Wheel-rail contact force measurement based on LC resonance sensing

Abstract

Railway transportation has become a popular means for passenger travel and goods delivery. Wheel-rail contact forces are important parameters in rolling stock, which determine the running safety and stability. Therefore, it is significant to take measures to measure them. The conventional approach is to employ strain gauges with wired connection to measure these forces which has strict requirements for the installation of strain gauges and signal transmission. Exploring wireless measurement methods is beneficial to replace the traditional way to lower the measurement complexity. Although methods such as gap sensors and digital image correlation (DIC) have been investigated, there remains a need to explore new approaches that are low-cost, wireless, and capable of multi-parameter measurement to extend the range of alternative techniques for wheel-rail contact force measurement. This thesis aims to investigate inductor-capacitor (LC) resonance-based sensing for wireless force measurement (WFM), which ultimately contribute to advancements in wheel-rail contact force measurement (WRCFM). The contributions of this work are summarized as follows: The first study is to investigate WFM using a rectangular LC resonance sensor and the LDC1614-based resonant frequency measurement system. The comparison with a strain gauge and aspect ratio of the inductance coil were studied. The results revealed that the rectangular LC resonance sensor realized the wireless force measurement with good repeatability and linear relationship between the resonant frequency (fres) and the loaded force, and a smaller aspect ratio caused greater directivity to enlarge the response range. The second work addresses the overlap challenge of the lift-off and force in sensor output, which can lead to misinterpretation, an orthogonal LC resonance sensor was designed alongside a multi-parameter measurement system based on the LDC1101-ESP32S3 that provided two resonant frequencies and two equivalent parallel resistances (Rp). By analysing their independence in response to lift-off and force changes, an 8-node coordinate transformation algorithm was proposed to effectively separate the lift-off and force. Validation results showed that the maximum relative error of 22.61% for forces and 1.66% for lift-offs, confirming the accuracy and robustness of the proposed system. The third investigation involves the enhancement in lift-off and sensitivity in WFM by integrating semiconductor strain gauge (SSG), magnetic resonance coupling (MRC), and an LDC1101-ESP32S3 based measurement system. By investigating parallel-parallel (PP) and series-parallel (SP) topologies under various lift-offs and SSGs, the approach achieved a sensitivity 5.5 Ω/kN at a lift-off of 33 mm for the PP topology with a 350 Ω SSG and ii demonstrated stable sensitivity near the critical coupling region. These advancements enable WFM with improved robustness and extend the operational range to lift-offs of tens of millimetres. The fourth research applies advancements from the third work to a rotating testing platform, focusing on signal stability against electromagnetic interference (EMI) and temperature fluctuations, as well as signal readout and analysis of signal characteristics. A differential LC resonator configuration combined with Kalman filter (KF) and moving average algorithm (MAA) significantly reduced the impact of temperature changes and EMI noise. According to the enhancement in hardware and software, a rotating testing platform was established to perform the signal readout and characteristic analysis at various lift-offs and Rvar (the simulated change in SSG). Results demonstrated that the changes in lift-off affected both ΔRp (the difference between the measured Rp and the reference Rp) and Δfres (the difference between the measured fres and the reference fres) while Rvar primarily altered ΔRp, demonstrating a proportional relation between the negative peak of Rvar and ΔRp with Δfres remaining stable. These findings highlight the potential of Δfres to measure lift-off and compensate for shifts in ΔRp, enabling improved performance in dynamic scenario. The research achievements for WFM presented in this thesis provide alternative approaches for WRCFM, offering the advantages of low cost, wireless sensing, and high sensitivity. Future work will conduct the actual testing in a wheel-rail system. The installation positions of LC resonators and the measurement system will be optimized according to stress distributions on the wheel, identifying the most susceptible areas. The sensing structure of LC resonators will be further developed into a double-D configuration to achieve uniform and concentrated eddy currents, enhancing sensing capability in terms of directivity and resolution. Additionally, the system will be integrated with an internet of things (IoT) platform and artificial intelligence to enable real-time feature extraction, data visualization, and monitoring of wheel-rail forces.