PTS Flowbench Forum

Hi and thanks for adding me to the forum.
I'd like to do some experimentation on optimizing a (regulation mandated) turbo inlet restrictor https://www.specialstage.com/photopost/ ... rawing.gif The dimensions of the 'choke' of the restrictor are stipulated by regulation and the downstream pressure recovery area is dictated by those regulations and the compressor inducer diameter but the upstream side is free. I'd like to experiment with different 'bell mouth' shapes to see if changing this geometry can either a. net a greater mass flow at 'choke' or b. net the same mass flow with lower downstream depression (i.e. working the turbo slightly less hard/lower turbo rpm/less pre turbine back pressure).
I have tried to do this in CFD (Ansys Fluent) which I have access to in my job but I'm not an analysis expert and can't get it to work properly so, being a mechanical engineer, I thought why not test it for real.
The trouble here is that most commercial flow benches, whilst technically capable of moving the sort of CFM required to choke a 34mm restrictor, will operate at 25 or 28" H2O (below atmospheric) test pressure. This is no where near sufficient to choke the restrictor. In order to move 28lbs/min (I think about 370 CFM at standard pressure/temp) I will need something like 0.5bar absolute test pressure.
Now, this is a big ask for any vacuum source and I'm considering options. This is where I'd like some advice.
Options I can think of are:
1. old turbocharger with turbine housing removed, pulley welded in place of turbine wheel and belt/pulley gearing to BIG electric motor (I'm predicting 3phase, 20-30hp)! - the motor would be a bit more than I want to spend and I don't like the idea of belts and potentially out of balance turbos running at 100k rpm!
2. A big capacity (by UK standards) normally aspirated engine free revving at about 4500-5000rpm - I know someone with a chevvy truck but it's not the most practical thing to use a truck as your pressure source and free revving the engine long enough to record results is not very kind to the truck.
Any suggestions welcome.

Not knowing a lot about building a flow bench, does anyone have any comments about the practicality of what I'm attempting and whether the PTS plans would be suitable for my purpose (with appropriate vacuum source)?
I like the look of the digital manometer but would it read down to the low test pressures I'm looking to work with? I do have data logging capability to read MAP sensors.
Otherwise I was thinking of using an analogue bourdon tube gauge for test pressure rather than a manometer (as unless I fill it with mercury a manometer would need to be 20ft tall)!
I've seen a couple of variations in diagrams (Wikipedia and elsewhere) for plumbing an inclined manometer - should it simply be one end of the manometer tube either side of the orifice plate? (a proper diagram would be really useful).
If the manometer is indeed plumbed ether side of the orifice plate, what I'm wondering is whether the distance from the orifice plate (on the downstream side especially) is critical because of the Vena Contracta/flow separation and resultant turbulent flow off the back edge of the orifice? Similarly I'd have thought that pipe diameter relative to orifice diameter would have an effect here???
How is the orifice plate/inclined manometer calibrated?
Sorry for so many questions, hopefully I've come to the right place for answers! Chris.

Hey Chris welcome to the forum.

Your estimate of .5 bar is approximately 2 to 20 times the pressure that most people flow test at. You're talking approximately 200 inches of water. Your testing ideal is at least twice the explosion strength ( personal injury ) of any commercially available flowbench and way beyond the capabilities of flowbench specific digital measuring equipment.

The required vacuum isn't available with vacuum motors. A large centrifugal fan is a a possibility.

I understand what you want to test and your correct in thinking that the air entry will improve the flow capabilities of the restrictor. A ram tube, bell mouth or any type of radius will help but only to a certain point. You can try different designs to work out what works the best but because of your restriction there will be a limit. A radius will improve it but it's not an infinite deal, what I mean is bigger and bigger radius won't equal more and more flow for long.

A conventional flowbench can seriously show you the results of this at a low depression without getting dangerous. It's a pretty simple test. A low test pressure will still show the difference that you need to see here.

Engines run way beyond 200 inches and bench testing at these levels just isn't practical although plenty have brought up the argument that 28" isn't realistic. A lot have tested at 60" and it's rumored some f1 and cup car development was done at 160. Changes in ports have shown different results at different pressures but an air entry is pretty easy to test.

If you want to test it at 'real' pressure, here's a bad idea. ( hillbilly method )
1. Acquire friends truck
2. Fix test piece to engine intake with vacuum gauge connection downstream of choke point.
3. Friend brings engine to a set rpm.
4. Change entry until lowest vacuum is seen. That's your least restriction.

Keep in mind the ambient temp, humidity, baro pressure, engine temp, rpm and throttle position will need to be in the exact same place each test ( near impossible ) for the changes to be reliable data.

On second thought just test at 28

Jim

Thanks for the tips. It would actually be an implosion if the flow bench apparatus couldn't take the 0.5 bar absolute pressure (vacuum) but it's an obvious concern.
It might not be necessary to go quite that low. I know that sort of dynamic pressure is seen just post restrictor but there is a bit of a pressure recovery area to the restrictor and the actual downstream pressure may not need to be quite so low.

