Structural Damage Identification Based On Time Series And Neural Network Under Ambient Temperature Variations

With the development of society and the progress of science and technology,the demand of human society is getting higher and higher.Large and complex structures such as space shuttles,high-rise buildings,offshore platform structures,sea-crossing bridges and so on are becoming more and more.In the course of their service,these structures will continue to be affected by the environment,especially the temperature,and then change the structure's force response.Due to the characteristic parameters of force response,the structural damage identification is often misjudged.However,the existing damage identification methods which eliminate or reduce the influence of temperature have structural limitations and can not accurately quantify the structural damage.In view of this situation,combining the applicability of time series to structure and the effectiveness of neural network damage step-by-step identification,this paper proposes a method of structural damage identification based on time series and neural network at ambient temperature.Benchmark model is used to study structural damage identification under temperature.The main content of the thesis includes:(1)The methods of structural damage identification based on time series and neural network are summarized,and their application scope and existing problems are pointed out.The mechanism of temperature effect on structure,the methods of structural damage identification considering temperature effect and their limitations are expounded.(2)The basic theory of time series and neural network is briefly introduced.Aiming at the influence of temperature on structural damage characteristic parameters,a method of structural damage identification based on time series and neural network under ambient temperature is proposed,and its realization process is introduced in detail.(3)Taking Benchmark frame structure as the research object,the damage of single component and two components is simulated by using Gaussian white noise excitation and introducing temperature effect,and the damage location and quantification of the structure are realized.The effectiveness of the method is verified.(4)The method of adding noise to structural damage identification is briefly described.The Benchmark frame structure is taken as the research object.The damage of single component and two components is simulated by using Gaussian white noise excitation and introducing temperature and noise effects.The damage location and quantity of the structure are realized,and the anti-noise ability of the method is verified.