摘要
目的:开展小微企业压力容器安全等级预测,有利于辨识高危设备并采取针对性技术和管理措施。方法:通过杭州地区设备定期检验与等级评定结果的统计,以染色机为例分析影响安全等级的缺陷致因,采用机器学习和岭回归法得出影响安全等级的9项主要特征值,进而采用K临近算法、决策树算法和GBDT算法进行安全等级预测。结果:超标裂纹和使用年限是影响安全等级的最主要因素,决策树算法准确率较高,适合用于安全等级预测。结论:采用机器学习和决策树算法可以较为准确地预测容器安全等级,从而为主管部门、检验机构和企业单位等精准识别高危设备,加强监管、检验和使用管理提供参考。
Aims:The prediction of the pressure vessel safety level in small and micro enterprises is conductive to identifying high-risk equipment and taking targeted technical and management measures.Methods:Based on the statistics of equipment periodic inspection and grade evaluation results of dyeing machines in Hangzhou,we analyzed the causes of defects affecting the safety grade.By using machine learning and ridge regression,9 main characteristic values affecting the safety grade were obtained.Then KNN,the decision tree and GBDT were used to predict the safety grade.Results:Exceeding standard cracks and service life were the most important factors affecting the safety level.The decision tree algorithm had a high accuracy and was suitable for safety level prediction.Conclusions:The machine learning and decision tree algorithms can predict the safety level of containers accurately and provide references for competent authorities,inspection agencies and enterprises to accurately identify high-risk equipment,strengthen supervision,inspection and management.
作者
赵明越
张庆浩
赵哲明
廖晓玲
陈涛
吴琳琳
ZHAO Mingyue;ZHANG Qinghao;ZHAO Zheming;LIAO Xiaoling;CHEN Tao;WU Linlin(College of Quality and Safety Engineering,China Jiliang University,Hangzhou 310018,China;Hangzhou Special Equipment Testing and Research Institute,Hangzhou 310003,China)
出处
《中国计量大学学报》
2021年第4期509-515,共7页
Journal of China University of Metrology
基金
国家重点研发计划项目(No.2018YFC0809000)。
关键词
压力容器
机器学习
预测模型
安全等级
pressure vessel
machine learning
predictive model
security level
