrial controller 32 continues to maintain the operating performance at the initial stage of commissioning, realize the immediate and targeted operation and maintenance of the control loop of the complex industrial process object 33, and ensure the stability of the industrial production quality of the industrial process object 33.

[0081] Compared with traditional operation and maintenance methods, the controller 32 operation and maintenance method proposed in the embodiment of the present invention avoids the tediousness and inefficiency of frequent manual maintenance, and can locate the cause of the decline in control performance of the controller 32 from the configured configuration parameters and the operating data of the industrial field control subsystem 30 where the controller 32 is located, thereby improving the performance of the control loop according to different reasons.

[0082] In the second embodiment of the present invention, a controller operation and maintenance device 10 is further provided. The controller operation and maintenance device 10 is used to operate and maintain a controller 32 that controls an industrial process object 33 in an industrial field control subsystem 30. Figure 2 , the controller operation and maintenance device 10 includes a first data reading module 11, a performance evaluation module 13, a mismatch detection module 12 and a control optimization module 14. The first data reading module 11 is used to collect configuration parameters and operation data of the industrial field control subsystem 30. The operation data includes control instruction data of the controller 32 and state data and an initial model of the industrial process object 33. The configuration parameters are parameters required for operation and maintenance of the operation performance of the controller 32, and the initial model is a mathematical model that conforms to the initial characteristics of the industrial process object 33. The initial model of the industrial process object 33 is a nominal model that conforms to the original characteristics of the industrial process object 33 determined during the design of the controller 32 in some of the operation and maintenance rounds, or it can be a recommended control model that conforms to the actual characteristics of the industrial process object 33 during the last operation and maintenance determined according to the operation data after the last operation and maintenance.

[0083] The performance evaluation module 13 is used to evaluate the operating performance of the controller 32 according to the configuration parameters and the operating data and obtain the performance evaluation result. When the performance evaluation result shows that the control performance of the controller 32 has deteriorated, the control optimization module 14 is used to perform the optimization and update process of the controller 32, thereby avoiding unnecessary monitoring and diagnosis processes, improving the operation and maintenance efficiency, and ensuring the effectiveness and timeliness of the operation and maintenance.

[0084] When the performance evaluation module 13 of the embodiment of the present invention performs performance evaluation on the operating performance of the controller 32 according to the configuration parameters and the operating data, the performance evaluation algorithm adopted includes but is not limited to the minimum variance control, the evaluation algorithm based on the linear quadratic Gaussian and other benchmarks, the evaluation algorithm based on the abnormality detection and the evaluation algorithm based on the data statistical index. The configuration parameters collected during the performance evaluation include the preset control performance change threshold, the benchmark data time period and the evaluation data time period, wherein the performance evaluation module 13 performs the performance evaluation on the operating performance of the controller 32 according to the configuration parameters and the operating data and obtains the performance evaluation result. The steps specifically include:

[0085] Obtain the benchmark operating data collected within the benchmark data time period, wherein the benchmark operating data is the operating data collected when the operating performance of the controller 32 is not abnormal. The operating data collected within the evaluation data time period is used as the operating data to be evaluated. According to the control performance change threshold, the operating data to be evaluated and the benchmark operating data, the operating performance of the controller 32 is evaluated and the performance evaluation result is obtained. The benchmark operating data collected within the benchmark data time period is the data when the operating performance of the controller 32 is better. The benchmark data time period is configured in the configuration parameters, and the controller operation and maintenance device 10 transmits the operating data of the industrial field control subsystem 30 collected within the time period specified by the benchmark data time period when the controller 32 is operating normally as the benchmark operating data to the performance evaluation module 13. The operating data to be evaluated collected within the evaluation data time period, such as the operating data within the period from 11:00 of a certain year / month / day to 14:00 of a certain year / month / day, is selected for performance evaluation. The configuration parameters also include a performance evaluation cycle, which is used to specify a time period for evaluating the operating performance of the controller 32, thereby determining how often the performance evaluation algorithm is run to evaluate the operating performance of the controller 32, such as running the performance evaluation algorithm once a day or a week to evaluate the operating performance of the controller 32. In this way, the operation and maintenance purpose of the controller 32 is achieved under the operation and maintenance modes of cost saving, instant access, and uninterrupted operation and maintenance, and the stability of industrial production quality is continuously maintained.

