Research On Multivariable Time-delay System Based On TC Controller And Extended State Observer

Actual control system is complex in the process of industrial structure which have multiple control circuit, the time-delay influence the normal operation of equipment, complex industrial plants have coupling problems in the process industrial structure, conventional control method could unable to achieve satisfactory results, in order to meet the demand of industrial process of high product quality, so this paper look for a kind of advanced control scheme. Conventional PID controller make it is not suitable for large time-delay system because of its parameter setting problem itself, at the same time the traditional Smith prediction control due to its reliance on forecast model, so the model mismatch problem is existed in practical application. In this paper, from the viewpoint of cybernetics, a control scheme is proposed based on information system model in view of the large time-delay system with TC and ESO. Research idea of this article includes four parts:(1) Control principle of TC controller is not based on accurate model information, it has simple control structure, and contains only two parameters. First of all, its control law expression is deduced, and puts forward a parameter setting method, secondly, the mathematical proof of TC controller control law is combining with linear time-invariant systems, the analysis about the feasibility of TC controller is made in large time-delay system from the angle of mathematics;(2) Aim at Smith prediction control model mismatch problem, when the forecast with the actual controlled plant model error exists, the disturbance observer is designed to compensate the interference error of the model mismatch, TC controller improve the robustness of the system at the same time, the single input single output large time-delay system is progressed for simulation compared with the conventional PID control, the superiority and feasibility of this control scheme is verified;(3) TC controller and extended state observer are application to the uncertain factors system, it is based on the system of compensation to reduce the coupling effect and improve the control effective, the two input two output time-delay coupling system carry on the simulation with conventional PID-Smith control strategy, the superiority and feasibility of the control scheme is demonstrated;(4) Two input and output control system eventually quantity expression is deduced in this paper, the RBF neural network have good ability of learning approach, at last the online learning RBF-ESO optimization algorithm is proposed.Multivariable systems with time-delay is controlled in TC-ESO, TC controller and extended state observer could estimate system, the simulation results show that the method reduces the affection for system coupling, and the control system has simple structure, it is easy to adjust the controller parameters, as well as helpful for engineering implementation.