Research On Wavelet-Genetic Algorithms Of Structure Damage Identification

Engineering structure damage will occur when in the long-term use by external environmental corrosion, material aging itself and fatigue load. When the damage accumulation goes to a certain degree, it will affect the normal use of the structure and even lead to the collapse. The more early discover and repair the damage, not only can greatly reduce the costs of structure, but also can prevent the unnecessary loss of lives and property. So, researching on the method of structural damage identification has important theoretical significance and engineering value.This paper is based on the theory of wavelet analysis and genetic algorithm. Wavelet analysis has excellent local analysis characteristics and features zoom in time-frequency domain. Genetic algorithm has strong adaptive, parallelism and robustness, etc and its global search ability is very strong. So, this paper is put forward the concept of wavelet-genetic algorithm based on the respective advantages of wavelet analysis and genetic algorithm and setting up a method of wavelet genetic algorithm which can both identify structural damage location and degree. First of all, by using the finite element analysis to solve the vibration modal of structure and using the db wavelet to make continuous wavelet transform, the damage location is identified by the modulus maxima of wavelet series. The paper uses element stiffness reduction factor as optimization variables of genetic algorithm and uses error function of vibration and frequency to construct objective function and simplifies the variables of the objective function through the determination of damage location.The objective function is optimized by genetic algorithm to determine damage degree of structure. This paper uses the beam as the research object and uses little damage, damage and different location of different damage degree to verify the effectiveness of the method proposed. The finite element model of beam containing damage is established. Modal analysis to the finite element is made to get modal parameters such as frequency and vibration mode. The paper makes continuous wavelet transform to displacement modal to get the wavelet coefficient map. According to the modulus maxima of wavelet coefficient map, the damage location can be pointed out. The objective function is simplified by the damage location and minimized by genetic algorithm to get the damage degree of damage location.This paper uses continuous beam as research object and three different finite element model of damage condition are established. The paper uses Lanczos method to analyze the finite element model to get modal parameters such as frequency and vibration mode. Continuous wavelet transform to displacement modal is made to get the wavelet coefficient map. According to the modulus maxima of wavelet coefficient map, the damage location can be pointed out. The objective function is defined by the error function of the vibration mode and frequency. The unknown variables of the objective function are simplified on the basis of the damage location.Then the objective function is minimized by the genetic algorithm to get damage degree of continuous beam. Results show that the method not only can effectively identify the damage location but also can accurately identify the extent of the damage. The method has important guiding significance to the beam damage identification.