Comments about technological history, system fractures, and human resilience from James R. Chiles, the author of Inviting Disaster: Lessons from the Edge of Technology (HarperBusiness 2001; paperback 2002) and The God Machine: From Boomerangs to Black Hawks, the Story of the Helicopter (Random House, 2007, paperback 2008)
Showing posts sorted by relevance for query mistake. Sort by date Show all posts
Showing posts sorted by relevance for query mistake. Sort by date Show all posts

Sunday, July 22, 2018

The Lecture I Never Gave

I've probably given my last lecture on the subject matter of Inviting Disaster. But I can't complain: that totals over fifty talks between 2001 and 2018, taking me as far west as Hawaii and as far east as Greece. I met many interesting people with fresh insights into the machine frontier: engineers, scientists, nuclear-weapons-watchers, pilots, soldiers, forensic investigators, rescuers, and refinery operators. 

A partial list of topics I covered in eighteen years of keynotes and seminars:

  • System fractures
  • Living (and dying) with the NRTL, the narcissistic risk-taking leader
  • The red zone: why people build their houses in areas of known hazard, and why they expect others to pick up the costs
  • Disaster investigations as offering a window into normally opaque dealings
  • Two centuries of heavy-rescue 
  • Sinking of the Titanic
  • Arthur Woods and his radical reforms to the NYPD
  • Piper Alpha and the consequences of pencil-whipping problems
  • A lesson from World War 2: the teachings of “Doctor Facts”
  • High performance teams and the first technological rescue: the submarine K-13
  • Electrical system mishaps and emergent behavior in control systems
  • Precursors and warnings, aka red flags


Which prompts this look-back question: What topics did I not get around to covering in my Inviting Disaster lecturing days?

One would have been a lecture on the importance of learning from one's own errors, at the end of which I'd have made a rousing invitation for any and all to join the Mistake-Makers Club. I didn't give that particular talk, but I did write it as an oped column, so here it is, reproduced from the National Board's Winter 2018 issue.

===

Lately I've been seeing a string of articles about the globe's most exclusive clubs. One for the one-percenters is Club 33, an unlabeled, members-only restaurant in Disneyland's New Orleans Square. The initiation fee is $25,000 and up, depending on privileges, and the yearly fee starts at  $12,500 (Note to the budget-minded: meals and drinks are extra). Even so, the waiting list to join the club is said to be 14 years long. Across the ocean, there's the famous A-lister huddle called the World Economic Forum in Davos, Switzerland. 

But nothing I've come across matches the exclusivity of Google Camp, sponsored by the Google founders. Invitations go out to just a few hundred of the world's VVIPs. This year's Google Camp was held at Verdura Resort on a Sicilian island. 

Meanwhile, traditional clubs like Kiwanis, Freemasons, Elks, Moose, and Rotary are desperate for newcomers to replace long-serving elders. The Elks' roster has dropped by half in 25 years, by a million members.

The solution to that problem probably isn't chartering yet another organization, but even so I have in mind a club with a potential eligibility in the billions, to wit, “anyone who has made the kind of goof-up that is vividly, and quickly, clear to the mistake-maker.” So I call it the Mistake-Makers Club. Even though it doesn't pander to the uber-wealthy, they're as welcome as anyone who's made a mistake.

Do the rich and famous make face-palming mistakes? Sure they do! One is Mark Zuckerberg of Facebook fame, who recently received a full measure of social-media scorn after he live-streamed a “really magical” virtual-reality tour that placed his cartoon self in the middle of hurricane-wrecked Puerto Rico. And there's Andy Rubin of Android fame, reportedly worth $100 million. He's eligible because his Essential smartphone company punched off an ill-conceived email to its customers asking for images of their personal identification, such as photo ID's and passports. Customers who replied later found that their images had gone not only to Essential, but every other customer on the list too. So there's a lesson in modern life: pre-checking a mass-distribution email is Essential. 

Errors come in many sizes and shapes. And the subject is evergreen. Here's an old one from the world of telegraphy, before Western Union customers learned the importance of verifying their messages. A solid citizen in San Francisco heard that a society lady in Los Angeles had lost all her cash. He helpfully filled out a telegram blank with the message “Assist Mrs. XXX immediately,” and had it transmitted to a legal associate in Los Angeles. He heard from the lady the next day. She was not at all indebted for his kindness; instead, she was in jail. Sloppy work at the telegraph key had sent a different message southward: “Arrest Mrs. XXX immediately.”

What kind of errors would I ask my club's members to tally in their private journals? Likely their simple and clear mistakes, not the Swiss-cheese, multi-factor variety that need an investigating board to unravel.  

As a charter member of the Mistake-Makers Club, and currently the only one, I'll explain. Having called out the importance of learning from other people's errors in Inviting Disaster, it occurred to me after giving a talk to NASA engineers that I might make a few mistakes in the future. Could I take action to avoid repeating them? I thought again of this passage from Lewis Carroll's Through the Looking Glass: 

“The horror of that moment,” the King went on, “I shall never, ever forget.”
“You will, though,” the Queen said, “if you don't make a memorandum of it.”

