Calling All Failure Analysts 5

[Yook]

I am currently working as a lone wolf materials engineer in a group of mechanical engineers. I perform failure analysis studies on a variety of failed machine components. This is my first job, and nobody working with me knows very much about the field of failure analysis. I basically make up my approach to every new problem I come across.

I decide which resources to use as reference literature. I decide the best means/tools to study the failure. I come up with my best assessment of the failure.

All of my work is done in the effort of being as useful as possible towards the goal of preventing said component from recurring failure.

My main issue is that I am going insane because I never know if I am correct in any of my conclusions. I feel as though this field is completely based upon "developing an eye" for things such as fractography, metallography, types of wear, etc. and I have no way to check my own work. I am comparing all my images to grainy images of different materials in ASM handbooks or journal articles.

I consider myself somewhat of a perfectionist and so really enjoy knowing that I am doing a decent job in my failure analysis, but I really have no way to gauge my success.
I find reference literature which contradicts previous things I've read every day. If I knew my references were sound I would be much more comfortable in my decisions as well.

Is it possible to grow as a failure analysis without anyone to check my work?

Perhaps it is the visual basis of these fields which is giving me so much trouble, as I would be more comfortable with using equations to calculate a stress or fatigue life.

I would much appreciate any advice on the matter.

 

[blacksmith37]

I recommend a book by Donald Wulpi; there are a few variations. They will have many photos of failed machinery components. I found it exceptionally helpful for machinery failures.

 

[EdStainless]

There is a professional Failure Analysis Society under ASM, they had papers and meetings at the MS&T conf that was a couple of weeks ago. Join, go to meetings, get to meet people and learn.
Perhaps you should consider hiring out some of the work, after you develop some relationships with some consultants. Consider it the cost of you getting some education.
The biggest problems that I have faced in failure analysis is getting my own people to do nothing. Just collect the pieces and hand them to me, don't try to see how they fit together, don't clean them, do nothing. When I did a lot of them I had a checklist that I used, just to remind me to confirm things as I went.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[romke]

failure analysis can be a very difficult task, a bit like forensic research into a serious crime. usually it is quite obvious how a part failed, but why that happened can have causes that have to do with materials involved, the way those materials were treated and machined, outside influences, suitability of lubricants involved etc. that list of contributing factors may also be a lot longer.

engineers involved in failure analysis also may be somewhat "single track minded", based on their education - a tribologist will look at a failure differently then a mechanical engineer, a chemical engineer or a machining specialist.

basically parts can fail due a number of reasons: incorrect design that causes higher stresses then calculated in certain areas, normal wear (each part will be designed for a specific mean time to failure, - a statistical type of calculation, some parts will last a lot shorter, others much longer), unfavourable operating conditions (shock loading, moisture, contact with chemicals, dusty and aggressive environments etc), misuse (by accident or on purpose), insufficient lubrication (which may also be designed in!) or due to material fatigue (the normal cause of failure for non conform tribologycal contacts like gears and roller bearings).

the best way to start out is to collect all the material involved in the state just after the breakdown - uncleaned and all of it. also data about the operating conditions right before failure like temperature, pressure, pollution etc should be obtained if at all possible. after that it is a good idea to look at the case not on your own but together with a number of specialists that may be more knowledgeable then you on specific subjects. opinions may then differ - but discussion might lead to a majority supported conclusion that no individual had drawn at first.

all in all it can be a very lengthy process. as far as education and books are involved: most of them are limited to specific types of failure or specific types of equipment. they can be very useful though, and some manufacturers of parts do offer literature that can be very interesting (bearing equipment manufacturers, gear manufacturers, engine builders etc)

nevertheless - it will not be easy at all. you learn by doing, the more failures you see the better you get. unfortunately failures today are not as frequent as they were say 40 years ago, thus the learning process will take longer then when i started out. at first you will often have no idea what to look for, but you will get better at it over time. it really is a specialist occupation where you need to learn to link what you see and what you measured or analysed to the failed parts in such a way that a plausible explanation results. and in some rare cases a plausible explanation may not be found....

