摘要
以磨粒分析为核心的铁谱技术已成为机械装备健康评估的关键手段。然而,受限于复杂、恶劣工况,机械装备磨损状态评估始终处于机理难辨识、严重程度难判断的困境。针对此问题,通过构建深度学习的磨粒链智能分割方法实现磨损宏观统计特征自动提取,结合磨粒全信息表征与BP神经网络建立多层磨粒类型辨识模型,实现摩擦副微观磨损机理的精准辨识。在此基础上,结合宏观统计特征与微观磨损机理共同建立机械装备磨损状态评估准则。结果表明:构建磨损综合评价体系能有效判断机械装备磨损严重等级,为机械装备预防性维护提供了关键指导信息。
Ferrography technology with wear particle analysis as the core has become a key method of mechanical equipment health assessment.However,due to the complex and severe working conditions,the mechanical equipment was in the wear state,and the evaluation mechanism was always difficult to identify and the severity was difficult to judge.To solve this problem,an intelligent segmentation method of wear particle chain based on deep learning was constructed to automatically extract the macro statistical characteristics of wear,and a multi-layer wear particle type identification model combined with the full information representation of wear particles and BP neural network was established to realize the accurate identification of the micro wear mechanism of friction pairs.On this basis,combined with the macro statistical characteristics and micro wear mechanism,the wear condition evaluation criteria of mechanical equipment were established.The test results show that the wear comprehensive evaluation system can effectively judge the wear severity of mechanical equipment,and provide key guidance information for preventive maintenance of mechanical equipment.
作者
周俊丽
周倩
ZHOU Junli;ZHOU Qian(China Energy Group Shendong Coal Group Quality and Technology Testing Center,Ordos 017209,China;The 39th Research Institute of China Electronics Technology Group Corporation,Xi′an 710065,China)
出处
《兵器材料科学与工程》
CAS
CSCD
北大核心
2022年第5期175-182,共8页
Ordnance Material Science and Engineering
基金
中国神华能源股份有限公司神东煤炭分公司项目(201391548017,201491548044)。
关键词
铁谱分析
磨粒链分割
磨粒辨识
磨损状态评估
ferrography analysis
abrasive chain segmentation
abrasive particle identification
wear state evaluation