Predictive Modelling of Hydrant Piping Corrosion Rates

Authors

  • Regan Rahadian Pambudi Department of Refinery Mechanical Engineering, PEM Akamigas Author
  • Umi Yuliatin Department of Logistik Oil and Gas, PEM Akamigas Author

Keywords:

Hydrant piping; Corrosion rate; Ultrasonic thickness testing; Multiple linear regression; Predictive modelling.

Abstract

This study develops an Ordinary Least Squares (OLS) multiple linear regression model to estimate corrosion rates in hydrant piping systems using ultrasonic wall-thickness measurements. Field inspection data from 147 pipeline segments at PEM Akamigas were analyzed using two thickness measurements as independent variables to examine their contributions to corrosion rate estimation. The regression model achieved a coefficient of determination (R²) of 0.874, indicating that the predictor variables collectively explain 87.4% of the observed variation in corrosion rates. The regression coefficients reveal different contributions from the two thickness measurements, with Thickness 1 exhibiting a stronger negative coefficient (−0.297) than Thickness 2 (−0.078). The estimated corrosion rates ranged from 0.0226 to 0.0343 mm/year, with ten segments identified as requiring priority maintenance attention, particularly within segments 53–108. These findings demonstrate the potential of regression-based analysis of ultrasonic inspection data to support pipeline condition assessment and maintenance prioritization. The proposed approach provides an interpretable statistical framework for identifying vulnerable pipeline segments and improving the management of industrial hydrant piping integrity.

Downloads

Published

2026-06-28

How to Cite

Predictive Modelling of Hydrant Piping Corrosion Rates. (2026). Luxfia: Journal of Multidisciplinary Research, 1(02), 187-201. https://risetcendikia.com/index.php/journal-luxfia/article/view/509
Google Scholar
Metrics & Citations
Indexed
Track citations and scholarly impact of this article directly on Google Scholar.
Search on Google Scholar