Search for a command to run...
Abstract The Midland Basin has been extensively developed with horizontal wells, which leads to a large percentage of new wells experiencing potential performance degradation. This paper will present an empirically derived formula for predicting degradation on the new (child) well, given distances and produced volumes from the existing (parent) wells. This has major impacts on well spacing decisions of new wells and calculations of remaining economic inventory. To calculate the degradation experienced by existing child wells, first the performance must be estimated for the child well's performance without any depletion. Calculating this baseline performance is accomplished through a neural network trained on all the public data in the Midland Basin, to account for variables including well spacing, frac size, lateral length and location. The actual child well actual performance (1-year cumulative oil) can then be compared against the predicted baseline and the degradation calculated. With this degradation percentage for each child well, an empirical equation to predict degradation can then be tuned. The equation developed here has inputs of distances in both the horizontal and vertical direction, as well as parent well volumes produced. This empirical equation was calibrated for each formation in the Midland Basin, based on all possible parent/child pairings in the basin (60,000 pairings). The results were different profiles of predicted degradation depending on the parent well formation. This formula can then give rapid estimates of degradation for potential new wells given both distances and produced volumes of parent wells. This provides a complement to the much more time intensive method of fracture and reservoir simulation to calculate this depletion degradation.