Applying Neuro-fuzzy Model to Predict Reservoir Properties from Seismic Attributes - A Case Study in an Oil Field in Iran

摘要

We applied neuro-fuzzy model to predict reservoir properties from seismic attributes. We used local linear neuro-fuzzy model tree (LOLIMOT) algorithm to train our model. This model uses well log data and seismic attributes in a well location for training. The trained neuro-fuzzy model is then used to predict reservoir properties from relevant seismic attributes in other wells. We used this method in an oil field in central part of Iran and predicted the porosity of reservoir from seismic attributes. The results are then compared with those obtained by neural network methods such as probabilistic neural network (PNN) and multi layer forward neural network (MLFN). Neuro-fuzzy model, comparing with traditional neural network, shows better performance in predicting reservoir properties from seismic attributes.