摘要
随着先进的仪器测量、大数据、人工智能等科学技术的发展,基于全生命周期的设备健康监测诊断对现代企业智能制造将起着越来越重要的作用。应力波传感器连续检测设备运动部件间的摩擦、冲击的电子信号,数据采集箱对信号进行处理、能量计算和幅值分析。借助于应力波能量趋势、应力波振幅直方图、应力波频谱等分析工具,系统对设备状态进行实时监测、预测故障的发生和其变化趋势,大幅度提高了设备运行和维修效率。基于人工神经网络的改进算法,对设备故障进行智能化的准确预测和定量分析,并提供设备健康诊断报告。试验表明,系统具有较强的鲁棒性和自适应能力,能更早地预测到故障的发生,更准确地判定故障的部位和类型,降低了维修成本,提高了运行效率,提高了生产安全性。
With the development of advanced instrument measurement,big data,and artificial intelligence,the equipment health monitoring and diagnosis system based on the life cycle is playing the more and more important role in the modern enterprise intelligent manufacturing. The stress wave sensor continuously detects the electronic signals of the friction and mechanical shock among the moving parts of the equipment. Then,the signal processing,energy calculation and amplitude analysis are utilized to mine the acquired data.With the help of analysis tools,the energy trend,the amplitude histogram and the spectrum of the stress wave can be monitored in real time. The occurrence of fault and its change trend can be predicted. In this way,the efficiency of equipment operation and maintenance is enhanced. The improved algorithm based on artificial neural network can predict and analyze the equipment fault intelligently and provide the diagnosis report of equipment health. Experimental results show that the system has strong robustness and self-adaptive ability,predict the occurrence of fault earlier,and determine the location and type of the fault more accurately. The maintenance cost is reduced. The operation efficiency and the safety of production are improved.
作者
吴天舒
陈蜀宇
吴朋
Wu Tianshu;Chen Shuyu;Wu Peng(College of Computer Science,Chongqing University,Chongqing 400044,China;Chongqing Chuanyi Automation Co.,Ltd.,Chongqing 401121,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2018年第8期204-211,共8页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61272399,61572090)
国家工信部智能制造专项(2015 No.82)资助
关键词
应力波检测
特征提取
大数据
人工神经网络
stress wave detection
feature extraction
big data
artificial neural network
