The Research Of Mutivariable Model Predictive Control For Integrating System

Integral is a typical unit in control theory, whose existence leads to the integral role. While error can be eliminated by introducing integral role, the system stability will be reduced. For the controlled object, many typical installations and equipments have integral links with the real target. Such kind of target is non-self-regulating and is open-loop unstable. But in predictive control, the closed-loop stability can usually be guaranteed with infinite predictive horizon. So, it is very significant to study the model predictive control by means of infinite predictive horizon.This paper studies the algorithms for the system's nominal stability, aiming at the special object of integrating system. The main contents can be summed up as follows:1) This paper introduces the single-variable and multi-variable state space model, a infinite predictive horizon method through which the optimization problem of objective function is converted into a quadratic optimization procedure with states constraints and input constraints, and the rules to ensure the controller's stability.2) In the process of dealing with integrating system by using the infinite predictive horizon, there will be encountered the situation that state constrains and input constrains are conflict with each other, which leads to the infeasibility of the system. For this problem, some methods are introduced to dispose the infeasibility of system with soft constraints. Then, when the restricted conditions are hard constrains, by introducing the concept of the condition set value, this paper propose a multivariable model predictive control for integrating system by online changing set value. The simulation results show that the present method is efficient and feasible.3) In order to settle the problem that the sampling rate of output is not often accordant with the holding rate of input in the complex industry process, on the basis of the model predictive control algorithm proposed, this paper further studies the multi-rate multivariable system in which outputs measured are less frequently than inputs updated, proposes the multi-rate multivariable model predictive control's state space model and then find out the optimal control consequence by means of infinite predictive horizon optimization strategy.