AI/ML applications for upstream oil and gas

More and more oil and natural gas producers are investing in Artificial Intelligence/Machine Learning (AI/ML) applications with mixed results. There’s lots of enthusiasm and a fair bit of vendor hype. Nonetheless, adding AI/ML functionality to existing applications or building entirely new AI/ML applications can produce significant benefits.

The prerequisites to AI/ML application success typically involve advancing the producer’s digital transformation with a focus on:

1. Gathering more operational data
2. Increasing the timeliness of data
3. Managing that data for higher accuracy and completeness
4. Integrating data from many sources effectively
5. Improving data accessibility, typically through better data analytics
6. Consciously promoting multi-disciplinary work

How can affordable AI/ML oil and natural gas applications can produce an attractive net benefit?

Increase profitable production

Most oil and natural gas producers regularly see production enhancement and plant throughput improvement proposals that look good, but don’t produce the promised net revenue increase after implementation.

Adding AI/ML functionality to production optimization applications can weed out proposals that won’t be profitable and improve recommendations that can be profitable. The benefits include:

1. Reducing investment for profitable proposals
2. Avoiding engineering time in developing unprofitable proposals

Reduce unscheduled downtime

Most oil and natural gas producers want to move beyond condition-based maintenance or fixing components once they break. They are frustrated by the variety and frequency of various outages. Too many production facilities operate materially below their nameplate capacity.

Adding AI/ML functionality to predictive maintenance applications can increase the number of days of advance notice of component failure. The benefits include:

1. Largely eliminating the risk of waiting for parts or staff to complete a repair
2. Reducing the length of an outage for repairs. That shorter time decreases revenue lost and generally reduces the cost of the repair

Reduce operating costs

Oil and natural gas producers are perpetually unhappy with their operating costs. With the benefit of hindsight, they see missed opportunities in operating cost statements.

Adding AI/ML functionality to analyze operating costs closer to real-time achieves the following benefits:

1. Identifying opportunities to reorganize work to reduce driving time
2. Reducing trucking costs for oil, clean water and produced water disposal
3. Improving the utilization of equipment
4. Reducing consumption of materials
5. Reducing component inventories

Improve equipment utilization

Oil and natural gas producers contract with service companies for drilling and service rigs. Low equipment utilization leads to lower service company net income and higher producer costs.

Adding AI/ML functionality to better categorize and analyze non-productive time and invisible lost time with more consistency and granularity achieves the following benefits:

1. Reducing travel for staff
2. Reducing mobilization time for heavy equipment and its crew
3. Improving work processes when rigs are in use
4. Improving the utilization of drilling and service rigs

Improve safety

Oil and natural gas facilities are complex and dangerous, with exposed rotary equipment, high pressure, high-temperature operations, large vehicles and hazardous chemicals. Producers invest significant resources to ensure their staff, suppliers, and neighbours are safe.

Adding AI/ML functionality to employee monitoring systems:

1. Focuses supervisors on important alerts and reduces the distraction of false alerts
2. Reduces lost-time incidents
3. Reduces injuries and deaths

Reduce GHG emissions

Oil and natural gas producers are working to reduce their greenhouse gas (GHG) emissions by:

1. Turning methane venting into CO2 flaring. That reduces GHG emissions by 95 per cent
2. Capturing methane venting and flaring into sales of natural gas – turning GHG emissions into revenue
3. Reducing fugitive GHG emissions from wellhead and processing equipment
4. Reducing their energy consumption

Adding AI/ML functionality leads to better measurement, estimating and reporting to identify additional opportunities to reduce GHG emissions.