基于神经网络与模糊逻辑的直接转矩控制系统

实现直接转矩控制 (DTC)的关键部件是定子磁链观测器和逆变器状态开关选择器 ,用训练好的神经网络取代U I N模型 ,模拟传统异步电动机DTC系统定子磁链观测器 ;把模糊控制算法引入状态开关选择器 ,用较少的模糊规则实现磁链误差、转矩误差和磁链角对逆变器开关的分级控制 .仿真结果表明 ,基于神经网络与模糊逻辑控制方法建立的直接转矩控制系统 ,具有较强的鲁棒性和良好的动态性能

Direct torque control (DTC) system based on neural network and fuzzy logic

The key components in DTC system are a stator flux observer and a inverter state selector. The stator flux observer in the DTC system of a conventional asynchronous motor is simulated by learned neural network replacing U-I-N model. Fuzzy control algorithm is introduced in the inverter state selector, the hierarchical control of the inverter state is implemented by the stator flux linkage error, and torque error and flux linkage angle and less fuzzy control rules are used. Simulation result shows that the DTC system based on neural network and fuzzy logic has a strong robust and good dynamic performance.