Womersley Number Simulator Back
Biomechanics Simulator

Womersley Number Simulator — Characteristic Number of Pulsatile Blood Flow

Interactive Womersley number alpha = R sqrt(omega rho / mu) calculator. Vary vessel radius, heart-rate, blood density and viscosity to inspect flow regimes, viscous penetration depth and the velocity profile.

Parameters
Vessel radius R
mm
Heart rate f
Hz
Blood density rho
kg/m³
Dynamic viscosity mu
mPa·s

$\alpha = R\sqrt{\omega\rho/\mu}$, $\omega = 2\pi f$. Aortic defaults: $R = 10$ mm, $f = 1$ Hz, $\rho = 1060$ kg/m³, $\mu = 3.5$ mPa·s give $\alpha \approx 13.8$ (inertia dominant).

Results
Womersley number alpha
Flow regime
Viscous depth delta
Peak Re (U=1 m/s)
Pulsatile velocity profile in the vessel cross-section

Blue lines = vessel wall at $\pm R$ / four curves = axial velocity $u(r,t)$ at four phases ($t = 0$, $T/4$, $T/2$, $3T/4$) / orange band = viscous penetration depth $\delta$. Small $\alpha$ gives Poiseuille parabolas, large $\alpha$ produces a central piston with a thin wall boundary layer and a phase lag.

Flow regime map (log alpha axis)

X axis = $\alpha$ (log scale 0.1 to 100) / green band = quasi-steady ($\alpha < 1$) / yellow band = transitional ($1 \le \alpha < 10$) / red band = inertia dominant ($\alpha \ge 10$) / yellow dot = current alpha. The curve is the normalised inertial index (phase lag).

Theory & Key Formulas

Definition of the Womersley number:

$$\alpha = R\,\sqrt{\dfrac{\omega\,\rho}{\mu}},\quad \omega = 2\pi f$$

Viscous penetration depth (Stokes layer) and its relation to $\alpha$:

$$\delta = \sqrt{\dfrac{\mu}{\rho\,\omega}},\quad \alpha = \dfrac{R}{\delta}\,\sqrt{2}\;\;(\text{approx.})$$

Peak Reynolds number (assuming peak velocity $U_\mathrm{peak}=1$ m/s):

$$\mathrm{Re}_\mathrm{peak} = \dfrac{\rho\,U_\mathrm{peak}\,(2R)}{\mu}$$

$R$ is vessel radius [m], $f$ is heart-rate [Hz], $\omega = 2\pi f$ is angular frequency [rad/s], $\rho$ is blood density [kg/m³] and $\mu$ is dynamic viscosity [Pa·s]. The flow is quasi-steady (Poiseuille-like) for $\alpha < 1$, transitional for $1 \le \alpha < 10$ and inertia dominant for $\alpha \ge 10$ (aorta at 1 Hz heart rate).

What is the Womersley number?

