Not logged inAmerican Welding Society Forum
Forum AWS Website Help Search Login
Up Topic Welding Industry / Inspection & Qualification / UT Inspection for cracks in turbine compressor blades
- - By Terry Vititoe 2 (*) Date 10-09-2008 12:14
I work for a company that is in the process of developing a UT technique for crack detection on the attachment base of o row compressor blades. The question I have does anybody know if I should set a DAC curve, or set the .060" hole to 80 present of scale. The thought I have is to use a 45 degree & 70 degree 5MHz. The material is a super alloy, and the possibility exists that this technique could be used for inspection of blades while assembled in the turbine.  ANY help would be greatly appreciated.
Parent - - By mroach (**) Date 10-09-2008 18:15
Terry, the first step in developing an ultrasonic test technique is to first know the acoustic qualities of the material to be tested. Second is to know the part geometry and how the sound will propagate from the front to back surface or through the part in general. Third, and most important, the type of defect you want to find. Size and orientation of the defect will be the determining factor for the test sensitivity.   Good luck on your project.
Parent - - By Bill M (***) Date 10-09-2008 18:52
Mroach is correct. 
It would be cool to develop a mock-up of the typical blade/hub configuration and include some actual or artificial defects, maybe some EDM notches, small drilled holes, notches, etc. and using that mock-up block, you can demonstrate ability for detection of indications, adjust and set sensitivity, and verify inspection system resolution.
Parent - - By Terry Vititoe 2 (*) Date 10-09-2008 18:58
We have 4 blades coming in the next 2 weeks from OEM, two are cracked the others aren't. I've asked that all blades be cleaned prior to my visual.  If this goes as planned one of the cracked blades will be cut up & and used as a Cal. block. I can make a rough drawing, and post it tommorrow so you can see what I'm up against.
Thanks for your input.
Parent - By dmilesdot (**) Date 10-10-2008 12:22
Is this area of your exam in the "Christmas Tree" area of the base?  Some blades of GE design had a sort of mortise and tennon design for engagement.  These things were a bear to examine with ut. Lots of geometry signals to sort out.  If I remember correctly we used a 1mhz, 30* refracted L wave. The material on the GE design was some kind of super alloy that GE gave it a GE number. This probably isnt very helpful but at least know that if it is this design,  I have sympathy for you. Its a tough exam to do without false calls.
Dave.
Parent - - By CWI555 (*****) Date 10-16-2008 14:49
I've set up a procedure like that before. It was not easy to say the least. I can say the angles you proposed, the frequency, and the use of a DAC curve will very likely fail. You cannot set up a proper DAC due to geometry as there will be multiple geometric signals that can mask indications, the angles you propose are to limiting, you will find you need multiple angles, as well as frequencies/modes of transfer. I would suggest you look into phased array for this exam, I suspect your cost are going to be prohibitively high if your getting full exam coverage and utilizing old school standard pulse echo machines.
Parent - - By Terry Vititoe 2 (*) Date 10-16-2008 16:37
I think that I haven't explained the attachment for these blades. they aren't your typical fir tree attachment. I'll try to post a drawing to better eplain
Attachment: UTcompressor.ppt (32k)
Parent - - By hogan (****) Date 10-16-2008 16:46
Terry,
I think that CWI555 has given good info on the phased array.
Parent - - By Terry Vititoe 2 (*) Date 10-17-2008 10:26
check out the attachment on the previous
Thanks
Parent - - By mroach (**) Date 10-17-2008 13:15
Terry, From the illustration the cracks look like they originate from an obvious stress area and are service/fatigue related. I have to agree with CWI555 and Hogan. Conventional ultrasound is going to be limited at best. The question to be answered at this point is what area of interest is more important? Is this a test to check the blade to see if problems are forming in a known suspect area and remove the part for further evaluation or test the part in general without removing it from service for predictive maintenance purposes? From experience I can tell you that the area you are looking at, depending on how tightly pressed or saturated with lubricant, a significant defect could be vey well hidden or misinterpreted due to geometry.  
Parent - - By Terry Vititoe 2 (*) Date 10-17-2008 13:55
The intention is to inspect 4 blade attachments (2 of which have known cracks) & develop a technique to inspect compressor blades still attached to the engine.  If successful one of the cracked blades would be cut up & used as a Cal. block

Comments welcome: Thanks I need all the help I can get
Parent - - By hogan (****) Date 10-17-2008 14:26
Can't see the forest thru the trees?
Parent - - By mroach (**) Date 10-17-2008 16:51
Yeah! Can hide a lot in a pile of grass!
Parent - - By Terry Vititoe 2 (*) Date 10-29-2008 12:17
I appreciate the constructive criticism, BUT would it be possible for you guys to hold down the colorful metaphors. OH I forgot I'm probably speaking to WELDERS, so I'll explain it this way, offer a SOLOUTION, or keep your snide remarks to yourself!!!!!!!
Parent - - By mroach (**) Date 10-30-2008 14:15
Welder? Not by any means. The comments were in regards to "hash", more commonly associated with material attenuation and mode conversion. The test setup comparing signal response from a known good part and a part with the crack type defect is a good start. The only problem with the setup and technique is that it will limit the test to the area where maximum reflection occurred. In other words the technique will be limited to that particular angle, at that sound path, at that depth and at that percent of signal reflection peak amplitude. It should work, if you are only interested in one area. From the illustration, it doesn't look like there is a whole lot of scanning room so flaw sizing would be limited at best. If your looking for an ultimate solution I would suggest recruiting the services of a UT level III in your area.
Parent - - By Terry Vititoe 2 (*) Date 10-30-2008 15:33
The Test pieces that have the flaws (the crack in my attachtment) are common on this type of blade & are on the suction side, and are always in the same spot due to fatigue. The welder remark was probably uncalled for, and I appologize, but I'm not good at reading between lines, especially with e-mail.
Parent - By mroach (**) Date 11-03-2008 19:09
Sounds like it may only be one specific area of interest. That would simplify things quite a bit.
Parent - By jwright650 (*****) Date 10-30-2008 16:01
Yes, the "trees"(<--extreme case), "grass"...all terms to describe the white noise associated with viewing a UT screen....just couple to anything, switch to a very narrow screen(ie 1" scale), crank up the gain...you'll quickly see what we were talking about....I really doubt is was meant as off color or derogetory. In this "grass" it is very possible to miss indications.

Take a quick look at this: The .060" side drilled hole is in all that "grass"(noise)......

Parent - By Shawn Potter Date 11-19-2008 07:13
Just out of curiosity, would it be possible to do a straight beam analysis of enough of the crack susceptible area to make your customer happy? I did a google image search for compressor blades and saw a few designs where this would be possible.
Up Topic Welding Industry / Inspection & Qualification / UT Inspection for cracks in turbine compressor blades

Powered by mwForum 2.29.2 © 1999-2013 Markus Wichitill