| There exist various constraints in real control systems. Constraint on manipulated variable and output variable is very common in complex industrial process. The research on the design of control system without constraint is mature. It is very difficult to apply the method on unconstraint control system to constraint control system. So, it is necessary to research control algorithm and design of control system with constraint.Model predictive control is a receding horizon control strategy and need no accurate mathematic expression of the object, which provides a systematic approach to handle the constraints. At the same time, predictive control is easily to cooperate with other advanced control algorithms, hence it show magnificent future in the application to the control of complex industrial process.A ubiquitous problem in control is that all real actuators have limited authority. Saturating nonlinear actuators is one of the common nonlinearities, and exists in many practical situations. Because of saturation, it is more difficult to analyze and design practical control system. For example, saturation on actuator leads to invalidation of manipulation due to drive locked. It has seen a great progress in the theory of saturation control system recent years, it needs to study it deeply in practical application. Based on the survey for research status, the research work has been expanded. The main results are given as follows:(1)Constraint types and methods for dealing with constraint are summarized and a survey for developing history and research status of design for constraint control system is provided.(2)Structure and principle of IMC is introduced. PID and IMC override control is studied as well as instruction for their strongpoint and defect separately. Because both of PID and IMC override control don't readily extend to general, multivariable, constraint control, constraint cascade controller is introduced in paper. The principle and design method of cascade constraint control are analyzed and studied. The output and control effort response of cascade constraint are identical to those of PID and IMC override control, but the cascade constraint control method is greater generality and limiting the setpoint to the control loop for the constrained output is conceptually simpler than override logic.(3)Based on study for MPC algorithm deeply, a new constrained MPC algorithm is proposed as well as the application of the algorithm on heat exchange system is validated. By studying and researching the disadvantage encountered in MPC algorithms, the paper introduces an advanced MPC algorithm which can deal with constrained process validly. At the same time, the paper introduce how constraints arise in process control systems and show how constraint handling can be integrated into the single-objective model-predictive control strategy. The formulation of constrained MPC can lead to situations in which there is no feasible solution to the QP generated. It allows some small constraint violation rather than declares infeasibility and shut the control system, which is called constraint softening. This approach would allow some constraint violation but the extent of the violation can be kept small by suitable choice of the weighting matrices used in the objective function. The sumulink of practical process supports the validity of constraint softening and using the constrained MPC can improve the performance of control system.(4) Actuator saturation will arouse the degradation in performance of industry process. It is an important research direction to design anti-wind-up controller. Based on the study for saturation performance of actuator, a method for designing of constrained controller is proposed, which uses the nonlinear implementation for the controller with anti-wind-up protection and uses slew-rate limitation as compensation. At the same time, to validate the design, we try to same mathematic and theory analysis. Then, we use an example to indicate the advantage of using the nonlinear compensation. |
PDF Full Text Request