Intelligent Identification Method for Misfires based on EMD-SVD and Ridge Features

Misfire incidents present substantial safety risks in mine blasting operations, making their precise and timely detection critical for operational safety. To address the challenges of low signal-to-noise ratios in complex mine vibration signals and subtle misfires often obscured by noise, this study develops a vibration signal processing method. The approach combines Empirical Mode Decomposition-singular Value Denoising (EMD-SVD) for primary feature extraction with time-segmented wavelet ridge analysis for misfire identification. Applied to a blasting operation case study at a Weinan metal mine, this method successfully identifies misfire events while overcoming the inadequate recognition accuracy of traditional methods in high-noise environments. Field vibration signal analysis confirms the method′s exceptional noise immunity, accurate feature identification, and reliable detection performance, establishing a novel technical solution for real-time intelligent misfire monitoring in mining operations.