Dissertation > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Automation systems > Automatic control,automatic control system > Computer control, computer control system

Research on Intelligent Control Strategy of Flatness for Tandem Cold Mill

Author LvQiDong
Tutor LiuJianChang
School Northeastern University
Course Control Theory and Control Engineering
Keywords flatness pattern recognition neural network predictive control model-reference adaptive control
CLC TP273.5
Type Master's thesis
Year 2008
Downloads 124
Quotes 3
Download Dissertation

Flatness is an important quality index of rolled plate and strip, and flatness control is the key technology of modern high precise rolling mills. As far as the technique is concerned, the key scientific problems in the world are pattern recognition and the realization of flatness intelligent control. Choosing the intelligent control strategy of flatness for tandem cold mill as the research object, the author has done a lot of theoretical and simulating research on flatness pattern recognition, flatness predicting online, and flatness close-loop control.Pattern recognition of flatness is very important for flatness control. For the problems of the traditional methods and the methods of fuzzy and neural network for flatness recognition, this paper puts forward an intelligent method based on static neural network, and established the model for flatness pattern recognition using Legendre multinomials as basic patterns, with only 3 input and 3 output signals based on fuzzy distances. The model is simple, fast, precise, its nodes have clear physical meanings, and is a useful new method for flatness recognition.In view of the time variability and the non-linear character in the hydraulic actuator of flatness control system, this paper put forward a model-reference adaptive control strategy based on fuzzy neural network. This strategy not only reserves the simple and realtime advantages of the model-reference adaptive control algorithm, but also overcomes the disadvantage influence brought in by nonlinearity in use of neural networks, and the performance and dynamic features of the actuator of flatness control system are greatly improved.In the rolling process, rolling gap takes to involve many nonlinear factors, such as bending force, rolling force, and other initial conditions.It is difficult to establish the precise expressions which can express the static and dynamic relations among the parameters using conventional mathematic modeling tools. This thesis establishes a flatness intelligent prediction model based on Elman network, which is not based on analytical principle and can solve many problems which are related with complicated system modeling.Intelligent control system of flatness control is established by combining the flatness pattern recognizing model, the flatness predicting model and the adaptive control strategy of hydraulic system. The simulation indicates that the system has strong adaptive ability to the environment and controls the flatness at the range of 0-3I. The proposed intelligent control strategy has important meaning on realizing flatness on-line real-time control and promoting the development of flatness control model.

Related Dissertations
More Dissertations