Adaptive Intelligent Guidance Control For Time-varying Large Time-delay Plant

There are many time-varying large time-delay phenomena in many control processes in the industrial field.For the control of time-varying large time-delay objects,it is usually necessary to identify each mode on-line,but the data processing ability of the controller is required to be high and may lead to the untimely' control,how to reduce the frequent on-line modeling operation in the control process is of great significance.In addition,it is helpful to the design of auxiliary control algorithm and the debugging of controller for the design of time-varying time-delay experimental platform in the laboratory environment to simulate the actual industrial scene.In view of this,the main work includes the following aspects:1.The large time-delay problem of time-varying large time-delay plant is studied first.Firstly,the parameter graphic method and the Hurwitz criterion are used to solve the PID parameters stability region of the system with and without time-delay respectively,from the point of view of stability region,the degree of control difficulty of time-delay plant is quantified.In order to make up for the deficiency of Smith prediction,.a three-stage control scheme is proposed,and the intelligent guidance control algorithm is designed based on the idea of auto disturbance rejection control.The dynamic compensation principle under model mismatch and parameter imbalance is described.Through the simulation comparison with Smith prediction algorithm.The results show that the algorithm is more suitable for the case of model mismatch,wider parameter domain and allowing time delay in the feedback loop.2.To study the control problem of time-varying and large time-delay objects.Firstly,in view of the advantage of the intelligent guidance control algorithm that allows model larger mismatch,a multi-modal adaptive control scheme based on intelligent guidance control is proposed.Then,different control schemes are designed according to the degree difference of the model mismatch caused by the time-varying object:the robustness of intelligent guidance control is used to solve small time-varying problems,the parameter of self-tuning controller is used to solve the problem of medium time-varying,online modeling and initial value optimization of controller parameters are used to solve large time-varying problems.Then,the mismatch threshold of intelligent guidance control is given in the experiment to determine the large,the middle and small time-varying boundaries.Finally,the fuzzy self-tuning algorithm based on intelligent guidance control rate is designed for the controller parameter self-tuning.The two-point method and recursive least square method are used for online modeling,and the particle swarm algorithm is used for the initial value optimization of the controller parameter.The results show that different control schemes can be adopted by the adaptive intelligent guidance control algorithm according to the different time-varying degree of the object,this makes it possible that the algorithm not only ensures the real-time control,but also avoids the computational burden brought by frequent modeling,and has certain intelligence and good performance.3.Based on the integrated test box,the experiment platform is designed and reformed.Firstly,the mechanism method is used to model the integrated test chamber.According to the physical meaning of the parameters,a pure physical control system with time-varying delay characteristics is constructed.Then,considering that it is difficult to make the designed controlled object complex enough by the transformation of pure physical structure,the digital hysteresis,second-order and time-varying transformation are carried out to realize the design of digital physical object.Finally,in order to further diversify the plant,a time-varying delay hardware in the loop simulation platform is built,which takes the test box as the controller and the abundant and changeable mathematical model in Simulink as the controlled plant.The platform design provides technical support for laboratory research and practical application of algorithms.