 

[IRstuff]

A critical starting point of these types of jobs is the creation a root-cause/fishbone diagram, either formally, or in your head. Either way, it's important that you do not jump to conclusions and wind up with confirmation bias. Each piece of evidence needs to be backed up with some level of analysis to valid the conclusions.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[mrfailure]

I salute you, Yook, for trying to understand failure analysis and understand the risks of not having anyone to give you a sanity check. One of the difficulties I've found that people in your situation have is is they may use the "follow the recipe" instead of understanding the meaning of their findings. The concern I would have would be the consequences if your findings don't correct the problem. Depending on the nature of your work, you also may not have the tools you need to perform said failure analysis (for example, do you have metallography equipment?). There is no shame in sending the work out to a metallurgical lab that specializes in failure analysis, and use that opportunity to learn more about the process.

You should definately take the ASM Failure Analysis course to better understand the process of failure analysis (it helped me way back when). Good references to have besides Wulpi include ASM Handbook Volumes 11 (Failure Analysis & Prevention) and 12 (Fractography), as well as the 2-volume Handbook of Case Histories in Failure Analysis.

Good Luck!

 

[racookpe1978]

Begin with "Why Buildings Fall Down" by Matthys Levy, Mario Salvadori
Then memorize:
"Blowback" by Paul Brennan (pressure vessel failures.)

"The Logic of Failure" by Dietrich Dorner

"Design Paradigms" by Henry Petroski

For more light hearted failures that only destroyed cultures and nations,

History's Worst Decisions, Stephen Weir (A good, general analysis of failure modes by the type of decision made or bias used.)
The Decline and Fall of the Roman Empire

Military Blunders, Saul David
Senseless Secrets, Michael Lanning
Arrogant Armies, James Perry
Great Military Disasters, Julian Spilsbury
Great Military Disasters, Michael Haskew
(You'd figure he would have realized that title was used before, wouldn't you?)
Great Military Disasters, Charles Messengers
(Another editor/writer who didn't research earlier books, right?)

 

[Tmoose]

" I would be more comfortable with using equations to calculate a stress or fatigue life. "

I believe you'll find a fair chunk of failures are caused by installation and manufacturing details, as opposed to the basic design.
It is tough to calculate and cipher my way out of those.

 

[mrfailure]

Failure analysis is a deductive process. I'll give you 3 rules that have served me very well over the years if all are applied simultaneously:

1: The Sherlock Holmes Rule: "Eliminate the impossible and that which remains, no matter how improbable, must be the truth".

2: Occam's Razor: When there are 2 or more explanations for an observed phenomenon, the simplest one will be the correct one.

3: Old Metallurgist Axiom: The material does not lie!

 

[WKTaylor]

I found the following very useful over-the-years...

AFP 127-1 Vol II Safety - INVESTIGATIVE TECHNIQUES [part of the USAF mishap investigation manual series... old but good for starters]

AFAPL-TR-73-74 FIRE AND EXPLOSION MANUAL FOR AIRCRAFT ACCIDENT INVESTIGATORS

AFWAL-TR-85-2057 AIRCRAFT MISHAP FIRE PATTERN INVESTIGATIONS

WRDC-TR-89-4060 FAILURE ANALYSIS HANDBOOK

WRDC-TR-90-4075 FAILURE ANALYSIS TECHNIQUES FOR THE EVALUATION OF ELECTRICAL AND ELECTRONIC COMPONENTS IN AIRCRAFT ACCIDENT INVESTIGATIONS

WL-TR-91-4032 COMPOSITE FAILURE ANALYSIS HANDBOOK [several sub-volumes/part under the base documents number]

