Ultrasonic Quantitative Testing Of Wood Hole-Defect Based On Energy Method

With the timber volume of production reducing and the demand's gaining, and this make non-destructive testing of wood is increasingly important. After discussing the shortages of existing analysis methods in non-destructive testing of wood, an approach based on Spectrum analysis and Wavelet theory is proposed for dealing with testing signal, then Neural Networks is used to testing wood defect.According to the regular of ultrasonic wave spreading in wood, expounds ultrasonic wave energy waning and frequency componential transforming, by the reason that the quality of timber variety. So Spectrum analysis and Wavelet theory is proposed for dealing with testing signal, and energy waning measure wood defect message.Spectrum analysis disposals defect signal: contrasting FFT and each power spectrum, choose to use the Welch method disposals defect signal. According to the power spectrum density judge variety regulation with the size and position of the hole defect. Taking each spectrum's energy varieties as the character inputs of the artificial neural network: originality signal are dispelled by 5 layer wavelet bundle, and receive 32 coefficient from low spectrum to high spectrum, then abstraction energy changers as the character inputs of the artificial neural network.In the identifying of nerve network stage: The processes of the hole position examination is similar with the examination of the hole size. Take each spectrum's energy varieties as the character inputs of the artificial neural network, training network. Make use of this network test Wood specimens, and can identify the timber hole the position and size.Result shows: Spectrum analysis has certain variety regulation with the position's variety and the size of the empty defects, and expresses defect spectrum abundance, so Wavelet analysis can used in further disposal. Wavelet analysis and neural networks combine together, can examining the position and size of the hole defect, achieving wood defect quantitative testing.