Neural Fuzzy System Development for the Prediction of Permeability From Wireline Data Based on Fuzzy Clustering

Abstract

Permeability is the most important tool for precise reservoir description and modeling. Despite the advances and modification in different methods for permeability evaluation such as well testing and well logging, the most exact method is core analysis, which is expensive and time consuming. Complexity, nonlinearity and vagueness are some reservoir parameter characteristics, which can be propagated simply by an intelligent system. With the advantage of fuzzy sets in knowledge representation and the high capacity of neural nets in learning knowledge expressed in data, the authors propose a neural fuzzy system to estimate permeability from wireline data in one of the Iranian heterogeneous oil reservoirs. After selecting suitable input, training the fuzzy model was made. The predictions of permeability values from the model were in good agreement with those values obtained from the cores. It was also observed that the fuzzy c-means cluster technique gives better prediction of permeability than neuro-fuzzy model without a clustering.