Self-powered and self-sensing intelligent suspension for application in heavy-haul trains

• A self-powered and self-sensing dual-function intelligent suspension is designed. • The suspension captures the low vibrational frequency and amplitude efficiently. • At 3.0 mm, 1.5 Hz, the RMS voltage is 10.43 V, and the power density is 4,192.84 W/m 3 . • The self-powered application is confirmed in the field test. • A self-sensing and train condition monitoring system is established. Energy issues have long been a key focus worldwide, particularly in the transportation sector, where energy consumption is substantial. This paper proposes a self-powered and self-sensing intelligent suspension (SSIS) based on energy harvesting, which captures low vibrational frequency and amplitude to address the power supply issue of heavy-haul train carriages and enable an onboard train condition monitoring system. The SSIS consists of three modules: a suspension vibration input module, a transmission module, and a generator module, and it generates electrical energy through an electromagnetic generator (EMG). The system was developed in MATLAB to study the mechanical characteristics and energy harvesting performance. Additionally, Simpack software was used to establish a coupled dynamics model of the vehicle-track-suspension system, allowing for the study of vibration and energy harvesting characteristics. Moreover, laboratory and field tests confirm that SSIS can effectively provide continuous power to LED lights and sensors. In field tests, the proposed system achieved a maximum voltage of 23.4 V, charging three parallel 10,000 μF capacitors within 45 s, then maintaining an output voltage above 3 V. Furthermore, the paper introduces the Long Short-Term Memory (LSTM) deep learning to enable intelligent monitoring of train operational conditions by analyzing and classifying the EMG voltage signal, with an accuracy rate of 95.06 %. Therefore, this paper presents an innovative and practical power supply solution, featuring a power density of 4,192.84 W/m 3 and intelligent monitoring in heavy-haul trains, with considerable potential for future applications.