Research On Key Technology And Applications Of Model Predictive Control Software

In this dissertation, the key technology and applications of model predictive control software are studied. For implementation and practical requirement, the model prediction and local steady-state optimization of model predictive control algorithm are improved. The method that simple local steady-state optimization algorithms of model predictive control that based on goal programming is presented. The control schemes that are applied to two chemical processes based on model predictive control are proposed. Implementations and simulations are given. Finally, robust parameter tuning of PID controller based on MiniMax optimization rule is presented.This dissertation is composed of follows:1. The model predictive control software - FrontAPC is introduced. Some characteristics and their implementation are presented in the multivariable model predictive control algorithm.2. In the state space framework, using parameter model and non-parameter model, the improved algorithm of model predictive control that can deal with unstable system by the improvement of model prediction and feedback is proposed. The simulation for long-time constant system and integration system is given. For long-time constant system, the improved algorithm can get good control performance use more little model horizon. For integrating plant, the improved algorithm also can get good control performance.3. The steady-state optimization algorithm that can accomplish all control requirements: hard constraint, soft constraint, economic optimization, overcoming multi-solution is presented comprehensively. The algorithm is presented with a single optimization base on goal programming, so the computation can be simpler.4. The density control scheme, nonlinear transform+model predictivecontrol, for slurry system of PTA plant is proposed. The single-value model predictive control that is so simple that it can run in DCS directly for over-damp plant is presented also.The control system is running for 5 months.5. In the research of advanced control and optimization for solvent dehydration tower, the method that judge current state is optimized or not is presented. Using FrontAPC, the simulation on advance control of the tower is described finally.6. The method of robust PID controller design is studied based on MiniMax rule. Simulation show the robust PID controller can get good controls performance in the scope of model uncertain.