基于深度学习的CFRP/TC4叠层结构制孔刀具磨损状态监测OA北大核心CSTPCD
Tool Wear Condition Monitoring Based on Deep Learning During Drilling CFRP/TC4 Laminated Structure
碳纤维复合材料(CFRP)和钛合金(TC4)因各具有优良的物理力学性能,其叠层结构广泛应用于航天工业领域.由于CFRP和TC4都属于典型难加工材料,且具有不同的机械和热学特性,因而在制孔过程中,刀具磨损较快,从而影响加工质量.为了保证钻孔质量、及时更换刀具,建立了一种基于卷积神经网络-长短期记忆网络(CNN-LSTM)的刀具磨损状态监测模型.该模型以与刀具磨损相关性较强的力、声发射信号特征作为输入,以刀具磨损状态标签作为输出,从而实现刀具磨损状态的监测.结果表明,该模型识别准确率高达97.222％,可以很好地实现CFRP/TC4叠层结构制孔过程中刀具磨损状态的监测.
Due to excellent physical and mechanical properties,carbon fiber reinforced plastics(CFRP)and titanium alloys(TC4)were often widely used in the aerospace industry as laminated structures.Since CFRP and TC4 were both typical difficult-to-machine materials,and had different mechanical and thermal properties,the tool wear was rapid during the hole-making process,which affected the machining quality.In order to ensure the quality of drilling and timely replacement of cutting tools,a tool wear condition monitoring model based on convolution neural network-long short time memory(CNN-LSTM)was established.The model took the feature of force and acoustic emission signals with strong correlation to tool wear as input and the tool wear condition labels as output to realize tool wear monitoring.The results show that the model has an accuracy rate of 97.222％,which can effectively monitor the tool wear status during the drilling process of CFRP/TC4 laminated structures.
江庆泉;李鹏南;邱新义;李树健;王春浩
湖南科技大学机电工程学院,湘潭 411201湖南科技大学机电工程学院,湘潭 411201湖南科技大学机电工程学院,湘潭 411201湖南科技大学机电工程学院,湘潭 411201湖南科技大学机电工程学院,湘潭 411201
金属材料
刀具磨损状态监测CFRP/TC4叠层特征提取深度学习
Tool wear condition monitoringCFRP/TC4 laminated structuresFeature extractionDeep learning
《宇航材料工艺》 2024 (5)
40-49,10
国家自然科学基金(52275423,52105442,51975208)
评论