Development of an Internal Corrosion Maintenance Management System to Enhance the Operational Availability of Natural Gas Midstream Pipeline Systems: A Case Study of Gasco's Mtwara to Dar Es Salaam Pipeline System

Internal corrosion significantly threatens midstream natural gas pipelines' reliability, safety, and efficiency, notably Tanzania's GASCO-operated 542 km Mtwara-Dar es Salaam pipeline transporting gas from Mnazi Bay since 2015. This study identifies key corrosion drivers, develops a predictive model for operational availability, and proposes a validated Internal Corrosion Management System (ICMS) to enhance performance. Using a mixed-methods approach involving surveys from 46 of 64 professionals at GASCO, TPDC, and EWURA, chemical analyses of 2,615 SCADA records, and machine learning modeling via XGBoost, Random Forest, and Support Vector Machine algorithms, the research highlighted major factors through Relative Importance Index analysis: solid/liquid contaminants (RII=0.932), moisture (RII=0.924), infrequent pigging (RII=0.852), design/material issues (RII=0.800), microbiological activity (RII=0.752), operational conditions (RII=0.728), and inhibitors (RII=0.708). The XGBoost model achieved superior performance with 87.43% accuracy, 87.21% precision, 87.32% recall, and 87.27% F1-score, outperforming Random Forest and SVM. Feature importance analysis identified the Corrosive Index (0.23), Fe₂O₃ percent (0.19), and Cl percent (0.15) as the most influential predictors. Model validation demonstrated robust performance with 91.67% severe case detection and 86.9% temporal F1-score. The developed ICMS, integrated with SCADA systems, provides real-time monitoring, automated severity classification, and maintenance prioritisation, supporting data-driven decisions to lower costs and ensure sustainable gas supply under EWURA regulations aligned with NACE SP0110, ASME B31.8S, and API 1160 standards