Cyclone design software, Barry

[ivan]

Well, it just shows you shouldn't speculate too much! Just after giving Pentz the benefit of the doubt in another post, I discover that while Penz's 'no development' claim might have been true 20 years ago, there has been a lot of recent computer driven development with the aim of increasing cyclone % separation, especially of smaller micron sized particles. This may, or may not consider energy efficiency - good separation with low horsepower...

Another problem: the composition of the incoming air /solid mixture is one of the design parameters, and of course the wood/air ratio varies from machine and cut to cut, also the range of particle sizes.

http://aerosol.ees.ufl.edu/cyclone/section08.html gives acess to a web based cylone design software you can try. Might be fun!

 

[Barry Burgess]

Thanks Ivan I will be playing with it.
Barry

 

[ivan]

Did a bit more digging last night and came up with some interesting stuff you might like, Barry.

Like most circular arguments (Pentz v World) it looks as if at least something on both sides is actually right...I found the following Phd thesis. published in 2004. If you did A level pure maths it won't look too foreign, mostly needs only a bit of algebra. The various intros, and summaries make very interesting reading and contain virtually no maths at all.

The point about Phd theses is that they have to be original work, so the writer's main conclusions have to add to the worlds' body of knowledge. Here's a couple of abstracts where he's setting the background for his work. For 'agricultural processing' read cotton industry and for 'abatement device' read cyclone.

In the agricultural processing industry, 2D2D (Shepherd and Lapple, 1939) and 1D3D (Parnell and Davis, 1979) cyclone designs are the most commonly used abatement devices for particulate matter control. The D in the 2D2D designation refer to the barrel diameter of the cyclone. The numbers preceding the D relate to the length of the barrel and cone sections, respectively. A 2D2D cyclone has barrel and cone lengths of two times the barrel diameter, whereas the 1D3D cyclone has a barrel length equal to the barrel diameter and a cone length of three times the barrel diameter.....Simpson and Parnell (1995) introduced a new low-pressure cyclone, called the 1D2D cyclone....The cyclone design procedure outlined in Cooper and Alley (1994), hereafter referred to as the classical cyclone design (CCD) process, was developed by Lapple in the early 1950s.

Further into the paper there's lots of tables of cyclone sizing incorparating his new analysis, including some very small ones. It seems the 1D2D configuration has the least flow resistance, suiting lower powered systems. Worth a download, methinks.

 

[Jake]

For 'agricultural processing' read cotton industry

When I was a teenager, I worked a couple of summers at a grain wholesalers during harvest. They have to clean and dry the grain as it comes from the field, which involves dropping it through sieves to get out big extraneous stuff and sucking loads of air through it to lift off lighter stuff like dust and weed seeds. Anyhow, all that air was sucked out and through big mother cyclones that would scare the living daylights out of Barry's (comparatively) tiny little one. They must have been 15/20 feet tall, plus stilts.

 

[Barry Burgess]

Ivan I have downloaded the PhD and will do some reading in bed. It was a very long time ago I did maths at a level and I also did it at Uni - there might be hope as I have seldom used it since.
Thanks again
Barry

 

[ivan]

Good on yer! Will do likewise over next few days, hope this will give some context to the stuff on Pentz's site too.

 

[Barry Burgess]

Ivan the main discussion is on the 1D3D and the 2D2D whereas Pentz's is 1D1.64D. He derives D from 3 times the pipe diameter but other than the length of the total cyclone I have missed how he came to 1.64 verses 3??
Barry
ps Maths in the PhD is OK - slowly plodding through

 

[ivan]

Barry, can't find any ref to 1D1.64D on the web unfortunately, don't know where Pentz got this from. However the thesis might give some idea where his design, and your 'accident'** fits in the scheme of cyclones, and after a skim of the 100 odd pages and a concentrated read of the summaries, I went back to Pentz's little CV in the introduction - it just doesn't seem likely he made it all up as some detractors appear to have implied.

I was interested to see how Bill Pentz (BP) compares with thesis 1D3D for your small vac cyclone. BP chooses a 15A Hoover whole house vac motor*. Current hoover bumf rates this @1600W/102cu ft min or ~ 2830 l/min. With a 1.5x3=4.5 sq" inlet, cyclone design inlet velocity is 3,200ft/min. It's 1D1.64D with D=6. From his drawing the ramp appears to turn the vortex once in 3", the height of the rectangular inlet. That's about 5 turns in all if this is maintained. The outlet is 3" dia. and 6"long.

Funnily enough the design inlet speed for the 1D3D cyclone is also 3200, so the inlet would have the same x setional area as BP's, and the reccomended airflow for D= 150mm is 93 cu.ft/min. with a 3" dia outlet also 6" long, and a vortex of just over 6 turns (thesis tables). Goodish match so far? The puzzle is how you get to shorten the cone - perhaps this is a funtion of the ramp. Maybe we should try an email to BP sometime.

