Estimating Original Hydrcarbor In Place using Neural Network

One of the methods that is very useful in estimating Original Hydrocarbor In Place (OHIP) for oil and gas reservoirs is Monte Carlo simulation. In routine Monte Carlo simulation, some predefined density functions used for constructing histogram s such porosity, saturation and Net to Gross (NTG). This process adds some uncertainty to the results due to density function fitting. In this paper, we introduce using neural network to construct histograms. For each parameter, a single network trained and then using uncertainty analysis the OHIP determined. The routine Monte Carlo simulation is not stable in which the calculated OHIP change considerably in each run. The most achievement of using neural network is the stability of this method. Enhancement in lowering the CPU time is the other preferably of applying neural network.