Dynamic Electric Motor Testing
of DC Motors by S. J. (Stu)
Mugford, B.Sc. (EE); P.Eng. and D. (Dumitru) Sarpe, B.Sc.,
Kadon Electro Mechanical
Services Ltd., Calgary, Alberta
www.kadon.ca
Presented at
PdM-2006
by: Noah Bethel, PdMA Corporation
Before we
jump into software data analysis specifics and test result data,
we felt it appropriate to first present a brief primer on DC
motors, so that we can better appreciate the data we are about
to look at.
After all,
what good is data if it does not make sense? Furthermore, we
have maintained for many years, that simply going out and
collecting data in the field, then coming back to the office to
interpret the data, is a flawed process. This is an important
difference between test instrumentation that efficiently, but
blindly collects and instrumentation that also efficiently
collects, but lets the technician see and evaluate the data, as
it is being collected, and enter notes, conclusions, and flags.
This is the argument for instrumentation that permits in-depth
troubleshooting if it is required. Let’s finish the job, not
leave it partly done!
We strongly
believe that we should be interpreting the data as we collect
it, this has so many advantages. If we see some data that is not
right, we are in the ideal position to follow up immediately
while we are in the plant, connected to the circuit. Follow up
action might mean taking additional, more specialized
measurements. It could mean simply entering the appropriate
condition code into the software, so that the motor and circuit
are flagged for extra attention. It might mean that the
motor/circuit are removed from service right now, for detailed
dismantling, inspection, and overhaul and repair if needed. It
might mean the motor/circuit is flagged for detailed
inspection/overhaul on the next shutdown. It might even mean
that the data is wrong, and should be checked and retaken
immediately. In our business there are few things worse than bad
data. Suspect data should be discussed within the group involved
(customer and vendor), so we can all learn from it. Certainly,
thinking about and analyzing data as it is collected makes the
job so much more interesting and challenging.
What is the
alternative? Collect the data blindly and quickly. Store it away
in the computer until you get time to look at it. The worst case
is when a machine fails, and then we look back at the data and
realize we could have caught the problem before it turned into a
blow; if we had looked and if we had understood what we were
looking at.
Download the entire
paper (1 meg PDF)
|