Vacuum pump sizing vs pipe conductance 3

[bernoulli313]

Hello all!

The engineering office I work for has been tasked with sizing a small vacuum distribution network for a pharma facility. None of us have experience with vacuum systems, and we are therefore working with a vacuum pump and piping supplier.

We can safely assume these guys know their stuff, but we want to do our due diligence anyway.

One thing we're trying to wrap our head around is the conductance of the pipe, and its impact on the pump selection. The supplier has sent us some flow/pipe lenght curves for different pipe diameters (see attached). My question is: should the duty point of the pump fall on the "with conductance" line (which varies with the pipe diameter), or can it fall between the "with conductance" and "without conductance", and, if not, why not? What information can we get from the "without conductance" curve?

This is probably an easy question for folks working with vacuum on a daily basis but for us, it's a whole different science.

Many thanks!

 

[Doug Hunter]

Hello,

Please clarify the X-axis of the chart. Is it pipe length?

Is conductance a measure of resistance for the piping, or something else?

I've not worked with these circuits, but I have worked with automotive vacuum pumps. The pumps are usually primarily specified by their displacement and rated by their evacuation curve for a fixed volume and drive speed. This evacuation curve would be rendered in terms of vacuum pressure vs. time. You may be more concerned with steady-state operation and maintaining a certain vacuum level with a certain amount of air leakages present.


Best regards,
Doug Hunter
Altarium Technical Consulting

https://www.altarium.us/

 

[LittleInch]

Interesting. This gives a bit of background.

https://www.leybold.com/en-uk/knowl...f-vacuum/how-to-calculate-vacuum-conductance.

"conductance" foxed me to start with, but is essentially flow restriction at very low pressures when actual velocity is high compared to volume flow due to the low pressure. not sure how they figure out the resistance or if they use a max velocity, but it's clearly something specific to the vacuum pump supply industry.

My understanding is that you want to be above the line with conductance line so that when you're trying to pull a vacuum on the vessel you're attached to you have a greater flow available to you when the pressure is higher than your required vacuum level. This diagram says you're attached to a 1000m3 tank - that's quite big!

One thing to ask though is what exactly is meant by "vacuum". This is normally expressed in in mbara or torr.

The graph you show doesn't show what pressure this is at though and I'm working on the bottom x axis being total length of pipe.

As everything there is a balance between having a large pump which can draw a tank down faster to start with, but then essentially get restricted by the size of connecting pipes versus one which is closer to the conductance line which might take longer, but still get there.

Do you have a time versus pressure chart provided?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bernoulli313]

Hi Doug,

Yes, it is pipe length. We are at about 50m of pipe.
From what I gather, conductance is the equivalent of pipe roughness,i.e. pipe resistance, for vacuum, except that it varies with lenght, pressure, diameter...

 

[bernoulli313]

Hi LittleInch,

(I've been reading your posts since FOREVER!)

It's a 1 m³ tank. ;-)

More data: 15 mbara, 987 lpm

This chart is the only one provided. We can ask. Attached is a chart with a different pipe diameter, with the duty point hitting the curve.

 

[LittleInch]

Ah,

Now that makes sense. The latest attached is all 40NB pipe, the first one had a load of 25NB pipe in it so the flow will be lower, hence take longer to get to your 15mbara if you're starting form 1000mbara.

Both still quote 1000m3 tanks, but maybe that's just so they can get a good graph - who knows?

So there is constant gas off at 15mbara? If you want to move 60m3/hr of vapour at 15mbara, it doesn't look to me like your smaller pipe option will cut it and you'll never get to 15mbara. But then maybe I'm reading this wrong.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Doug Hunter]

bernoulli313,

Thanks for clarifying. I've got nothing to add to what LittleInch is advising.

Best regards,
Doug Hunter
Altarium Technical Consulting

https://www.altarium.us/

 

[katmar]

I strongly suspect that the X-axis on the charts is the pressure in mbara and not the pipe length. I see in both cases the operating point is at an X value of 15 and this corresponds to the pressure and not to the pipe length.

I have not seen this type of chart with the two curves before. You should ask the supplier how the "including conductance" curve is derived. I tried a whole bunch of pressure drop options and none made sense. If you are paying money to these people the least they can do is explain what they are trying to sell you.

You are being a whole lot more charitable than I would be when you say "We can safely assume these guys know their stuff". I need to be convinced of this in each case before I place an order.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[bernoulli313]

Katmar,

yes, indeed. Pressure and not pipe length. Duh!, in hindsight.

That's what you get when all you obtain is a screenshot with no further explanations.
Our client is actually buying directly from the supplier, but they involved the supplier after installing the pipe network, and now the supplier is saying the pipes installed are undersized.

I will come back here when/if I hear back from the supplier on the curves.

Oh, and thank you, LittleInch for the Leybold link!

60m3/hr of vapour at 15mbara is what we'll see at the pump (diversification included). But the DN25 is going to rooms with lower use (approx 120 lpm).

 

[LittleInch]

That makes sense.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[pierreick]

Hi,
To add to the discussion.
I'm not familiar with the terminology "conductance"! Q=Delta P*C
Interesting to perform a google search.

https://indico.cern.ch/event/264020/attachments/467855/648232/CERN-Basic-Vacuum-2012-lecture.pdf

Good luck.
Pierre

 

[katmar]

If the main pipe is 41 m long and 25 NB Sch 40 then I would agree with the supplier that it is too small.

