Conducting a Failure Investigation 4

[metalman8357]

Hi all,

I've recieved a couple small (.25" diameter) mandrels from production and I've been asked to look at the root cause for failure. The mandrels are broken in half, and I want to analyze the fracture surface to determine the failure mechanism. I'm wondering what the best way to go about this is? I have mounting capability, micro hardness, macro hardness, an inverted and a stereo microscope. I've only been working for a few months as a materials engineer out of college, and I would appreciate some guidance on this project (I work for a small company, no one else has a metallurgy background). I know that if the failure is fatigue, I should be able to see some beach marks, but I wouldnt know the best way to setup and visualize this. Any suggestions would be extremely helpful. Thanks again!

 

[NickE]

!!!!! DO NOT PUT THE BROKEN PARTS BACK TOGETHER !!!!

First use your stereo scope and examine at various magnifications the fracture surface. There's an ASM book about failure analysis, buy it and read it. Look carefully at the micrographs.

make sure to get as much details about how and when the mandrels broke, be sure to not implicate anyone. Unless it's gross negligence there's no reason to assign blame, usually.

take your time, look at lots of references, if there's a university with an engineering department near you go to their library and spend an afternoon reading the books you can't afford.

failure takes practice, there are people here who are really good at it...

http://www.asminternational.org/portal/site/

http://www/store/storeList2/?category=,T02,&categoryName=Failure Analysis

Nick

 

[metengr]

One other option to consider that may be of benefit to you as a young materials engineer is to send the failed mandrels to a local metallurgical lab and spend some time at the lab following the analysis. Some time back, when I first started I followed a senior metallurgist and learned how to perform a proper failure analysis. I found the hands-on experience of great help and after watching and participating I was able to review reference books and get more out of them.

 

[swall]

Rule #1--always work from non destructive to destructive analysis. Reserve half of the fracture for future work. Document everything with photographs. I rather like metengr's idea of finding an outside lab and working with an experienced metallurgist from whom you can learn.

 

[TVP]

All of the above are good suggestions. Here are some more:
1. Join ASM International
2. Have your employer purchase authoritative references such as ASM HANDBOOKS (Vol 11 is Failure Analysis and Prevention, Vol 12 is Fractography)

The following links are for some excellent reference material that is freely available on the web:

Understanding the Surface Features of Fatigue Fractures: How They Describe the Failure Cause and the Failure History

http://www.asminternational.org/pdf/spotlights/jfap0502p011.pdf

Fractography of Metals and Plastics

http://www.imrtest.com/cms/repository/media/ArticleFractography.pdf

Failures of Engineering Components Due to Environmentally Assisted Cracking

http://www.asm-intl.org/pdf/spotlights/FailuresEAC.pdf

Visual Examination and Light Microscopy

http://www.georgevandervoort.com/frac_fail_pdf/Fractography_Vol12.pdf

Fatigue Fracture Appearances

http://www.asm-intl.org/pdf/spotlights/FatigueFracture.pdf

Fatigue Cracking and Fractography

http://www.lbl.gov/ritchie/Teaching/MSE212/FCP07.pdf

Techniques of Failure Analysis

http://highered.mcgraw-hill.com/sites/dl/free/0073398144/934758/Ch14TechniquesOfFailureAnalysis.pdf

How to Organize and Run a Failure Investigation (Table of Contents)

http://www.asminternational.org/content/ASM/StoreFiles/ACFAB60.pdf

Chapter 1 What Is a Failure?

http://www.asminternational.org/content/ASM/StoreFiles/Chapter_1_WEB_2.pdf

 

[metalhoosier]

Metalman,

In addition to the great suggestions above, I would highly recommend attending the ASM Metallography for Failure Analysis taught by Frauke Hogue. It is a hands on course that would directly answer your question.

MetalHoo

 

[tomwalz]

Plus 1 on what NickE said about blame.

You can do a lot of good work with a $350 digital camera from Costco and a $150 Dino-Lite if that is all the budget you have.

Understand that failure can be a complex issue.

