An interpretable machine learning model for failure pressure prediction of blended hydrogen natural gas pipelines containing a crack-in-dent defect

Guojin Qin, Chao Zhang, Bohong Wang, Pingan Ni and Yihuan Wang

Energy, 2025, vol. 320, issue C

Abstract: The present study developed an interpretable hybrid machine learning-based model to predict the failure pressure of blended hydrogen-natural gas (BHNG) pipelines with crack-in-dent (CID) defects. With extreme gradient boosting (XGBoost) as the fundamental predictor, the whale optimization algorithm (WOA) was used to optimize its hyperparameters. The developed model, WOA-XGBoost, was used to train the machine learning-based prediction model with finite element modeling data. Three benchmark models were trained for comparative reasons to predict the failure pressure. The developed WOA-XGBoost model demonstrated superior predictive performance with a coefficient of determination (R2) of 0.986. The sensitivity analysis showed that when n_estimators is 200, the mean values of R2, RMSE, and MAE are 0.986, 0.275, and 0.203, respectively. Model interpretability was enhanced through SHapley Additive exPlanations (SHAP), which visualized global and local feature contributions. Results indicated that H atom concentration significantly affected failure pressure predictions, with a mean absolute SHAP value of 0.91. The proposed model provides a robust and innovative solution for the integrity management of BHNG pipelines with complex defect configurations.

Keywords: Blended hydrogen-natural gas pipelines; Failure pressure prediction; Machine learning; Model interpretable; Crack-in-dent defect (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544225010436
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:320:y:2025:i:c:s0360544225010436

DOI: 10.1016/j.energy.2025.135401

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().