Original article was published on artificial intelligence
Oil rigs are one of the most dangerous work sites on the planet. You can never be too careful, and the safety standards of the oil industry frequently undergo changes to ensure the safety of the workers. But, to err is human. So the oil rig industry is slowly moving towards human independent safety systems by deploying AI-based systems. In the next section, we look at a couple of companies and how they are employing next-gen technology.
Roolloos Uses AI For Workers Safety
Rolloos — a Dutch oil and gas company — has partnered with NVIDIA to leverage their hardware to train deep learning algorithms. These algorithms are used to gather insights and develop warning systems based on the feed that is obtained from the onboard cameras. To develop, what the company calls, a Red Zone System, they began by mounting cameras on the rig and capturing 60 days worth of various types of operations performed in all types of weather and conditions. The company then employed human annotators to label every person and object incoming video feed and used that information to build an algorithm that identifies people in a complex rig environment.
Today, the model is so accurate that it can point out at locations within 18 inches and run the whole sequence under 250 milliseconds. Speed is crucial when it comes to warning systems. In the case of false positives or false alarms, the data can be fed to the algorithm to make it better. And, all of this could be accomplished due to NVIDIA as its hardware was running onsite because there is no internet connectivity on rigs.
Rolloos uses NVIDIA data centre GPUs for inference on an NVIDIA T4 Tensor Core GPU.
How Seadrill Uses LIDAR Technology
Seadrill, an international offshore drilling company, has partnered with the Marsden Group to develop a safety solution to assist with Red Zone Management on the drill floor. As a part of the solution, they use a LIDAR system that provides 360-degree coverage of an area generating approximately 700,000 sensor readings per second that are analysed in real time. The algorithms then identify and track the location and movement of people and equipment.
When a worker enters the restricted red zone, the warning systems go off and alert the personnel. The company is also planning to use this tool as a back-up anti-collision system monitor for the drilling package equipment. They have even partnered with Microsoft for using their cloud platform Azure to onboard data science operations.
Over the last few decades, several offshore accidents have led to environmental disasters and the deaths of workers. But with the rise of automation, companies are hoping that future accidents can be reduced.
Not only for the safety measures but also machine learning is now being widely embraced by the traditional oil companies to identify the reserves. For example, as part of a corporate-wide digital transformation, British Petroleum is embracing artificial intelligence to change the way the company works. To this end, BP scientists now explore the potential of new energy deposits using Microsoft Azure Machine Learning service and its automated machine learning capabilities. They build more finely tuned, accurate models in dramatically less time, helping them better gauge available hydrocarbon reserves.
“Safety remains our top priority, and we see a wide range of possibilities for AI to help us make our operations safer,” says Kennedy. “We also anticipate dramatic increases in productivity as we learn to work smarter with AI.”
-Diana Kennedy, Vice President, Strategy at BP
BP has created an AI Center of Excellence within the company to nurture new ideas and see them through to fruition as they learn to work smarter with AI.
Provide your comments below