| 引用本文: | 赵文清,王 强,牛东晓.基于贝叶斯网络的电抗器健康诊断[J].电力自动化设备,2013,33(1): |
| ZHAO Wenqing,WANG Qiang,NIU Dongxiao.Reactor health diagnosis based on Bayesian network[J].Electric Power Automation Equipment,2013,33(1): |
|
| 摘要: |
| 针对电抗器的特高压绝缘、局部过热、振动和噪声这3个问题，采用5级健康状态诊断法，并充分考虑电抗器预防性试验数据及其变化量。首先建立评估电抗器健康状态的分层模型，通过该模型评估电抗器的历史、当前状态，然后通过无偏GM（1，1）模型预测电抗器的未来状态，并确定分层模型中电抗器各个参数的阈值和分值，最终建立基于贝叶斯网络的电抗器健康诊断模型。实例验证了所提模型的正确性和可行性。 |
| 关键词: 电抗器 贝叶斯网络 模型 无偏GM（1，1） 健康诊断 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金资助项目（70671039，61074078）；河北省自然科学基金资助项目（E2009001392）；中央高校基本科研业务费专项资金资助项目（12MS121）；山西省电力公司科技项目（XZGDKJ2012005） |
|
| Reactor health diagnosis based on Bayesian network |
|
ZHAO Wenqing1,2, WANG Qiang1, NIU Dongxiao2
|
|
1.School of Control and Computer Engineering，North China Electric Power University，Baoding 071003，China;2.School of Economics and Management，North China Electric Power University，Beijing 102206，China
|
| Abstract: |
| Five-level health diagnosis method is applied to detect three reactor problems：UHV insulation，local overheating and vibration & noise，which considers the preventive test data of reactor and their variations. A layered model is established for evaluating the historic and current health status of reactor，and its future status is predicted with the unbiased GM（1，1） model. The threshold value and score of each reactor parameter are determined and the reactor health diagnosis model based on Bayesian network is finally established. Its validity and feasibility are verified by examples. |
| Key words: electric reactors Bayesian networks models unbiased GM（1,1） health diagnosis |
