Fan energy fundamentals

5 posts / 0 new
Last post

Paul & others:

I just wanted to say off the bat: the EDR resource Paul linked us looks
to address the heart of my query directly (and thank you so much!), but
it's difficult for me to summarize further right now...

I think I'll need to re-read the appropriate parts of this document a
few times over and arrive at a deeper understanding of actual
methods/sequencing strategies for static pressure reset before I come to
any definite conclusions regarding VAV fan curve.

The query I'm trying to nail down is what assumptions and specific
system behaviors the DOE2 library/default curves and particularly the
prescribed 90.1-2007 baseline curve represent, and from there answer
what's fair game insofar as defining/modifying a different sort of
EIR-FPLR curve for a proposed model. I have the understanding that the
associated fan energy savings between various approaches are potentially
big and real, but I'm scratching the surface of how to actually model it
to a realistic degree.

If anyone has pondered the same topic or is a few strides ahead of me
and wishes to share any thoughts/guidance, please do not hesitate =).

~Nick

NICK CATON, P.E.

Nick-Caton's picture
Offline
Joined: 2011-09-30
Reputation: 805
Nick's concern about VAV fan curves is not trivial. I am working on a LEED model using Baseline System 5 (PVAV). The Baseline fan energy using the 90.1 G3.1.3.15 VAV FPLR curve is 50.8% more than the fan energy using the default eQUEST "Variable Speed Drive FPLR" curve. YMMV of course but that is a big difference. This penalizes the Baseline for my project - the total electrical use increases 8.9% in the Baseline vs. only a 4.1% increase in the Proposed model when switching to the same G3.1.3.15 curve. Bill
Bishop, Bill2's picture
Offline
Joined: 2011-09-30
Reputation: 0

I agree it is not trivial but I am really surprised by that 50.8% number. I think the curves are relatively similar at the high end and most systems rarely spend time on the low end where the curves do differ considerably. Is there a big difference in the system volumes, specifically are the baseline fans bigger in volume and thus always operating at a lower ratio than the proposed?

Generally, I assume that the curve in 90.1 was supposed to represent a VAV VFD curve and that my VAV VFD fans should not be penalized against it. VAV vs. CAV is trickier and brings the temptation to switch your rationale based on which side your project falls on.

Paul Riemer, PE, LEED AP BD+C

Paul Riemer's picture
Offline
Joined: 2011-09-30
Reputation: 0
Paul, The 50.8% difference is Baseline w/default eQUEST VAV curve vs. Baseline w/ASHRAE 90.1 curve. An hourly report for one of the systems shows it averages around 0.50 or 0.55 PLR. Using the curve coefficients shown (below) for the two curves, power at 0.5 PLR is .196 full load power for the eQUEST curve and 0.373 full load power for the ASHRAE curve. That's a 90% difference right there. Bill
Bishop, Bill2's picture
Offline
Joined: 2011-09-30
Reputation: 0

I've also been surprised at the magnitudes involved, but these kinds of
savings do seem very real. One fundamental I had to wrap my head around
(confirmed with some controls servicing contacts) is that large supply
fans with VSD's should rarely operate at or near full capacity when a
system is being controlled correctly.

Small differences as the curves approach zero really add up over 8760
hours however. The EDR resource linked below is really a better teacher
than myself, but if you'll compare the curves I posted below, you'll
notice they're actually very different at most PLR's. Actually I can
overlay these pretty quickly with my personal curve spreadsheet:

When looking at these: consider the same system operating at any given
PLR: The difference may seem small, but at 20% PLR the system with the
library curve is drawing double the energy of the 90.1 curve. Similar
story but inversed around the 40% mark. Bill's model producing a ~50%
difference in fan energy consumption between the curves is not outside
the realm of reason considering this -his hourly PLR distribution
through the year and minimum flow ratios would factor in, of course.

I more strongly do not believe these two particular curves are
apples-to-apples however. My present understanding is the library VSD
curve (the "Nike swoosh") represents a VSD fan system without static
pressure reset (as prescribed by 90.1). The other curve shape (umm...
"upturned banana?") does look like fan system curves incorporating a
static reset sequence, but it's still weird that it actually approaches
zero at zero (well technically 0.0013)... .

This is something I haven't reconciled with reality just yet... One
thought is it would appear to represent a perfectly efficient duct
system with zero static pressure to overcome at low flows - perhaps that
means for a level playing field we're supposed to assume the same for a
proposed curve? Is there language in 90.1/users' manual that explicitly
says something along the lines of "duct/pipe distribution losses are not
supposed to be modeled?"

Still mulling this over... =)

~Nick

NICK CATON, P.E.

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