A controller performance adaptive optimization method based on intelligent algorithm

Abstract

The present invention relates to the field of data processing technology, and in particular to a controller performance adaptive optimization method based on an intelligent algorithm, comprising: obtaining a real-time sampling value sequence of the controller; obtaining a predicted value sequence; obtaining an error change rate at each moment; obtaining a predicted deviation of the predicted value sequence relative to the real-time sampling value sequence; obtaining a corrected error change rate at each moment based on the predicted deviation of the predicted value sequence relative to the real-time sampling value sequence and the error change rate at each moment; obtaining an error change rate at the next moment relative to the current moment; obtaining a real-time PID differential time parameter at each moment; and obtaining a PID differential time parameter adjusted in advance for the next moment of the current moment. The present invention improves the real-time response speed of the controller by adaptively adjusting the PID differential time parameter in real time based on the real-time response time of the controller and the error change rate of the predicted response time.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a controller performance adaptive optimization method based on an intelligent algorithm. Background Art

[0002] Modern control systems often need to operate stably in highly variable and unknown environments. Technological advancements are driving increasingly stringent performance requirements for control systems, including fast response, high precision, and robustness. Given the inherent nonlinearity, time-varying nature, and parameter uncertainty of systems, optimizing the controller to rapidly respond to input signals is crucial for improving control system performance. This optimization can reduce system response time and improve dynamic performance, which is particularly important in time-sensitive applications.

[0003] Traditional PID controllers and linear control algorithms generally adjust the controller's response time to input signals by manually adjusting PID parameters or setting parameter thresholds. However, if the controller is in a dynamically changing environment, or the application environment and input signals are unstable, the above methods may not be able to achieve timely and accurate responses, thereby reducing the controller's operating efficiency. Summary of the Invention

[0004] The present invention provides a controller performance adaptive optimization method based on an intelligent algorithm to solve the existing problems.

[0005] The controller performance adaptive optimization method based on intelligent algorithm of the present invention adopts the following technical solutions:

[0006] The present invention proposes a controller performance adaptive optimization method based on an intelligent algorithm, which includes the following steps:

[0007] Get the real-time sampling value sequence of the controller;

[0008] Obtain the predicted value sequence; obtain the error change rate at each moment based on the response time of the real-time sampling value sequence and the predicted value sequence at the same moment;

[0009] Based on the difference in response time of the corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, as well as the overall difference in response time changes, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is obtained; based on the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence and the error change rate at each moment, the corrected error change rate at each moment is obtained;

[0010] According to the difference between the corrected error change rates at adjacent moments, the error change rate at the next moment relative to the current moment is obtained;

[0011] The real-time PID differential time parameter of each moment is obtained; the PID differential time parameter adjusted in advance for the next moment of the current moment is obtained according to the real-time PID differential time parameter of each moment and the error change rate of the next moment relative to the current moment.

[0012] Furthermore, the specific steps of obtaining the real-time sampling value sequence of the controller include the following:

[0013] Use an oscilloscope to measure the time difference of signal changes and obtain the real-time sampling value sequence of the controller.

[0014] Furthermore, the specific steps of obtaining the predicted value sequence are as follows:

[0015] The Holt double parameter smoothing method is used to obtain the corresponding predicted value for each sampling value in the real-time sampling value sequence to obtain the predicted value sequence.

[0016] Furthermore, the error change rate at each moment is obtained based on the response time at the same moment in the real-time sampling value sequence and the predicted value sequence. The corresponding specific calculation formula is: in, Indicates the Error rate of change at each moment; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment.

[0017] Furthermore, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is obtained based on the difference in response time of corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, as well as the overall difference in response time changes, including the following specific steps:

[0018] According to the difference in response time of the corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, the average difference in response time between the predicted value sequence and the real-time sampling value sequence at the same time is obtained;

[0019] According to the overall change difference of the response time in the real-time sampling value sequence and the predicted value sequence, the standard deviation of the real-time sampling value sequence, the standard deviation of the predicted value sequence and the covariance of the predicted value sequence and the real-time sampling value sequence are obtained, and then the deviation determination coefficient of the predicted value sequence and the real-time sampling value sequence is obtained;

[0020] Based on the difference mean value and the deviation determination coefficient, combined with the differences between the DTW distances of the real-time sampling value sequence and the predicted value sequence and the ideal response time series, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is mapped; the prediction deviation is positively correlated with the difference between the difference mean value and the DTW distance, and negatively correlated with the deviation determination coefficient.

