Prediction of wool fibre breaking strength based on acoustic emission signal parameters and combined noise reduction method

The accurate original signals of fibre fracture and the veritable correlation between the signal parameters and the fibre tensile stress are investigated in this paper based on a suitable noise reduction method. The physical properties of materials can be indirectly characterised through fracture acoustic emission signals, such as fibres. However, it makes the characterisation difficult and inaccurate due to the fracture signals of most fibres are weaker than the background noise. A suitable noise reduction method based on complete ensemble empirical mode decomposition with adaptive noise and wavelet packet transform denoising is proposed to reduce the noise of the fibre fracture signals and obtain the exact raw signals. The fibre fracture signals are decomposed into several intrinsic mode function series and categorised into two groups. The two groups of series are denoised by different thresholds to minimise the level of noise. The two sets of denoised series are reconstructed as the original fibre fracture signals. The results show the proposed combined denoising method can reduce the noise components and well preserve the components of fibre fracture signals. The signal parameters denoised by the combined noised reduction method have a better correlation with the fibre breaking strength and higher prediction accuracy.