I might be wrong, but I'm just not convinced that a conventional flow bench (I have access to a Superflow 600) will give anything like realistic results as the mass flow/velocity will be so much lower (than critical flow) and the Reynolds number changes?
Surely any fluid has inertia, right? It will be less inclined to change direction at higher velocities. Think of the 'drop' of a bullet, a bullet of the same mass will see less drop at higher velocity. I reason that there would be greater flow separation (off the back of the choke) at higher air speeds. What effect the upstream geometry has on this is what I want to prove.
I have seen a Youtube video of exactly these type of restrictors being flow bench tested in Germany and I simply don't believe the (comparative) results (based on empirical observation of what 'works'). I have the exact dimensions of a WRC restrictor from a couple of years ago and he is a long way off with his stubby little design compared to a €50millon/year team.
If I have the dimensions of a works restrictor, why not just use that you may ask? Well, I don't have room in my installation and need to make it a bit more compact but I still want to optimise it within my packaging constraints.

The 'least restriction' test sounds plausible but don't forget, we're dealing with compressible flow. Would the upstream geometry affect how the 'ideal gas' is compressed? I understand that, at choked flow, further downstream depression will not increase mass flow and only an increase in upstream density will increase mass flow at choke, but is it possible that the upstream geometry is actually acting to increase the density/compress the air? Or is it more a case of simply 'smoothing' the flow into the smallest diameter of the restrictor?

The other thing is, watching this https://www.youtube.com/watch?v=JbqloCTvTrc
(aside from the fact it says all have the same inner diameter when one clearly has 33mm written on it and another 34mm)
there is absolutely no way the difference between the best and worst restrictor is 52% flow at choke. If you were making 275bhp on restrictor #1 (perfectly feasible) you'd jump to over 420bhp? I'm not buying it I'm afraid...

I believe turbocharged aircraft can maintain sea level power up to typically 10,000 feet where the air density is roughly half. But then the exhaust turbine will be exhausting into the same lower pressure and should actually work better at altitude.

I have no personal experience with any of this, but on the surface it looks like about half an atmosphere might be about the low practical design limit for compressor inlet pressure. Anything lower might be more than the turbo can make up for. So what you are suggesting sounds realistic.

A long time ago I worked at a Government materials testing laboratory, and they had a vacuum source that consisted of a fairly large roots blower driven by a 30Hp motor, and I remember a demonstration, and hearing one of their sonic vacuum nozzles suddenly go completely silent as it went sonic. This was about thirty five years ago, but the sonic nozzle itself would fit in the palm of your hand, the bore size would have not been that large possibly 35mm as a wild guess. It had a fairly short curved bell mouth as I remember, and the whole thing was quite compact.

The only other place I have seen a similar large vacuum pump, a roots blower driven by a 15Hp electric motor was in a desalination plant at an Antarctic base. This boiled sea water at low pressure and at only slightly warm temperatures, and worked at a similar low pressure.

The Forum digital manometer should work, but it would need different pressure transducers fitted, to suit the much higher differential pressures.
I do not know if the software calibration and number crunching would go off scale, Rick or Bruce might be able to advise ?

That is about all I can offer.

Yeah for sure, compressor designs have a practical pressure ratio limit, you'll see this if you look at compressor map where the vertical axis is the pressure ratio (and horizontal is airflow).
The thing I've never quite understood (as a benefit) with running turbos at altitude (whether that be aircraft or something like Pikes Peak) is as the atmospheric/barometric pressure reduces, to maintain 'boost pressure' you surely have to keep the wastegate closed longer and spin the turbo faster; this means that (ideally) the whole turbo (compressor) design needs to be optimised for the chosen operating conditions else you risk over speeding the thing or operating outside of the maximum efficiency area of the compressor map? I guess with the aircraft, they might use some extra power for takeoff but really they can keep the wastegate open at lower altitudes and only use the turbo to its full potential at higher altitudes.
A supercharger as vacuum source may be easier to gear to an electric motor (or indeed an IC engine) given the power required to drive it.
As an aside, regarding a sonic nozzle, I can see how in the case of a restrictor such as I describe above or any other venturi, the maximum velocity will be present at the smallest diameter of the venture and a the diameter tapers up again downstream of the throat the velocity will slow and the dynamic pressure will recover. The mass flow will be limited once the velocity in the choke is sonic.
However in a convergent-divergent rocket nozzle the velocity downstream of the throat increases and continues to accelerate as the diameter increases - this to me goes against Bernoulli and is something I can't get my head around.

Crice,

The project sounds interesting and in my opinion is more like testing venturi designs, sonic nozzles or or even de Laval nozzles which are all about the inlet and outlet shape. At the pressures and velocities you are speaking of even a machining burr will cause turbulence that can throw the whole system into turbulent choke. If it were me i would do allot of research on valve seat design and how minimal angle changes and widths have such a huge effect on pressure recovery and thus port CD. For what you are trying to do it will require a good test fixture that is repeatable to a very small differential, the good news is that A to D conversion costs are falling and 16 bit A-D (65535 bits) is cheep so measuring down to .003" H20 is not out of the question.

