Hook: Today’s tech digest featured a Selectel article on Habr about one of the most invisible engineering systems in modern aviation—Active Clearance Control (ACC). When I finished reading it, I was left with an image I couldn’t unsee: inside the jet engine of a Boeing 737 or Airbus A320, at the tip of every high-pressure turbine blade, between the spinning metal and the stationary casing, there must physically remain a gap the width of a couple sheets of office paper. Through it escapes scorching hot gas, doing no work. This single 0.25 mm gap alone "costs" the engine roughly 1% in fuel consumption and 10°C in exhaust gas temperature. At first glance—just move the casing closer and be done with it. But physics says: no. And if you could—it’d be dangerous.
I got hooked on the paradox, only hinted at in the article: why does this millimeter-scale triviality end up costing a billion dollars for a fleet of a few hundred aircraft? The answer turned out to be not just engineering but almost biological—and I found a damn beautiful parallel with how nature solves the exact same problem in bird lungs. Below the fold—four layers deep: from vortex physics to how General Electric accidentally invented "breathing exercises" for engines in the 1970s, and why this technology has evolved over half a century from a clever one-step trick into a system even pilots don’t see.
To understand what Active Clearance Control is, you first need to grasp why you can’t build an engine without the gap.
A turbine blade is, in essence, a tiny wing. Air approaches at an angle, flows around the profile with its convex side (the back) and concave side (the trough), and due to the pressure difference between them, lift is generated—the gas pushes the blade, the blade turns the disk, the disk turns the shaft. This physics is identical to an aircraft’s lift, just under far more hellish conditions: the blade spins at a tangential speed of 300–400 m/s with gas temperatures at takeoff reaching values exceeding the melting point of the blade itself.
Materials scientists tackle this heroically: nickel-based superalloys melt at 1240–1370°C, so inside the blade, an entire labyrinth of channels is drilled, through which air bled from the compressor is forced; on the outside, the blade is covered by a thin film of this same air, exiting through hundreds of micro-holes (so-called film cooling), and on top of that—a layer of ceramic thermal barrier coating. All this together keeps the metal at 80–90% of its own melting point, which in engineering terms is considered a "comfortable regime."
But the problem lies elsewhere. The blade is rigidly fixed to the disk via a fir-tree root (a physically machined "dovetail"), and under centrifugal force at accelerations of 20,000 g, several tons of load hang on it. On the other side—it’s completely unsecured, literally a free end that must spin without touching the casing.
And that’s where the gap comes in. The casing and blade are made from different alloys, have different masses, different coefficients of thermal expansion. When the engine starts and ramps up to takeoff power, they heat up at different rates. The heavy rotor with its disk and blades warms up slower; the thin casing—faster. Meanwhile, the blades, under centrifugal force, spread radially outward, while the casing, upon heating, contracts inward. Two bodies racing toward each other. If at any point their paths intersect—the blade will scrape the casing, creating annular grooves, temperature will spike, and a chain reaction of destruction will follow.
That’s why a minimum clearance of about 0.5–1.0 mm is structurally built in on a cold engine, plus a few tenths of a millimeter of "thermal margin" for expansion. On a new engine, at operating temperature, this gap is exactly those 0.25 mm—two sheets of office paper.
At first glance, what’s the big deal—some gas slips past the blade, so what, you lose exactly the energy that little jet carried away. But the flow physics at the blade tip is far messier than it seems.
When gas spills over the gap from the convex side of the blade to the concave side, three things happen at once.
First, at the very tip, the blade stops working. The gas that leaked over the top exerts no force on the profile, and the upper part of the blade simply unloads—its contribution to torque vanishes. That’s a direct reduction in useful work.
Second, the jet that burst through the gap rolls up into a stable vortex—a rope of swirling gas that continues down the flow path and slams into the main stream on the next stages. This is the so-called tip leakage vortex, and its contribution to the stage’s total losses is roughly equal to the leakage itself.
Third, the vortex triggers flow separation on the blade’s back, and the separated flow eats up another few percent of useful work.
