molecalculator

Moles & Molar Mass

Pressure-Based vs Density-Based Solvers in CFD

Guide 01-2081 · Updated · Step by step

Pressure-based and density-based solvers are two approaches used in computational fluid dynamics (CFD) to solve the equations governing fluid flow. Pressure-based methods are widely used for incompressible and low-speed flows, while density-based methods are often effective for compressible flows with strong density changes, such as high-speed gas flow; neither is universally better.

What the Two Solver Names Mean

A CFD solver approximates the conservation laws for mass, momentum and, when relevant, energy on a computational mesh. The two labels describe how the numerical method couples pressure, velocity and density while advancing toward a solution. They do not describe different physical laws: both approaches can solve the same governing equations, subject to the software’s formulation, models and settings.

In a pressure-based formulation, pressure is central to enforcing mass conservation. This is especially important for incompressible flow, where density is treated as constant or changes only slightly and pressure is not obtained from a simple compressible equation of state. In a density-based formulation, density is among the primary variables advanced through the conservation equations, with pressure linked to density and energy through an equation of state for compressible flow.

How a Pressure-Based Solver Works

For incompressible flow, a tentative velocity field may not satisfy continuity: fluid could numerically accumulate in one cell and disappear from another. A pressure-correction procedure adjusts pressure and velocity so that the corrected flow satisfies mass conservation. Algorithms in the SIMPLE and PISO families are common examples, though implementations differ.

In broad terms, the solver performs the following operations during an iteration or time step:

  • Estimates velocity and other flow variables using the current pressure field.
  • Builds a pressure-correction equation from the requirement that the updated velocity obey continuity.
  • Corrects pressure and velocity, then updates additional variables such as temperature or turbulence quantities.
  • Repeats the process until residuals and relevant physical quantities meet the chosen convergence criteria.

Pressure-based solvers can also handle compressible flow. They may use density and an equation of state while retaining pressure-velocity coupling methods suited to a broad range of speeds. Therefore, “pressure-based” does not mean “incompressible only.”

How a Density-Based Solver Works

A density-based method solves conservation equations for quantities such as density, momentum and total energy as a coupled system. Pressure is then calculated from the thermodynamic state, commonly using an equation of state. Coupling the equations can be advantageous when pressure waves and changes in density are important, as in transonic or supersonic flow.

Many density-based methods use flux calculations designed to capture shocks and other steep changes without spreading them excessively across the mesh. Their behavior depends on the numerical flux, spatial discretization, time integration, limiters and boundary conditions—not just on the solver label. Density-based methods can also be applied to low-speed cases, but may need preconditioning or other treatments because the equations become numerically stiff when flow speed is far below the speed of sound.

How to Choose Between Them

Start with the physics and the capabilities of the CFD software rather than assuming one solver is always superior. Consider whether density variation, compressibility and shock waves matter, along with the flow speed, expected temperature changes and the type of calculation.

  • Incompressible liquids or low-speed ventilation: A pressure-based solver is often a practical starting point because pressure-velocity coupling directly addresses continuity.
  • Compressible flow with shocks or high-speed gas dynamics: A density-based solver is often a natural choice, especially when the energy equation and pressure waves are central.
  • Low-Mach-number gas flow with meaningful heating: Either approach may be suitable. Use a formulation that accounts for the required density and temperature changes, and compare convergence and accuracy.
  • Mixed regimes or uncertain behavior: Check the solver documentation, run a controlled comparison if feasible, and verify results against conservation balances, mesh refinement and suitable reference data.

Mach number—the flow speed divided by the local speed of sound—is a useful guide, not a complete decision rule. Compressibility effects are often small at low Mach numbers, but heating, cooling or large pressure changes can still make density variation important. Conversely, a nominally compressible case may not require a particular solver if the software’s pressure-based formulation handles it accurately.

Convergence, Accuracy and Common Problems

A solver that reports convergence has not automatically proved that its answer is physically correct. Residuals measure numerical equation imbalances; they should be considered alongside quantities such as mass flow in and out, forces, pressure drop and energy balance. A result can look stable while remaining sensitive to mesh resolution, boundary conditions or modeling assumptions.

Common difficulties include:

  • Slow or unstable convergence: Check mesh quality, boundary conditions, initialization, time-step size and relaxation or coupling settings. A smaller time step or better initial field may help, but does not fix incorrect physics.
  • Shock smearing or oscillations: In compressible calculations, review the discretization, shock-capturing settings and mesh resolution. A shock is a thin physical transition that may require substantial local refinement.
  • Poor low-speed convergence in a density-based method: The large gap between flow and acoustic speeds can make the equations stiff. Low-Mach preconditioning or a pressure-based alternative may be appropriate if available.
  • Misleading agreement between solvers: Two methods can share the same incorrect geometry, boundary conditions or turbulence model. Agreement is useful evidence, not independent validation by itself.

Comparison at a Glance

The distinctions below are general tendencies, not guarantees about every software package.

  • Primary numerical emphasis: Pressure-based methods emphasize pressure-velocity coupling; density-based methods advance density and related conserved quantities as a coupled system.
  • Common strength: Pressure-based methods are convenient across incompressible and many low-speed applications; density-based methods are often well suited to compressible flows and shocks.
  • Low-Mach behavior: Pressure-based methods commonly perform well; density-based methods may require specialized treatment.
  • Compressible behavior: Both can model compressible flow when formulated for it. Performance depends on implementation and case setup.

These are numerical approaches, not competing physical models. Turbulence, combustion, multiphase flow and heat transfer are separate modeling choices that may be combined with either solver, depending on the software.

Frequently Asked Questions

Is a pressure-based solver only for incompressible flow?
No. It is commonly used for incompressible problems, but many pressure-based formulations also solve compressible flows. Confirm that the selected formulation includes the physics your case needs.

Is a density-based solver always better for supersonic flow?
Not automatically. Density-based methods are often effective for supersonic flows and shocks, but accuracy and convergence also depend on the numerical scheme, mesh, boundary conditions and physical models.

Can I use a density-based solver for a liquid?
Sometimes, if the software and model support the liquid’s relevant behavior. For many ordinary liquid-flow calculations, a pressure-based approach is a straightforward choice; the solver name alone does not determine whether a setup is valid.

Does solver choice change the conservation laws?
No. Both approaches are numerical ways of solving conservation equations. They can differ in variable arrangement, coupling and discretization, but should represent the intended physics when configured appropriately.

How do I know which result to trust?
Check convergence, mass and energy balances, mesh sensitivity and agreement with experimental or established reference data when available. Solver-to-solver agreement can add confidence, but cannot replace validation.

Related

General steps for current mainstream versions; menus can differ by device and version. See the disclaimer or report an error.

More in this section

The whole Moles & Molar Mass section →