Lesson 4 of 11

Aerodynamics

How the combustor sculpts its own flow: swirl and vortex breakdown create the recirculation zone that holds the flame; jets in crossflow feed and quench it; and the pattern factor grades the result.

7 min3 live simulations
you are here
The question

How can spinning air make it flow backwards?

The answer is the single most important aerodynamic trick in a gas turbine combustor, and it is the reason the flame in Lesson 1 can survive.

Spin, and the low pressure it leaves behind

Picture air leaving the swirler: a ring of angled vanes around the fuel nozzle. Each parcel now moves forward and around. Rotating fluid needs a force pointing inward to keep it on its circular path (the same reason you feel pushed outward on a merry-go-round), and in a fluid that force can only come from pressure. So the pressure must be higher at the outside and lower at the centre:

∂p∂r=ρ W2r\frac{\partial p}{\partial r} = \frac{\rho\, \cair{W}^2}{r}

where WW is the tangential (swirl) velocity. Strong swirl carves out a low-pressure core on the axis.

Now follow that core downstream. As the swirling jet spreads out into the wider liner, it slows its spin (angular momentum is conserved while the radius grows), so the pressure deficit on the axis weakens. That means the pressure on the axis rises along the flow direction. Downstream is at higher pressure than upstream, and the fluid on the axis, which has little forward momentum to begin with, gets pushed back upstream.

That is Vortex breakdownThe abrupt reversal of flow on the axis of a strongly swirling jet as its swirl decays and axial pressure rises., and the bubble of reversed flow it creates is the Central recirculation zoneA bubble of reversed flow on the combustor axis, created by vortex breakdown, that returns hot products to ignite fresh mixture..

ExploreSpin the air until it turns around
central recirculation zone
45°
Sₙ = ⅔ · (1 − h³)/(1 − h²) · tan θ, hub ratio h = 0.5
Swirl number0.78strong
Axis velocity (min)-1.24 U₀reversed flow
0246810-101distance downstream (nozzle diameters)u / U₀reverse flow
Vortex breakdown! A robust recirculation zone sits just downstream of the nozzle, continuously feeding hot products back to ignite fresh fuel. This is where production swirlers live.
Challenges0/2
  • Find the weakest swirl that still reverses the flow on the axis.
  • Swirl so hard the recirculation bubble reaches the fuel nozzle.
Half the combustor in side view. Particles are coloured by temperature; gold is fresh fuel. Increase the vane angle and watch the flow on the axis stop, then reverse.

The CRZ is the combustor's pilot light. Hot, radical-rich products circulate back toward the nozzle and meet the fresh fuel–air mixture along a shear layer, igniting it continuously. Everywhere along that boundary the flame is being relit, thousands of times a second.

Measuring swirl

The strength of the swirl is captured by the Swirl number SₙRatio of the axial flux of angular momentum to that of axial momentum. Above ~0.6 the flow on the axis reverses.: the ratio of the axial flux of angular momentum to the axial flux of axial momentum (times a radius):

Sn=GϕGxR=∫0Rρ UW r2 drR∫0Rρ U2 r drS_n = \frac{G_\phi}{G_x R} = \frac{\int_0^R \rho\, U W\, r^2\, dr}{R\int_0^R \rho\, U^2\, r\, dr}

For a flat-vaned annular swirler with hub-to-tip radius ratio hh and vane angle θ\theta, this integrates to a handy result:

Sn=23 1−h31−h2 tan⁡θS_n = \frac{2}{3}\,\frac{1-h^3}{1-h^2}\,\tan\theta

Going deeper
Deriving the vane-swirler swirl number

Assume flat vanes turn the flow uniformly, so that across the annulus from Rh=hRR_h = hR to RR the axial velocity UU is constant and the swirl velocity is W=Utan⁡θW = U\tan\theta. Then

Gϕ=2πρ∫RhRU (Utan⁡θ) r2 dr=2π3ρU2tan⁡θ (R3−Rh3)G_\phi = 2\pi\rho \int_{R_h}^{R} U\,(U\tan\theta)\, r^2\,dr = \tfrac{2\pi}{3}\rho U^2 \tan\theta\,(R^3 - R_h^3)

Gx=2πρ∫RhRU2 r dr=πρU2(R2−Rh2)G_x = 2\pi\rho \int_{R_h}^{R} U^2\, r\,dr = \pi\rho U^2 (R^2 - R_h^2)

Divide, and normalise by RR:

Sn=GϕGxR=23 1−h31−h2 tan⁡θS_n = \frac{G_\phi}{G_x R} = \frac{2}{3}\,\frac{1 - h^3}{1 - h^2}\,\tan\theta

With h=0.5h = 0.5, a 45° swirler gives Sn≈0.78S_n \approx 0.78, comfortably into the recirculating regime.

And in the widget you can see the thresholds that matter:

  • Sn≲0.4S_n \lesssim 0.4: weak swirl. The flow spins but never reverses, and there is nothing to hold the flame.
  • Sn≈0.6S_n \approx 0.6 and above: strong swirl. Vortex breakdown creates a robust CRZ; this is where production swirlers live, typically Sn≈0.6S_n \approx 0.6–1.01.0.
  • Sn≫1S_n \gg 1: the bubble grows and creeps upstream onto the nozzle, where hot products can overheat and coke the injector tip, and the pressure loss climbs.

