Noise reduction for LDC resonant eddy current using multi-scale dual-parameter reinforced-learning fusion

Abstract

This paper presents a novel methodology for mitigating vibration-induced distortions in rail surface defect detection through eddy current testing. We developed a dual-parameter eddy current sensor utilizing the LDC1101 (inductive digital converter). The sensor exploits the fundamental relationship between parallel impedance (R_P) and inductance (L) to establish a dual-parameter fusion model that effectively reduces vibration interference during detection. Sensor sensitivity is enhanced through Litz Wire implementation. Our numerical simulations compare the defect detection capabilities of Litz Wire versus single-core coils, while examining the Litz Wire’s response characteristics during dual-parameter detection. These analyses also elucidate the mechanisms underlying anomalous R_P responses. Laboratory and field trials validate the methodology’s efficacy in suppressing vibration interference during eddy current detection, demonstrating the probe’s reliability for surface defect detection in rail production.