基于朴素贝叶斯的阀门关闭不合格原因评估  

Valve Closing Nonconformance Cause Assessment Based on Naive Bayes

作　　者：杜胜东 DU Shengdong(Datang Northwest Electric Power Test and Research Institute,Xi’an 710018,China)

机构地区：[1]大唐西北电力试验研究院,陕西西安710018

出　　处：《自动化仪表》2022年第6期38-42,共5页Process Automation Instrumentation

摘　　要：通过对某电科院阀门关闭测试报告中汽轮机阀门不合格原因和描述的历史信息进行分类整理,将不合格原因评估转为文本分类任务。建立基于自然语言分词处理与朴素贝叶斯分类算法结合的不合格原因评估模型,实现对新增阀门关闭不合格原因的准确判断。通过验证与对比改进,朴素贝叶斯分类模型能够对新增不合格原因进行评估分类。测试中,控制器和网络相关原因准确率能够达到98%以上,其他原因高于85%。该方法有效结合传统纸质报告信息与贝叶斯分类技术的优势,能够对汽轮机阀门复杂的关闭不合格因素进行有效判断。通过贝叶斯分类模型实现对汽轮机阀门不合格原因的判断,可减少人员依赖、加快评估速度、提升评估准确度,并且对智慧电厂中的智能运维与检修起到一定示范作用。Through string out the historical information of the nonconformance causes and description of steam turbine valves in the valve closing test report of a certain institute of electrical science and technology,the nonconformance causes assessment is turned into a text classification task.The nonconformance cause evaluation model based on natural language word segmentation and naive Bayes classification algorithm is established to accurately judge the cause of new valve closing.Through verification and comparison,the naive Bayes classification model can evaluate and classify the new nonconformance causes.In the test,the accuracy rate of controller and network related reasons can reach more than 98%,and that of other reasons is higher than 85%.This method effectively combines the advantages of traditional paper report information and Bayes classification technology and can effectively judge the complex closing disqualification factors of steam turbine valves.To realize the judgment of nonconformance causes of steam turbine valves through Bayes classification model can reduce personnel dependence,speed up evaluation,improve evaluation accuracy,and play a certain demonstration role for intelligent operation and maintenance in smart power plants.

关 键 词：发电厂 朴素贝叶斯 汽轮机 阀门 故障诊断 

分 类 号：TH-17[机械工程]