Features of long-term health monitored strains of a bridge with wavelet analysis

This paper analyses the five years' monitored strains collected from a long-term health monitoring system installed on a bridge with wavelet transform. In the analysis, the monitored strains are pre-processed, features of the monitored data are summarized briefly. The influences of the base functions on the results of wavelet analysis are studied simultaneously. The results show that the db wavelet is a good mother wavelet function in the analysis, and the order N should be larger than 20, but less than 46 in decomposing the monitored strains of the bridge. According to the strain variation features of concrete bridge, the proper decomposition level is 4 in the wavelet multi-resolution analysis. With the present method, the strains caused by random loads and daily sunlight can be accurately extracted from the monitored strains. The decomposed components of the monitored strains show that the amplitudes of the strains caused by random loads, daily sunlight, and annual temperature effect, are about 5 mu epsilon, 25 mu epsilon, and 50 mu epsilon respectively. The structural response under random load is smaller than the other parts. (C) 2011 The Chinese Society of Theoretical and Applied Mechanics.