Redistribution of Steam Injection in Heavy Oil Reservoir Management to Improve EOR Economics, Powered by a Unique Integration of Reservoir Physics and Machine Learning
| Autor: | Stylianos Kyriacou, Pallav Sarma, Dakin Sloss, Paul Orland, Mark Henning, Ganesh Thakur |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Physics
Engineering Petroleum engineering business.industry Steam injection Heavy oil reservoir 02 engineering and technology Redistribution (cultural anthropology) 010502 geochemistry & geophysics 01 natural sciences Reservoir simulation 020401 chemical engineering Reservoir engineering Reservoir modeling Reservoir management 0204 chemical engineering business 0105 earth and related environmental sciences |
| Zdroj: | Day 2 Thu, May 18, 2017. |
| DOI: | 10.2118/185507-ms |
| Popis: | The application of a novel modeling and optimization approach is presented, demonstrating the impact of quantitatively optimized steam redistribution in mature heavy oil fields. Results are presented for a steamflood in the San Joaquin Basin in California, demonstrating significant savings of steam and operational costs and significant production increase, ultimately increasing net present value (NPV) by at least 10%. The new approach, termed Data Physics, is based on a novel combination of state-of-the-art machine learning methods with the partial differential equations of reservoir fluid flow, as present in reservoir simulators. Because the presented approach honors reservoir physics, it provides long term predictive capacity and always ensures physically realistic solutions. Further, this method is inherently low order and continuous scale, and therefore orders of magnitude faster than traditional approaches. Combined with an advanced data assimilation algorithm that merges a modified ensemble Kalman filter with quadratic programming, this approach allows rapid and simultaneous integration of production, injection, temperature, completion, maintenance and log data for large fields with thousands of wells. The paper first describes the Data Physics modeling approach as applied to steamfloods. The paper then describes how the model is fit to historical data. A portion of the historical data is used as training data, and the rest is used as test data to understand the predictive capacity of the model. This is demonstrated through both statistical tests and validations of the model's ability to predict incremental production due to the steamflood, and performance of newly drilled wells. Second, the paper describes how the fitted models are combined with advanced multi-objective optimization algorithms and cloud computing to consider thousands of scenarios to optimize steam redistribution. Third, actual field results are presented from the continuous modification of steam injection throughout the reservoir. For the field in this case study, production comes from poorly consolidated sands within the Antelope Shale member of the Miocene Monterey formation with porosity averaging 30%, permeability averaging 2,000 mD and net thicknesses typically between 50 and 300 feet. Structural dip is steep at approximately 60 degrees. The reservoirs are shallow, with depths ranging from 200 - 600 feet TVD. Oil gravity is approximately 13° API. Reservoir pressures are well below bubble point and average 50 - 100 PSI. The most significant new finding is that steam redistribution can be quantitatively optimized rapidly to maximize short and long-term EOR economics. This is important, particularly for mature fields, in order to maximize recovery amidst varying commodity prices. The novelty of the new model is its combination of speed of data integration (less than a week) and runtime (minutes) with long-term predictive accuracy (years or decades). This is due to the unique integration of reservoir physics with fast data-driven methods. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|