🙋
Hemodynamics textbooks always mention the Womersley number — how is it different from the Reynolds number, and why do pulsatile flows need a separate dimensionless group?
🎓
The Womersley number $\alpha = R\sqrt{\omega\rho/\mu}$ measures the inertia-to-viscosity ratio on the time scale of the oscillation. Reynolds is built from the mean velocity, but pulsatile flow has its own oscillatory inertia that needs its own dimensionless group. With this tool's defaults ($R=10$ mm, $f=1$ Hz, $\rho=1060$ kg/m³, $\mu=3.5$ mPa·s) you get $\alpha = 13.79$, the typical aortic value at a 1 Hz heart rate. That sits firmly in the inertia-dominated regime.
🙋
If $\alpha$ is large the velocity profile is no longer a Poiseuille parabola, right? The four-phase curves in the vessel canvas look nearly flat at the centre.
🎓
Exactly — that is the "annular effect." Large $\alpha$ makes the core move as a single slug, while only a thin wall layer (the viscous penetration depth $\delta$) carries shear. At default values $\delta = 0.725$ mm, just 7 percent of the 10 mm radius. The wall fluid is held back by viscosity, so it lags the core in phase, and there are even moments (around $t=T/4$) when the core and the wall move in opposite directions.
🙋
If I drop the vessel radius to 0.5 mm $\alpha$ becomes 0.69 and the regime switches to quasi-steady. So this is essentially capillary behaviour?
🎓
Right. In small arterioles or distal coronary branches with $\alpha < 1$, each instant looks like a fresh Poiseuille parabola. That is the regime where the classic resistance law $\Delta P = 8\mu L Q/(\pi R^4)$ applies. Big vessels with $\alpha > 10$ are the opposite — full pulsatile inertia — and CFD modelling has to use the Womersley analytical solution or solve the unsteady Navier-Stokes equations.
🙋
When I push the heart rate to 3 Hz (180 bpm, intense exercise) $\alpha$ shoots up to 23.9. So exercising makes the aorta even more inertia-dominated?
🎓
Yes. $\alpha$ scales like $\sqrt{f}$ so tripling the heart rate multiplies $\alpha$ by about 1.73. Exercise also raises peak velocity from 1 m/s to 2–3 m/s, so peak Re climbs from 6000 to over 15000, occasionally entering turbulence. In trained athletes pulsation and turbulent bursts coexist in the aorta — exactly why Womersley analysis is core to evaluating cardiovascular risk and to designing prosthetic valves or stents.

FAQ

Linearising the Navier-Stokes equations for a rigid pipe and assuming $u(r,t) = U(r) e^{i\omega t}$ reduces the problem to a Bessel ODE. The solution is $U(r) = \dfrac{G}{i\omega\rho}\left[1 - \dfrac{J_0(\alpha i^{3/2} r/R)}{J_0(\alpha i^{3/2})}\right]$, where $G$ is the pressure-gradient amplitude and $J_0$ is the Bessel function of the first kind. As $\alpha \to 0$ this collapses to the Poiseuille parabola; as $\alpha \to \infty$ it becomes a plug flow with a Stokes boundary layer at the wall. This tool uses a qualitative interpolation between Poiseuille and central piston for legibility — for quantitative analysis use the Bessel solution or unsteady CFD.
Any oscillatory pipe flow uses it: piston-pump hydraulic lines, internal-combustion engine intake/exhaust ducts, wave pumps (artificial hearts), oscillating water columns in ocean engineering and pulsed drug delivery devices. For example, an IC engine intake at $f = 50$–$200$ Hz, $R = 25$ mm, with air ($\rho \approx 1.2$ kg/m³, $\mu \approx 1.8 \times 10^{-5}$ Pa·s) yields $\alpha \approx 130$–$260$, deep into inertial behaviour. That is exactly where designers exploit wave resonance for inertial supercharging. MEMS micropumps, acoustic streaming, oscillation corrections for capillary viscometry and many other devices also depend on this number.
For a rigid pipe $\alpha = R\sqrt{\omega\rho/\mu}$ is enough. For elastic vessels you also need the pipe wave speed $c$ (Moens-Korteweg $c = \sqrt{Eh/(2\rho R)}$) — the full Womersley-Witzig elastic analysis. The parameter $M^2 = (\omega R/c)^2$ enters, and wave reflection / pulse-wave shapes become important. In real aortas pulse-wave speed is 4–10 m/s with $f = 1$ Hz, $R = 10$ mm, so $M \approx 0.006$–$0.016$ is tiny and the quasi-steady wave approximation holds, leaving $\alpha$ alone to control the basic flow structure. As arteries stiffen with age, $c$ rises and pulse-wave velocity (PWV) becomes a clinical biomarker of cardiovascular risk.
Match both $\alpha$ and Re in any similarity study. When $\alpha$ is large the wall Stokes layer $\delta = \sqrt{\mu/(\rho\omega)}$ must be resolved with 5–10 cells, so $\Delta r \le \delta/5$. With aortic $\delta \approx 0.7$ mm you need $\Delta r \approx 0.1$ mm. Use 100–200 time steps per period ($\Delta t \le T/100$), drop 3–5 startup cycles and analyse the 5th–10th cycle to capture the periodic solution. Reach for LES or k-omega SST only when Re_peak exceeds about 4000; otherwise laminar suffices. In ANSYS Fluent or OpenFOAM (transientSimpleFoam / pimpleFoam) the standard inlet boundary is the Womersley analytical velocity profile imposed as a time-dependent map.

