A Study Of Adaptive Technology Based On Model-Free Theory

At present, the controllers used in industrial process control are mostly classical PID regulators which were developed in the 1940s. The basic idea of designing this kind of controllers is based on linear system construction. Despite of generally ensuring stability in many applied systems, this design cannot guarantee control precision and parameter adaptability is poor. It cannot overcome unknown strong nonlinear influence and disturbed effect. Modern control theory has been developed greatly in recent years, but it meets difficulties on many applying occasions, for many mathematical models of controlled objects cannot be suitably given. Therefore, it is of great theoretical significance and applicable value to find out various control methods and algorithms combining with the characteristics of industrial process. These control methods and algorithms are supposed to have low demand to models, compute simply on-line, and are with better comprehensive control effect.Model-free control algorithm and controller are a new kind of controller, which combines modern control theories with classical PID. They are independent of the mathematical model of controlled objects, and can ensure the stability of closed-loop system; they are with fine control effect and the algorithm is simple.The paper introduces the new control methods and conducts a research on the design problems and concrete improvement methods of model-free control. It presents that by combining tracking differentiator with model-free adaptive controller (MFAC) and joining their advantages, it is possible to design a controller fit for such complicated systems as large-scale time delay system, time-variation system and strong coupling system. This controller realizes the fast and non-overshoot control on strong disturbance system and large-scale time delay system by arranging transitional process of tracking differentiator. The paper carries out a contrast emulation research between the MFAC with the tracking differentiator and the non-linear PID controller. The advantages of the controller can be seen directly and the applications of model-free controller to industry in terms of parameter optimization are theoretically explained the mechanism of these advantages and briefly introduced.