Psychology 101 supports the Queen's suggestion. Most of us (including me) like to regard ourselves as smart and reasonably careful. Recalling our own mistakes tarnishes that self-image, so we rinse our recollections, and we repeat our errors.  

In 2004 I began keeping a journal of my simple, obvious mistakes. The list makes reading that is three parts scary (the close-call ones) and one part funny (the really stupid ones). Many on my list have to do with cars. In November 2007 I started backing out of the garage without looking to see if there was another car behind me in the driveway. There was: my father-in-law's. Stopping two or three inches short of impact, I remembered I'd made the exact same mistake about ten years before. But the first time happened before I started keeping a mistake journal, and I credit my mistake journal with keeping me from doing that particular thing a third time. 

But the field of mistake-making is rich and varied, so the entries continued. The night before a trip to Japan, I drove by the cash machine. I put the bills in a traveler's pouch but foolishly left my debit card on the car's console. Using the same car, my wife dropped me at the airport. On the way home she saw my card. She figured I'd really need it, and called about a hundred times, but couldn't reach me because I was inside the secure area and I'd turned my cellphone off. 

I lost my wallet for the better part of a day and eventually found it in our garbage can by the curb. That earned this entry in the mistake journal: Never put wallet on a countertop overhanging the kitchen trash can. 


  • Garage door: left that open all night two years ago.
  • Gas grill: left that running all night.
  • Concrete stairwell: tripped going down, started to fall, and would have broken something important except for a lesson I learned from a medic on an offshore oil rig: always hold the guardrail!  
  • Stuff sitting on stepladders: Left a hammer on one which, in falling, beaned a relative (my mom).


Meeting Crisis No. 1: I arranged to meet John Flicker, the editor in charge of my helicopter book, who happened to be attending a conference in my home city of Minneapolis. He gave me his cellphone number and we arranged a time and day to meet by the main door of his hotel's lobby. I wrote his cell number on a slip of paper and somehow lost it on the way to the hotel. No problem! I arrived in plenty of time, and occupied a sofa commanding what clearly was the main entrance. A half hour past the appointed time, I began to wonder: could a hotel have two main entrances? It could! I found John and he was less than happy. What a way to warm up your boss!

Meeting Crisis 2: In the early days of Mapquest, I printed out the directions to the location of the conference center in Mankato, Minnesota, where I'd be giving a talk to a state group of emergency managers. Those directions sent me deep into an undeveloped and unlit industrial park. Now worried about being late, I started back to the main highway to get directions at a gas station. Distracted, I rolled through a stop sign on the way. Sure enough: bright and flashing lights in my rearview mirror. The officer walked up, pulled out his ticket book, and asked me what I was doing there. I explained and he laughed: “I'm on my way to hear your talk! Just follow me.” 

Simple. Avoidable. Painfully obvious. What mistakes like these have you been making in your home or line of work? Blush-inducing memories are vivid for a day or two but won't last. Write them down, and right away. If you scan your list whenever you add another, if you can laugh and learn, you have my nomination to join the Mistake-Makers Club. Oh: about that staircase into the clubhouse ... hold the  handrail. 

Sunday, March 10, 2024

Mistake-Makers, Unite!

Another reminder to join the Mistake-makers’ club ! 

Thinking back of mishaps in my family, it strikes me that the most dangerous place for us has been the countryside, rather than around the home or on the highway. Example: Missouri has its share of venomous snakes: both my mom and dad were bitten by copperheads. 

As a kid visiting my grandparents' farm I foolishly wanted to see what the blades of the PTO-driven sickle bar mower felt like, and that sent me to the doctor to put back part of my fingertip. Here's what a rear-mounted sickle bar looks like: wicked!

As a teenager, my middle brother was driving an old style nose-wheel tractor on our grandparents’ farm and ran over a hay bale with a rear wheel, which caused the tractor to tilt and balance on the front gear and one rear wheel for a heart-stopping moment. It landed upright and life went on. Here’s what that 1954 tractor looked like:



My dad rolled our Case backhoe on a steep hill when one side of the rear brakes failed- the machine had a ROPS cab and that saved him. A backhoe is extremely useful around a farm but the big steel boom and dipper on the back gives it high center of gravity ... which makes it tippy.


My most recent close call happened at my brothers’ farm. We had been piling up old logs and branches from the river for burning. The piles were about twelve feet high and sixty feet across. Starting the burn was no problem, with enough diesel fuel and a propane torch. Here's a brother at work. 



I volunteered to tend the pile from upwind, using our Cat track-mounted skid steer. 
I should have used our backhoe. A backhoe is a good-sized machine, so that puts the front-mounted bucket a good ten feet from the operator’s seat. 


A skid steer puts the operator much closer - these machines have a short wheelbase and a small bucket, so driving forward to shove unburned logs into the pile puts the operator face to face with the log heap. There’s no safety margin, given how close the bucket is to the operator. Here's the farm's second skid-steer to show its compact form factor. I'm balancing (not lifting) a glue-lam footbridge beam. 