WL-TR-95-4004 AIRCRAFT MISHAP INVESTIGATION HANDBOOK FOR ELECTRONIC HARDWARE

ATSB TRANSPORT SAFETY INVESTIGATION REPORT - Aviation Research and Analysis Report – B20070191 Final - Aircraft Reciprocating-Engine Failure: An Analysis of Failure in a Complex Engineered System

https://www.atsb.gov.au/media/29980/b20070191.pdf

ICAO Doc 9757 AN/965 Manual of Aircraft Accident and Incident Investigation - Part III Investigation

Engine Failure Analysis - Internal Combustion Engine Failures and Their Causes [SAE] ISBN-10: 0768008859

Component Failures - Maintenance and Repair - A Tribology Handbook, SAE R-137

NOTE. MOST bearing manufacturers [plain/ball/roller/spherical] have technical literature on their websites related to bearing failure analysis and failure prevention... and usually include lubrication aspects.

Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[Yook]

Thanks very much to all those who have replied. My apologies for not checking up on this post much in the last month.

Wil Taylor, I will definitely look into the references you have listed as well.
MrFailure. I unfortunately do not have a metallography lab. I do have access to an SEM for an hourly cost and do have a stereoscope at my desk. Unfortunately, my situation has not improved much in the past 2 months since posting this either. Either my situation is such that I cannot really succeed here or I just do not have a knack/interest in failure analysis after all.

I find that I have the most trouble with answering the question "so what?" when I am performing an analysis.

As an Illustration:

Say I have a fractured 304 rod which failed at a weld toe and shows signs of reversed bending fatigue with a moderate amount of red/brown corrosion in crevices of the weld and across the entire fracture face.

I can go to my maintenance engineer colleagues with these findings and tell them that there is likely a corrosion fatigue mechanism at work here. For examples sake, assume that my current understanding of materials science and failure analysis means that this is as much as I understand about the failure.

Now I must put myself in the shoes of my ME colleague and ask "so what?". How does this information lead to a meaningful design change to prevent future failures.

If neither of us knows enough about eachother's field or about the rather hazy subject of corrosion fatigue then my work in preparing a failure analysis report yields no fruit. There is an understanding gap and I am not sure which field I need to become more proficient in to bridge it.

I see three paths that I can personally take to try and jump this gap in understanding and it is here where I am often lost:

1. I can dig deeper into fractography, metallography, metallurgy references and try to extract more information from the failure itself, but it will often be in purely MSE terms.

2. I can learn more about the mechanical engineering design process someone may have gone through to create this part and hopefully in doing so reveal some assumption which was in error.

3. I can learn more about weld engineering and metallurgy in hopes of understanding where we may have erred in our procedures or design.

I usually end up following path 1 above because there are a number of people in the group who have knowledge in 2 and 3 so I assume that it must be my understanding of MSE which is too shallow to provide enough information to jump the gap of failure cause to solution. When I do this I end up moving away from practical references like Wulpi and then feel as though I am adrift in an ocean of case studies for edge scenarios which may or may not apply to my situation(remember I have no one to check my work or guide me:/) and just become overwhelmed. I have no clue how people have the time to sift through references or familiarize themselves with even a single 1800 page ASM handbook(which usually only yields a starting point).

Does anyone else find themselves at the crossroads I described above? I understand that failure analysis is regarded as a highly interdisciplinary field, but I honestly cannot comprehend how someone could ever become close to proficient in any one field in a single lifetime given the vast amount of information out there. I struggle with knowing how far I need to go into each field to start providing more useful reports to my colleagues and work towards actual solutions.

 

[weldstan]

Determining metallurgical cause of failure can be somewhat easy. Operating history (imposed loads, temperatures exposures, corrosion service, initial installation and irregularities experienced therein, etc.) of the failed component is also essential.

 

[mrfailure]

Hi Yook. Your forthright posting listing the limitations you are currently facing tells me you have exactly the right type of mindframe suitable for failure analysis. Experience as a materials engineer over time will make you feel more confident when performing failure analysis.

