Eddy Current Testing vs Magnetic Particle Testing 3

[norzul]

Hi Guys,

Quite recently, we had a turbine blade broken during the no load mechanical running test (NLMRT) at supplier factory, prior to shipment to site.

The supplier standard practice is the followings:
1) After fabrication/machining of the blade, a 100% magnetic particle test (MPT) will be conducted by the suppplier
2) The blades together with the rotor will undergo an overspeed test.
3) A random (25% of the total blades) will be tested using MPT to detect any irregularities and defects after the test.
4) NLMRT then be conducted. If successfull, the final inspection and transport to site.

This is their normal process, however due to the failure of the blade during the initial NLMRT, one of our consultant advised that the supplier to carry out (100%) eddy current test (ECT) on the blade.

The supplier was very reluctant and not recommending to proceed with the ECT due to the followings:
1) Time consuming and not reliable results as compared to that of MPT
2) ECT can only detect the surface or very close to surface defect i.e. quite similar to MPT
3) They don't have the expertise (since their standard practice is MPT) to carry out the ECT, therefore need to outsource....cause more time and cost.

What do yu think? What are the pros and cons of ECT and MPT.

Thanks

NORZUL

 

[metengr]

I work with a fleet of GE and W turbines in our company. I concur with the above statements. The nature of the blade failure should be investigated. It could be one or more factors that have absolutely nothing to do with the use of ECT. For maximum sensitivity in the detection of surface flaws, wet fluorescent MT is about the best method.

 

[norzul]

Dear metengr,

What is GE's practice?

So yu say, for surface crack detection, MPT is better than ECT? Why? Appreciate if yu could elaborate further

What about deep internal crack, what method is recommended? Ultrasonic?

Thanks

norzul

 

[metengr]

Norzul;
In dealing with replacement of large steam turbine blades either supplied thru a turbine OEM (like GE) or thru a 3rd party supplier, we have always specified wet fluorescent MT. The reason is cost, time, and detectability of surface flaws. In 99% of the cases related to blade failures in service, high cycle fatigue is the culprit and in these cases it is a surface defect or stress riser that causes the problem (corrosion pit, machine mark, or sharp radius).

The only caveat here is that the entire blade needs to be accessible for a proper wet fluorescent MT, especially if the failure was located in the root. ECT can detect surface and subsurface flaws in turbine wheels and when the blades are installed. ECT is cumbersome. Ultrasonic testing can also be used but this is quite complicated and cumbersome.

When we have GE or W supply new blades they are always subjected to a wet fluorescent MT as their SOP (standards operating practice). Again because the blades are fully accessible.

If you had a blade failure, you need to have this analyzed to determine the cause of failure. Based on these results, an appropriate course of action can be determined. IF the blade failure was in the root or at the wheel attachment, ECT could be the better alternative while the blades are installed. Personally, if the blade failure was in the root attachment area, I would have the OEM remove the blades and inspect each one to assure this problem has been eliminated.

 

[stanweld]

When a blade fails as described, an analysis of the failure mode is required. The blade may have been improperly heat treated, the wrong material, damaged during installation, etc.

We had some very early compressor blade failures a few years ago when the blade manufacturer removed the peened compression area of the leading edge by machining too much material.

 

[rmw]

Metengr and stanweld,

see thread666-142518 for some background on what is going on here.

rmw

 

[metengr]

Thanks rmw. Very interesting information.

Ok, if the failure of the last stage blade occurred in the tie wire hole, the only question I would have is if the tie wire is brazed to the airfoil.

During overspeed testing and with the windage effects described in the other thread, the last stage blades were indeed subjected to flutter or resonance conditions.

Usually, for new equipment you can have an infant mortality rate in these types of situations. Why one blade and not others? Usually, the weak link(s) are surfaced under adverse operating conditions and, with all other things being equal being new equipment, subsequent NDT of the remaining blades in this row makes sense.

As a minimum, I would perform 100% wet fluorescent MT on the last stage blades with care emphasized at the tie wire holes. If the tie wires are brazed, ECT could be considered because crack initiation would occur at the edge of the tie wire hole due to stress concentration. However, from what I remember about past failure analysis I performed on several of our GE turbine buckets (that developed cracks at brazed tie wire holes), the critical crack length before final fracture of the airfoil was well beyond the fillet size of the brazed tie wire. This means that existing fatigue cracks can be detected using a surface examination method like wet fluorescent MT. If the critical crack size was less than the size of the braze fillet, you most likely would have to remove the braze fillet and remove the last stage blades for 100% examination (which I would endorse, as well) of the tie wire holes.

 

[norzul]

Dear Metengr,

The tie wire or we call it "damping wire" is not brazed to the blade. It is only touching the blade. see photo at

http://technical2005.blogspot.com/

My questions:
1) I've been researching the internet and my conclusion is, ECT and MPT will give the same result i.e. able to detect surface and "near" surface crack. Based on your posting, yu mentioned that wet fluorescent MPT will give max sensitivity i.e. better than ECT? Is that the case?

2) After the NLMRT, is it a practice in GE to open the casing and carry out the NDT on the blade? Even though API 612 does allow but someone told me, it is not really necessary to open the casing. Your opinion.

3) What we are planning to do after the NLMRT is just carry out the MPT on the last stage blade. This is only on the "suction" side of the blade i.e. one side only. The other side not possible to do any MPT unless open the casing. What do yu think?

Thanks

norzul

 

[metengr]

Norzul;
I viewed the failed blade on the web site you provided. Here are the answers to your questions below;

1. Yes, based on the location of failure and for your specific application, I would definitely stay with wet fluorescent MT. My rationale is that cracks will initiate off of the tie wire hole from fatigue due to local stress concentration. If there are any cracks in the remainder of the last stage blade tie wire holes, they should be detected on the suction or pressure side face of the airfoil and would be thumbnail in shape; meaning you really would not have the service conditions for initiating subsurface flaw. I was more concerned with a brazed tie wire and in that situation using ECT.

2. No. However, what you had was serious enough in my opinion to require supplemental inspection/NDT of the machine.


3. Having a single failure in the last stage blade row, I would perform a wet fluorescent MT of the entire row focusing on the tie wire holes with the casing off. I would bet that the reason this blade failed was local stress concentration from too sharp of an edge. Care must be taken for breaking all corners on tie wire holes.

 

[norzul]

TQ Metengr...valuable recommendation, another star to yu..