Title:
MODEL AUTONOMOUS LEARNING METHOD SUITABLE FOR PREDICTIVE CONTROL IN PROCESS INDUSTRY
Document Type and Number:
WIPO Patent Application WO/2024/016556
Kind Code:
A1
Abstract:
A model autonomous learning method suitable for predictive control in process industry. The model autonomous learning method suitable for predictive control in process industry is implemented by taking a multi-input single-output system as a basic subsystem, and comprises: step 1, model setting and learning parameter auto-selection; step 2, parameter distribution modeling; step 3, data acquisition and processing; step 4, model parameter learning; step 5, model parameter check; and step 6, model prediction error comparison. According to the model autonomous learning method suitable for predictive control in process industry, the dynamic characteristics of a process can be autonomously learned online; the type of a parameter to be learned can be automatically selected according to a preset model parameter, so that the parameter to be learned has real physical significance, and in a learning process, a preset model has a constraint effect on a result of model learning so as to improve the reliability of the model; learning results of a parameter are checked, and a prediction result is evaluated, thereby further improving the reliability of the model; the use of a parameter dead zone improves the stability of model parameters, and control effect fluctuation caused by frequent change of the model parameters is reduced.
Inventors:
CHU JIAN (CN)
LIU LEI (CN)
FENG KAI (CN)
WANG JIADONG (CN)
LIU LEI (CN)
FENG KAI (CN)
WANG JIADONG (CN)
Application Number:
PCT/CN2022/136000
Publication Date:
January 25, 2024
Filing Date:
December 01, 2022
Export Citation:
Assignee:
SUPCON TECH CO LTD (CN)
International Classes:
G05B13/04
Foreign References:
| CN114924489A | 2022-08-19 | |||
| CN112000001A | 2020-11-27 | |||
| CN101782743A | 2010-07-21 | |||
| CN106991212A | 2017-07-28 | |||
| CN109521677A | 2019-03-26 | |||
| CN110174841A | 2019-08-27 | |||
| CN111797992A | 2020-10-20 | |||
| CN114488785A | 2022-05-13 | |||
| US20190303767A1 | 2019-10-03 |
Other References:
YANDING WEI: "Research on Convergent Speed of Iterative Learning Control", MECHANICAL & ELECTRICAL ENGINEERING MAGAZINE, no. 5, 30 September 1999 (1999-09-30), pages 178 - 180, XP093131754
Attorney, Agent or Firm:
HANGZHOU YUXIN UNITED INTELLECTUAL PROPERTY CO., LTD. (CN)
Download PDF: