Structural Damage Identification Based On Time Series Analysis

After the earthquake, hurricane, the heavy snow, debris flow and other largeexternal excitation, whether the structure is damaged? Is it can still satisfy the securityrequirements? The technology of structural damage identification is a problem for thepresent structure engineering. Even severe external shocks did not occur, buildingstructure in the use of the process, constantly affected by changes in temperature, windand other factors, building materials may also appear damage and premature aging.Although this phenomenon does not immediately lead to the destruction of the wholestructure, but it is a potential danger for the safety of the structure, with the passage ofthe time, due to the influence of stress concentration, fatigue and other factors, make thedamage structure part of the continuous expansion and increase, bearing capacity of thewhole structure gradually decline, eventually led to the damage of the structure. Alongwith our country’s existing buildings grow with each passing days, and the frequentnatural disasters in recent years, inventing a set of efficient and reliable method ofstructural damage identification to ensure the safety of people’s life and property, is abrook no delay mission for the scholars of civil engineering. Structural damageidentification method based on the time series analysis is in the condition of unaware ofthe vibration modal and the natural frequencies of the structure, completion of thestructure damage state real-time analysis. The theoretical core of this method is to findthe relationship between the auto-regressive coefficients of the ARMA models and thelayers-shear structure’s acceleration response time series. The auto-regressivecoefficients are regard as an inherent characteristic of structural vibration, and it isapplied to the research of the properties of structures and the law of structural vibration.The main works of this paper are as follows:1) Self compiled MATLAB programs to obtain the six layers-shear structure’sacceleration response time series in the white noise excitation, through the computernumerical simulation. We establish the ARMA models, in order to get theauto-regressive coefficients under different condition of our structure.2) Using the method of auto-regressive coefficients sensitivity factor and themethod of acceleration vector to analysis the auto-regressive coefficients and completethe damage location function for our layers-shear structure and feasibility study on twokinds of methods. 3) Using the trained identification matrix Z for quantitative identification of thelayers-shear structure’s stiffness damage degree, to complete the simulation applicationof structural damage identification methods based on time series analysis. And make theerror analysis of quantitative identification of matrix method.