Application Of Fuzzy Control And Neural Computing Technology To Shape Control

Our country's steel rolling industry has developed at full speed in recent years, and our country has become one of the countries whose output of steel are most in the world. But currently sizable disparity exists in our country's steel production as for steel rolling technology compared with international advanced level, and the economic losses brought by bad shape are great every year. Shape precision is one main quality index of cold rolled strip and is also an important factor which determines its competitiveness in the market. The urgent demand of improving the quality of cold rolled strip has made shape control be an important problem that faces our country's steel enterprises. Factors that affect shape are very complicated, and at the same time greater difficulty exists too in the shape measurement technique. These bring quite great difficulty to the automatic shape control. In the closed loop shape control system, pattern recognition of shape defects and shape control are two relevant important links. The real shape must be recognized to establish rational control strategy and the recognition results must be quantificationally provided to the next control link with promissory parameters. An approach to shape pattern recognition based on Radical-Basis-Function neural network has been put forward. Lerande orthodoxy polynomial is selected as standard shape patterns, Euclid distances between the stylebooks to be recognized and six standard stylebooks are used as input of the network, three eigenvalues of shape are used as output, and shape pattern recognition model has been established based on orthogonal least squares learning algorithm. This approach has clear physical meaning and is brief and simple, and the time for training is also very short. Real shape defects are recognized using the model obtained by this approach. The simulation results indicate that the model can recognize complicated complex shape defects and has fast recognition speed and high precision. The shape control system of Universal Crown rolling mill has the characteristics of inertia and lag, and the mathematical model of its bending intermediate rolls system has strong time-variant property and uncertainty. A fuzzy logic controller based on Genetic Algorithm has been devised aimed at this point. GA is employed to optimize the membership functions of the Fuzzy Logic Controller and the initial values of the quantifying factors and the scaling factor. And the quantifying factors and the scaling factor are adjusted on-line according to the output of the fuzzy control query table. The controller is applied to the bending intermediate roll system of the UC rolling mill. The simulation results indicate that the quadratic component of the shape defects can be rapidly and efficiently controlled, the FLC optimized by GA has better control performance than conventional FLC and it has self-adaptive capability to a certain degree.