Test fixture, first off i would follow Jim's idea of flowbench testing even up to 60" to get your head around what this thing is doing and what crude changes have as an effect on delta P, I would also at this point look at bench total current draw as a metric. Then scrap the conventional flowbench idea find yourself a cheep used Roots supercharger and a 20 to 30 HP lawn tractor engine. You will be flowing in one direction only so the design does not need to be complicated but you will need some type a settling chamber maybe a 24" x 24" or bigger piece of storm pipe. Throttle control and a good sized flywheel should give you decent stability from RPM fluctuations at ideal test pressure and make it more repeatable. I also think you need to consider using multiple metrics on you data analysis, flow, pressure, velocity (multiple points), current draw, engine rpm, weather data etc..

So the goal as i see it, is really not trying to measure an improvement in actual flow but an improvement in pressure recovery efficiency! Remember choke is a limiting flow concept but choke can happen prematurely due to turbulence created on either side of the restriction that fundamentally reduce the effective CSA.

Just some thoughts,

Rick

I guess with the aircraft, they might use some extra power for takeoff but really they can keep the wastegate open at lower altitudes and only use the turbo to its full potential at higher altitudes.

The wastegate has the usual spring, but it works opposing a vacuum filled aneroid bellows. At sea level there is zero boost, as the bellows sucks away all the wastegate spring seat pressure and you get full rated sea level engine power. As you climb, the inlet manifold stays at sea level pressure up to the point where the wastegate is full closed at perhaps 10,000 feet. Above that power falls of quickly.

Its not usually practical to add extra boost at sea level, because the engines are air cooled and already running at some safe cylinder head and piston temperature design limits for sustained climb conditions. Just as you say, the turbo is optimised for maximum altitude to maintain sea level manifold pressure. It also sometimes pressurises the aircraft cabin to sea level pressure as well.

However in a convergent-divergent rocket nozzle the velocity downstream of the throat increases and continues to accelerate as the diameter increases - this to me goes against Bernoulli and is something I can't get my head around.

The fact that its still burning on the way out may have a significant effect on classical nozzle airflow theory.

Antilag turbo systems, and afterburning gas turbines also move the goal posts around a fair bit.

Thinking back through 35 years and that very brief sonic nozzle demonstration. The roots blower and its motor were located in a small remote hut for noise reasons, with about fifty feet of large bore pipe to the laboratory building. The depression created in the test laboratory just had a steady continuous air roar until it went sonic. There was no evidence of pulsing or the usual drone associated with roots blowers that are generating a positive pressure.

If you think about a roots blower on the suction side, its quite different to the explosive back flow you get on the pressure side as the rotor tips are uncovered. The intake side should make a much smoother vacuum pump, than the violently pulsing flow on the pressure side.

Anti-lag is one thing I do know a fair bit about. Contrary to popular belief, most systems (especially those deployed on gpN or ‘clubman’ cars will not allow a turbo to generate boost at low engine speeds where there is insufficient exhaust energy to drive the turbine. As such it only prevents the turbo slowing down too much while ‘off thottle’. I put off throttle in inverted commas because with a jacked open throttle setup the throttle is never fully closed and engine torque and idle speed are controlled with a combination of ignition retard and fuel (and/or ignition) cuts.
Very few systems I know of allow actual combustion of fuel (to any great degree) in the exhaust manifold and do not do so as a torque augmentation system. The exception to this is the ‘rocket’ system used exclusively by Prodrive on the Subaru WRC cars which has a combustion chamber in the exhaust manifold. I have the only version of this that I am aware of outside of a genuine WRC car and the only one running on a commercially available ECU rather than the TAG McLaren ECUs used on the WRC (and F1 cars) at the time.

Crice wrote:...2. A big capacity (by UK standards) normally aspirated engine free revving at about 4500-5000rpm - I know someone with a chevvy truck but it's not the most practical thing to use a truck as your pressure source and free revving the engine long enough to record results is not very kind to the truck. Any suggestions welcome.

I'm thinking a truck engine might reach redline before reaching the flow you need unless it was loaded enough to require more throttle opening.

Years ago (maybe 1960s?), there was an American company (possibly Ingersoll Rand) which bought V8 engines from Ford, and used custom camshafts and manifolds to run one bank of cylinders as an inline four-cylinder engine, and the other bank as a four-cylinder, two-stroke-cycle compressor. No change to the crank, rods, pistons, valves, cooling or lubrication. The compressor side took in air but no fuel, and exhausted into a tank, while the engine side ran normally with a carb and exhaust system.

I didn't pay much attention at the time I saw this thing, because I didn't envision an application, but I remember marveling at the low number of parts changed to do the conversion. This would seem to be close to what you need, if you could smooth the pulses. You wouldn't even need the tank on the exhaust side, just vent to an open pipe. It would be self-contained without torque or weird couplings to manage, and could run be mounted on a stationary pallet.

Check back in with what you decide to do. This is an interesting project!

Paul in Utah