Research On Networked Control System With Dynamic Matrix Control

Networked Control Systems (NCSs) are systems with a feedback loop closed via a real-time network. In this way, equipments in the control system will not have the geographical limitations. As a result, NCSs are widely used in large-scale distributed control systems, tele-operation, intelligent traffic systems and satellite control systems, etc. NCSs have received increasing attentions in recent years. Meanwhile, Dynamic Matrix Control (DMC), which developed firstly in 1970s, is also widely used in the area of process control. The most important advantage of DMC is that instead of modeling the object well, the users just need to find its step response, which appears much easier to fill some plants they could not understand them well. This dissertation mainly focus on the combination of NCSs and DMC, especially how to deal with the random transmission delay and packet dropout. The main contents obtained in this dissertation are as follows:1. For a class of less time-sensitive systems, random time delay ischanged to fixed time delay by setting cache in controller and actor. Both controller and actor will record the data and will not work until the maximum time delay comes. The system will decrease its complexity greatly in this way. The calculation method of step response and impulse response when time delay occurs is discussed firstly. The form of controller will not change by replacing them. Then a robust stability sufficient condition for DMC with fixed time delay by applying Jury's dominate coefficient lemma is derived. Finally, some numerical simulations are given to demonstrate the theoretical results.2. For a class of high time-sensitive systems, it is not appropriate to transform random time delay to fixed time delay. In order to deal with this problem, we predict time delay by using the approach and feedback correction to promise the system performance. Controller design method with random time delay is discussed in this paper. It is not necessary to update large amount of data by using mathematical expectation to predict time delay. The stabilization problem is transformed to uncertain parameters problem by using step response coefficient to predict future output and a robust stabilization sufficient condition is obtained with some monotonicity lemma and Jury's dominate coefficient lemma. Finally, a numerical simulation which shows the output of a system with random time delay is given to demonstrate the theoretical results.3. To deal with the packet dropout, DMC has its special advantages compared with traditional control method. DMC will give a vector of control volume once, though only the first will be employed in the algorithm while the others are discarded, which could help to improve the performance when packet dropout occurred. Improved control strategy is given in detail in this paper through an example. Then a robust stabilization sufficient condition is obtained under some restrictions. Finally, a comparison of improved control method in this dissertation and traditional control method is given to demonstrate the theoretical results through a numerical example.Finally, the conclusion and further research topics are given.