Structural damage identification using the similarity measure of the cloud model and response surface-based model updating considering the uncertainty

This paper proposes a novel structural damage identification method by using the similarity measurement of the cloud model based on response surface model updating, taking into account of measurement noise. Cloud model, as an important cognitive computing model, can realize the bidirectional cognitive transformation between qualitative concept and quantitative data based on the theory of probability and fuzzy set. Therefore, the numerical characteristics of the cloud model are introduced to quantify the measurement noise of the structural response. Then, a model updating technique is used for damage diagnosis. The cloud generator is employed to generate response samples, then combined with the fast-computation feature of the response surface method, an inverse optimization problem is established for predicting parameters corresponding to each sample. Finally, the max boundary based on the cloud model is then used to calculate the parameter similarity of each element under different scenarios, and the mean value of the similarity index of each element is used as the threshold value to identify the location of the damage. And the damage severity is constructed using the variation of the expectation of damage parameters. Numerical study on a 5-story steel frame is conducted to investigate the accuracy and robustness of the proposed approach. The effects of noise level, number of original samples and number of cloud droplets on the damage identification results are discussed. Experimental verification on a laboratory steel frame structure model is conducted to further validate the accuracy of the proposed approach. The results show that the proposed method gives enough accuracy in damage identification of not only numerical but also laboratory structure with single and multiple damage scenarios.