Neural Network Approach to Predict Average Reservoir Pressure

Abstract

The average reservoir pressure is an important parameter in petroleum engineering which is 
utilized in almost all reservoir and production engineering calculations. Normally, the average reservoir 
pressure is obtained from build-up test and can be measured only when the well is shut in. Besides that, it 
requires an update from time to time. It also has a great economic impact caused by shutting in the well during 
the entire test. However, the objective of this study is to establish a neural network model that can map certain 
relationship that controls previous production performance of the reservoir to predict current average reservoir 
pressure without shutting in the wells.