But first, a rant. Whenever volumetric flows of gas are given it is essential that the basis be given with the flowrate. It is meaningless to say 60 m3/h without stating whether it is free air, actual conditions or standard conditions.

My calculations indicate that the pressure drop in the 25 NB pipe would be too high to achieve a flow of 60 m3/h at 15 mbara. You would only achieve a pressure of around 30 mbara in the 1.0 m3 chamber. And 60 m3/h at 30 mbara is a very different kettle of fish than 60 m3/h at 15 mbara.

I see that the supplier has a field on their charts for the gas load to be specified in kg/h. This is a positive sign and I would encourage you to work in mass units to avoid all the ambiguity of volumetric flowrates.

Early in my career I was told off very severely for specifying a compressor in mass flow units because "nobody does it that way". Nevertheless I have continued to specify gas flows in mass units, usually also specifying the volume units to keep the others happy. A kg is always a kg, but a m3 is not always a m3.

If the main pipe is increased to 40 NB the pressure drop from the chamber to the pump will only be around 1.5 to 2.0 mbar and I would say this is a far better choice.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[katmar]

I see that while I was typing up my rant, Pierre has posted a perfect example of what I was getting at.

Each branch of hydraulics, from fire water to vacuum systems, developed their own arcane terminology and methods of problem solution. Back in the day this allowed them to solve routine problems with charts, graphs and circular slide rules (but these only in the most technologically advanced fields). Unfortunately these practices have persisted, and continue to confuse everyone. Those days are gone, it's time to move on and use the modern tools.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[LittleInch]

Basically "conductance" is the same as pressure drop as far as I can see.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[katmar]

According to the reference LittleInch gave, the conductance is not the pressure drop but it is the multiplier that must be applied to the pressure drop to calculate the flow rate in volumetric terms.

Under vacuum conditions the density of the gas is usually low enough to ensure that the Reynolds number is less than 2000 and the flow is laminar. In laminar flow we can use Poiseuille's Law so the flow rate is related to the pressure drop by

Flow = [ (pi/64) * diam^4 / (viscosity x length) ] * pressure drop

This makes all the terms in the square brackets the conductance.

The caveat is that Poiseuille's Law applies to incompressible fluids and gases can only be treated as incompressible if the pressure drop is small (<10%) relative to the absolute pressure.

This was probably a useful concept in the days of slide rules and log books, but I seriously question its relevance today.

But LittleInch is correct that with regard to the charts produced by the pump supplier when they refer to "including conductance" they mean that the values take the pressure drop through the piping into account.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Compositepro]

High-vacuum pumps are rated in volumetric flow at actual inlet conditions. They are basically positive displacement pumps that move a constant volume per unit time. Flow stays constant until you get to where the vapor pressure of the lubricant or off-gassing of the materials of the vacuum chamber overwhelm. Low vacuum pumps are limited by design features such as unswept volume or back-flow leakage.

High-vacumm is below about 0.1 to 0.01 Torr where flow is due to molecular diffusion and not due to pressure differences. The goal is usually to limit the number of gas molecules per cc. It is a very different and specialized field. As I recall pipe conductance is proportional to the fifth power of pipe diameter. This is where the phrase "you cannot store a vacuum", used in another thread, applies. When you just need suction, different concepts apply.

 

[bernoulli313]

Katmar,
All flows are actuals. I thought "actual" when typing the values, but didn't actually type "actual".
Thanks for your rant. If only the world could agree on one system of units, that would be nice. And use dots for decimal points and commas for thousands separator!

Everything above is usefull stuff! But we're still trying to understand the curves supplied, and what it means when the duty point does not fall on the conductance line. As written before, I'm wating for the supplier (Busch, for information) to get back to me.

Thanks again, all.

 

[katmar]

To answer your original question - the operating point must fall on or below the "including conductance" curve.

As I stated earlier, I have not worked with these conductance curves before, and although the Leybold web page is quite comprehensive it is a nightmare of unit conversions, nomograms and assumptions. To work through all of that without making an arithmetical error would be difficult. Let me show you how I would calculate this problem.

My assumption is that you want a pressure of 15 mbara at the chamber and the temperature is 20C. The desired flow is an actual 987 litres per minute of air at these conditions. Following my own rules, the first step is to convert this flowrate to a mass flow and this turns out to be 1.06 kg/h.

The next step is to work out what pressure would be required at the pump suction. So we need to calculate the pressure drop through 60 m of 40 NB pipe with an inlet pressure of 15 mbara and a flow of 1.06 kg/h. The calculator says 2.9 mbar so let's take it as 3.0 and look what happens at the pump when the suction pressure is 12.0 mbara. The "excluding conductance" curve shows that the flow is a little more than 81 (actual) m3/h. Converting 81.0 m3/h at 12 mbara and 20C to a mass flow gives 1.15 kg/h.

This is more than the 1.06 kg/h required and the pump looks adequate to me. If the supplier gives the full workings for the "with conductance" curve we will be able to see why the curve falls slightly below the operating point. The small difference could be due to my assumption of 20C as the base temperature.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"