综合灰色模型在电力变压器故障预测中的应用

Application of Comprehensive Gray Model in Fault Prediction of Power Transformers

通过改进的综合灰色模型对电力变压器油中溶解气体浓度进行预测,进而对电力变压器的故障进行预测,为电力变压器状态检修工作的顺利开展提供数据支特。该模型在传统灰色模型的基础上加以改进,对原始数据弱化处理以减少其随机性,通过残差修正模型提高预测精度。利用此模型在某供电公司220 kV主变压器的油中溶解气体历电数据的基础上进行预测,发现预测结果与油中气体分解物含量的发展趋势基本一致,证明该预测模型具有较强的实用性。

Tile paper proposes a comprehensive gray model ba^d method to predict the dissolved gas concentration in power transformer oil and then the potential fault in the power transformer can be predicted. Firstly, the non-equal interval time sequence is converted into equal interval. Secondly, class ratio test is applied to the sequence, and weakening treatment is used for the unqualified sequence. Finally, the GM （1, 1） method is used to calculate the sequence used for fault prediction. The precision test of the calculated sequence is carried out and residual modification is made for the unqualified sequences. The result shows the model is of high practicality through predicting the dissolved gas concentration in the power transformer oil based on the historical data from a power supply company.