Adaptive Control And Its Application Research Based On Characteristic Model

A control system is usually supposed to be a definite model in the traditional control system. So the model is created as an accurate one and the controller is designed according to the dynamic and static performance index. However, we can't normally get the accurate models of many systems, especially some complex or time-varied ones. Consequently, this paper proposed to base on the characteristic modeling theory, and create characteristic-model adaptive controller combined with the control performance requirements. The characteristic model is utilized to estimate the unknown parameters of the system to design the controller. For the control system being affected by all kinds of disturbances and uncertainty, the variable structure controller, adaptive controller based on extended state observer or RBF neural network compensation were presented and the characteristic-model adaptive control algorithm ran successfully on the SIMATIC PCS7, which was used to control the discharge port temperature of a tubular heat exchanger. The detail studies are as follow:Aimed at a kind of parameter-unknown control problem, a variable structure control (VSC) method based on the tracking-differentiator (TD) is presented by creating a characteristic model. The characteristic model is created by the method of least squares (LS) with forgotten factor to estimate the plant's the parameters. When the system is steady, the output of the model is consistence with the original one. Considering there are always various disturbances in complex conditions, the VSC are applied to control the plant with its good quality of insensitive to the disturbances and uncertain parameters.Aimed at the control problem of a class of unknown parameters, time-invariant and high order system, a low-order time varying difference equation is presented to represent the system characteristics based on the characteristic modeling theory. The extended states observer (ESO) which is independent with the accurate plant model can be used to estimate the system's all disturbances, while the least square method with forgetting factor is used to estimate the system's parameters. In the steady state, the output of the characteristic model is equivalent to the one of the system. Therefore, the controller can be designed for compensating the total disturbances using ESO.Aimed at the control problems for a kind of unknown parameter object, a control method of character modeling is proposed which is based on Neural Networks (NN) compensation. The time-variable difference equation of the plant is built according to the characteristic modeling theory. The method introduces a NN supervisor controller to compensate the dynamic modeling error of the characteristic model with an adaptive control law so as to improve the dynamic performance. At the same time, when the system is disturbed, the feedback control based on golden-segmentation control law could immediately overcome the disturbance.The tubular heat exchanger has the characteristics of time delay, strong inertia, and its dynamic characteristic will change in different industrial conditions. Therefore, general control methods can't keep a good dynamic and static performance of the plant due to the update of the parameters not timely. The characteristic of the heat exchanger is studied and an intelligent adaptive control method based on the characteristic model is proposed. SIMATIC PCS7 is used to control the temperature of the discharge port. The results show that the heat exchanger control system can be well adaptive to the change of industrial conditions and has a good dynamic performance.