基于GRNN的永磁同步电动机直接转矩控制

摘要

研究了一种基于泛化回归神经网络(GRNN)的永磁同步电动机(PMSM)直接转矩控制(DTC)策略.在策略中,利用设计的 GRNN 控制器替换传统 PMSM-DTC 转矩及磁链滞环控制器和开关状态选择表,通过 GRNN 对输入数据不断优化选择能力,它可以有效地解决由于传统 Pang-Pang 型滞环控制器引起的系统转矩脉动过大问题.同时,通过仿真验证了基于新型 GRNN 控制器的 PMSM-DTC 系统的控制性能.仿真结果表明:通过与常规的 PMSM-DTC仿真结果相比较,新策略仍然具有较好的快速动态特性,尤其是系统稳态性能得到了明显改善.%A novel generalized regression neural network based PMSM-DTC intelligence implementation scheme was investigated in the paper. With the proposed GRNN controller,the voltage vector switching table in the conventional PMSMDTC scheme was improved. At the same time, the torque ripple is largely reduced and the dynamic performance is kept as good as that of the original scheme. By simulation model, the comparison between the basic DTC and the proposed DTC were carried out to verify the proposed scheme. Simulation results show that the proposed strategy has a much better steady state performance while keeping a good dynamic performance.