[0086] When the performance of the controller 32 is evaluated, the embodiment of the present invention takes the evaluation algorithm based on data statistical indicators as an example to introduce a specific implementation process of the performance evaluation of the controller 32. The basic idea of ​​the evaluation algorithm based on data statistical indicators is to compare the changes in the operating performance of the controller 32 in the evaluation data time period relative to the benchmark data time period by calculating the benchmark operating data in the benchmark data time period and the statistical indicators related to the operating data to be evaluated in the evaluation data time period, and then point out the performance improvement space of the controller 32 under various changes, wherein the statistical indicators include but are not limited to the control error standard deviation, online rate, covariance, saturation rate, in-range rate, number of changes, minimum variance performance indicators, etc. Taking the evaluation and monitoring method based on covariance as an example, when the performance evaluation module 13 performs performance evaluation on the operating performance of the controller 32 according to the control performance change threshold, the operating data to be evaluated and the benchmark operating data, the main implementation steps are as follows:

[0087] 11) Select the output data with better performance in the benchmark data time period as the benchmark data, obtain the operation data in the monitoring phase (i.e., the evaluation data time period), and the user sets the control performance change threshold in the configuration parameters. The control performance change threshold includes a performance deterioration threshold and a performance improvement threshold;

[0088] 12) Calculate the covariance matrix of the output variables of the industrial process object 33 in the industrial field control subsystem 30 in the reference data time period and the covariance matrix of the evaluation data time period, perform generalized eigenvalue decomposition on the covariance matrix of the reference data time period and the covariance matrix of the evaluation data time period, obtain the matrix eigenvalues ​​of each covariance matrix, and calculate the overall performance index of the monitoring stage;

[0089] 13) Perform generalized eigenvalue statistical inference to obtain the confidence interval of the matrix eigenvalues ​​of each covariance matrix in the corresponding characteristic direction, compare it with the performance deterioration threshold and performance improvement threshold set by the user, and judge the change of performance in the corresponding characteristic direction;

[0090] 14) By vector projection, determine the main output variables that cause the characteristic direction of deterioration or improvement of the performance of the controller 32.

[0091] The mismatch detection module 12 is used to determine whether the industrial process object 33 has a model mismatch according to the configuration parameters and the operation data when the performance evaluation result is that the operation performance of the controller 32 is abnormal, and obtain the model mismatch detection result. In the embodiment of the present invention, when the mismatch detection module 12 determines whether the industrial process object 33 has a model mismatch according to the configuration parameters and the operation data, the model mismatch detection algorithm used includes but is not limited to the correlation analysis method, the system identification method, etc. The configuration parameters used by the mismatch detection module 12 include the pre-set model mismatch threshold and the mismatch detection data time period when the model mismatch detection algorithm detects the model mismatch. When determining whether the industrial process object 33 has a model mismatch according to the configuration parameters and the operation data, the collected operation data within the time period specified by the mismatch detection data time period can be selected, and the model mismatch of the industrial process object 33 can be determined according to the model mismatch threshold and the selected operation data. For example, the data within the period from 12:00 of a certain year / month / day to 15:00 of a certain year / month / day is selected for mismatch detection. The configuration parameters also include a mismatch detection cycle, which is used to specify a time period for determining whether a model mismatch occurs in the industrial process object 33 and to determine how often the mismatch detection process is run, thereby automatically determining the operating performance of the controller 32 based on data within a specific time period, and achieving regular automatic monitoring and diagnosis of the operating performance of the controller 32.

[0092] The control optimization module 14 is used to determine the type of influencing factors affecting the operating performance of the controller 32 according to the model mismatch detection results, and determine the operation and maintenance strategy for optimizing the operating performance of the controller 32 based on the type of influencing factors, so as to optimize the operating performance of the controller 32 according to different influencing factors to improve the performance of the control loop.

[0093] The controller operation and maintenance device 10 also includes a model updating module 15 and a first data writing module 16, wherein the control optimization module 14 determines the type of influencing factors affecting the operation performance of the controller 32 according to the model mismatch detection result, including: when the model mismatch detection result is that the industrial process object 33 has a model mismatch, the type of influencing factors is determined to be that the real characteristics currently possessed by the industrial process object 33 are different from the initial characteristics, that is, the initial characteristics such as the structure and parameters of the industrial process object 33 itself have changed significantly. Then the control optimization module 14 determines the operation and maintenance strategy for optimizing the operation performance of the controller 32 based on the type of influencing factors, including: the control optimization module 14 obtains the recommended control model determined by the model updating module 15 according to the operation data, and the recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object 33. When determining the recommended control model that conforms to the real characteristics of the industrial process object 33 according to the operation data, the model online identification algorithm used includes but is not limited to the least squares method, the prediction error algorithm, the subspace method, etc. Then, the control optimization module 14 determines the first recommended controller parameters that match the design indicators of the controller 32 based on the recommended control model, and sends the first recommended controller parameters to the first data writing module 16. The first data writing module 16 feeds back the first optimization result consisting of the first recommended controller parameters and the recommended control model to the controller 32, so as to use the first optimization result to optimize the operating performance of the controller 32, thereby ensuring that the controller 32 continues to maintain the operating performance at the initial stage of operation when the structure, parameters and other characteristics of the industrial process object 33 change.