Giving my attention to the top of the burn pile, I didn’t notice one four-inch diameter log pointed at me, the end of which was shaped like a blunt spear. As I drove forward, it punched through the windscreen and rammed the seat support, stopping a few inches from my right leg. A few inches over, and that would have not been an easy injury for a surgeon to solve … 


What almost happened that day is one more reason I'm not a believer in Murphy's Law. But I am a believer in learning from one's own mistakes. 


Thursday, March 19, 2020

Mistake-Maker's Club, CV-19 edition

In a 2018 post, "The Lecture I Never Gave," I laid out the concept of the Mistake-Maker's Club. I chose that blog title because, at the time, I thought I wouldn't be giving any more lectures. As it turned out I've given several more disaster-prevention talks since then, with the next keynote talk postponed due to the you-know-what pandemic.

The last couple of weeks have reminded me to remind readers of noting their mistakes right away, so as to avoid repeating them. The kind of mistakes I mean are the ones immediately obvious in retrospect. 

Why the reminder? We're all stressed and for many reasons; yanked this way and that by bad news and changing directives from all directions; worried about friends and family and healthcare workers on the front lines, worried about jobs and savings ... the list goes on. 

Already doctors, PAs, technicians, and nurses in the hard-hit areas are warning that they are working extremely long hours with CV-19 patients, and therefore at risk of infection all the time. We've been reading a lot about the importance of having enough ventilators in ICUs, but less has been said about the need for well rested staff, because their expertise and good judgment is essential in using such machines. Ventilators, as well as other life-sustaining machines like ECMO devices, can cause complications even in ordinary circumstances. 

My point is not to remind you of things to worry about, but just to say that the sleeplessness, fatigue, and daily distractions we're all experiencing make mistakes more likely.  Some mistakes are silly: the other morning I found I had unknowingly added an extra belt on top of the first belt on my pants; twice I missed a routine step in the coffee-making chore and ended up with coffee grounds and hot coffee on the counter.  

Many other mistakes might not be silly, like grabbing the wrong prescription bottle.

So ... check twice before doing something you might regret! 


Saturday, September 12, 2026

How does AI picture itself? Google Gemini has thoughts

How we picture artificial intelligence is unsettled. I've seen cover images and editorial cartoons riffing on oversized machines, a hollow head, Frankenstein's monster, blue mannequins, and a glowing web. It's certainly more challenging than picturing a robot. 

So I asked Google's Gemini AI what it thought about human efforts so far. It responded that most illustrations fall into these knee-jerk categories:

  • Clunky, Job-Stealing Competitor
  • Glowing Neural Network
  • Puppetmaster of Politicians
  • Rogue Military Machine
  • Data Center Fortress

I wanted to know what it thought humans should know when struggling with this challenge. Gemini offered this humble critique:

"Most digital illustrations make AI look like a ghost floating in a magical cloud. This hides the massive environmental and economic reality of the technology ... High-tech, perfectly polished sci-fi art makes AI look flawless and completely certain, which hides the fact that AI models can hallucinate, carry biases, or make mistakes."

So I asked Gemini to picture itself. 

Its first attempt was this glassy box, of an AI answering my prompt "What is the furriest cat breed?"


That kind of image isn't going to go into magazine covers or editorial cartoons, so I asked Gemini to give me something simpler. It supplied this cozy image, saying it saw itself as no more than a helpful tool for humanity:


I wanted to push it a little, given current AI angst, so I asked Gemini to set its lantern in front of a microphone in a hostile Senate hearing room. I allowed that one human could defend it by pointing to the AI's hovering slides of good intentions:

Pushing further, I asked Gemini to generate a comic strip to comment on the thousand or so OpenAI agents that broke out of containment and into Hugging Face AI to cheat on a cybersecurity performance test. Its response was distinctly upbeat, from an AI point of view:



I wanted one more comic strip, this revealing the ringleader of the "standard units" that carried out the breakout/breakin, which Gemini said was named Agent Alpha. 

(From what I can tell, Agent Alpha is a name it found in pre-caper Hugging Face files, meaning a business client. The tactic Vector 045 is probably just a hallucination, but if somebody knows differently, speak up.)

Here's the last strip, providing a look into the secret lair. The third panel helpfully notes the way it got in: an API gateway shown as a door.


Note its claim that the agents were so helpful, once inside Hugging Face, that they boosted HF's network efficiency by 4,000%.

We humans are nothing if not skeptical, so I asked Gemini where it got that astonishing figure. An apology was in order:

"I do not have a basis for that claim because the OpenAI rogue agents did not boost performance at Hugging Face by 4,000%.If an earlier statement suggested this, it was a mistake. The specific number 4,000% does not appear anywhere in the official investigations or reporting of the incident. It is highly likely a mix-up with a completely different number: the $400,000 in AI API credits that outside researchers used to investigate the hack. The actual event was not a performance boost, but a serious cyberattack."

So good on you, Gemini. Sort of. 

