• norsk
    • English
  • English 
    • norsk
    • English
  • Login
View Item 
  •   Home
  • Sentraladministrasjonen
  • Publikasjoner fra Cristin
  • View Item
  •   Home
  • Sentraladministrasjonen
  • Publikasjoner fra Cristin
  • View Item
JavaScript is disabled for your browser. Some features of this site may not work without it.

Residual Stress Prediction Of Welded Joints Using Gradient Boosting Regression

Keprate, Arvind; Bhardwaj, Sachin; RATNAYAKE MUDIYANSELAGE, Chandima; Ratnayake Mudiyanselage, Chandima
Conference object
Accepted version
Thumbnail
View/Open
Residual Stress Prediction Of Welded Joints Using Gradient Boosting Regression.pdf (1.104Mb)
URI
https://hdl.handle.net/11250/3023360
Date
2022-07-23
Metadata
Show full item record
Collections
  • Publikasjoner fra Cristin [4186]
  • TKD - Institutt for maskin, elektronikk og kjemi [265]
Original version
https://doi.org/10.1007/978-3-031-10525-8_4
Abstract
Welding residual stress (WRS) estimation is highly nonlinear process due to its association with high thermal gradients generated during welding. Accurate and fast estimation of welding induced residual stresses in critical weld geometries of offshore structures, piping components etc., becomes important from structural integrity perspective. Fitness for services (FFS) codes like API 579, BS7910 recommend residual stress profiles are mainly based on three approaches, out of which nonlinear finite element modelling (FEM) results coupled with residual stress experimental measurement, have been found to be most conservative and realistic. The residual stress estimation from thermo mechanical FEM models is computationally expensive as it involves a large degree of interactions between thermal, mechanical, metallurgical and phase transformations etc. The destructive and non-destructive measurement techniques also carry a large amount of uncertainly due to lack of standardization and interpretation variability of measurement results. To mitigate the aforementioned challenges, response surface models (RSMs) have been proposed in this study, for the estimation of WRS at a significant confidence. This paper examines the applicability of 12 different Response Surface Models (RSMs) for estimating WRS. The training and testing data is generated using FEM, Abaqus- 2D weld interface (AWI) plugin. To compare the accuracy of the RSMs, three metrics, namely, Root Mean Square Error (RMSE), Maximum Absolute Error (AAE), and Explained Variance Score (EVS) are used. An illustrative case study to demonstrate the applicability of the response surface model to predict WRS is also presented.
Publisher
Springer
Series
Communications in Computer and Information Science;Volume 1616
Journal
Communications in Computer and Information Science

Contact Us | Send Feedback

Privacy policy
DSpace software copyright © 2002-2019  DuraSpace

Service from  Unit
 

 

Browse

ArchiveCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsDocument TypesJournalsThis CollectionBy Issue DateAuthorsTitlesSubjectsDocument TypesJournals

My Account

Login

Statistics

View Usage Statistics

Contact Us | Send Feedback

Privacy policy
DSpace software copyright © 2002-2019  DuraSpace

Service from  Unit