[0094] Among them, the control optimization module 14 determines the type of influencing factors that affect the operating performance of the controller 32 according to the model mismatch detection result, including: when the model mismatch detection result is that the industrial process object 33 does not have a model mismatch, the type of influencing factor is determined to be that the external environment in which the industrial process object 33 is currently operating is different from the initial operating environment. For example, there is an external disturbance in the operating environment of the industrial process object 33. Then the control optimization module 14 determines the operation and maintenance strategy for optimizing the operating performance of the controller 32 based on the type of influencing factors, including: determining the second recommended controller parameters that match the design indicators of the controller 32 according to the initial model, and sending the second recommended controller parameters to the first data writing module 16, and the first data writing module 16 feeds back the second optimization result composed of the second recommended controller parameters to the controller 32, so as to use the second optimization result to optimize the operating performance of the controller 32, so as to ensure that the controller 32 continues to maintain the operating performance at the initial stage of commissioning when the operating environment of the industrial process object 33 changes externally.

[0095] In the third embodiment of the present invention, a controller 32 operation and maintenance system is also provided, which includes the controller operation and maintenance device 10, the data acquisition and storage subsystem 20 and the industrial field control subsystem 30 provided in the second embodiment, and the industrial field control subsystem 30 includes the controller 32 to be operated and maintained and the industrial process object 33 controlled by the controller 32. Among them, the controller operation and maintenance device 10 is used to operate and maintain the controller 32. The data acquisition and storage subsystem 20 is used to collect and store the operation data generated by the industrial field control subsystem and the data generated when the controller operation and maintenance device 10 operates and maintains the controller 32, that is, the controller operation and maintenance device 10 exchanges data with the industrial field control subsystem 30 through the data acquisition and storage subsystem 20.

[0096] Specifically, the controller operation and maintenance device 10 in the embodiment of the present invention is a cloud-based intelligent operation and maintenance subsystem combined with a cloud computing platform, and has potential large-scale application and replication of technical achievements in industries such as cement, solid waste, and steel. In the embodiment of the present invention, the data acquisition and storage subsystem 20 is used to collect and store the operation data generated by the industrial field control subsystem 30 and the optimization result data generated by the cloud-based intelligent operation and maintenance subsystem. The cloud-based intelligent operation and maintenance subsystem evaluates, detects, and optimizes the control performance of the running industrial controller 32 based on the collected field operation data, and outputs the optimization result data. The industrial field control subsystem 30 adjusts the parameters of the controller 32 based on the optimization result data sent down and continues to operate. The embodiment of the present invention is based on the cloud platform and the configured data acquisition and storage subsystem 20 for scheduling, calculation, and storage, so that the entire operation and maintenance system is not limited by specific physical storage and computing devices, and has the advantages of cost saving, instant access, uninterrupted, and data security.

[0097] Combination Figure 4 As can be seen from the operation and maintenance principle block diagram of the operation and maintenance system shown, the controller operation and maintenance device 10 as a cloud-based intelligent operation and maintenance subsystem includes a performance evaluation module 13, a mismatch detection module 12, a model update module 15, a control optimization module 14, a first data reading module 11, a first data writing module 16 and a monitoring interface module 17.

[0098] Among them, the performance evaluation module 13 is configured with a first input interface, a controller performance evaluation algorithm and a first output interface. The first input interface receives the operating data of the industrial field control subsystem 30 collected by the first data reading module 11, and the configuration parameters required for operation and maintenance configured by the user in the monitoring interface module 17, wherein the operating data includes the control instruction data of the controller 32, the status data and the nominal model of the industrial process object 33. Figure 4The process control model in is the nominal model or the recommended control model recommended by the model update module 15 during the last operation and maintenance. The first output interface outputs the performance evaluation result generated after the controller performance evaluation algorithm evaluates the operating performance of the controller 32. The configuration parameters are mainly the parameters required for the performance evaluation module 13 to perform performance evaluation and the mismatch detection module 12 to perform mismatch detection. The configuration parameters include parameters such as performance degradation threshold, performance improvement threshold, model mismatch threshold, data time period for performance evaluation, data time period for mismatch detection, and algorithm operation cycle. The configuration parameters are input by the user from the monitoring interface module 17. Since the performance evaluation / mismatch detection is based on the collected input and output data, it is necessary to select a segment of data for evaluation / mismatch detection, wherein the data time period for performance evaluation or mismatch detection is: for example, selecting the data from 12:00 on a certain year / month / day to 15:00 on a certain year / month / day for performance evaluation or mismatch detection. The operation cycle of the algorithm is used to specify how often the performance evaluation module 13 runs the evaluation algorithm and how often the mismatch detection module 12 runs the mismatch detection algorithm, such as once a day or a week. That is, the data time period is used to determine the operation frequency of the performance evaluation algorithm and the model mismatch algorithm in the operation and maintenance system.

