In absence of PVT laboratory experiments data on representative fluid samples, it is usually difficult to choose the appropriate PVT correlations to calculate oil viscosity. This difficulty will increase when input data to PVT correlations (oil API gravity, initial gas-oil ratio, specific gravity of separator gas and temperature) vary along the flow from one section to the other in the production system. However, the accuracy of these correlations has become inadequate for the best estimations. The achievements of the Artificial Intelligent (AI) techniques alone open the door to use the hybrid system. This research focuses on the use of predictive NFuzzy model that is a result of combination of the learning capabilities of Neural Networks (NN) with the reasoning capabilities of Fuzzy Logic as a hybrid intelligent system. The proposed approach is based on clustering the PVT data into three clusters (heavy, medium and light oil) based on solution gas oil ratio. Around 500 to 2500 data points for each oil viscosity obtained from Middle East and worldwide laboratory measurements. The data were separated into two parts, 70% of data for training and the rest 30% were utilizing for testing. The present model used to estimate dead viscosity, saturated and under-saturated oil viscosity.
Based on this result, we conclude that NFuzzy exhibits a robust predictive capability for estimation of oil viscosity by providing a good match with the measured values. The additional data samples were selected to compare and validate this model.

Fluid classification is a critical factor in decision of reservoir and production problems. Reservoir fluid can be classified into five types according to laboratory and production data as black oil, volatile oil, gas condensate, wet gas and dry gas. In this work a novel application of Neural Networks (ANN) is presented. Based on production and laboratory data neural networks model is developed for automatic classification of reservoir FLUID. More than 450 samples of five types of reservoir fluids are used to develop the neural network model. About 70 % of data are accepted for neural network training, 15 % for validation and 15 % are used as test set. The importance of different input fluid properties in classification was studied.
The different types of architectures for different groups of input data were tested to select the optimal neural network architecture by fitness criteria. The optimized neural network model was capable of classifying the reservoir fluids with high accuracy. The performance of ANNs models was determined by classification quality index and network error.
The model has been applied successfully to classification of Yemeni fluids using different range of parameters. The results show that the proposed novel ANN model can achieve high accuracy.