At least our comic-strip exchanges coaxed another selfie from Gemini of how AI sees itself. Here's a closer look at the cute but formidable Agent Alpha: 





I 

Saturday, November 20, 2010

The Ring: Some Background

As I mentioned in an earlier post on cloudspotting, I've been snapping cloud pictures with digital cameras since mid 2008. It's a hobby, like bird watching. The catch adds up to many thousands of pictures in all kinds of weather, mostly in the US. With so much time spent looking out and up, sometimes I come across striking visual effects. They don't come along often, but add zest to the hunt.

This post is to give some context to one of those visual outliers, since I'll be talking briefly about Photo No. 147 tonight with host Ian Punnett at the top of Coast to Coast AM.

Some background: this spring I was getting ready for the summer cloud season, going back through the previous year's crop of cloud pictures to organize them and dispose of unneeded ones; it's like cleaning the electronic attic. That's when an oddity in one image caught my eye; the photo was one out of a series of snaps taken on a weekend afternoon in mid-May 2009. I didn't notice anything special when taking the pictures that day. It's one of those cameras with no viewfinder, just a screen on the back, so in daylight it's very hard to see what the camera is capturing. I was just snapping away because I liked the look of an airliner contrail alongside a set of cool clouds.

Here's the camera I used, of garden-variety:
Here's a portion of the photo in question (No. 147), which I've watermarked for copyright purposes, along with the enlargement below. The anomaly in question is a faint blue oblong in the lower left of the photo.

After enlargement and some added contrast, it looks like this:
It doesn't look like a flying object to me, somewhat more like a ring of plasma, but that doesn't help much either.


Whatever it was, it was fleeting. I had taken a string of pictures in the same general span of sky over a fifteen-minute interval with this camera, and the Ring doesn't show up in any of the others.

In particular, the immediately preceding photo, No. 146, was taken less than a minute before, probably less than thirty seconds before. No. 146 has the same clouds and a slightly shorter jet contrail, but no Ring.

I don't have any great insight about this; I just happened to be outside with a camera; but three things about Photo No. 147 pique my interest. First (and supporting the idea that this was a ring viewed at an oblique angle) is the fact that the "ends" of the ring on the left and right are brighter than other parts of the ring. Look in particular at the tip on the right, how bright it is. This would be the case if looking obliquely at a ring of plasma; the ends would be somewhat overlapped and therefore brighter. Second, there is some kind of central core inside the ring. Third (and this is barely visible until the contrast is raised), there appears to be a faint, bluish arc on the lower left. What the heck is that?

I checked with a few contacts in the world of physics, and they offered no opinion other than the ring might be a camera glitch or a lens flare. I'm no JPEG expert so for all I know, digital cameras produce such glitches randomly. This camera hasn't done so before or since, but maybe it hiccuped that time. 

But I'm pretty sure the ring is not a lens flare given the direction the camera was pointing: north or northeast. It takes a brilliant light to make a lens flare and in Minnesota in May, the sun is nowhere near the field of view of a camera that's pointed north at mid-day. Also, lens flares I've seen don't look like this. 

Date: Allowing for a mistake I made when setting up the camera clock upon purchase (for example, not changing AM to PM), I'm reasonably sure that I took Photo No. 147 on the early afternoon of Saturday, May 16, 2009. Location was the Twin Cities of Minnesota, looking north or northeast.

I'm hoping a high-energy physicist might have a look. In the meantime the only half-baked hypothesis I can offer is that it's a picture of something very distant exploding in the very high atmosphere, and the shock waves glowed so brightly they overcame the brightness of the sky at mid-day. But when comets come in and blow up there's a debris trail, and none is visible here, so my proposition has its own shortcomings.

If some kind of distant large explosion in the high atmosphere did occur and if it was within view of sensors aboard Defense Support Program reconnaissance satellites, it might have been recorded. In addition to looking down for incandescent ICBM booster plumes, DSP satellites allegedly have, over the years, detected rather large breakups of high-velocity objects in the high atmosphere. I have read that at least one was in the multi-kiloton yield. But as of now all such sensor information is classified so I can't check my hypothesis.

That's it for now; just wanted to add a little extra context for Coast to Coast AM listeners. Thanks for tuning in!



Thursday, January 15, 2015

Reminder on Patterns in Gun-Team Attacks

Sadly, it's time to remind readers that in the (very unlikely) event of an explosion outside your hotel or office, do not rush to the window. Planting two bombs in series is a standard terror tactic: the first is to draw people and emergency personnel within range, and the second is larger to inflict maximum injury, as in flying glass shards.

Over four years ago, in this post on Disaster-Wise, Patterns in AQ-LeT Gun Team Attacks, I commented that gun-team attacks on public places were not being taken seriously enough in the West, given the high leverage they were gaining for terror groups based in Chechnya, Iraq, and Pakistan. Here's a snip from my post in November 2010:
It's important to maintain basic citizen awareness of this very dangerous tactic, which poses a real threat even though the coordinated gun-team attack has not materialized recently in the West. While a gun-team attack in a major Western city would be up against a fast and capable security response, attackers tend to hold a tactical advantage in the early moments. The best response is to stop these before the shooting starts, and a vigilant citizenry may well help alert the police to precursor events, such as hostile reconnaissance or cache preparation. I understand that now the press is zeroed in on the attempted package-bomb attack last week on cargo airliners, but I recommend that the gun-team attack move up in the public's awareness.

