📖 ABSTRACT/OVERVIEW
The integrity of Nigeria's petroleum pipeline network is compromised by a combination of ground deformation, corrosion, sabotage, and operational overpressure, resulting in recurrent spills, explosions, and significant environmental damage, yet spatially comprehensive deformation monitoring of pipeline corridors remains absent from the operational toolkit of the pipeline infrastructure operators. This dissertation develops an original methodology for critical petroleum infrastructure protection through InSAR time-series deformation monitoring and machine learning-based anomaly detection applied to Nigeria's major transcontinental pipeline corridors. Sentinel-1 SAR data covering three major pipeline routes totalling 2,700 kilometres are processed using the MT-InSAR SqueeSAR algorithm to derive ground deformation time series at persistent scatterer and distributed scatterer points along the corridors. A novel unsupervised deep learning anomaly detection architecture, the Pipeline Deformation Signature Autoencoder, is developed to distinguish pipeline-relevant deformation patterns from background geodynamic signals at continental scale. The architecture is trained on a labelled library of known pipeline failure and deformation events compiled from published incident databases. Field validation campaigns at twenty-two accessible pipeline sections verify detected anomaly locations against physical condition indicators. Results demonstrate detection of twelve previously unreported anomalous deformation zones along the surveyed corridors, with four subsequently confirmed by operator inspection to have measurable pipe displacement. The dissertation advances InSAR operational applications to critical infrastructure management and proposes a continuous pipeline monitoring framework for the Nigerian National Petroleum Company Limited. Keywords: InSAR, pipeline monitoring, deformation, machine learning, critical infrastructure.
Need Complete Chapters of the Above Topic?
Get high-quality, Zero-AI research materials with current citations.
Request via WhatsApp 💬