Research On Dynamic Matrix Control Algorithms And Itsapplication For Multivariable Time-Varying Delay Systems

Time delay widely exists in industrial systems such as industrial process, mechanicaltransmission and fluid transfer system. Due to the difference of the system’s structure itself andsome factors like the interference of the external environment, time delay may appeared asdifferent properties such as random, bounded, time varying and uncertain, which make thedesign and analysis of multivariable systems quite difficult.Dynamic Matrix Control (DMC) is a typical model predictive control algorithm, which iswidely used in industrial areas such as process control as a result of its simple structure, goodperformance, flexible application and some other advantages. For a class of multivariabletime-varying delay systems, the research of DMC algorithm and its application for networkedcontrol systems and inverted pendulum device is investigated in this thesis. The mainachievements are as follows:Firstly, DMC algorithm for a class of multivariable time-varying delay systems is designedand analyzed. Considering the character of time-varying delay, system is converted into a systemwith fixed bounded delay. By using a multi-input and multi-output (MIMO) state-space model,and optimizing the performance index function which includes delay information, the DMCcontroller for time-varying delay is derived. Then the sufficient conditions for asymptoticstability of the controller are given based on the Lyapunov stability theory and LMI method.Further more, considering there is a model mismatch between the prediction model and thesystem model, the problem of design and analysis of the DMC algorithm’s robustness isresearched, and the sufficient conditions of robust stability are determined.Secondly, the application of DMC algorithm described above is researched in networkedcontrol systems (NCSs). Considering a class of networked control systems with boundedtime-vary delay in both forward and backward channels, transform the model into the form ofmultivariable time-varying delay systems, and the DMC controller is designed. Make a doubletank water level networked system as an example, the validity of the algorithm is proved by simulation experiments. The results show that, the networked multivariable DMC algorithm mayhave an effect of compensation on the network-induced delay, and the algorithm is robust stable.Finally, the DMC algorithm designed for the control of inverted pendulum is presented andrealized. By using the Newton-Euler modeling method and the linearization method, thestate-space model of the inverted pendulum is derived. Considering the fact that datatransmission of the system may be influenced by time-varying delay, the DMC algorithm isdesigned. Combined with the inverted pendulum platform, the derived DMC algorithm iscompiled into the controller codes. Through a series of real-time control experiment, theeffectiveness and practicability of the proposed results are demonstrated.