How can so many engineers get it so wrong sometimes?
Why do structures and machines fail? Surely the engineers studied the possibilities during their design?
It's an old saw that says "we learn more from our failures than our successes."
How do engineers ensure that lessons learned from one failure are not forgotten?
Books for the general reader
Henry Petroski, called To Engineer is Human
Not a particularly technical book, but it certainly does a good job of explaining why some engineering projects fail, and why others succeed.
Good for providing a "layman" reader with an appreciation of how engineers manage design risk in any kind of project, either a building, city bus, or a spacecraft. I would recommend it to anyone asking questions about how a particular design could fail, how a particular project could be late or unreliable, or trying to understand how a rational design process could go wrong despite the designer's best efforts.
When studying the subject of "engineering failures" or looking for the "lessons learned", we can start with a number of very high-profile failures or accidents, which also have the distinction of being thoroughly documented, such as the following:
Spaceflight
Rogers Commission Report, INVESTIGATION OF THE CHALLENGER ACCIDENT
https://www.gpo.gov/fdsys/pkg/GPO-CRPT-99hrpt1016/pdf/CHRG-101shrg1087-1.pdf
COLUMBIA ACCIDENT INVESTIGATION BOARD REPORT
http://spaceflight.nasa.gov/shuttle/archives/sts-107/investigation/CAIB_medres_full.pdf
The two space shuttle accidents occurred decades apart, at different phases of flight, with very different physical causes.
Yet, if you read the two reports, striking similarities come up right away.
The details of each accident illuminate some of the reasons for the similarities, but going too far into the nuts and bolts will cloud the matter.
Stepping back (as the investigators did in both of these investigations) and looking at NASA as an organization, they discovered that the ways that people managed risk was allowed to change beyond reasonable limits, and checks on decision making were never put in place (even after the first accident). Here we see again how engineering is human exercise, and how we work with each other is just as important as how we work with our tools.
Aviation
There are permanent boards that exist to investigate accidents, and you can view their accidents online:
USA: NTSB: www.ntsb.gov
Canada: TSB: http://www.bst-tsb.gc.ca/eng/
Europe does not have a united air accident agency, like EASA which regulates all of Europe's safety regulations.
Instead, each nation maintains their own investigation board. For instance:
UK: AAIB: https://www.gov.uk/government/organisations/air-accidents-investigation-branch
The scope of accident investigation agencies vary from country to country, too, depending on economics and population.
In Canada, for instance, the TSB is mandated to investigate air, road, rail, ocean, and pipeline accidents, while the UK example is only charged with examining aircraft accidents. It is important to note that the agencies that make safety policy are NOT the same agencies that investigate accidents.
In any given country, they are typically government agencies, set up in such a way that neither reports to the other, and ideally, the accident investigators report directly to an elected branch of government.
Each of these agencies publish accident reports, most directly to their websites which the public can read.
You can read each investigation report one at a time, but if you read enough, you discover patterns:
There are types of accidents that happen over and over again; there are ways that components will fail more often than others; there are mistakes that humans keep on making.
These patterns are important, because the agencies that set the safety standards look at these trends and make new policies from them.
Engineers who are interested in learning from the mistakes of others can learn a lot from these reports, and the author's opinion on this matter is not confined to just aerospace engineers.
Marine and Rail Transportation
Modes of transportation other than aviation do have bodies that are organized to investigate accidents, and create safety standards, even if the mandates are not as clear as they are in aviation.
In many countries (such as the USA and Canada, as mentioned before) the same agencies that investigate aviation accidents are also charged with investigating major transportation accidents.
Other countries (such as the UK) there is a separate board to investigate each of Rail and Marine accidents.
Reports of these accidents are also publicly available through the respective websites.
Road Transportation
I think the safety agencies of the world just gave up early on, trying to study every automobile accident - they are dramatically more frequent than other categories of transportation accident.
Governments continue to track statistics on accident rates, leave individual accident investigations to local police authorities (who have their own systems for tracking accident causes and investigation techniques).
Also not handled the same way is the manner in which failure data is communicated to the automobile manufacturers. That's a whole article in itself. They seem to rely on voluntary reporting and supplier testing.
When it comes to studying accident rates, statistics, and promulgating safety findings to the public, here is a major difference from aviation, in the USA.
The National Highway Traffic Safety Administration (NHTSA) is in the business of writing the rules AND studying accident causes.
While there is a certain convenience in this arrangement, the proximity of the NHTSA investigations to the politics of the DOT is precisely the reason that many aviation investigation boards do NOT report to the government regulator of aviation.
For a contrary opinion in automobile safety, it is agencies like the Insurance Institute of Highway Safety (IIHS) that have some clout.
Since their work is funded by insurance agencies, their findings are reports are not, by design, beholden to any government agenda.
