Waveform inversion via reduced order modeling

TL;DR

This paper presents a new waveform inversion method using reduced order models (ROM) of the wave operator, which improves robustness and convexity of the inverse problem, enabling better velocity estimation from seismic data.

Contribution

The paper introduces a data-driven ROM approach for waveform inversion that simplifies computation and enhances convexity, making the inverse problem more tractable and robust to data inaccuracies.

Findings

ROM-based misfit function exhibits better convexity than traditional least-squares.

The method can handle approximate data matrices via interpolation and reciprocity.

ROM inversion shows improved convergence from poor initial guesses.

Abstract

We introduce a novel approach to waveform inversion, based on a data driven reduced order model (ROM) of the wave operator. The presentation is for the acoustic wave equation, but the approach can be extended to elastic or electromagnetic waves. The data are time resolved measurements of the pressure wave gathered by an acquisition system which probes the unknown medium with pulses and measures the generated waves. We propose to solve the inverse problem of velocity estimation by minimizing the square misfit between the ROM computed from the recorded data and the ROM computed from the modeled data, at the current guess of the velocity. We give the step by step computation of the ROM, which depends nonlinearly on the data and yet can be obtained from them in a non-iterative fashion, using efficient methods from linear algebra. We also explain how to make the ROM robust to data inaccuracy. The ROM computation requires the full array response matrix gathered with collocated sources and receivers. However, we show that the computation can deal with an approximation of this matrix, obtained from towed-streamer data using interpolation and reciprocity on-the-fly. While the full-waveform inversion approach of nonlinear least-squares data fitting is challenging without low frequency information, due to multiple minima of the data fit objective function, we show that the ROM misfit objective function has a better behavior, even for a poor initial guess. We also show by an explicit computation of the objective functions in a simple setting that the ROM misfit objective function has convexity properties, whereas the least squares data fit objective function displays multiple local minima.