Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
Shurafa2

Sr. Member
Registered:
Posts: 78
Reply with quote  #1 

What is your answer to the below vibration question and *** why ***?

This page is from "Fault Analysis and Correction". Part 5 of Vibration Analysis Certification Exam Preparation Package. Amazon Link

Drop me a private message, if you are interested in more details about the booklets.

Regards- Ali M. Al-Shurafa

#certification #rotatingequipment #conditionmonitoring #vibrationanalysis #reliability #machinery #machines #vibration #diagnosis

Vibration Analysis Certification Exam Preparation Package Certified Vi.png 

 
Attached Files
pdf Vibration Analysis Certification Exam Preparation Package Certified Vi.pdf (4.27 MB, 20 views)

DP

Member
Registered:
Posts: 9
Reply with quote  #2 
My answer to this is that the plots are constructed in a way that are completely unrealistic and meant to confuse the reader/test taker.

I think I get the point with the units in Plot 1 being msec and mils and Plot 2 using sec and mils p-p.  Identifying that and seeing Plot 1 has the data (kind of) centered around 0 mils indicate that Plot 1 is a waveform and Plot 2 is a trend.  What is confusing is that you can generate a DC coupled time waveform so the waveform will not be centered around 0 mils. Also, what was the point of using the exact same curve offset vertically?  Plot 1 doesn't resemble periodic vibration even if you assume many frequencies present.  Plot 2 doesn't resemble a realistic trend plot with the amplitude varying +/-200% in just a few seconds. 

Questions like this one discourage good vibration analysts from pursuing certification.
fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #3 
agree second its a overall trend, only overalls has unit type rms, peak, pp

yes both plots are similar, in real life you wont find a wave similar to an overall, but the point of the questions is your understanding about units and type in the digitalization of the signals.

both plots been realistic i dont know?

plot 2,  is posiible in real life to find overall that varies 500% over the period of 5 seconds... beat frequency?

plot 1, i think is posible to a complex signal to look like this.

Alex

Avatar / Picture

Sr. Member
Registered:
Posts: 140
Reply with quote  #4 
I must say I 100% agree with DP. I have never seen a trend like on the second plot. The question is (at least) very strange IMO and itself causes another questions. Luckily I have never had to deal with such type of questions in my diagnostics career.  
Noknroll

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 734
Reply with quote  #5 
This has confused some very experienced analysts here so what hope does a rookie cat1 have
EDIT:to be fair it might make more sense with course teachings around that question fresh in the students mind
Alex

Avatar / Picture

Sr. Member
Registered:
Posts: 140
Reply with quote  #6 
The problem is not because the question would be hard. The problem is with the question itself. Mr Shurafa, please tell me from your practice, which software has the similar trend display as yours. I am talking about the trend which has seconds on X axis and which curve is interpolated. I have never seen anything like your plot with Emerson, Pr├╝ftechnik, Adash, Erbessd.
    
MarkL

Avatar / Picture

Sr. Member
Registered:
Posts: 920
Reply with quote  #7 
I too came across this question the other night whilst reading the book and it had me very confused, it doenst look like any trend or waveform I have ever seen.

Shurafa2

Sr. Member
Registered:
Posts: 78
Reply with quote  #8 
Interesting and challenging comments. Thanks a lot for the critique. I'll try to rewrite down the weaknesses/incorrectnesses to ensure I understood your comments. 

  1. This kind of data is not seen in real-life. It doesn't resemble realistic periodic vibration even if you assume many frequencies present. 
  2. This kind of data is not seen in real-life. It is unrealistic that an amplitude varying +/-200% in just a few seconds.
  3. DC coupled time waveform can be generated so the waveform will not be centered around 0 mils. This will match with Plot 2.
  4. Because the question is confusing, it discourages good vibration analysts from pursuing certification. The question is very strange and itself causes other questions. This (question) has confused some very experienced analysts so it is not suitable for Cat 1.
  5. Both plots are similar (identical?), in real life, you won't find a wave similar to an overall.
  6. No common software has a similar trend display as shown in the question i.e the trend which has seconds on X-axis and whose curve is interpolated.

Regards- Ali M. Al-Shurafa
fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #9 
here is a real data plot like 1.


img00052.png
 
Here is a overal plot over time in seconds like plot #2, sorry dont have them in mills.


img00053.png

Allen Plymon

Member
Registered:
Posts: 27
Reply with quote  #10 
When i look at the two time waveforms, my first thought were AC & DC coupled portrayals of the same measure - with a bit of vertical sizing error (perhaps the correction but I'd expect a more obvious change)  between the two.

Allen Plymon
electricpete

Sr. Member
Registered:
Posts: 517
Reply with quote  #11 
I think the intended answer is D:  plot 1 is TWF and plot 2 is an "overall"   

What is supposed to lead us in this direction is the time scale (msec in first an sec in 2nd) along with the vertical sale (mils in the first and mils pk-pk in the 2nd_

I agree with DP, fburgos etc it's a contrived question....  The similar appearance of the plots is very odd considering one is apparently supposed to be a TWF and one is supposed to be an overall trend on a different timescale.  Also note the trend of an overall on such a short timescale depends on the processing parameters (over what timescale is it recomputing the overall). It would have been better if they made the overall trend look completely different.  fburgos gave a good real life example.

