摘要
针对传统直接转矩控制(DTC)系统由于进行复杂运算时延而降低系统性能和不便于硬件实现的缺陷,介绍了一种新的基于BP神经网络的电压矢量控制器以取代常规的状态选择器。控制器的输入信号为电磁转矩误差、定子磁链、正反转信号和区间号,通过所设计的BP神经网络加以映射,得到逆变器的开关状态输出信号。仿真实验结果验证了新型电压矢量控制器设计的正确性。与传统DTC系统的仿真结果对比,表明所设计的矢量控制器能有效减少转矩脉动,提高系统性能,降低传统状态选择器硬件实现的复杂性,并且具有较强的鲁棒性。
Aiming at the traditional direct torque control（DTC） ＇s time delay and difficult application by hardware caused by the complex calculation, this paper presents a novel control for selecting the voltage space vector based on BP neural network. This controller uses electromagnetism torque error, stator flux linkage, normal-reverse signal and interval signal as the input, through the designed BP neural network, we could obtain the inverter＇ switch-status output signal. The simulation test configures the designed controller＇ correctness. Compared with the hardware＇ complexity, and has a good robustness.
出处
《微电机》
北大核心
2008年第1期21-23,30,共4页
Micromotors
