Troubleshooting HVAC Systems: The Domino Effect of Problems

In a previous post, I mentioned that we had been having nuisance
low refrigerant temperature safety trips
in the chillers
serving a new library on a new construction project project where I
have been working with the commissioning team to bring the building
and its systems on line.  At the time, I was looking at the
possibility of one of the refrigerant circuits perhaps needing to
have its charge level “tweaked”. The chilled water system in
question was a primary/secondary, variable flow type system. 
Under peak conditions, the system supply temperature needed to be
39°F to control relative humidity in critical areas in the
library.  This supply temperature was only a few degrees above
the 36°F setting associated with the chiller’s freeze
protection safeties and thus, the system is very sensitive to
fluctuations in leaving water temperature when running at the
design supply temperature.  

In this post, I’m going to experiment with using a video capture
of my computer screen to convey a troubleshooting sequence in
addition to actually discussing the topic noted in the subject
line. Recently, I have been experimenting with
Camtasia Studio 5
a software package that allows
you to capture your computer screen along with an audio stream to a
video file for replay in a number of common formats. It seems like
it could be a very good way to capture relatively complicated
technical content and make it available to others versus writing an
informal report or a memo with a zillion embedded pictures. So,
lately, I have been experimenting with it for just that purpose and
will include a video clip later in this post to explain the
troubleshooting sequence alluded to in the title line.

In any case, after tweaking and adjusting on a number of fronts,
things seemed to be going along fairly well with the chilled 
water system unless we tried to force the system to run under an
extremely low load condition. In fact the discussion had turned
from a focusing on issues with the chilled water system to focusing
on whether the test we were performing really represented a
condition that would actually occur vs. a condition that was
technically possible by, in the real operating environment
unlikely. (I’ll probably have more to say about that discussion in
a subsequent post). After all, did it really matter if the system
exhibited operating difficulties when running against a load
condition that would never realistically exist?

That entire conversation went out the window one week when we
experienced three trips under what appeared to be relatively normal
operation with a significant load in the context of the system in

Robert Schram, the lead control system programmer on the project
began to explore what was going on during the trip that occurred
toward the end of the week and discovered a software bug in the
algorithms associated with AH-C which appeared to be a likely
culprit. Specifically, 

Robert discovered that a subtle nuance of the
software was causing the the control sequence to shift into and out
of economizer mode quite rapidly under a certain operating
. When this happened, the chilled water
valve for AHU-C would stroke through 70% or more of its normal
actuating range in less than a minute, which in turn, caused a
major variation in the distribution system flow in the same time

The variation in distribution system flow caused the
a major change in the primary chilled water loop
(the loop that circulated flow through
the chiller evaporators) in the same time frame and the chiller
capacity control system simply could not respond quickly enough,
resulting in a chiller safety trip. Robert modified the software to
eliminate the problem, but what was bothering me was the fact that
when the first trip occurred, the valve serving AH-C appeared to be
very stable up until the point at which the chiller tripped.

So, while I was pretty sure Robert had found and
corrected a significant problem and a likely cause of our nuisance
trips, I was not sure that we had found all of the
causes. So, I started to look more closely at what happened during
the first trip that occurred that week. This video is what I sent
to the commissioning team to walk them through what I discovered
and also suggest a test that would verify the hypothesis that I
developed as the result of my investigation.

The bottom line is that I was convinced that a minor (4-5% of
span) persistent hunt in the AH-B chilled water valve control loop
was causing a significant (20-25%) variation in flow on the
distribution system that ultimately rippled out to cause a chiller
trip on occasion. To test the hypothesis, I suggested that we
simply take control of the AH-B chilled water valve and lock it
into a fixed position and see if the problem went away. The result
of the experiment is illustrated below.

As you can see, (in, what I might add is a very gratifying
affirmation of physics) when Gary Kawabuchi, one of our field
engineers, took manual control of the valve and caused it to hold a
fixed position (red arrow in the picture above), the instability
that had existed in flow and temperature went away. Upon seeing
this, Robert dug into the control software for AH-B and discovered
that the the dew point control loop and the discharge temperature
control loop associated with the system were “fighting” with each
other, causing the valve to hunt and triggering the flow
instability in the distribution system that rippled out to
temperature instability in the primary (evaporator) flow loop and
occasionally caused a chiller trip.

Having made adjustments in both the AH-C and AH-B control processes
associated with their respective chilled water control valves, we
have now run several weeks with out a nuisance chiller trip
(knock-on-wood, or perhaps more appropriately for this issue,
knock-on-black-iron-pipe). So, a couple of bottom lines.

The HVAC systems we deal with in our buildings are
highly integrated and interactive.
A small issue in
one system can ripple out and trigger a problem in a different
system, often cascading through multiple control loops and/or
physical loops in the process. For me personally, this is part of
the intrigue and fascination associated with working with building
systems. If you found the problem discussed in this post
intriguing, then you may have some commissioning provider lurking
around in your genes and you should explore that; the industry is
growing and could probably use you.

When confronted with multiple oscillating parameters
that seem to be inter-related
forcing one of the
parameters to hold steady and observing the result can be a
valuable trouble shooting trick. Bear in mind this must be applied
with caution. For instance, locking the preheat valve down to fully
closed on a make up air system when the outdoor temperatures are
below freezing may or may not identify instability in its control
loop as the cause of instability elsewhere in the system. But it
just about guarantees that you will freeze the coil, an event that
will eclipse any success you may have in diagnosing the instability
you had a targeted with your test.

Solving problems that can ripple across multiple
will often require the cooperation of parties
from multiple disciplines/subcontractors/trades to resolve. This
can be a difficult pill to swallow at the end of a new construction
project when most of the folks on the construction team would just
as soon move on to the next project while the folks on the
operating team inheriting the facility are a bit intimidated by the
prospect of dealing with a building they are unfamiliar with that
is exhibiting a funky, intermittent operating problem. Having a
subcontractor who, like Robert, is willing to jump in and resolve
the problem when presented with some data pointing them in the
right direction can be a real benefit for all.

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s