Fast and Accurate Traveltime Inversion for Locating Passive Seismic Events

In this work, a fast and accurate nonlinear traveltime inversion is presented for estimating hypocenter location parameters and origin time of a passive seismic event. To reduce the computational cost, a coarse three-dimensional gridded velocity model is adopted in which traveltimes are calculated using the eikonal equation. First, a systematic grid search algorithm is applied to minimize the objective function constructed by fitting the model response traveltimes to a finite set of observed data using an l_1 norm, as the derivative is not required. The hypocenter location parameters and origin time are obtained as the grid search technique seeks a global solution for the nonlinear and multimodal objective function. Second, the conjugate gradient algorithm is applied to improve the accuracy of the derived location parameters and origin time estimates. The conjugate gradient algorithm is immune to local minima since the starting estimates, obtained by the grid search method, are close enough to the global minimizers. To prove this inversion concept, the algorithm is implemented to a synthetic case study where the model is generated using parameterization that could represent a recording setup used during hydraulic fracture monitoring of a tight hydrocarbon reservoir, or in a CO2 sequestration project.