Bayesian Inversion For Microseismic Velocity Optimization And High Precision Event Location

Microseismic monitoring technology has become one of the most important methods to evaluate hydraulic fracturing in unconventional oil and gas exploitation.The velocity model in downhole monitoring has a serious effect on the accuracy of the microseismic image.In the traditional method,the initial velocity model is established based on acoustic logging signals,and then the velocity model is further optimized by using the perforation shot information.Due to the limited number of perforation shots and narrow coverage of ray paths,this method has poor stability in the inversion process,and it may cause large event location errors in microseismic monitoring in complex formations.It is difficult to accurately describe the fracture morphology of hydraulic fracturing.To improve the location accuracy of microseismic events,we introduced more effective constraints with the microseismic events to the velocity optimization.To get a flexible velocity model structure and improve the event location accuracy,we carefully discussed the velocity optimization method in complex geological conditions.Also,it was committed to meeting the needs of field microseismic monitoring.The main achievements are as follows:To solve the problem of large initial velocity model error and low event location accuracy under complicated geological conditions,we first introduced the Bayesian transdimensional inversion to downhole velocity calibration.During the inversion process,the layer number,the layer depths,and the velocity values were updated simultaneously,and finally a sparse velocity model was obtained under the constraint of the perforation shots and microseismic events.Both theoretical and field examples showed that our method was less affected by acoustic logging information,and the location accuracy of the microseismic events can be effectively improved.The simultaneous inversion of velocity model and microseismic event locations still faces the risk of coupling on the condition of fitting the observed travel times,some systematic deviations may be unavoidable.To avoid the errors in microseismic event location results,an Incremental Pseudo-Master(IPM)method was proposed in this paper.It combines the advantage of the master-event location method and the Bayesian inference.In the IPM method,the accuracy of the microseismic event locations and the velocity model were iteratively improved,and the number of model parameters was reduced to improve the inversion stability.The synthetic and field tests all represented that the IPM method was suitable for velocity optimization under different geological conditions,and the microseismic event location accuracy can be largely improved.When the strong anisotropy in the hydraulic fracturing area seriously affects the location accuracy of the microseismic events,we deduced the anisotropy parameters of VTI(Vertical Transverse Isotropy)media based on the IPM method.In the fixed dimensional inversion,the IPM method successfully predicted the uncertainty of the anisotropic parameters.In the transdimensional inversion,the IPM method successfully obtained the sparse equivalent anisotropic velocity model,which effectively improved the location accuracy of the microseismic events.To solve the problem that the MCMC(Markov Chain Monte Carlo)sampling method in Bayesian inversion needs many iterations and the calculation process needs a lot of time,we developed an efficient raytracing algorithm for isotropic and VTI media.It combined the simple principle of the shooting method and the strong applicability of the shortest-path method.The algorithm was simple and efficient,and the high precision travel times and ray paths can be deduced from sparse nodes.It was beneficial to efficiently obtain the optimized velocity model,which facilitates real-time microseismic monitoring.The velocity optimized method proposed in this paper overcomes the problems of the traditional method,such as less constraint information,poor applicability of velocity model,large event location error,and low inversion efficiency.It is suitable for microseismic monitoring under complex geological conditions.The location error is reduced to ten meters,and the inversion efficiency is improved by tens of times.It is suitable for field hydraulic fracturing production and provides a strong guarantee for improving unconventional energy oil and gas recovery.