Research On Improved Model Predictive Control And PID Control Optimized By Model Predictive Control

Predictive control is a class of model-based optimal control algorithms which appeared in 1970s, which has great potential in the process of complex industrial process optimization control. On the one hand, it is convenient to build a model. On the other hand, it has excellent tracking performance and strong robustness. With the development of science and technology, the complexity of the process of industrial production is more and more high. At the same time, with the development of society, the market competition is becoming more and more fierce. Further research on predictive control method with higher quality is of great significance.PID control is the most widely used controller in the industrial process, which has great vitality. The main reason is that PID control is simple, high reliability, wide adaptability and easy implementation advantages. But with the development of the industry, the complexity of the object continues to deepen, especially for complex systems, time-varying uncertainty, large time delay. Conventional PID control is difficult to meet the control performance requirements. If we can design a controller combining the advantages of predictive control and simple structure as PID control, this will be helpful to improve the production efficiency.The main contents of this paper include:(1) Combining the PID with the multivariable generalized predictive control, an improved multivariable GPC algorithm named multi variable PIDGPC algorithm is proposed. According to the structure of incremental PID, the objective function of GPC is improved, and a multi variable GPC algorithm with PID structure is derived. This algorithm combines the feedback structure of PID and the predictive function of multi variable GPC, and it does not increased the amount of computation compared with the multi variable GPC. Ball mill pulverizing system MATLAB simulation results show that multi variable PIDGPC algorithm can achieve good control effect in the condition of model matching and model mismatching. Compared to the multivariable GPC, proposed algorithm is with the advantages of no overshoot, model adaptability, stronger anti-interference performance, and its quality has a certain degree of improvement.(2) Combining fractional PID and multivariable dynamic matrix control algorithm, this paper proposes an improved multi variable DMC algorithm named multi variable fractional order PIDDMC algorithm. According to the structure of incremental fractional PID, the objective function of DMC is improved, and a multi variable DMC algorithm with fractional PID structure is derived. The robustness and stability of multivariable fractional PIDDMC algorithm is proved by Lyapunov second stability theorem. MATLAB simulation further demonstrate the effectiveness of the proposed algorithm.(3) The parameters of PID controller are optimized by GPC, a PID algorithm based on GPC optimization algorithm named GPC-PID is proposed. This algorithm keeps the simple structure like traditional PID, and it has excellent control quality as GPC. The simulation results of circulating fluidized bed temperature show that GPC-PID compared to the traditional PID control, the control quality is greatly improved, and proposed algorithm is with the advantages of set point tracking time, no overshoot, adaptability and strong ability of suppressing interference.