[0099] The mismatch detection module 12 is configured with a second input interface, a model mismatch detection algorithm and a second output interface. The second input interface receives the configuration parameters and operation data collected by the first data reading module 11, and the second output interface outputs the mismatch detection result generated after the model mismatch detection algorithm performs model mismatch detection on the industrial process object 33.

[0100] The model updating module 15 is configured with a third input interface, an online model identification algorithm and a third output interface. The third input interface receives the operating data collected by the first data reading module 11 and the mismatch detection result output by the second output interface of the mismatch detection module 12. The third output interface outputs the recommended control model to the control optimization module 14 and the first data writing module 16.

[0101] The control optimization module 14 is configured with a fourth input interface, a control parameter recommendation algorithm and a fourth output interface. The fourth input interface receives the recommended control model, and the fourth output interface outputs the recommended controller parameters that match the design indicators of the controller 32 determined by the control optimization module 14 using the control parameter recommendation algorithm according to the recommended control model.

[0102] The monitoring interface module 17 is configured with a fifth input interface, result display, trend analysis, parameter editing and a fifth output interface, etc. The fifth input interface receives operation data, evaluation results, mismatch detection results and optimization results, and the fifth output interface sends the configuration parameters input by the user to the first data reading module 11.

[0103] The data acquisition and storage subsystem 20 includes a data cloud storage module 22, a second data reading module 23, and a second data writing module 21. The second data reading module 23 collects and obtains operation data from the industrial field control subsystem 30, and obtains recommended controller parameter data from the cloud intelligent operation and maintenance subsystem. The data cloud storage module 22 is responsible for storing the operation data and the recommended controller parameter data. The second data writing module 21 writes the operation data and the recommended controller parameter data to the cloud intelligent operation and maintenance subsystem and the industrial field control subsystem 30.

[0104] The industrial field control subsystem 30 includes a third data reading module 31, a set value input module 35, a controller 32, an industrial process object 33 and a third data writing module 34. The controller 32 is configured with a sixth input interface, a control algorithm and a sixth output interface. The sixth input interface is the state data, command set value and optimization result of the recommended controller parameters of the industrial process object 33, and the sixth output interface outputs the control command data obtained by the control algorithm operation to the industrial process object 33. The industrial process object 33 module is configured with a seventh input interface, an industrial process object body and a seventh output interface. The seventh input interface receives the control command data issued by the controller 32, and the seventh output interface outputs the state data.

[0105] The main implementation process of the operation and maintenance method based on the cloud intelligent operation and maintenance subsystem in the embodiment of the present invention is as follows:

[0106] Combination Figure 4 and Figure 5 It can be seen that the operation and maintenance method includes four main steps, namely control performance evaluation, model mismatch detection, model online identification and control parameter optimization. Among them:

[0107] 1) During the control performance evaluation, based on the operation data of the industrial field control subsystem obtained by the data acquisition and storage subsystem and the configuration parameters input by the user, the operation performance evaluation module 13 uses the performance evaluation algorithm to obtain the operation status evaluation result of the controller 32, and determines whether the operation performance of the controller 32 is abnormal based on the evaluation result. Figure 5 It can be seen that if no abnormality occurs, no maintenance is performed and the industrial field control subsystem 30 continues to run. If an abnormality occurs, the model mismatch detection process is entered and steps 2) to 4) are executed. Among them, the performance evaluation algorithm includes but is not limited to an evaluation algorithm based on minimum variance control, linear quadratic Gaussian and other benchmarks, an evaluation algorithm based on abnormality detection and an evaluation algorithm based on historical data statistical indicators.

[0108] The configuration parameters collected by the performance evaluation module 13 during performance evaluation include a preset performance change threshold, a reference data time period, and an evaluation data time period. The performance evaluation module 13 performs performance evaluation on the operating performance of the controller 32 according to the configuration parameters and the operating data and obtains the performance evaluation result, specifically including:

[0109] The benchmark operation data collected in the benchmark data time period is obtained, wherein the benchmark operation data is the operation data collected by the cloud intelligent operation and maintenance subsystem when the operation performance of the controller 32 is not abnormal. The operation data collected in the evaluation data time period is used as the operation data to be evaluated. According to the control performance change threshold, the operation data to be evaluated and the benchmark operation data, the operation performance of the controller 32 is evaluated and the performance evaluation result is obtained. The benchmark operation data collected in the benchmark data time period is the data when the operation performance of the controller 32 is good. The operation data to be evaluated collected in the evaluation data time period, such as the operation data from 11:00 in a certain year / month / day to 14:00 in a certain year / month / day, is selected for performance evaluation. The configuration parameters also include a performance evaluation cycle, which is used to specify the time period for performance evaluation of the operation performance of the controller 32, thereby determining how often to run the performance evaluation algorithm to perform performance evaluation on the operation performance of the controller 32, such as running the performance evaluation algorithm once a day or a week to perform performance evaluation on the operation performance of the controller 32. The operation and maintenance objectives of the controller 32 can be achieved under the operation and maintenance modes such as cost saving, instant access, and uninterrupted operation and maintenance, so as to continuously maintain the stability of industrial production quality.

