Fast integration method for averaging polydisperse bubble population dynamics
Spencer H. Bryngelson · 2025
Read Fast integration method for averaging polydisperse bubble population dynamics by Spencer H. Bryngelson (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-22920.txt.
Introduction
Spencer H Bryngelson tackles a costly math problem in the study of bubbly flows, where gas bubbles of many sizes move in a liquid. Models need average values from the full spread of bubble sizes. Those averages change over time before they settle, and that early phase matters for real machines.
Computing the averages is slow because the sums swing up and down fast. Bryngelson offers a faster integration method that cuts the work at each point in a simulation. Engineers who model bubbly flows can use the approach to run larger cases without losing the time-changing detail they need.
Catalog record
| CREATOR | Spencer H. Bryngelson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-22920
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Fast integration method for averaging polydisperse bubble population dynamics Spencer H. Bryngelson Abstract Ensemble-averaged polydisperse bubbly flow models require statistical moments of the evolving bubble size distribution. Under step forcing, these moments reach statistical equilibrium in finite time. However, the transitional phase before equilibrium and cases with time-dependent forcing are required to predict flow in engineering applications. Computing these moments is expensive because the integrands are highly oscillatory, even when the bubble dynamics are linear. Ensemble-averaged models compute these moments at each grid point and time step, making cost reduction important for large-scale bubbly flow simulations. Traditional methods evaluate the integrals via traditional quadrature rules. This approach requires a large number of quadrature nodes in the equilibrium bubble size, each equipped with its own advection partial differential equation (PDE), resulting in significant computational expense. We formulate a Levin collocation method to reduce this cost. Given the differential equation associated with the integrand, or moment, the method approximates it by evaluating its derivative via polynomial collocation. The differential matrix and amplitude function are well-suited to numerical differentiation via collocation, and so the computation is comparatively cheap. For an example excited polydisperse bubble population, the first moment is computed with the presented method at 10 − 3 10^{-3} relative error with 100 times fewer quadrature nodes than the trapezoidal rule. The gap increases for smaller target relative errors: the Levin method requires 10 4 10^{4} times fewer points for a relative error of 10 − 8 10^{-8} . The formulated method maintains constant cost as the integrands become more oscillatory with time, making it particularly attractive for long-time simulations. Mechanistically, the transient behavior of the moments is set by two effects: resonance detuning across bubble sizes, which causes phase mixing of oscillations, and viscous damping, which removes radial kinetic energy. The proposed formulation isolates the oscillations while keeping the remaining terms smooth, so accuracy does not deteriorate at late times. 1 Introduction Bubbly flows occur in numerous engineering applications, including ship propulsion systems [ 1 ] , hydraulic machinery [ 2 ] , spillway flows [ 3 ] , and biomedical applications such as lithotripsy and histotripsy [ 4 , 5 , 6 ] . In these systems, polydisperse bubble populations result from nucleation, breakup, and shedding of larger vapor regions [ 7 , 8 , 9 ] . The ensemble of bubble size oscillations characterizes the associated collective bubble dynamics. Individual bubble dynamics can even cause shock waves comparable in magnitude to those in the carrier flow [ 10 , 11 , 12 ] . Further, small numbers of bubbles or small void fractions can change large-scale pressure wave propagation [ 13 ] , and polydispersity meaningfully influences the acoustics in bubbly liquids [ 14 , 5 ] . For practical use in ship and biomedical settings, the implementation requires consistent boundary conditions at inlets and walls, prescribing the driving pressure signal at the inflow or outflow, specifying the bubble size distribution, and ensuring the ensemble moments are compatible with the carrier-flow boundary conditions. Computational modeling of these flows presents challenges due to the range of scales. Individual bubble radii span from microns to millimeters [ 11 ] , while bubble or vapor clouds and turbulent structures extend to meters or larger depending on the application [ 15 ] . The temporal scales are disparate: bubble natural frequencies and collapse times occur on timescales of microseconds, while flow observation times extend to seconds or more [ 11 ] . This scale separation makes direct numerical simulation of resolved flow dynamics computationally expensive or intractable for engineering applications. To address this challenge, ensemble-averaged sub-grid modeling approaches view the individual bubble dynamics as less important than the statistical properties of the bubble population [ 16 , 17 , 18 , 19 ] . This view is referred to as two-way coupled Euler–Euler modeling and remains valid as long as the void fractions are sufficiently small. Euler–Euler ensemble average models result in equations for the continuous liquid phase that are forced, in a two-way coupled manner, by statistical moments of bubble properties, such as radius and radial velocity. The bubble variables become stochastic, Eulerian fields. These methods contrast with Lagrangian approaches that track individual bubbles, or particles more broadly, as points with their own dynamics [ 20 , 21 , 22 ] . Recent advances in Eulerian multiphase closures [ 23 , 24 ] develop well-posed models for polydisperse flows, while open-source frameworks [ 25 , 26 , 27 ] provide computational tools for modeling the multiphase flow. The ensemble-averaging approach is illustrated schematically in fig. 1 , where the statistical properties of a bubble population determine the mixture-averaged flow properties. Figure 1: Schematic of ensemble-averaging approach for polydisperse bubble populations. Individual bubble dynamics are replaced by statistical moments of the bubble size distribution, which couple to the continuous liquid phase through mixture-averaged properties. The ensemble-averaged approach requires the computation of statistical moments of the bubble size distribution, which involve integrals over the bubble population. Alternative approaches to handling polydispersity include quadrature-based moment methods (QMOM) and conditional quadrature methods (CQMOM) [ 28 , 29 ] , which evolve the moments directly through transport equations. These methods are often used in the particle-laden flow community and, likewise, require closure assumptions for unclosed moments as well [ 29 ] . The ensemble-averaged approach computes moments from the bubble dynamics. Still, a computational challenge arises from the polydisperse nature of the bubble population: bubbles of different equilibrium sizes oscillate at different natural frequencies. These frequencies are typically uncoupled from one another, and their distributions lead to complex oscillatory behavior in the bubble size distribution over time [ 14 , 30 , 31 ] . At a physical level, a polydisperse population behaves as an ensemble of damped radial oscillators. The spread of natural frequencies across sizes drives dephasing and cancellation in the population-averaged moments, while viscosity sets the decay rate of these transients. In the context of moment computation, this oscillatory behavior entails highly oscillatory integrands and, thus, expensive computation via traditional quadrature methods. Conventional approaches to this problem employ large numbers of quadrature points to resolve the highly oscillatory integrands [ 32 ] , or accept that the moment closure may not be accurate. However, this strategy becomes computationally expensive; each additional quadrature point requires solving an additional advection partial differential equation (PDE). For example, achieving acceptable accuracy can require thousands of quadrature points, as well as thousands of additional PDEs to accompany the Navier–Stokes-like PDEs. These quadrature requirements mean the computational cost of moment evaluation exceeds that of the flow solver. The challenge of computing integrals with very oscillatory integrands is not unique to bubbly flow modeling and has been recognized in the numerical analysis community. Traditional quadrature methods, including Gaussian quadrature, exhibit poor convergence properties when the integrand oscillates more rapidly than the quadrature rule resolves, as they have not yet reached their asymptotic convergence rate. Filon 1928 first addressed this limitation by proposing trading numerical integration for differentiation. Levin 1996 , Levin 1997 extended the concept by developing an approach for constructing differential equations with non-oscillatory solutions that are solved efficiently using standard numerical methods. Later developments include adaptive Filon methods [ 36