改进灰色模型在变压器故障预测中的应用

Improvement of grey model to forecast the transformer faults

针对传统的灰色GM(1,1)预测模型只适用于对较强指数规律序列进行预测的局限,对传统的GM(1,1)模型进行了改进,通过引入一个m点均值算子,将波动的原始序列生成一个近似指数变化的新序列,并建立等维新息模型,缩小灰平面,从而实现对具有波动性质的序列进行有效的预测.通过对变压器油液的C2H2体积分数预测结果表明,改进GM(1,1)模型对波动的变压器油色谱液数据有良好的逼近效果,且预测精度高于传统的GM(1,1)模型.

Traditional GM（1,1） model could used only in the sequences with more obvious exponential characters. An improved GM（1,1） model is proposed, which makes the fluctuant sequence to a exponential rule sequence by introducing a operator of m points. A new information model with constant dimension was built to reduce the grey plane. Thus the fluctuant sequence can be forecasted well. The forecasted C2 He consistence in transformer＇s oil shows that the forecast sequence of improved GM（1, 1） model has a good approach to the fluctuant sequence of transformer oil chromatogram and its forecast precision is higher than that of traditional GM（1,1） model.