Research On Random Time Delay Compensation Methods For PROFIET Networked Control System

The increasing need for big data transfer, remote control and distributed device communication due to development of industrial applications has made study of networked control systems(NCSs) a crucial topic of discussion nowadays. However, as the networked control system brings advantages such as resource sharing, maintenance cost-saving and flexible usage, it inevitably leads to data drops and time delays which may greatly degrade system performance to some extent. This makes the analysis of NCS even more complicated and demanding when traditional control methods cannot be applied directly to the design of networked control systems. Thus several attempts have been made in this paper to propose strategies to compensate delay effect on the performance of NCS, which is of great significance for practice.Firstly, this dissertation studies the characteristics of time delay in the networked control system, and analyzes effect of forward delay and feedback delay on the networked control system performance and stability under different parameters and points out the reason of negative performance. Using the Matlab/Truetime simulink toolbox to simulate network modules and nodes, this paper studies the factors which influence the network delay, including network protocol, node-drive and the network bandwidth rate of the interference. Then the simulation results illuminate the characteristics and performance of NCS under different network protocols as well as the principle of different node driven approach.According to the random delay, this paper designs Fuzzy-PID control method and Smith predictor control method, and analyzes these two methods in theory. In terms of the same controlled object, two proposed methods are tested separately in the simulation of NCS to compare the compensation effect of random delay. By changing the model of controlled object, simulation results prove that Smith predictor greatly relies on the precise model of controlled object. To improve the robustness, a Fuzzy-smith compensation method is designed, in which Fuzzy self-tuning of open-loop gain is introduced to the conventional smith predictor. At last, the Fuzzy-Smith method has been tested on multiple controlled objects in the simulation, it verifies that this method is able to reduce the effect of delay and improve the performance of synchronous control system.Experiment platform is built based on the industrial Ethernet PROFINET network: controlled object in NCS is simulated by using National Instruments Semi-physical simulation system based on RTW/x PC principle, network load is accomplished by data communication between multiple PLC S7-300 s, and the system monitoring is realized by the Win CC software.Finally, experiments are done to verify characteristics of feedforward and feedback delay in the theory analysis and random delay is achieved by a delay measurement program. Several comparative experiments under different proposed control methods are done, and experiment results prove that delay compensation algorithm could reduce the impact of delay on the performance of NCS.