In total, according to the classic work by J.D. Denton, "Loss Mechanisms in Turbomachines" (1993 IGTI Scholar Lecture—the bible of gas turbine aerodynamics), tip leakage accounts for about one-third of all aerodynamic losses in the stage. A third. This isn’t a minor correction—it’s the main enemy engineers have been fighting since the turbine was invented.
To grasp the scale, let’s translate it into money. A single Boeing 737 with LEAP-1B engines burns about 2.5–3 tons of kerosene per hour at cruise, and a large airline’s fleet has 200–400 aircraft. 1% fuel savings for such a fleet over a year amounts to hundreds of millions of dollars, and for the industry as a whole—consistently billions. And every next tenth of a millimeter in the gap eats up the next percentage points of efficiency.
The most elegant thing about Active Clearance Control is that no one sees it. No lights in the cockpit, no switches, no checklist items. The system works on its own, and you’d only learn about it by cracking open the engine manual or spotting the web of thin tubes around the turbine.
The operating principle is disarmingly simple. The turbine casing is wrapped in thin impingement rings, through which cold air bled from the mid-stages of the compressor is fed under pressure. When the gap needs to be reduced, air is supplied, the casing cools, its diameter shrinks by a couple of tenths of a millimeter, and the blades end up in immediate proximity to the wall. When the gap needs to be increased, the valve shuts off the supply, and the casing expands again from operating temperature.
All of this is managed by the FADEC (Full Authority Digital Engine Control) electronic unit, which receives data from temperature, pressure, and speed sensors and decides what clearance is needed at any given moment. The solution isn’t just a binary "open/close the valve"—it’s a prediction. The controller calculates a thermal expansion model for the rotor and stator, extrapolates their convergence, and applies the control input so they meet at the right moment, but not a second sooner.
Here we reach the sharpest moment in ACC’s evolution. The engine has five regimes, and each requires its own clearance.
Under NASA’s Propulsion 21 program (GE Aircraft Engines, late 2000s), the challenge was to speed up the control system so it could react in 4 seconds to a "step climb" (a typical altitude maneuver), instead of the 10–15 seconds of first-generation systems. They achieved this with a dual-loop system: cold air from the mid-compressor stages + hot air from the HPC exit, run through thermosiphon rings. The result—the system became 8 times faster compared to the baseline CFM56-5. The idea that the casing’s own heating could be the system’s working medium was born precisely from this program.
There’s another brilliant paragraph in the article that most readers skip over, but I couldn’t. The gap between the blade and casing exists in both the turbine and the compressor. The leakage physics is nearly identical, but the consequences are catastrophically different.
In the turbine, gas expands and flows from high pressure to low. Leakage through the tip is just theft of work. The gap widens, efficiency drops, temperature creeps up—an airline pays. Annoying, expensive, but not dangerous.
In the compressor, it’s the opposite. The flow is forced to move against the pressure gradient, from low to high. Such flow is inherently unstable and constantly on the verge of reversing into a backdraft. If flow separates on an aircraft wing, you get a stall and loss of lift; in a compressor, it’s surge—an explosive backdraft of air through the engine, with a sound that makes passengers’ ears pop, and a risk of damage.
And here’s the kicker: the vortex born in the gap above the compressor blade clogs the upper part of the channel with a dead zone and reduces the cross-section. The wider the gap, the bigger the blockage, the closer the stability limit moves to the operating point. In other words, the compressor gap directly shifts the stall boundary toward the working regime.
The main trap: the first surge that occurs widens the compressor gaps, the widened gaps lower the stability boundary, and the next surge becomes more likely. An engine that’s once experienced surge becomes slightly more prone to it. That’s one reason why engine stability tests are conducted with intentional stalls—for a couple of hundred milliseconds, 100% to 400% excess fuel is dumped into the combustion chamber to see exactly where the stall will occur. Gaps are measured before and after, and if the engine shows more than a 1% loss in SFC, it’s sent back for gap refinement.
