EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Predictive Modelling of Weld Pore Size in TIG Welding of ASTM A36 Mild Steel: A Comparative Study of Response Surface Methodology and Random Forest

Collins Eruogun Etin-Osa

Path of Science · 31.10.2026 · 12 (10) · 3001

DOI: 10.22178/pos.135-3 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Path of Science, online erschienen am 31.10.2026. Verfasst von Collins Eruogun Etin-Osa. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Porosity is a critical defect in Tungsten Inert Gas (TIG) welding that significantly compromises the mechanical integrity and fatigue life of welded joints. This study investigates the influence of welding parameters, specifically welding current, arc voltage, and shielding gas flow rate, on the pore size of ASTM A36 mild steel TIG butt welds. We used a Central Composite Design (CCD) to run 20 experiments and measured the resulting pore sizes using a 1200× digital microscope. We developed and compared two predictive models: Response Surface Methodology (RSM) and Random Forest (RF). The RSM quadratic model showed strong fit, with an R² of 95.03%, an Adjusted R² of 90.55%, and a predicted R² of 83.93%. ANOVA results revealed that voltage, along with the quadratic effects of current and gas flow rate, are the most significant factors influencing pore size. The Random Forest model yielded a moderate fit, with a fitted-line coefficient of determination (R²) of 82.0%. Comparative analysis confirms that the RSM model outperforms the RF model in this experimental domain, providing a reliable tool for optimising TIG welding parameters to minimise weld porosity.

Zitieren

APA

Etin-Osa, C. E. (2026). Predictive Modelling of Weld Pore Size in TIG Welding of ASTM A36 Mild Steel: A Comparative Study of Response Surface Methodology and Random Forest. Path of Science, 12(10), 3001. https://doi.org/10.22178/pos.135-3

Vancouver

Etin-Osa CE. Predictive Modelling of Weld Pore Size in TIG Welding of ASTM A36 Mild Steel: A Comparative Study of Response Surface Methodology and Random Forest. Path of Science. 2026;12(10):3001. doi:10.22178/pos.135-3

BibTeX
@article{EtinOsa2026,
  author = {Etin-Osa, Collins Eruogun},
  title = {Predictive Modelling of Weld Pore Size in TIG Welding of ASTM A36 Mild Steel: A Comparative Study of Response Surface Methodology and Random Forest},
  journal = {Path of Science},
  year = {2026},
  volume = {12},
  number = {10},
  pages = {3001},
  doi = {10.22178/pos.135-3}
}
Kostenlos starten →