CLIENT
The client is an oil and gas exploration and production company operating in Germany and Austria. The company's journey began in 2016 when founded as a joint venture between a mining and investment company. After strategic shifts and acquisitions, the entity evolved into a sustainability focused oil and gas producer in September 2021.
CHALLENGES
The challenge was to replace outdated legacy systems, take the first steps towards digitalization, and move away from manual processes, especially Excel-based workflows. The client faced specific challenges, including the need for an automated back-allocation system (PDMS – Production Data Management System), a rod pump solution, and a water control advisory system (POV – Proof of Value).
OBJECTIVE
Balance production injection patterns | To achieve optimized water injection, the goal was to keep the voidage replacement ratio (VRR) within defined limits on each reservoir block of the field while maximizing oil production. The manual optimization process was challenging and time-consuming, prompting the need for a standardized and automated water injection advisory system.
The client’s field presented unique characteristics with around 45 producers and 11 injectors. Producing well water cuts ranged from low to high with gross production in the low to moderate range. Some wells produced primarily water but needed to be operational in winter to regulate system temperature. Historically, insufficient emphasis was placed on achieving balanced production/injection patterns across the field.
SOLUTION
- Automation | A comprehensive approach was taken, unifying data from over 6 sources, implementing an automated back allocation system with reporting capabilities, and establishing surveillance and monitoring for production and injection wells. Additionally, an automated water injection advisory system was introduced.
- P# Logic Driven | Best engineering practices were captured in P# with subject matter expert involvement. Accurate production and injection volumes were crucial for developing a trustworthy water advisory solution. Recommendations included redistributing available water injection volumes, shutting in wells, or bringing some inactive wells into production, while adhering to constraints such as VRR per block, maximum injection rate, and water availability in the field.
- ML Driven CRM Waterflood Prediction Model | Using a Capacitance Resistance Model (CRM), a Machine Learning algorithm employed for history matching and forecasting, enabled connectivity analysis through the utilization of production and injection data. The objective function, used for the optimization exercise was represented with Net Present Value (NPV), considering oil price, water production cost, and injection cost. Proposed injection rates for the next three months were generated, advising the optimum injection rates that fulfill the well, reservoir and field constraints.
RESULT
The disparate legacy data sources were consolidated into a unified data model. The enhanced water injection advisory system achieved balanced production/injection patterns. The water injection advisory POV was accepted by the client, relieving engineers from manual processes as they could now focus on advised actions.
