Tuesday, June 26, 2012

The Dangers of Complex THINGS

I just finished reading a book The Age of the Unthinkable by Joshua C. Ramo.  The book highlights the need for different thinking to resolve complex problems we now see in the world today, and how the old ways of dealing with them are not serving us well.  This book threw off a large number of ideas I am still trying to digest.

My great friend Ed sent me a link yesterday that provided the spark for writing this article.  Thanks, Ed!  Hey, Ed, you’re finally famous!  The story is about an upgrade of an oil refinery (named the Motiva Refinery) in Port Arthur, Texas, to make this refinery the biggest in the USA.  Note that there have been NO new refineries built in the USA in 30 years, and yet, we still have some over-capacity in oil refining.  The interesting pair of thoughts in this story is how expensive it was to do this BIG upgrade and what happened, an accident:

A $10 billion project, and it looks like the owners (Royal Dutch Shell and Saudi Aramco) are going to have costs of over $1 billion for the accident.  One small chemical spill caused this.

But, as we have all seen, all it takes is one small thing to shut down something very big…  This article will be limited to discussing complex things, I will take up discussions of other complex matters (like the financial system and international relations) at another time.


The below are examples of complex THINGS where something went wrong and had big bad consequences (Black Swans in some cases):

n  Ed’s refinery piece above
n  Fukushima, Chernobyl and close calls re nuclear power plants
n  The BP oil spill in the Gulf of Mexico
n  The SuperSonic Transport plane killed in the early 1970s
n  Major weapons systems
n  The Space Shuttle
n  The Super Collider project in Texas
n  Stuxnet and complex software in general

I will now discuss each in turn.

It is worth your while (if you find this article interesting) to go and read about the refinery above.  Because it is (was) the biggest in the USA, it is perhaps the most COMPLEX one as well.  Size of a refinery does not necessarily correlate 100% with complexity, but I would be pretty sure that this would be right up there near No. 1 in complexity.


Nuclear power plants are inherently VERY complex.  I have been told so by a physicist I know.  It is very tricky handling all the moving parts…  The people running the plants must be well-trained and alert at all times.  We have had two disasters with nuclear power plants and an almost-disaster right here in the USA (Three Mile Island, back around 1979 or so).  Fukushima and Chernobyl can both be classified as real live Black Swans.  OK, yes, building nuclear power plants along seismic fault lines should have been seen as an error, but we have nuclear power plants built along fault lines in California and in flood plains (remember the nuclear power plant that was almost flooded not too long ago in Nebraska I believe?).  I have to believe that other nuclear power plants have been built in other areas that are not well chosen (there is one less than 40 miles north of New York City).

Fukushima has caused an enormous dislocation in Japan.  They cannot adequately re-settle the population dispersed from the Fukushima area, there is no spare land available in Japan.  It will take trillions of dollars to clean up the mess.  Who knows what the likely health effects ARE already among the effected population?  If things get worse, will they try to evacuate TOKYO?  Where would the 30,000,000 people go?  (My answer: let a bunch of the best of them into the USA!).  What if the dire scenarios happen (the control rods catch fire and spread radioactive plutonium all over the Northern Hemisphere)?

Chernobyl caused death counts believed to be in the 100,000s.  It caused birth defects and other health problems among the living.  The immediate area around Chernobyl (northern Ukraine now) is too radioactive to be habitable.

And my reading says that we came “this close” to a BIG problem at Three Mile Island.  There have been a number of near misses at other US nuclear power plants as well. I n recent years there was a set of problems at a plant in Toledo, Ohio that almost had BIG problems because of the regional electrical blackout.


Another disaster happened in great part due to being very complex.  BP was drilling an offshore oil well in deep water to a very great depth below the bottom of the Gulf there (the infamous Macondo well).  Drilling in deep water has its own set of special engineering challenges, I worked (briefly) on offshore rigs as a young guy, so I know this to be true.  And drilling deep holes also has its own separate challenges.  Here we got both.

BP was drilling in this challenging environment and was considered a little “sloppy” by other oil company standards.  Save a buck here, speed up drilling there, cut some corners elsewhere.  This kind of thinking is very dangerous in environments that are complex and where the consequences of a blowout were very high.  Something like 20 lives were lost and a HUGE amount of oil came spewing into the Gulf.  If the drilling and production had gone as planned, this would have been considered a major discovery.  Lots of oil at very high pressure, a great discovery if everything had been done right and gone according to plan.

Instead, a number of small things went wrong (including modifications of the blowout preventers, a crew not paying attention, BP’s pressure on Transocean to “make hole”, etc.) led to this disaster.  All of this is still being fought out in court, and we still do not understand the total environmental damages…  My lil ol suspicious mind would tell me that they are still lying to us.  I have at least one friend who will not eat any seafood from the Gulf of Mexico.


As a young space cadet teenager in the early 1970s, I was dismayed when I heard the news that the SuperSonic Transport plane (to be bigger and better than the Anglo-French Concorde) was killed.  This was another extremely complicated machine…  It ran WAY over budget…  Seeing how the Concorde turned out to be a big waste of money for France and England, I am sure that killing the project then saved many billions of dollars for a plane of dubious worth.


In a very similar way, we see the same in major weapons systems.  “Everybody” wants to have the best weaponry to destroy our enemy’s weapons should it come to that.  Technological change forces our military to have to upgrade different weapon systems every now and then (time period varies depending on the weapon).  Even relatively uncomplicated things can get expensive and be of dubious value (eg, the M-16 automatic rifle when first introduced in Viet Nam).  Here is a list of major Defense Department projects with huge cost overruns:

The F-22 fighter, a bad plane my pilot friend tells me
The planned F-30 fighter, even worse
The C-17 transport plane, running WAY over budget and of limited use
Aircraft Carriers, are these worth it if the Chinese (or Iranians) can sink them?
The Osprey (vertical take-off plane that the Marines decided they did not want)
Almost all new helicopter projects cost too much.

The list goes on, but my expertise is limited.  Of course, almost ANYTHING the Pentagon gets involved with, almost “by definition” now, becomes MUCH more expensive than planned on.  This would be par for the course with the military-industrial complex.


Again in a similar way we had our Space Shuttle project.  We lost two of them, one via a bad rocket (frozen O-ring), the other with heat tiles peeling off.  The Space Shuttle was a very complex piece of equipment, and it is not clear to me if we benefited or not with this program.

There is no follow-on program to getting us back into space.


Yet another tremendously complex thing was the abandoned Super Collider project in Texas.  This was to be the world’s most powerful particle accelerator.  I believe the circular ring was seven miles in diameter.  The idea was to have this huge machine move particles and smash them together at such high speeds as to better learn more about sub-atomic particles and the rules of the universe at a very small scale.  The project was killed because of its high cost.

The Europeans already have a pretty big one along the French-Swiss border (CERN).  Bully for them.


Finally, although arguably not a thing, is complex software.  The “Stuxnet” virus we (and/or the Israelis) unleashed on Iran is something which may have a big blowback…   Stuxnet, no doubt, is now being looked at by China (et al) for eventual recycling into the USA.

In general, complicated software is prone to bugs, which are often VERY difficult to eradicate…

In addition, complex software can often yield incorrect results.  This too could be considered dangerous.  Who knows what kinds of bugs lurk in (for example) aircraft control software in the newest “fly by wire” airplanes?  Or drones?


And these are just THINGS!  The complexity of the above things pales beside the complexities of the financial markets, international relations and challenges of understanding the environmental and ecological systems we live in.  Perhaps I will look into those in future articles.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.