Evaluating Geotechnical Hazards For Long-Distance Gas Pipelines Across Pakistan’s Northern Areas

Abstract

The geotechnical hazard assessment framework is vital to the safe operation of energy pipelines by ensuring that the structure of these pipelines does not suffer in the mountainous areas. This paper describes an open-source framework to determine terrain stability along potential long-distance gas pipeline corridors constructed in northern Pakistan, emphasising the Muzaffarabad area. The method combines machine learning algorithms, geographic information systems (GIS), and conventional geotechnical indicators to define the high-risk zones of landslides and seismic vulnerability. Based on a regionally specific data set containing 1,212 samples, two ensemble approaches, Random Forest and XGBoost, were learned against topographical, hydrological, geological, and seismic characteristics. Random Forest has the highest percentage accuracy (76.95%) and ROC, AUC (0.8384), which makes it a resilient model to apply in a terrain classification exercise. Flow accumulation, elevation, and precipitation were identified as significant factors of slope failure by feature importance analysis. A composite hazard index was computed, and pipeline segments were assigned Low, Moderate, High and Critical risk zones. Despite that, it is essential to note that the scores on segments 6-9 were critical at a frequency of more than 90% of observations. In addition, the seismic threat was simulated by taking synthetic Peak Ground Acceleration (PGA) and Factor of Safety (FOS) values and showing high correlation with the hazard zones predicted with machine learning. A blend of ML predictions and geotechnical thresholds has permitted a successful multi-modal validation. The suggested approach, which is entirely written in Python, allows one to pilot a reproducible, scalable, and region-specific approach to mitigating infrastructure risk. It gives practical information to engineers and planners practising in geologically sensitive locations and enables sustainable energy infrastructure planning in line with national safety and sustainability objectives.