Velocity analysis of surface seismic data with VSP data as constraints

Abstract

The full waveform inversion with the surface seismic data has the advantage of wide inversion range, but it has poor resolution than with the VSP data. In contrast, inversion using the VSP data gives a high resolution result, but it has spatial limit of positioning the well. In this paper, we have developed the damped least square algorithm using the interval velocity from VSP data as constraints to solve the full waveform inverse problem in frequency domain. In addition, we compared and analyzed the inversion results with and without VSP constraints. The velocity model computed from inversion using VSP constraint shows more accurate velocity structure of subsurface. Additionally, since the damped least square method calculates the sensitivity matrix unlike nonstationary iterative methods such as the steepest descent method, we use this information to select proper frequencies which improve the efficiency of the inversion.