Application of AI with computer vision for monitoring and detection of leaks at oil and gas facilities

by George
Problem: Leaks occur at large oil and gas facilities, which can lead to significant environmental and economic losses. Traditional leak detection methods can miss small or slow leaks, resulting in a loss of up to 3% of the total oil produced per year.

Solution: Implementation of an AI-based system with computer vision that monitors territories and equipment in real time. The system analyzes the video stream from the cameras, recognizing signs of leaks, such as uncharacteristic changes in the environment or on the surface of the equipment. In addition, the system is able to identify the source of the leak and automatically send an alarm to operators and security service.

Result:
• Leak detection has accelerated by 80%, allowing you to take measures to eliminate them faster.
• Leaks that previously went unnoticed due to their small size are now effectively detected, which has led to a reduction in total losses to 0.5% of the total volume.
• Due to early detection and elimination of leaks, environmental damage and related fines have been reduced by 90%.