Failure may be a single event or a series of events. The series of events may be different events or a series of the same events. 1. A new part can fracture on first use. 2. A part can wear down and then fracture. 3. A part can wear down then weaken through a series of successive impacts then finally then finally twist apart.

I am much more a carbide guy than a steel guy but I have found it very handy to keep notes in the language used by the operators.

Here is my carbide list. It doesn’t seem very complete or elegant but it does seem to help solve a lot of problems.

17 Factors That Can Contribute To Carbide Wear
1. Wear – the grains and the binder just plain wear down
Wear is often assumed to be straight abrasive wear. The just plain rubbing of material against material always plays a part. However there can be a whole lot of other factors involved.
2. Macrofracture – big chunks break off or the whole part breaks
3. Microfracture – edge chipping
4. Crack Initiation – How hard it is to start a crack.
5. Crack propagation - how fast and how far the crack runs once started
6. Individual grains breaking
7. Individual grains pulling out.
8. Friction Welding between the carbide and the material being cut
9. Physical Adhesion – the grains get physically pulled out. Think of sharp edges of the grains getting pulled by wood fibers.
10. Chemical adhesion – think of the grains as getting glued to the material being cut such as MDF, fibreboard, etc.
11. Chemical leaching that will dissolve the binder and let the grains fall out. As with any chemical reaction of this sort the acids create a salt that protects underlying binder until the salt is abraded away so grain size and binder chemistry are also important.
12. Rubbing can also generate an electrical potential that will accelerate grain loss.
13. Part deformation - If there is too much binder the part can deform.
14. Metal fatigue – The metal binder gets bent and fatigues like bending a piece of steel or other metal
15. Heat – adds to the whole thing especially as a saw goes in and out of a cut. The outside gets hotter faster than the inside. As the outside grows rapidly with the heat the inside doesn’t grow as fast and this creates stress that tends to cause flaking (spalling) on the outside.
16. Compression / Tension Cycling - in interrupted cuts the carbide rapidly goes though this cycle. There is good evidence that most damage is done as the carbide tip leaves the cut and pressure is released.
17. Tribology – as the tip moves though the material it is an acid environment and the heat and friction of the cutting create a combination of forces.


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[stanweld]

If you know what the material is supposed to be, find out if it is - perform chemical analysis; e.g. if the material was supposed to be AISI 4340, and it is actualy AISI 1020, the rest of the failure analysis (cause of fracture) is moot. One of the first things to consider when I performed failure analysis of steel provided by one of our customers was whether the involved material was indeed manufactured by us. Many a time, it was not; in which case, end of further analyses.

 

[mrfailure]

A couple of thoughts:

- Make sure you get background about the operation, service conditions, history, design, specified material, etc on the mandrel. It is critical to understand the nature of the part you are analyzing if you are going to conduct a useful failure investigation.

- Identifying failure mechanism is relatively easy. What you really need to keep in mind is why the mandrel failed (root cause) and think about what corrective actions could be taken to prevent future failures.

- The most useful step in a failure investigation is what I call the "ponder" step: Look at the pieces and think about why they are the way they are. Figure out which features are relevant and which are red herrings (such as damage that occurred after the mandrel broke). Take photos showing what you see (and feel free to go overboard on this - JPEGs don't cost anything). This visual analysis of the as-received failure will dictate the rest of the investigation.

- From what you have described, you will probably have to send some of the work out to a local lab anyway. It would be normal to check chemical composition to determine if the material was what it was supposed to be. You may also need to perform mechanical testing (tensiles, Charpy impact), and possibly SEM work (fractography, semi-quantitative composition).

- One never becomes proficient at failure analysis without actually trying. However, the stakes in the investigation may be too high in your organization to risk an incorrect or improperly performed analysis. I echo the sentiments of everyone else that you are safest to hire a local lab experienced in industrial failure analysis, but if you do see what you come up with based on lab results to see if you reach the same conclusion they did. That will be useful in getting acquainted with failure analysis.

Good luck!