Calibrating one-dimensional velocity model for downhole microseismic monitoring using station-pair differential arrival times based on the differential evolution method

For microseismic monitoring, in order to obtain accurate microseismic locations, a good estimation of the velocity model for the region covering the monitoring stations and microseismic sources is crucial. Perforation shots are generally used to estimate a velocity model suitable for microseismic location. However, the origin times of the perforation shots are generally not accurate. To mitigate the effect of inaccurate origin times of perforation shots on calibrating the velocity model, we propose to search for a velocity model fitting for station-pair differential arrival times instead of absolute arrival times from perforation shots. Another advantage of using station-pair differential arrival times is that waveform cross-correlation can be used to estimate more accurate differential times because of waveform similarity among stations for perforation shots. Due to high nonlinearity of the objective function for estimating one-dimensional velocity model, the differential evolution (DE) method for solving high-dimensional global optimization problems is utilized in the optimization. Compared to the grid-search method, the DE method is much more efficient. Synthetic tests based on a downhole microseismic monitoring system show the effectiveness of the proposed method to recover the velocity model. We also test our DE-based method by using perforation shots for a real microseismic monitoring project. Compared to the well sonic velocity model, station-pair differential arrival times are better fitted and perforation shots are also more accurately relocated with the calibrated velocity model. (C) 2016 Elsevier B.V. All rights reserved.