Font Size: a A A

Research On Network Control System Based On Model Predictive Control Algorithm

Posted on:2012-04-14Degree:MasterType:Thesis
Country:ChinaCandidate:L YuFull Text:PDF
GTID:2178330338451660Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With the widespread use of computer networks and control technology and network technology continues to evolve, the network control system came into being. Since the introduction of the network created a new control problem, in this paper, with a random time delay for the network control system control issues has researched.In this papers ,an improved generalized predictive control algorithm has been proposed based on the generalized predictive control algorithm. The algorithm avoids solving the Diophantine equation, but also simplifies the problem of matrix inversion. With the improvement of the order of the controlled system, the algorithm could save time. At the same time the stability of the algorithm was proved. Finally, the algorithm is successfully applied in the networked control systems with a random time delay.Taking into account the classical PID control algorithm has been widely used in industry. And the algorithm in order to achieve good control effect must be offline or online tuning PID regulator parameters, but it is very difficult for less experienced technical staff or time-varying systems. To solve this problem Yamamoto.T who combined generalized predictive control algorithm with the PID control algorithm proposed GPC-based PID control algorithm. The algorithm is not very suitable for network control system. Because the network control system and the delay system are not a simple equivalent when the actuator uses time-driven and the random network delay from actuator to controller greater than a sampling period. Therefore, this paper presents an improved GPC-PID control algorithm which is more suitable network control system.Due to the improved GPC - PID algorithm is combining PID control algorithm with GPC algorithm, so the algorithm is only suitable for linear systems. To solve the problem, finally, the paper introduces the artificial neural network identification method. The generalized predictive model parameters of the nonlinear system are successfully identified, to make using the improved GPC-PID control algorithm in nonlinear network control system possible.In order to further test the study in this paper, the corresponding simulation analysis in the proposed algorithm is given. The simulation analysis shows that improved generalized predictive control algorithm can not only obtain good control effect and save time than generalized predictive control algorithm. Through the improvements to the GPC - PID control algorithm simulation result is proved that the algorithm has obtained good control effect in the presence of random time-delays network control system,at the same time, avoid the problem of offline and online setting PID parameters. Finally, through simulated analysis to nonlinear network control system, the simulation results show that the artificial neural networks introducing into the generalized predictive model parameter identification make improved GPC - PID control algorithm can achieve effective control for nonlinear network control system.
Keywords/Search Tags:network control, random time delay, generalized predictive control, BP algorithm, nonlinear, improved generalized predictive control, improved GPC-PID control
PDF Full Text Request
Related items