基于遗传算法-人工神经网络的发电机定子铁心补偿预测模型研究

Research on Predictive Model for Compensation of Generator Stator Core Based on GA and Artificial Neural Network

在发电机定子铁心的叠装过程中,现场工程师根据压铅试验及自身经验来修正预补偿表,并实时放置补偿片。由于人为补偿经验及数据无法储存,导致补偿方案及结果不一致,质量无法准确管控。利用遗传算法和人工神经网络建立了发电机定子铁心叠装补偿的智能决策模型。这一模型可利用已积累的人为补偿经验,根据实时压铅数据来优化发电机定子铁心的补偿方案,从而将工程师的发电机定子铁心补偿经验固化,并实现发电机定子铁心生产质量的智能管控。

During the stacking process of the stator core of the generator, the field engineer corrects the pre-compensation table according to the pressure lead test and the self-experience, and places the balancing tab in real time. As human experience and data for compensation can not be stored, the compensation programs may be inconsistent with the outcome so that the quality can not be controlled accurately. An intelligent decision model for compensation of iron core of generator stator during stacking process was established by using genetic algorithm and artificial neural network. This model can be used to optimize the compensation scheme for the iron core of the generator stator by utilizing the accumulated experience of artificial compensation and the pressure lead data in real-time. This approach can solidify the engineer＇s experience on the compensation of the iron core of the generator stator and achieve intelligent control of the production quality of the iron core of the stator.