Photo courtesy of the U.S. Library of Congress

Talking with Dr. Layne Karafantis of the Smithsonian National Air and Space Museum about human factors engineering initiatives in the maintenance of complex aerospace systems.


Gridium:     Hello everyone, and welcome to this conversation with Dr. Karafantis, curator, modern military aircraft in the Aeronautics Department of the Smithsonian’s National Air and Space Museum.

Dr. Karafantis received her PhD in the history of science and technology from the Johns Hopkins University, specializing in aerospace history, the history of American military technologies, and suburban and urban histories.

My name is Millen, and I’m with Gridium. Buildings use our software to fine-tune operations.

Layne and I will be discussing a paper she co-authored with Daniel Volmar of Harvard University titled “Idiot-Proofing the Air Force: Human Factors Engineering and the Crisis of Systems Maintenance”.

It’s such a pleasure to be speaking with you today, Dr. Karafantis.

Dr. Karafantis:        Well, thank you Millen, and thank you for your interest in my work. I’m glad to be a part of what I’m seeing as a growing trend among scholars to consider maintenance of systems to the same extent that we investigate innovation.

Gridium:             Yeah, same here. So, let’s get started with this question and starting from the top: what is human factors engineering?

Dr. Karafantis:        I think most people would be more familiar with the term ergonomics which is sometimes used interchangeably with human factors engineering, particularly outside of the US, but there’s something specific I think to human factors engineering as opposed to ergonomics, and that is, ergonomics often makes people think of making machines or objects that are going to be more comfortable or user friendly, whereas human factors engineering goes beyond that.

Yes, it’s a way of designing objects or spaces, like say a cockpit, with consideration taken to the pilot’s body using anthropometric data.

An example of this might be measuring the distance an arm can reach when subjected to various G forces, or something like that. Human factors engineering also considers how to make human-machine interaction more efficient, such as figuring out how much information a pilot could process at one time, which inserts a bit of psychology into the process. Human factors engineering is highly interdisciplinary and it integrates aspects of physiology, social or cognitive psychology, operations research, and quite a few other fields.

Gridium:             Okay. I know that you mentioned some examples like color-coding connectors and fasteners. What are some others?

Dr. Karafantis:        So, color-coding is an easy way to make (we’ll stay with the cockpit example) a pilot’s flying experience more efficient because those signals would quicken response time but also, uniform placement of pilot controls is a form of implementing human factors into the cockpit.

I mean, up through almost the Second World War, military aircraft (well, also commercial and private planes, but let’s stick with the military)… military aircraft did not have uniform placement of controls, meaning dials, switches and so on. So, every time the pilot went into this new plane they might find themselves having to learn an entirely new setup or–and this is where human factors becomes very important–­­­the pilot might incorrectly assume that the controls are laid out in the way that they’re familiar with, and that is how accidents happen.

Gridium:             Ah, yes.

Dr. Karafantis:        So, actually the growing impetus for human factors engineering is safety.

After World War II the field grew due to the increasing number of complex safety-critical or mission-critical systems, particularly in aerospace. These, in turn, have become more prevalent due to our emerging technologies, mainly anything involving nuclear energy weapons, as well as computing systems that enable far-flung operations that could be conducted in real-time.

Gridium:             Okay, so granted, color-coding sounds simple, but I was surprised to learn about the role human factors engineering played in the Damascus incident. What happened here?

Dr. Karafantis:        Right, so for your listeners who don’t know this story, in 1980 there was a broken arrow–that’s what we call accidental events involving nuclear weapons.

This broken arrow happened at a Titan missile silo on farmland near a little town of Damascus, which is about 50 miles north of Little Rock, Arkansas. These missiles, by the way, are operated by the Strategic Air Command within the Air Force. So, the story goes: two enlisted maintenance workers were checking a low-pressure warning on a missile and someone dropped a nine pound socket from their socket wrench. Then, this socket fell seven stories, pierced the skin of the missile, and that ruptured the first stage propellant tank.

Now, yeah… It was a problem, the site had to be evacuated, but the fumes eventually caused a huge explosion which blew a 740-ton blast door off the top of the silo, and also projected this nine-megaton warhead about 100 feet from the opening.

Gridium:             Wow.

Dr. Karafantis:        Yeah. But, I mean, thankfully the warhead safety measures worked and it didn’t leak any radioactive material, but one person died and some 20 others were seriously injured.