[0110] When the performance of the controller 32 is evaluated in the embodiment of the present invention, an evaluation algorithm based on data statistical indicators is used as an example to introduce a specific implementation process of the performance evaluation of the controller 32. The basic idea of ​​the evaluation algorithm based on data statistical indicators is to calculate the statistical indicators related to the benchmark operating data of the benchmark data time period and the operating data to be evaluated in the evaluation data time period, and compare the changes in the operating performance of the controller 32 in the evaluation data time period relative to the benchmark data time period, and the changes include deterioration, basically unchanged or improved, and then point out the performance improvement space of the industrial field control subsystem 30 under various changes, wherein the statistical indicators include but are not limited to the control error standard deviation, online rate, covariance, saturation rate, in-range rate, number of changes, minimum variance performance indicators, etc. Taking the evaluation and monitoring method based on covariance as an example, the main implementation steps are as follows:

[0111] 11) Select the output data with better performance in the benchmark data time period as the benchmark data, obtain the operation data in the monitoring phase (i.e., the evaluation data time period), and the user sets the control performance change threshold in the configuration parameters. The control performance change threshold includes a performance deterioration threshold and a performance improvement threshold;

[0112] 12) Calculate the covariance matrix of the output variables of the industrial process object 33 in the industrial field control subsystem 30 in the reference data time period and the covariance matrix of the evaluation data time period, perform generalized eigenvalue decomposition on the covariance matrix of the reference data time period and the covariance matrix of the evaluation data time period, obtain the matrix eigenvalues ​​of each covariance matrix, and calculate the overall performance index of the monitoring stage;

[0113] 13) Perform generalized eigenvalue statistical inference to obtain the confidence interval of the matrix eigenvalues ​​of each covariance matrix in the corresponding characteristic direction, compare it with the performance deterioration threshold and performance improvement threshold set by the user, and judge the change of performance in the corresponding characteristic direction;

[0114] 14) By vector projection, determine the main output variables that cause the characteristic direction of deterioration or improvement of the performance of the controller 32.

[0115] When evaluating the performance of the controller 32, the embodiment of the present invention takes the evaluation algorithm based on the minimum variance control benchmark as an example to introduce the specific implementation process of the performance evaluation of the controller 32. The basic idea of ​​the evaluation algorithm of the minimum variance control benchmark is to find the ideal minimum variance of the process output of the current control system by analyzing the process history data and the disturbance model, and use this as the evaluation benchmark to point out the performance improvement space of the control system under various working conditions. The main steps are as follows: based on the historical closed-loop output data, build a time series model, perform step response expansion, solve the minimum variance control output item, based on the real-time closed-loop output data, solve the control error variance, calculate the minimum variance performance index, and when the minimum variance performance index is significantly less than 1, it means that the performance of the controller 32 may be abnormal and there is room for improvement.

[0116] 2) During model mismatch detection, in one round of mismatch detection, according to the performance evaluation results, the original or user-input nominal model of the industrial process object 33 when the controller is put into operation, the configuration parameters input by the user and the operating data of the industrial field control subsystem obtained by the data acquisition and storage subsystem 20. The operating model mismatch detection module 12 adopts the model mismatch detection algorithm to obtain the accuracy test result of the nominal model, and evaluates whether the original model has a model mismatch based on the test result. Among them, the model mismatch detection algorithm includes but is not limited to the correlation analysis method, the system identification method, etc. Taking the system identification method as an example, the core principle is to use the collected operating data, configuration parameters, and evaluation results, combined with the nominal model to obtain the prediction error data of the nominal model, and then combine the input data and the prediction error data to perform system identification to construct the error model, and finally compare the error model with the nominal model to obtain the size of the mismatch to evaluate whether the nominal model is mismatched. The specific steps are:

[0117] 21) combining the collected input and output data with the nominal model of the industrial process object 33 determined during the controller design, and constructing prediction error data through model simulation;

[0118] 22) Combining the input and prediction error data, construct an extended data matrix of the error model at past and future times;

[0119] 23) By expanding the data matrix, the expansion coefficient matrix of the error model is obtained;

[0120] 24) By expanding the coefficient matrix, the system matrix of the error model is solved;

[0121] 25) The system matrix of the error model is compared with the nominal model to obtain the mismatch index that measures the difference between the models;

[0122] 26) According to the mismatch index and the mismatch threshold set by the configuration parameters, it is determined whether the nominal model is mismatched. If the mismatch index is greater than the mismatch threshold, it is considered that the model is mismatched.

