Damage detection of bridges under changing environmental temperature using the characteristics of the narrow dimension (CND) of damage features

Abstract

Latent environmental variable-based methods are popular for diagnosing structural damage in bridges under varying environmental temperatures. However, it is still a shortage of researches on distinguishing and selecting the damage features suitable for this type of method in order to improve the validity of damage detection in terms of their application to actual bridges. To address this issue, a method based on the characteristics of the narrow dimension (CND) of damage features is proposed for damage detection in actual bridges under changing environmental temperatures. First, the definition of the CND of damage features is presented in detail by analysing the influencing factors of latent environmental variable-based methods for damage detection. Second, based on the correlations among damage features, an index is presented for determining whether the damage features possess CND. If so, these damage features are suitable for latent environmental variable-based methods, which are beneficial for detecting damage to bridges while considering environmental influences; if not, then other types of methods must be considered in using the above damage features. Consequently, combined with novelty detection and variational mode decomposition, a method based on CND is implemented for detecting the damage to actual bridges. Finally, examples of numerical simulations and actual bridges are provided to demonstrate the effectiveness of the proposed method.