Using waterflood performance to characterize flow-units in a heterogeneous reservoir [electronic resource] /
Abstract (Summary)
USING WATERFLOOD PERFORMANCE TO CHARACTERIZE FLOW-UNITS
IN A HETEROGENEOUS RESERVOIR
Vamsi Krishna Alla
Simulation of waterflood performance is used to verify the characterization of flow-units
in a heterogeneous reservoir. BOAST98 software was used for simulation of waterflood
performance. Two adjacent five spot patterns in the Stringtown field in West Virginia
were selected to be evaluated in this study.
The predicted permeability along with the porosity from Artificial Neural Network
(A.N.N) was used for primary characterization purpose and identification of flow-units in
the two five spots patterns.
Two alternate simulation models were developed. The simulation model was developed
using the available data and those predicted using A.N.N. to verify that the flow-units
model was an accurate description of the field. One model had two layers representing
two flow units whereas the second model had three layers representing three flow units.
The output from the simulation runs was compared with 10 years of actual production
history of both the wells. A close history matching was obtained with about 15-25% of
total volume water injection rates into the individual patterns from each of the 6 wells.
Though both the models were good in predicting the output, the second model which was
developed for three flow units was better.
iii
Bibliographical Information:
Advisor:
School:West Virginia University
School Location:USA - West Virginia
Source Type:Master's Thesis
Keywords:oil field flooding reservoir engineering neural networks computer science
ISBN:
Date of Publication: