Help me build a flow Bench for a formula car!

Discussion on general flowbench design
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Help me build a flow Bench for a formula car!

Post by Tony »

Flow figures are not absolute, they must be corrected back to standard pressure and temperature to be meaningful.
Usually that is standard sea level pressure and (I think ?) 20 degrees Celsius.

So what you do to calibrate your bench, is use a calibration plate that flows stated xxx CFM at sea level pressure and at standard temperature.

If you then adjust your own flow manometer slope to read correctly with the supplied calibration plate you are set.

What the actual flow through the calibration plate is, may vary from day to day, because ambient conditions change.
But your bench should always read the same, because the same air flows through the calibration plate, as the measurement orifice, and they are both subject to the identical ambient variations.

If you use something other than an orifice plate to measure flow, you need to start thinking about how you are going to correct that "other" measurement device so you end up with a final flow figure corrected back to standard sea level pressure and temperature.

By using an orifice plate as both a flow standard, and another orifice plate for flow measurements, they automatically track each other, and provide automatic compensation back to the standard conditions under which the flow standard was certified.

Its really just a sneaky trick, but it solves ALL of the problems of day to day variations in a very simple practical way.
Also known as the infamous "Warpspeed" on some other Forums.
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Help me build a flow Bench for a formula car!

Post by Tony »

There are two completely separate issues here.

The first is to build a flow bench that is repeatable over a very wide range of constantly changing ambient conditions.
Repeatability is the key if you are going to be able to reliably measure very small increments of change over readings that are separated over time.

Repeatability of measurement is absolutely crucial to success in developing real parts for your race car.

The other issue is measurement accuracy, which is a completely different thing to repeatability.
Your bench could be out by 25% it could read way out compared to any other flow bench, but that has nothing to do with its usefulness as a practical development tool.

If your bench is wrong by 25.000% and it stays wrong by 25.000% over its entire life, you have a precision tool for measuring small CHANGES in what you are testing.

The ratiometic bench is so powerful because the same day to day ambient variations affect both the port and measurement orifice in the same way. No compensations or corrections are required.

We are not interested in absolute ultimate calibration accuracy, it is completely secondary to the purpose of using an airflow bench as a development tool for making iterative changes to some part you are trying to optimise.
Also known as the infamous "Warpspeed" on some other Forums.
Philip_Robinson
Posts: 7
Joined: Mon Nov 28, 2016 6:33 pm

Re: Help me build a flow Bench for a formula car!

Post by Philip_Robinson »

Thanks, I think I'm starting to wrap my head around all of this. So I don't need a fully accurate bench to calculate the flow coefficient? just a repeatable one?

And I just made a huge realization about the design of my flow bench. In the build document I followed the author said that it was necessary to have laminar flow before the orifice plate and put in place a grid designed to straiten the flow. I was planning to do the same with some sort of honeycomb. But then I found out about baffle plates.

So I DON'T want laminar flow right? This seriously makes me question the knowledge of the guy who designed this bench and makes me wonder if I should just buy the plans from Bruce.
Last edited by Philip_Robinson on Mon Dec 05, 2016 1:25 pm, edited 1 time in total.
Philip_Robinson
Posts: 7
Joined: Mon Nov 28, 2016 6:33 pm

Re: Help me build a flow Bench for a formula car!

Post by Philip_Robinson »

But Wait, this website says turbulence is bad and we want Laminar flow. Now I am confused.
http://www.omega.ca/literature/transact ... -DIFF.html
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Help me build a flow Bench for a formula car!

Post by Tony »

What you want is for the measurement orifice to "do its own thing" unperturbed by anything that may create unnatural up stream air flow conditions.

How up stream air actually flows into an orifice is really interesting, and you can observe the effect with a piece of wool or string. Amazingly, most of the air approaches the hole radially over the flat surrounding area, and then falls over the edge into the hole. The air definitely does not rush at the hole from straight ahead from a distance as you may imagine it does.

In fact if air is forced axially towards the orifice from straight ahead, it totally screws up what the air naturally wants to do.
And its why placing an orifice in a pipe is such a very bad idea. What the air really wants to do is enter a very large settling volume where the energy contained in swirls and eddies can dissipate. This volume just cannot be made too large, but obviously there are some practical limits.

The measurement orifice ideally needs to be located right in the centre of a large flat area well away from any corners.
The baffle plate is there to minimise any disturbance to all this radially inflowing air. The design of the baffle plate is not hugely critical within reasonable limits, but leaving it out completely would be a big mistake.

Another consideration is the design differential pressure drop across the measurement orifice. If its made very low the velocity of air through the orifice will be low, and more easily perturbed by any up stream turbulence or disturbance. This can seriously effect bench repeatability because even the slightest up stream upset due for example, in positioning the head over the test hole can influence the readings.

At the other extreme, a very high design orifice pressure drop across the measurement orifice will promote very high velocity, all that fury and violence at the measurement orifice will dwarf any minor up stream upsets, so that is good.
But it has the big disadvantage of hugely increasing the total pressure drop through the bench, and the horsepower required, and electrical input to power the air blower.

A compromise is required here, and many flow benches operate around the 12 to 16 inches region. The PTS bench uses 16 inches, and that has proven over time to be an excellent choice.

All of this is quite flexible, but the PTS designed bench has evolved over many years and a huge amount of testing by many people all around the world. Ideas have been suggested and tried, and gradually the PTS design has evolved into a very mature and trouble free flow bench that anyone can build. Everyone makes slight changes, that is human nature, but the basic design plans are well worth studying, as they represent an enormous amount of development experience.
Also known as the infamous "Warpspeed" on some other Forums.
storm
Posts: 289
Joined: Mon Feb 23, 2015 6:10 am
Location: In the glorious Hunter Valley

Re: Help me build a flow Bench for a formula car!

Post by storm »

I could sit and read your posts all day Tony 8-)
Every job is a self portrait of the person who does it.
Autograph your work with excellence.
Post Reply