Leak Detection in Oil Pipelines
The client, a prominent oil transportation operator, aimed to improve their leak detection system to reduce false positives and enhance sensitivity in detecting real leaks.
In collaboration with our partner Sciata, PRESAGE designed and implemented an innovative system based on AI prediction models. This system successfully reduced false positives by up to 90% and demonstrated the capability to detect leaks as small as less than 1% of the flow through simulated tests.
The solution utilizes an intelligent middleware that combines various proprietary machine learning algorithms written in Python. These algorithms connect to SCADA data and feed into a user-friendly front-end, allowing users to transparently view and comprehend the reasons behind each generated alert.
- Cost reduction by avoiding expensive interventions associated with false positives.
- Prevention of ecological damages and litigations by detecting leaks in a timely manner.
- Decreased stress and fatigue for operators through the reduction of false positives in SCADA.