Aerodynamics Measurement Criteria

I think the only criteria that must be kept in mind is that the pressure sensing holes must be smooth and burr free and obviously the airflow sensing hole must point upstream in the airflow while the static pressure sensing holes must be at right angles to the air flow.

I don't think the actual hole dimensions matter as there is no airflow through the sensing holes..?

John.

Dear Electroman

Your approach is possitive but it is not sufficient for design point of view. I think there are some more critical dimentions are to be considered. Please think and suggest. Eagerly I am waiting for you reply.

Thank you

MOHD. SAEED KHAN

I think your appraoch to this conference forum is a little brusque shall I say...

What do you think the critical dimensions you mention are then??

Perhaps if you provide the 'critical' dimensions you require we might be able to answer your questions...

John.

I suggest that instead of designing a Pitot tube you should rather select and buy one of Pitot tubes available in the market (like Dwyer) (they are rather inexpensive). In their website, you will find specs and also the velocity versus delta p graphics.

When you will see these graphics, you will find out that you have to measure the pressure differential and the temperature as well.

At this point, you should make your own calculations as to the accuracy, precision and resolution you need for your measurements, in order to select the instruments to read the variables values.

With this requirements, you can go for the inexpensive instruments or you can buy the more expensive ones. You can use portable meters (like the electronic micromanometers from Alnor or any other reliable manufacturer which include Pitot measurements, i.e. converts the pressure differential into velocity) or dedicated sensors (like MKS, Setra or any other suitable for the requirements)

Anyway, you should qualify the measuring point as this is the most important design feature you should take into account. Any tiny misplacement of the Pitot tube will lead you to erronous measurements, much larger than the errors from the measuring meters. The Pitot also has its own measurement error (not negligible) and is not so accurate when measuring low velocities.

Please do your homework and read the application notes you will find in many websites.

Dear Mario

I know there are numerous type of PITOTSTATIC TUBES are readily available in the market but being a Mechanical Engineer we should know how it is desigened to teach others, so, I don't want go to short cut. Design means design. Please think again and suggest.

with regards

MOHD. SAEED KHAN

Any special reason for choosing a complicated way to do this? why not just install mini anemometer, easly available no mods needed Directly calibrated and so on. After all it is the wind speed you want to measure not the Pitot system carrier, (an airplane. )for which it was originally designed.

Wangito.

Dear Wangito

I also don't want go through the complicated way. The test section of our wind tunnel is 600mm x 600mm. We also put an airfoil, bluff bodies, car modles etc. for pressure, lift and drag measurements. That is why I want to design a pitotstatic tube. So please concentrate on design and suggest.

with regards

Thank you

MOHD. SAEED KHAN

Your tunnel is a nice size for school use: small enough to allow relatively high speeds without high energy consumption, but large enough to filter out some of the problems inherent in making small scale models. Ironically, the best place to design a really effective, highly optimized pitot tube is in a wind tunnel (or in a reliable 3d CFD program). The interactions between the pitot and the object being measured will have a large influence on measurement accuracy, For example, if you want to sample speeds near a foil surface, then the presence of the pitot tube itself interferes with the measurement. The angle of flow near an object will not be parallel to the wind tunnel wall, and if you want to measure flow relative to the surface in question, (rather than relative to the "straight" flow path) you will need to be able to very accurately adjust pitot tube angle to the slipstream angle.

You will want to be able to move the tube around to almost any location within your tunnel. Certainly, you will want the smallest and most streamlined tube possible to minimize interference effects. I'd suggest an airfoil shape for the strut on which the pitot tube mounts, probably just 3mm - 4mm thick by perhaps 15 mm in chord, given sufficiently stiff material to avoid vibrations (unidirectional carbon fiber, for example). If you keep the strut aligned accurately with the flow to be measured, then the pitot orifice could be simply drilled into a flat spot on the leading edge of this airfoil. On the other hand, if the flow is not accurately aligned, then you'd improve accuracy by making the tube more conventional, with the orifice at least two chord lengths ahead of the airfoil strut.