BP's full size cylone appears to fit the 1D3D model if you choose 900 cu ft/in*** at 3,200 ft./sec, which gives D~18" and inlet area of 40 sq". Indeed BP talks about a 1D3D design (cyclone plan,paragraph headed cyclone design) pointing out the problem of a motor+4D+wastebin giving overall height too great for normal room. His revised design shorter, and with less flow resistance.

* Equating to UK vacs, a 1000W Numatic or small Festool does 2400l/min, Alto1500W 3600l/minand my 2 motor Numatic 4800l/min. Where does the vac you tried fit in? Also did you have a smooth transition from vac hose to cyclone so flow into the 1.5x3" inlet was laminar, and not turbulent? Also the Dyson has a twin concentric cyclone, the outer (just a little bin) removes the bigger stuff so the fine dust removing inner doesn't block up.

** Not a rudery-I mean the happy but accidental meeting of flowerpot and drum.

*** Folk grumble BP is confusing. I suppose he does rabbit on (If I were he, I'd be paranoid about dust too) However when he goes on to talk about 1900 cu.ft/min motor/impeller, he is talking about just that. Flow will drop to ~800/900 cu.ft/min when this is attached to the cyclone, filters, and pipework, so this is the figure the BP cyclone is designed for.

 

[Barry Burgess]

Ivan none of BP's cyclones either the large or the small use ti 1D3D model - he uses the 1D1.64D ie a diameter of 18" gives a cone length 52.52". The 1D2D model discussed in the PhD requires the least airflow and appears at first glance to be the least efficient but the test material was fly dust and manure dust not anything like wood??
Barry

 

[Barry Burgess]

* Equating to UK vacs, a 1000W Numatic or small Festool does 2400l/min, Alto1500W 3600l/minand my 2 motor Numatic 4800l/min. Where does the vac you tried fit in? Also did you have a smooth transition from vac hose to cyclone so flow into the 1.5x3" inlet was laminar, and not turbulent? Also the Dyson has a twin concentric cyclone, the outer (just a little bin) removes the bigger stuff so the fine dust removing inner doesn't block up.

** Not a rudery-I mean the happy but accidental meeting of flowerpot and drum.

Ivan I used a 2000W Morphy Richards - who knows what it produces. With my other non BP cyclone I also tested with a 600W and a Dyson which produced bad results.
The transissoins from round pipe to rectangular in the inlet are similar in both the Flower pot and the baby cyclone, They are both not that smooth but I heated the rectangular pipe and fitted the round pipe in. Filled in the gap at the top and bottom.
This is the inside of my non BP cyclone -

All I had to play with was the length of the exit pipe to the vacuum
Barry

 

[ivan]

Barry, the spreadsheet and diagram don't, no indeed. However in BP's blurb, just above his cyclone plans, he says this:

To build an 18" diameter cyclone with an optimal cone length of three times the upper cylinder diameter you get a 72" too tall cyclone for those of us with 8' ceilings. Adding a 30" dust collection drum below and up to 24" of motor-blower above leaves a standard 8' (96") ceiling "vertically challenged"!

and goes on

For wood dust the best cone length occurs when it is 3 times the upper cylinder diameter. That also gives the best particle separation until you get to a much taller cyclone. Unfortunately, that creates a too-tall cyclone for many woodworkers who have an 8' ceiling. The next best reversal point is when your cone is 1.64 times that upper diameter. That's what I choose to use in my design to get a cyclone that will just barely fit under an 8' ceiling with a small dust bin.

Thus he starts from a 1D3D and then modifies his ideas from this point to get his 1D1.64D design which also has a bit lower resistance (changed inlet and ramp, shorter cone)

I'm not wanting to argue with BP at all; I thought it would be useful to understand cyclone design better, and to see where his design fits in, in the face of transatlantic criticism, and maybe help see why your cyclone works bettr than you expected.

 

[ivan]

I think 1D2D is what BP says the commercial designs are, so they work with lower powered motors (cheaper) and being shorter, fit parcel size of common US carriers. I got interested in this config. till I read it doesn't suit fine dust- make a good chip collector though!

You said your small cyclone blocked up? I wodered if this was because the flow was too turbulent so waste was picked up from the bottom of the cone

 

[Barry Burgess]

Ivan I think the 2" exist is too small and if I re built it I would use a 3" outlet and make the cone 3D not 1.64D. My test is a little unfair as I take the collection bucket from the big cyclone and use that as the test dust(about 30 liters). The big one picks up chucks of wood and a number of leaves if I vacuum up the workshop. The non BP cyclone works fine but the baby cyclone has a problem.
Barry