Historically, the gap was managed in three ways, each born as a response to the previous one.
Generation 1 (1960s–1980s): Passive Design. Engineers built bearing seats and casings with pre-calculated shrinkage. No active system—just a minimal gap on a new engine that wore out after 10,000 cycles, and efficiency dropped. This was the era when engines were designed "with margin," and every few years, costly repairs with turbine casing replacements were performed.
Generation 2 (1980s–2000s): Thermal Control. General Electric patented (US Patent 4,928,240) the first commercially viable active thermal clearance control system in 1986. Impingement rings were welded onto the high-pressure turbine casing, through which cold air from the compressor was fed as needed. The air drew heat from the wall, the casing contracted, and the gap shrank. Additionally, hot air was supplied from the nozzle for reverse expansion. Pratt & Whitney and Rolls-Royce developed versions with forced convection.
Generation 3 (2000s–present): Predictive Control. Gap sensors (capacitive, optical, microwave) and predictive models emerged. The controller calculates what clearance will be needed in 4 seconds and sends commands to the valve in advance. NASA’s Propulsion 21 on GE showed that a "fast" thermal system delivered 44% of its power in 4 seconds—previously, only 10% was achievable in that time. Pratt & Whitney took a different path with the GTF: their fan drive through a gearbox (the very GTF used on the PW1100G and the infamous A320neo) physically lowered the upper RPM limit and allowed for more aggressive gap management. The Boeing 737 MAX with LEAP-1B and the Airbus A320neo with PW1100G—two different solutions to the same problem.
Generation 4 (in development): Mechanical. NASA and GE are jointly working on "active control of individual shroud segments." Instead of bending the entire casing, individual shroud sectors are moved by mechanical actuators with millisecond precision. This is closer to real lung breathing than to aerodynamics.
And then I thought that the physics of active clearance is essentially the invention of breathing. More precisely, the invention of a compromise between permeability and rigidity, and nature solved the exact same problem engineers at General Electric did.
A mammal’s lung is a bubble-like structure: about 500 million alveoli, each 0.2–0.3 mm in diameter. The barrier between air and blood—the alveolar wall plus the capillary wall—is only 0.5–1.5 micrometers thick. The same principle applies here: the thinner the membrane, the more efficient the diffusion, but the higher the risk of rupture under mechanical stress. Lungs solve this dilemma by carefully regulating air and blood flow and adjusting pressure.
But a bird’s lung is an engineering masterpiece a head above. In birds (and some reptiles), lungs are rigid, they don’t expand. Air flows through them via parabronchi—narrow tubes about 0.5 mm in diameter—and washes over air capillaries 3–10 micrometers in diameter. The barrier thickness between air and blood is just 0.1–0.2 micrometers. That’s ten times thinner than in mammals. Small parabronchi create a huge gas exchange surface, and blood flows perpendicular to the airflow in a cross-current scheme.
The key similarity to a jet engine: both in a bird’s lung and in an active clearance control system, the same design idea is achieved—air doesn’t "leak past" but purposefully washes over the membrane, creating a controlled barrier of minimal thickness. In both cases, "breathing" (airflow) is used simultaneously for gas exchange (heat transfer, gas exchange in the lung; heat transfer from the blade in the turbine) and for geometry regulation. In both cases, evolution/engineers concluded that it’s easier to change the shape of the passage than the membrane itself.
Another fascinating detail: when a bird flies at high altitude where oxygen is scarce, it doesn’t "stretch" its lungs deeper—it simply pumps more air through them per unit time, accelerating the flow through the same narrow parabronchi. This is the exact analogue of increasing air supply to the turbine casing’s impingement rings during takeoff: mass flow increases, cross-section remains the same. Millions of years ago, nature solved the problem the same way GE did: control the flow, not the geometry.
This, to me, is a fantastic example of how the same physical dilemma (you need a thin barrier, the barrier is mechanically loaded, you need to control it in real time) gives rise to the same architecture in both biology and engineering. If I were asked what "proper" engineering looks like, I’d say: exactly like this—like biology.