[0123] 3) In the case of model mismatch, the model is identified online through the model update module 15. According to the results of the model mismatch detection and the operation data of the industrial field control system obtained through data acquisition and storage, the model update module 15 is operated to adopt the model online identification algorithm to obtain a recommended control model that is closer to the current real characteristics of the industrial process object 33. The recommended control models include finite step response models, transfer function models, and state space equation models. Among them, the model online identification algorithm includes but is not limited to the least squares method, the prediction error algorithm, the subspace method, etc. Taking the subspace method as an example, the core principle is to use the collected input and output data to perform parameter identification to obtain an estimated model of the characteristics of the industrial process object 33. The specific steps are:

[0124] 31) Combining input and output data, constructing the extended data matrix of the past and future time, the input and output data are the operation data and the mismatch detection result respectively;

[0125] 32) By expanding the data matrix, the expansion coefficient matrix is ​​obtained;

[0126] 33) By expanding the coefficient matrix, the system matrix of the identification model is solved;

[0127] 34) Based on the combined forgetting factor, the nominal model and the identification model are weighted to obtain a recommended control model that is closer to the current true characteristics of the industrial process object 33.

[0128] 4) The control optimization module 14 has two working modes when optimizing control parameters:

[0129] 41) If the model mismatch detection module 12 detects that the model of the industrial process object 33 is mismatched, it means that the structure, parameters and other characteristics of the industrial process object 33 itself have changed significantly. Then, according to the recommended control model identified and output by the model update module 15, the operation control optimization module 14 adopts the control parameter optimization algorithm to obtain the recommended controller parameters that match the design indicators of the controller 32. The recommended controller parameters include the proportional coefficient Kp, integral coefficient Ti, differential coefficient Td of the PID controller, the feedback gain of the linear quadratic controller LQR, the smoothing factor, equivalent deviation, closed-loop reference time, model feedback correction parameters, etc. of the model predictive controller MPC.

[0130] 42) If the model detection shows no mismatch, then the external operating environment of the industrial process object 33 has changed significantly. Then, based on the results of the nominal model and performance evaluation, the operation control optimization module 14 uses a control parameter optimization algorithm to obtain recommended controller parameters that match the design indicators of the controller 32.

[0131] The control parameter optimization algorithm adopted by the control optimization module 14 includes but is not limited to the critical proportionality method, the internal model control method, etc. Taking the critical proportionality method as an example, the core principle is to use the collected input and output data to perform parameter identification to obtain an estimated model of the characteristics of the industrial process object 33. The specific steps are:

[0132] 121) Based on the recommended control model obtained by identification, a closed-loop control loop is constructed to determine the critical stability limit of the loop;

[0133] 122) According to the setting formula and the design index of the controller 32, the recommended controller parameters of the controller 32 are calculated, and the nominal model is a mathematical model of the original characteristics of the reaction industrial process object 33 determined when the controller 32 is put into operation.

[0134] The specific steps of this operation and maintenance method are as follows:

[0135] Step S1 , obtaining operation data generated by the operation of the industrial field control subsystem 30 from the data acquisition and storage subsystem 20 through the first data reading module 11 .

[0136] Step S2, based on the read-in operating data and the configuration parameters input by the user, periodically schedule or single-schedule the operation performance evaluation module 13, and evaluate the various performance indicators of the controller 32 in the industrial field control subsystem 30 through the performance evaluation algorithm running in the performance evaluation module 13 to obtain the evaluation result of the performance of the controller 32.

[0137] Step S3, determining whether the performance of the controller 32 is abnormal, if the performance of the controller 32 is abnormal, executing step S4, otherwise executing step S9.

[0138] Step S4, based on the read-in operating data and the configuration parameters input by the user, the mismatch detection module 12 is run periodically or once, and the model mismatch detection algorithm running in the mismatch detection module 12 is used to detect the changes in the model characteristics of the industrial process object 33 in the industrial field control subsystem 30, and obtain the detection results of the model on which the controller 32 parameters depend.

[0139] Step S5, determining whether the model characteristics of the industrial process object 33 are mismatched, if the model characteristics of the industrial process object 33 are mismatched, executing step S6, otherwise executing step S9.

[0140] Step S6, based on the read-in operating data and the result of the model mismatch detection, trigger the operation of the model updating module 15, and obtain a recommended control model that is closer to the latest characteristics of the industrial process object 33 through the model online identification algorithm running in the model updating module 15.

[0141] Step S7, based on the recommended control model obtained by online model identification, the control optimization module 14 is run, and the recommended controller parameters of the controller 32 are obtained by combining the parameter recommendation algorithm running in the control optimization module 14 with the various performance indicators designed for the controller 32.

