Research on the Method of Well Test Model Diagnosis Based on BP Neural Network

摘要

10 typical well test theoretical models were studied this paper, then extracted pressure derivative and wavelet transform of pressure derivative as the characteristic curve, and according to self-organization, self-learning and adaptive characteristics of neural network, making use of modular neural network structure, constructed the appropriate BP neural network classifier model in order to diagnose 10 well test models. Finally, the method is feasible proved by instances.