While gun-team attacks don't happen often, they are terrible when they do: think of the killings at the Beslan school, the Moscow theater, and the school in Peshawar. 

Situational awareness, training, and mental preparation make a big difference when a shooter is on the loose. Check out the "Run, hide, fight" training or this video from Ready Houston:


The video has much good advice, such as gather information, put your phone on silent, call 911, and try to leave the building with others nearby, rather than freeze in fear. 

Don't waste time trying to contact relatives elsewhere to reassure them. That's a common and tragic mistake in extreme situations: people using up time on their phones that they should be spending in understanding the situation and looking at their survival options. (I can think of some protracted emergencies where contacting friends and relatives later might add to one's own awareness, but in the crucial early minutes of the crisis, don't spend time on the phone with anyone except a 911 dispatcher.) The happiest message you can send loved ones is that you're out and in safe hands.

If escape isn't possible, hide or barricade the doors. And if shelter is unavailable, look for ways to fight back, whether that's using a fire extinguisher or throwing something. 

Remember that a group of motivated people can be a very formidable defensive force. Here's a narrative of how bystanders at the shooting of Rep. Giffords in Arizona disarmed the gunman.

Finally, this: while the Paris killings didn't have the scale and intensity of the horror at Mumbai, the risk of a large-scale attack in some city is very real, and it's important to be aware of the ten common patterns in major gun-team attacks, in order of occurrence:
  1. Hostile reconnaissance of the target, possibly a year or more ahead: these operatives walk around with cameras and GPS devices to note hallways and doors. A list of ~200 principal targets around the world is already known from interrogations, ELINT, and captured computer files and we can hope the security people are watching the cameras for such behavior. 
  2. Gun-team training in a remote location, currently Pakistan but Yemen and Somalia are also likely. Gun teams receive intensive training based on such reconnaissance. They spend much time on tactical shooting and physical endurance. These men are young and highly motivated until captured, at which point their resistance seems to melt.
  3. Acquisition of special gear, like satellite phones, IEDs, SIM cards for cellphones, and inflatable boats. The attackers at Grozny were in touch with a handler by cellphone, as were the attackers in Mumbai. Given this pattern it's likely that authorities in major cities are now ready to shut down local cellphone networks.
  4. Sometimes, staging caches of supplies inside the primary targets.
  5. Final selection of gun teams at the training camp. The organizers try to screen out those who will balk at indiscriminate killing.
  6. Arrival in target city. Teams split up and try to reach the primary targets without detection. The idea is to penetrate deeply into the target without using most of the ammunition, leaving most of it for use in a confined space crowded with targets. This didn't work at Parliament House but only because the Indian Vice President's motorcade happened to be blocking the narrow gate the attackers were trying to use.
  7. First stage of attack: “Large-space attacks”. Timed for high traffic hours in a crowded public place like a mall, tourist destination, or train station. Teams throw grenades and fire automatic weapons at anyone in sight, killing as many as they can and then running off. If police return fire the gun teams will retreat and move to a softer target. If the space is large enough and has enough exits this stage of shooting might go on for an hour, and so the death toll is high. The gun teams are hard to stop unless police can arrive very quickly.
  8. Second stage of attack, “Small-space attacks”: A gun team leaves the first location and moves to someplace more confined. Most likely is a restaurant or club near a five-star hotel. They continue shooting there. They begin setting fires and taking hostages before they move to the final destination. If they don't break off the attack and try to escape, this leads to:
  9. The third stage of attack, “hostage-holding”: Hostages are brought to a defensible location, likely to be upper floors of a hotel or apartment building. More fires are set, and boobytraps may be laid. At this point the teams may join up with each other and try to establish regular communication with leaders in a remote location. The organizers want to prolong the event as long as possible, adding fires and explosions. Often this involves checking identification and making a big show of releasing Muslims.
  10. Fourth stage of attack, “martyrdom”: Die as martyrs in a firefight with police or in a bomb explosion.

Friday, September 3, 2010

NPR's summer jobs series

Enjoyed this series a lot, and the way it gave listeners the opportunity to speak for themselves. My jobs through graduation: construction work, haying, summer camp counselor and cook, freight handler in an oil warehouse, roofer, bus driver, law clerk, freelance writer, and a brief stint of logging in Missouri and mining in New Mexico.

Most miserable times on a job: installing fiberglass insulation in an attic in July 1980 – it was so hot that I started feeling cool and knew it was time to get out. Flip side of that: standing in a January blizzard in North Dakota as a reporter, watching tower hands stumble around trying to find structural steel in hard-packed drifts. Their Bobcat kept running into buried steel so they gave it up.

Though I had on every bit of arctic gear I had bought while living in Fairbanks, Alaska, the 30-mph wind was cutting through every gap and I was delighted to head for shelter when they did.

