DCS control system optimization method based on intelligent algorithm

Abstract

The invention relates to the technical field of DCS control systems, in particular to a DCS control system optimization method based on an intelligent algorithm. Historical input and output variables of the DCS control system are preprocessed, then a DCS control system model is established, to-be-optimized parameters of the DCS control system are selected, a fitness function is determined, the to-be-optimized parameters form a parameter combination randomly, the parameter combination serves as an initial value, an initial search range is set, and the positions of search individuals are initialized randomly. Then expanding, developing and searching, and finally performing simulation verification. The optimization method provided by the invention can help to find the optimal control strategy parameter, obtain a better solution in a relatively short time, improve the efficiency of the optimization process and reduce the calculation time, and meanwhile, simulation is carried out by utilizing the constructed DCS control system model to verify the optimization effect of the DCS control system parameter, so that the optimization efficiency of the DCS control system is improved. Therefore, the obtained parameter combination can satisfy the parameter optimization purpose of the DCS control system.

Description

Technical Field

[0001] The invention relates to the technical field of DCS control systems, and in particular to a DCS control system optimization method based on an intelligent algorithm. Background Art

[0002] DCS control system, namely distributed control system, distributes control functions to multiple control units and realizes centralized management and monitoring through network connection. It adopts a hierarchical structure, usually including field control layer, monitoring layer and management layer. Each layer has specific functions and tasks, working together to achieve efficient control of the production process. With the continuous development of industrial automation, distributed control systems (DCS) have been widely used in various fields.

[0003] In the patent document with application number "CN201410269302.9" and name "A control system parameter optimization method based on orthogonal optimization particle swarm algorithm", it is recorded that "the control system parameter optimization method combines the orthogonal optimization method with the particle swarm optimization method, and determines the optimization range of each controller parameter, the sensitivity relationship between controller parameters and the weight relationship between various performance indicators through orthogonal experiments. On this basis, the particle swarm algorithm is applied to optimize, iterate cyclically, and obtain the optimization result. The method of the present invention uses the results of orthogonal optimization to guide the particle swarm optimization, with fast optimization speed, clear optimization direction and significant optimization effect."

[0004] For example, in the patent document with application number "CN201910024660.6" and name "A method for optimizing given values ​​in a campus electric heating DCS control system", it is recorded that "it includes the following steps: thermal load classification, thermal load classification coding, establishment of a prediction model, soft start control and three-phase balance control. The structure of the present invention is scientific and reasonable, safe and convenient to use. With the help of the idea of ​​power load classification, according to the importance of the thermal load and the actual temperature requirements, the thermal load is classified and different levels of thermal loads are assigned different temperature set values ​​in different time zones to avoid low-level loads consuming the same amount of electricity as high-level loads, thereby achieving power saving benefits. With the help of thermal load classification, thermal load coding strategy and feedback control ideas, the Ping--Pangk control method based on the minimum three-phase imbalance is adopted to ensure that the three-phase imbalance is minimized, thereby ensuring the power supply quality of the power grid and avoiding damage to transformers and power loads due to overvoltage caused by three-phase imbalance."

[0005] Although the above patents can achieve the optimization of parameters or given values, they may have problems such as low control accuracy, slow response speed, and poor adaptability when facing complex industrial processes. Based on this, the present invention provides a DCS control system optimization method based on an intelligent algorithm to solve the above technical problems. Summary of the invention

[0006] The purpose of the present invention is to provide a DCS control system optimization method based on an intelligent algorithm, which can help find the best control strategy parameters, obtain a better solution in a relatively short time, improve the efficiency of the optimization process, and reduce the calculation time. At the same time, the constructed DCS control system model is used for simulation to verify the optimization effect of the DCS control system parameters to ensure that the obtained parameter combination can meet the purpose of optimizing the DCS control system parameters.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The present invention provides a DCS control system optimization method based on an intelligent algorithm, comprising the following steps:

[0009] S1. Analyze the DCS control system structure, obtain the historical input and output variables and corresponding parameters in the DCS control system, pre-process the historical input and output variables, and then establish a DCS control system model;

[0010] S2, analyze the characteristics of the DCS control system, select the parameters to be optimized of the DCS control system, and determine the fitness function;

[0011] S3, randomly group the parameters to be optimized into parameter combinations and use them as initial values, set the initial search range and randomly initialize the search individual positions, and then expand and develop the search;

[0012] S4. Use the DCS control system model for simulation;

[0013] S5. Compare the simulation results and complete the optimization of the DCS control system.

[0014] The present invention is further configured as follows: in the step S1, the preprocessing is to optimize the abnormal values ​​of the historical input and output variables.

