In the absence of 3D dynamic models for conceptual reservoir development and production forecasting, reservoir engineers are left with only one fallback option – the use of tank material balance models which may not fully capture the physics of flow in complex reservoir systems. In order to reasonably account for the varying factors that may affect production performance, a bit of ingenuity must be applied. This paper describes the workflow involved in developing a comprehensive integrated asset model that is able to predict production performance under different reservoir configurations. First, using available petrophysical information, a method is presented to estimate recovery factor ranges and the corresponding residual saturations which is then used to constrain material balance models in prediction of ultimate recovery. The reservoir is then divided into drain sections with conceptualized target wells and the entire reservoir system is modeled as a connected multi-tank model with varying in-place volumes. The final steps involve building an integrated asset model consisting of wells and facility equipment and generating production forecasts. In this study, key factors that impact the complex multi-tank model behavior and the resulting production forecasts are outlined. Boundaries were then set for these factors and multiple simulation runs generated for proxy model building using Neural Networks. However, it was noted that most of these factors were temporal and did not influence the overall cumulative production overtime. Hence for this study a fixed time period was used to evaluate uncertainties and generate forecast ranges. This paper will be beneficial to reservoir engineers tasked with developing strategies for reservoir exploitation and forecasting pre-drill production profiles. It will be of special interest to engineers inclined to automation, script writing and the use of machine learning in subsurface analysis.

The existing decline curve analysis (DCA) equations, some with valid theoretical justifications, cannot directly react to changes in operating conditions. Thus, they all assume constant operating conditions over the flowing life of a well. This however is an obvious oversimplification. This paper begins by briefly reviewing Gilbert's equation for flowrate prediction and then the C-curve and Logistic growth model DCA theories. The above review serves to identify well key flow indicators (KFI) and performance drivers. Subsequently, a forecasting approach which involves building artificial neural network (ANN) frameworks and training them on well KFI data is presented. Using trained ANNs, production forecasts were generated for three oil wells in the Niger-Delta producing from separate reservoirs under different flow regimes. The results were compared to forecasts from traditional DCA methods and material balance simulation, as well as with future production from the wells themselves. The results indicated that trained ANNs are capable of generating better performance curves than traditional DCA, with forecasts tying closely with results of material balance simulation and measured future well production rates. The ability of trained ANNs to evaluate the effect of changes in operating conditions (i.e. FTHP, GOR and water-cut) on production profiles and reserves drainable by wells, allows for scenario forecasting which is invaluable in field development planning. This is illustrated with field cases. This paper also presents a novel approach to evaluating the optimal hyperparameter configuration (i.e. the number of layers, neuron count per layer, dropout, batch size and the learning rate) required to minimize the loss function whilst training an ANN on any given dataset. This should prove invaluable to engineers and geoscientists integrating deep learning into sub-surface analyses.