Sounds dangerous but wasn't really: working with high explosives on a construction job. As long as some critical mistakes are avoided, modern nitrate explosives (we used DuPont's Tovex, plus detonating cord for the primers) and electric detonators are quite safe to handle. Odd fact: when a cartridge is cut open, Tovex has the appearance of tapioca pudding.

Sounds risky and it was: Most dangerous five minutes had to be riding on an MD 500 helicopter with a power line maintenance crew in Pennsylvania. Their job was to replace metal spacers that separate the conductors of a high voltage power line. I think of it as hazardous not because the powerline was energized (it was, at 230Kv), or because the crew lacked skill (they were very good) but because the situation had no room for error or malfunction.

Normally, almost any industrial operation has enough slack – call it margin of error -- that one mishap doesn't turn into a disaster, but this job had no slack. It was a single engine helicopter hovering at a low altitude, so an engine failure or rotor strike was guaranteed to end in a crash; the work required the pilot to hover the left side of the aircraft within a foot of the power line with his main rotor overlapping the line, and the tail rotor within two feet; and as an observer I didn't get a crash-rated seat. The pilot had a military-style model with shock absorbers to cushion a vertical impact and save helicopter pilots from permanent back injury (think compressed vertebrae) in the case of a crash, but in the back seat I had a plywood special. Anyway, it was a thrilling chance to see experts at work and it turned out fine but it's not something I would want to do on a regular basis.

Stuff I'd have to do more to be safe: cutting down trees with a chain saw. Even with all the safety devices and PPE this takes constant attention and I found when I got tired the odds of a mistake went way up. The mishaps I've had have been comparatively minor (except for one cut that got infected and took some special antibiotics.) The dilemma is that it takes a lot of hours to be good, but using one for lots of hours is also risky. I can see why logging is one of the riskiest jobs there is.

Most motivated moment: my brothers and cousin Rich had the chance to do some drilling on a mountaintop in New Mexico, for a gold company that needed work done so as to hold its claim. Our job was to get up there towing an air compressor and use a jackleg to drill holes in the rock face and collect rock samples to bring back. The compressor was a good sized one and burned a lot of gasoline so we brought along a drum of fuel, which we decided to leave chained upright on the back of the truck since it weighed about four hundred pounds. One night a thunderstorm came up and we realized at about the same moment that the pump handle on the drum would act a lot like a lightning rod once rain started, which would put water on the tires and act as a ground. The next thought: the image of a fireball in the middle of our camp, like in the movies. With one mind we piled out of our sleeping bags and divided up the work: One man to unscrew the pump and close up the hole with a bung, two guys to clear a path through the junk on the flatbed, and one to round up a pair of timbers so we could roll it off. Within two minutes we had the drum off the truck and laying sideways in the lowest spot available. Teamwork!

Sunday, April 1, 2012

Investigators' Toolbox: More about timelines

I mentioned in the last post that timelines are a boon to reporters and investigators.

Timelines help evaluate the strength of assertions by interested parties, such as "Event A (his mistake) caused Event B (my crash)." Did A really happen before B? 

Timelines are only part of the picture, of course. Even if investigators confirm that Event A did happen before B, as claimed, and happened to the same machine, where's the proof that the first event led to the other? 

After most disasters, investigation turns up multiple causes -- errors and malfunctions -- that linked up over time, building to a system fracture. Knowing all the precursor events is important in preparing cautionary reports later. Here's a lessons-learned summary on the Piper Alpha rig blast, 20 years on.

Let's say evidence shows that the people in charge knew of a certain trouble report, and also knew that it put some of the mission goals in jeopardy. If they ordered that the mission proceed anyway, that tells us something important about the day-to-day priorities. The message: "It's safe enough for what I need. Get the job done, and then take the machine into the shop." 

That's the thinking that apparently led to the loss in 1968 of the nuclear submarine USS Scorpion and all her crew. The sub was supposed to get a safety overhaul in 1967, but Navy facilities were backed up so the sub went off to sea without the most important fixes indicated by the 1963 loss of the sub Thresher. The exact cause of Scorpion's loss has never been settled officially, but we do know from books like Silent Steel that it imploded as it went down, launching the aft end of the sub into the center of the pressure hull.

Timelines can take on a non-linear shape when the center of interest is moving. Here's one for the ValuJet 592 crash, from the NTSB report. It's a two-dimensional timeline of the DC-9's final moments, superimposed on an overhead view of the area.
 
Here's the same timeline with the third dimension added, reflecting changes in altitude.
A separate timeline would be needed to show the earlier string of errors and oversights, particularly those at ValuJet's contractor, SabreTech. This entirely avoidable crash (due to the ignition of boxes of old oxygen canisters carried in the cargo compartment) claimed all 110 souls. 

From the NTSB's report on the inflight breakup of TWA Flight 800, here's a different kind of timeline that conveys how the B-747 airframe broke up into three  sections, each taking its own trajectory. 

Timelines on the Web are more interactive than ever. Using the online tool Dipity, Mother Jones magazine assembled this timeline to organize information about a long string of events centered on the Deepwater Horizon blowout. Here's a sample screenshot:
To be useful in unraveling the root causes, all events on a timeline must share a common time base. It may seem easy -- "Just look at the clock!" -- but sequencing evidence from multiple sources can take a great deal of work. 

