【摘要】 永磁直线同步电动机(Permanent magnet linear synchronous motor,PMLSM)齿槽力是影响电动机性能的主要因素之一,特别是在高精度,低速情况下,问题尤为突出。根据q<1分数槽绕组PMLSM结构,采用有限元法计算齿槽力的影响,建立以径向基神经网络为基础的PMLSM齿槽力预估器,其学习算法首先采用快速模糊C均值算法(Accelerated fuzzy C-means,AFCM)对数据进行聚类,选取基函数传播因子,再由最小正交平方算法(Orthogonal least squares learning algorithm,OLSA)选取中心矢量,该预估器与带动量的BP网络(Back propagation neural net- work,BPNN)预估器相比较表明,能够在加快网络学习速度的前提下,保证精度,缩小网络规模,提高网络分类能力。试验结果表明,采用q<1分数槽绕组PMLSM能够有效地减小齿槽力的影响。预估器的建立,能够在设计阶段对PMLSM齿槽结构参数进行有效地预估,使得电动机在满足推力波动指标条件下,实现快速敏捷设计,提高PMLSM的整体设计水平。更多还原

【Abstract】 The cogging force is of great impact to the efficiency of permanent magnetic liner synchronous motor(PMLSM) especially in high precision and low speed. According to the fractional slot with q<1 structure of PMLSM, FEM is used to analyze the influence of cogging force. Supposed estimator based on radical basis functional network(RBFN) is presented by improved algorithm. To select the right spread factor of base function, the accelerate fuzzy C-means(AFCM) is used in data clustering. Then, OLSA is used to choose the center vector from the clustering center. Comparing to the estimator based on back propagation neural network(BPNN) with momentum method, the novel estimator increases the clustering of neural network with boosting learning rate. Results show the fractional slot with q <1 structure effectively reduces the influence of cogging force in PMLSM. Through the estimator based on RBFN, the parameters of the PMLSM can be evaluated in the design period. By satisfying the standards of cogging force ripple, the estimator achieves the agility demand and improves the design level of PMLSM.更多还原