As former oil and gas operators, we knew there must be a better way to run field operations especially when we looked at how technology was improving other well-monitoring and production accounting processes.
We set out to build a solution to manage the modern challenges in the oilfield. Our goal: build the industry’s most advanced SaaS-based operations platform to achieve operational excellence for North America’s producers.
EZ Ops oilfield software began by digitizing oilfield operations
First developed in 2015, EZ Ops oilfield software is a central information hub for operation teams to share field data and communicate in real-time. It equips operations teams to automate scheduling and maintenance activities such as chemical injections, pigging and equipment inspections to minimize downtime, reduce trips, and optimize production.
EZ Ops also helps upstream producers reduce their ESG risk with automatic alerts, monitoring oil tanks to prevent leaks, and streamlining regulatory compliance management and reporting.
And we’ve seen incredible success. To date, EZ Ops has saved our customers around $48-million in time and operating costs and reduced carbon dioxide emissions by 4,880 tonnes.
But to become the industry’s leading SaaS-operations platform, we needed to go beyond field data capture by adding machine learning and artificial intelligence capabilities.
In 2021, EZ Ops partnered with the Alberta Machine Intelligence Institute (Amii), one of the world’s leading centers of artificial intelligence and machine learning, to truly optimize oil and gas operations management.
Beyond field data capture: how EZ Ops is using machine learning for optimal oilfield operations management
Prioritizing operator actions to improve oil and gas production
Oilfield operators typically plan their day based on routine visitation schedules, guided by their personal experience and available information at the time. But managing field operations in this way is time-consuming and can lead to more costly operations and operator error, as the amount of data to analyze grows.
EZ Ops generates a list of high-priority tasks for operators to action to take the guesswork out of their day. Now, we’ve added artificial intelligence to make that task prioritization even more efficient.
Our machine learning model identifies and prioritizes the actions of operators based on the impact on oil and gas production. By using data recorded both by frontline operators with information collected by the wells themselves plotted over time, the model finds trends and patterns that give a better understanding of what tasks to move to the top of the list.
Using hybrid reinforcement learning techniques, EZ Ops’ new dynamic routing and scheduling capabilities provide a prioritized task list from the model’s suggestions. Operators can then decide whether to add them to their schedule or not.
The outcome? Operators can efficiently plan their daily schedule in under three minutes with certainty they are doing the highest value work during their shift, with the least drive time and lowest greenhouse gas emissions.
Predicting future costs and production
EZ Ops already captures a significant amount of historical production data. But we wanted to use machine learning methods to better predict customers’ future production and costs.
The challenge? Oilfield production generates massive amounts of data and different data types including rate-time, pressure, and well-log data. To be able to predict future production and optimize well performance, we have to merge two different data sources – software interface data and well data, which come from different origins with different formats and have different frequencies.
Fortunately, most problems come with some solutions in machine learning. We turned to a method called Principal Component Analysis (or PCA) to address this challenge. And with advanced regression modelling, our machine learning algorithms can forecast the production rate and other valuable information about a well.
The benefit of having a precise prediction of future production and costs? Producers can adjust prices more efficiently, and operators get insights on upcoming events such as an increase in oil well pressure or the exhaustion rate of the well’s parts.
Amii case study: EZ Ops uses machine learning to get the most out of oil and gas wells
Curious to dive deeper into how EZ Ops enhanced our AI and machine learning capabilities? See the full case study.
This case study was developed in collaboration with the AI Pathways Partnership, made possible with funding from Prairies Canada.