Gridium:             One of the issues here, as I understand it, was the use of the “wrong tool”, but it’s more nuanced, isn’t it?

Dr. Karafantis:        Yes, yes. Air Force command decided that they would blame this series of unfortunate events on the fact that the maintenance workers were using a socket wrench instead of a torque wrench.

So, recently a technical order had come down instructing the workers to only use a torque wrench to loosen the oxidizer valve, but these men had been using a socket wrench without any issue in their minds for a really long time, so when they didn’t have a torque wrench on them, they opted to use the same tool that they had been using.

And, really, I think the choice in tool is moot because the incident wasn’t caused by the workers using the wrong tool, it was caused because they dropped the socket.

The order to use the torque wrench had nothing to do with the possibility of mishandling the socket in the socket wrench.

Gridium:             So, let’s ratchet it up a few levels: the interplay between headquarters management and onsite technicians also played a big role here, isn’t that right?

I think this is visible by the action taken against the techs and the action that was not taken against management.

Dr. Karafantis:        Yes. Yes, exactly. The Air Force really did scapegoat the little guys.

They cited one man for dereliction of duty, others were committed to mental hospitals presumably for reasons related to morale, I think. All the official reports squarely blamed the maintenance workers’ supposed inability to use correct tools or their deviation from procedure.

These were the reasons that they cited that the incident occurred. I mean, no one at SAC headquarters was fired, but they deserved blame. It wasn’t just these workers’ fault. The onsite maintenance team, in fact, had wanted to vent the fuel by opening the blast door, the one that eventually got blown off in the explosion, but headquarters immediately rejected their suggestion.

Furthermore, in the larger picture, mid-ranking officers had already known that the Titan was troublesome to maintain, it was a very troublesome missile, and yet headquarters still blamed human error rather than assigning any blame to equipment.

Dr. Karafantis:        So, the final thing I think I said is, officers knew that the Titan was troublesome to maintain and they still blamed human error rather than assigning any blame to the equipment.

Gridium:             So, the power plays at work in Damascus incident, it’s been argued, are also fundamentally embedded in the relationship between the designer of the technological system and the maintainer of that system. How did Alan Swain, a maintenance engineer at Sandia National Labs, which by the way is in my hometown, describe this relationship?

Dr. Karafantis:        Hometown of Albuquerque or Livermore? It seems long ago now, but I wrote a master’s thesis about how Sandia National Labs impacted urban areas in the post-war years. So, I spent a lot of time in both those cities doing research and enjoyed both for their own reasons.

Gridium:             That’s Albuquerque, New Mexico.

Dr. Karafantis:        Albuquerque, fantastic, I love green chile.

Gridium:             Me too, good choice.

Dr. Karafantis:        Anyway, yes Alan Swain did try to explain why these designers resisted feedback from maintenance engineers in the design of Air Force systems.

He didn’t think that the designers failed to consider the human elements in their designs, but that they were considering it in a flawed and pretty condescending way.

That is, Swain thought that any mistakes were too easily assigned to the operators and that designers failed to own up to the possibility that their designs could be responsible for the error, or consider that the intelligence of the worker might actually be an asset.

Gridium:                 I was struck by this example, why do you think headquarters thinks that maintenance technicians need to be reminded not to walk on bombs?

Dr. Karafantis:        (Laughs) Right. Well, according to Swain, at one military installation, these maintenance technicians were regularly walking on a large bomb, presumably a nuclear weapon…

Gridium:             Wow.

Dr. Karafantis:        …and that was with an aircraft–right? And headquarters was obviously perturbed by this, so they wanted caution signs to be printed on the bombs, so that this would stop.

But this reaction really overlooked a fundamental problem while addressing a symptom, in my opinion. This what Swain said, and I agree… I mean, headquarters might have thought to asked themselves, “Why are these workers walking on the bombs?” or “Do they need a work stand?” things like that, instead of putting something that might be a redundant factoid of writing “caution” on the side of the bomb.

Gridium:             There is an interesting story with Jeff Bezos, the CEO of Amazon, who was informed of an accident at a warehouse where a worker had damaged his thumb by placing a backpack on a conveyor belt, and so the question is, “Why was the backpack on the conveyor belt?” and it goes on from there.

