An Innovative Approach of CNN-BiGRU Based Post-Earthquake Damage Detection of Reinforced Concrete for Frame Buildings
Ashish Nagila, P. Saravanakumar, S. Pranavan, Ravi Goutam, Dinesh C. Dobhal, Gagan Singh
Abstract
The new and using an updated post-seismic damage assessment method, the residual deformations the damaged structure underwent during the earthquake are taken into consideration. Estimates of maximum deformations are made using both local and global residual deformations, as well as signs of damage that can be seen with the naked eye. The unique aspect of this technique is that it can account for both rotations and displacements at the same time. Uncertainties due to both stimulation and damage are directly considered. The resulting maximum displacements estimates can assist decide if the investigated structure is usable or repairable. Image preprocessing, feature extraction, and model training have received the bulk of attention as of late. Images are gathered in advance of an event and processed via preprocessing. For feature extraction, the proposed system used data on roof whole detection as well as on building height. CNN, BiGRU, and CNN-BiGRU models will be used to evaluate trainee progress. The suggested model outperforms several well-known alternatives, including CNN and BiGRU.