Online Crack Detection in Laser Cladding Using Wavelet Threshold Denoising

Early detection of defects in laser cladding is crucial to minimize material waste and ensure product quality. This paper presents an online crack detection method based on acoustic emission and wavelet threshold denoising. A microphone is employed to capture sound signals generated during the laser cladding process, which contain valuable information about crack formation.

To enhance the signal-to-noise ratio and extract meaningful crack information, a wavelet threshold denoising technique is implemented. The optimal parameters for the denoising method are determined by evaluating the signal-to-noise ratio (SNR) and root mean square error (RMSE). The chosen parameters include the Db12 wavelet basis function, 3-level wavelet decomposition, and hard thresholding function.

A customized wavelet coefficient thresholding method is then applied to denoise the crack signal, effectively reducing noise while preserving critical crack information. This customized approach exhibits superior denoising performance compared to other signal denoising methods. The results demonstrate the feasibility and effectiveness of custom wavelet threshold denoising for online crack detection in laser cladding, paving the way for real-time process monitoring and quality control.