Damage alarming analysis for ancient wood structures based on wavelet packet energy spectrum

Abstract

Purpose – The purpose of this paper is to establish the damage alarming indexes for ancient wood structures and study the damage sensitivity and noise robustness of these indexes under random excitation. Design/methodology/approach – Xi’an Bell Tower is taken as a case in this paper to simulate the damage of ancient wood structures through finite element (FE) simulation and determine the satisfactory damage alarming indexes with wavelet packet energy spectrum. Findings – The results of this paper show that: 1) the damage alarming indexes can effectively identify the damage of ancient wood structures, each index with a different damage sensitivity; 2) the energy ratio deviation is greater than the energy ratio variance and is close to the maximum variation of energy ratio; 3) the energy ratio deviation has a better alarming effect than the energy ratio variance during the initial period of the damage. With the accumulation of the damage, the energy ratio variance outperforms the energy ratio deviation; 4) the sensitivity of the energy ratio deviation and variance varies from positions, changing from the highest to lowest at the mortise-and-tenon joints, the beam mid-span and the plinth; 5) if signal to noise ratio (SNR) is 40db or larger, the indexes can accurately identify the damage of ancient wood structures. As SNR increases, the indexes will have an increasingly higher sensitivity and certain ability to resist noise. Research limitations/implications – The FE model is simpiy, it does not completely reflect Xi’an Bell Tower. Practical implications – It will provide a theoretical basis for the damage alarming of Xi’an Bell Tower. Social implications – It makes structural health monitoring through structural vibration response under ambient excitation a new research field in damage detection as well as a positive way of ancient architecture protection. Originality/value – This paper studies the damage alarming effect on ancient wood structures from different wavelet functions and wavelet packet decomposition levels. To study the effect under white noise environment, this paper adds Gaussian white noise with a SNR of 10, 20, 30, 40 and 50 db to the acceleration response signal of intact structure and damaged structure.