Closed Loop Head calculations

[jtimmer]

I have a quick general questions. For a chilled water system that is closed loop (the supply goes out from the pump through the supply branches which feed different cooling coils and then all feed back to the return branch and back to the pump) when I am calculating the pressure drop to get the head for the pump do I total the loss for each branch or only the worst case branch. I have found conflicting data that says "only the worst branch since the circuits are in really in parallel to each other and the pressure drop per circuit will always be the same as the total zone pressure drop". Just curious here because my thought was the pump had to overcome the total loss from each branch as that is what it is pushing against in essence.

 

[HPost]

You only need to consider the worst branch. The head loss is cumulative in series and the pressure in each branch should reach an equilibrium.

HPost CEng MIMechE

 

[jtimmer]

So, let me show this in a quick diagram to make certain I got it right. In the attached picture, if the pump is supplying water to all the Air handlers (AHU's) (the system is fully open), then the pump head for this system would only need to be calculated for the losses from the worst case scenario. Which means if the red highlighted loop is the worst case loop then that is the one I need to find the losses for (and of course include the chiller and any valves in the system).

But I can ignore the other AHU's?

Thanks for the help

 

[LittleInch]

It all depends IMO, whether the amount flowing through AHU 1, 2 & 3 is a fixed flow or actually varies.

Assuming it's the same for each one and has been set=up that way, then in order to get the overall flow resistance in a parallel branch system you need to break it into sections between each branch. Hence chiller to AHU1 branchhas all flow, AHU1 to AHU 2 has all flow minus AHU 1 and so on then with the addition of whatever the pressure loss is across AHU4 start increasing the flow in each branch until you get back to full flow fromAHU1 branch to the pump.

So in essence yes, you follow the line of greatest resistance, but that resistance needs to take account of the additional flow in some sections due to the water for each of the other units.

The difficulty in reality is that it is not easy to set up the system such that equal flow goes through all the units as this requires a larger differential pressure from delivery branch to return branch across AHU1 than it will across AHU4. What you sometimes see is that he flow in the return header starts at AHU1 flowing to AHU 2 etc before returning to the pump.

Depending on your diameters and physical distances of your branches, equalising flow through each AHU can be quite easy or quite hard.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[HPost]

This changes things.

You'll need to add process valves to modulate the flow so that you get some flow through AHU4. Valves installed on the other AHUs will increase the head on the pump.

With the valve creating a pressure drop equal to the pressure loss across the pipework leading to AHU4, and the same setting on each other AHU, the pressure drop is thus equal across all areas again. Hence you only need to consider the highest pressure in the system.

Hpost CEng MIMechE

 

[jtimmer]

First off the flow through each AHU is fixed but they are not all the same GPM. Each AHU has a certain GPM based on the room(s) it is cooling.

I think the way you have described it is how I did the calculation.

I started at the pump and took into account the total system flow in the discharge pipe, I used the total system flow for pressure loss calcs up to the first tee to AHU#1, then from that branch to the brach of AHU#2 I used the total system flow minus the flow that went through AHU#1 and so on until I get to the supply pipe for AHU#4 which only has the GPM for that AHU in it. Then on the return side I did the same type of thing back to the pump. Then to find the worst case scenerio I summed up each loop and the one with the highest pressure drop is the one I have to size the pump for (with saftey factors/calculation error factors added in). Does that sound right?

I agree, the branches right near the pump will have to burn off quite a bit of pressure while the ones at the end of the run could be starved of water if the system is not balanced correctley. I am finding that this is where the Cv values for the contral valves come into play in order to get the right GPM and pressure through each AHU.

 

[LittleInch]

Sounds right.

As I said, also think about whether your return header can start at AHU1 and flow in the same direction as the supply header. It'll make your control much easier and depending on the distance only need a extra bit of pipe from AHU4 to the pump and a small increase in pump power. All depends on the relative location, but if it's not finally designed yet it will save a lot of time during commissioning and also mean one AHU shouldn't be starved of cold water during peak flow and they should all start to work more or less at the same time during start-up

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[jtimmer]

Thanks for the quick help littleinch and hpost.