论文部分内容阅读
针对在聚合反应温度控制中的时变、热惯性、热滞后、被控对象模型难以确定等特点,以及常规PID控制在非线性、时变、滞后系统中的局限性,提出了一种单神经元自适应控制与Smith预估器相结合的控制方法。该方法利用神经网络的非线性映射和自学习能力,来实现控制参数的在线整定。利用Smith预估算法来解决系统的大滞后。该控制器具有使系统超调量小、调整时间短、鲁棒性好等优点。文中并给出了基于数字信号处理器(DSP)和单片机的双CPU的温度控制系统的设计。
In view of the characteristics of time-varying, thermal inertia, thermal hysteresis and difficulty in determining the controlled object model in the polymerization temperature control, as well as the limitations of conventional PID control in nonlinear, time-varying and hysteresis systems, a single neural Control Method Based on Metamodel Adaptive Control and Smith Predictor. The method makes use of neural network’s nonlinear mapping and self-learning ability to realize online tuning of control parameters. Use Smith’s predictive algorithm to solve the system’s large lag. The controller has the advantages of small system overshoot, short adjustment time and good robustness. In the paper, the design of dual CPU temperature control system based on digital signal processor (DSP) and singlechip is given.