Pump Controls 3

[flexiblycool]

The Sequence of Operation of the chilled water system pumps (depicted in the attached sketch) is that when there is demand for cooling, the Primary Pump comes on. Upon proof of flow from the flow switch, the chiller associated with this primary pump comes into operation. If the cooling demand is not met with one chiller, the second Primary pump and its chiller comes into operation and likewise upon further increase of cooling load, the third primary pump and its chiller comes into service.
Likewise when one primary pumps is automatically switched off based on returning chilled water temperature, the flow in the system reduces to 400 GPM from 600 GPM. The fan coil units that are still in service will now see less than design flow. If two primary pumps are switched off based on returning chilled water temperature, the flow in the system reduces to 200 GPM and the fan coil units that are still in service will get much less flow than design flow for their capacity.
The Secondary Pumps are continuous duty constant speed pumps.
This scheme of things therefore appears flawed. This system would be fine for the 2 way valves, but not for the 3 way bypass valves. So the proposal is that instead of removing the primary pumps from service, let all the primary pumps stay in duty regardless of demand and just switch the chillers on or off to match the load. Is this acceptable?

 

[BigInch]

Are you saying that the secondary pumps never get switched off.
I'm not clear on how the 2-way, or 3-way valves affect the problem, but my expertise is not cooling.
Runing all the primary pumps all the time sounds like the most expensive option you could think of.

Independent events are seldomly independent.

 

[flexiblycool]

Big Inch: Much thanks for response.
Yes the Secondary Pumps are neither VFD nor do they switch off!
If the fan coil units were equipped with 2 way valves, upon meeting of the cooling demand, these two way valves would close and thus stop the chilled water from entering the coil. The GPM needed would therefore reduce and there would be no problem if the Primary Pumps cum chiller could be stopped by sensing either the return water temperature or by the pressure switch that would sense the 2 way valves of many fan coil units have closed causing the increase in pressure.
The remaining in duty Primary Pumps could continue to handle the GPM requirements of the fan coil units whose 2 way valves were still open.
In the case of 3 way valves, this would not be possible, because the pressure of the system would not rise significantly due to the bypassing of chilled water from the coils of fan coil units. Furthermore, the secondary pumps are not VFD so they will continue to run even when the demand is very little. So if the Primary pumps are switched off based on return water temperature, the other pumps would continue pumping water to all the fan coils (both in which the water is being bypassed by the diverting 3 way valve and those in which the cooling demand has not been met). So the GPM that the remaining (on duty) primary pumps will churn out will be shared by the off duty fan coil units bypass leg as well as the fan coil units in which the thermostat has not diverted the flow due to the cooling demand not having been met. But the GPM that will be seen by these On Duty FCUs will be less than needed for cooling because of the needless sharing by the off duty FCUs. So the only way of circumventing this scenario might be to keep all primary pumps running and only switching off the chillers depending on return water temperature.
Your feedback will be appreciated.

 

[flexiblycool]

Big Inch:
Another consideration is as follows:
Since the secondary pumps are constant duty pumps, their cancellation and replacement with an untrimmed pump impeller (to give a higher head) is being contemplated. However, there is another reservation. The actual GPM being handled by each secondary pump is 300 GPM instead of the 200 GPM shown in the sketch. This 300 GPM each means 900 GPM for all three secondary pumps whereas the total GPM of all three primary pumps combined is 600 GPM. Is this 300 GPM due to diversity factor in which the Primary pumps are sized for diversity whereas the secondary pumps are sized for the maximum possible demand?.

 

[BigInch]

That's a little more clear. It's late here, so let me think about it some more tomorrow.

Independent events are seldomly independent.

 

[flexiblycool]

BigInch: Flexibly cool is thankful and waiting

 

[BigInch]

by the pressure switch that would sense the 2 way valves of many the fan coil units (that) have closed causing the increase in pressure.

I don't think that is necessarily true, since once one primary pump shuts off and the flow drops to 400, the pressure drop required for the 400 gpm flow in the circuit should be roughly 1/2 of what it was at 600 gpm. A drop to 200 gpm would result in roughly 10% of the pressure drop at 600 gpm. Even though the pump diff pressure would tend to increase, that could be economically controlled with only a discharge pressure control valve and on/off switches on each pump.

The secondary pumps, if none shut off, would see their capacity at 600 gpm (presumably 200 gpm each), 66% of capacity at 400 gpm, and 33% capacity at 200 gpm. 33% might not be very efficient for those pumps, but so be it for now. Unless you cut off one fan coil circuit completely, then the next etc., your stuck with that mode of operation.

It is not clear what type of cooling must be maintained. Must it be cooling in all units continuously, or if the demand is unit by unit, ie. you can shut off one fan coil, then the other and the last. Is the system for room cooling where some cooling must be maintained on a warm day, or is it for freezer cooling, where one unit can be emptied of its contents and turned off, for example, then the next, etc. What you need to do there might have a significant effect on what measures you should take to do it.

