文本信息挖掘技术及其在断路器全寿命状态评价中的应用

发布时间：2018-04-12 21:33

本文选题：全寿命状态评价 + 检修消缺　；参考：《电力系统自动化》2016年06期

【摘要】：电网企业记录了大量故障与缺陷中文文本,这些文本蕴藏了丰富的设备健康信息。但迄今为止,鲜有电力领域的文本信息挖掘技术研究。以断路器全寿命状态评价为应用研究背景,探索了电网中文文本挖掘方法。首先,根据断路器状态评价的研究现状,提出了构建文本挖掘与全寿命状态评价模型的关键问题。然后,构建了包含文本挖掘信息的全寿命状态评价模型,通过基于隐马尔可夫法(HMM)的文本预处理与向量化、自主区间搜索k最近邻(KNN)算法的文本分类和比率型状态信息融合模型完成了断路器全寿命健康状态指数的展示。最后,采用某电网公司实际缺陷文本构建算例。算例表明,文本挖掘技术实现了相似缺陷的相关性学习,比率型信息融合模型能更全面真实地展示健康状态评价的历史流。
[Abstract]:Power grid enterprises record a large number of fault and defect Chinese text, which contains a wealth of equipment health information.But so far, there are few research on text information mining technology in power field.Based on the life state evaluation of circuit breakers, a Chinese text mining method is explored.Firstly, according to the status quo of circuit breaker status evaluation, the key problems of constructing text mining and life state evaluation model are put forward.Then, a whole life state evaluation model including text mining information is constructed, and text preprocessing and vectorization based on Hidden Markov method (HMMM) is proposed.The text classification and ratio state information fusion model of autonomous interval search k nearest neighbor KNN algorithm are used to display the lifetime health state index of circuit breakers.Finally, the actual defect text of a power grid company is used to construct an example.Numerical examples show that text mining technology can realize the correlation learning of similar defects, and the ratio information fusion model can show the historical flow of health status evaluation more comprehensively and truly.
【作者单位】：浙江大学电气工程学院;国网金华供电公司;国网浙江省电力公司电力科学研究院;
【基金】：国家电网公司科技项目~~
【分类号】：TP391.1

【参考文献】