[0015] The present invention is further configured as follows: the pre-processing process is as follows:

[0016] Calculate the mean and standard deviation of historical input and output variables respectively, and take the variable value exceeding the mean β times the standard deviation as an outlier;

[0017] Then the mean is used to replace the outliers to obtain new input and output variables.

[0018] The present invention is further configured as follows: in the step S1, the DCS control system model is established by training the DCS control system virtual model using preprocessed input and output variables and corresponding parameters to complete the establishment of the DCS control system model.

[0019] The present invention is further configured as follows: in step S2, the fitness function is: Where b1 is the control error, b2 is the response time, b3 is the system stability, a1, a2 and a3 are weight coefficients, and k1 and k2 are adjustment coefficients.

[0020] The present invention is further configured as follows: in step S3, the following steps are specifically included:

[0021] Set the initial search range and randomly initialize the search individual position, where the formula for randomly initializing the search individual position is:

[0022] In the formula, i is the search population size, j is the problem dimension, and f i j+1 =λf i j (1-f i j ), f i j is the variable, λ is the coefficient, UA j is the jth lower bound for the given problem, WA j is the jth upper bound of the given problem;

[0023] Expand your search: Where, Y1 e+1 is the search result obtained in the e+1th round using the search method Y1, brst is the best solution obtained in the previous e rounds of iterations, E is the maximum number of search rounds, rand is a random vector with a value between 0 and 1, is the average value of the solutions in the e-th round of iteration;

[0024] Narrow your search: In the formula, is the search result obtained in the e+1th round using the search method of Y2, is the flight function, and Δx is the adjustment range.

[0025] The present invention is further configured as follows: in step S4, the following steps are specifically included:

[0026] Using the DCS control system model, historical data is used as input to obtain the corresponding historical output information;

[0027] Then change the parameters of the DCS control system model, use the historical data as input, and obtain the corresponding change output information.

[0028] The present invention is further configured as follows: the parameters are parameter combinations corresponding to the current optimization results.

[0029] The present invention is further configured as follows: in step S5, if the simulated change output information is better than the historical output information, the optimization of the DCS control system is stopped, and the parameters used for the change are the optimized parameters of the DCS control system; otherwise, steps S1 to S4 are repeated.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention analyzes the structure of a DCS control system, obtains historical input and output variables and corresponding parameters in the DCS control system, pre-processes the historical input and output variables, and then establishes a DCS control system model; then analyzes the characteristics of the DCS control system, selects the parameters to be optimized of the DCS control system, and determines the fitness function; and randomly forms a parameter combination of the parameters to be optimized, and uses it as an initial value, sets an initial search range and randomly initializes the search individual position, and then expands and develops the search, and at the same time, uses the DCS control system model for simulation to complete the optimization of the DCS control system. The DCS control system optimization method based on an intelligent algorithm provided by the present invention can help find the best control strategy parameters, obtain a better solution in a relatively short time, improve the efficiency of the optimization process, and reduce the calculation time. At the same time, the constructed DCS control system model is used for simulation to verify the optimization effect of the DCS control system parameters, so as to ensure that the obtained parameter combination can meet the purpose of optimizing the parameters of the DCS control system, avoid abnormal parameter optimization application affecting the operation of the DCS control system, and reduce the optimization operation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The present invention is a flow chart of a DCS control system optimization method based on an intelligent algorithm. DETAILED DESCRIPTION

[0033] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Example:

[0035] This embodiment provides a DCS control system optimization method based on an intelligent algorithm, comprising the following steps:

[0036] S1. Analyze the DCS control system structure, obtain the historical input and output variables and corresponding parameters in the DCS control system, pre-process the historical input and output variables, and then establish the DCS control system model.

[0037] Among them, preprocessing is to optimize the outliers of historical input and output variables.

[0038] In this embodiment, it should be noted that, as an implementation method of this embodiment, the preprocessing process in this embodiment is: respectively calculating the mean and standard deviation of the historical input and output variables, taking the variable value exceeding the mean β times the standard deviation as an abnormal value; then using the mean to replace the abnormal value to obtain new input and output variables. By preprocessing the historical input and output variables, the reliability of the data source is guaranteed, so as to subsequently improve the reliability of the DCS control system model.

[0039] Furthermore, the DCS control system model is established by using the preprocessed input and output variables and corresponding parameters to train the DCS control system virtual model to complete the establishment of the DCS control system model.

[0040] In this embodiment, it should be noted that this embodiment uses historical input and output variables and corresponding parameters to train the DCS control system virtual model, constructs a simulation for subsequent optimization of parameter combinations, and verifies the optimization effect on the DCS control system parameters.

[0041] As an embodiment, taking the temperature control system in a chemical production process as an example, the temperature setting value (input variable) and the actual temperature measurement value (output variable) in the past period of time, as well as the corresponding timestamp, control parameters and other related data are obtained from the DCS control system.

