Parameter evaluation of low-permeability reservoir

More and more attention is paid on low-permeability reservoirs nowadays. This paper proposes an neural network (ANN) method to calculate permeability of porous media. By improving the algorithm of BP neural network, convergence speed is enhanced and better result can be achieved. A four-layer BP network is constructed which can effectively calculate permeability from well log data. Spontaneous Potential (SP), Resistivity of Deep Later Log (RLLD), Resistivity of Micro-gradient Log (Rmt), Resistivity of Micro-normal log (Rmd), Interval Transit Time of Acoustic Log (AC) and Resistivity of Shallow Later Log (RLLS) are selected as the inputs, permeability is selected as the output. 35 and 40 units are respectively used in the two hidden layers. 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 0.046 mu m2 and the average relative error is only 1.93%. For practical application, the average relative error between calculated permeability and actual permeability is also as low as about 10.0%.