基于支持向量机回归的智能交流接触器PWM分时控制参数优化

PARAMETER OPTIMISATION OF PWM TIME-SHARED CONTROL FOR INTELLIGENT AC CONTACTOR USING SUPPORT VECTOR MACHINE FOR REGRESSION

智能交流接触器控制模块PWM占空比的选取要依据设计者经验,造成PWM占空比组合不是最优的,影响接触器吸合动态特性。针对这种情况,提出PWM分时控制策略及支持向量机回归参数优化方法。在不同PWM分时控制组合下测试智能交流接触器动态特性,根据实验数据采用支持向量回归机方法建立PWM控制模型并优化控制参数,应用该模型对预测样本进行评估分析。实验结果表明,PWM控制模型能准确反映接触器吸合动态参数与PWM控制组合之间的内在规律性,对智能交流接触器控制方案的优化设计具有参考价值。

The selection of PWM duty cycles of intellectual AC contactor control module is according to the experience of designers, and this results in the combination of PWM duty cycles not being optimal, and affects the dynamic pull-in characteristics of the contactor as well. In light of this case, we presented in the paper the time-shared control strategy of PWM and the SVM regression parameter optimisation method. We tested the dynamic characteristics of intelligent AC contactor with different combinations of PWM time-shared control. According to experimental data we employed the method of support vector machine for regression to build PWM control model and to optimise the control parameters, and applied the model to assess and analyse the prediction samples. Experimental results indicated that the PWM control model could precisely reflect the inherent regularity between dynamic pull-in parameters of the contractor and PWM controls combination. This has the reference value on optimising the design of intelligent AC contactor control scheme.