The Application Of Fuzzy Machine Learning In Laminar Cooling Process On The Runout Table

Hot strips are cooled by laminar flow water on the runout table between the finisher and the coiler. This process is used to control the strip from the finishing temperature to the coiling temperature according to technical requirements and obtain good mechanical properties for the strip. Modeling and control of the laminar cooling process is difficult because of its complex nature (e.g. highly nonlinear, time varying), frequent variations of operating conditions and the difficulty of measuring the strip temperature online. In a hot strip mill, control of the laminar cooling process is essential to product quality and safe operation of the mill.AFS (Axiomatic Fuzzy Set) fuzzy logic system is a new fuzzy logic system. In the framework of AFS theory, a new algorithm of determining membership functions for fuzzy concepts according to original date and information and propose AFS fuzzy logic is presented, which make the construction of the fuzzy membership function and the fuzzy logic system more objective. AFS theory is primarily applied in the field of data mining,case-based reasoning,pattern identification and consensus diagnosis.This research was a part of the 973 project "Research on theory and algorithm of real-time intelligent control for complex industrial processes" (2002CB312201) supported by the China National Key Basic Research and Development Program. Research on key issues in the modeling and control of the laminar cooling process on the runout table is conducted in this thesis. Detailed works are as follows:(1) In chapter 1, first, a survey is given about the laminar cooling process and the development of its technology. The industry's background of the problem is also introduced. What is the "Machine Learning" and the application of case-based reasoning (CBR) is presented in the second part of chapter 1. The brief overview about AFS (Axiomatic Fuzzy Set) fuzzy logic system is given in chapter 2.(2) In chapter 3, in the frame of AFS theory, a new improvement CBR approach based on the classical Machine Learning is proposed. The application of this new CBR approach in the parameter identified is given to solve the problem of the laminar cooling's controller. The paper also gives the result of the simulink result.(3) In Chapter 4, a survey of the AFS theory's application in the extraction of fuzzy rule is given. This approach is consisted of fuzzy logic clustering algorithm, fuzzy closensess degrees and converse extraction of fuzzy rule. The extraction of fuzzy rule based on AFS fuzzy logic is used in the feedback controller of laminar cooling process on the runout table and gets the primary result by matlab. In the end of the paper, the development of AFS fuzzy logic's application is presented.