Eventos Anais de eventos
COBEM 2023
27th International Congress of Mechanical Engineering
On the Prediction of Critical Heat Flux via Generalized Additive Models (GAMs)
Submission Author:
Guilherme Borges Ribeiro , SP , Brazil
Co-Authors:
Renan Santos Barbosa, Camila de Souza, Guilherme Borges Ribeiro
Presenter: Renan Santos Barbosa
doi://10.26678/ABCM.COBEM2023.COB2023-0399
Abstract
The critical heat flux is the heat flux value in the boiling or cooling process in which the heat transfer decreases, and the heated surface temperature rises rapidly due to factors such as the presence of vapor films and bubbles. Due to the difficulties in elaborate experiments, and disagreements about measurement and evaluation techniques related to CHF, some methods try to predict CHF, like, lookup tables, physical correlations, and machine learning. These methods use experimental variables to estimate the CHF value, like pressure, temperature, and mass flux. Generalized additive models (GAMs) are a type of regression model that extends the applications of generalized linear models (GLMs), giving the possibility to model nonlinear associations between the response variable (CHF) and their predictors. GAMs use smoothing functions to estimate the relations allowing the GAMs models to be more flexible in modeling nonlinear relationships and control predictive aspects like bias and variance to prevent overfitting. This work presents the results obtained from the use of generalized additive models (GAMs) on the prediction of critical heat flux on water-related data. Moreover, this work evaluates the capability of GAMs models in predicting CHF compared to other methods like the lookup table and machine learning models.
Keywords
CHF, Flow boiling, machine learning, Generalized Additive Models