It seems to me you are in a situation where you face multiple challenges. First is confidence and general knowledge. I would suggest getting some training. ASM has a failure analysis course (their most popular course) which is taught in person or on-line. That will help with organization and understanding. You can also network by joining the ASM Failure Analysis Society, which will let you connect to the failure community. You also can ask opinions on-line through this forum or the Metallurgical forum in Eng-Tips, or from the Failure Analysis forum on Linked-In.

The main challenge you face is you currently don't have much in the way of tools to help you understand the failures you get. You could at least identify the failure mode of reversed bending visually in the example you gave (and not all failure modes are as easy to identify with only visual and/or SEM), but you cannot evaluate across-section through the weld and fracture to determine if failure was related to welding, or perform microhardness profiles through the weld to characterize material, or collect chemical compositions of the weld and base metal. Not your fault - you just don't have the basic tools you need to help your company identify root cause and corrective actions, which is why you get the "So what?" response. This is why I would recommend subcontracting out to a local metallurgical laboratory and at least initially having them do the metallurgical analysis.

Finally, it is important to recognize that failure analysis is usually an interdisciplinary activity with multiple people with various expertise. No shame in not knowing everything (I certainly don't!)

 

"Lone wolf" -- good observation; failure analysis is not done very well by committees.
"I am going insane" -- normal situation at times. (Suggest you spend less time among engineers )
"Perfectionist" -- helpful but it can be paralyzing.
I believe jobs are suited to the person in terms of how they think and work, and even their personality. I suggest OCD is a useful attribute for a failure analyst (don't ask me how I know this). I view failure analysis as more of a calling than a career; to do it well requires a serious long term commitment, and attainment of a massive and varied breadth of knowledge.

Failure analysis is really no different than detective work, and the Sherlock Holmes quote is very apropos. You start out with all possibilities on the table and systematically eliminate them. You then construct a 'story' that leads from observations and testing through analysis through conclusions, connecting them in a logical chain. What you want to provide is:

1) How it happened;
2) Why it happened;
3) What can the client do to prevent or mitigate in future.
Many people doing FA only provide the first one or two items; the recommendations are where the real value lies.

I look at the skill set for doing failure analysis as a three-legged stool:

1) Knowledge of physical metallurgy, how things are manufactured and processed, including welding.
2) Knowledge of damage mechanisms and fracture modes including corrosion, fractography, thermal degradation, etc. API RP 571 - Damage Mechanisms Affecting Fixed Equipment in the Refining Industry is an excellent reference.
3) Knowledge of the application and service environment of the failed component. Since the varieties are nearly infinite, you need to be a quick study on this. For this reason I have specialized, and also why I might consult a 3rd party expert (don't ignore mechanics and technicians who have done and seen it all).

You will need skills in systematic, deductive reasoning, and be able to communicate your findings in a way that the non-expert can understand. I follow the K.I.S.S. principle in writing, and my personal standard for formal reports is 'lawyer-proof'. I suggest you make up a questionnaire to use at the initial interview step where you obtain not just the immediate circumstances but (hopefully) the history. For testing, be aware of indirect tests that can help give information; e.g., gear oil analysis; water analysis. Also, I have seen the scanning electron microscope used routinely just to pad the invoice when it was not technically justified. I use it more for confirmation or when it is the only tool that will work, and never just to go fishing (and besides it is useless as a tool if the user has inadequate fractography skills).

The ASM course mentioned by another poster is a very good suggestion. I would also read as many FA reports as you can find; there are several compendia offered by ASM. The older ASM Handbook on failure analysis (red sleeve) is excellent, and better than later editions (green) because it groups failures by environment (boilers, gears, etc.). ASM also publishes a journal on the subject.

Good luck!


