文章提出了一种基于综合灰色模型的电力变压器油中溶解气体浓度预测的方法,进而预测电力变压器的故障.首先对非等间隔时间序列进行等间隔处理,然后对序列进行级比检验,对于级比检验不合格的序列进行弱化处理,使用GM(1,1)得到预测序列,检验预测序列精度,对精度未满足要求的序列使用残差 GM(1,1)进行残差修正,通过预测某超高压公司电力变压器的油中溶解气体的历史数据,证明了该预测模型具有较强的实用性.%This paper presents a method based on a synthesis gray model to predict the power transformer oil dissolved gas concentrations, then predict the failure of power transformer. First, do equal-space processing to the non-equal-space time series, then carries on level compare examination to the sequence, carries on attenuated processing regarding to the unqualified sequence, after qualified we use GM(l, 1) to obtain the forecast sequence, then exam the forecast sequence's precision, to the sequence that not satisfied the request to the precision, we use residual GM(1,1) to revise residual. Through the forecast of a ultrahigh voltage company's power transformer's oil dissolved gas's historical data, we find this forecast model has strong usability.
展开▼
