|
|
|
Aluminum Mill Stand Monitoring Application
|
|
An
Aluminum Company in the southern U.S. was looking for a way to
detect and give warning of conditions which would lead to mill
stand
resonance.
Maars Inc. monitoring equipment was installed and the
results are the following
|
Trend of four
Resonance events
As a test
the mill operator ran five junk rolls through the mill stand ,
each to breakage of the roll.
Aluminum went everywhere on the breaks, and cleanup of
the mill after a break required about 30 minutes to 1 hour of
downtime for the mill. This slide shows data taken one of the
junk rolls going through the mill.
The operator ran the mill into resonance four times, as
you can see with the trend shown above. On the fourth time
into the resonance, the aluminum broke. Before breaking, the
aluminum in the mill stand would vibrate up and down as the
aluminum was accelerated and decelerated between the mill
stands rolls. The last trend peak is wider than the others
because the roll broke, and the mill was instantaneously
unloaded, causing a speed up on the rolls and then a reduction
in speed as the operator brought down the mill.
|
 |
Trend of four
Resonance events pasted to Excel
Here we show the data copied into Excel and
graphed for the trend. This is an example of how the
Pathfinder software is compatible with the Microsoft Office
suite of products. Here you can see each of the
chatter/resonance events in detail.
Here is the
trend and FFT of the resonance event as shown in PathFinder
software.
|
|
Using
Scenario Trend to determine frequency range for Trend.
This
slide shows the “Scenario Trending” function of the
PathFinder software, and this function was the key to
unlocking the mystery. This is a powerful feature of the
software. When a database is defined, you have the option of
including frequency bands in the unit. You can then generate
trends of the frequency bands. This is a normal operation that
most condition monitoring people are using to perform analysis.
But what if you did not include the correct band while
you were performing the data collection? With the PathFinder
software, you can define a band of frequencies AFTER the data
collection, and the software will generate a trend plot of the
new frequency band! This function is not available on any
other condition monitoring software package we have ever seen.
In this case, we were looking for a frequency band that
would give the operator some warning when he was about to hit
the chatter/resonance of the mill stand. This is a
trial-and-error process, but after testing several frequency
ranges we were able to zero in on a particular frequency range
that gave us this information.
|
|
New
Trend seems to precede Resonance event.
Note
that the frequency range that we selected appears to give us a
warning of impending chatter. The generated trend is the top
display, and the bottom trend is the overall trend. AGAIN,
this top trend was generated AFTER the data collection by
entering in a frequency range and asking the software to go
through the collected waveforms to generate the trend. The
trend generation is very fast, usually taking only about 10-12
seconds. This is a very powerful feature of the PathFinder
software.
New
Trend accurately predicts entry into resonance 1 min before
vibration amplitude increases ( 4 out of 4 times!).
|
|
Here we
have overlaid the new trend, shown in pink, over the overall
trend we showed earlier, in blue. Notice that the new trend
predicts the chatter event correctly 4 out of 4 times! The average warning time was about one minute before the
chatter/resonance event caused the vibration amplitudes to
reach the critical levels.
|
|
|

|