基于小样本性能数据的电气设备可靠性评估与预测方法研究  

Research on Assessment and Prediction of Electrical Equipment Reliability Based on Small Sample Performance Data

作　　者：周凯[1] 丁坚勇[1] 田世明 卜凡鹏 朱天曈 ZHOU Kai;DING Jianyong;TIAN Shiming;BU Fanpeng;ZHU Tiantong(School of Electrical Engineering, Wuhan University, Wuhan 430072, Hubei Province, China;Beijing Key Laboratory of Demand Side Multi-Energy Carriers Optimization and Interaction Technique （China Electric Power Research Institute）, Haidian District, Beijing 100192, China)

机构地区：[1]武汉大学电气工程学院,湖北省武汉市430072 [2]需求侧多能互补优化与供需互动技术北京市重点实验室(中国电力科学研究院有限公司),北京市海淀区100192

出　　处：《电网技术》2018年第6期1967-1973,共7页Power System Technology

基　　金：国家863高技术基金项目(2015AA050203)~~

摘　　要：提出了基于电气设备小样本性能数据的可靠性评估与预测模型。首先采用改进的Bootstrap法对小样本性能数据进行扩容,并进一步分析了设备性能变化的分布函数。在此基础上提出了设备性能可靠性的概念,并由此计算出设备在不同时刻下的可靠性。然后基于形态相似距离对上述可靠性值进行了曲线拟合,最后基于曲线预测值与预测误差对设备的可靠性进行预测估计。该方法不仅在数据扩容上优势明显,而且还能抑制＂野值点＂对拟合结果的影响。算例分析结果说明了所提方法在小样本数据条件下进行可靠性评估与预测的有效性与实用性。This paper presents a reliability evaluation and prediction model based on electrical equipment performance data of small sample. Firstly, an improved bootstrap method is used to expand the performance data of small sample to analyze distribution function of the performance. Based on this, definition of equipment performance reliability is proposed to calculate the reliability in various time instants. Then reliability curves are fitted based on morphological similarity distance. This method has obvious advantages in data expansion, and can inhibit influence of ＂abnormalities＂ on fitting results. Finally, equipment reliability is estimated based on predicted value and prediction error. Case study results show effectiveness and practicability of the proposed method based on small sample data.

关 键 词：小样本性能数据 BOOTSTRAP 可靠性评估预测 形态相似距离 

分 类 号：TM73[电气工程—电力系统及自动化]