Wireless Rail Fastener Looseness Detection Based on MEMS Accelerometer and Vibration Entropy

In this paper, we present an automatic inspection system based on micro motion sensors for detecting the 'looseness' of rail fasteners. The system is composed of a low-power MEMS accelerometer and a Global System for Mobile Communications (GSM) unit, and can detect loose fasteners and upload the results to a cloud server in real-time. Finite Element Method (FEM) is used to analyze rail vibration characteristics in the vertical direction as the rail is excited by mechanical pulse inputs. On this basis, the Chao-Shen Entropy theory was applied to identify fastener looseness reliably. In addition, field experiments were also conducted on Datong-Qinhuangdao Railway, the longest coal-transport railway in the world. The experimental results show that fastener looseness can slow down the attenuation of rail vibration in the time domain and produce a large entropy value. Using the method of amplitude entropy, loose fasteners can be identified reliably for looseness factor >60%. The proposed system has been experimental validated to enable real-time detection of railway fastener looseness and could potentially bring significant benefits for the day-to-day maintenance of railways.