I have a question regarding 90.1 2004 Appendix G baseline chilled water pump
control. Section 3.1.3.10 says that baseline CHW systems must be
primary/secondary, with a constant flow primary loop and a variable flow
secondary loop. The secondary pump will either ride the curve or be
controlled by a VSD depending on square footage. In the case I currently
have in my hands, the pump rides the curve. No questions yet.

Section 3.1.3.11 goes on to state that each chiller shall have its own
condenser water and chilled water pumps, interlocked to operate with the
chiller. This is clear and reasonable for the condenser water and primary
chilled water pumps, but what about secondary chilled water pumps? For my
current application (one chiller, pump riding curve), my question simply
boils down to whether I should turn the secondary pump off when the chiller
is off, or if I should run it all of the time.

But to get into it a little more, here are three approaches I see:

1. Model one secondary pump for each chiller, and turn each pump off
when its associated chiller turns off. Section 3.1.3.10 is titled Chilled
Water Pumps, and that section includes both primary and secondary pumps, so
when the next section references "chilled water pumps" that should include
both primary and secondary. However, this does not reflect the typical
reality of one secondary pump serving multiple chillers, and the common
interpretation that only a primary pump has one single associated chiller.

2. Model one secondary pump regardless of how many chillers there are,
and turn it off when all chillers are off. This is perhaps most similar to
what you might expect in a 'real' system. If you only have one chiller (as
I currently have), this is exactly the same as approach 1.

3. Model one secondary pump regardless of how many chillers there are,
and never turn it off. Many buildings are unfortunately operated this way.
This approach significantly increases baseline energy when the pump is
riding the curve.

I have always used approach 2 without putting much thought into it. But now
my results are highly dependent on which approach I take, and I see enough
ambiguity that arguments could be made for approach 3.

Your thoughts?

Thanks,

Christian Kaltreider, LEED AP