That said, my strong preference would be to simply make a few and experiment, beginning with just the strut (support) in the tunnel to check for vibration (flutter). Remember that on many older airplanes, the pitot tube was just a hunk of copper tube, bent by hand into the general direction of the airflow. Granted, aerodynamics is not rocket science (it's more complicated than rocket science) but it is also not as demanding as brain surgery.

Personally, for instruction purposes, I think there is some merit in classic, fully mechanical balances and actual manometer tubes. However, that arrangement makes data logging a manual task -- but even that can be a good thing, for learning purposes. A great many things can go straight from CFD to production without ever having passed through a wind tunnel, and I believe there is a tendency to rely too much on software and too little on brain power. For example, the Aptera is a small vehicle that is in prototype here in the US. The promoters claim very high MPG figures (330 mpg at 65 mph). When I calculated the Cd required to achieve that figure, I came up with .06. Coincidentally, this is the figure that their designer had quoted, I later found out. But as you know, .06 is about the drag you would expect from an airfoil section spun around into a body of rotation (a torpedo, in effect) and anything with wheels, doors, air intakes, mirrors, etc, etc, won't be that streamlined. I'm sure their number came from a CFD program, inappropriately applied. In short, there's no substitute for real hands-on experience, and mechanical balances and manometers help solidify concepts for students, I think. But that's just my opinion, without any real hard evidence to support it.

Have fun, and get your students involved in the design process (if I've correctly jumped to the conclusion that you are involved in education).

I'd agree with all the previous remarks: namely that a mini anemometer might be a better choice, but if you want very high speeds (which you will need in a small tunnel) then a pitot is worth considering. The pitot tube itself can be remarkably crude, with hole size and even finish being non-critical. (There is flow through a pitot tube only when the speed changes, and then only enough flow to move the instrument.) You'll need to play with calibration, and you'll need to be especially careful to locate the pitot tube or other measuring device where it is out of the way of the airstream -- in other words, so it is not in a position that will cause turbulent flow when you are looking for laminar flow.

Buy one; why "reinvent the wheel"?

The important thing is to do a traverse across the tunnel to obtain a representative view of the velocity at each point in the duct. If the Reynolds number is sufficiently high, as suspected for a wind tunnel, then a single measurement may suffice. The pitot tube measures the impinging velocity of the static streamline. The difference in head at the attached manometer indicates the velocity at that point. Check out a fluid dynamics textbook and go from there.

Why does the operator of a wind tunnel not know the airspeed inside it at any time?

One of the main concerns in flow measurement is the error introduced by placing the instrument, Pitot or anemometer, into the fluid flow. A total process measurement uncertainty need to be calculated, so one may be able to discern actual flow variations from variations due to different sources of errors (evaluated trough the calculated expanded uncertainty).

Metrolog, I don't think inserting a pitot tube into a 600 x 600 mm square tunnel is going to cause any noticeable error in reading...

More importantly is the laminar flow of the air and temperature as well as density i.e. the pressure, if the readout is going to be in litres / min the ambient pressure and temperature must be corrected for.

John.

the problem will be : how are you going to calibrate your pitot static tube? without calibration it is no better than putting your head in the wind and trying to gauge the wind speed. i suggest you can refer the website of nist. they have international standards on designs of pitot static tubes and also procedures for calibrating it,. however is is important to correct for temperature and pressure and also take the total uncertainty calculations as well. there is going to be noticeable error in measuing wind speed if you simply poke your pitot tube into the tunnel. better find the average veocity profile and take the avg value.

Sirius, there are standard look up tables giving you the air velocity related to pressure, so calibration is easy with a pitot tube.

Almost every other method you do need to calibrate each instrument, I would guess that is probably why the pitot tube is the simplest method of measuring air speed...

John.

Although I usually agree with practically everything you say, John, this is not one of those times. Some of the issues:

Speed across the tunnel is not constant, even in a well-designed tunnel.

The probe creates its own pressure gradients even well-forward of the probe itself, and the nature of these gradients change in a non-linear way (and often in a non-logarithmic way as well).

The probe operates at its own Reynold's number, distinct from the Reynold's number at which the model is operating. (In other words, there can be bubbles forming around the probe that render it inaccurate at certain Reynold's numbers that are important for the model.)

