A simple rolling mill model with linear quadratic optimal controller

Abstract

The tandem rolling of metal is a complex engineering process whose optimization represents a challenging control problem. The current technology relies on complex nonlinear models that are used to develop control strategies to address the various control problems encountered. Such models however, necessitate extensive computational requirements for simulation. Also, in various instances, access to the detailed models may be unavailable, or computational capability may be limited, yet it is desired to obtain some rough indication of performance of specific control systems when coupled to the mill. Thus the need arises for a simplified model (which can be easily implemented on a PC) for testing of control concepts. This paper fulfills such a need by providing a simple linearized model of the mill. An example of the usefulness of the model is presented by coupling to a linear quadratic optimal controller with a simple method to mitigate disturbances in incoming metal thickness. Satisfactory results are demonstrated by a simulation.