Hi All,

I have a couple of questions with respect to modeling specific chillers.

1) I am trying to model a centrifugal chiller with an economizer. The

chiller in question is a centrifugal chiller running on R123 with an

economizer installed to it. Do I just need to get data from the chiller

manufacturer and input the curves into the model? Is this the right way to

go?

2) How do you model multi stage centrifugal chillers without VFD's?

Thanks,

Sunil Nair

When the chiller runs on economizer it will use a separate set of curves and be modeled as a distinct chiller.

One set of curves will be difficult to model both modes of operation.

DSE Mobile

Sunil,

My understanding of an economizer for a multi-stage centrifugal is that some of the liquid refrigerant flashes as it passes through the expansion device. This flash gas, I believe is introduced to the input side of the second stage impeller. Since the economizer is internal to the chiller, one set of curves should be sufficient as there is only one mode of operation as I understand it. You didn't specify, so I'm assuming you are not referring to a water-side economizer.

I haven't custom defined a chiller yet in eQUEST, however, it appears that several performance data points from the manufacturer are required for eQUEST to generate the curves if you select this method. The other is to generate your own curves. Many years ago when I used to work in VisualDOE, I found the calculation methodology to define chiller curves from the Pacific Gas & Electric web site, which I used to set up a spreadsheet to calculate the curve coefficients. This was originally set up for constant speed chillers. A number of months ago I added a spreadsheet tab to handle variable speed chillers. To create the 3 sets of curve coefficients, at least 10 full load data points at different chilled and condenser water temperatures were required from the manufacturer in order to map the compressor. There was no max or min limit for the part-load data points as I recall.

If you'd like a copy of the spreadsheet, I can email it to you, however, you might want to check the eQUEST manual or DOE-2.2 reference manuals regarding directly entering data points into eQUEST and let the program figure out the curves for you.

My best,

Mike Busman

Dave and Mike,

Thanks for your inputs.

Mike,

Yes, I am referring to the multistage chiller where the liquid is flashed

to the inlet of the second stage impeller. As you have mentioned this

process is internal to the chiller.

My thought initially was I may have to model multiple compressors per

circuit for a chiller without VFD. I did not find this option in eQUEST. As

I think about it now, my case would be quite different from having multiple

compressors per circuit.

I am in the process of obtaining data points for the chiller from the

manufacturer so I can input the data points into eQUEST.

I'll keep this thread updated.

Sincerely,

Sunil Nair

You should be able to look at the manufacturer data plotted vs ECWT and load and determine if a single set of curves (one modeled chiller) or two are needed to accurately model power usage by the chiller over the range of load. (if you can get the data)

David S. Eldridge, Jr., P.E., LEED AP BD+C, BEMP, BEAP, HBDP