Thermal-vibration correlation study for high-temperature superconducting maglev intelligent monitoring based on back propagation neural network analysis

Abstract The internal temperature rise inside the high-temperature superconducting (HTS) superconductor (SC) resulting from irregular magnetic field (MF) above the permanent magnet guideway is a major factor contributing to the decline of levitation performance. Real-time monitoring of the temperature rise inside YBCO SC is an important issue for the safe operation of the maglev train systems. However, the existing temperature rise testing method involves destructive intrusion less or more, easily affected by strong MF, occupying limited space and sensors prone to detachment. This paper innovatively proposes a non-contact internal temperature rise testing method combining artificial intelligence (AI) methods. Vibration is common signal of a maglev train system, which inspires to establish a fundamental thermal-dynamic levitation force synchronous testing device for YBCO SC. Then, a set of temperature rise-vibration dataset exposed to different alternating MF frequencies is created. The wavelet transform is chosen to extract the frequency band energy of vibration, and the backpropagation neural network is used to identify the corresponding temperature rise. The recognition accuracy can reach over 99.9%, which firstly proves the effectiveness of AI algorithms in the thermal-vibration correlation analysis for the HTS maglev system. The results can provide the foundation reference for the intelligent monitoring and fault diagnosis of thermal-dynamics stabilities of HTS maglev train in the future.