Talk:Failure analysis

Untitled
This article currently discusses only failure analysis in electronics. It should also describe its application to mechanics, structural engineering, business management, military strategy, and possibly other fields. Neon   Merlin   16:54, 5 August 2006 (UTC)


 * Good comment Merlin, maybe if it gets on Wikiproject engineering lists engineering editors will notice and improve it. --Blechnic (talk) 05:46, 23 April 2008 (UTC)


 * This is an important subject in regards to engineering. Ultimately the goal of engineers is to prevent failures, and as such analysis of failures is tantamount to its own field.  When this subject is expanded to all fields of engineering (besides electrical), it will be very relevant. --Markozeta (talk) 18:56, 18 June 2009 (UTC)

Wiki Education Foundation-supported course assignment
This article is or was the subject of a Wiki Education Foundation-supported course assignment. Further details are available on the course page. Student editor(s): TSprague.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 21:07, 16 January 2022 (UTC)

SQUID based microscopes
Take alook at User:Slicky/Microscopy in science, which has a subsection relating to Scanning Magnetic Flux Microscope. DFH DFH (talk) 07:53, 11 April 2008 (UTC)
 * The leading company is this field is Neocera, Inc.. DFH (talk) 08:08, 11 April 2008 (UTC)

No Fault Found
No Fault Found needs to be unmerged from this page. As a topic it is growing in importance and it is thought that NFF has very high cost implications for the maintenance industry in all sectors.JPelham (talk) 16:34, 14 April 2016 (UTC)

A draft article has been submitted Draft:No Fault FoundJPelham (talk) 09:58, 15 April 2016 (UTC)
 * The draft article has now been accepted and I believe the NFF elements should be removed from failure analysis and moved there as the current explanation confuses the two distinct topics. NFF can occur with no associated failure and the analysis in that case is not failure analysis.  I've updated the wording temporarily to reflect but in the long term it makes senses to split the topics and just briefly describe NFF as sharing techniques and redirect the reader to the NFF article if they wish to read more.JPelham (talk) 13:25, 26 April 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Failure analysis. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20051214155340/http://www.nenastran.com/newnoran/conferencePaper2/08_CPFiniteElementImplementationAdvancedFailureCriteriaComposites.pdf to http://www.nenastran.com/newnoran/conferencePaper2/08_CPFiniteElementImplementationAdvancedFailureCriteriaComposites.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 15:02, 27 September 2017 (UTC)

Structural failures
DennisK SE (talk) 18:56, 18 March 2021 (UTC)

Although the talk page claims the article is only about electronics, the 1st example is a structural failure (bridge failure). I got here from trying to research limit states in failure analysis. A current wiki article equates limit state design with load factor resistance design (LRFD) which may be how some people think of limit state DESIGN but, in my disorganized mind, limit states have nothing to do with design by any method. I think the world engineering community should agree to a definition of limit state, at least as far as mechanical and structural engineering is concerned, as the state of stress in an element just prior to failure. In 2-dimensions, one can find this state from a Mohr's Circle which shows the combined normal and shear stresses on an element. In 3D things are a lot more complicated but 3D is where our designs live and, hopefully, perform.

I'd like to throw in a couple of real life examples of failed structures which, I hope, will help illustrate how, in this day and age, structures can fail and what we might consider doing about that.

BOAT AND RV STORAGE FACILITY

The first example is unknown to everybody outside of a small community in Idaho. In 1993 a storage facility collapsed during a snow storm. I analyzed the failure for the buildings owner. The facility was built for winter storage of toys---boats and RVs---in an area that normally has 2 to 4 ft of snow on the ground all winter. Since it was desirable to be able to drive through the short dimension of the building, the building designer chose to not have braced bays in the long direction (the building was 510 ft long but only wide enough for 2 bays of RV parking). The company designing the building employed only one engineer and he was not involved in this project as he was leaving the company. However, a fellow employee asked him what size a steel column, so tall, supporting a specific distributed load over a specific area of the building, should be. The engineer answered (from the tables in the old AISC) that a 5x5 hollow tube, 1/4" thick, would suffice. Unfortunately, the engineer was not privy to all the germane info (as sometimes happens when we aren't being vigilant). In order to drive through the building, without incurring the cost and complexity of a moment resisting frame structure, the designer used an inverted pendulum design---the columns were intended to be fixed at the base and free to move and rotate at the top. Bottom fixity was via relatively large rectangular footings---longer in the long direction of the building (the short direction was braced).

As anyone who knows anything about structural design can tell you that column design is based on stability and stability is determined by effective, not actual, length. The effective length of an inverted pendulum column is twice its actual length. So the columns were seriously over stressed. Then, 2 different drawings were produced (I saw both). One had the 5x5 tubes. The other showed 4x4 tubes. The 2nd drawing was used for construction. It is easy to say that someone pocketed the diff in cost between the 2 columns but they did so with, literally, zero understanding of steel construction. Steel tubes (called steel tubes at the time and hollow structural sections now) that were 5" x 5" with a 3/16" thick wall would have weighed less than the 4x4x1/4 columns while being quite stronger and difficult to tell from the originally specified columns.

The steel structure fell down---laying down in the long direction---on the night of the heaviest single snow storm to hit the valley in several years. Some of the columns were buckled near the base (plastic failure) and some had pulled out the anchors from the footings. Since there was a little wind pushing the building in the direction of collapse, and since there were no braces in the long direction, most people determined that wind caused the collapse. That wasn't technically true. Given the effective length of the originally specified columns, they would have been grossly over stressed at failure. With the substitution to smaller columns, the columns used had no code approved capacity, in theory they couldn't support their own weight, yet they managed to hold up the structure dead load and an appreciable amount of snow, before they failed. The snow load on the building the next morning was around 40 psf (measured with a local grocery store scale) the design roof snow load was 80 psf.