[0021] Furthermore, the specific calculation formula for the average difference between the response time of the predicted value sequence and the real-time sampling value sequence at the same time is: in, Represents a sequence of predicted values With real-time sampling value sequence The average of the differences in response time at the same moment; Represents a sequence of predicted values Middle The predicted response time of each sampling point; Represents a real-time sampling value sequence Middle Response time of real-time sampling of each sampling point; Indicates the number of sampling points.

[0022] Furthermore, the deviation determination coefficient between the predicted value sequence and the real-time sampling value sequence is obtained, and the corresponding specific calculation formula is: in, Represents a sequence of predicted values With real-time sampling value sequence The coefficient of determination of deviation; Represents a sequence of predicted values With real-time sampling value sequence covariance of Represents a sequence of predicted values The standard deviation of Shows real-time sampling value sequence The standard deviation of .

[0023] Furthermore, the predicted deviation of the predicted value sequence relative to the real-time sampling value sequence and the error change rate at each moment are used to obtain the corrected error change rate at each moment. The corresponding specific calculation formula is: in, Indicates the Error change rate after time correction; Represents a sequence of predicted values Relative to the real-time sampling value sequence The prediction bias of Indicates the The error rate of change at each moment.

[0024] Furthermore, the error change rate at the next moment relative to the current moment is obtained based on the difference between the corrected error change rates at adjacent moments. The corresponding specific calculation formula is: in, Indicates the next moment relative to the current moment The error rate of change; Indicates the Error change rate after time correction; Indicates the Error change rate after time correction; Indicates the current time The rate of change of error after correction; Indicates the current time The corresponding sequence value.

[0025] Furthermore, the PID differential time parameter adjusted in advance for the next moment of the current moment is obtained based on the real-time PID differential time parameter at each moment and the error change rate of the next moment relative to the current moment. The corresponding specific calculation formula is: in, Indicates the PID differential time parameter adjusted in advance for the next moment from the current moment; Indicates the PID differential time parameter at the current moment; Indicates the next moment relative to the current moment The error rate of change.

[0026] The beneficial effects of the technical solution of the present invention are:

[0027] By adaptively adjusting the PID differential time parameter in real time based on the controller's real-time response time and the error rate of change in the predicted response time, the controller's response speed accuracy and system stability are improved. Specifically, the controller's real-time sampling value sequence is obtained and predicted using the Holt dual-parameter smoothing method to obtain a predicted value sequence that better reflects future trends and improves prediction accuracy. Next, based on the response time difference between the real-time sampling value sequence and the predicted value sequence, the error rate of change at each moment is calculated, accurately describing the error's evolution over time and enabling timely adjustment of the control strategy. The prediction deviation is then calculated by comprehensively considering the response time difference and its changing trend between the real-time sampling value sequence and the predicted value sequence, providing a more comprehensive assessment of the deviation between the predicted and actual values. The error rate of change at each moment is corrected using the prediction deviation, minimizing the impact of the error on control accuracy and improving the accuracy of the error rate of change. By analyzing the difference between the corrected error rates of change at adjacent moments, the error rate of change at the next moment is predicted, enabling the controller to respond in advance and improving the real-time performance of the control system. Finally, the PID differential time parameters are dynamically adjusted according to the corrected error change rate, so that the controller can maintain the best response performance in a changing environment, reduce the overshoot and oscillation of the system, and improve the stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0029] Figure 1 This is a flowchart of the steps of a controller performance adaptive optimization method based on an intelligent algorithm of the present invention;

[0030] Figure 2 It is a schematic diagram of the fitting curve of the real-time sampling value sequence;

[0031] Figure 3 Schematic diagram of the fitting curve of the predicted value sequence. DETAILED DESCRIPTION

[0032] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a controller performance adaptive optimization method based on an intelligent algorithm proposed in accordance with the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0033] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0034] The following describes in detail a specific solution of a controller performance adaptive optimization method based on an intelligent algorithm provided by the present invention with reference to the accompanying drawings.

[0035] See also Figure 1 , which shows a flowchart of a controller performance adaptive optimization method based on an intelligent algorithm provided by an embodiment of the present invention, the method comprising the following steps:

[0036] Step S001: Acquire the real-time sampling value sequence of the controller.

[0037] The purpose of this embodiment is to optimize the response time and accuracy of the controller, so it is necessary to collect data on the real-time response time through an oscilloscope.

[0038] Specifically, an oscilloscope is used to directly measure the time difference of the signal change to obtain the real-time sampling value sequence of the controller; the fitting curve of the real-time sampling value sequence of the controller is as follows: Figure 2 As shown, the horizontal axis is the sampling time, and the vertical axis is the real-time response time of the controller. Set the sampling start time to time 0, and the current sampling time is time.