[0142] In step S8, the recommended control model identified by the online model and the recommended controller parameters optimized by the control parameters are used as optimization results, which are transmitted to the data acquisition and storage subsystem 20 through the first data writing module 16, and then sent to the industrial field control subsystem 30 by the second data writing module 21 of the data acquisition and storage subsystem 20. The industrial field control subsystem 30 reads the optimization results through the third data reading module 31, and uses the read optimization results to update the model and parameter configuration of the controller 32.

[0143] Step S9, continuously operate the industrial field control subsystem 30, and regularly schedule the corresponding configuration modules in the cloud-based intelligent operation and maintenance subsystem to evaluate, detect and optimize according to the configuration parameters, so as to ensure that the controller 32 continues to maintain the operating performance at the initial stage of operation.

[0144] All core modules, algorithms and data included in the controller 32 operation and maintenance system proposed in the embodiment of the present invention are stored, scheduled and calculated in the cloud, which is not limited by specific physical storage and computing devices, and has the advantages of cost saving, instant access, uninterrupted, data security, etc. The controller 32 operation and maintenance method consists of control performance evaluation, model mismatch detection, model online identification and control parameter optimization, forming a complete and fully autonomous closed-loop operation and maintenance process.

[0145] The intelligent operation and maintenance method and operation and maintenance system based on cloud computing technology that are automatically operated in the embodiments of the present invention can form a complete fully autonomous closed-loop operation and maintenance process in the cloud. The operation and maintenance system includes a cloud operation and maintenance subsystem, a data acquisition and storage subsystem 20, and an industrial field control subsystem 30. Compared with the traditional operation and maintenance method, multiple industrial field controllers 32 can be connected to the cloud to achieve large-scale automatic operation and maintenance and monitoring scheduling at the group level, which can expand the scale of operation and maintenance and monitoring, effectively reduce the threshold of operation and maintenance, and improve the level of automated management of industrial production.

[0146] Therefore, the operation and maintenance system provided by the embodiment of the present invention can avoid the tedious and inefficient problems existing when a large number of people participate in the operation and maintenance of the controller 32, and can reduce the work pressure of personnel and improve the quality of operation and maintenance when operating and maintaining the controller 32. The performance of the control loop of the controller 32 is comprehensively improved from the parameters such as the initial model of the industrial process object 33 and the control instruction data of the controller 32, thereby improving the adaptability of the controller 32 to the complex and changeable working conditions of the industrial site, improving the control effect of the controller 32 on the industrial process object 33, and reducing the operation and maintenance cost of the controller 32. Real-time and targeted operation and maintenance of the control loop of the complex industrial process object 33 is achieved, and the industrial production quality of the industrial process object 33 is kept stable.

[0147] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A controller operation and maintenance method, characterized in that: The controller (32) controls an industrial process object (33) in an industrial field control subsystem (30), and the operation and maintenance method comprises: Collecting configuration parameters and operating data of the industrial field control subsystem (30), wherein the operating data includes control instruction data of the controller (32) and state data and an initial model of the industrial process object (33), wherein the configuration parameters are parameters required for operating and maintaining the operating performance of the controller (32), and the initial model is a mathematical model that conforms to the initial characteristics of the industrial process object (33); According to the configuration parameters and the operating data, a performance evaluation is performed on the operating performance of the controller (32) and a performance evaluation result is obtained; When the performance evaluation result indicates that the operating performance of the controller (32) is abnormal, determining whether a model mismatch occurs in the industrial process object (33) based on the configuration parameters and the operating data, and obtaining a model mismatch detection result; Determining the type of influencing factors affecting the operating performance of the controller (32) according to the model mismatch detection result, and determining an operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors; The step of determining the type of influencing factor affecting the operating performance of the controller (32) according to the model mismatch detection result comprises: When the model mismatch detection result is that the industrial process object (33) has a model mismatch, determining the type of influencing factor is that the real characteristics currently possessed by the industrial process object (33) are different from the initial characteristics; Then, the step of determining an operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors comprises: Acquire a recommended control model determined according to the operating data, wherein the recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object (33); A first recommended controller parameter matching the design index of the controller (32) is determined according to the recommended control model, so as to optimize the operating performance of the controller (32) using a first optimization result composed of the first recommended controller parameter and the recommended control model.

2. The method according to claim 1, characterized in that The step of determining the type of influencing factor affecting the operating performance of the controller (32) according to the model mismatch detection result further includes: When the model mismatch detection result is that the industrial process object (33) has no model mismatch, determining the type of influencing factor is that the external environment in which the industrial process object (33) is currently operating is different from the initial operating environment; Then, the step of determining an operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors comprises: Second recommended controller parameters matching the design index of the controller (32) are determined according to the initial model, so as to optimize the operating performance of the controller (32) using a second optimization result formed by the second recommended controller parameters.

