Mathematical Theory of Compressible Viscous Fluids

Compressible viscous fluids occupy a central place in continuum mechanics. Their behaviour is governed by a system of nonlinear partial differential equations that couple mass, momentum, and energy conservation with constitutive laws for stresses and heat flux. Mathematically, the theory poses deep challenges: the equations are hyperbolicparabolic, they admit shocks and other singularities, and the existence of global classical solutions is unknown in general. The present article summarises the principal mathematical structures, the main analytical results, and the open problems that continue to stimulate research.

1. Governing Equations

Let \(\Omega\subset\mathbb{R}^d\) (\(d=2\) or \(3\)) be a bounded domain, \(t\ge0\) time, \(\rho(t,x)\) the density, \(\mathbf{u}(t,x)\) the velocity, and \(\theta(t,x)\) the absolute temperature. The compressible NavierStokesFourier (CNSF) system reads

\[\begin{aligned}\partial_t\rho + \nabla\!\cdot(\rho\mathbf{u}) &= 0,\\[4pt]\partial_t(\rho\mathbf{u}) + \nabla\!\cdot(\rho\mathbf{u}\otimes\mathbf{u})       + \nabla p &= \nabla\!\cdot\mathbb{S},\\[4pt]\partial_t(\rho e) + \nabla\!\cdot(\rho e\mathbf{u})       + \nabla\!\cdot\mathbf{q} &= \mathbb{S}:\nabla\mathbf{u},\end{aligned}\]\]

where \(p=p(\rho,\theta)\) is the pressure, \(e=e(\rho,\theta)\) the internal energy, \(\mathbb{S}\) the viscous stress tensor, and \(\mathbf{q}\) the heat flux. For a Newtonian fluid

\[\mathbb{S}= \mu(\nabla\mathbf{u}+\nabla\mathbf{u}^\top)          +\lambda\,(\nabla\!\cdot\mathbf{u})\mathbf{I},\qquad\mathbf{q}= -\kappa\nabla\theta,\]

with shear viscosity \(\mu>0\), bulk viscosity \(\lambda\ge -\tfrac{2}{d}\mu\), and thermal conductivity \(\kappa>0\). The equations are supplemented by initial data and suitable boundary conditions (e.g., noslip velocity and prescribed temperature).

2. Entropy and Thermodynamics

Thermodynamic consistency is encoded in the entropy inequality. Introducing the specific entropy \(s(\rho,\theta)\) and using the Gibbs relation \( \theta\,{\rm d}s = {\rm d}e + p\,{\rm d}(1/\rho) \), one obtains the balance

\[\partial_t(\rho s) + \nabla\!\cdot(\rho s\mathbf{u})       + \nabla\!\cdot\!\left(\frac{\mathbf{q}}{\theta}\right)      = \frac{1}{\theta}\Bigl(\mathbb{S}:\nabla\mathbf{u}        - \frac{\mathbf{q}\cdot\nabla\theta}{\theta}\Bigr)\ge0 .\]

The righthand side is nonnegative thanks to the positivity of \(\mu,\lambda,\kappa\), providing a crucial apriori estimate that underlies most existence theories.

3. Weak Solutions and Existence Theory

Because classical solutions are only known locally in time, the community has focused on weak (distributional) solutions. A function \((\rho,\mathbf{u},\theta)\) is a weak solution if it satisfies the integral forms of the three balance laws together with the entropy inequality for all smooth test functions compactly supported in spacetime. The most celebrated result is due to Lions (1998) and later refined by Feireisl, Novotn and coauthors. Under the structural assumptions:

• Equation of state of the form \(p(\rho,\theta)=R\rho\theta + a\rho^\gamma\) with \(\gamma> \frac{d}{2}\),
• Viscosities and conductivity bounded away from zero and infinity,
• Finite initial energy,

they proved the existence of globalintime weak solutions for any finite energy data. The proof relies on: (i) a careful regularisation (artificial pressure, artificial viscosity), (ii) uniform estimates derived from the energy and entropy inequalities, (iii) compactness arguments exploiting the effective viscous flux, and (iv) the renormalised continuity equation of DiPernaLions.

4. Strong Solutions and Conditional Regularity

When the initial data are smoother and satisfy a compatibility condition, localintime strong solutions exist (MatsumuraNishida, 1980). A major open problem, often called the compressible NavierStokes regularity problem, asks whether such solutions can be extended globally under natural physical bounds (e.g., bounded density and temperature). Partial progress has been achieved via conditional regularity criteria: if the density stays away from vacuum and the gradient of velocity belongs to a critical Lebesgue space, then the solution remains smooth (Sun, Wang, and Zhang, 2011). These criteria resemble the celebrated BealeKatoMajda condition for incompressible flows.

5. LowMach and Incompressible Limits

Another rich line of inquiry studies the limit as the Mach number \(Ma\to0\). Formal scaling shows that solutions of the CNSF system converge to solutions of the incompressible NavierStokes equations, provided the initial data are well prepared (no fast acoustic oscillations). Rigorous justification has been achieved using relative entropy methods, which quantify the distance between a compressible solution and an incompressible target. The analysis reveals how acoustic waves are damped by viscosity and how the pressure decomposes into thermodynamic and incompressible parts.

6. Numerical Approximation

Designing stable discretisations that respect the underlying entropy structure is essential for reliable simulations. Finite volume and discontinuous Galerkin schemes equipped with entropystable fluxes have become standard. Recent work emphasises structurepreserving methods: they inherit a discrete version of the energyentropy balance, which guarantees unconditional stability for large time steps and prevents nonphysical oscillations near shocks.

7. Open Problems and Research Directions

• Uniqueness of weak solutions. The weaksolution theory yields existence but not uniqueness. Convex integration techniques have produced infinitely many admissible weak solutions for certain equations of state, suggesting that additional admissibility criteria (e.g., maximal entropy production) may be required.
• Vacuum states. While the LionsFeireisl theory handles vacuum in the density, the behaviour of temperature near vacuum remains delicate. Understanding regularity and possible singularities when \(\rho\to0\) is an active topic.
• Longtime dynamics. For bounded domains, whether weak solutions converge to equilibrium and at which rate is largely unknown. The entropy inequality provides a Lyapunov functional, but quantifying its decay demands new functional inequalities.
• Coupling with other physics. Realistic models involve elasticity, chemical reactions, or multiphase interfaces. Extending the rigorous theory to such coupled systems poses substantial analytical challenges.

8. References

1. P.-L. Lions, Mathematical Topics in Fluid Mechanics, Vol.2, Oxford University Press, 1998.
2. E. Feireisl, Dynamics of Viscous Compressible Fluids, Oxford University Press, 2004.
3. R. Danchin, Global existence in critical spaces for compressible NavierStokes equations, Invent. Math., 2000.
4. T. Sun, Y. Wang, and X. Zhang, On the regularity criterion of threedimensional compressible NavierStokes equations, J. Differential Equations, 2011.
5. R. S. Strickland, Entropystable finite volume methods for compressible NavierStokes equations, Comput. Fluids, 2020.
6. A. J. L. Shee, Relative entropy and the lowMach limit for compressible viscous fluids, Arch. Ration. Mech. Anal., 2019.