[0039] It should be noted that the real-time response time of the controller will fluctuate due to the influence of network delay and processor load, and the response time can be optimized by adjusting the PID parameters of the controller. In PID control, there are usually three parameters that need to be adjusted, namely the proportional coefficient, the integral time, and the differential time. These three parameters directly affect the performance and response time of the controller. The differential control acts on the rate of change of the error, which can improve the response time and stability of the system. An excessively large differential time parameter will cause the system to respond faster to rapidly changing input signals or reference signals, but may increase the system's sensitivity to noise. Therefore, this embodiment performs real-time adaptive adjustment on the differential time parameter in PID control to optimize the real-time response time.

[0040] Step S002: Obtain the predicted value sequence; obtain the error change rate at each moment based on the response time at the same moment in the real-time sampling value sequence and the predicted value sequence.

[0041] It should be noted that since the method of adjusting the response time through a fixed differential time parameter threshold has low sensitivity in complex application environments, in order to optimize the controller response time, real-time adjustment is performed through adaptive differential time parameters. Since the error change rate of the response time determines the size of the differential time parameter, the differential time parameter should be adjusted based on the size of the error change rate as much as possible. To make the error change rate more accurate, the real-time error change rate can be adjusted by calculating the difference between the actual response time series and the predicted response time series. The error change rate at the current moment is then adaptively determined based on the adjusted real-time error change rate, thereby obtaining the real-time PID differential time parameter. To calculate the real-time error change rate of the controller response time at each sampling moment, a real-time predicted response time value is required. The Holt dual-parameter smoothing method can be used to calculate the corresponding predicted value sequence based on the sampled values, and then the real-time error change rate is determined based on the predicted value sequence. The Holt dual-parameter smoothing method is a time series prediction method used to process data with trends and seasonality. It is an extension of the Holt linear trend method and includes two smoothing parameters: the smoothing level value and the smoothing trend value.

[0042] Specifically, the smoothing level value preset in this embodiment is , smoothed trend value , is described as an example, and other implementations may be set to other values. The Holt double parameter smoothing method is used to obtain a corresponding predicted value for each sample value in the real-time sample value sequence; the Holt double parameter smoothing method is a well-known technology and will not be described in detail in this embodiment; what needs to be explained is that the predicted value at the first sampling moment is equal to the actual sample value; a fitting curve diagram of the predicted value sequence is established, such as Figure 3 As shown, Figure 3 The middle dotted line is the fitting curve of the predicted value sequence obtained by the Holt double parameter smoothing method; the specific calculation formula for the error change rate at each moment based on the real-time sampling value sequence and the predicted value sequence is: in, Indicates the The error change rate at the moment, what needs to be explained is, Starting from 2, the error change rate at the first moment is set to 0; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment.

[0043] What needs to be explained is that Indicates the The error value at time, Indicates the The error value at time, It means the The error change rate at a certain moment, that is, the error change rate is reflected by the ratio of the error values ​​at adjacent moments.

[0044] Step S003: Based on the difference in response time of the corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, as well as the overall difference in response time changes, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is obtained; based on the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence and the error change rate at each moment, the corrected error change rate at each moment is obtained.

[0045] It should be noted that after obtaining the error change rate at each sampling moment, since the error change rate is obtained based on the predicted value sequence, and there is a difference between the predicted value and the real-time sampling value, in order to minimize this difference and make the size of the error change rate more accurate, the overall deviation between the real-time sampling value sequence and the predicted value sequence is analyzed, and the error change rate at each moment is weighted and corrected based on the overall deviation. Then, the error change rate at the next moment relative to the current moment is determined in combination with the corrected real-time error change rate. Among them, the deviation degree represents the deviation of the overall predicted value of the response time from the overall real-time sampling value. The larger the deviation degree, the greater the deviation of the predicted value from the real-time sampling value, and the greater the degree of correction of the real-time error change rate. Therefore, the error change rate can be corrected by the deviation degree to make the adjustment more accurate.

[0046] Specifically, the specific calculation formula for calculating the prediction deviation based on the overall difference between the real-time sampling value sequence and the predicted value sequence is as follows: in, Represents a sequence of predicted values Relative to the real-time sampling value sequence The prediction bias of Represents a sequence of predicted values Middle The predicted response time of each sampling point; Represents a real-time sampling value sequence Middle Response time of real-time sampling of each sampling point; Indicates the number of sampling points, that is, the number of sampling moments; Represents a sequence of predicted values With real-time sampling value sequence covariance of Represents a sequence of predicted values The standard deviation of Represents a real-time sampling value sequence The standard deviation of Represents a sequence of predicted values and ideal response time series The DTW distance, Represents a real-time sampling value sequence and ideal response time series DTW distance, where the ideal response time series Direct acquisition based on the sampling time, which will not be described in detail in this embodiment; Represents a mapping function.

