Multiscale/Multiphysics Modeling and Simulation of Bubble Collapse: Plasma, Gas, Liquid and Solid Interactions
Abstract
Cavitation is the transformation of liquid into vapor, driven by depressurization. In the naval context, cavitation is a highly undesirable phenomenon that produces noise, hull vibration, material erosion, and performance degradation in propulsors. Despite significant progress in understanding cavitation flow physics, some aspects of cavitation remain poorly understood, including (1) plasma generation during bubble collapse, (2) the mechanisms driving the erosion of solid surfaces during bubble collapse, and (3) the impact of dissolved gases on bubble dynamics. The proposed research will systematically study the multiphysics-multiscale phenomena involved in the collapse of multicomponent bubbles, focusing on plasma formation and the role of adjacent deformable solid surfaces.To accurately describe the multiphysics nature of the problem, we propose a hierarchy of models, based on first-principles that include multiphase-multicomponent mixtures with liquid, gas, and plasma physics. The proposed model involves physics that spans a wide range of time and length scales. In order to solve the problem efficiently, we develop amultiscale computational algorithm that integratesthe variational vultiscale method with superresolution techniques based on artificial intelligence. Under our framework, the solution to partial differential equations is obtained through the direct-sum of a fine-scale solution and a coarse-scale solution. The former is obtained via approximation of an inverseproblem using either a trainable Residual Graph Convolutional network (Res-GCN) or aseries expansion. The latter is obtained via a non-trainable finite-element constrained graph convolutional network. The variational multiscale framework allows us to develop the final algorithm as a finite element method, which makes the extension to fluid-structure interaction straightforward. The solid will be modeled as a fully nonlinear material at finite deformation. We will apply this model to investigatebubble and plasma dynamics of ultrasound and laser-induced bubbles near a solid surface.The successful undertaking of this project will leapfrog the US Navy#s capabilities to predict cavitation-induced erosion. The proposed research has far-reaching impact for DoD problems related to cavitation in ship and submarine propellers. This project has potential to be the basis forhighly-sought predictive models of cavitation-induced erosion as well as technological advances to delay and/or control cavitation erosion, thus contributing to maintaining the stealth superiority of the US Navy.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 13, 2025
- Source ID
- N000142512096
Entities
People
- Hector Gomez Diaz
Organizations
- Office of Naval Research
- Purdue University
- United States Navy