Buzz-saw noise from big turbofans at high thrust

Question

This topic has been addressed many times on this and other forums. I have seen (and often not fully understood) some deep technical analyses, incuding manufacturers' research labs, since this is a significant component of take-off noise. (-And all that extra mechanical energy might have adverse effects on blade and liner life.)

Answers generally seem to include the notion of fan tips going supersonic, sometimes with complications vie reflections from stator structures. But the origin of the specific low-audio frequency is not addressed anywhere I've seen. How does a many-bladed fan spinning at a few thousand RPM generate a (harmonic-rich) fundamental around 50 Hz?

In the past few years, I've submitted a possible theory to a couple of other forums, but never had it accepted by moderators. I'm hoping that this group might be helpful.

Thinking about potential acoustic resonances in a turbofan engine, one presumably wants to find the structural element of greatest size in at least one dimension. The core might be considered to be as long as the whole engine, but there are many interruptions in the flow path. The bypass casing, however, might make a nice organ pipe at about the right frequency, excited either asynchronously or as a deep submultiple of the rotating blade-shock frequency.

One phenomenon that this might explain is the slight drop in bass note when power is reduced after takeoff. Fan and core shafts presumably slow down very quickly. But if the main sound is from a pipe resonance, that shouldn't be affected. However, reducing RPM also reduces compression in the bypass channel. This presumably reduces temperature, hence speed of sound, making the pipe acoustically longer and lowering resonant frequency.

Is this silly, or perhaps obvious? The idea may appear elsewhere, but I haven't found it.

Thanks for your thoughts.

Answer 1

What I'm missing is why the buzzing sound from such a symmetrical object as a tightly bladed fan should be one-per-revolution.

It's not. "Buzz saw noise" as I would define it, has a fundamental frequency at the fan blade passing frequency. i.e. if there are 20 fan blades, then "buzz saw noise" has a fundamental at 20/rev. As you mentioned, this noise is mostly due to the shock wave generated as the tips of the fan blades go transsonic.

Now there is noise generated by an engine at 1/rev, but it has nothing to do with the fan blades. This noise is generated by the unbalance of the rotor. Because the rotor is never perfectly balanced, there is a force between the rotor and the bearings at 1/rev. This force gets transmitted structurally up from the bearings, through the frames and cases, through the mounts, through the wing, and eventually into the fuselage. Throughout this whole path, it's not noise, its a structural vibration. It can become acoustic noise when the fuselage motion starts to cause the air inside the cabin to start vibrating. But you can also just feel it. In programs I worked on, the airframers would put accelerometers on the seat tracks to determine how much 1/rev vibration the passengers would feel in their seats.

Answer 2

How does a many-bladed fan spinning at a few thousand RPM generate a (harmonic-rich) fundamental around 50 Hz?

According to its certification documentation, the Pratt&Whitney PW1100G has a low-pressure rotating speed $N_1$ of 10'047rpm. This low-pressure stage is connected to the fan via a 1:3.0625 gearbox, therefore giving a fan speed of 10'047/3.0625=3'280rpm. This translate into some 3'280/60=55rps (revolutions per second) aka Hz. Now, due to the impossibility of building a perfectly balanced rotating machine (even if the blades might have a difference of only a couple of milligrammes), the rotor vibrates at this particular frequency therefore generating the relevant noise which gets to the cabin through the air; plus, since the turbofan is connected to the airframe, there's also a structural path letting these vibrations travel till the cabin.

So, even without taking into account any aerodynamis phenomenon, only from a pure mechanical point of view the fundamental "noise" frequency of a modern jetliner's turbofan is just in the range that you have expected.

Obviously for the low-pressure stage the fundamental frequency is 3.0625 times as much i.e. some 170Hz. And for the high-pressure stage rotating at 22'300rpm, the fundamental frequency is 22'300/60=370Hz.

At all these fundamental frequencies we have to add their whole multiples and especially the frequencies given by the fundamentals times the number of blades: for the fan for example the frequency 55x19=1kHz should also be important.

And then there are all the noises having an aerodynamics source but, as you pointed out, that's should have been already addressed by other answers on this site.

As a side note it's nice to know that the whole airplane's structure and all the components installed on the aircraft should be designed and tested so that they have no eigenfrequency close to those fundamental frequencies in order to avoid possibly catastrophic resonance phenomena.