[0047] What needs to be explained is that Represents a sequence of predicted values With real-time sampling value sequence The average difference in response time at the same moment is used to reflect the average deviation of the predicted value from the real-time sampling value. The closer the value is to 0, the smaller the deviation of the predicted value from the real-time sampling value is, and the smaller the corresponding prediction deviation is. is the deviation determination coefficient of the two data sequences, representing the predicted value sequence With real-time sampling value sequence The degree of data fitting is The larger the value, the more the predicted value sequence Relative to the real-time sampling value sequence The higher the degree of fit, the smaller the corresponding prediction deviation. Represents the two series and the ideal response time series The difference in the DTW minimum matching distance of the two sequences is more similar to the DTW minimum matching distance of the ideal response time, that is, The smaller the value, the more similar the two sequences are. Relative to the real-time sampling value sequence The smaller the prediction deviation is. The mapping function maps the result to the range [-1, 1].

[0048] Furthermore, by predicting the value sequence Relative to the real-time sampling value sequence The prediction deviation is used to adjust the real-time error change rate to eliminate the data deviation caused by inaccurate prediction values, thereby obtaining the corrected error change rate at each moment. The specific calculation formula is as follows: in, Indicates the Error change rate after time correction; Represents a sequence of predicted values Relative to the real-time sampling value sequence The prediction bias of Indicates the The error rate of change at each moment.

[0049] At this point, the corrected error change rate at each moment is obtained.

[0050] Step S004: Obtain the error change rate at the next moment relative to the current moment based on the difference between the corrected error change rates at adjacent moments.

[0051] It should be noted that since the actual response time value at the next moment is unknown, the error change rate at the next moment cannot be calculated. Therefore, the error change rate at the next moment relative to the current moment can be calculated based on the changing trend of the error change rate at historical moments. This is equivalent to predicting the error change rate at the next moment relative to the current moment based on the changing trend of the error change rate at historical moments. If the changing trend of the error change rate at historical moments is positively correlated and the larger the trend, the greater the error change rate at the next moment relative to the current moment, and vice versa.

[0052] Specifically, the specific calculation formula for the error change rate at the next moment relative to the current moment is as follows: in, Indicates the next moment Moment relative to the current moment The error rate of change; Indicates the Error change rate after time correction; Indicates the Error change rate after time correction; Indicates the current time The rate of change of error after correction; Indicates the current time The corresponding sequence value.

[0053] What needs to be explained is that Represents the average trend of the error change rate of any two adjacent sampling moments before the current moment. If The value of is positive and the larger it is, the greater the error in the controller response time is. On the contrary, if the value is negative and smaller, it means that the controller response time error is negative and larger. The value of decreases.

[0054] Step S005: Obtain the real-time PID differential time parameter at each moment; obtain the PID differential time parameter adjusted in advance for the next moment of the current moment based on the real-time PID differential time parameter at each moment and the error change rate of the next moment relative to the current moment.

[0055] It should be noted that after obtaining the error change rate at the next moment relative to the current moment, the PID differential time parameter depends on the size of the error change rate. Therefore, the real-time differential time parameter can be adaptively adjusted in advance according to the real-time error change rate to optimize the response time of the controller.

[0056] Specifically, the real-time PID differential time parameter at each moment is obtained (directly obtained through real-time monitoring of the PID controller), and the specific calculation formula for the PID differential time parameter adjusted in advance at the next moment of the current moment is as follows: in, Indicates the PID differential time parameter adjusted in advance for the next moment of the current moment, that is, PID differential time parameter at the moment; Indicates the PID differential time parameter at the current moment; Indicates the next moment Moment relative to the current moment The error rate of change.

[0057] Furthermore, by optimizing the response time of the controller by adjusting the PID differential time parameters in advance between the current moment and the next moment, adaptive real-time optimization of the controller performance is achieved.

