High Secondary Chilled Water Pump head

[Buildtech2]

I am currently designing a chilled water plant of total cooling capacity of 2400 tons. There are 4 nos duty water cooled ammonia chillers of capacity 600 tons each. The chillers are water cooled type with evaporative condensers.The total chilled water loop is quite large about 2000 m including both chws and chwr.The pumping configuration is constant primary and variable secondary system.The total chilled water flow is 3600 gpm with 16 degree delta-T. Before performing actual pump head calculations using spread sheet, I always do some rough calculation using thumb rule as follows but I foudn that pump head for the secondary loop is quite high and I have to take a decision if primary / secondary pumping configuration is suitable for this plant or should I consider primary, secondary and tertiary pumping system to minimize the head. Any though on this would be highly appreciated.

Total length of chilled water supply = 1000 m
Total length of chilled water return = 1000 m
Pipe friction loss = 3.5 ft / 100 ft
Total chilled water pipe length = 2000 m = 6560 ft
Total friction loss = 1.5 x system length (ft) x friction rate (ft/100 ft)
Pipe friction rate = 3.5 ft / 100 ft
Total friction loss = 1.5 x 6560 x 3.5 / 100 = 344 ft

Rules of Thumb – BSRIA 2001
0.36 X circuit length =
Pressure in kpa and length in meter.
Total circuit length = 2000 m
Pump head = 0.36 x 2000 = 720 kpa = 241 ft .wg.

 

[trashcanman]

Lots of variables:

1. What is the gpm flow rate?
2. What type of pipe, age, condition, etc?
3. For a system this size, I would not use "Rule of thumb". Do a formal calculation using Darcy-Weisbach with Churchill's friction factor.
4. Where did the 1.5 * system length come from?
5. A diagram would be helpful.
6. Does the system use glycol? That can affect friction rate.

 

[EnergyProfessional]

I hope the plant size itself was determined by other means than rules of thumb....

You should actually calculate the pressure drop with a software. At minimum use ASHRAE Fundamentals methods and also calculate fittings, valves etc.
And take into account chillers have cold water, table values assume standard water of 68°F etc. This also impacts pump performance (pump manufacturers have software to account for that).

 

[Buildtech2]

I always perform proper hydraulic calculations using spread sheet but before I do that, I always check by thumb rule the approximate head which was found to be much higher.
There will be 4 duty pumps in parallel and one standby and the flow rate for each pump is 900 gpm.
The thumb rule equation is: Total friction loss = 1.5 x system length (ft) x friction rate (ft/100 ft.
The system use chilled water but not glycol.
The plant size was determined by proper cooling load calculations with HAP 5.0 by adding the the load of all the AHU's and FCU's. However, diversity was not considered in the chiller load.

 

[LittleInch]

Pipe size?
Flow velocity?
Put in a bigger pipe if you're worried about pressure but 7 to 10 bar pressure loss for only two kilometres sounds like too small a pipe to me. I think that's what your second ROT is telling you - economic pressures will make your pipe a bit bigger whilst lowering pump and pumping costs.

For something which works a lot of time, don't ignore the OPEX cost for all that extra pumping. You can often make a case for a bigger pipe once you throw in a year or two of energy saving.



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

 

[quark]

240 ft for 3600 gpm is too high, unless your pipe diameter is less than a feet.

 

[Buildtech2]

Noted. I will perform hydraulic calculations to verify the actual head.

 

[nglty]

Few things you will need to consider

1. Pumps in parallel do not add up in GPM. i.e 4 pumps at 900 GPM each will give you no more than around 1600 to 1800 GPM.
2. You should look into one single pump, using 3500 RPM motor. Or may two approx. 2800 to 3000 GPM Pumps with 3500 RPM to get your 3600 RPM. Yes, with 240 feet of head or whatever you calculated. That would be your most efficient design.
3. 3500 RPM of this size will be noisy as hell, you cannot stand in the room without hearing protection, make sure enclosure is sound protected, and have some ear plugs at Mechanical room entrance ready for workers.

Google parallel Pumping and you will see the problem of 4 pumps at 900 GPM each. Use B&G software and use multiple pumps in parallel and you will see the max GPM you get.

 

[quark]

1. Pumps in parallel do not add up in GPM

. That needs some explanation.

If your system resistance is calculated for full flow and each pump is selected with a head equivalent to system resistance (for example, system resistance is 100 feet at 3600 gpm, then each pump should be selected at 900 gpm and 100 feet), there is no reason why the flow shouldn't add up.

 

[Artisi]

quark, you are correct in challenging that statement.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[nglty]

Quark and Artisi

The flow does not add up, because two pumps in parallel follow the system curve.

Even when I indicated to google Parallel pumping, you did not have the curiosity to check before questioning the statement.

Artisi - as you say, it is indeed a capital mistake to theorise before one has data. Read-on.

https://energy.gov/sites/prod/files/2014/05/f16/optimize_parallel_pumping.pdf

https://www.engineeringtoolbox.com/pumps-parallel-serial-d_636.html

 

[LittleInch]

Well you can both be correct.

If you take quarks data then you can split the duty point into as many pumps as you want ( four is bit much for me) and when you run all the pumps together then the duty point will add up to 4 x one pump.

However nglty is also correct if you have one pump which is rated for a particular duty point and then when you add another identical pump in parallel it won't get you double the flow if the pump intersects the system curve without modification (i.e. no VFD, no control valve etc, as per the link provided.

Everything depends on the shape of your system curve and pump curve so it's not possible to get a fixed view of these things.


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

 

[nglty]

Sure, you can have multiple pumps, but each pump will have a HP close to the single pump in order to get the GPM. Which doe snot make sense.

Try select pumps in parallel with B&G software and you will see, instead of just talk guys. Get the data and then come back and show us how the four pumps will deliver the flow.

 

[moideen]

I have a related question…OP says one pump is standby out of 4 pumps. Each pump 900 gpm. then how flow come 3600 gpm, 900*3=2700gpm. Or when sizing the pipe ,will consider the standby also?

 

[LittleInch]

No he said 4 duty and one standby = 5 pumps in total.

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

 

[Buildtech2]

Both Quark and nglty are correct.
According to nglty, when one pump is already operating and when we add another similar pump in parallel, the flow does not become double because both the pumps operate at the same system curve which means without changing the system resistance, you are simply adding a similar pump in parallel, therfore the overall system resistance has increased which cannot give double flow rate.

According to Quark, all four pumps are simultaneously operating at the same system resistance which will easily add up the flow to 4 x 1 pump.

The point which I am trying to highlight here is if you want the system flow to be doubled by adding similar pump in parallel, the system resistance to be reduced which means increasing the pipe size.

 

[moideen]

I have same experience that recently during the checking of condenser pump flow rate, I found one pump is lower than design flow rate. The 1350 gpm design / name plate and found pushing only 620 gpm. Then I started second pump, found the flow rate comes down! one pump only 340 gpm and the flow rate with two pump around 710 gpm. One pump is also not designed against the system head. Some colleges argued that why two do not increased the total flow. This matter already had discussed on another thread.