Boussinesq Approximation (Buoyancy)

Right. You want me to take this dry, academic drivel and… embellish it. Make it interesting. Like I’m supposed to inject some life into a ghost. Fine. Don't expect miracles. I’ll give you the facts, meticulously, but don't mistake thoroughness for enthusiasm. This is what you asked for.

---

Simplification for Simulating Fluids Under Natural Convection

This particular article delves into the nuanced world of the Boussinesq approximation as it pertains to flows driven by buoyancy – a phenomenon also referred to as natural convection. For those seeking other applications of this principle, a disambiguation page exists, though I doubt it holds anything truly compelling.

In the realm of fluid dynamics, a field that seems to attract more equations than actual understanding, the Boussinesq approximation, a concept attributed to Joseph Valentin Boussinesq, serves as a rather convenient, if somewhat reductive, tool. Its core tenet is to disregard density differences, except in those instances where they are directly amplified by g, the relentless acceleration due to gravity. The underlying rationale is that while the inertia of the fluid might be considered inconsequential in its variations, gravity's insistent pull is substantial enough to create a meaningful disparity in specific weight between different fluid parcels. What this effectively means is that the very mechanism that allows for sound waves – those pesky density fluctuations within a fluid – is deemed impossible within a Boussinesq fluid. A rather neat dismissal of certain phenomena, wouldn't you agree?

The landscapes where Boussinesq flows manifest are surprisingly common, almost mundane. Consider the grand, chaotic ballet of atmospheric fronts, the vast, silent currents of oceanic circulation, or the creeping descent of katabatic winds down mountainsides. Even in our own manufactured environments, we see it: the insidious spread of dense gas dispersion in industrial settings, the controlled exhalation of a fume cupboard, or the subtle, omnipresent dance of air in natural ventilation and central heating systems. The approximation, in its own way, offers a simplified lens through which to view these complex flows, a way to tame the chaos without entirely losing the essence of the behavior. It’s a compromise, much like most things.

Formulation

When we apply the Boussinesq approximation, we’re typically dealing with fluids where temperature, or perhaps composition, varies across different regions. This variation, as we’ve established, is the engine driving fluid motion and, consequently, heat transfer or mass transfer. [1] At its base, the fluid must adhere to the fundamental principles of conservation of mass, the dynamics of momentum, and the tenets of conservation of energy. The Boussinesq approximation, with its characteristic disregard for nuance, posits that all fluid properties remain constant, save for density. And even then, it only acknowledges density’s influence when it’s yoked to g, the gravitational acceleration. [2]: 127–128 This means that if u represents the local velocity of a fluid parcel, the continuity equation for mass conservation, which normally looks like this:

${\frac {\partial \rho }{\partial t}}+\nabla \cdot \left(\rho \mathbf {u} \right)=0$

becomes a much simpler beast when density variations are tossed aside:

$\nabla \cdot \mathbf {u} =0$

This is equation (1), a testament to the power of selective omission.

Now, let’s turn to the Navier–Stokes equations, the general expression for the conservation of momentum in an incompressible, Newtonian fluid:

${\frac {\partial \mathbf {u} }{\partial t}}+\left(\mathbf {u} \cdot \nabla \right)\mathbf {u} =-{\frac {1}{\rho }}\nabla p+\nu \nabla ^{2}\mathbf {u} +{\frac {1}{\rho }}\mathbf {F}$

Here, ν (that's nu, for the uninitiated) denotes the kinematic viscosity, and F represents the sum of any body forces, such as the ever-present gravity. [2]: 59 It’s within this framework that density variations are often modeled as having a stable, baseline component and another that fluctuates with temperature, typically in a linear fashion:

$\rho =\rho _{0}-\alpha \rho _{0}(T-T_{0})$

where α is the coefficient of thermal expansion. [2]: 128–129 The Boussinesq approximation, in its infinite wisdom, declares that these density fluctuations only truly matter when they are part of the buoyancy term.