This system seems basically to be designed for one and only one flowrate with room temperatures being maintained by turning on, or off, 1 or 2 chillers. It also doesn't seem that you could do much about it, other than adding a control valve downstream of the discharge of the primary pumps (or make them all VFD driven) to save some energy there, and perhaps making a similar installation on the secondary pumps, if you still wanted to squeeze more efficiency out of them too. What is the object of this exercise?

Independent events are seldomly independent.

 

[LittleInch]

I've read this a few times now to see if it makes sense and ultimately I agree with you and BI that the system as it currently is designed is intended to work with a fixed flow rate, whereby the 3 way valves on the FCU control cooling by bypassing, presumably via some orifice or other device to simulate the pressure drop accross the fan unit. However this is being compromised by the on /off nature of the primary pumps. Use of secondary pumps like this is also a bit odd, especially when the first one only raises pressure from 70 to 75 psi.

As I see it you have a couple of alternatives - raise the pressure in the system by increasing head from the primary pump such that the pressure at the far building is always > 75 psig and then introduce pressure control valves to limit each building to 75 psig and run the primary pumps continuosly and just turn the chillers off if not required.

Alternatively you could replace your three way valves with isolation valves (two way valves) and do the same as above to get rid of the secondary pumps, but turn off the primary pumps if flow rate falls ( flow meter required) or return temperature falls.

You really need to have the main control either at the coils or the primary pumps, not both. The chillers should control by maintaining the outlet temperature from the chillers.

One point though - your description at the start "the chiller associated with this primary pump " implies that each primary pump is directly connected in line with a chiller, wheras your diagram indicates that the outlet of the pumps is commoned up BEFORE the inlet to the chillers, not after which is implied by your description. Which is correct?. To achieve a constant flow, does flow bypass or go through each chiller regardless but is not being chilled? There is a discrepancy here which is also causing conflict in the control system design.

My motto: Learn something new every day

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

 

[BigInch]

Yes I was struggling with that issue too, but just assumed that the diagram was correct.

Independent events are seldomly independent.

 

[flexiblycool]

LittleInch & BigInch: I am impressed and grateful for the overwhelming and meaningful feedback. Let me get back to the drawing board with more details on the schematic for posting it on the Forum.
Meanwhile, Thanks again.

 

[LittleInch]

The other thing that needs to be considered / understood is how the system curve varies with flow rate. You state that each pump contributes 200 gpm. Unless there is a very high fixed element of friction losses (the AHU / FCU) then friction losses rise as a square fo the flow rate so adding more identical pumps in parallel doeasn't increase flow by the same amount each time and one pump can over speed unless controlled due to lack of resistance to a low flow rate. As your diagram has 100 psi output and then 50 psi at the furthest building, this implies that pipe friction losses are quite a large proportion of total losses and hence will affect the primary pump performence.

My motto: Learn something new every day

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

 

[flexiblycool]

LittleInch: Thanks much.
The schematics that was attached with the thread was to depict the system in a very general way. From the depth of analyses received from the Forum, I am convinced that the schematics must be revamped to reflect the installation with more detail and more accuracy and actual numbers. So please bear with me for the next little while till the Autocad specialist can churn out another sketch perhaps within a day or 2.

 

[BigInch]

We are thankful ... and waiting

Independent events are seldomly independent.

 

[flexiblycool]

Please refer to the attached sketch for following discussion.
1. The primary pumps rated for 200 GPM are probably downsized taking into account the diversity factor, and this might explain the flow of 280 GPM in the secondary pumps. However, the system might malfunction because secondary pumps are not VFD and therefore unable to give flows lesser than 280 GPM.
2. Due to the absence of the Hydraulic Decoupler, the system can hardly be called Primary-Secondary Pump System. The so called secondary pumps are barely boosting the pressure to take care of the pressure drop in the building they serve. (Would there be any advantage to retrofit a hydraulic decoupler given that the pumps are constant speed, and what would be its location.?).
3. Each chiller has 2 circuits, in fact two small chillers are packaged as one.
4. The primary pumps are not dedicated for a particular chiller. Any pump can come on to cater a chiller that cuts into service.
5. The Secondary pumps are NOT VFD and they are continuous duty. In other words there appears to be no advantage of providing these secondary pumps.
6. The bypass line shown on the chillers seems redundant, as it is equipped with a normally closed manually operated butterfly valve. If it is converted to a motorized valve that opened when a chiller cut out of service, the system would at least become a constant volume system, thereby not starving any fan coil units of design GPM. However, another simpler method would be to just let the water run through all the chillers regardless of whether they are ON or OFF.
7. Advice about how to retrofit the system to make it functional and energy efficient would be appreciated.

 

[BigInch]

Closed loop system could be a candidate for VFD, but with 3 pumps you can do almost as well just using on/off controls for all pumps. With 3 parallel pumps, on/off will normally get you reasonable economic coverage of flowrates from 30%-100% of the 820 max, but if you have no control valve on the primary pump discharge, you will be mostly hovering around either 280, 560, or 820 gpm flowrates. Using a control valve, or VFD would make it possible to directly target intermediate flowrates. The energy savings between CVs or VFDs may not be very significant and both of those options will come at a higher cost than operating the pumps in on/off mode at the "interger" flowrates, but you might want to do some calculations to prove to yourself that what I'm telling you is true.

