Application of Multi-Layer Feed Forward Neural Network (MLFNN) for the

M.F. Mahmood, Z. Ahmad


The geophysical formation evaluation plays a fundamental role in hydrocarbon exploration. Porosity is one of the main parameters that determine the amount of oil present in a rock formation. Accurate determination of porosity is a difficult problem due to failure in understanding of spatial porosity parameter distribution. Multi-layer feed forward neural network (MLFN) proved to be a powerful tool for mapping porosity across the whole field and proved to be a powerful tool for mapping complicated relationships in reservoir. In MLFN three layers are involved that is an input layer, an output layer and a variable number of hidden layers. Input for training eight external attributes are used which are P-impedance, S-impedance, density, fluid, lithology impedance, lamda-rho, mu-rho, and Vp/Vs. Five nodes are used in hidden layer and one output node for mapping total porosity of Badin gas field. In this study 3D cube of Badin field and 3 wells are used. The findings proved competence of multi-layer feed forward neural network in the porosity prediction process with an average error of 0.014 [v/v] and the correlation coefficient of 0.91 and helped in studying the lateral variations in the porosity along the reservoir. The A sands show same porosity values along both the well locations, while for B sand the porosity value decreases from Zaur-01 to Chakri-01 well while for C sand the porosity value increases from Zaur-01 to Chakri-01 well.

Full Text:



B. Goodway, T. Chen and J. Downtown, Improved AVO fluid detection and lithology discrimination using Lame petrophysical parameters; lambda rho, mu rho and lambda/mu fluid stack, from P and S inversions, Society of Exploration Geophysics Expanded Abstracts, 16, pp.183-186, 1997.

D.P. Hampson, J.S. Schuelke and J.A. Quirein, Use of multi-attribute transforms to predict log properties from seismic data, Geophysics, vol. 66, no. 1, pp. 220-236, 2001.

D.H. Han, A. Nur and D. Morgan, Effect of porosity and clay content on wave velocity in sandstones, Geophysics, vol. 51, pp. 2093-2107, 1986.

R.B. Latimer, R. Davison and P. van Riel, An interpreter's guide to understanding and working with seismic-derived acoustic impedance data, The Leading Edge, vol. 19, no.3, pp. 242-256, 2000.

M.D. McCormack, Neural computing in geophysics, The Leading Edge, vol. 10, pp. 11-15, 1991.

T. Masters, Practical Neural Network Recipes in C++, Academic Press, London. p.. 493, 1993.0

B.H. Russell, The application of multivariate statistics and neural networks to the prediction of reservoir parameters using seismic attributes, Ph.D. Dissertation. University of Calgary, Alberta, 2004.

R.K. Shrestha, Reservoir characterization of high impedance sands in the Ada field, North Louisiana, USA. Society of Exploration Geophysicists Spring Symposium Technical Program vol. 11, pp. 5-16, 2008

T. Todorov, R. Steward, D.P. Hampson and B.H. Russell, Well Log Prediction Using Attributes from 3C-3D Seismic Data, Expanded Abstracts, vol. 8, pp. 1574-1576, 1998.

M.D. Zoback, Reservoir geomechanics, University Press, Cambridge, p. 449, 2007.


  • There are currently no refbacks.