JX Nippon Improves Asset Evaluation Process Efficiency, Reliability, and Collaboration with Agile Reservoir Modeling Workflow
As part of its initiative to accelerate digital transformation of upstream business, JX Nippon successfully applied Agile Reservoir Modeling in the DELFI cloud environment to speed up and increase the reliability of asset evaluation and promote collaboration between the teams.
JX Nippon needed to complete a reevaluation of producing assets within three months, including uncertainty analysis using the subsurface model to optimize well placement. Lack of time and resources presented a challenge. The existing subsurface model could not reproduce actual production results, so the accuracy of asset evaluation was questionable. In addition, because of COVID-19, teams needed to work from multiple locations across borders.
The Agile Reservoir Modeling workflow explores a broader uncertainty space—geologically and dynamically—simultaneously and in a shorter time. The workflow narrows down the range of uncertainty by comparing it with history to obtain a subsurface model that is effective for multiple predictions incorporating uncertainty.
With the scalability of high-performance computing (HPC), the cloud-based environment can complete the entire process from generating multiple geological realizations, searching for valid models by filtering by the history data, predicting the forecast performance, and optimizing the development plan in ultrafast time.
The cloud-based environment enables real-time collaboration between teams from multiple locations across borders—even team members working from home because of COVID-19.
With improved efficiency, collaboration, and reliability, the asset reevaluation was completed within three months based on quantifiable facts.
More than 900 geological realizations were generated within two weeks, including model validation through dynamic simulation.
More than 200 simulations for well placement sensitivity studies were completed within a week.
The uncertainty range of predictions was quantified with 200 realizations within 8 hours using 50 virtual machines.
More than 900 realizations were explored, and the uncertainty range was narrowed down. Ten history-matched models were used for forecasting and development planning.
More than 10 engineers from JX Nippon and Schlumberger collaborated on the cloud environment from more than five locations to complete the project on time.
The scalable cloud environment dedicated to this project enabled instant data sharing of over 2TB of data, eliminating the need for time-consuming data transfers via email or other file transfer systems.