Obviously, the wind direction initiated the collapse but the building would have fallen without any wind. If you imagine a forest of steel columns being seriously overloaded, they will tend to buckle in a direction dictated by weakness caused by variations in the cross section of the individuals columns. However, being in a braced system in one direction, they could only buckle in one of two directions. Eventually, enough columns would have swayed in the same direction that the entire building would collapse in that direction. (If the building had relied on inverted pendulums in both directions, it likely would have twisted as it collapsed, instead of falling in one direction like a row of dominos pushed over.) At the time, I showed, via two methods, that both the stability (increasing deflection in a column due to P-Delta effects---a stable system will reach an equilibrium state as each successive analysis produces smaller and smaller deformation---an unstable system will blow up---each successive trial will produce larger deformation) and overall stress (increasing moment due to P-Delta effects eventually exceeds the capacity of the column causing a plastic hinge to develop), would approach a limit state at about the snow load on the building in the morning, after the snow had stopped falling. Had I not had other work, I could have made calculated guesses of the snow load on the structure at the time of failure and looked at the forces due to axial load combined with a small (and unknown) wind load. That seemed like wasted effort given that the structure was doomed to fail by morning even without wind to create the initial deflection.

Among the things we should take away from this example is that, at least in 1993, the steel codes were so conservative that a building built with no theoretical capacity was able to hold itself up along with at least 20 psf of snow. Given that, why should we ever have failures? Human error! To err is human. In the normal course of design and construction, there are many checks on the process. Someone different in the office checks the design to avoid the human tendency to miss-see the same evidence in the same way through repeatedly looking at something. There is also supposed to be another check performed by the building department that has jurisdiction. In this particular case, the building official (called a building inspector) was a shoe sales man with no experience and no plan check certification. The building drawings were not sealed by a registered design professional and no design calculations were ever submitted.

COLLAPSE OF AN OFFICE TOWER ON 9-11

The next example is presented with almost no structural insights although there are plenty of reports on the collapse that one can find on line. On 9-11 we lost 3 buildings---the two big towers and a smaller one in the same complex. The smaller tower did not get hit by an airplane. it was struck by flaming debris from one of the collapsing big towers and fire took it down.

The 1st thing one should note is that most people think the probability of having a major damage event combined with a major fire is remote enough that the results shouldn't be considered. That has never been true. Natural disasters are frequently accompanied by fires usually due to broken gas mains, ruptured oil or chemical tanks, etc. In 9-11, the 1st building to collapse took out the water supply for all the surrounding area. The smaller building relied too heavily on a sprinkler system that no longer had enough water to charge the system.

The 2nd thing to note is that the resulting fire was significantly hotter than we had previously expected for fires in office buildings. The reason for that is something called fire load. Fire loads for different occupancies were determined through tests in several counties back in the 70s. In the 70s, office paper was generated manually by secretaries typing on duplicate or triplicate forms. We had copy machines but the digital revolution had not made it quite so easy to generate documents. Now we can print reams of paper with the click of a button so we tend to do so. Also, in the 70s, office furniture tended to be metal or wood. Now it is plastic or a sawdust plastic matrix product all of which burns more readily and produces more toxic smoke than traditional materials. Add to that the fact that the digital equipment itself is highly combustible and full of nasty heavy metals and you see the problem. None-the-less, some engineers still maintain that it is impossible for a steel building to fail in a fire.

What's the lesson here? Are we supposed to protect buildings from commercial airplane strikes? I might argue, yes. Shortly after WWII an Army medium bomber smacked the Empire State building. It was coming in, not going out, so it had less fuel load, but it smacked the building square. One engine went through the building and fell to the street on the far side. Flaming aviation fuel poured through offices and down elevator shafts but we did not loose that building. remember we are taking about the failure of a building that was not hit by an airplane in the 9-11 event. The impact and fire experienced by the Empire State Building would have been (I say without any effort to verify) significantly more severe than what happened to the shorter office tower on 9-11. What was the difference? Two things. One, structural design was more robust in the past because we did not have calculators with 8 digits 9most of which are meaningless if ou understand significant figures) and, two, structural engineers understood the necessity of protecting metal from fire back in the day---something they adamantly refuse to consider today. Are you comfortable knowing that turn of the century (the real one, not the last one) design and construction was better than current practices? It's a trend driven by greed---cheaper, faster, flimsier and sell it off as condos when it's done.

Relevancy of median salary and expected soft skills for failure analysis engineers?
The text in the Failure analysis engineers section seems odd and out of place (emphasis mine):

Structural Engineers and Mechanical Engineers are very common for the job. More specific majors can also get into the position such as materials engineers. Specializing in metallurgy and chemistry is always useful along with properties and strengths of materials. Someone could be hired for different reasons, whether it be to further prevent or liability issues. The median salary of a failure analysis engineer, an engineer with experience in the field, is $81,647. A failure analysis engineer requires a good amount of communication and ability to work with others. Usually, the person hired has a bachelor's degree in engineering, but there are certifications that can be acquired.

I would not expect to see information like this in an article on a technical subject, and the salary info will quickly become outdated.

RequestSelector (talk) 20:59, 9 November 2023 (UTC)