Finite Element Modeling and Assessment of Acoustic Scattering by Unexploded Ordnance in Variable Seafloor Environments
Dr. Aaron Gunderson | Applied Research Laboratories: The University of Texas at Austin
The goal of this study will be to develop a three-dimensional (3D) finite element (FE) model capable of predicting near-field and far-field scattering by unexploded ordnance (UXO) at arbitrary burial within variable profile sediment environments. This model will be validated using existing experimental scattering data from UXO surrogates in seafloor environments, implementing measured bathymetry, interface roughness, and sediment geoacoustic properties into the model. This project will build on earlier work in which a 3D FE model template was constructed, capable of solving for scattering from targets at variable exposure through a flat interface. The template takes advantage of a hybrid near-field/far-field evaluation technique and a numerically determined Green’s function approach to maximize computational rapidity and wide applicability. Incorporation of interface variability into the model opens the possibility of training UXO classifiers using databases populated with modeled UXO response in different seafloor environments, assisting with UXO detection and recovery missions. Experimental data from multiple sites will be used to evaluate model accuracy over a range of environmental conditions and UXO types and sizes.
The flat interface FE model template applies a nontraditional scattering formulation, which greatly improves model accuracy, but complicates the solution process when the interface is not perfectly flat. Several methods have been considered for incorporating interface variability and roughness into the model, and these methods will be applied and tested for accuracy and efficiency. The project will begin with incorporation of large-scale bathymetry and interface variability into the model and will progress to inclusion of small-scale interface roughness. Experimental validation will occur after each stage. Far-field scattering will be determined through the numerical Green’s function approach and the hybrid nearfield/far-field technique. This technique limits the size of the physical domain required for FE computation, thereby permitting 3D models to be solved as rapidly as possible while avoiding the symmetry constraints imposed by two-dimensional (2D) models. Known target/environment geometries and measured geoacoustic properties from benchmark experiments can be incorporated for high experimental representation within the model.
While SERDP has previously funded investigations into 2D FE modeling of buried UXO scattering, 3D FE modeling efforts remain far less explored. Unlike 2D models, 3D models allow for asymmetric target and environment geometries, permitting consideration of UXO structural asymmetry, UXO tilt relative to the interface, asymmetric interface variability/roughness spectra, and the presence of nearby clutter. 3D FE methods can model any UXO in any configuration within any environment with high accuracy. The models will also be unique in their use of a numerical Green’s function approach to determining the far-field scattering. This approach is significantly more accurate than using analytic halfspace solutions or other approximations to the Green’s function in complex environments and does not require knowing or applying the Green’s function before solving. Once validated, the FE model template can generate synthetic data for in situ UXO modeling and classifier training, bypassing the need for costly and time-consuming experimental data acquisition.