Damage assessment of a full-scale bridge based on the response surface method

Abstract

As a combination of statistical and mathematical techniques, response surface models have been recently found to be capable of substituting FE models in model updating iterations by using explicit mathematical functions to represent the relationship between the inputs and outputs of a physical system. However, the literature related to this topic is still scarce despite the wide employment of the response surface method in many engineering realms such as chemistry and industry. Due to that, this paper attempts to propose a systematic damage assessment procedure based on the model updating strategy using the response surface method. Instead of the qualitative evaluation traditionally used, here the 2(k) factorial design is employed to screen out non-significant updating parameters by quantitative statistical analysis, which considerably improves the screening reliability. Meanwhile, the central composite design is adopted to construct response surface models substituting original FE models during updating. The proposed method is used to detect the damage existing in an experimental full-scale bridge. The results demonstrate the merits of this method in its easy implementation and high computation efficiency, especially for the bridge case.