A real-time performance evaluation method for single-loop control systems in tobacco processing

Abstract

The invention relates to the related technical field of control over the tobacco processing process, in particular to a performance evaluation method suitable for a single loop control system in the tobacco processing process. The method is characterized by including the steps of selecting a certain single loop control system in the tobacco processing flow to serve as an object to be evaluated, identifying a time sequence model of operation data in real time through the least square method with variable forgetting factors so as to improve the identification accuracy of the model, and conducting real-time rolling analysis on the operation data in a sliding window mode so that the real-time performance of performance evaluation can be improved. By means of the method, the forgetting factors change while the fitting error of the model changes. When the difference between the fitting error and the expected error is large, the identification accuracy is improved by appropriately decreasing values of the forgetting factors; when the difference between the fitting error and the expected error is small, the identification accuracy is improved by appropriately increasing values of the forgetting factors. By means of the real-time identification capacity of the method, the performance evaluation level in the tobacco processing process can be improved.

Description

technical field [0001] The invention relates to the related technical field of tobacco processing process control, in particular to a performance evaluation method for a single-loop control system in the tobacco processing process. Background technique [0002] Tobacco processing is a continuous production process. After a long period of production and operation, the performance of each control loop will decline to varying degrees. How to use the process production data to evaluate the control performance, find out the potential degradation of the control system performance in time, and then take preventive improvements Measures to improve process control level and production and maintenance efficiency are urgent problems for process control personnel. [0003] The tobacco processing process has many steps, and each step contains many different types of control loops. The single loop is still the most basic and common control loop. Harris found the feedback invariant item ...

AI Technical Summary

Problems solved by technology

[0004] However, this conventional minimum variance performance evaluation method faces the following main problems in practical application: First, this method uses commonly used identification methods (such as least squares; Burg method, etc.) to identify the time series model of the operating data , the accuracy of the identified model is not high, which affects the accuracy of performance evaluation; second, due to the limitation of the identification method, the operation data is usually analyzed and evaluated in batches, and there is a lack of real-time rolling analysis of production data. Not real-time

Images