Research On Time Delay Compensation For NCS

With the development of computer communication network technology, as well as the constant expansion and the complexity of the continuous improvement of industrial systems,the traditional point to point control method has been unable to meet the complex industrial control systems. Networked control system (NCS) has been rapidly applied to a variety of systems and remote control for its advantages, such as resource sharing, distributed architecture, less connecting cables, low cost, maintenance convenience, and so on. Because of the existence of signal transmission delay induced by network, congestion and packet loss and other issues, the traditional control theory based on the conditions of no transmission delay and modulation is no longer applicable to the NCS. In terms of the actual control system, it is very difficult to establish a precise mathematical model of plant. Many scholars are focusing research on NCS, in order to overcome uncertainty of plant and system's sensitivity to interference. Some problem is stilly not to be solved, though a lot of result was got. Therefore, the following problems are studied in this dissertation:1.The causes of network time delay are described. Some methods are introduced to solve the delay, such as augmented deterministic discrete-time model, queuing method, perturbation method, stochastic optimal control method, robust control method,fuzzy logic modulation method, model-based control method and so on.2.By object reference parameter adaptive law based on Lyapunov stability theory, the compensate system structure is designed for NCS. It means a packet loss occurs that the data transmission time beyond a range. In this condition, the former data can be used to predict the lost packets. The stability of NCS is analyzed.3. An object reference parameter adaptive compensator is adopted to solve the plant model uncertainty and dead-time. This compensator can be used for all uncertain plant with large dead-time which satisfy some stability conditions. In the compensating system, in order to make the model match the being controlled plant, the predictive output follow that of the uncertain part of it and the model output trace that of it, a group of models are used to imitate the uncertain part of it. 4. The true time toolbox is used as a NCS simulation platform, and the NCS model is constructed by it. NCS simulation has been done on the platform. Simulation results prove that the NCS based on an object reference parameter adaptive compensator has good power in dead-time compensating and adaptive tracing.