VFD on Existing System 2

[ChrisConley]

Hello All,

I'm looking into the possibility of a variable primary flow chilled water system. I've heard all the warnings about minimum flow, but the client would still like to investigate. The system is a retrofit ~750tons, water-cooled chiller. We are replacing the existing chiller, cooling tower, primary pumps and condenser pumps.

My issue is that the secondary loops that feed off the primary loop are all pump driven coils with three-way mixing valves for capacity control.

What I would like to do is control the primary pump to run according to the number of secondary pumps that are operating.

Right now the existing system circulates 2400 gpm if even one small 80 gpm coil is calling for cooling. Please no comments on the flows as the original system was designed for a much larger load – a load they have yet to achieve in 40 years of operation. We will be reducing the primary flow to match the load profile and new chiller.

Has anyone ever controlled a pump VFD like this? Also, does anyone have any warnings about variable flow through the chiller beyond the standard? We are also considering variable speed on the chiller and/or the cooling tower.

 

[quark]

Picked up this good link from other threads,

http://www.automatedbuildings.com/news/mar02/art/hrtmn/hrtmn.htm

Also check the other papers from the same author. The issue is extensively discussed with pros and cons.

I have one question(other than flowrates). Why can't you have variable speed secondary pumping system? A 3 way valve can be converted into a 2 way valve by closing the bypass port(and I did it in the past)

What I would like to do is control the primary pump to run according to the number of secondary pumps that are operating.

If it is a 3 way valve system(constant flow secondary pumping) how are you stopping any of the secondary pumps? Is it when you shut off the AHU?

Regards,

 

[Drazen]

I have installed VFD run by flow sensor as mandatory demand from client. 2 months after startup, we added new rule: when temperature in secondary falls in, say maximum acceptable range of 4-8 degrees C, VFD swithces to minimum constant flow, which is larger then calculated varible flow for that range (flow which would be adjusted by VFD without that additional rule). That way we stopped pump failures (3 in that two months) caused by endless variating of flow.

 

[ChrisConley]

Thanks for the excellent link.

The pumps are being controlled to run on a schedule, temperature is then controled through the three way valve. The schedule varies enough that the system rarely sees 100% flow, in fact usually sees 55-80% flow.

The reason I haven't been considering variable speed secondary flow is that there are 15-20 secondary pumps remotely located, all relatively new. The scope of the project already has us replacing the primary pumps so it would be 'easy' to incorporate variable primary flow.

 

[quark]

During low load conditions primary chilled water see low temperature difference and with flow decrement, the temperature difference across chiller will be lower. Variable speed chiller is a good option in this case. dp sensors react faster than flowsensors though the control band is a bit higher.

I have yet to find out how a flow sensors controls the vfd when higher flowrates are required. Can it be done without the use of temperature control?

 

[SAK9]

Most chiller suppliers do not recommnd flow to go down less than 85 % of design .You can go lower only if your client is willing to take the risk.I have heard of chiller tube bursts due to freezing( no direct experience though).

Theoretically when you reduce the flow,the chiller capacity control should come into play and reduce the refrigerant flow .The chiller capacity control needs to respond as fast as the flow reduction to avoid freezing in the tubes.This is not practical as the chilled water leaving temperature must drop first for the capacity control to get into action .So there is a time lag between flow reduction and capacity control.As I understand, the rate of water temperature drop is faster than the capacity control.This will lead to the chiller cutting off on the antifreeze thermostat(if it works!).No wonder then why manufacturers would not want you to vary the flow!

I certainly would not recommend fitting the the cooling tower fan with a VFD.In one installation I tried it,I could not the frequency down more than 2 hz the reason being the propeller fan capacity sharply drops off with any reduction in speed.But other things such as local weather etc may help you achieve better results.The best approach would be to try it on one tower and see what results you get.Most likely again you may not get much help from the manufacturer!

 

[ChasBean1]

Chris,

If you want to control the primary pump based on the number of secondary pumps, why not just remove the primary pump?

Maybe I'm not understanding the objective well enough, but if you want to control primary plant flow to optimize energy use without taking into account the critical parameter (secondary supply temperature) something is amiss...

CB

 

[sterl]

1) Make usre you know aht the Chiller Manufacturer would consider a Minimum Flow through the chiller. If the required head of the primary pump for full load is only marginally larger than the chiller Delta-P at that minimu flow, you are largely wasting your time.

2) Put all the secondary pumps on the same Power Circuit with a CT. Run your Primary Pump at a fast enough speed to maintain the flow above, and then devise your P & ID loop to increase that speed according to whatever the current sum for the running secondary pumps might be.

3) you can also approach the topic with a 3-way valve on the chiller to short loop the water and keep the chiller flow above the minimu, then speed up you pump according to the Delta-T across an operating chiller.

In the overall analysis it will come down to How near perpendicular you system curve will be at reduced total secondary flow to the (BEP-at reduced speed) Curve for the pump...Otherwise you will keep pushing the pump into zones where it really doesn't wwant to be.

 

[imok2]

The traditional arguments for desiring constant flow through
chiller evaporators no longer carry much weight; most modern
microprocessor-based chiller controls can effectively deal with upsets due to variable flow. Moreover,constant-flow primary designs cannot respond to the need to put more CHW through chillers in the event that the distribution
system returns low CHW ÆT to the central plant.
A variable-flow design with pumps either oversized and controlled by VSDs or banked can respond
to low ÆT central plant syndrome. Thus, for the same
reason that we as HVAC designers provide freezestats upstream of cooling coils, nonoverloading
motors to drive pumps and fans, and tube pull space at chillers, boilers, and air-handling units,
we need to design chilled water plants that can anticipate the possibility of low CHW ÆT and
respond to it. Therefore, The author believes
it’s time to put primary-secondary pumping back into our
tool bag of applications to address specific design situations and adopt a new paradigm for
chilled water system design. HPAC
For a full read go here:

http://www.kirsner.org/kce/media/pdfs/PriSecChilWatr.pdf