Study on the Division of Strata Sequence Based on Combinatorial Neural Network
Jun Du, Rui Liu · International Conference on Electric Information and Control Engineering · 2012
Using geophysical well logging data to divide the strata sequence is an important step in the research of oil and gas reservoir geology. However, because of the complicacy in strata sediments, there isn't a distinct range on the well logging curve between two different strata and we also cannot find a clear and definite mathematical formula to distinguish the strata sequence. Considering the excellent capacity of solving complicated mapping relationship, we proposed a kind of combinatorial neural network based on RBF and BP to ascertain the sequence of strata. Then, by using the data which had been divided artificially as the inputs to the model, results from combinatorial neural network and other algorithms were compared to define the most appropriate method. From the final results presented in this paper, the superior capacity of the combinatorial neural network can be seen clearly.