融合粗糙集与灰色理论的电力变压器故障预测

Fault Prediction of Power Transformer by Combination of Rough Sets and Grey Theory

结合粗糙集和灰色理论的各自特点,提出一种用于变压器故障预测的新方法。基于粗糙集的知识获取方法,获得改进的三比值诊断决策表,并简化决策表,建立最小诊断规则;分别建立决策表中三比值的灰色预测模型,通过灰色模型对特征气体的比值进行预测,获得预测的特征气体比值的状态特征,对照最小诊断规则,得出预测的故障类型,并结合规则置信度和预测的状态特征对该故障类型的支持数确定其发生概率,这种方法通过结合预测气体比值状态特征和诊断规则,尽早准确地发现潜伏的故障类型,从而可以预先有针对性对变压器进行检修;进行变压器故障预测实例分析,预测结果证明该方法的有效性和正确性。

A new transformer fault prediction method based on rough sets and grey theory was presented. The improved three-ratio attribute decision table was constructed and simplified by the knowledge acquisition method based on rough sets, and the minimal rules were obtained. Then, the prediction models of three ratios in decision table were constructed respectively. The ratios of feature gases could be predicted by grey model and their future state feature was obtained. According to the minimal rules, the incipient failure could be predicted, and its probability was acquired by combination rules＇ credibility with the number of the failure acquired from predicted feature of gases＇ ratios. The method can detect incipient failure early and correctly by combination predicted future state feature of gases＇ ratios with minimal rules According to predicted incipient failure, pertinent repair can be done early. Finally, the effectiveness and correctness of this method were validated by the result of fault prediction examples.