Gong to Shurafa's summary points, I've added my reaction to each below in [CAPITAL LETTERS]
  1. This kind of data is not seen in real-life. It doesn't resemble realistic periodic vibration even if you assume many frequencies present.  [I THINK THIS IS BETTER CAPTURED IN POINT 5] 
  2. This kind of data is not seen in real-life. It is unrealistic that an amplitude varying +/-200% in just a few seconds.  [I'M NOT SURE I AGREE.  IT DEPENDS ON THE TWF AND THE PROCESSING]
  3. DC coupled time waveform can be generated so the waveform will not be centered around 0 mils. This will match with Plot 2.  [POSSIBLY, BUT TYPICALLY I EXPECT TWF TO BE CENTERED ON ZERO AND OVERALL TO BE ALWAYS POSITIVE... THIS SEEMS LIKE A MINOR POINT WITH RESPECT TO THE PURPOSE OF THE QUESTION]
  4. Because the question is confusing, it discourages good vibration analysts from pursuing certification. The question is very strange and itself causes other questions. This (question) has confused some very experienced analysts so it is not suitable for Cat 1. [I AGREE]
  5. Both plots are similar (identical?), in real life, you won't find a wave similar to an overall.  [I AGREE]
  6. No common software has a similar trend display as shown in the question i.e the trend which has seconds on X-axis and whose curve is interpolated. [I DON'T AGREE.  ON OUR ENTEK DATAPAKS I KNOW WE HAD A LIVE TREND GRAPH OF OVERALL WHICH WE COULD WATCH FOR STABILIZATION AFTER PLUNKING DOWN THE ACCELEROMETER AND BEFORE PRESSING STORE.  THERE MAY BE THE SAME ON CSI 2140, I'M NOT SURE]
Sinski

Avatar / Picture

Sr. Member
Registered:
Posts: 339
Reply with quote  #12 
Quote:
Originally Posted by electricpete
I think the intended answer is D:  plot 1 is TWF and plot 2 is an "overall"   

What is supposed to lead us in this direction is the time scale (msec in first an sec in 2nd) along with the vertical sale (mils in the first and mils pk-pk in the 2nd_

I agree with DP, fburgos etc it's a contrived question....  The similar appearance of the plots is very odd considering one is apparently supposed to be a TWF and one is supposed to be an overall trend on a different timescale.  Also note the trend of an overall on such a short timescale depends on the processing parameters (over what timescale is it recomputing the overall). It would have been better if they made the overall trend look completely different.  fburgos gave a good real life example.

Gong to Shurafa's summary points, I've added my reaction to each below in [CAPITAL LETTERS]
  1. This kind of data is not seen in real-life. It doesn't resemble realistic periodic vibration even if you assume many frequencies present.  [I THINK THIS IS BETTER CAPTURED IN POINT 5] 
  2. This kind of data is not seen in real-life. It is unrealistic that an amplitude varying +/-200% in just a few seconds.  [I'M NOT SURE I AGREE.  IT DEPENDS ON THE TWF AND THE PROCESSING]
  3. DC coupled time waveform can be generated so the waveform will not be centered around 0 mils. This will match with Plot 2.  [POSSIBLY, BUT TYPICALLY I EXPECT TWF TO BE CENTERED ON ZERO AND OVERALL TO BE ALWAYS POSITIVE... THIS SEEMS LIKE A MINOR POINT WITH RESPECT TO THE PURPOSE OF THE QUESTION]
  4. Because the question is confusing, it discourages good vibration analysts from pursuing certification. The question is very strange and itself causes other questions. This (question) has confused some very experienced analysts so it is not suitable for Cat 1. [I AGREE]
  5. Both plots are similar (identical?), in real life, you won't find a wave similar to an overall.  [I AGREE]
  6. No common software has a similar trend display as shown in the question i.e the trend which has seconds on X-axis and whose curve is interpolated. [I DON'T AGREE.  ON OUR ENTEK DATAPAKS I KNOW WE HAD A LIVE TREND GRAPH OF OVERALL WHICH WE COULD WATCH FOR STABILIZATION AFTER PLUNKING DOWN THE ACCELEROMETER AND BEFORE PRESSING STORE.  THERE MAY BE THE SAME ON CSI 2140, I'M NOT SURE]


There is a trend graph for overall in the 2140. 
fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #13 
Quote:
Originally Posted by Sinski


There is a trend graph for overall in the 2140. 


yes in analyze module, not shure if you can choose overall from route (f12) , but shure is in home, f7, f1 select overall, f2 to selec a mode 1hz to 80kHz or fmin fmax


I've used schenck, vibroport40, vibrotest60 and viboport80, and all had overall trend graph (overall vs time)
DP

Member
Registered:
Posts: 9
Reply with quote  #14 
I apologize if my response came off as overly critical, that wasn't my intent. Thanks for putting together another resource to help those pursuing certification.  

The fact that there is discussion/debate by lots of RSG's (really smart guys) raises a red flag about a question for a Category I review.

If I understand this correctly, the intent is to assess whether or not the test taker can identify the difference between a TWF and an overall trend.  In that case make the plots look like a typical TWF and a typical overall trend that a Cat I analyst should be able to assess.

Simplify this by:
  1. Make the curve in plot 1 look like a sine wave  with one or a couple of frequencies with the curve centered about 0 in amplitude.
  2. A typical trend plot that is of any use for most machinery should have a useful time scale.  It may make sense to change the time axis units in Plot 2 to minutes or days.  A machine with load/speed variation or a machine experiencing a certain fault may warrant inspection of trends with the time interval in seconds but I wouldn't expect a Cat I analyst to be digging into something like that.
Again, there are exceptions to every generalization we could make to what a certain plot type should look like.  There are reasons why I might evaluate a DC coupled TWF or an overall amplitude trend over a period of a few seconds.   I just think the goal should be to remove confusing details and try to keep test/review questions close to realistic data.

One other question... what did you mean by "correction".  Is that supposed to be "compensation"?

 
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.