牛国成1,2,胡贞1,胡冬梅2.基于SVM与物元信息熵的变压器健康度分析与预测[J].湖南大学学报:自然科学版,2019,(8):91~97
基于SVM与物元信息熵的变压器健康度分析与预测
Analysis and Prediction of Transformer Health Index Based on SVM and Matter Element Information Entropy
  
DOI:
中文关键词:  变压器  光声光谱  复合物元  AHP  关联熵  健康度  支持向量机
英文关键词:power transformer  photoacoustic spectroscopy  complex matter element  Analytic Hierarchy Process(AHP)  correlation entropy  health index  Support Vector Machines(SVM)
基金项目:
作者单位
牛国成1,2,胡贞1,胡冬梅2 (1. 长春理工大学 电子信息工程学院吉林 长春130022 2. 北华大学 电气信息工程学院吉林 吉林 132021) 
摘要点击次数: 21
全文下载次数: 20
中文摘要:
      为实现变压器运行状态的定量分析和预测,提出了利用变压器中溶解气体结合变压器典型故障类型建立变压器健康度的立体交叉复合物元,分别利用层次分析法AHP(Analytic Hierarchy Process)和信息熵值法确定影响变压器健康度的主、客观权重,利用物元-最大信息熵来定量分析变压器健康度.提出了利用支持向量机SVM( Support Vector Machines)预测变压器未来的运行状况,采用交叉验证的网格搜索法(K-fold)、遗传算法 (Genetic Algorithm GA)和粒子群算法 ( Particle Swarm Optimization PSO)优化支持向量机的参数,建立最佳预测模型,该方法为变压器的故障排除、检修决策和在线预估提供了数据支持.
英文摘要:
      In order to realize the quantitative analysis and prediction on the operation state of the transformer, the interchange complex matter element was built between dissolved gases in transformer oil and typical faults. Analytic Hierarchy Process (AHP) and maximum information entropy were used to determine the subjective and objective weights influencing the transformer health level, respectively. The quantitative analysis of the transformer health level was proposed based on matter element maximum information entropy. The Support Vector Machines (SVM) algorithm was adopted to predict the operation condition of transformers, the parameters (c and g) were optimized by grid-search, Genetic Algorithm (GA) and Particle Swarm Optimization(PSO),and the optimal prediction model was established. This method provides a good guiding value for the elimination of transformer faults, overhaul decisions and online predictions.
查看全文  查看/发表评论  下载PDF阅读器
关闭