Fuzzy neural networks control for hydraulic AGC system of aluminum cold rolling mill

Abstract

The hydraulic AGC system of aluminum cold rolling mill is directly related to the quality and effectiveness of cold rolling aluminum sheet strips. Traditional PID control becomes difficult to satisfy the necessity of improving the control performance of cold rolling mill. High precision, simple and effective control strategies are very important for the hydraulic AGC system of cold rolling mill. For the complex control problems, it is also desirable to integrate neural networks into fuzzy control so as to simplify and automate the specification of linguistic rules. This leads to good adaptation, good robustness and less dependency on the precise model of the control system. In the paper, a fuzzy neural networks control has been developed and applied to the screw-down mechanism control of hydraulic AGC system of cold rolling mill. The simulation experiment results verify the superiority of the proposed compound control to the conventional PID control in the static and dynamic control performance. The merits of both fuzzy control and artificial neural networks are well used in the hydraulic AGC system of aluminum cold rolling mill.