一种适用于运行风险评估的元件修复时间概率分布  

A Novel Probability Distribution of Component Repair Time for Operation Risk Assessment

作　　者：宁辽逸[1] 吴文传[1] 张伯明[1] 

机构地区：[1]电力系统及发电设备控制和仿真国家重点实验室(清华大学电机系),北京市海淀区100084

出　　处：《中国电机工程学报》2009年第16期15-20,共6页Proceedings of the CSEE

基　　金：国家自然科学基金项目(50707013);国家重点基础研究发展计划项目(973项目)(2004CB217908)~~

摘　　要：确定元件在未来短时段内的瞬时状态概率是电力系统运行风险评估研究中的关键问题之一。首先,通过定性分析和定量计算说明描述元件维修时间的概率分布类型、参数对元件的瞬时状态概率值有较大的影响,并说明可靠性研究中采用指数分布描述维修时间不适用于运行风险评估计算。接着提出维修时间概率分布的选型需要满足的3个必要条件。然后,提出一种新的概率分布——"叠加指数分布",其密度曲线为"铃形",可很好地表征故障维修时间的分布特点,且使瞬时状态概率易于计算。采用现场实际数据的算例结果显示,叠加指数分布能非常好地符合实际样本,得到的元件瞬时状态概率也相对来说最真实。Calculating transient state probability of components in future short time duration is one of the key issues in power systems operation risk assessment. Firstly, the impact of repair time＇s probability distribution on instantaneous state probability of components was evaluated by qualitative analysis and quantitative calculation in this paper, which proves that exponential distribution which is universally used in the reliability research is not suitable to formulate repair time in operation risk assessment. Secondly, three necessary conditions were proposed to choose proper probability distribution for repair time in this paper. Lastly, a novel probability distribution named ＇splice exponential distribution＇ was presented for repair time, which satisfies the above conditions. The shape of probability density of splice exponential distribution likes a bell, which can describe the distribution characteristics of component repair times well, and the calculation of the transient state probability becomes easier. The numerical test result using real-life data showed that the splice exponential distribution matched the real-life data well, and the relevant transient state probability was most factual.

关 键 词：电力系统 运行风险评估 瞬时状态概率 维修时间 概率分布 

分 类 号：TM77[电气工程—电力系统及自动化]