Neural Approach for Calculating Permeability of Porous Medium

Permeability is one of the most important properties of porous media. It is considerably difficult to calculate reservoir permeability precisely by using single well-logging response and simple formula because reservoir is of serious heterogeneity, and well-logging response curves are badly affected by many complicated factors underground. We propose a neural network method to calculate permeability of porous media. By improving the algorithm of the back-propagation neural network, convergence speed is enhanced and better results can be achieved. A four-layer back-propagation network is constructed to effectively calculate permeability from well log data. Spontaneous potential, resistivity of deep lateral log, resistivity of micro-gradient log, resistivity of micro-normal log, Interval transit time of acoustic log and resistivity of shallow lateral log are selected as the inputs, and permeability is selected as the output. There are 35 and 40 units used in the two hidden layers, respectively. During the training course, the correlation coefficient between the calculated permeability and the standard pattern is as high as 0.9937, the average absolute error between them is 0.046μm² and the average relative error is only 1.93%. For practical applications, the average relative error between the calculated permeability and actual permeability is also as low as about 10.0%.