Application Research On Multi-model Adaptive Control In Overheated Steam Control System

Dynamic characteristic of power plant thermal process is close related to operation conditions of power units. For peak load unit, its process model and structure parameters have obviously fast time-varying characteristic while load changing. It cannot get satisfactory control result using robust or adaptive control scheme for controlled objects of this sort. Multi-Model adaptive control scheme is exclusively designed for this sort of objects.Neural networks have the ability of approach any nonlinear function. Configuration optimizing RBF neural networks have characters of short modeling time, simple structure and good generalization, This merit makes them applied in thermal processes widely. In this paper, configuration optimizing RBF neural networks are introduced into Multi-Model adaptive system for improving the performance,the main contents are as following.Reviewing the history of generation, development, application of Multi-Model adaptive control and neural networks; introducing the theory of Multi-Model control and Neural networks and its applications in process control fields And prospecting further Neural networks applications in process control.A fast resource optimization networks (FRON) is proposed based structure characters of RBF networks and used in spray-desuperheater modeling. Combining Neural networks and Multi-Model control in control schemes, improving the performance of the control system at the condition of small dimension of the static model collection.The characteristic of the developed Multi-Model control scheme lies in identifying the mismatch degree and compensating control defect according to the bias between output and destine, which make property of Multi-Model control system fine during all working condition, And according to overheated steam property of hysteresis, inertia and time-varying, a Multi-Model control in series scheme is adopted,which take advantage of mature and reliable PID. The scheme is a good feasible control tactics in engineer application.