Toward reliable adaptive controllers for industrial process control.

摘要

In this dissertation we explore some of the issues involved in synthesizing adaptive control algorithms suitable for use in the process industries. We begin with a discussion of the process control needs of the process industries and focus on the way in which adaptive control techniques might fit into the industrial setting. The discussion focuses on specific problems which the industrial setting presents from the viewpoint of adaptive control, as well as specific instances in which adaptive control is especially useful. This points out several requirements to be met by the adaptive control system.; Next, we review some basic concepts involved in adaptive control. A survey of the adaptive control literature is presented, with special emphasis on the variety of algorithms available, their strengths, and their weaknesses from the point of view of industrial applicability. A review of the literature on robustness of adaptive control systems is undertaken as well. Several key factors which tend to enhance desirable properties in an adaptive control system are identified. The discussion leads in a natural way to a "recipe" for synthesizing adaptive control systems with desirable properties.; With this "recipe" in mind, a structural description of a prototype self-tuning control system designed to include these "elements of success" is offered. The final chapter of this thesis is devoted to exploring ways of including these ideas in the synthesis of an adaptive control system. These concepts are illustrated through a series of instructive examples depicting typical situations in adaptive control application and suggesting specific strategies for coping with various difficulties in each case.