DEVELOPMENT OF EOR PREDICTIVE MODEL WITH HYDROCARBON GAS INJECTION

• Main Author: Putra, Dananjaya
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/47260
• Institution: Institut Teknologi Bandung

Processes for oilfield development have to be optimized to increase the cumulative oil production. A process of developing oil fields that is successful in increasing oil production is hydrocarbon gas injection. Hydrocarbon gas injection has been widely applied in several fields, both in offshore and onshore fields. The utilization of hydrocarbon gases as injection fluid has shown very good results, with a significant increase in the oil production. However, the evaluation method for injection gas hydrocarbon needs a simulation model. The evaluation with simulation requires a lot of time and cost, therefore a fast and accurate evaluation method is needed to predict the performance of hydrocarbon gas injection.. The Artificial Neural Network (ANN) model used as predictive model, with the ability ANN model could calculate a complex problem. ANN is a method that developed by applying the principles of brain performance. ANN method can predict a complex case, with some input values are reservoir parameters that have been determined. Reservoir parameters based on 20 parameters with the number of experiments is 5,482 data, with the injection pattern used in the simulation model is 5-spot. The ANN model was developed to calculate the cumulative of oil production, initial oil production time, maximum flow rate, time at maximum oil flow rate, and oil flow rate at the end of the simulation. From all of these values based on ANN models will form a production profile which will be compared with the production profile of the simulation model.