Abstract
In a brownfield with 55 years of oil production located in the middle of the Amazon region in Ecuador, with important operational challenges influencing operational continuity and production optimization, it is essential to optimize operational workflows, expedite decision-making processes, and reduce implementation times. Innovating within the field presents a daily challenge, implementing technological solutions is critical for maximizing oil production and the return on investment of an asset. Located in the remote Ecuadorian Amazon Region, this mature field represents 19% of the national oil production; supported by more than 10 years of waterflood recovery, the main artificial lift system is electro-submersible pumps (ESP) which represents 95% of the asset's lifting mechanism. With increasing flow assurance challenges, the traditional evaluation time for a new lifting technology takes no less than 2 months, resulting in an associated deferred production inherent with every decision-making process. This document narrates how the introduction of state-of-the-art telemetry and communication systems paired with novel cloud-based platforms supported the field optimization strategy in both maintaining operational output and increasing production.
Data science is an interdisciplinary field that uses scientific methods, processes, algorithms, and systems to extract knowledge and insights from structured and unstructured data (Harvard Data Science Review, 2024). According to the definition, the workflow developed to select candidates involved several operational variables that were made available by a cloud corporate system as well as traditional files that are processed and uploaded manually. Currently, the field has 260 wells, each one with no less than 10 variables that involve production data, depth, tortuosity (deviation and dogleg), economic evaluation, casing ID, production potential, rigless operations, downtime, and mechanical failures. This information was incorporated into a collaborative Data Science platform which facilitated the digitalization of complex engineering evaluation processes beyond standard programming. The final solution provided a scalable digital screening workflow to reduce the time spent on decision-making.
Comparing a manual workflow created using a traditional tool like Excel to an automated workflow using data science tools revealed time savings. The automated process takes approximately 70 hours, compared to a traditional 480 hours for the manual process every time a new evaluation is required. This reduction translates into 85.5% of time savings. In addition, the digital candidate selection flow counts with real-time data updates. This leads to the conclusion that the outcome will be an updated candidate portfolio, tailored to the current conditions of each oil-producing well within the asset. It offers speed and efficiency, operating continuously 24/7 without interruptions. It ensures high consistency and accuracy by adhering to predefined instructions, significantly reducing the likelihood of errors.
The digital screening workflow is also adaptable to any type of analysis intended to be carried out within the asset, which means the model will receive new instructions but will retain the original structure. This model has been successfully tested in selecting high tortuosity wells to implement a new shorter and more efficient ESP design, resulting in a portfolio of 20 candidates that represent an opportunity to increase about 2000 oil barrels per day (BOPD). This paper describes the challenges, solutions, and benefits of implementing a Digital Candidate Screening Workflow through Data Science Tools.