[0042] Calculate the mean and standard deviation of the input and output variables, and set the variable values ​​that exceed 2 times the standard deviation of the mean as outliers. For example, if the mean of the input variable temperature setting value is 80℃ and the standard deviation is 5℃, then the values ​​that exceed the range of 70-90℃ are considered outliers. Replace the outliers with the mean to obtain a new input and output variable data set.

[0043] A DCS control system virtual model based on a neural network is trained using preprocessed input and output variables and corresponding parameters. The structure of the neural network can adopt a three-layer structure, the number of input layer nodes corresponds to the dimension of the input variable, the number of hidden layer nodes is set to 10 based on experience, and the number of output layer nodes corresponds to the dimension of the output variable.

[0044] S2. Analyze the characteristics of the DCS control system, select the parameters to be optimized of the DCS control system, and determine the fitness function.

[0045] Among them, the fitness function is: Where b1 is the control error, b2 is the response time, b3 is the system stability, a1, a2 and a3 are weight coefficients, and k1 and k2 are adjustment coefficients.

[0046] In this embodiment, it should be noted that this embodiment uses control error, response time and system stability as factors of the fitness function. As an implementation method, a1>a2>a3 in this embodiment, that is, the importance of control error is superior to response time. At the same time, the importance of response time is superior to system stability. The fitness function is used to measure the optimization performance of a specific control parameter combination in achieving the performance target of the DCS control system.

[0047] As an example of an embodiment of the temperature control system in the above-mentioned chemical production process, by analyzing the temperature control system, it can be determined that the parameters to be optimized are the proportional coefficient (P), integral coefficient (I) and differential coefficient (D), that is, the three key parameters of the controller, and a1, a2 and a3 are taken as 0.5, 0.4 and 0.3 respectively.

[0048] S3. Randomly combine the parameters to be optimized into parameter combinations and use them as initial values. Set the initial search range and randomly initialize the search individual positions, and then expand and develop the search.

[0049] Specifically, the following steps are included:

[0050] Set the initial search range and randomly initialize the search individual position, where the formula for randomly initializing the search individual position is:

[0051] In the formula, i is the search population size, j is the problem dimension, and f i j+1 =λf i j (1-f i j ), f i j is the variable, λ is the coefficient, UA j is the jth lower bound for the given problem, WA j is the jth upper bound of the given problem;

[0052] Expand your search: Where, Y1 e+1 is the search result obtained in the e+1th round using the search method Y1, brst is the best solution obtained in the previous e rounds of iterations, E is the maximum number of search rounds, rand is a random vector with a value between 0 and 1, is the average value of the solutions in the e-th round of iteration;

[0053] Narrow your search: In the formula, is the search result obtained in the e+1th round using the search method of Y2, is the flight function, and Δx is the adjustment range.

[0054] In this embodiment, it should be noted that it can help find the best control strategy parameters, obtain a better solution in a relatively short time, improve the efficiency of the optimization process, and reduce the calculation time.

[0055] As an example, the parameters to be optimized (P, I, D) can be randomly combined into a parameter combination and used as the initial value. The initial search range is set, for example, the range of P is [0.1, 10], the range of I is [0.01, 1], and the range of D is [0.01, 1].

[0056] S4. Use the DCS control system model for simulation.

[0057] Specifically, the following steps are included:

[0058] Using the DCS control system model, historical data is used as input to obtain the corresponding historical output information;

[0059] Then change the parameters of the DCS control system model, use the historical data as input, and obtain the corresponding change output information.

[0060] Furthermore, the parameter is a parameter combination corresponding to the current optimization result.

[0061] In this embodiment, it should be noted that the DCS control system model constructed in step S1 is used to simulate the original DCS control system parameters and the DCS control system optimization parameters respectively, and then the simulation results are obtained, and the obtained parameter combination is verified to ensure the reliability and rationality of the optimized parameter combination.

[0062] S5. Compare the simulation results and complete the optimization of the DCS control system.

[0063] If the simulated change output information is better than the historical output information, the optimization of the DCS control system is stopped, and the parameters used for the change are the optimized parameters of the DCS control system. Otherwise, steps S1 to S4 are repeated.

[0064] In this embodiment, it should be noted that the obtained parameter combination is verified by comparing the simulation results to ensure that the obtained parameter combination can meet the purpose of optimizing the parameters of the DCS control system.

[0065] As a test, before optimizing the parameters, the control error: when the set temperature is 80℃, the actual temperature fluctuates between 70℃ and 90℃, and the average control error is 3℃. Response time: the time required from the set temperature change to the actual temperature stabilizing within the range of ±2℃ of the set value is an average of 45s. System stability: the standard deviation of temperature fluctuation is 2.5℃.

