预测性维护算法混合多域测评技术研究
Research on Hybrid Multi-Domain Evaluation Technology for Predictive Maintenance Algorithms
王成城 1王金江 2王凯 3田琨 4张来斌2
作者信息
- 1. 中国石油大学(北京)安全与海洋工程学院,北京 102249;机械工业仪器仪表综合技术经济研究所,北京 100055
- 2. 中国石油大学(北京)安全与海洋工程学院,北京 102249;国家市场监督管理总局重点实验室(油气生产装备质量检测与健康诊断),北京 102249
- 3. 中国石油大学(北京)安全与海洋工程学院,北京 102249
- 4. 中国石油大学(北京)安全与海洋工程学院,北京 102249;中国石油集团安全环保技术研究院有限公司,北京 102206
- 折叠
摘要
预测性维护(PdM)是智能制造背景下的运维模式,体现了人工智能技术的应用.PdM广泛应用于工业自动化领域,通过故障诊断和预测开展自动化设备的运维,以有效降低设备停机时间、提高生产效率.但目前PdM技术的应用缺少性能评估的指标体系,存在无法判断算法优劣的问题.通过回顾PdM性能评估指标体系的相关研究可知,部分学者提出了初步的评估指标体系,但由于定量指标缺少测试技术的支撑而导致该体系难以推广和实施.采用可被标准化的在线与离线融合评估指标体系,实现了PdM算法的有效测评.通过对多类型设备故障诊断算法进行定量指标测试验证,有效评估了指标体系的适用性.该研究为工业自动化仪表及装备的故障诊断和寿命预测算法测评提供了新思路与新方法.
Abstract
Predictive maintenance(PdM) is an operation mode in the content of intelligent manufacturing,which reflects the application of artificial intelligence technology.PdM is widely used in the field of industrial automation to carry out the operation and maintenance of automated equipment through fault diagnosis and prediction,in order to effectively reduce the downtime of the equipment and improve the production efficiency.However,the current application of PdM technology lacks a performance evaluation index system,and there is the problem of not being able to judge the advantages and disadvantages of the algorithm.By reviewing the related research on PdM performance evaluation index system,it can be known that some scholars have proposed a preliminary evaluation index system,but the lack of testing technology to support the quantitative indexes makes it difficult to promote and implement the system.The effective measurement of PdM algorithms is achieved by adopting an online and offline fusion evaluation index system that can be standardized.The applicability of the index system is effectively evaluated by testing and validating the quantitative indexes on multiple types of equipment fault diagnosis algorithms.This research provides new ideas and new methods for the evaluation of fault diagnosis and life prediction algorithms for industrial automation instruments and equipment.
关键词
预测性维护/算法测评/诊断和预测/在线测试/离线测试/工业自动化Key words
Predictive maintenance(PdM)/Algorithmic evaluation/Diagnosis and prediction/Online testing/Offline testing/Industrial automation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据