Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
Ash

Member
Registered:
Posts: 2
Reply with quote  #1 
Hi!

 I had a basic doubt on Vibration analyses to ask. Any directions from the group members would help. Vibration data captured and saved on systems like OSI PI,is usually based on a change of value, i. e only captured in PI when a change occurs. In this process, there is a lot of loss of data points especially from a frequency domain perspective (if I am looking for a 1 KHz signal in the frequency domain, I would expect at least millisecond level data.) Are there techniques which can help finding these missing points? 

Thanks!
Ash
Ralph Stewart

Avatar / Picture

Sr. Member
Registered:
Posts: 235
Reply with quote  #2 
I do not know anything about the brand of equipment you use, but I do know that EMERSON (CSI) has an option to "save all data" or "only store if in alarm" or "store no data".

Have you checked to see if you have any choices to pick what you want it to do?

Thanks and Have a Great Day,
Ralph

__________________
Ralph Stewart
AlertAnalytical.com
Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #3 
My understanding is the OSI PI system only handles static values such as temperatures, pressures, and overall vibration levels. It does not save any analog or digitized data for frequency analysis. The sensor or monitoring system would have to have a separate data path to store or transmit dynamic data. You might be able to adjust the way the PI system stores values, such as saving X number of data points prior to an alarm value. I believe you will have to use other instruments or monitoring systems for vibration waveform and spectrum analysis. For example, this week I worked on a steam turbine at a power plant that had a vibration monitoring system similar to the GE-Bentley vibration monitor that provided overall values to the PI system. We had to plug into the BNC connectors that provided AC signals to get the required vibration data including shaft centerline, shaft orbits, waveforms, and spectrum. You can have your PI and eat it too![smile]

Walt
Ash

Member
Registered:
Posts: 2
Reply with quote  #4 
Thanks a bunch Ralph and Walt! I was looking at data from the historian alone, and it captures aggregated data for good reasons. It's good to know that. For more granular data, looks like I would have to have other data paths for dynamic data.
Thanks for the insights!

Cheers,
Ash
Ron Stiemsma

Sr. Member
Registered:
Posts: 84
Reply with quote  #5 

We take date from out BN 3500 racks and put it into PI.  You can get the overall, 1X, 1X phase, 2x, 2X phase and gap.  Maybe even relative and absolute data depending on the type of sensors you are using.  This will give you a lot of info even without the spectral data.  On a big turbine where most of the spectral data shows up as 1X, 2X and sub-sync, I have a good idea what the spectrum looks like by just reviewing this data.  In your original post, if you are just looking at overall a 1 KHz signal may not show up at all.  Especially if it is reading in mils pk-pk to make operators happy.

Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.