Chiller Curves

8 posts / 0 new
Last post

I am attempting to construct custom chiller curves, and even though it seems
simple enough to accomplish I have run into a snag. The manufacturer will
only supply me with 4 points and not 6 points, like equest likes. Any
suggestions beyond complaining to the manufacturer some more?

thanks,

--
Rob Hudson

rdh4176's picture
Offline
Joined: 2011-02-14
Reputation: 0

Hi Rob,

Have you looked on their website to see if they have any technical
specifications posted? Many manufacturers do, you just have to look around a
bit. Otherwise if you have any information for a similar chiller that you
could use to see how it unloads.

Best,

Carol

cmg750's picture
Offline
Joined: 2010-10-05
Reputation: 0

I have a new Chiller Curve question/situation. I have created a custom
chiller curve for the EIR = f(PLR) as a cubic function using the following
data inputs:

1 .3313
.75 .2437
.5 .2201
.25 .1928

create a hourly report summary for the Corrected Chiller EIR and compare it
to the default EIR equation for this entry, i get 0.034 for the default and
0.072 for my curve at 25% ratio. Why is it that the default is better than
my curve? I have looked at the equation for the default entry and got the
following equations:

EIR = a + b x = c x^2
a = .04812248
b = .69573420
c = .23493889

at 25% this equation results in 0.2367, and my equation gets 0.1928

Is there any other calculations going on in the background that is causing
some problems? I have attached an excel snapshot of the different EIR and
the corrected EIR values for two different chillers in my program. Neither
is currently using the default EIR = f(PLR) curves, they are using custom
curves.

Thanks in advance,

--
Rob Hudson

rdh4176's picture
Offline
Joined: 2011-02-14
Reputation: 0

Hi Rob!

Well to start, let's not assume every chiller out there MUST be more
efficient than the default curve at every point. Better to assume they
can and should differ.

With that disclaimer out of the way, you have noted your coefficients
generated for your custom PLR curve should produce a lower correction
factor at 25% (confirmed in the Var 9 column you attached).

What I think you're missing is that there is indeed more than one
correction factor being calculated every hour, and more than one curve
in play to produce those factors that you need to consider. The extra
column you chose to include in your spreadsheet hints at this (Var 10).

Recommended DOE2 reading that should fill in these specific gaps for
you: Volume 2: Dictionary > HVAC Components > CHILLER > Chiller Energy
Consumption

There much further discussion on this topic in the archives as well, if
you just can't get enough ;).

~Nick

NICK CATON, P.E.

Nick-Caton's picture
Offline
Joined: 2011-09-30
Reputation: 805

Prior to your response, i got a little impatient and found some of your past
articles in the archives after a lot of digging.

I found out that the EIR equation is actually a % of the EIR that needs to
be taken into account. So the equation should output a value from 1 to 0 to
multiply the design rated EIR by to get the usage at that part load. there
was a method that i found that, when the manufacturer doesn't provide you
with enough to populate all three of the custom curves, you can simply use
the EIR f(PLR) curve, which i did and got the results i expected to see.

Thanks for your response, both past and present.

Rob

rdh4176's picture
Offline
Joined: 2011-02-14
Reputation: 0

Awesome!

I'm always happy to hear my contributions help out =). My first thought
was that you might not have picked up on the "normalized" aspect of the
factors (1 to 0 as you're saying), but I couldn't be sure. I suggested
the DOE2 help entry below because it covers that point among other
things.

Glad to hear the ship is sailing smoothly again...

~Nick

NICK CATON, P.E.

Nick-Caton's picture
Offline
Joined: 2011-09-30
Reputation: 805

Hi Rob,

It looks like you all ready solved your problem, which is good, but I'm
going to send you a comment anyway.

When I create a curve, I always use "raw data points" instead of
"coefficients". The reason I do is that I can always obtain them and I don't
have to worry about making a calculation mistake because eQUEST takes the
raw data points and calculates the coefficients for me. If I want to see
what they are, I can look in the .bdl file.

The only trick to raw data points is that for everything except
refrigeration compressors the dependent, in this case PLR, needs to be
normalized to either the ARI design, or the actual design if it is
different.

So for your chiller curve you would input: Independent Dependent

1.0 1.0

0.75 0.7356

0.50 0.6643

0.25 0.582

Anyway, that's my 2 cents.

Carol

cmg750's picture
Offline
Joined: 2010-10-05
Reputation: 0

Hi Rob,

Part 2 of the Energy Design Guide resources program on HVAC Simulation goes
into detail on how to create custom chiller curves. The guide can be found
here:
http://www.energydesignresources.com/media/2654/EDR_DesignGuidelines_%20HVAC_Simulation.pdf

The thing that trips a lot of people is that EIRFPLR is actually a ratio of
power draw, not efficiency.

For example:
EIRFPLR (50% PLR) = kW (50% PLR) / kW (full-load)

*Robby Oylear, LEED**?** AP BD+C*

Robby Oylear's picture
Offline
Joined: 2011-09-30
Reputation: 202