Answer 3

Above frequency spectrum plot, earlier used in this answer, was cc-d from a now defunct link and shows fundamentals of 75 Hz, way below the blade passing frequency. @DanielK points out that the 75 Hz fundamentals coincide with the vibrations of the fan rotor unbalance, amplified by the fuselage:

This vibration energy is transmitted as structural vibration through the engine structure, through the pylon and the wing, and then causes the fuselage itself to vibrate at that frequency. This causes the interior of the cabin to become a giagantic loudspeaker, at the rotor frequency.

The fan unbalance loads on the structure are a function of thrust provided by the fan, so reduction of power after TO reduces both the fan frequency and -amplitude.

Answer 4

How does a many-bladed fan spinning at a few thousand RPM generate a (harmonic-rich) fundamental around 50 Hz?

Easy, but you'll need to bear with me :)

For the electrical grid to function, the loads on the grid must all be met by standing up generating capacity - or you get into annoyances like rolling blackouts, or uncontrolled grid collapse.

However, loads on the grid are highly variable through time - peaks occurring a few hours a day, being different on weekdays and seasonally. In summer, load at 6 pm can be 3 times load at 2AM (we're talking pre-electric cars; those may really help fill in usage valleys.)

Why are we talking about electricity?

To cover peaking loads, they have peaking units aka "peakers". The most popular peaker is an airliner jet engine core turning a generator instead of a fan, and run on methane instead of Jet A in most cases. These "aero-derivative" units are lightweight, movable and can be serviced by the aviation supply chain. (notably, 40 years ago, peakers were locomotive engines, which are optimized for 1000 RPM = 50Hz / 3.) For instance

• Pratt & Whitney GG4/FT4 (24 MW) <- J75/JT4 - Delta Dart, F105, B707, DC-8
• Pratt & Whitney GG8/FT8 (25-30 MW) <- JT8D - A6, A4, B727, B737 original, DC-9
• GE LM2500 (20 MW) <- TF39 (early CF6) - early C-5 Galaxy
• GE LM6000 (40-56 MW) <- CF6-80 - B767, A310, 747-400, MD-11, C-5M
• GE LM9000 (66 MW) <- GE90 Boeing 777

Some of these had longer careers as stationery/marine engines than as aircraft engines! GE will still sell you a new LM2500.

Mind you, it's not like they design for peaking only and then try to fit it to aviation as an afterthought. But it's not the other way round, either. If they feel that engine fits the stationary generator market, the peaking application is on their mind from day one. Note that 1500 and 1800 RPM are also useful for generation since 4-pole generators can be used.

Almost looks like a scene at an FBO.

The biggest markets for peaking units use European style 50 Hz grid frequency. As such, jet engine designers have incentive to design the N1 shaft speed to have 3000 RPM (50 Hz) as a "sweet spot". Pretty much simple as that.

Now they do make heavier-built turbine engines, probably marine derivative, that have better long-term durability - the tradeoff being more costly, and too heavy to easily ship offsite for rebuilds, chaining you to the local labor supply and skill levels. A lot of peaking plants run only a few hundred hours per year and such engines would be costly overkill. If you're thinking "It's a darn shame to have so much costly infrastructure that is so rarely used", I'd agree. And if you're thinking "Wait, electric cars and home batteries could be a game-changer", I agree too. But it's easier than that; we just haven't gotten organized.

Answer 5

I have a woodfire burner in my sitting room with a 25 foot long 5" diameter flue pipe. When lighting this with a firelighter block, the flue pipe will start to resonate so strongly like a 25' long organ pipe that it shakes the whole house (solidly built, all brick farmhouse dating from 1837), One can control the volume of this resonance by varying the draft to the fire but the frequency remains pretty constant and as the draft is reduced, the resonance will suddenly cut off completely. I can therefore well believe that the low frequency droning on high thrust (CFM 56 engines and their close relatives seem the most prone to this) is due to organ pipe type resonance of the tube that forms the channel for the air from the ducted fan. The exciting factor for this may be variations in the air flow from the fan but the basic frequency would be a function of the length of the tube and the intensity controlled by the volume of air passing through the tube.