Why Safety Takes Time

I suppose it’s really our own fault because of the way we teach science.

We give you labs constructed to get the right answer on the first try.   We have you measure things you already know the value of.   We tell you that things were invented by a single person on a specific date.

We give you an absolutely bogus model of the scientific method, putting it into a nice pretty, linear flowchart that is perfectly designed for multiple-choice tests. (Right, from HowStuffWorks.)

In reality, the scientific method looks a lot more like the figure below than the one to the right.  Visit the webpage to read the text on the flowchart.  It’s worth it because a lot of the boxes lead to explosions and/or lab fires.  It may be NSFW — depending on where you work. If you work in a lab, you’ve already said all those words today, I bet.

This blatant misrepresentation of how science and engineering actually work is probably why the blogosphere and all electromagnetic radiation methods of communication (that would be radio and television) are full of people calling on NASCAR to ‘fix’ their ‘safety problem’ before we go back to Daytona and preferably before Phoenix.

Sorry, folks.  Science and engineering just don’t work that way.  We’re talking about research here – which means we’re dealing with things we don’t understand completely.  You can’t set a timetable for when you’re going to discover something.  It happens on its own schedule.

You may hear about R&D, which is research and development.  Research is figuring stuff out.  Development is making it happen.  Research is “what type of barrier do I need to stop a 200-mph car?”  Development is making the barrier that comes out of the research.

You can’t do ‘D’ before you do ‘R’.  Einstein once said “If we knew what we were doing, it wouldn’t be research”.  By definition, research is trying to answer questions for which we don’t already know the answer.

Who Does Motorsports Safety Research?

The car companies all do extensive safety research — but not at 180 mph.  Their results, which are on passenger cars, can’t be extrapolated to high speed racecars, so there’s not much useful information there for racetracks.

There are a couple university groups that do motorsports safety research, but you can count them on one hand.  There are a few consulting companies that are involved with motorsports safety research (sometimes as an offshoot of defense-department research) but again, it’s a small number.  The FIA is a centralized automotive organization focused on F1 that does extensive research in Europe for open- and closed-cockpit cars.

Safety research is inherently expensive.  Take barrier research.  Start by getting a racecar – which is an expensive proposition to start with – and then figure out how you’re going to accelerate a driverless race car into a wall at a very specific angle without a driver.  You have to set up high speed cameras, high speed and high strength data acquisition equipment and it all has to work right.  The companies that are looking at protective gear for soldiers can get funding from the Defense Department (at least before sequestration they could) and some of those results can be applied to motorsports.  The FIA dips into the coffers of the F1 enterprise.  It doesn’t make sense for auto manufacturers to pay for motorsports safety research, since little of it will impact safety in passenger cars.  That pretty much means that motorsports sanctioning bodies, tracks and teams have to foot the bill for safety research.  And it’s a really big bill.

It’s big in part because we not only have to do experiments, we have to design experiments.  Anyone got a spare F1 car we can crash into a barrier for testing?

I have a great story in my book, The Physics of NASCAR, from Dean Sicking, the inventor of the SAFER barriers about their first attempt to demonstrate to the Indycar people that they could crash an Indycar into a barrier.  It wasn’t pretty.  It also wasn’t successful — the first time.  They learned from their miscues and — years later — they had the first SAFER barrier design.  The SAFER barrier research started in 1998.  The first SAFER barrier was installed at Indy in 2002.  It took that long to figure out what the right design for the barrier was.

Research also never really ends.   They’re looking at how to make portable SAFER barriers for street circuits, SAFER barrier gates, cheaper SAFER barriers for road courses, and how to protect the ends of pit walls.  Even when one problem is mostly solved, there’s lots of fine tuning and specialization to be done.

The Problem with Multi-Car Accidents

With all the computer power we have to do simulations, shouldn’t we be able to use a computer to predict in advance what types of accidents are most likely?  Let’s say you want to create a computer program that would generate all the possible types of accidents.  Start with a single car.  Consider all the places on the track where an accident could happen and the types of accidents (hitting the front of the car first, hitting the rear of the car first, hitting  the pit road wall, getting airborne, etc.).  I bet you could list a hundred possibilities (at least) without having to think too hard.  When you get done with your list, let me know.

Now list all the possibilities for two cars.  It’s not two times the number of possibilities for one car because the two cars can interact in many different ways.

There were twelve cars involved in the Nationwide accident at Daytona last Saturday.  There is no way anyone could go through all of the possible outcomes of a twelve car pileup.  The problem with plate tracks is that they cars are running in a bunch and therefore the probability of a lot of cars being involved in an accident increases relative to, say, Martinsville, where most of the accidents are a few cars at most.

Here’s the other problem.  If I want to simulate a program that shows me how a fence will respond if a car hits it, I need real-world data on which to base my simulation.  So at some point, you’ve got to have a way to propel racecars into catchfences.

The Most Important Data Comes About When Something Doesn’t Work

You don’t learn anything if you don’t get data.  You don’t get a lot of useful data in motorsports safety research when everything works perfectly.  The only time you get real data is when something goes wrong.

The number of accidents at any one track is actually pretty small, so the amount of data we have is very limited.  NASCAR has an accident database where they’ve collected every bit of video, data from crash recorders and anything else they can get their hands on.  Every bit of data gives them a little more information on which to base future work.  But the most valuable data, unfortunately, comes when something doesn’t work.  This is true for every type of safety equipment:  barriers, HANS devices, seats, roof flaps, and yes, catchfences.

Priorities:  Ends and Openings

Most people in motorsports safety will tell you that the most challenging problem right now is building protective gear that is also required to move.  Window nets, for example, need to keep speeding car parts out of the car, but must be rigged so that they driver can lower the window net very quickly if necessary.  Ends and openings are the most challenging aspects of track design. One of the challenges to putting SAFER barriers on the inside of tracks is that there have to be openings for emergency vehicles to get on the track.  This requires SAFER barriers on hinges that can’t be damaged if the car happens to hit the hinges.

Saturday’s accident involved a pedestrian crossing gate – gates are the literal weak links in the catchfence design.  I suspect that if you had the exact same accident happen on a different point of the track, the results would have been very different.  It was the unfortunately perfect storm.  I’ve addressed most of the simple suggestions for catchfences in a previous blog post and those arguments apply here.

Where We Need to Focus

If we gear our efforts to simply preventing the type of accident that happened Saturday, we’ve missed the whole point.  The odds of that exact same type of accident are very small.  Then a different kind of accident happens and now we all run to prevent that specific type of accident.  Remember that a racetrack is a system:  it’s the cars, it’s the drivers’ safety equipment, the configuration of pit road, the track shape and size, and the catchfence.  All of these things work in concert with each other and focusing on any one of them misses the point.  It’s a complicated system and it takes time to figure it all out.

I wish I could say that, at some point, motorsports will be perfectly safe – for the drivers and the fans.   But it’s unrealistic.  There will never be any guarantees.  We can just do the best we can.

Related Articles

• My hometown newspaper, the Milwaukee Journal-Sentinel, has a nice article on safety by Dave Kallman
• Catchfence science after the Las Vegas Dan Wheldon accident
• A HuffPost Live discussion on the Daytona incident, as well as some interesting comments about who owns footage from a sporting event and when a sporting event becomes news.