Globally, leading oil and gas producers are beginning to reap benefits from their investments in data science and machine learning. Most of these investments have been focused on core production operations. Today, machine learning and data analytics are widely deployed for improving oil well discovery and precision drilling.
As the next step, practical artificial intelligence (Ai) solutions in other operational areas are also being explored, such as:
Reducing impact on the environment is another important part that machine learning, Ai, and computer vision can provide for oil and gas producers. Computer vision, for example, plays a huge role in monitoring valve banks and other critical infrastructure for human error as well as wear and tear leading to equipment failure.
Cost efficiencies are another part to consider in the effort to promote advanced facilities to help protect the environment. By saving costs on maintenance failures with the help of computer vision, budgets can then be turned to helping clean up from the aftermath of drilling or other ways, such as research and development of green energy, to keep the environment secure.
ElectrifAi offers a variety of computer vision-based solutions for workplace safety, infrastructure monitoring, and equipment surveillance. Below are a few examples of how we can help the oil and gas industry with computer vision.
Fingerboard Latch Monitoring
Fingerboard latch monitoring on a drilling rig is a critical element of operations that ensures safety and prevents incidents from occurring that could otherwise cost millions of dollars. Most rigs have deployed or are in the process of deploying CCTVs to provide footage of the latches; this solution, however, still requires a human to confirm and validate latch positions.
This is an error prone and expensive problem as real time confirmation across multiple fingerboard latches is required to continue operations. Why is this such a concern? If a latch remains closed while it should have been open, the latch can break or fingerboards may come down.
ElectrifAi’s computer vision models are trained on latch open/close positions and provide real time visual confirmation of latch status. This reduces the need for human spotters, increases safety, and creates an enormous cost benefit.
Stickup Height Measurement
As the world continues to search for untapped oil reserves, drilling at new depths has become the norm. But more than ever, these opportunities come with the need for the highest level of precision and no margin for errors. When drilling depths exceed 30,000 feet, operations are on around the clock. Speed and precision are key for optimal drilling operations.
Stickup height measurement ensures pipe joints are executed at the right time with minimal loss of operations. This is where ElectrifAi’s trained computer vision models track and indicate the stickup height in real time is the most valuable.
Floating Cap Storage Tank Monitoring
The floating roofs innovation in oil storage tanks has been embraced by the industry. Floating roofs minimize the volume of vapor as well as reduces the emission of vapors from tanks, thereby making operations much safer. Estimates suggest there are over 30,000 floating cap oil storage tanks of varying capacities worldwide.
Floating roof tanks need to quickly detect conditions such as inclination, the presence of hydrocarbons and/or water, sand, snow. Depending on the weather, the floating roof could suffer a malfunction. While satellite solutions have been touted as the next frontier to solve this problem, satellites are very expensive and are only as good as the frequency at which images can be sourced.
Storage tank operators need real time solutions based on current images that they control. Drones can be used to fly over tanks and collect images from all sides of the tank. This viable solution is being deployed to determine whether the floating roof is indeed horizontally aligned or is beginning to show signs of inclination. At that point a maintenance crew would need to be sent for repairs.
Using ElectrifAi’s computer vision models, the drone video can be processed very quickly to output only those roofs that actually need to be reviewed. Time to value receiving this information is key. The solution does not require humans to review all the video footage captured by the drones. A tank operator could execute this solution everyday as opposed to a defined frequency of manual checks.
Red Zone Monitoring
Drilling operations are complex with multiple people surrounded by heavy equipment, working in hazardous spaces. Red zone monitoring are designated high-risk areas on the drill floor where the crew needs to perform specific tasks.
ElectrifAi’s computer vision machine learning models allow real time monitoring of equipment and people to determine incursions or dangerous behavior.
Valve Control Bank Monitoring Solution
Trained computer vision ML models can recognize value positions. Using this solution for real time cross checking of valve positions against related work order can help to properly set values. The solution highlights and sends an alert if the value pattern does not match the expected pattern from the work order. Preventing human errors, time and cost associated with cross checking can save companies thousands of dollars.
Conclusion
Our computer vision solutions drive workplace safety and cost reductions. The sample of use cases described here are not exhaustive. We have created many computer vision use cases that go beyond the normal range of computer vision abilities yet seen.
If you would like to learn more about our pre-built machine learning models and computer vision offerings, reach out today! Schedule a custom demo to find out how we can help you accelerate your time to value.
