Mechanical Vibration Standard

[engrdeepee]

Hello everyone. Our company is currently operating a 50 year 900 kW Franics turbine hydroelectric power plant. One of our main objective is to come up with a vibration monitoring plan to maintain the machine. As part of the monitoring plan, we needed to establish a vibration monitoring range to determine if the generator can run unrestrictedly. However, we are currently discussing if ISO's mechanical vibration standards that would suffice since the plant's operation manual has been gone throughout time. Any tips from you guys. We currently considering ISO 10816-3 but I am still a bit skeptical about it. We have installed sensors above the turbine bearing to have it monitored in real time. We really just need the correct standards to establish the right values. Thank you.

 

[Strong]

What type of sensors are you using on the turbine bearing? What sensors are you going to use on the generator? What is the machine shaft speed? Do you plan on making detailed diagnostic measurements or simply rely on overall vibration levels compared to the ISO severity guideline as the methodology to meet your objective? I have performed diagnostic measurements, conducted routine vibration surveys, and install permanent monitors on hydro-turbines. My approach would be to understand the machine first before installing a permanent monitor and trying to set alarm levels.

Walt
Walter Strong on LinkedIn.com

 

[IRstuff]

Seems to me that this is more like a SPC application, and you should be building your limits around what is correct for your equipment and not from some arbitrary thing in a dusty spec. To that end, part of the process of developing the limits is to collect in-situ data as soon as, and as often as, possible, particularly if the system is operating "correctly" at the moment.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[Strong]

"particularly if the system is operating "correctly" at the moment."
Diagnostic measurements are needed to assess the current condition of the machine. How can one trust statistical overall amplitude values or trending when the starting point has not been verified?

Walt

 

[ronfrend]

Both IRstuff and Strong are correct. The old standards are really obsolete - I only use them as a tool for acceptance testing.
To do this properly first carry out a full diagnostic analysis on your turbine. I assume you are using plain bearings so I would also start an oil analysis programme too.
Once you are happy with the condition of the machine then you start carrying out routine monitoring. I tend to use 2 alarm levels
First (Alert) at (Average + 2 std deviations)
Second (Warning) at (Average + 3 std deviations).
You will probably find these settings will give you false positives so what I then do is take the 2 threshold values and multiply them by 50% for Alert and 100% for Warning.
The figures I give you here work for most machines with variations in process conditions so my recommendation to you is to start with a fairly low level to start with and work out what your "normal" deviations are and fine tune from there.
If you have a vibration protection system on the machine (such as Bentley 3500) I would also set up to record maximum amplitude at critical speed during runup and rundown. That will give you the best indication of excessive bearing clearances. Please note that this has to be recorded under "normal" conditions as spending too much time at critical is "a bad thing".


Ron Frend

http://www.predicon.net