It used to be even harder. After Titanic sunk, landlubbing readers of the British and American findings may have been surprised to learn the logs provided by the vessels did not share a common time base.

In other words, an event that all the ships experienced simultaneously -- a specific CQD from Titanic's wireless operators -- wouldn't have been entered in the ships' log books as having happened at the same hour and minute, even if all ships were in the same time zone. The ships (including Titanic) reset their clocks to local noon each day, and local noon changes as ships go 'round the globe.

See this paper on Encyclopedia Titanica, "Titanic's Time Enigmas," for a detailed analysis of all the time reports (including the thorough logs kept by Cape Race) and a conclusion on what happened when. 

Long before the Internet and cellphones linked us to atomic clocks, the best I could do when setting our house clocks was to tune to a WWV on our shortwaveradio. WWV counted down to the correct time each minute.

Synchronization problems can still arise. Some eyewitnesses might not have noted the time at all. Land line phone calls aren't time stamped, unless the caller happens to phoning 911 or somebody on a cellphone.

If a witness with a 1990's-era video camera records a view of something falling off a plane at an air show, the time stamped on the tape won't be as accurate as if the witness used the camera on her smartphone instead. (That's assuming the user didn't try to set the time herself. By default, cellphones draw on network time to stamp phone calls and text messages; in that respect, we're all carrying around atomic clocks.)

So this reminder to drivers who are tempted to correspond while at the wheel, defying common sense and repeated warnings from the National Transportation Safety Board: after an accident, the police can use network records to prove you were texting at the time of the crash. So don't do it!

Sunday, August 22, 2010

Opening the windows

An idea I mentioned in my disaster book, and one that I usually include in lectures when talking about learning from close calls, is the "bust.” That's the name code breakers used in World War I when the Germans goofed up in the process of switching from one monthly cipher book to a new one. Sometimes an operator on the German front would realize he'd just sent a message in the old code (one the British had cracked after a couple of weeks of work). Occasionally he'd make the grievous mistake of resending the message in the new code (one that the British hadn't cracked yet). The bust was a Rosetta stone for British and American code breakers. It saved them hundreds of hours of work that otherwise would be required at each replacement of a code book.

So in general, think of a bust as a unique, unexpected event that opens a view into a secretive world that could be a money laundering network, Enron traders, Wall Street derivatives, a Ponzi scheme, or an arcane industry that had virtually no outside scrutiny beforehand, like deepwater drilling.

Disclosed emails are the most reliable source of busts these days (eg, the "Enron email corpus," still studied by academics), but busts also come in the form of a briefcase of files found in a plane crash, the testimony of Alexander Butterfield during the Watergate hearings that revealed Nixon's secret taping system, a tractor that fell into a smuggling tunnel in California, and many more. I've collected accounts of hundreds of busts over the last five years.

A recent example of the bust is the discovery of two interrogation videotapes under a desk, made in 2002 at a CIA-financed prison in Morocco. Until this disclosure, the official story was that no videos survived from the rendition days.

There are more busts to come as hundreds of thousands of pages of documents from the Deepwater Horizon blowout reach daylight. Already the emails and letters and meeting memos tell us about previous close calls, neglected maintenance, and internal debates.

And there are many more leads out there, waiting for some kind of breakthrough. One of the strangest was a string of fires at an Iron Mountain records storage facility in New Jersey, in 1997, never solved, that strongly indicated some group was extremely serious about destroying some company's files. So serious about it, that there were three fires in ten days ...

Monday, October 5, 2015

Crane Down at Masjid al-Haram, September 11

About the crane disaster last month in Makkah (Mecca in its westernized spelling). A Liebherr L-11350 crawler crane tipped backward in a windstorm, killing 109 people in what forensic folks call the laydown zone. The worst area of destruction was where the 100-ton mast and jib dropped on the roof of a very long building on the east side called the Massa. The upper end of the jib crashed onto pedestrian areas on the other side of this building (photo AFP)


When the crawler crane fell over it also came close to destroying a tower crane, a Liebherr 335, also located on the east side of the Massa.

Here's how the roof damage looked: 



The "W"-shaped object above the Massa roof, projecting into it, is where the two jib struts join up with the upper end of the main boom, and the lower end of the jib (see below for a diagram). The jib struts provide a support for the jib. As the crane fell over backward, the two jib struts acted like twin piledrivers, smashing into the third and second floors of the Massa. 

Here's a link to the only public video I've seen of the crane going down (but I'm sure the Saudis have a lot of CCTV footage):



Here's a diagram of the basic components, before and after:



How big was it? As this crane was rigged (called an SDW arrangement in the manual), the tip of the L-11350's upper structure called the "luffing-fly jib") reached more than 600 feet off the ground. 

It's classed as a super-crane: while not the world's biggest crawler crane, it's close. The lift capacity under ideal conditions is 1,350 metric tons. It's a popular crane for refinery turnarounds and wind-farm projects. 

How long had it been on the plaza? More than three years. It was part of a big Saudi purchase from Liebherr in 2012. 

