Publication Type : Conference Paper
Url : https://doi.org/10.1007/978-981-33-4324-5_6
Campus : Coimbatore
School : School of Engineering
Year : 2021
Abstract : Present study focuses on developing region-specific New Generation Ground Motion Prediction Models using Artificial intelligence technique for North India purely based on a measured ground motion data from specific region. Simple single hidden layered feed forward multilayer perceptron networks with back-propagation learning algorithm are used. A total of 280 data points of recorded strong motion data from the Kangra and Uttar Pradesh (UP) arrays, made available by the Program for Excellence in Strong Motion Studies (PESMOS), were used to train these networks. The first model predicts Moment Magnitude for a given Hypocentral Distance and Peak Ground Acceleration. The second model predicts Peak Ground Acceleration (PGA) for a given Hypocentral Distance (HPD) and Moment Magnitude (MM). Performance analysis, Uncertainty analysis and analysis of interactive effects have been done to test the reliability of the generated models. Optimization analysis was also performed to predict possible inputs of the models for a given set of outputs. Models have performed reasonably well for the given amount of non-linearity in the data.
Cite this Research Publication : Huang H., Ramkrishnan R., Kolathayar S., Garg A., Yadav J. S. (2021) Development of Region-Specific New Generation Attenuation Relations for North India Using Artificial Neural Networks. In: Garg A., Solanki C.H., Bogireddy C., Liu J. (eds) Proceedings of the 1st Indo-China Research Series in Geotechnical and Geoenvironmental Engineering. Lecture Notes in Civil Engineering, vol 123. Springer, Singapore.