Low Flow and Pump Problems with odd Placement of Chiller in multiple chiller/pump system

[jtimmer]

I have inherited a chilled water system with a unique design that is having problems. The system has four 15 Ton chillers that are water cooled. Each chiller has its own pump; the original design is for 3 chillers and 3 pumps to be online at any one time. The unique aspect to me is that 3 of the chillers are located together at the end of the system and the final chiller is connected in the middle of the system and there is just one single loop with no segregation or zones to it. See the diagram. The problem they are having is that the chiller/pump combination in the middle seems to starve the chillers at the end of the system. When the middle pump is not running the other pumps (even when all 3 are running) have a hard time keeping the system cool.

As I look at the system I believe this is due to a number of factors:

First, the expansion tank is connected at the inlet side of the middle chiller and will have little effect on helping the suction pressure at the back pumps due the distance and pressure losses. The return line goes back to the other chillers first then even father back to the pumps and by that time it seems to me that the expansion tank pressure will not help.

Secondly, the system (as far as I can tell) is designed for 3 chillers and pumps to supply the water needed, however the operating procedure that is currently used appears to be to run all the 4 pumps regardless of which 3 chillers are operating. My analysis of this is: since the flow control devices are constant flow valves and the control valves are 3 way bypass valves at each air handler, if the middle pump is running it can very likely starve the 3 pumps at the end of the system because there is only a set amount of water that will flow, does that make sense. I am not totally sure why the flow rate is not up to what it should be based on the constant flow fittings in the system however.

My question then, aside from getting the operating conditions correct (only run 3 pumps at any one time), what is the best way to correct the suction pressure on the back 3 pumps. I was thinking of adding a pipe off the expansion tank that runs to the common suction header on the back 3 pumps , but wanted to know what the thought is on this, will teeing the expansion tank into two location in the same system have any negative effects? Is there a simple way this could be tested with minimal system intrusion?

Thanks

 

[dbill74]

If there is just the one expansion tank and IS sized for the entire system, then it is improperly located. The system would do well to have the expansion tank on the main return line.

Based on the information provided, my guess is that the "middle" pump is improperly sized. What is the head/flow of that single pump compared to the three main pumps? I'd be curious to know if it was added as a retrofit project (it looks this way), and if so what was the reasoning for adding it. Knowing this could expose other problems that need to be addressed.

It almost looks like someone tried to add a "secondary" loop to a single loop system and didn't know what they were doing. When you say "Each chiller has its own pump", are you referring to the main pumps or does each chiller have a separate pump as one of its components?

Doing a google search for Secondary chiller piping will yield several PDF's from ASHRAE that should help with understanding.

With little else to go on, I'm betting you will need to redo things at the "middle chiller" to put it on a "primary loop" (verbiage sounds backwards to me, but that is the convention). Repipe the middle chiller to come off the supply CHW and dump its chilled water back into the same supply line, these connections should be close together and no take-offs or anything between on the main supply line. The pump for this new primary loop would be sized for flow through the new loop with just enough head to get the water through loop pipng and the chiller.

 

[jtimmer]

Yes the expansion tank is sized for the entire system, and as far as I can tell the system was designed this way. All the pumps are identical, from the flow rates and pump curve/design point the pumps were sized to provide system flow with 3 pumps online.

What I meant by "Each chiller has its own pump" was simple referring to the fact that there are 4 pumps and 4 plants, the back 3 pumps in essence serve the back 3 plants, though they are all tied to the same discharge and suction headers.

 

[dbill74]

So yea, that's what I thought you were talking about.

With all 4 pumps being equal in terms of flow and head, the "middle pump" (on the right side of your sketch) is fighting with the other three. Simple way to test and verify is to disable able it and the associated chiller for a day and see if the system maintains flow and temperatures at the AHU's (I bet it will).

Assuming this test does work, you then need to change the function and performance of the middle pump.