基于多源状态参量融合的油纸绝缘套管缺陷预警方法研究

Research on the Defects Early Warning Method of Oil-Paper Bushing Based on Multi-Source State Parameter Fusion

油纸绝缘套管是大型电力变压器不可或缺的外部连接组件,若其发生故障影响巨大,有必要对油纸绝缘套管的缺陷进行及时预警。结合上百例油纸绝缘套管故障案例及专家经验,构建了基于多源状态参量融合的套管缺陷预警信息决策表,并结合遗传优化算法对决策表进行约简,实现了油纸绝缘套管决策规律的提取,形成套管缺陷预警知识,最终建立了套管缺陷预警模型。经实际应用,模型可实现对油纸绝缘套管的缺陷预警。该研究建立的油纸绝缘套管缺陷预警模型可随着数据量的积累与丰富,实现迭代优化,挖掘隐含的套管缺陷预警知识。研究成果可为高压套管智能化运维提供理论依据。

Oil-paper bushing is an indispensable external connecting component of large power transformers,which has a great impact if a fault occurs.It is necessary to timely warn the defects of the oil-paper bushing.Based on hundreds of oilpaper bushing failure cases and expert experience,this paper constructs a casing defect early warning information decision table based on multi-source state parameter fusion,and reduces the decision table with genetic optimization algorithm.Further,the decision rules of oil-paper bushing are extracted and bushing defects early warning knowledge are formed.Finally,the bushing defects early warning model is established.Through practical application,the model can realize the defect early warning of oil-paper bushing,and the positive judgment rate is not less than 80%.With the accumulation and enrichment of data,the oil-paper bushing defects early warning model established in this paper can realize iterative optimization,and mine the hidden bushing defects early warning knowledge.The research results can provide a theoretical basis for the intelligent operation and maintenance of high voltage bushing.