Research On Time Delay Compensation Control Strategies For Networked Control System Based On Predictive Control

Networked control system (NCS) has the advantages of resource sharing, easy installation and maintenance, remote operation and control, high reliability, which has a wide range of the application areas such as process control, aerospace, robotic, remote teaching and telemedicine, etc. Due to the constraints of network bandwidth and network communication rate, network delay will inevitably exist in NCS in the process of mutual exchange of information, which will cause adverse effect on the performance of NCS. Therefore, how to weaken the adverse effect on the performance of NCS caused by network delay and explore an advanced control strategy has important theoretical significance and practical value.The subject is based on the Natural Science Foundation of Shandong Province: research on industrial network control system synchronization control performance and control strategy (ZR2009GM027). According to the analysis and research on a large number of domestic and foreign related literature, and on the basis of a deep analysis on the characteristic of network time delay in NCS, a network delay compensation strategy based on generalized predictive control (GPC) is presented to weaken the adverse effect on the performance of NCS caused by network delay.In order to analyze the effect of network time delay on the performance of NCS, the TrueTime toolbox is applied to establish a NCS simulation model in which a PID algorithm is applied to control a direct motor, and the performance of NCS under the cases of different time delay are simulated respectively. Simulation results show that the existence of network delay will bring adverse effect on the performance of NCS, and excessive delay may even make NCS become unstable.Based on a deep analysis of the principle of GPC algorithm, a network control system simulation model of GPC is established in order to study the regulation effect of GPC algorithm on the performance of NCS existing time delay and model mismatch, , and the performance of GPC algorithm under the cases of different time delay and model mismatch are given. The simulation results show that GPC algorithm has a good control performance to NCS with fixed time delay, but a bad performance with random time delay. The simulation results also show that GPC algorithm still has a good control performance to NCS with a certain degree of model mismatch and a strong robustness to model mismatch, but a worse performance to model mismatch caused by random delay jitter. An approach by measuring average network delay and introducing it into the predictive model of GPC controller is presented to weaken the problem of model mismatch. Simulation results prove the effectiveness of the method.Aiming at the problem that only using GPC algorithm can't guarantee a good control performance of NCS with random time delay, a time delay compensation strategy combined by GPC algorithm and time delay compensator is presented. The strategy includes both the GPC algorithm for multi-step predictive characteristic and the advantage of time delay compensator that can select the most appropriate control data, which can effectively weaken the bad influence of uncertain time delay on the performance of NCS. The effect of the strategy under the circumstances of different random time delay is simulated. Simulation results show that the strategy can gain a better result under the circumstances of different random time delay and ensure a good performance of NCS.On the basis of simulation studies on GPC algorithm, the implementation process of GPC algorithm on network control experimental platform is given, which provides some theoretical bases for establishing actual experimental platform and the verification of actual effect in future.