分析1万t多向模锻水压机分配器驱动系统的结构特点,提出--~oo智能超差二次调整控制算法。该算法应用神经网络控制和模糊控制对转角误差进行分段调整,由ε-NN神经网络输出提前量给开关伺服阀断电,采用模糊控制使伺服阀再次得电进行细调。仿真和实际应用结果表明,分配器能不振荡准确到位,转角误差由原来的-9°~9°变为-3°~3°;智能超差二次调整控制算法非常适用于含有带死区的继电特性环节的不确定性位置控制非线性系统,实现无振荡快速高精度位置控制。
The characteristic of driving system for the dispenser used in ten-thousand-ton multidirectional die forging hydraulic press was analyzed. An intelligent control algorithm of two-times regulation of angle error was proposed, which combined neural network control and fuzzy control. Advanced output of neural network control turns off the servo valve switch. Fuzzy control turns on the servo valve switch again for more precise regulation. The simulation and practical results demonstrate that dispenser can track reference angle value quickly and precisely without oscillation; the angle error is reduced from -9°~9°to-3°~3°; the proposed intelligent control algorithm is applicable to the uncertain nonlinear system with relay character element having dead zone for high-speed and high-precision position control without oscillation.