For the fuzzy distribution of well logging curves, it is very difficult to partition lithology. In order to analyze the well logging response curves in the process of lithology identification, escape particle swarm optimization (EPSO) was proposed to be combined with Elman neural network (ENN), and EPSO-NN hybrid algorithm was produced. The nonlinear dynamical identification system was constructed based on EPSO-NN. The system was employed to adaptively identify the well logging lithology. The results of the engineering cases indicate that this system is feasible and effective in the lithology ...
