Fans in Series

[Engineer6512]

One of my colleagues is designing a system that has two centrifugal fans in series. His question to me is how the static pressure is developed in the duct.. for example, each fan is generating 5000 cfm at 12" W.G. There is approximately 20 feet of duct before the first fan, and 20 feet of duct between the two fans.

If we were to install pressure gauges every 5 feet along the entire duct, what would the pressures look like? Would we see -12" vacuum in the first inlet duct? or would it be -24"? Would we see +12" right up to the inlet of the 2nd fan?

Would it be different for non-centrifugal fans?


Steamguy2

http://www.SteamPlantEngineering.com

- Discussion & resources for professionals in the Power Generation Industry

 

[lilliput1]

You need to do a trial and error solution to find the system curve that matches the series system curve of the two fans. Since the fans are in series and running at the same speed, each fan would see the same CFM and pressure rise. Assume a CFM and make pressure drop calculations of each of the (3) duct segments (before, between & after). Include inlet and discharge losses. See SMACNA duct design manual. Then assume other CFM. The pressure drop across the duct segment now will be PD2 = (CFM2/CFM1)^2 x PD1.
Find the combination of CFM and 1/2 PD total that will fall on one fan curve. You can use the above squared CFM relationship to find the presure drop increment as the air travels through the ductwork.

 

[ko99]

Assuming you've done as lilliput suggests and find:
1. 5000 CFM at 12" is the operating point of each fan
2. The two 20 foot duct sections are identical

Then the pressure at 5 foot intervals would look something like:

--------------------------------------------------------
<---|fan|-12 -9 -6 -3 0 |fan|-12 -9 -6 -3 0 <---
--------------------------------------------------------

ko (

http://www.ecooling.biz)

 

[quark]

12"WG is the differential pressure and not appears at the suction totally. If I take the duct size to be of 1'x1' then there will be about 0.6" WC of pressure drop at suction, for 20ft, and the discharge pressure can't be more than 12-0.6 = 11.4"WG(So, suction of first fan will be 0.6"WG below atmosphere). There is further drop of 0.6"WG in the second fan inlet. So the suction of second fan is 10.8"WG above atmosphere.

 

[ko99]

Good point, quark. I ignored the fan inlet and exit losses.

ko (

http://www.ecooling.biz)

 

[Engineer6512]

So what I would actually see is this:

--------------------------------------------------------
22.8<--|fan|10.8 11.1 11.4|fan|-0.6 -0.3 0 <---
--------------------------------------------------------

Correct?

Steamguy2

http://www.SteamPlantEngineering.com

- Discussion & resources for professionals in the Power Generation Industry

 

[AEBuck]

Determining the static pressure profile along the complete length of a duct system with two fans in series is no different than developing such a profile for one fan, except there will be a second rise in static pressure profile due to the second fan (assuming the intersection of the system curve and the combined fan curve falls within the operating range of both fans). Lilliput1 correctly described the procedure for determining flow in the duct system for equally-sized fans. Starting with 0" total pressure and 0" static pressure near the entrance to the duct system, simply use this flow to determine positive and negative changes in total and static pressure along the duct length.

Neglecting duct intake, duct discharge, and fan inlet and exit losses, and assuming no changes in duct size for the moment ...

If there were 20' of duct upstream of the first fan, 20' between the fans, and none downstream of the second fan, then the static pressure at each fan inlet would be -5".

If there were 20' of duct before, between and after the fans, then the static pressure at the first fan inlet would be -3.33 and the static pressure at the second fan inlet would be -1.67 (each equal duct segment experiences a loss of 3.33", while each fan adds 5").

The procedure is the same regardless of fan type, but please respect fan manufacturer's recommendations for fan inlet and outlet duct configurations - which differ with fan type. For example, flow at the entrance to the second fan must be fully-developed following turbulence generated from the outlet of the first fan. Failure to respect these distances will require Advil, money and possibly a cold call to replace that lost client.

 

[ko99]

Steamguy2,

Your chart is not correct based on your initial description "20 feet of duct before the first fan, and 20 feet of duct between the two fans". As AEBuck said, the exit pressure should be 0".