Since insurance agencies benefit from highway safety by lowering the insurance cost of settlement payouts, they have some "skin in the game" that should ostensibly make their reporting conservative.
Any attempt to compare findings or activities of the NHTSA against the IIHS is unknown to me, but would be VERY interesting.
Energy
The national and international focus becomes more local, when looking at the regulation of the energy sector. This leads to a more local focus when things go wrong, and when accidents are investigated.
From time to time, national investigations or national approvals become necessary approval of certain projects (just take the Keystone XL pipeline upgrade for an example), but most of he time, state or provincial regulations apply.
Accidents that don't garner national attention will be investigated only at a local level. Here is an example of a pipeline leak into a major river in Alberta, Canada.
This particular accident (and others in the same year) did not garner the attention of the national TSB, even though the mandate of the TSB includes investigation of pipeline accidents.
For an example of a pipeline accident that did warrant TSB investigation, this one from 2013 involved a natural gas explosion that would have been fatal if it had occurred in a populated area.
Since the TSB did not investigate the local pipeline leak, but did investigate the gas pipeline explosion, I wonder what trends and statistics they use when they survey the safety record of the industry.
For a much higher profile energy accident, it's easy to find the Three Mile Island accident and the investigation that ensued, created by then-president Jimmy Carter, which examined the causes and provided a report 6 months later. There were, simultaneously, many other state- and national investigations, each with their own focus of attention. Much has been learned from that accident, and it is widely believed by nuclear engineers that modern nuclear power stations would not suffer the same type of accident as a result of the stuck valve at TMI. Even in my industry, aviation, those lessons are considered valuable. For example, component reliability statistics, originally compiled for nuclear power plants, are adapted for use when estimating the reliability of certain components of aircraft. Reliability statistics tend to be quite difficult to come by, so even sources outside the aviation industry are treated with appreciation.
Chemical
Investigations of accidents in an industrial setting often invoke workplace health and safety laws.
In the USA, the Chemical Safety Board (CSB) is mandated to investigate chemical accidents at industrial sites, but they are not aligned with any regulatory body, despite being an agency of the US Federal government. In this regard the CSB is an arm's length investigative body, and the way it carries out its mandate is similar to that of the NTSB in aviation.
In the UK, the Health and Safety Executive (HSE) investigates and publicly reports on accident findings. They also carry out industry-wide research and root-cause analysis.
The HSE is an umbrella agency that includes nuclear, environment, and natural resource regulators in the nations of the UK. Here we find their investigative power is not at arm's length, but they do have the power to call a public inquiry.
Note that the reports published on the webpage are not up to date. This newsletter is much more current: the Loss Prevention Bulletin.
Civil
Not a civil engineer myself, so I won't propose any authoritative sources, unless others suggest them.
It just goes without saying that the construction of buildings, roads, bridges, and other engineering works of infrastructure have suffered many infamous failures.
Each failure receives some notoriety in the press, in political circles, and tends to lead to judicial action in local courts.
If enough people are injured or killed, political action may create an investigation/inquiry board, usually in an ad-hoc manner.
Here is an example: Elliot Lake Algo Mall Collapse Inquiry(This link was current in 2016, but will probably become broken in a few years. As I said: "ad-hoc")
My (informal) observation is that most, if not all, failures related to building design are examined primarily in the local legal system.
The findings of the investigation get wrapped up in the court's ruling, and while the causes are rigorously examined, no effort is made to find patterns, evaluate industry practices, or make reports to other bodies of government or industry.
This is different from other engineering endeavors such as transportation, which have permanent investigation bodies with a formal mandate to analyze accidents from a broad range of perspectives, and to report directly to government.
For a long time in history, building codes have been local regulations, and only more recent efforts to make national standards have created some kind of "universal" code.
Even so, local jurisdictions continue to modify the national code to their purposes, often to alleviate something not applicable in the area (such as earthquake protection in the central prairies).
So from an outsider's perspective, the Building Code isn't as "National" as it claims to be. I invite clarification or correction on this matter, and will add citations if any are offered.
Many high-profile civil/structural disasters (both natural and human-caused) are covered in the weekly magazine Engineering News-Record (ENR). Here is an editorial "round up" for the year 2016 in the ENR.
As news, investigations, and other developments occur, the magazine publishes updates. This can go on for months. The magazine is well respected in the Engineering and Construction communities; it has been published since 1917.
From time to time, ENR will publish books which are compilations of their magazine articles. These present an interesting outline of how and when information on a disaster became available.
In one of their books, Construction Disasters, Steven S. Ross (1984), subjects include the Tacoma Narrows Bridge (1940), Kemper Arena Roof Failure (1979), and Hyatt Walkway Collapse (1981).
