Finite-Aperture Ultrasound Field Reproduction in Coarse-Grid FDTD via Operator-Consistent Source Reduction

Full-waveform inversion (FWI) typically employs finite-difference time-domain (FDTD) methods to solve the wave equation under the point-source assumption. For computational efficiency, the coarsest grid that satisfies dispersion and stability requirements is generally preferred. However, this grid may not resolve the finite aperture and spatial phase characteristics of a transducer, thereby introducing imaging artifacts and errors. This paper proposes an Operator-consistent reduced source that maps a local fine-grid radiation field to node-specific forcing histories for a prescribed coarse-grid operator. These histories are reused in repeated simulations without online fine-grid computation. On a 0.50 mm grid, the method reduced root-mean-square (RMS) field relative error by 96.4% compared with the Direct coarse-grid aperture source. In controlled FWI with independent observations, the sound-speed root-mean-square error was 2.14 m/s, compared with 4.94 and 5.17 m/s for the Point source and the Direct coarse-grid aperture source, respectively. Hydrophone measurements, blind ring-water validation, and three repeated phantom acquisitions further assessed field reproduction and reconstruction performance. The method enables coarse-grid FDTD simulations to represent the radiated fields of finite-aperture transducers whose device-scale features are not directly resolved by the target grid.