Your paper seems to argue that Mr. Swain didn’t get it quite right. Why isn’t it enough for the designers to simply design technological systems that are straightforward to maintain?

Dr. Karafantis:        Well, it’s circular, right?

People are not uniform and people can be unpredictable.

I mean, it’d be a tremendous challenge to design for everyone in a way that a technological system is accessible and straightforward to maintain. I mean, that’s placing an awful lot of responsibility on the designer. The Air Force tried to mitigate this problem somewhat in its selection of certain people for particular types of jobs, but there are a lot of real and metaphorical moving parts here.

Gridium:             It looks like the Air Force put a lot of effort into studying both designers and maintainers and I find the historical trends in your paper to be fascinating, really.

How did the Air Force end up with a maintainability crisis in the mid to late 1950’s?

Dr. Karafantis:        After World War II, the Air Force, which was newly formed in 1947… the Air Force found its cost and complex systems rising and this was happening while simultaneously its labor and talent pools were shrinking.

Attrition was high, particularly for skilled workers; they couldn’t keep people enlisted because these workers would receive their training in the military and then choose not to reenlist and instead get a far higher-paying job in private industry.

So, between 1959 and 1961, the Air Force lost more than 18,000 workers who worked in maintenance specialties that were considered highly technical.

Gridium:             Yeah, that is remarkable, and I understand why the Air Force would kick off a study. And in 1961, as I understand it, they found that a technological system can cost anywhere from 10 to 100 times in maintenance as it had in initial procurement.

So, if that’s true, wouldn’t it seem that maintenance would itself receive a ton of special attention?

Dr. Karafantis:        Right, you would think so, and it did begin to receive some more attention from the Air Force, but by the late 1950s, early 1960s the Air Force was publishing reports on maintainability and on human factors and these were promoted as ways of keeping costs down.

Gridium:                 Yeah. Well, I wonder if history is repeating itself. In the built environment much research has been done to show that building engineers are nearing retirement and the pool of new talent is shallow.

What did the Air Force do when it recognized that it desperately needed to reconfigure its maintenance staffing?

Dr. Karafantis:        Well, the Air Force and all military branches had always paid quite a bit of attention to selection and job placement, such as testing recruits for intelligence and aptitude for particular tasks and so on.

But, in the 1950s, the Air Force created new Air Force specialty codes and career tracks that were exclusively dedicated to very specific maintenance positions. Now, they also however, tried to reengineer tools and the physical environment in ways that would enable more people to be qualified for these new types of positions.

So, this lead to the Air Force’s interest in human factors engineering because they were crudely trying to idiot-proof operations so that they could use whatever stock of personnel they could muster.

Gridium:                 Is it also true that the Air Force worked on improving the maintainability of the systems they purchased by setting up guidelines?

Dr. Karafantis:        When you look at how developed the field of human factors engineering is today, these books of guidelines would seem very crude in their content, almost so obvious as to be insulting, in my opinion.

But, they were new for their time, and they were basically do’s and don’ts for designers.

Things like: this is how to design to guard against pinched cables, or these parts should be placed in this fashion so that the maintenance worker can access them easily…things like that.

Gridium:                 The aviation psychologist, John Flannigan, he advocated for the technicians.

Do you think that the technological system designers are at all convinced when they hear about the benefit and the importance of thinking about the maintenance worker in their designs?

Dr. Karafantis:        (Laughs) I’m not sure about that.

I get a distinctly smug impression from designers from the research that I’ve conducted.

Flannigan wanted designers to consider just even the mere possibility that the technicians could be assets with expert judgment that was gained through experience, and I second this sentiment.

Why not capitalize on the knowledge of the people who actually use the equipment, right?

Gridium:                 Yeah.

Dr. Karafantis:        So, I mean you do see some give in favor of trusting operators more in today’s version of applied human factors, but in the 1960s the designers had largely convinced themselves that they were the ultimate and even sole experts.

Gridium:                 Is it fair to say that the overall shape–and I got this impression from reading your paper–of the Air Force’s maintenance journey was one with complex systems and fewer and fewer resources to dedicate towards maintenance?

Dr. Karafantis:        Certainly, and there’s almost a conflict here because you were having a decreased personnel level which is limiting the Air Force operation in some ways, but in other ways there is an ambition to cut the amount of personnel they need to have, in the first place, to keep costs down.