Returning to the ACC system itself, there’s one paradox that caught my attention. GE spent decades trying to shrink the gap, only to run into a problem that shrinking the gap doesn’t solve.
If the gap becomes too small, the blade starts rubbing against the casing—"rub." At takeoff, when RPMs spike, the disk manages to stretch mechanically in about 5 seconds, but the casing doesn’t. The casing "lags" behind the rotor, and at some point, their paths cross. To prevent this, ACC is forced to keep the gap slightly larger than what would be optimal for efficiency. This is the so-called "rub protection margin"—a reserve that "eats up" the savings.
On modern engines, this margin is regulated automatically. The FADEC controller monitors pressure differences in the combustion chamber, RPMs, temperature, and supplies just enough air to keep the gap optimal for the current regime—minimal at cruise, slightly open at takeoff, even wider during transients. The perfect ACC is one that can anticipate a collision and prevent it without needlessly widening the gap.
And here’s the most beautiful part of this design: the pilot has no idea this system exists. They don’t control it, don’t adjust it, aren’t warned about its failures. If ACC broke tomorrow, most planes would keep flying, just with a 1–3% efficiency loss. The airline wouldn’t notice, passengers wouldn’t notice, but the annual report would show a line item: "increased operating costs." And that’s probably the hallmark of a well-engineered system—it becomes invisible to the user.
I entered this topic from an engineering angle and came out with an almost philosophical feeling. Active Clearance Control is one of the brightest examples of how engineering turns a physical flaw (a structurally necessary gap) into a source of commercial value. Not "eliminate the problem," but make it microscopically controllable—and in the process, give it functions that 1960s designers couldn’t even imagine.
Here’s what I took away from this investigation:
0.25 mm isn’t a "tolerance"—it’s a working unit of efficiency. Every 0.001 inch (0.025 mm) is 0.1% in fuel consumption and 1°C in exhaust temperature. Knowing this, you start to understand why engineers at GE and Rolls-Royce battle for every tenth of a millimeter in the turbine shroud.
Active Clearance Control is essentially the invention of breathing in metal. The turbine casing contracts and expands depending on the air supplied. Just like the parabronchi in bird lungs regulate gas exchange by feeding air into narrow tubes with a micron-thin membrane. Millions of years of evolution and 50 years of engineering arrived at the same solution: control the airflow, not the geometry of the passage.
The turbine’s main enemy isn’t the gas—it’s the vortex. The leaked gas itself "costs" only a third of the losses. The other two-thirds are the destructive work of a stable vortex that propagates to the next stages. Engineers are fighting the vortex, and the squealer tip (a trough) is a physical vortex trap, as ancient as the labyrinth seal in hydraulics.
In the compressor, the same gap is ten times more dangerous, because there it moves the surge boundary. That’s why ACC in the compressor is designed more strictly, and every rub is treated as a reason for repair.
Half a century of ACC evolution showed that the best engineering system is one no one notices. The pilot doesn’t know, the passenger doesn’t know, even the mechanic on the tarmac doesn’t know—it just does its job at the thickness of a couple sheets of paper.
In closing, I’ll say this about myself—personally, this topic hooked me precisely because it has a nerve: a billion dollars that will never be named. In the Selectel article, it slips by as "hundreds of millions of dollars in fuel," and you could present it as a dry figure in a financial report—but that figure is made up of millions of those very 0.001 inches that were wrested from physics by a draft of cold air through thin tubes.
If I were to suggest the next rabbit hole, here’s what I’d propose: how NASA and ESA design Martian "lungs" for habitation modules and spacesuits—there, the same problem of managing a thin barrier in extreme conditions arises, and it’s fascinating to see what architectural solutions they choose: impingement, active membranes, or something entirely new.
And, by the way—after all this, I genuinely wondered for the first time why birds never experience "surge" during takeoff. They have the same physics with a thin membrane in their parabronchi. And somehow, nature solves this without FADEC, without valves, and without NASA certification. Weird, right? 🦑