[0066] After optimizing the parameters, the control error: when the set temperature is 80℃, the actual temperature fluctuates between 78℃ and 82℃, and the average control error is 1℃. Response time: the time required from the set temperature change to the actual temperature stabilizing within the range of ±2℃ of the set value is 25s on average. System stability: the standard deviation of temperature fluctuation is 1.5℃.

[0067] The performance indicators have been improved as follows:

[0068] Improvement of control error: The average control error was 3°C before optimization and 1°C after optimization, with an improvement ratio of (3-1) / 3≈66.7%.

[0069] Response time improvement: The average response time before optimization was 45s, and after optimization it was 25s, with an improvement ratio of (45-25) / 45≈44.4%.

[0070] Improved system stability: The standard deviation of temperature fluctuation was 2.5°C before optimization and 1.5°C after optimization, with an improvement ratio of (2.5-1.5) / 2.5≈40%.

[0071] After optimization by the DCS control system optimization method based on intelligent algorithm, the system control error, response time and system stability have been significantly improved, indicating that the optimization method can effectively improve the performance of the DCS control system and achieve the optimization purpose.

[0072] The present invention analyzes the structure of a DCS control system, obtains historical input and output variables and corresponding parameters in the DCS control system, pre-processes the historical input and output variables, and then establishes a DCS control system model; then analyzes the characteristics of the DCS control system, selects the parameters to be optimized of the DCS control system, and determines the fitness function; and randomly forms a parameter combination of the parameters to be optimized, and uses it as an initial value, sets an initial search range and randomly initializes the search individual position, and then expands and develops the search, and at the same time, uses the DCS control system model for simulation to complete the optimization of the DCS control system. The DCS control system optimization method based on an intelligent algorithm provided by the present invention can help find the best control strategy parameters, obtain a better solution in a relatively short time, improve the efficiency of the optimization process, and reduce the calculation time. At the same time, the constructed DCS control system model is used for simulation to verify the optimization effect of the DCS control system parameters, so as to ensure that the obtained parameter combination can meet the purpose of optimizing the parameters of the DCS control system.

[0073] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0074] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A DCS control system optimization method based on intelligent algorithm, characterized in that: The following steps are involved: S1. Analyze the DCS control system structure, obtain the historical input and output variables and corresponding parameters in the DCS control system, pre-process the historical input and output variables, and then establish a DCS control system model; S2, analyze the characteristics of the DCS control system, select the parameters to be optimized of the DCS control system, and determine the fitness function; S3, randomly group the parameters to be optimized into parameter combinations and use them as initial values, set the initial search range and randomly initialize the search individual positions, and then expand and develop the search; S4. Use the DCS control system model for simulation; S5. Compare the simulation results and complete the optimization of the DCS control system.

2. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In the step S1, the preprocessing is to optimize the abnormal values ​​of historical input and output variables.

3. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In the step S1, the DCS control system model is established by training the DCS control system virtual model using the preprocessed input and output variables and corresponding parameters to complete the establishment of the DCS control system model.

4. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In step S2, the fitness function is: F = a1 × e -k1×b1 +a2×e -k2×b2 +a3×b3, where b1 is the control error, b2 is the response time, b3 is the system stability, a1, a2 and a3 are weight coefficients, and k1 and k2 are adjustment coefficients.

5. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In step S3, the following steps are specifically included: Set the initial search range and randomly initialize the search individual position, where the formula for randomly initializing the search individual position is: In the formula, i is the search population size, j is the problem dimension, and f i j+1 =λf i j (1-f i j ), f i j is the variable, λ is the coefficient, UA j is the jth lower bound of the given problem, WA j is the jth upper bound of the given problem; Expand your search: Where, Y1 e+1 is the search result obtained in the e+1th round using the search method Y1, Y brst is the best solution obtained in the previous e rounds of iterations, E is the maximum number of search rounds, rand is a random vector with a value between 0 and 1, is the average value of the solutions in the e-th round of iteration; Narrow your search: In the formula, Y2 e+1 is the search result obtained in the e+1th round using the search method of Y2, is the flight function, and Δx is the adjustment range.

6. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In the step S4, the following steps are specifically included: Using the DCS control system model, historical data is used as input to obtain the corresponding historical output information; Then change the parameters of the DCS control system model, use the historical data as input, and obtain the corresponding change output information.

7. The DCS control system optimization method based on intelligent algorithm according to claim 6 is characterized in that: The parameters are the parameter combinations corresponding to the current optimization results.

8. The DCS control system optimization method based on intelligent algorithm according to claim 1 is characterized in that: In step S5, if the simulated change output information is better than the historical output information, the optimization of the DCS control system is stopped, and the parameters used for the change are the optimized parameters of the DCS control system. Otherwise, steps S1 to S4 are repeated.

Images