CAJ | 학술논문

代替传统化学方法,研究采用局部放电来评估变压器油纸绝缘的老化状态具有理论意义,并且在变压器滤油后也存在实际应用价值。文章在单因子加速热老化试验的基础上测量了不同老化阶段试品的局部放电,并利用因子分析方法从局部放电29个统计特征参数中提取了10个主成分因子向量。引入单位圆描述了油纸绝缘的老化状态,建立了以老化半径R为单一目标的三层BP神经网络评估模型。分别采用遗传算法GA和LM算法对网络进行了训练,并对测试样本进行了老化状态识别,结果表明GA-BP网络较LM-BP网络对老化具有更高的识别率。
대체전통화학방법,연구채용국부방전래평고변압기유지절연적노화상태구유이론의의,병차재변압기려유후야존재실제응용개치。문장재단인자가속열노화시험적기출상측량료불동노화계단시품적국부방전,병이용인자분석방법종국부방전29개통계특정삼수중제취료10개주성분인자향량。인입단위원묘술료유지절연적노화상태,건립료이노화반경R위단일목표적삼층BP신경망락평고모형。분별채용유전산법GA화LM산법대망락진행료훈련,병대측시양본진행료노화상태식별,결과표명GA-BP망락교LM-BP망락대노화구유경고적식별솔。
It is theoretically significant to study on the aging condition assessment of oil-paper insulation based on partial discharge （PD） instead of traditional chemical methods. Moreover,practical application value still exists after refining transformer oil. In this paper,PD of the test samples with different aging degrees was measured after a thermally accelerated aging experiment. Then ten principal component factors were extracted from 29 statistical parameters. The aging condition of oil-paper was characterized utilizing a unit circle. Using the aging radius R as the single objective,an assessment model was established based on a three-layer BP neural network which was trained by genetic algorithm （GA） and Levenberg-Marquardt （LM） algorithm respectively. Finally,the test samples were recognized by the trained GA-BP and LM-BP network. It has been found that GA-BP network has the advantages of higher recognition rates comparing with LM-BP network.