Abstract
Gear is widely used in mechanical transmission system, but it is prone to failure, which seriously affects the performance of equipment. In order to realize the diagnosis and classification of gear faults, we used Convolutional Neural Networks (CNN) to extract time-frequency image features of its vibration signal. CNN can extract the characteristics of time-frequency signals from vibration signals and identify gear faults accurately. However, due to the large training data set, CNN training costs too much time. According to the characteristics of Graphics Processing Unit (GPU), Compute Unified Device Architecture (CUDA) can improve the speed of CNN algorithm and reduce the time consumption. Therefore, this paper proposes a method based on GPU-CNN for gear fault diagosis. The experimental results show that the method can effectively shorten the training time and significantly improve the operation efficiency.
Export citation and abstract BibTeX RIS
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.