3. The method according to claim 1, characterized in that The configuration parameters include a preset control performance change threshold, a reference data time period, and an evaluation data time period, wherein the steps of evaluating the operating performance of the controller (32) and obtaining a performance evaluation result based on the configuration parameters and the operating data include: Acquiring reference operating data collected within the reference data time period, wherein the reference operating data is the operating data collected when the operating performance of the controller (32) is normal; Using the operating data collected within the evaluation data time period as the operating data to be evaluated; According to the control performance change threshold, the to-be-evaluated operating data and the reference operating data, the operating performance of the controller (32) is evaluated and a performance evaluation result is obtained.

4. The method according to claim 3, characterized in that The configuration parameters also include a preset model mismatch threshold and a mismatch detection data time period. The step of determining whether a model mismatch occurs in the industrial process object (33) based on the configuration parameters and the operating data includes: The operating data collected during the mismatch detection data time period is acquired, and whether a model mismatch occurs in the industrial process object (33) is determined based on the model mismatch threshold and the operating data.

5. The method according to claim 1, characterized in that: The configuration parameters also include a performance evaluation period, which is used to specify a time period for performing a performance evaluation on the operating performance of the controller (32).

6. The method according to claim 5, characterized in that The configuration parameters also include a mismatch detection period, which is used to specify a time period for determining whether a model mismatch occurs in the industrial process object (33).

7. A controller operation and maintenance device, characterized in that: The controller operation and maintenance device (10) is used to operate and maintain a controller (32) that controls an industrial process object (33) in an industrial field control subsystem (30), and the controller operation and maintenance device (10) comprises: A first data reading module (11), a performance evaluation module (13), a mismatch detection module (12) and a control optimization module (14); The first data reading module (11) is used to collect configuration parameters and operating data of the industrial field control subsystem (30), wherein the operating data includes control instruction data of the controller (32) and state data and an initial model of the industrial process object (33), wherein the configuration parameters are parameters required for operating and maintaining the operating performance of the controller (32), and the initial model is a mathematical model that conforms to the initial characteristics of the industrial process object (33); The performance evaluation module (13) is used to perform a performance evaluation on the operating performance of the controller (32) according to the configuration parameters and the operating data and obtain a performance evaluation result; The mismatch detection module (12) is used to, when the performance evaluation result indicates that the operating performance of the controller (32) is abnormal, determine whether a model mismatch occurs in the industrial process object (33) based on the configuration parameters and the operating data, and obtain a model mismatch detection result; The control optimization module (14) is used to determine the type of influencing factors affecting the operating performance of the controller (32) according to the model mismatch detection result, and determine the operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors; The system further comprises a model updating module (15) and a first data writing module (16), wherein the step of determining the type of influencing factor affecting the operating performance of the controller (32) according to the model mismatch detection result by the control optimization module (14) comprises: When the model mismatch detection result is that the industrial process object (33) has a model mismatch, determining the type of influencing factor is that the real characteristics currently possessed by the industrial process object (33) are different from the initial characteristics; Then, the step of the control optimization module (14) determining the operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors comprises: Obtaining a recommended control model determined by the model updating module according to the operating data, wherein the recommended control model is a mathematical model that conforms to the real characteristics of the industrial process object (33); A first recommended controller parameter matching the design index of the controller (32) is determined based on the recommended control model, and the first recommended controller parameter is sent to the first data writing module (16). The first data writing module (16) feeds back a first optimization result consisting of the first recommended controller parameter and the recommended control model to the controller (32), so as to optimize the operating performance of the controller (32) using the first optimization result.

8. The controller operation and maintenance device according to claim 7, characterized in that: The system further comprises a first data writing module (16), wherein the control optimization module (14) determines the type of influencing factor affecting the operating performance of the controller (32) according to the model mismatch detection result, comprising: When the model mismatch detection result is that the industrial process object (33) has no model mismatch, determining the type of influencing factor is that the external environment in which the industrial process object (33) is currently operating is different from the initial operating environment; Then, the step of the control optimization module (14) determining the operation and maintenance strategy for optimizing the operating performance of the controller (32) based on the type of influencing factors comprises: Determine second recommended controller parameters that match the design indicators of the controller (32) based on the initial model, and send the second recommended controller parameters to the first data writing module, and the first data writing module (16) feeds back a second optimization result consisting of the second recommended controller parameters to the controller (32) to optimize the operating performance of the controller (32) using the second optimization result.

9. A controller operation and maintenance system, characterized in that: include: The controller operation and maintenance device (10) as claimed in claim 7 or 8; An industrial field control subsystem (30), the industrial field control subsystem (30) comprising a controller (32) to be operated and maintained and an industrial process object (33) controlled by the controller (32); The controller operation and maintenance device (10) is used to operate and maintain the controller (32).

10. The system according to claim 9, characterized in that It also includes a data acquisition and storage subsystem (20), which is used to acquire and store operating data generated by the industrial field control subsystem (30) and data generated when the controller operation and maintenance device (10) operates and maintains the controller (32).

Images