【摘要】 机械设备正朝着大型化、精密化和自动化的方向发展,机械系统也因此变得越来越复杂。考虑到机械系统可能会发生无特征的灾难性故障,因此机械故障的自动检测是一个巨大的挑战。然而,现有的故障检测方法在对高度复杂的工业系统进行故障类型识别时,误诊率较高,无法给出准确的故障诊断结果。针对这一问题,本文以滚动轴承这一机械设备关键部件作为研究对象,提出一种基于混合域残差注意力网络的故障诊断方法,旨在结合深度卷积神经网络自动学习表示的优点,并配合通道注意力机制和空间注意力机制的关键特征提取能力,提高故障检测性能。实验结果表明,所提出的方法能够准确地检测轴承故障类型,在准确度指标方面优于其他方法。更多还原

【Abstract】 Mechanical equipment is developing in the direction of large-scale, precision and automation, and mechanical systems are becoming increasingly complex. The automatic detection of mechanical faults is a great challenge considering that those mechanical systems may suffer from featueless catastrophic failures. However, existing fault detection methods have a high misdiagnosis rate when identifying fault types in highly complex industrial systems, and cannot produce accurate fault diagnosis results. To solve this problem, a new fault diagnosis method based on mixed domain residual attention network is proposed in this paper, which takes rolling bearing as the research object as it is the key component of mechanical equipment. This paper aims to improve the performance of fault detection by combining the advantages of automatic learning representation of deep convolutional neural network and key feature extraction ability of channel attention mechanism and spatial attention mechanism. Experimental results show that the proposed method can accurately detect bearing faults and is superior to existing state-of-the-art methods.更多还原