This paper presents the reservoir modeling and an improved engineering workflow to perform a technical evaluation of an EOR miscible hydrocarbon gas injection project using horizontal wells. This project was performed in the Fateh Mishrif, a large carbonate reservoir in a Middle East offshore operating environment. A new methodology to generate synthetic forecast of EOR oil and gas profiles is proposed. The paper describes the technical approach followed to assess the impact of uncertainties and the process used to carry them forward into a probabilistic uncertainty analysis.
Experimental Design and proxy model techniques with associated algorithms, uncertainties, bias, and accuracy vs. simulation models have been discussed through numerous papers. These papers usually present discrete results such as Cumulative oil, STOIIP, or NPV though proxy allowing a simple and time saving solution compared to resource consuming compositional simulations.
The new tool developed here allows for rapid re-computing of a full “synthetic” EOR oil and gas profile (or Response Function) build on 6 essential parameters (EOR breakthrough time, oil ramp up, peak oil, peak oil time, decline, cumulative oil), which are identified as the typical signatures of EOR oil and gas response to miscible gas injection. The EOR profiles are designed for patterns representing the Field Development Plan. The 6 parameters are individually computed through their own proxy models and are a function of 19 inputs parameters. Each proxy model is built based on Experimental Design simulations with Latin-hypercube method which runs hundreds of simulation cases based on the 19 uncertainty parameters. Iterations process with a sufficient number of runs ensures the reliability of the proxy model (Polynomial Regression) vs. simulations with minimizing error.
The response of EOR oil and gas can be predicted as a function of a range of uncertainty with parameters such as rock properties, PVT, as well as operational parameters such well performance, gas injection rate, and WAG management. A work based on lab data, simulation history match, and literature has been achieved for the description, selection, and settings of the boundary values of each parameter to minimize bias.
Response Functions are ready for roll up process for the EOR Field Development Plan through a spreadsheet and for decision making by use of uncertainty analysis.