Neural network modelling and classification of lithofacies using well log data: A case study from KTB borehole site

摘要: SUMMARY A novel approach based on the concept of super self-adapting back propagation (SSABP) neural network has been developed for classifying lithofacies boundaries from well log data. The SSABP learning paradigm applied to constrain by parameterzing three sets data, that is, density, neutron porosity and gamma ray obtained German Continental Deep Drilling Project (KTB). A multilayer perceptron (MLP) networks model was generated in a supervised feed-forward mode training published core sample total 351 pairs input output examples were used self-adaptive weight bias values appropriately updated during each epoch according gradient-descent momentum scheme. actual data analysis suggests is able emulate pattern all KTB identify correctly. comparisons maximum likelihood geological sections with available information existing geophysical findings over area suggest that, addition known main units, namely paragneisses, metabasites heterogeneous series containing partly calc-silicate bearing paragneisses-metabasites alternations former volcano-sedimentary sequences, technique resolves more detailed finer structures embedded bigger units at certain depths region which seems be some significance. efficacy method stability results also tested presence different levels coloured noise. test designed topology considerably unwavering up 20 per cent correlated noise; however, adding noise (∼50 or more) degrades results. Our analyses demonstrate renders robust means classification complex successions borehole thus may provide useful guide/information understanding crustal inhomogeneity structural discontinuity many other regions.