基于遗传程序设计的变压器油中溶解气体含量预测  

Prediction of Gases Content Dissolved in Power Transformer Oil Based on Genetic Programming

作　　者：胡资斌[1] 朱永利[1] 董卓[1] 

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003

出　　处：《电力科学与工程》2011年第7期7-11,共5页Electric Power Science and Engineering

摘　　要：为了对变压器的运行状态和潜伏性故障进行有效预测,提出了基于遗传程序设计(GP)的变压器油中溶解气体浓度的预测方法。结合变压器运行实例,分别给出了变压器油中7种主要气体的预测模型,并与灰色模型及其改进形式进行比较。对比分析结果表明,在同为小样本训练数据的情况下,GP比灰色模型的预测精度更高。In order to predicting the operational status and the latent faults of a power transformer effectively. A new method to forecast the dissolved gas＇s concentration in transformer oil based on Genetic Programming is proposed. With a running instance of a power transformer, prediction models are established for Seven major gas in power transformer oil, then compared with the grey model and it＇s improved form. The contrast analysis show that, GP model has higher prediction accuracy than grey model while each model used small sample to train data.

关 键 词：遗传程序 油中溶解气体 浓度预测 

分 类 号：TM711[电气工程—电力系统及自动化]