But yes, to go back to this question about complex systems, after World War II, the complexity of systems rose dramatically and you didn’t have the skilled workers that you needed in order to maintain most of many Air Force’s operations, so it was a time where they were trying to find ways to get around this problem and that’s how efforts such as commissioning studies for maintainability and human factors engineering came into play and into the forefront for the Air Force.

Gridium:                 Was the Air Force bashful about trying to get to a place where they needed fewer maintenance techs?

Dr. Karafantis:        No, they were pretty unabashedly willing to say we would prefer not to have as many workers and automate as much as possible.

And to a certain degree the designers–this is going back to the more smug designers that I mentioned before–would like to say that we should automate. If it were possible to automate everything then they would choose to go in that direction and eliminate any possibility of human error.

What I think is problematic about that type of thinking is an automated system is designed by a human, so even if you have a complete automated system you’re always going to have an element of humanity within their creativity and the opportunity for error.

I don’t even want to get started on a list of errors that have occurred due to automation, even within the military systems.

Gridium:                 Right. And, it doesn’t look like we’re ever–speaking of which–that we’re ever quite done with this challenge.

I just pulled from the recent headlines that the Navy and the Air Force are both currently struggling to resolve hypoxia problems in F-18 and F-35 fighter jets.

Admiral Scott Swift, the Pacific Fleet Commander, stated something along the lines of the complexity of aircraft, human machine interfaces, and the unforgiving environment that they’re operated in, will always tend to have issues between operations and human physiology.

Are you surprised to hear the Pacific Fleet Commander talk about this?

Dr. Karafantis:        Right, it’s pretty bold to be making that admission.

I find that part to be a little surprising, but the fact that we’re always going to be having episodes that involve accidents or failed equipment or human error or that certain interfaces are still needed to be perfected.

I mean… I think those are just obviously things that are still under investigation and production, especially within the military, trying to find new ways to mitigate error and to enable people to operate machines safely, including these jets. But no, I’m not surprised that these are still ongoing concerns, but more surprised that he’s willing to admit it.

Gridium:                 Yeah. And that interplay between human and machine design and maintenance is such a fascinating one.

Kim Stanley Robinson–one of my favorite science fiction novelists–wrote about a world where a spaceship managed by an artificial intelligence and equipped with versatile 3D printers might one day be capable of maintaining itself.

Could you ever imagine a world where an aircraft carrier or a stealth bomber or a space station might someday be capable of something like that?

Dr. Karafantis:        Right. I mean, there was a space station like that, it was called the Event Horizon. (Laughs)

No, this is funny and it reminds me that just recently here at the Smithsonian we’ve changed our name of our Internet to Skynet.

Would you be able to have these machines that are capable of maintaining themselves? Sure, I mean, to what extent? But then it becomes more philosophical, at least from my perspective, to what extent do you want that automation?

You have also some kind of moral obligations to think about. Is this going to free people to do more highbrow acts or would it maybe even take away that sort of–you mentioned

Zen and the Art of Motorcycle Maintenance. So, would this maybe even take away the Zen of conducting maintenance tasks that I think a lot of people, including myself, find.

So, possible? Yes. But, desirable?

I’m really not so sure about that. Also, this reminds me of a discussion I’ve been having every once in awhile with all these self-driving cars. People have been asking me: do you think one day we’ll be able to have a commercial airliner just have the plane fly itself?

I tell them we do have that, but at least with my generation and older…I don’t think there’s going to be that psychological buy-in where you’re comfortable without having a pilot in the pilot seat.

There’s something about having–even if it is a fully automated process–there’s something about having a person that is even just monitoring the situation, so there’s a psychological component to automation that I think we’re wrestling with a lot these days and it’s pretty interesting stuff.

Gridium:             Well that is a lot to think about and I want to say I found this conversation quite interesting. Thank you so much, Dr Karafantis, for taking the time today to chat with us!

Dr. Karafantis:        You got it, Millen, it was my pleasure.

Gridium:             Great.

About Millen Paschich

Millen began his career at Cambridge Associates, trained in finance at SMU, and has an MBA from UCLA. Talk to him about bicycling, business, and green chile burritos.

2 replies on “Color-coding complex maintenance”

  1. A wonderfully insightful interview about human-machine interactions and design in the 20th century. Thanks so much for publishing!

    1. Thanks for letting us know!

Comments are closed.

You may also be interested in...