Jets in crossflow: the liner's surgical tools

The swirler handles the dome, but much of the air enters through rows of holes in the liner, as jets fired sideways into the hot flow. Each row has a job:

  • Primary jets feed and shape the primary zone. Opposed jets from the outer and inner walls collide on the axis and help close the recirculation from downstream.
  • Intermediate jets supply the oxygen to burn out CO, H₂ and soot, while keeping the gas hot enough for the burnout to finish.
  • Dilution jets, the largest holes, must drive cold air all the way into the hot core and quench it to the turbine's temperature.

A jet fired into a crossflow bends over like smoke from a chimney in a wind. How far it penetrates depends overwhelmingly on the Momentum-flux ratio JρⱼVⱼ² / ρₘVₘ² — the jet-to-crossflow momentum ratio that sets how deep a liner jet penetrates.:

J=ρjVj2ρmVm2,yd≈0.89 J0.47(xd)0.36J = \frac{\rho_j V_j^2}{\rho_m V_m^2}, \qquad \frac{y}{d} \approx 0.89\, J^{0.47}\left(\frac{x}{d}\right)^{0.36}

ExploreAiming the air jets
hot gas from the primary zone →combustor centrelinejets meet on the axis
20.0
J = ρⱼVⱼ² / ρₘVₘ² — raised by a bigger liner pressure drop
0.100
Penetration96% of height1.5 liner heights downstream
VerdictReaches core
Good penetration: the cold jets reach the combustor core, where the hottest gas lives. Now they can mix it out.
Challenges0/2
  • Make opposed jets meet on the axis while keeping J under 8 (a cheap liner pressure drop).
  • Fire a single-wall jet so hard it overshoots and hits the far wall.
Air jets fired from the liner wall into the hot crossflow bend over as they go. How deep they reach is set almost entirely by the jet-to-mainstream momentum-flux ratio J.
Estimate first
The liner pressure drop is 3% of a 30-bar, 800 K compressor delivery. Roughly how fast does air shoot through the liner holes?
1101001,000
Drag along the line to place your guess — each tick is ten times the last.

Notice what controls JJ. The jet velocity comes from the pressure drop across the liner, so penetration is bought with pressure loss. The jet air is also cold and dense relative to the 2000 K mainstream, which helps a lot: ρj/ρm\rho_j/\rho_m can be 2–3.

Check your intuition
Dilution jets are penetrating too little and leaving a hot core. Which change most directly fixes it?

Pressure loss: the price of mixing

Every vane and every hole takes a toll. The overall combustor pressure loss:

ΔPPt=Pt,in−Pt,outPt,in\frac{\Delta P}{P_t} = \frac{P_{t,\text{in}} - P_{t,\text{out}}}{P_{t,\text{in}}}

is typically held to 3–6%. Within it, the liner pressure drop is what drives the jets and the swirler; the rest is lost in the diffuser and annuli. More liner drop means faster jets, better penetration, more turbulence and shorter, more stable flames. It also means a less efficient engine. That window of 3–6% is where the mixing benefit and the cycle penalty balance for most engines.

The report card: pattern factor

Everything in this lesson (swirl, jets, pressure drop, even the diffuser's uniformity) is graded in one place: the temperature distribution at the combustor exit, as the first turbine vanes see it.

PF=Tmax−TˉexitTˉexit−Tinlet\text{PF} = \frac{T_\text{max} - \bar T_\text{exit}}{\bar T_\text{exit} - T_\text{inlet}}

The Pattern factor(T_max − T_mean)/(T_mean − T_inlet) at the combustor exit: how much hotter the worst hot streak is than average. compares the single hottest spot with the average temperature rise. Turbine cooling must be designed for the hottest spot, so a high PF is paid for directly, in turbine life or turbine cooling air.

ExploreThe report card: exit temperature traverse
PF = 0.23peak 1965 K
typical
Longer dilution zone / higher liner ΔP → better mixing
18
Pattern factor0.23(T_max − T_mean)/(T_mean − T₃)
Hottest spot+215 Kabove the 1750 K mean
Acceptable but costly — the turbine must be cooled for the hottest streak, not the average.
Challenges0/2
  • Bring the pattern factor below 0.15.
  • Show how one clogged injector can push a well-mixed combustor above PF = 0.3.
The ring is the combustor exit, as the first turbine vanes see it. Each burner leaves a footprint; dilution mixing smooths the footprints out. A partly clogged injector unbalances its neighbours and creates a hot streak.

Typical modern values are around 0.15–0.25. Notice how a single partly blocked injector creates a hot streak far out of proportion to its size: that is why fuel-injector uniformity and deposits (Lesson 6) matter so much.

Recap
What to carry forward
  • Swirl creates low pressure on the axis; as the swirl decays downstream, the axial pressure rises and drives flow backwards. This is vortex breakdown and the CRZ.
  • The swirl number SnS_n grades swirl strength: < 0.4 weak, ~0.6–1.0 strong and useful, ≫ 1 risks overheating the nozzle.
  • Liner jets penetrate according to the momentum-flux ratio JJ, which is paid for with liner pressure drop.
  • Total pressure loss is a budget of 3–6%, spent to buy mixing.
  • The pattern factor measures exit non-uniformity; typical values are 0.15–0.25, and a single bad injector can dominate it.

We now have a flame and a flow to hold it. Next question: how do we know the combustor is working, and at what point does it stop?

Finished this lesson?Mark it complete to track your progress through the course (saved in this browser).
Up next · Lesson 5Combustion PerformanceWhere exactly does a combustor stop working?Keep going →