关键词: 温度控制, 模型预测, 自适应 PID

Abstract: For a specific temperature system, although most control methods can achieve good control effect,they also increase a lot of parameter adjustment process and other preliminary work. A generally applicable adaptive PID ( Proportion Integration Differentiation) control method combined with model predictive control is proposed. This method includes a simple modeling method. The predicted system output is obtained through the model, and the parameter change of PID controller is guided by the principle of minimizing the sum of square error between the predicted system output and expectation in the future. The method is implemented on the industrial controller PLC(Programmable Logic Controller). It is verified that compared with the traditional PID,due to the parameter adaptation guided by the model prediction, its control accuracy is higher, and its maximum error does not exceed the minimum resolution of the sensor after the output is stable. The convergence speed is faster, the average adjustment time is shortened by about 30% , and the average rise time is shortened by 20% ~40% . It has strong stability, and the maximum dynamic deviation is the same order as PID. Compared with
other model prediction methods, this method gives full play to the robustness of PID method, and has the characteristics of easy implementation and universal application.

Key words: temperature control, model prediction, adaptive proportion integration differentiation (PID)