Chiller Optimization 2

[gopack85]

I'm having some issues with my chillers short cycling. We have four chillers with fixed flow primary pumps, variable flow secondary pumps, and an open bypass. When the secondary flow requirement increases beyond the primary pumps another chiller is staged on since we get reverse flow through the bypass = increasing supply temps.

After the next chiller is staged on, the building coils close and the secondary flow drops which stages down a chiller. This cycle repeats depending on our stage up/down delays (currently at 1 hour).

From what I've read, it sounds like I need to install a check/control valve on the bypass so it won't allow reverse flow and mix warm return water with cold supply. If I do this, how will I know when to stage another chiller on? It seem like the only indication would be if the coils in the buildings went to 100%. Any ideas?

 

[CarolinaPE]

I would check the installation schematic against ones put out as "how to" by ones like the following:

http://jmpco.com/page.cfm?sp=schematic&app=6

 

[tys90]

Is it possible your secondary pumps are reacting too fast to the change in pressure? This could cause it to overshoot the necessary flow required to satisfy coil demand and cause a chiller to come online too soon. Same thing would happen in the reverse.

Another option would be to look at 2 parameters before starting a second chiller. This may be calculated chiller load (fairly easy to get) or actual load if you can read it from the chiller controls. Don't allow a second chiller to run unless the first one is 95%+ loaded for "X" amount of minutes. There are any number of parameters you could read or calculate in addition to having reverse flow in the bypass. If both conditions aren't met, you don't start a second chiller.

 

[SAK9]

What are the chiller and field temperature differentials(Delta T)?Are they closely tracking design?
Do you have any coil valves at 100% before the additional chiller comes on?
The bypass is meant to act bidirectionally and putting a check valve beats that purpose.

 

[MTES93]

The old check valve topic.... Check out

http://wwww.kele.com

look under resources/industry-product reference for Gil Avery Articles. I believe he discusses this topic.

I've done it measuring system tonnage ~ flow x delta T and comparing to which and how many chillers you bring on while maintaining CWS set point. There are a number of ways I suppose. I think controlling on CWS temp alone can cause issues similar to what you describe.

 

[ChasBean1]

The primary/secondary setup you mention (now ‘old school’—the world has gone variable primary) is robust and will work. Do not change to new school because it’s not worth it. That’s a long story and another topic.

Forget about the flow through the bypass, or “decoupler” that you mention. Do NOT put a check in that line.

If you have a BACnet or LON interface, stage to the second, third, fourth chillers by each chiller’s percent load above ~95% for 15 minutes. Stage down one chiller at a time when each chillers’ percent load is less than 50%.

If you don’t have that interface, use a secondary loop supply temperature, return temperature, and flow meter. If the secondary loop supply water is 2-6°F (your choice, depending upon how critical dehumidification is) above set point for 15 minutes, stage on another. When the calculated tonnage is 90% or below the operating chillers' ratings, stage each chiller down after 15 minutes. For example: if there are three 500 ton chillers and three are running, stage back to two when total tonnage is below 900; stage back to one when total tonnage is below 450.

 

[cry22]

I do the same thing as Chassbean, may I add two cents:
1. Make sure you understand what Chassbean means: control your chiller with sensors/flow meter on the secondary loop, not on the primary loop.
2. I would also monitor chiller amp draw the same way as tonnage monitoring described by Chass.

I saw others who put a bidirectional flow meter on the decoupler for additional monitoring of direction flow through the decoupler

 

[gopack85]

Thanks for the input. Right now the chiller temp SP is 40F as due to incorrectly sized coils in one of the critical rooms, any temp higher than 42F to those coils will set off humidity alarms.

I do have them on a LON network with percent load. It's critical that a chiller stage up if the supply temp starts to rise which I monitor. I don't have a flow meter in the decoupler but I know the size of the primary pumps and how much total flow is in the system so I can calculate if its reverse flow.

The big question was when to stage down, and I think i'll experience with the total load being below 50%. Maybe add in some logic as well to look at positive flow through the decoupler.

 

[EnergyProfessional]

I'd doublecheck on that RH setpoint. Maybe someone arbitrarily set a low value for that alarm for not an actually good reason? that one zone driving your entire system may have to be eliminated (rogue zone elimination)

 

[BronYrAur]

I am fighting the same issue. Although I agree on paper with the ideas presented here, my problem rapidly escalates when my secondary flow exceeds primary. Once I become secondary-strong, the supply temperature raises quickly. My design temp is 36 deg, with a 20 deg delta. When I become secondary-strong, the temp can easily reach 40 deg. When that happens, control valves all start to open up more which raises my secondary flow even higher.

By the time a chiller comes online, I may still be secondary-strong, which tries to bring on another chiller. Then after a short time, I find myself too primary-strong which often knocks out a chiller (or chillers) on low temp. This causes a secondary-strong condition, and the whole vicious cycle repeats itself.

I don't mean to redirect this thread about my situation but it seems very similar to gopack85's. I have constant volume primary depending on how many pumps (and chillers) are running, and variable secondary controlled by delta-P. As long as I am primary-strong ... without being too primary-strong ..... everything runs well. But when something happens to cause a secondary-strong condition, it quite often takes several hours for the plant to settle back down.

 

[LowDeltaTSolver]

See AHSRAE Journal October 2014 Article titled "Simplified Chiller Sequencing."

Link

The article describes a chiller sequencing approach for a primary/secondary chilled water system that is being used for a university campus chilled water distribution system.

The control approach minimizes flow in the decoupler piping and describes a chiller sequencing control approach.

What are the cooling coil entering design chilled water temperatures?

What is the distribution chilled water supply temperature when the next chiller is started?

What is the chilled water control valve control sequence of operation when the space relative humidity sensor is above the space high humidity set point?

 

[CelsoIII]

I agree with the comments regarding staging your chillers based on the percent load of each chiller. I would make the trigger point to stage up or down match the chillers performance curve. Example: if your chillers sweet spot is between 60% and 80% load, I would program the next chiller to come on line if the first chiller had been running at 85% load for x amount of time. Vise versa, if you have two or more chillers running, they should stage down if the last chiller brought on line is loaded to 60% or less for x amount of time.
To further reduce electrical operating costs I would look at trending outside air temperature for the day and make sure to stage up a chiller before entering the high peak demand period for the day based on outside temperature forecast. Staring a chiller (without VFD) will create a KW demand spike that may cost hundreds of dollars compared to staring the same chiller one or two hours earlier.
In order to improve chiller staging in addition to considering time delays, the system automation should consider CHW flow rates, chiller loads, and supply return temperatures at a minimum. In some cases even cooling valve status (avg % open or closed) could be used also to make the decision to stage up or down.