基于故障现象文本的水轮机故障诊断研究
Research on Fault Diagnosis of Hydraulic Turbines Based on Fault Phenomenon Text
刘海洋 1祝迪 2劳鹏飞3
作者信息
- 1. 国能四川大渡河双江口水电开发有限公司,四川 阿坝 624600
- 2. 湖北清江水电开发有限责任公司,湖北 宜昌 443000
- 3. 武汉三联水电控制设备有限公司,湖北 武汉 430200
- 折叠
摘要
为诊断水轮机故障,基于文本挖掘与机器学习技术,提出了一种基于故障现象文本的水轮机故障诊断模型.该模型首先使用 Word2vec将故障文本映射到向量空间,然后将提取的文本特征输入到XGBOOST分类器中进行故障诊断;此外,还使用KmeansSMOTE算法来弥补由于数据不平衡造成的分类误差.在一个真实的数据集中进行了模型性能验证,结果表明所提的故障诊断模型的综合性能优于其他对比模型.
Abstract
This article applies text mining and machine learning techniques to the fault diagnosis of hydraulic tur-bines,and proposes a fault diagnosis model for hydraulic turbine based on fault phenomenon text.Firstly,the model maps the fault text to a vector space using Word2vec.And then,the extracted text features were input into the XG-BOOST classifier for fault diagnosis.In addition,the KmeansSMOTE algorithm was used to compensate for classifica-tion errors caused by data imbalance.The performance of the model was validated on a real dataset.The results show that the proposed fault diagnosis model has better overall performance than other comparative models.
关键词
文本挖掘/故障诊断模型/过采样算法/水轮机Key words
text mining/fault diagnosis model/oversampling algorithm/hydraulic turbine引用本文复制引用
出版年
2024
NETL
NSTL
万方数据