Condition Identification of Bolted Joints Based on Wavelet Analysis

In this paper, discrete wavelet transform (DWT) is used to analyze the acceleration signals near to the loose bolt, in order to study the characteristics of bolted joint condition. Firstly, an experimental system is built based on the NI data acquisition equipment. Secondly, based on qualitative analysis of the loose process of the bolted joints and spectrum analysis of response signals, an appropriate wavelet function is chosen and decomposed levels are determined. Lastly, according to the amplitude of each level, the contour diagrams are drawn and the Root Mean Square (RMS) of detailed coefficients at the sensitive level is calculated. Experimental result shows that both DWT contour diagrams and RMS at sensitive level have significant difference when preload changes.