Real-world applications

Cardiovascular CFD: Aortic aneurysm rupture risk, carotid stenosis and ventricular flow studies all live in the $\alpha > 10$ inertia-dominated regime. Steady Poiseuille analysis is inadequate; unsteady Navier-Stokes pulsatile CFD is mandatory. Codes such as ANSYS Fluent, SimVascular, CFX or ABAQUS use patient-specific MRI/CT geometries, with cardiac flow waveforms as inlet conditions, integrated over 5–10 cycles. The aortic default of this tool ($\alpha = 13.79$, Re_peak = 6057) is exactly that target.

Mechanical circulatory support devices: VAD (ventricular assist devices) and ECMO blood pumps either produce or accept pulsatile flow, so Womersley similarity is at the heart of their design. Data from animal trials at $\alpha = 5$–$15$ are extrapolated to human patients by matching $\alpha$ and peak Re simultaneously. Designers evaluate hemolysis risk (wall shear above 150 Pa breaks red cells) and thrombosis risk (residence time above 1 s triggers coagulation). Manufacturers such as Nipro, Terumo and Abbott use this workflow as their standard process.

Pulse-wave velocity (PWV) and arterial stiffness: PWV is a clinical biomarker for early atherosclerosis closely tied to Womersley analysis. Healthy adults have PWV in the 5–8 m/s range, and progression to 12–15 m/s indicates arterial stiffening. The increase reflects wall stiffness $E$, recovered from Moens-Korteweg $c = \sqrt{Eh/(2\rho R)}$. In Japan PWV measurement is performed routinely in health checks as a "vascular age" index, a direct application of Womersley fluid mechanics to public health.

Engine intake / exhaust manifold design: Four-stroke engines have intake $\alpha = 50$–$300$ — extreme inertia — and exploit wave dynamics through inertia supercharging. End-of-intake suction pulses reflect into pressure pulses that arrive in time for the next intake stroke. F1 and MotoGP teams tune intake runners to a quarter-wave resonance, gaining 5–10% volumetric efficiency at Womersley numbers around 100. 1D acoustic / wave-action codes such as GT-POWER, AVL Boost or Wave are the standard tools.

Common misconceptions

The most common mistake is assuming that "blood flow equals Poiseuille flow." That is only valid in vessels smaller than about 0.5 mm radius where $\alpha < 1$; in the aorta and main coronaries it fails badly. Setting $R = 0.5$ mm in this tool gives $\alpha = 0.69$ (quasi-steady), where the Poiseuille law $\Delta P = 8\mu L Q/(\pi R^4)$ applies. At $R = 10$ mm the Poiseuille resistance can overestimate the actual pressure gradient by more than 50% because phase lag and the central piston take over. Clinical hemodynamic models must switch formulas by vessel size.

Another frequent error is the belief that "large $\alpha$ means turbulence." $\alpha$ is the inertia-to-viscosity ratio for oscillation, not for the mean flow. The aorta has $\alpha = 13.8$ but is mostly laminar; only the systolic peak can be partially turbulent when Re_peak exceeds about 4000. Using $\alpha$ alone to declare turbulence is a category mistake — always evaluate $\alpha$ and Re_peak together. The default values here give Re_peak = 6057, right on the boundary where short turbulent bursts can appear during systole.

The third pitfall is to assume that "the Womersley solution works for any vessel." The classical Womersley analytical solution assumes a straight, rigid, axisymmetric pipe with Newtonian fluid and a single oscillation frequency. Real arteries have branching, curvature, elastic walls and non-Newtonian rheology (low-shear viscosity rise due to red-cell aggregation). For instance Dean vortices appear in bends and bifurcations, which the linear theory cannot capture. The Womersley number remains an essential teaching tool but patient-specific clinical CFD requires geometry-resolved meshes, unsteady Navier-Stokes and rheology models such as Carreau-Yasuda.