摘要
为了探讨朴素贝叶斯分类在仪表故障判断领域的应用价值,通过将某核电厂压力表故障的历史信息进行分类汇总,将故障的判断转换成文本分类任务,结合朴素贝叶斯分类算法和自然语言处理建立故障的分类模型,实现对新增故障的准确判断。通过验证,朴素贝叶斯分类模型能够对新增故障进行判断分类。测试中需要进行校验类故障准确率能够达到95%以上,其他类故障准确率高于70%。传统故障判断一般是由人来完成,通过贝叶斯分类模型实现对故障的判断,可减轻人员劳动强度,提高工厂维修自动化水平。
In order to investigate the application value of Naive Bayesian classification in the field of instruments fault judgment.By classifying and summarizing the historical maintenance information of pressure gauge fault in a nuclear power plant,the fault judgment is converted into a text classification task.After that the naive Bayesian classification algorithm and natural language processing are used to establish a diagnosis model to achieve accurate judgment of the new faults.It is proven the model can realize the fault judgment function,the accuracy rate which needs calibration can reach above 95%,while the others are above 70%.Traditional faults judgment is generally completed by human beings,while the bayesian classification model can reduce the labor intensity of personnel and improve the level of maintenance automation.
作者
周自强
ZHOU Ziqiang(Liaoning Hongyanhe Nuclear Power Co.Ltd. ,Dalian,Liaoning 116000,China)
出处
《南华大学学报(自然科学版)》
2020年第2期21-24,33,共5页
Journal of University of South China:Science and Technology
关键词
朴素贝叶斯
分类
仪表
故障判断
Naive Bayesian
classification
instruments
fault judgment