Semester

Spring

Date of Graduation

2020

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Petroleum and Natural Gas Engineering

Committee Chair

Shahab Mohaghegh

Committee Member

Mehrdad Zamirian

Committee Member

Kashy Aminian

Committee Member

Sam Ameri

Abstract

Data-Driven Reservoir Modeling (DDRM), commonly referred to as Top-Down Modeling (TDM), is a relatively new and cutting-edge approach to the traditional numerical reservoir modeling and simulation techniques. DDRM uses artificial intelligence and machine learning in tandem to construct full-field models using measured data instead of calculations that refer to equations derived from averaged values and type curves. TDM allows all of the measured data from a field to be combined and used towards generating predictions of the production on a well by well basis for a specific field. Due to TDM not using the traditional physics-based approach, it is subjected to a plethora of criticisms within the industry. Therefore, the purpose of this thesis is to confirm the capabilities of TDM versus data synthetically generated using a Numerical Reservoir Simulator (NRS). To do this, the fluid flow through porous media will be modeled via the use of a traditional NRS; this way, everything is known about the reservoir in question. The data generated will then be exported and used towards the construction of the TDM. To complete the proposed objectives of this thesis, an application will be used to aid in the development of a TDM. All of the data used in order to develop and history match the TDM will have been generated via the NRS; this is done to confirm the abilities of TDM forecasting existing wells behavior. Once the TDM has been constructed; the forecast data will be compared to that from the NRS to validate the ability of the TDM.

Download

Included in

Engineering Commons

Share

COinS