A method for monitoring cable force using resonance-enhanced magnetoelastic with the correction of the hysteresis effect

The magnetoelastic method enables effective monitoring of local cable forces independent of the complex boundary conditions of the cable. Nevertheless, the inherent hysteresis effect can lead to measurement inaccuracies in cable force assessment. This study established the theoretical relationship between the peak-to-peak value of the cable force and the induced voltage under hysteresis. This indicates that the measurement errors in the cable force caused by hysteresis effects are related to the load state of the cable. Various types of cyclic loading experiments were conducted to explore the relationship between hysteresis error and load states in real-world scenarios. The correlation between the characteristics of the induced voltage signals over time and the load states was analyzed. Characteristic values of the induced voltage signals for correcting hysteresis errors over time were proposed. The cable force was predicted using a support vector machine model optimized by particle swarm optimization. The results indicate that compared to traditional linear regression methods for cable force monitoring, the method accounting for corrected hysteresis errors reduced the error by 75.4%. The maximum relative error at tensile forces exceeding 30 kN is 3.98%.