A Neural Network-Based Sensitivity Analysis of Reservoir-Related Parameters during Laser Perforation in Downhole Conditions in Limestone

摘要

High power lasers are capable of cutting and removing rocks efficiently and they are considered as one of the appropriate substitutions for conventional shaped charge perforation methods due to- its superiorities over the current shaped charge methods that the most important one is considerable permeability increase and no need to have costly post-perforation operations to reduce the effect of formation damage caused by perforation. To optimize the laser perforation in limestone in both technical and economical aspects and improve the efficiency of laser perforation in limestone, the effectiveness of the important parameters should be considered. In this paper, a neural network approach has been used for backward elimination sensitivity analysis of the effective reservoir-related parameters during laser perforation in limestone. For this purpose, firstly a feed-forward with back-propagation neural network has been developed to predict the specific energy according to the effective parameters like lasing time, saturation, confining pressure, axial pressure and pore pressure. The data is related to around 100 laser perforation laboratory tests on limestone core samples. Then sensitivity analysis was done and finally the effectiveness of each parameter was determined successfully.