Complex lithology automatic identification technology based on fuzzy clustering and neural networks

The complex mineral composition and special lithology cause great difficulty to well logging evaluation in volcaniclastic rock reservoir. At present, the accuracy of Lithological discrimination is very low. Fuzzy clustering method combined with Back Propagation (BP) neural network are applied to recognize lithology using logging data of volcaniclastic reservoir in H basin, based on the layer-wise method for logging curves combining intra-layer difference method with clustering analysis method. The results of the application show that the recognized lithology results are in good agreement with the result of core description. The coincidence rate of accuracy is more than 80%.