Independent events are seldomly independent.

 

[LittleInch]

Flexibly cool,

Nice diagram, but doesn't really add much to the info you've already given I'm afraid and I'm not actually sure what your problem is any more. Does the system work, but perform badly?, Does it work at all? or are you just trying to figure out how it will work?

~anyway my comments to your points
1) There is a discrepancy between the primary and secondary pumps, but just because a pump says it's duty is 280, doesn't mean it will do that for a centrifugal pump. It could do anything from 75 to 350+ gpm depending on the head it sees.
2) What is a hydraulic de-coupler?. In my opinion you are better off raising the primary head to over 75 psig at the furthest building and introducing pressure controllers to each building and doing away with the secondary pumps as it is difficult to control and spec them properly, especially for a fixed speed unit
3)Ok - I assume the balancing valve is a manualy adjustable control valve aimed at eqaulising flow between the two chillers?
4)OK
5)The secondary pumps are there to maintain pressure in each building, but this is a poor way of doing it and looks a little to me like a fix to a problem which the original deisgner or operator found
6)I think you need to let the water flow through the chillers regardless of operating or not so that you can balance the system
7)To me you have a disconnect between the design of the buildings fan coil units which appears to use / be designed for a constant flow and constant pressure of chilled water and then bypasses units when cooling is not required and a primary flow system which varies flow depending on return water temperature and may then starve a building of chilled water which it needs. Therefore I think you need to remove the secondary pumps, upgrade the primary pumps in terms of head and fit pressure regualting valves at each building inlet so that each buildign gets the same water supply. Run the sytem continuosly at 600 gpm and flow through all the chillers, just turning them on as return temperature rises. You should be able to rotate the chillers so that they get an equal amount of run time.

Maybe not the most efficient, but the alternative is to change the fan coil units to isolation valves instead of bypass and then control water flow by turning on pumps or using a VFD to maintain a set pressure in the discharge leg of the primary pumps and again get rid of the secondary units by increasing the size of the primary pumps and using pressure controllers to regulate flow between the buildings evenly. This would leave the chiller controls as is, but means a lot of work to change all the bypass valves and revise your control logic.

I think you've got too much installed equipment which is ion conflict with each other to make it the most efficent system, but it will work a lot better....

My motto: Learn something new every day

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

 

[flexiblycool]

BigInch & LittleInch
Thanks to both of your gentlemen.
Best Regards
flexiblycool

 

[LittleInch]

No problem.

Did it help? Are you going to change anything?

We all like a bit of feedback about our comments...

My motto: Learn something new every day

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

 

[flexiblycool]

LittleInch: You have all the right to seek my feedback.
I found your feedback (and of THE other respected members of the Forum) most meaningful to help chalk the strategy which is as follows:
The secondary pumps are acting only as boosters and the terminology Secondary Pumps is superficial.
If the head of the Primary Pumps is sufficient 'as is', the Secondary Pumps can be removed immediately for a cost saving from the O&M point of view. In fact is the head of the primary pumps turns out to be sufficient, the money of the secondary pumps should in all fairness be refunded to the Owner. The pump head calculations must have been incorrect and these were not reviewed carefully by any party. The only justification that might be put forward in favor of secondary pumps might be that they will take care of increased system resistance due to clogged FCU strainers. This argument might not be tenable because routine FCU maintenance should take care of such contingencies and not the secondary pumps.
The Primary Pumps must never be shut down. This seems to be very rigorous for the primary pumps but fortunately there is a fourth pump that is used as standby so that could relieve one pump at a time.
The water must be allowed to circuit through the idle chiller to keep it a constant volume system.
The alternative recommendation would be to make it a true primary-secondary(vfd) system with 2 way valves on the FCUs.
Question: The proper location on the hydraulic decoupler with reference to the new sketch will be appreciated.
Thanks & Best Regards

 

[LittleInch]

I had to go look this up, but the link below confirmed what I thought. In order to run the primary circuit at a standard flow, you need to have some sort of variable flow control which can go from 0 to 100% of flow to create a primary loop. The true secondary pumps are normally VFD to provide a variable flow of water to your variable load. To make it work well yu really need to monitor flow coming out of the primary pumps and then adjust the de-coupler to maintain a fixed flow. As the secondary ciruits take more then the de-coupler needs to adjust to reduce flow and allow more to flow via the buildings assuming they work on a 2 way valve principle. This also only works when the primary flow is greater than all the secondary flows combined.

Therefore you need to create two loops either at the far end connecting the far building supply line to the return line or doing it at the start of the circuit. The way yours was drawn up was a straight through system with the "secondary" pumps acting like a booster pump.

Sometimes also seems to be a called a neutral bridge.

http://hpac.com/air-conditioning/improving-efficiency-variable-primary-flow

My motto: Learn something new every day

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