Research on Gear Early Wear Fault Diagnosis Based on the Sparse Attention Mechanism

Abstract

In the field of gear fault diagnosis, it is of great significance to effectively diagnose early wear faults of gears. However, early wear fault has weak characteristics and is difficult to diagnose. Thus, a gear early wear fault diagnosis model based on the sparse attention mechanism is proposed to solve this problem. This model uses a new sparse attention mechanism combined with a convolution neural network to improve the traditional sectional attention mechanism and locate the specific fault frequency. The gearbox fault simulation test data is used for test verification. Compared with other diagnosis methods, this new method can achieve a more accurate and comprehensive diagnosis, reduce analysis cost and obtain sensitive fault characteristic frequency under the same sample condition and calculation cost. The conclusion of this method can provide data support for gear maintenance.

Keywords

• Gear wear;Fault diagnosis;Convolution neural network;Attention mechanism