Ten lessons, one machine. Can you design a combustor that passes every test at once?
No multiple choice. Just a mission, a set of limits, and every trade-off you have learned, all fighting each other at the same time.
How a combustor engineer thinks
Real combustor design rarely starts with the flame. It starts with the mission: the pressures and temperatures the engine cycle will impose, and the limits the combustor must live within. Then the designer works through a surprisingly consistent chain of reasoning, and you now know every link of it.
- What does the turbine need? The exit temperature fixes the overall fuel–air ratio. Everything else is about where the fuel burns and how the air is shared (Lessons 1 and 2).
- Where must the flame sit on the bell curve? Near stoichiometric for stability and relight; far from it for NOₓ. The flame strategy is the answer to this one question (Lessons 2 and 9).
- How is the air budget divided? Primary, dilution and cooling all claim the same fixed air flow, and every claim is a trade (Lessons 1, 4 and 8).
- Is it stable everywhere, including idle and altitude? The Damköhler number and the relight kernel have veto power (Lesson 5).
- Does the fuel arrive as a fine mist? Otherwise low-power efficiency and smoke suffer (Lesson 6).
- Will it sing, melt, or flash back? Acoustics, liner life and premixer safety are pass/fail, not nice-to-haves (Lessons 7, 8 and 10).
The studio
Choose a mission, then shape your combustor. The drawing redraws as you design: the swirler vanes tilt, dilution jets thicken with their share of air, the flame changes colour with its temperature, cooling films turn into effusion holes, and the exit temperature profile on the right flattens or spikes with your pattern factor. Each failing test tells you why, and which lesson holds the fix.
Design reviews
Stuck? These reviews explain one route to certification for each mission. Try on your own first: the struggle is where the understanding forms.
Going deeper
Airliner: why the obvious design fails
Start from the default conventional combustor and look at the report. NOₓ fails because the primary zone sits at φ ≈ 1 at 880 K inlet, right at the peak of the bell. Idle efficiency fails because a simplex nozzle atomises poorly at low flow. The liner fails because slot films can't protect a wall next to 2600 K gas.
One route: switch to RQL. Give the primary zone less air, so it runs genuinely rich (φ around 2), and choose an airblast injector for fine sprays and less smoke. Then you'll meet RQL's notorious weakness: a rich, sooty zone radiates hard, and the compressor "coolant" is already at 880 K. Effusion + TBC with a generous cooling share gets close; a SAF blend, with its low-soot flame, often tips it over the line. That's Lesson 8 and Lesson 10 shaking hands.
The other route is lean staging: send roughly 70% of the air through the premixer so the main stage is truly lean, keep a modest pilot (enough for relight at 9 km), add resonators for the inevitable dynamics, and watch the liner, which now has little air left for cooling. No conventional design can pass: the stoichiometric primary zone makes too much NOₓ however you tune it.
Going deeper
Fighter: short, hot, and relightable
The fighter's enemies are size and relight at 12 km. A long, lightly loaded combustor fails the length test, and a lean-staged design needs too much length to stay efficient and stable. A conventional design comes tantalisingly close but usually falls just short on relight. RQL is the natural fit: its rich pocket is easy to ignite, strong swirl gives the kernel a recirculation zone to live in, and a short, compact liner keeps the weight down. Watch smoke: a rich primary needs a good spray and strong mixing to burn its soot off, and a SAF blend helps. Notice that the fighter's generous NOₓ limit barely matters: its hard problems are elsewhere.
Going deeper
Power plant: the lean premixed obstacle course
Only lean staging can get under 9 g/kg. Send most of the air through the premixer (70% or so) so the main stage is genuinely lean (φ ≈ 0.55–0.65), and use very little pilot fuel. That creates three new problems, each from a different lesson:
- Combustion dynamics (Lesson 7): lean premixed flames sing. Add Helmholtz resonators.
- Flashback (Lesson 10): keep the main stage lean and the premixer fast.
- Liner cooling (Lesson 8): with 70% of the air in the premixer, there is little left for cooling. But the flame is also far cooler, so modest effusion cooling can suffice.
Size is not a concern for a power plant, so use the length to buy residence time for efficiency at part load.
The whole picture
Follow one last parcel of air through the machine, and notice that you now know why each thing happens.
It leaves the compressor at 150 m/s and is slowed by the diffuser (Lesson 3), which trades a little pressure for a flow the flame can live in. The swirler spins it into a recirculation zone (Lesson 4) that anchors a flame governed by the bell curve and the Arrhenius tail (Lesson 2), inside a stability loop set by the Damköhler number (Lesson 5). The flame is fed by a spray whose droplet sizes decide how completely it burns (Lesson 6), it may couple with the chamber's acoustic modes (Lesson 7), and it is fenced in by a liner that survives only because films, coatings and a clean flame keep the heat out (Lesson 8). What leaves carries an emissions signature written by the path each parcel took through φ-space (Lesson 9). And the fuel itself is changing beneath all of it (Lesson 10).
Every one of these is in tension with the others. That tension is not a flaw in the technology. It is what engineering near physical limits looks like, and you have just done it yourself.
- Temperature is the master variable: through Arrhenius and Zeldovich it controls stability, efficiency, NOₓ and durability at once.
- Time is the currency: every performance limit compares time available with time needed (Da, d², quench, residence).
- Air is the budget: every zone, jet and film is a claim on the same fixed air flow.
- Local beats average: NOₓ, soot and hot streaks are set by individual parcels, not means.
- Everything couples: a change to fuel, spray, swirl or cooling ripples through stability, acoustics, heat transfer and emissions.
If you can now look at a combustor design decision and see the physics behind it, and the price it pays somewhere else, the course has done its job.