Structural Damage Identification Research Based On D-S Theory And Cloud Model

Nearly half a century,the building structures area has been developed swift and rapidly.As increasingly buildings has been built and used in past two decades,the structural damage identification has become a focus of the study in civil engineering field.Different degrees of damage will happen in building structures owing to the influence of all kinds of loads such as wind load and earthquake load in the process of normal use.On the other hand,the building materials will continuously ageing as the years goes by,which leads to local damage of structure.As the damage of structure will influence the safety,durability and applicability of buildings,it is necessary to locate and identify structural damage with scientific damage identification method.In this paper,modal strain energy and acceleration inner product vector are used to construct different cloud models.After respective identification results are obtained,improved D-S theory is adopted to fuse the different results and make the final identification result.This paper firstly introduce the theory and current research status of cloud models,as well as the concept of cloud model,the digital characteristics of cloud model,the basic properties and the application of cloud model in various fields.The cloud model reflects the fuzziness and randomness of things,which ensures the feasibility of the application of uncertainty reasoning.In this paper,the primary application of the normal cloud model is mainly discussed and a single condition single rule cloud generator is constructed with the grey cloud model theory to transfer uncertainty.To research the damage identification effect under different noise levels and different damage conditions of cloud model,model strain energy is used to construct cloud model.At the same time,in order to prove the cloud model based on other indicators can also effectively identify the damage,acceleration inner vector product is calculated to construct another cloud model to identify structure damage.In this paper,based on the knowledge of structural dynamics,the acceleration inner vector product can be calculated by the cross correlation function in T=0,and the cross correlation function can be calculated by the acceleration response of the structure under the external random excitation load.Meanwhile,D-S theory is applied to eliminate the accidental error and false positives caused by a single information source.Since traditional D-S theory is defective when dealing with the evidence of high conflict,which leads to the results unreliable and unreasonable,distance function is introduced to improve and optimize the classical conflict coefficient K.The identify results by improved D-S theory are more accuracy and have stronger anti-noise capability.By comparing and analyzing the expectations and entropy of damage identification results,the reliability and stability of this new damage identification method is proved.Finally,the new method has great application and research prospect in the field of damage identification.