茶叶杀青机模糊RBF神经网络PID温控系统设计与试验

Design and Experiment of the Temperature Control System of the Fuzzy RBF Neural Network PID in Tea Fixing Machine

针对滚筒式杀青机温控系统具有时变不确定非线性的特点,采用常规PID控制难于满足控制要求,利用模糊控制的良好收敛性和对模糊量的运算优势,以及神经网络自学习、自适应的特性,将常规PID控制与模糊控制、神经网络结合起来,提出一种基于模糊RBF神经网络的PID控制策略,实现了对PID参数的实时在线整定。MATLAB软件仿真与试验结果表明,模糊RBF神经网络PID控制与常规PID控制相比,系统具有更好的动静态特性和抗干扰性能,温度控制误差在±2℃范围内,能很好地满足茶叶杀青工艺对温度的控制要求,保证了茶叶的杀青质量。

Due to the characteristics of time-varying uncertainty and nonlinearity of the temperature control system of rotary fixing machine, the conventional PID control parameters are difficult to meet the control requirements . Based on the good convergence of fuzzy control, computing advantages of fuzzy quantity and the self-learning and -adapting characteristics of neural network, a PID control strategy combined PID control, fuzzy control and neural network was proposed to achieve real-time online tuning of PID parameters. The simulation and test results of MATLAB software show that the fuzzy-RBF neural network PID control had better dynamic and static characteristics and anti-jamming performance than the conventional PID control. The temperature control error was within ±2℃, which well met the temperature control requirements of the tea fixation process and ensured the quality.