Forward modeling-free full waveform inversion with well calibration

Marcelo Guarido, Laurence R. Lines, Robert James Ferguson

Full waveform inversion (FWI) has the goal to find the Earth’s model parameters that minimize the difference of acquired and synthetic shots. It is a powerful tool to automatize some complex processes. However, when we talk about seismic data, we are talking about huge datasets, and the FWI shows to be very hard to be applied in large scale, as it requires a large number of synthetic data and migrations at each iteration. In this work, We are proposing an approximation for the FWI that is forward modeling-free, requiring only the migration of the acquired data, which can be pre or post stack, and the optimization driven by a sonic log calibration. We tested the approximation for acoustic inversion in a synthetic 2D Marmousi survey. It is stable and leads to detailed inverted models with reduced computational costs.