Determining the water saturations in thin-bedded turbidites using wire-line logs is difficult; errors in Sw calculation frequently result in uneconomical completions. Consequently, current Brushy Canyon completion decisions include expensive core information to provide an acceptable indicator of oil saturation in order to compensate for the Sw calculation problem. Completion decisions can be improved and less core data is needed using a new method that correlates wire-line logs with core measured bulk volume oil (ΦSo).
A neural network was trained and tested using density and neutron porosity plus shallow and deep resistivity logs as input variables. The neural network was trained to predict the ΦSo product from whole core analysis.
The trained and tested neural network was then used to estimate ΦSo in 25 additional Brushy Canyon wells that were not used in the training, but had the same four wire-line logs. A ΦSo cutoff of 22 units was determined and values greater than the cutoff were summed through the perforated interval in each well. The summed bulk volume oil of the 25 wells was plotted versus the first year’s total production. The plot suggests that ΣΦSo greater than 20,000 units will usually result in an economical new well or reentry completion.
During the course of optimizing the neural network architecture, valuable insights into network architecture design were gained. For this type of study, less complex architectures produced robust testing results, indicating that the solution, though non-linear, is still reasonably simple.
The method should be useful when evaluating behindpipe completion opportunities in the Brushy Canyon interval of the Delaware Sands in the Permian Basin. Re-completion costs are lower than new well costs; thus thin zones with high values of ΦSo are potential targets.