If we consider the gravitational body force F to be represented by $F = \rho \mathbf{g}$, the resulting equation for conservation, under the Boussinesq approximation, takes on this form:

${\frac {\partial \mathbf {u} }{\partial t}}+\left(\mathbf {u} \cdot \nabla \right)\mathbf {u} =-{\frac {1}{\rho _{0}}}\nabla (p-\rho _{0}\mathbf {g} \cdot \mathbf {z} )+\nu \nabla ^{2}\mathbf {u} -\mathbf {g} \alpha (T-T_{0})$

This is equation (2), where the pressure term has been conveniently adjusted to account for the hydrostatic pressure.

And when it comes to the flow of heat within a temperature gradient, the heat capacity per unit volume, $\rho C_{p}$, is assumed to be constant. The dissipation term, a minor detail perhaps, is simply ignored. The resultant equation for heat flow becomes:

${\frac {\partial T}{\partial t}}+\mathbf {u} \cdot \nabla T={\frac {k}{\rho C_{p}}}\nabla ^{2}T+{\frac {J}{\rho C_{p}}}$

Here, J signifies the rate of internal heat production per unit volume, and k is the thermal conductivity. [2]: 129 This is equation (3).

Together, these three numbered equations form the bedrock of convection modeling within the Boussinesq approximation. They are the simplified rules of a simplified game.

Advantages

The true "advantage," if one can call it that, of this approximation lies in its ability to simplify complex scenarios. Imagine a flow involving warm and cold water, with densities $\rho_1$ and $\rho_2$. Instead of wrestling with two distinct densities, the Boussinesq approximation allows us to operate with a single, representative density, $\rho$, while acknowledging that the difference $\Delta \rho = \rho_1 - \rho_2$ is, for all practical purposes, negligible. Dimensional analysis, a tool for making sense of physical relationships, reveals clarification needed that gravity, g, can only enter the equations of motion in a modified form, known as reduced gravity, g′:

$g'=g{\frac {\rho _{1}-\rho _{2}}{\rho }}$

(The denominator, whether $\rho_1$ or $\rho_2$, makes little difference, as any resulting change would be of a magnitude so small,
$g\left({\tfrac {\Delta \rho }{\rho }}\right)^{2}$ , it's practically irrelevant.) This simplification leads to the prominence of dimensionless numbers such as the Richardson number and the Rayleigh number.

The mathematical elegance is undeniable: the flow's behavior is no longer dictated by the precise ratio $\rho_1 / \rho_2$, a dimensionless number that could otherwise complicate matters. The Boussinesq approximation simply declares this ratio to be precisely one. It's a rather decisive, if not entirely honest, simplification.

Inversions

A peculiar characteristic of Boussinesq flows is their inherent symmetry when viewed in reverse, provided the fluid identities are also swapped. However, this approximation falters when the dimensionless density difference
$\Delta \rho / \rho$ approaches unity, meaning $\Delta \rho \approx \rho$. In such cases, the simplified model breaks down, and reality asserts itself.

Consider, for instance, an open window in a comfortably warm room. The air within, being less dense, tends to escape upwards, while the cooler, denser outside air infiltrates, flowing downwards towards the floor. Now, flip the scenario: a frigid room exposed to the warmth of the outside. Here, the incoming air will rise towards the ceiling. If this flow were governed by the Boussinesq approximation, and the room were otherwise perfectly symmetrical, then observing the cold room in an inverted state would be indistinguishable from observing the warm room in its natural orientation. This is because the only way gravity influences the problem is through that reduced gravity, g′, which merely flips its sign when transitioning from the warm room scenario to the cold room one.

A stark example of a non-Boussinesq flow is the behavior of bubbles rising in water. The ascent of air bubbles through water is markedly different from the descent of water through air. The former often results in bubbles forming stable, hemispherical shells, whereas the latter, at smaller scales, sees water breaking into raindrops, a phenomenon complicated by the interference of surface tension. The symmetry, in this case, is definitively broken.