Structural Damage Identification Using High Dimensional Model Representation

Abstract

Any engineering structure is subject to various internal and external factors which may cause wear or malfunction due to deterioration, an incorrect construction process, lack of quality control or environmental effects. To be able to observe these changes in the material and to react in a proper way before serious damage is caused, the implementation of a damage identification system is crucial. In the past, many methods have attempted to identify damage by solving an inverse problem, which inevitably needs an analytical model. However, often the construction of these analytical model requires considerable effort in building a mathematical framework with acceptable level of accuracy and reliability which makes these approaches less attractive. In order to circumvent this complexity, this work presents a computationally efficient approach in structural damage identification (SDI) using high dimensional model representation (HDMR). In general, most of the structural systems are simulated with the help of finite element (FE) models to predict static as well as dynamic behaviour of the systems with different boundary conditions. Therefore the FE models have to be in tune with the experimental observation to facilitate any modifications in the systems so that the future responses can be accurately predicted, and subsequently utilized in design optimization. Thus, finite element model updating (FEMU) is effective in improving the correlation between predicted and observed ones by correcting the inaccurate modelling assumptions. The proposed methodology involves an integrated finite element modeling, development of response surface model using HDMR, establishment of objective function, and minimization of the function using genetic algorithm. An attempt has been made to reduce the computational effort with increase in the accuracy of updated parameters. The proposed methodology is applied in model updating of a simulated beam and an existing reinforce cement concrete (RCC) box culvert structure. The results have demonstrated that the HDMR based FEMU is a good candidate featuring computational efficiency. Further to validate the proposed methodology in SDI, threecase-studies (an experimental beam, a frame structure and a bridge structure) have been considered. The damage patterns, locations and severity obtained using the proposed methodology are compared with the experimental results available in literature, and are found to be in good agreement. Based on the study conducted, it can be concluded that the HDMR based FEMU in SDI is computationally efficient.