[0058] At this point, this embodiment is completed.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A controller performance adaptive optimization method based on intelligent algorithm, characterized in that: The method comprises the following steps: Get the real-time sampling value sequence of the controller; Obtain the predicted value sequence; obtain the error change rate at each moment based on the response time of the real-time sampling value sequence and the predicted value sequence at the same moment; Based on the difference in response time of the corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, as well as the overall difference in response time changes, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is obtained; based on the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence and the error change rate at each moment, the corrected error change rate at each moment is obtained; According to the difference between the corrected error change rates at adjacent moments, the error change rate at the next moment relative to the current moment is obtained; Obtaining the real-time PID differential time parameter at each moment; obtaining the PID differential time parameter adjusted in advance for the next moment of the current moment based on the real-time PID differential time parameter at each moment and the error change rate of the next moment relative to the current moment. The specific steps of obtaining the real-time sampling value sequence of the controller are as follows: An oscilloscope is used to measure the time difference of the signal change to obtain the real-time sampling value sequence of the controller. The specific steps of obtaining the predicted value sequence are as follows: The Holt double parameter smoothing method is used to obtain the corresponding predicted value for each sample value in the real-time sample value sequence to obtain the predicted value sequence. The error change rate at each moment is obtained according to the response time at the same moment in the real-time sample value sequence and the predicted value sequence. The corresponding specific calculation formula is: in, Indicates the Error rate of change at each moment; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment; Indicates the Predicted response time at the moment; Indicates the The response time value of the real-time sampling at the moment, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is obtained based on the difference in response time of the corresponding sampling points at the same moment in the real-time sampling value sequence and the predicted value sequence, as well as the overall difference in response time changes, including the specific steps as follows: According to the difference in response time of the corresponding sampling points at the same time in the real-time sampling value sequence and the predicted value sequence, the average difference in response time between the predicted value sequence and the real-time sampling value sequence at the same time is obtained; According to the overall change difference of the response time in the real-time sampling value sequence and the predicted value sequence, the standard deviation of the real-time sampling value sequence, the standard deviation of the predicted value sequence and the covariance of the predicted value sequence and the real-time sampling value sequence are obtained, and then the deviation determination coefficient of the predicted value sequence and the real-time sampling value sequence is obtained; Based on the difference mean value and the deviation determination coefficient, combined with the differences between the DTW distances of the real-time sampling value sequence and the predicted value sequence and the ideal response time series, the prediction deviation of the predicted value sequence relative to the real-time sampling value sequence is mapped; the prediction deviation is positively correlated with the difference between the difference mean value and the DTW distance, and negatively correlated with the deviation determination coefficient.

2. The controller performance adaptive optimization method based on intelligent algorithm according to claim 1, characterized in that: The specific calculation formula for the average difference in response time between the predicted value sequence and the real-time sampling value sequence at the same time is: in, Represents a sequence of predicted values With real-time sampling value sequence The average of the differences in response time at the same moment; Represents a sequence of predicted values Middle The predicted response time of each sampling point; Represents a real-time sampling value sequence Middle Response time of real-time sampling of each sampling point; Indicates the number of sampling points.

3. The controller performance adaptive optimization method based on intelligent algorithm according to claim 1, characterized in that: The specific calculation formula for the deviation determination coefficient between the predicted value sequence and the real-time sampling value sequence is: in, Represents a sequence of predicted values With real-time sampling value sequence The coefficient of determination of deviation; Represents a sequence of predicted values With real-time sampling value sequence covariance of Represents a sequence of predicted values The standard deviation of Represents a real-time sampling value sequence The standard deviation of .

4. The controller performance adaptive optimization method based on intelligent algorithm according to claim 1, characterized in that: The error change rate after correction at each moment is obtained based on the prediction deviation of the prediction value sequence relative to the real-time sampling value sequence and the error change rate at each moment. The corresponding specific calculation formula is: in, Indicates the Error change rate after time correction; Represents a sequence of predicted values Relative to the real-time sampling value sequence The prediction bias of Indicates the The error rate of change at each moment.

5. The controller performance adaptive optimization method based on intelligent algorithm according to claim 1, characterized in that: The error change rate of the next moment relative to the current moment is obtained based on the difference between the corrected error change rates at adjacent moments. The corresponding specific calculation formula is: in, Indicates the next moment relative to the current moment The error rate of change; Indicates the Error change rate after time correction; Indicates the Error change rate after time correction; Indicates the current time The rate of change of error after correction; Indicates the current time The corresponding sequence value.

6. The controller performance adaptive optimization method based on intelligent algorithm according to claim 1, characterized in that: The PID differential time parameter adjusted in advance for the next moment of the current moment is obtained based on the real-time PID differential time parameter at each moment and the error change rate of the next moment relative to the current moment. The corresponding specific calculation formula is: in, Indicates the PID differential time parameter adjusted in advance for the next moment from the current moment; Indicates the PID differential time parameter at the current moment; Indicates the error change rate at the next moment relative to the current moment.

Images