Development and Evaluation of An Integrated AI-Based Technology for Pipelines Integrity Assessment

Abstract

In this study, an integrated framework of Artificial Intelligence (AI) based Digital Twin for predictive integrity assessment, corrosion monitoring, pipeline maintenance optimization and disruption reduction was developed and evaluated for both piggable and unpiggable pipeline system under Nigeria environmental conditions. The challenges faced with pipeline corrosion, poor operational condition, vandalism, inadequate cathodic protection system and environmental hazards in shallow water, deep offshore, Niger Delta swamp and land terrain environment were the focus of the study. The methodology was based on the adoption of a unified digital twin architecture that combines the Smart Pig inspection data, AI-based corrosion prediction, robotic crawler ultrasonic thickness measurement, Distributed Acoustic Sensing/Distributed Temperature Sensing (DAS/DTS) fiber optic monitoring system, anomaly detection algorithms, environmental risk analysis and predictive maintenance decision support models. The study involved 300 pipeline segments. The results of the integrity assessment revealed that many pipelines had operating pressure margins that were safe, and that no pipeline segment was forecast to experience greater than 80% pipeline wall loss over the next 12 months. The Mean Absolute Error value for the AI-based corrosion prediction model was 1.520 mm, indicating that the model has a good predictive capability. Out of the 300 pipeline segments assessed, 157 were determined to be low risk, 137 medium risks, and 6 were high risk, and the average Digital Twin Integrity Index was 69.12, and the average Remaining Useful Life was 22.11 years. The study also identified 24 pipeline sections that were considered anomalous, 7 pipeline sections that were suspected to have a leak, and 38 locations where cathodic protection was not working. In addition, proactive integrity management strategies resulted in cumulative disruption and environmental impact scores that were reduced from 4844.0 in reactive management to 983.8, which is about 79.7% less than reactive management. The report found that the use of AI-powered digital twin systems, predictive analytics, fiber optic surveillance, and anomaly detection can greatly enhance pipeline reliability, predictive maintenance planning, structural integrity assessment, and environmental sustainability in complex oil and gas operating environments.

Keywords

References