You will often want to probe all around the model itself -- that is really the main reason for using a probe. If all you wanted to do was measure lift or drag in a general sense versus angle of attack, (or other model orientation) then a calculated wind speed based on the motor speed, ambient air pressure, etc. gets you nearly close enough (once you have calibrated that). But often, you will want to probe all around a model to find just what the velocity distribution is -- only by knowing that, can you change the model in a way that "cures" a velocity distribution "problem."

There are strong interference affects between two airfoils operating in proximity (and in a tunnel, one of those airfoils is the probe). Therefore, it is often very difficult to know whether you are measuring the actual airflow at a particular point around the model (as it would be if the probe were not there) or if you are measuring mainly the influence of the probe itself. The difficulties are enough to make grown men cry.

If I were to think of an analogy, it would be like trying to measure a very low level fluctuating electrical signal, by probing with a big inductive multimeter -- where the meter not only does not tell you what you hoped to find out, but actually renders the circuit virtually inoperative.

Ironically, Mohamed might have the advantage of having access to a really good CFD program, so that he can see what the flows should be at various points around a model, and then see if his probe accurately measures those flows. I say ironically, because the gold standard is still (in my opinion) real windtunnel time vs CFD time if your model can be expected to generate turbulent flow or separations. Ordinarily, you use the wind tunnel to evaluate the CFD, rather than the other way around. (Although we are now, after all these years, at a point where opinion is split, I think -- but I dunno, I'm kind of a luddite, with a strong preference for touching and feeling things.)

Ken, I bow to your superior knowledge on this topic...

It is not one I've had experience of. Although the Question didn't state that the probe was to be used around a test model, if it did I would agree with what you said about the probe affecting the result...

I thought the air speed to be measured was in an open 600 x 600 mm tunnel with nothing else to cause vortices or any deviation from a nearly laminar flow.

I agree the wall boundary later effect will make the volume flow smaller than measured, but that should be a constant? Maybe?

Of course you are right, if the question was about measuring air flow around a test object then its a different matter entirely.

John.

John,

No need to bow whatsoever: "superior" knowledge is a stretch -- a real stretch. I've worked with wind tunnels a bit (and practical application of aerodynamics to real world things, like car bodies, is my "thing") -- but much of what I've said is theoretical, rather than (necessarily) practical. It is also based on an assumption that the original poster is using the tunnel for instructional purposes -- which I am assuming by its size. Then (as the assumptions further pile up) it is interesting to show students, I think, the real local velocities, which vary substantially, close to the thing being tested. It's especially interesting, I think, to see pressures close to models of things like car bodies, where flow is less predictable (for a student, or someone like me) than around a wing section. (Eventually, some people get to the point where they are human CFD programs.... more or less -- and stuff that seems weird to mere mortals seems perfectly predictable to them.)

If the original poster is using the tunnel for non-instructional purposes, then a lot of what I've said is a little "over-the-top" in comparison to the actual requirements. The stuff you say starting with "I thought ..." might very well be true -- in which case the original poster is probably wondering "What planet is this guy from?" re my detailed post.

For a real view of what works from someone who has loads of real-world experience as well as a very strong theoretical basis, you can look at Tom Speer's post above. I've mentioned elsewhere that I value scientists (in response to those who think scientists are a bunch of conspirators and hoax propagators) and that several have served as mentors and have been inordinately generous with knowledge. Tom is among those who are very active in helping others understand aerodynamics, and has also done a lot of hands on stuff (like sailing and designing interesting sailing things). In fact, he is widely known as the God of Aerodynamics, and you must bow in his presence. (OK, the last is a stretch, but he is an all around good guy who knows far more about this stuff than I do.)

For a wind tunnel, it may not be necessary to use a pitot-static probe. A simple pitot tube may be just as good. You can measure the static pressure at a port in the tunnel wall.

The pitot tube itself can just be a simple, square-ended tube placed far enough into the tunnel to be away from the wall boundary layer. Such a tube will read an accurate total pressure as long as the local flow angle is within 10 - 15 degrees of the tube axis, so alignment is not very critical. Since the total pressure is constant everywhere the flow is irrotational, only a single probe is necessary and it really doesn't matter very much where you put it.

