Publication:
Computational study of foreshocks in the Burridge-Knopoff earthquake model using machine learning

No Thumbnail Available
Authors
Ortega Chagueza, Jesus G.
Embargoed Until
Advisor
Ramos , Rafael A.
College
College of Arts and Sciences - Sciences
Department
Department of Physics
Degree Level
M.S.
Publisher
Date
2023-12-13
Abstract
The study of mechanical models of earthquake faults is important for understanding the different behaviors observed in real earthquakes. The model considered for this work was introduced by Burridge and Knopoff, consisting of blocks connected by linear springs in contact with a moving rough surface. A numerical simulation of the model was implemented, revealing a variety in event size that follows a power-law distribution. As the next step, a database of artificial earthquakes was created with the purpose of training an artificial neural network (ANN) model capable of estimating the magnitude of the events generated by the simulation. The ANN models show satisfactory results in estimating the magnitude of artificial seismic events; however, there are still significant aspects to be discovered before they can be effectively applied to real seismic events.

El estudio de modelos mecánicos de fallas sísmicas es importante para comprender los diferentes comportamientos observados en los terremotos reales. El modelo considerado para este trabajo fue introducido por Burridge y Knopoff, y consiste en bloques conectados por resortes lineales en contacto con una superficie rugosa en movimiento. Se implementó una simulación numérica del modelo, revelando una variedad en el tamaño de los eventos que sigue una distribución de ley de potencias. Como siguiente paso, se creó una base de datos de terremotos artificiales con el propósito de entrenar un modelo de red neuronal artificial (RNA) capaz de estimar la magnitud de los eventos generados por la simulación. Los modelos de RNA muestran resultados satisfactorios en la estimación de la magnitud de los eventos sísmicos artificiales; sin embargo, aún existen aspectos significativos por descubrir antes de que puedan aplicarse de manera efectiva a eventos sísmicos reales.
Keywords
Earthquakes,
Burridge and Knopoff,
Artificial neural network,
Numerical simulation,
Machine learning
Usage Rights
Except where otherwise noted, this item’s license is described as Attribution-NonCommercial-ShareAlike 4.0 International
Cite
Ortega Chagueza, J. G. (2023). Computational study of foreshocks in the Burridge-Knopoff earthquake model using machine learning [Thesis]. Retrieved from https://hdl.handle.net/20.500.11801/3622