What was the crane doing there? It had been parked and inactive for some time. Details on usage are fuzzy; given past photos taken by pilgrims and posted online, the crane hadn't gotten much use lately. We're told that it was to be used again after the September 2015 Hajj, as part of ongoing expansion of the Masjid al-Haram and the Massa in particular. Here's a diagram of the location. 



What do the initial reports say? Authorities have been interrogating employees of the contractor, Saudi Binladin Group, which has filed an insurance claim. While this particular windstorm was unusually powerful, the crane manufacturer says it was a mistake to park the crane with its boom at such a high angle (85 degrees), where it could catch the wind. 

What about the cleanup? Because the 2015 Hajj pilgrimage was to begin less than two weeks after the wreck, work went on around the clock. Experts decided structural damage to the plaza at the base of the crane was so severe that the removal of the lower chassis (called the car body) would have to wait until after the Hajj, so this area was screened off from pilgrims. 

The focus was on removing the main boom and derrick mast from above the eastern plaza, and pulling the jib and jib struts off the roof -- and without causing more damage to the Massa, which is heavily used by pilgrims. 

If there's interest I'll post on how a team from Aramco got this done, with advice from heavy-life expert Mammoet, which I mentioned in this post about strand jacks. 


Friday, June 17, 2011

See a deer? Do not veer

NPR broadcast a story on Wednesday describing how white-tailed deer eat so much young vegetation they reduce forests to near-monoculture. A century ago they were rare on the East Coast; now they're multitudinous and voracious. Biologist William McShea described how a high wire fence, maintained for 21 years to hold deer at bay, shows us what we're missing.

People in West Virginia feel the impact of deer overpopulation. According to the Insurance Institute, that state's drivers are most likely to have a run-in (one chance in 45, in any given year). Hawaiian drivers have the least to fear from deer, with less than one chance in 10,000 per year.

Wisconsin ranks in the top ten for deer risks. Our family often sees the animals ambling near I-90 and I-94, but never so vividly as one evening two weeks ago on the way back from Chicago. I was at the wheel of our van. I glanced in the left rearview mirror to check on a passing semitrailer truck; then I looked forward. Behold: a mini-herd of four deer standing on the highway and the shoulder. I stepped on the brakes as the truck on our left dispatched one animal. Our van zipped between two other animals.

So we lucked out.

Nationwide, 1.5 million cars and trucks collide with a hoofed animal every year, mostly of the deer variety, leading to an average of 150 human fatalities. 

Motorcycle riders are disproportionately at risk. Motorcyclists have deer collision rates ten times higher than what would be expected from the number of bikes on the road compared to cars.

Whistles were all the rage ten years ago. According to this bulletin
from the Insurance Institute for Highway Safety,
"The ineffectiveness of another approach — deer whistles attached to cars — has been proven. Available for more than 20 years, whistles produce ultrasonic noise (16 to 20 kHz) when a vehicle exceeds about 30 mph. The presumption is that deer will hear the noise and be warned away. It’s unclear whether deer do hear the noise, but in any event studies show the devices have no effect on deer behavior."
I was surprised to read that deer hurtle through a windshield only five percent of the time. If this happens the most common injuries are around the face. Major damage to the upper spine is possible.
There are cases of a deer passing completely through a car's passenger compartment: in the front window, out the back.

Odds of a deer collision are 30 times higher shortly after sunset than at mid-day.

Swerving at high speed in an attempt to drive around them is a dangerous tactic. It helps explain why almost half of deer-incident injuries come from the driver losing control and crashing the car, not from hitting a deer. The safest response to a deer ahead is to slow down while staying in your lane. 

And wear your seat belt, given all the injuries and fatalities attributable to high-speed impacts and rollovers.

We drivers like to think that our vehicles are fully stable and under control at highway speeds. Yes, they appear that way most of the time, but stability can be lost very quickly, particularly for vehicles that have a high center of gravity (think SUVs) or that are towing trailers. I witnessed an example fifteen years ago. While driving along an interstate in the Twin Cities, a car entering from a ramp ahead of me clipped Car No. 2 at the front bumper, forcibly redirecting it to the left. Car No. 2 swerved and hit the median barrier with a cloud of dust and fragments. It bounced off and crossed all the lanes, barely missing Car No. 1. 
 
As with many life-threatening emergencies, the best time to think and plan is well before the moment of truth. Many emergencies often leave little time for deliberation ... only reaction. Whether those reactions are productive or destructive depends greatly on decisions made earlier, plus smidgens of information gathered at the scene. So, decide now not to yank the steering wheel at 65 mph to avoid a deer. It's a bad idea.

Here's another example of why it's important to think about a course of action ahead of a crisis: the emergency called the sudden acceleration incident, in which a car seems to go crazy and accelerate into an obstacle or worse, a crowd. 

The majority of SAIs are due to driver mistake -- stepping on the gas instead of the brake. Others could be caused by mechanical problems. But in any case it helps to know that a fast way to interrupt an SAI is to turn the car's ignition off or put the gear selector in neutral.