针对内置式永磁电机铁心饱和以及交叉耦合效应对电机性能影响较大的问题,提出采用BP神经网络拟合电机的最大转矩/电流控制曲线.通过对永磁电机动态方程的分析,指出交叉耦合效应不仅影响电机的动态特性,而且改变了其稳态转矩.传统的解析方法只能通过非线性交、直轴自感引入铁心饱和的影响,而无法考虑两者的交叉耦合效应.构建了用于完成电流幅值与最佳相位角之间非线性映射的BP神经网络,计算仿真结果表明,随着定子电流幅值的增加,传统的解析计算的最佳电流相位角与有限元计算结果的偏差也越大,而神经网络的拟合结果与有限元计算结果的最大误差不到1°.%The saturation and cross-coupling effects in interior permanent magnet ( IPM ) motors have great impact on the motor performance. To take both effects into account, a method of fitting the maximum torque per ampere (MTPA) curve by BP neural network was presented. It can be derived from the dynamic equations of IPM motors that the dynamic torque and the steady-state torque are influenced by the cross-coupling effect. However, only the variations of direct- and quadrature-axis self-inductances are normally considered in the traditional analytic methods. A BP neural network was constructed to map the nonlinearity between the current amplitude and the optimal phase angle of current. The computational and simulation results show that the errors between the analytic computational result and the result calculated by finite element method ( FE A) increase with the current amplitude, while the maximum error between the fitting result and the result calculated by FEA is less than 1 degree.
展开▼
