摘要
提出了一种将直接转矩控制技术和RBF神经网络自整定PID相结合的新方法。系统外环的输入为速度误差及输出转矩,通过神经网络控制器,输出目标转矩,送入60 kW四相8/6结构的开关磁阻电动机直接转矩调速系统内环。仿真结果表明,这种控制方法不仅解决了常规控制方法因电机数学模型难以精确确定而导致无法确定控制参数的问题,且具有上升时间短、超调小、抗干扰能力强的优点。
A control which combined direct torque control and RBF neural network-tuning PID was proposed. Input for the system was the speed error and the outer torque, through the neural network controller, and the output was target torque which gave a four-phase 8/6 structure switched reluctance motor whose capacity was 60 kW. Siumlation results show that the problem is solved that the control parameters can not be determined by common control method because motor mathematical model is difficult to be determined. This method has short rise time,the small overshoot and strong anti-interference.
出处
《微特电机》
北大核心
2012年第1期54-57,共4页
Small & Special Electrical Machines
