Time-lapse data matching using a recurrent neural network approach

Time-lapse seismic data acquisition is an essential tool to monitor changes in a reservoir due fluid injection, such as CO 2 injection. By acquiring multiple surveys the exact same location, authors can identify by analyzing difference data. However, analysis be skewed near-surface seasonal velocity variations, inaccuracy, and repeatability parameters, other inevitable noise. The common practice (cross equalization) address this problem uses part of which are not expected design matching filter then apply it whole data, including area. Like cross equalization, train recurrent neural network (RNN) on parts excluding area infer reservoir-related RNN learn time dependency unlike that processes based local information obtained window. determine method various examples compare with conventional filter. Specifically, we start demonstrating ability approach two traces test prestack 2D synthetic Then, verify enhancements 4D signal providing reverse migration images. measure using normalized root-mean-square predictability metrics find that, some cases, our proposed performed better than approach.