Damage indicator for building structures using artificial neural networks as emulators

Abstract

Damage indicator for building structures using artificial neural networks (ANN) requiring only acceleration response is proposed. The ANN emulator used for emulating the structural response is tuned to properly model the hysteretic nature of building response. To facilitate the most realistic monitoring system using accelerometers, the acceleration streams at the same location but at different time steps were utilized. The prediction accuracy could be raised by the increment of number of acceleration streams at different time steps. In our proposed approach, damage occurrence alarm could be obtained practically and economically only using readily available acceleration time histories. Based on the numerical simulation for a 5-story shear structure, the adaptability, generality and appropriate parameter of the neural network were studied in. The damage is quantified by using relative root mean square (RRMS) error. Variant ground motions were used to certify the generality of this approach. The appropriate parameter of the neural network was suggested according to variant values of damage index corresponding to the different parameters.