Regression Models for Estimating the Stress Concentration Factor of Rectangular Plates

Estimating Stress Concentration Factors (SCF) guarantees resistance and durability criteria in structures and design components. Failure to correctly identify the SCFs could lead to premature material failure. In this chapter, eight regression models were used to predict the SCF. The regression mode...

Descripción completa

Detalles Bibliográficos
Autores: Rogelio Florencia, Jose Alfredo Ramirez Monares
Tipo de recurso: capítulo de libro
Estado:Versión publicada
Fecha de publicación:2024
País:México
Institución:Universidad Autónoma de Ciudad Juárez
Repositorio:Repositorio Institucional de la Universidad Autónoma de Ciudad Juárez
OAI Identifier:oai:uacj.mx:oai:cathi.uacj.mx:20.500.11961ir-30442
Acceso en línea:https://doi.org/10.1007/978-3-031-66731-2_17
Access Level:acceso abierto
Palabra clave:Stress concentration factor
Rectangular plates
Polynomial curve fitting
Artificial intelligence
Regression models
Random sample consensus
Ridge regression
LASSO regression
Elastic Net
Random forest regression
Support vector regression
Polynomial regression
info:eu-repo/classification/cti/7
Descripción
Sumario:Estimating Stress Concentration Factors (SCF) guarantees resistance and durability criteria in structures and design components. Failure to correctly identify the SCFs could lead to premature material failure. In this chapter, eight regression models were used to predict the SCF. The regression models were multiple linear regression, random sample consensus, ridge regression, LASSO regression, elastic net, random forest regression, support vector regression, and polynomial regression. The models were trained on a dataset resulting from a two-dimensional Finite Ele ment Analysis from the Finite Element Method for different values of the parameters: large, width, and circular hole radius in a tensile plate. Least squares polynomial equations were fitted to these design points. The performance of the models was compared using the MSE, RMSE, MAE, MAPE, and R2 metrics. The random forest regression performed the best.