Chiller ROI - HELP!

10 posts / 0 new
Last post

I've been attempting to calculate a simple ROI for a new 250 ton chiller
our client wants to install to replace his older chiller. I ran the
simulation for the default "high rise hotel" construction which, to my
luck, is quite similar to the actual building. The detailed results,
however, are not making sense to me. Specifically, when I look at
Equipment Loads and Energy Use for the Chiller, the Elec Use and Cool
Load don't coincide given the default Chiller efficiency of .67kw/ton.
For example they give a Chiller Elec Use of 535453 KWH and a Cool Load
of 3294 MBTU. However, according to my calculations, at .67KW/TON,
535453 KWH should give me closer to 9590 MBTU. I see most of the hours
are in part load range, but does the efficiency drop off that
dramatically at part load that it would explain the discrepancy? Which
numbers should I base my analysis on, the Cool Load number or the Elec
Use number.

Also, on the same page (Equipment Loads and Energy Usage), it gives me a
peak load of 2436 KBTU/HR. However, when I go to the section titled
"Building Peak Load Components" the peak cooling load is significantly
less (1619 KBTU/HR)

At this point, I have no idea what any number represents, since I don't
know how it was calculated.

I greatly appreciate any help! Thanks.

Respectfully,

Brett Sanicola

Brett Sanicola's picture
Offline
Joined: 2011-09-30
Reputation: 0

KW/ton drops off with decreasing part load, and since the chiller isn't
always operating at ARI full load conditions, you can't just divide
total chiller energy consumption (in KWH) by the total cooling load (in
btu) and hope to get anywhere near the full load rated KW/ton.

As you go about this ROI study, make sure that when you are analyzing
your proposed chiller replacement, that you create a custom EIR vs PLR
curve based on manufacturer's data, otherwise you're using a default
DOE-2 curve that isn't necessarily correct.

Furthermore, to make the study worthwhile, the existing chiller model
really needs to be normalized as best as possible to actual building
operating conditions. You can get away without doing it for a simple
comparison, but it would be more accurate if you normalized to utility
bills.

James Hansen, PE, LEED AP

James Hansen's picture
Offline
Joined: 2011-09-30
Reputation: 200

Thanks! I was starting to conclude that as well, but then I spoke to
Carrier, and they said that there is no way that it would drop off that
significantly. For example, they said they have a unit that is full
load rated at .6KW/ton and .7KW/ton down to 25% part load. The numbers
I'm getting is nearing 2KW/ton, even though the full load rating remains
at .67KW/ton. I did mention that the vast majority of my hours are
operating below the 10 percent part load, but he still said it wouldn't
be that high. Any thoughts?

Respectfully,

Brett Sanicola

Brett Sanicola's picture
Offline
Joined: 2011-09-30
Reputation: 0

Brett,

to get better part load data and efficiency of your proposed chiller
replacement, call your chill sales engineer and ask them to generate a part
load performance report for you in 10% unloading increments, then you can
see the decrease in performance efficiency for each part-load step of the
chiller. Also be sure that you recognize if you are looking at RATED
Conditions (which means chiller performance at ARI standard conditions), or
at DESIGN conditions (which are your project's specified operating
conditions--this would account for part load efficiencies if you were
running a 12-or 14 Degree delta-T across your chiller.)

Before you set up a new performance curve you will need to get this data
supplied from the manufacturer.

Pasha

Pasha Korber-Gonzalez's picture
Joined: 2011-09-30
Reputation: 600

IPLV and NPLV values are not perfect by any means but I think you should get comfortable with what they represent before you try to use a detailed energy model to overcome their short comings.

Paul Riemer's picture
Offline
Joined: 2011-09-30
Reputation: 0

I reread your email and something you said struck a chord. I'm assuming
the default curve, which I'm currently using, falls off drastically at
part load. Hence the reason for my results. Am I assuming correct?
Also, how would I change the curve? Can I do that in the wizard or will
I need to access the detailed interface?

Respectfully,

Brett Sanicola

Brett Sanicola's picture
Offline
Joined: 2011-09-30
Reputation: 0

Brett, attached are some example kw/ton curves for existing chillers on a
recent project. Chiller #1 is a 250 ton unit, and chillers 2 thru 4 are 550
ton units. Chillers 1 and 2 are ~ 15 years old, and chillers 3 and 4 are
newer with better efficiencies. All 4 chillers are water cooled. The
curves show efficiencies with and without Adjustable Frequency Drives. It's
hard to tell the difference between the "w/AFD" and "w/o AFD" curves, but
the "w/AFD" curves have a lower kw/ton except at full load.

Jim Marsh, PE, LEED AP

Jim Marsh's picture
Offline
Joined: 2011-09-30
Reputation: 0

Thank you all. I see that the below 20% load, where I was predominantly
operating, the KW/Ton shoots up drastically. Thanks, that explains my
results.

Respectfully,

Brett Sanicola

Brett Sanicola's picture
Offline
Joined: 2011-09-30
Reputation: 0

Here's a custom curve from a McQuay frictionless chiller that I just
used on a project:

TONS

PLR

KW/TON

250

1

0.64

225

0.9

0.56

200

0.8

0.509

175

0.7

0.451

150

0.6

0.4

125

0.5

0.346

100

0.4

0.368

75

0.3

0.413

50

0.2

0.495

25

0.1

0.697

Like you said, when you start to get into the 10%-20% ranges, KW/ton
starts to suffer, and this is a VFD chiller with magnetic bearings, so
it's better than most.

James Hansen, PE, LEED AP

James Hansen's picture
Offline
Joined: 2011-09-30
Reputation: 200

Is this at a constant condenser water temperature, or is there condenser water relief?

John Aulbach's picture
Offline
Joined: 2011-09-30
Reputation: 1