Also, be aware that AEBuck's first example assumes the fans operate at 5" each and I assumed they were 12" each. Both of these are guesses. Otherwise we are in agreement

Also note that we both neglected duct and fan inlet & exit losses and assumed equal loss in each 20' section.


ko (

http://www.ecooling.biz)

 

[AEBuck]

Sorry about that 5 versus 12. Slap me silly for not checking the facts.

 

[lilliput1]

This setup is not real. there should be the equipment that would be eating up the fan head. There should be automatic damper to close when the fans are off. There may be a requirement for discharge stack. The fan type, inlet and discharge size should be given. Well I don't have a fan catalo with me but I do have my SMACNA base duct design program. I inputed the setup described by Steamguy2 and I assumed SWSI centrifugal fans 20" dia. inlet and 16"x20" discharge. I just selected sizes at around 2250 fPM velocity. Next I inputed various sizes of ducts with the lenghts per Steamguy2's description. I included duct inlet loss, 20' run of duct, transition to 1st fan inlet, 1st fan discharge transition to the duct, 20' of duct, transition to 2nd fan inlet and free fan discharge (assuming backward inclined blades). The duct size I came up with that would have total 24" wg pressure drop is 8.95". The duct entrance loss is 1/2 the velocity head per SMACNA. My calculated losses are as follows: 4.1" for entrance, 3.9" for the 20' straight run, 5.6" for the transition to fan inlet, 0.4" for the fan discharge transition to the duct, 3.9" for the 20' duct, 5.6" for the transition to the 2nd fan inlet and 0.5" for the BI fan free discharge. It is ridiculouse to have 8.95" duct. The pressure drop taken up by the equipment served should be included. Note that you must start both fan at the same time or else the lag fan wheel will be spinning in the opposite direction and the fan amp draw will exceed its rating once started with blades spinning backwards.

 

[friartuck]

Am I missing something here?

There are 2 fans connected to a system. The system hasn't yet been fuly described. After the 2nd fan, where does it go? Does it continue for several hundred metres absorbing the massive system pressure. (I am assuming it is supplying of course)

We also don't now the duct sizes between the fans and the duct inlet sizes. If these are large enough, the pressure drop will be negligible regardless of what pressure the fans will achieve. The fans will simpy interact on their fan curves with the system curve (Which also needs to be calculated)

So, we might have a revised system operating duty nothing like anyone has yet mentioned.

If however life was perfect, and the designers had designed the system to match the fans, i.e. the ductwork PD was 24" wg. (I would like to see that operating, I bet the duct will throb a bit).
Then we still need the system fan outlet details.

We might have something like this

atmospheric inlet
-2" wg up to the first fan
add 12" wg from fan 1 = +10"wg.
-2" wg from 20' of duct between the 2 fans=8"wg
add 12"wg from 2nd fan=20"wg

the 20" wg would then have to be absorbed by the rest of the system

PS, this is just a guess, you have to draw the fan curves (In series you add the pressures but keep the volume constant)

Then you estimate a system pressure drop based on your design airflow rate.This gives you a PD.
Once you have this, create a system curve using the equation PD=KxVolxVol

You work out K from knowing vol and the design PD.

Substiture several vols in the equation to give you a set of PDs fpr a range of vol flows.

Plot this on your fan curve and it will give your operating condition which in turns tells you the PD across the fans etc etc.



Friar Tuck of Sherwood

 

[Engineer6512]

Yes Friartuck.... this is what I intended... I didn't mention this in my original post, but I am designing the discharge system to be for 24" W.G. Sorry if I wasn't clear in my original post.

Steamguy2

http://www.SteamPlantEngineering.com

- Discussion & resources for professionals in the Power Generation Industry

 

[lilliput1]

You can do this duty with one fan. Check out New York Blowers. They have units can can deliver up to 84" wg. If dual fans are deswired for reliability, consider two in parallel. This would provide about 60% capacity if one fan is down. If you need continuous duty, consider a 3rd fan as spare. Because of the high pressure the fans will be noisy. Sound attenuators may be required at inlet & discharge side. You need automatic isolation dampers for each fan. How are fans controlled? should they be variable speed? Are filters required? Are heating or cooling coil required? What is the process/equipment eating up the 24"wg.? You have to add all equipment losses and duct friction loss to determine the required fan total pressure. Will the heat gain from the fan be acceptable? You will need about 40 hp.