Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #1 
Have a customer with a number of large fans, and after the recent failure of one of the wheels we are looking at the best way to inspect the remaining wheels. These are large wheels (10,000 lbs, 7 ft diameter, 4 ft wide) and fairly complex weldments, overlaid with chrome-carbide liners.

Would 2-channel impact testing, if done in a uniform manner on all wheels, reveal at least the existence of cracks (for further testing)? I would think that significant differences in the FRFs, or a lack of coherence (impact on one side, measure on the other) might indicate a significant crack.

Have any of you done this before? I don't want to propose this if it has little chance of being effective.

__________________
"The trend is your friend"
plongson

Member
Registered:
Posts: 19
Reply with quote  #2 
Hmm, dunno Rusty. Never even thought about using vibe technology for crack evaluation. I'm not saying it can't be done but I'd hate to be the first boots on the ground charging $100/hr to "experiment".

It would be a great research paper but I'd feel better showing up on site with a proven method...

Stuff like Zyglo from Magnaflux for instance.

http://magnaflux.com/Magnaflux/Products/Liquid-Penetrant-Testing.htm

__________________

Vibe Category/Level III
Owner of Longson Services Inc.

RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #3 
Because of the chrome-carbide overlays, many of the base welds are hidden since the liners are seam-welded along the edges, so not visible for visual inspection. Would dye-penetrant or magnaflux testing work in this situation?
__________________
"The trend is your friend"
Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,397
Reply with quote  #4 
I agree with Plongson.  I've seen crack testing done a lot with the dye penetrant product. Even small cracks really stand out.  Paper machine roll journals and reel spools are regularly tested this way.  I further think that by the time vibration provides a confident difference to be able to diagnose a crack the wheel or shaft is probably nearly in 2 pieces.  I have not tested my theory - it is but a gut feel.  Be suspicious of stainless steel since that cracks - not a matter of if but when due to the lack of fatigue limit.  Yet so many manufacturers still use stainless.  In one instance one of these fan impellors exploded and left a hole in the roof far above. Luck no one was nearby. rgds

RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #5 
Keep in mind 1/2 of the welds are "hidden" -- how do you test that with dye penetrant? Also, not talking about using "vibration" to identify a wheel with cracks, but rather 2-Channel mpact testing, done during a shutdown. Actually with a 4-channel 2140 you can measure the impact on 'A' and the response on the other 3 channels.
__________________
"The trend is your friend"
Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #6 
The impact method can work, but you have to find the vibration mode that is most influenced by a crack and the location of interest. The Modal Shop and other sell flaw detection systems based on impact testing.

Walt
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #7 
Thanks for the input. I will be onsite for the shutdown anyway, so will give it a go using the equipment I have, which is 5 or 6 hammer sizes (all of them I think) though I doubt the big sledge or the little tiny one would be of any use. I'll report back with results (or lack thereof).
__________________
"The trend is your friend"
Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #8 
Rusty,

Many years ago I conducted an impact test on the blades of a large (12 foot diameter) flue gas recirculation fan. One of the blades had broke, but the others remained intact. The hammer of choice would be the 1 pound head impact hammer from PCB or similar. My approach would be to get the impact response from several locations near the blade tip on each blade and check for similarities and differences. Based on these results, you can proceed on more detailed modal testing. Bring your lunch, because you'll be inside the fan for quite a while!

Walt
WWST

Sr. Member
Registered:
Posts: 86
Reply with quote  #9 
Rusty,

I have used this technique successfully a few times. The cracks in the attached case history were verified by NDT. I have never used it on fans, but I'm sure it couldn't hurt to try and find the obvious outliers, then verify with NDT or monitored carefully for amplitude/phase changes. This technique allowed them to quickly identify in 15 minutes the ones that needed full testing, which required 12-16 hours of disassembly/assembly.

I have a powerpoint if you would like it. Maybe I'll upload to YouTube and link here later.


corn processing centrifuge cracks 1.jpgter. 2-fabricated-replacement.jpg  CentrifugeBasket cracked.jpg  coherence and FRF good and bad.jpg    corn processing centrifuge cracks.jpg 


RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #10 
Bill, exactly what I had in mind. Just a "first pass" as I'm not sure they plan to inspect all the fans. The other thing I don't track closely enough is influence coefficients when I balance these fans. I know about what they are so don't make a point of recording them since I do these so often. That would be a help as significant cracks (in the wheel) should charge the coefficients.

Thanks for the info above, and yes, the pp would be helpful.

And I need to get with you this week on a remote system. Looking at installing on 22 fans (at least), but they want to try before they buy.

__________________
"The trend is your friend"
WWST

Sr. Member
Registered:
Posts: 86
Reply with quote  #11 
No problem, Rusty I'll email it. We can do up to a 2 week demo if they are interested. We can also talk about your commission [smile] . I have a test site in NW AR now, I need to visit soon.
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #12 
Bill, thanks, but there wouldn't be any commission on a sale to one of my existing customers -- I feel that would be a conflict of interest. If it works well, and keeps them out of trouble, that's payment enough for me.  However, if your customer in NW Ark needs additional reliability services, then we could certainly discuss that!
__________________
"The trend is your friend"
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #13 
A bit late to the game, but my group has used bumps on turbine wheels to detect defects. I've only done this on a bridge, but I've read some of their reports. If you plan to use the different blades as controls for eachother, you are going to be getting a lot of false positives. For how highly engineered these wheels are, the fn's you are looking for are counterintuitively, highly variable, up to 5-10%. That's more variation than you'd expect to see in error from a cracked weldment, unless (as stated before), the thing is about to fly apart anyway.

For wheels this large though, I would be surprised if they did not take a base note before commissioning so that you have a reference for an [assumed] fully intact wheel with specific fn's for each blade. This is probably the only feasible way of defect detection with bumps.
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.