Hilbert-Huang Transform Based Nondestructive Testing Of Wood

During the paste few decades, stress-wave-based techniques have been investigated extensively and have shown promise for predicting the mechanical properties of wood materials. But many studies analyze the quality of the wood properties by the stress wave propagation velocity or propagation time. The detection accuracy is low. Therefore, some advanced feature extraction algorithms are needed for stress-wave-based wood test techniques to pave the way for further discovery in physics and nature.This paper applied EEMD-HHT method to the stress-wave-based nondestructive testing of wood. Two new algrithm were put forward aiming to obtain an improved stress wave signal with reduced artifacts and withdraw the weak feature hidden in the signals. The research of this paper are as follows:(1) The priciples of the Hilbert-Huang transform was studied. Then the differences between EMD and EEMD, as well as the differences between the Hilbert specture and the Fourier specture were analysised. A stress wave signal acquisition system utilizing the piezoelectric sensor and a DAQ instrument were also designed in this paper. The signals used in this study were collected on both the healthy wood sample and the decay wood sample.(2) When the hammer hits the wood, the generated stress waves are always followed along with a mixture of dilatational waves and shear distortions. It is difficult to extract useful information directly from the raw stress wave signals. In this paper, an ensemble empirical mode decomposition (EEMD) based approach with the aim of signal denoising was proposed and applied on stress wave signals. Butterworth low pass filter, EEMD-based low pass filter and EEMD-based thresholding filter were used to compare filtering performance. Studies have shown that the method is effective for very noisy signals, the noise can be reduced effectively (MSE obtained is0.02) even in cases where the signal quality is low (SNR value is-5dB).(3) Based on the Hilbert analysis of stress wave signals collected from the healthy wood sample and decay wood sample, the frequency band energy ration and the intrinsic mode energy ratio were compared. It has been proved that the intrinsic mode energy ratio and the frequency band energy ratio based on Hilbert spectrum have shown promise for predicting the inner decay of wood.