You can confirm this yourself by moving the probe to various locations. If you see a variation across the width of the tunnel, then something upstream is shedding a wake that is reducing the total pressure there.

If your tunnel is of the open circuit in-draft type, then you know in advance what the total pressure will be - it's the static pressure outside the tunnel.

The static ports can be arranged along the length of the test section to measure the pressure at each station. Static pressure is constant across a boundary layer, so the pressure at the wall will be the same as the pressure outside the boundary layer.

The classic text book on wind tunnel instrumentation is Alan Pope's "Low-Speed Wind Tunnel Testing". I'm sure you can find it in just about any engineering library.

Hi Tom:

I appreciate your stopping by. I tend to be closeted away working on this project or that and don't work with either CFD or wind tunnels every day, so I am feeling a little out of touch with current technology.

Seems to me that,for many things (pipe fittings maybe being the most basic), CFD programs are considered as good as a wind tunnel or water tunnel in the sense that if you are off by 10% the risk is low. So, for many things, you could go from CFD to production without going into a tunnel.

For car bodies, it has recently become marketable to clean things up -- but still styling often (usually) overrides aero, so that a Prius with its .26 Cd is considered very good (even though its is only 10% better than many other sedans). Then there is the Aptera, which claims .06, which strikes me as impossibly low for anything on wheels -- I'd guess it might really be .09. I suspect their number comes from a CFD program inappropriately applied -- but maybe I'm all wet. Has CFD progressed to the point that wind tunnels are quickly becoming obsolete, even for complex shapes like a car body, where you have to expect fully turbulent flow in places?

At Boeing, where the cost of 1% additional drag can be huge, is CFD now sufficient for going ahead with production without using a wind tunnel?

Dear Dr Tspeer

Thank you so much for your suggestion. I have to go through the book Low Speed Wind Tunnel testing by Alan Pope's, but the dimentions (how much should be inner and outer diameter for the inner tube and in comparision of this how much should be inner and outer diameters of outer tube) are not clear in this book. More over the position of static holes are also not clear. If you have these data kindly send immediately.

with regards

MOHD. SAEED KHAN

I hate to admit how old this is . . .

I have a 1967 AMCA Standard 210-67 that gives detailed dimensions and geometric proportions for pitot-static tubes and flow nozzles, traversing patterns, etc. Let me know how I can send you an image (somehow I'm not smart enough yet to attach to this format).

No need to worry about how old it is -- I'm sure you bought it a used book store recently as an antique.

I have a couple old (mature? seasoned? wise?) texts and references that I love to go through. Abbott and Doenhoff can tell you a tremendous amount in just the couple of minutes that it takes your computer to boot up!

I wish I had more time (and a wife that understands) to puruse older technical volumes. When I really want to have a hoot, I will open up a 1918 SAE handbook. The "dead men" had interesting standards and methods!

Look in the internet for specific standards regarding the design of pitot tubes. Yes, they are ruled by some ASME or something like this. I remember seeing one once, but I'm sorry I just cannot remember.

Basically, you'll measure total and static pressures, and with the difference and the geometric conditions of the wind tunnel, you'll be able to know the air speed.

But I also would like to remember that an air flow in a closed channel is not uniform in the entire cross section. Even if you are designing the wind tunnel to have a turbulent flow with a more flat speed profile, its always a good idea to have several pitot tubes with independent pressure measurements to read different distances from the wall. With this care you'll be able to better determine the speed profile.

Two options for accounting for the speed profile:

first is to take measurements at various points within the flow stream and average the results. AMCA has charts for round, square, and rectangular points to measure and profile. Depending on how many points you average, it can be time consuming.

second is to make up an array of pitot-static devices and connect them together to get an average. Many years ago, a plant I was in bought an air handler system with a couple of averaging pitot static systems from Cambridge Scientific. This system was in a 42" by 84" duct, and I believe it had an array of 16 or 24 points.

The ACMA standard I did forward to Mr. Khan also included a measurement array for round duct. I'm sure a Google search might turn up a similar chart for square duct. I believe his system was 600mm square.