"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[blakmax]

My specialisation is in adhesive bond failure forensics (but I have worked beside and with metallurgists in failure analysis) so my sympathies are with you. In my field the data base is just about non-existent and I have found a number of official crash investigations involving adhesive bond failures where the conclusions drawn are either dubious or totally in error. If ever you stray into that field you may find these references of interest:

Link
Link

http://www.adhesionassociates.com/p...f Adhesively Bonded Structures and Joints.doc

Link

I was so frustrated with the level of knowledge even from "specialists" that I created my own course to teach people how to identify bond failure characteristics, their causes and the consequences to structural performance of adhesive bonds.

If you can get over the starting hurdles and establish your own self confidence in your assessment philosophies then it can be a truly rewarding field.

Good luck

Blakmax

 

The Chemical Safety Board publishes high quality reports on serious industrial failures.

https://en.wikipedia.org/wiki/U.S._Chemical_Safety_and_Hazard_Investigation_Board

http://www.csb.gov/investigations/completed-investigations/

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[tomwalz]

People – look at the people

Look for a LEAN, Six Sigma, TQM program or similar gone terribly wrong.

Currently I am in a world where companies are implementing programs to eliminate non-productive operations. Unfortunately it is very hard to justify preventive maintenance as a productive operation.

I was once at a major aircraft company about carbide breakage. They had no data on what broke when and how. They kept track of tool breakage for a month. Number one cause was "Just Broke". However, 90% of the breakage was in left hand tools. The brazers were being blamed. They suggested I looked at the grinding operation. When right hand tools were ground, the coolant went up and over into a collector. When left hand tools were ground, the coolant hit the operator in the chest. This identified the problem. The manager told me that I did not get credit for the solution since it was a grinding problem and I was a brazing expert.

In another case, a machine was powered by acetylene gas on its wat to a troch. In this happy facility, I was not allowed to talk to anyone unless a union rep and a manager was present. On the way to coffee on of the employees slid next to me and whispered, “They make us use the gas until tank is completely empty. As the gas gets low the pressure goes to hell.” He was absolutely correct.`

I once gave a speech and I said, “Whatever you do, do not do this.” I looked up and there was a tall, very young guy in the back. The expression on his face clearly said “Shit, I better quit doing that.”



Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[rwwmetweld]

Your frustration and concern are legitimate. Resources abound, many of which have been mentioned by others, but there's nothing like practice under mentorship to develop your FA skills. As I congratulate you for understanding already, we ALL need peer review, especially early on in our careers. I've seen many painful examples of what happens when it is absent. Since you don't have an experienced mentor immediately available, the alternative is to network, network, network. Become active on a committee, like the ASM Failure Analysis Society, and get to know your fellow committee members. And don't limit yourself to rubbing shoulders with just metallurgists. Branch out! The better you understand the design and manufacturing aspects of what you're examining, the more effectively you'll be able to connect it to the metallurgical aspects. And for a consultant, which is essentially what you are, that is what it's all about.

Best regards,
Robert

https://www.linkedin.com/in/robert-warke-p-e-8b59753

 

[rwwmetweld]

Yook, I just noticed your second, longer post. You definitely have the right kind of brain for this kind of work. It's a struggle, and you're struggling. It's normal! The worst thing to do when you lack mentorship and peer review is to become overconfident in your interpretations, and then start publishing drivel that makes people like me roll their eyes. The ability to think critically, and to critique your own logic train, is absolutely key to your long-term success.

You're doing what you need to do. Just keep doing it, and *network.*

Best wishes!

Robert

P.S. We've been trying to hire someone just like you, but recruiting is forbidden on Eng-Tips. ;-)

 

[Dhurjati Sen]

Dear Yook,

Throw your problems in this forum. This is a fantastic forum with no profit motive and the answers are really impersonal. Lots of experts would offer their years of experience just for the love of the subject and the job.

People like you would keep this forum active and you would get enriched with years of experience of the experts.

Regards.

DHURJATI SEN