Eventos Anais de eventos
COBEM 2017
24th ABCM International Congress of Mechanical Engineering
An Inverse Heat Conduction Study Using Time Travelling Regularization
Submission Author:
Elisan dos Santos Magalhães , SP , Brazil
Co-Authors:
Natan Araujo Moura Leite, Rodrigo Gustavo Dourado da Silva, Elisan dos Santos Magalhães, Sandro Metrevelle Marcondes de Lima e Silva
Presenter: Rodrigo Gustavo Dourado da Silva
doi://10.26678/ABCM.COBEM2017.COB17-1439
Abstract
The Time Travelling Regularization (TTR) is an inverse methodology that is used together with an optimization technique, in this case with the Golden Section method. One of its great advantages is that it allows reducing the noise on the output data on inverse models. In order to compare its versatility, this work presents an alternative way to determine the heat input in a thermal analysis of the heat transfer occurring in a stainless-steel plate. The inverse methodology was based on the Golden Section method used together with the Time Travelling Regularization (TTR). As the direct model, the one-dimensional heat diffusion equation was solved through Finite Differences on a MATLAB code. In order to validate the methodology, a set of lab-experiments were conducted on Stainless Steel AISI 304 samples. The heat input was minimized through the TTR and Specification Flux method. The TTR and Specification Flux methods presented similar results. The heat flux obtained through these methods were very close to each other. As the number of future time-steps is increased the noise on the output data is reduced. The similarity between the results of the TTR and Specification Flux techniques and the low temperature residuals validated the proposed methodology.
Keywords
Time Travelling Regularization, Inverse problems, Optimization, Heat transfer
