Waterflooding is an important method in optimizing field development, and best-in-class waterflood management practices increase the economic value of the field, and zero waste management efforts.
Mature waterfloods often present significant reservoir management challenges.
Conventional approaches such as operational spreadsheets and reservoir simulation are ineffective for actively managing waterflooding; either too simplistic or challenging to (re)calibrate in a short timeframe for operational decision-making.
In this webinar, an in-depth
look is taken at how an AI-driven, cloud-native waterflood management technology optimally blends physics-based and data-driven approaches for fast and reliable subsurface modeling.
The application characterizes the reservoir through a heterogeneous, dynamic, directed graph, depicting connections between injector and producer nodes.
These super-fast simulations are generated using a graph neural network (GNN).
The GNN captures spatial and temporal patterns and leverages critical physics to improve model accuracy and generalization.
The proposed methodology was successfully applied to many reservoirs worldwide and delivered significant value in terms of CAPEX/OPEX reduction and
overall improvement in reservoir operations.
In this webinar, the results are presented for a carbonate field with more than 150 wells, 60 years of history, and a 50% water cut.
The GNN model was trained with test-set (the last 12 months of production history was held out a test set) accuracy of 90% and then used to optimize the waterflooding strategy for the next six months.
The optimal scenario resulted in a 26,100 STB/D increase in oil production without a drastic change in water production level, only by adjusting the injection rates and producers' operational conditions without drilling or major workovers.
The modeling approach presented
in this webinar has many benefits for actively managing waterflooding and significantly improving workflow efficiencies:
- compared to conventional workflows, the time spent building and (re)training the model is demonstrated to be 90% faster
- super-fast simulations with GNN
- improved model accuracy/generalization leveraging a physics-informed machine learning
- more robust decision-making through uncertainty quantification
- significantly shorter decision cycles in waterflood operations for well control optimization to increase oil recovery and/or reduce water production
