Recent advances in neural networks have provided computers (and machines) with intuition - the ability to produce a reasonable result to a problem which is intractable, or unreasonably hard, to deal with by formal logical means. Neural networks can learn complex nonlinear relationships, even when the input information is noisy and less precise. Neural networks have made strong advances in pattern recognition, classification of noisy data, nonlinear feature detection, market forecasting and process modelling. These abilities make the neural network technology very well suited for solving problems in the petroleum industry.

The aim of this paper is to give a synopsis of the areas of petroleum technology in which neural networks have been used with success, and to discuss other potential areas of application. Some of the examples include seismic pattern recognition, permeability predictions, identification of sandstone lithofacies, drill bit diagnosis and analysis and improvement of gas well production. Neural networks technology could also contribute significantly to the analysis, prediction and optimisation of well performance, integrated reservoir characterisation and portfolio management.

Although there has been some advances on the design of optimal network structures, it is still largely an art to determine the best paradigm. The paper highlights key factors in the design or selection of neural networks and the limitations of the commonly used models. It also discusses the use of hybrid expert networks which combine neural networks with rule-based reasoning and fuzzy logic for general decision-making operations.

You can access this article if you purchase or spend a download.