Every once in a while I read something from an animal rights group that basically says that all drug testing on animals can be replaced with mathematical/computer models. I have to wonder what background they have in computer modeling and biological systems. I’ve seen the effects of inaccurate assumptions used in computer models, and they’ve cost millions of dollars, years of effort, and the loss scientific data. A bad assumption in a biological computer model could cost a life.
I spent 14 years of my professional life flying one-of-a-kind science satellites. I got payload operations requirements from the scientists, planned spacecraft operations, monitored vehicle health, wrote command sequences, and so on. Our department was getting ready to launch two satellites within a couple of months of each other. The satellite I was preparing for was to be launched on the standard version of the launch vehicle, while the other satellite would be launched first on a stretched version of the rocket. The standard version had been used a few times in the past, while the stretched version would be on its maiden flight.
The satellite flying on the stretched version of the rocket was interesting in that the payload planners used dice. And not just the kind you find on the craps tables in Las Vegas. They also used 8-sided and 12-sided dice. There may have been some others, too – I don’t remember. The payload scientist wanted his instrument placed into random configurations each orbit, and what better way to get a random configuration than to throw dice? The phrase “has anyone seen the planning dice” was sometimes heard around the complex during shift change.
The big day for the launch of the stretched version of the rocket came. A group of us were watching live video from the chase plane used to monitor the launch. During the flight we noticed that the rocket seemed to “porpoise” in flight. The rocket contrail had a definite wave shape. Then pieces of the rocket started falling off. Several million dollars worth of spacecraft pieces ended up on the bottom of the ocean. Scientists who spent years developing instruments were in tears. My satellites launch was put on hold until the cause of the failure could be determined.
During the hiatus I met with some of the engineers working on my satellite. They were privy to the launch failure investigation and told me what led to the crash. When the company decided to build a stretched version of the rocket so that they could launch heavier spacecraft, they decided to skip wind tunnel testing and instead use a computer model. They felt that they already had a proven version of the standard rocket. The engineers had done complete wind tunnel testing on the standard version and had developed computer models that matched the wind tunnel testing. Modifying the computer model to account for the increased size of the stretched rocket would be easy, they wouldn’t have to verify the results in the wind tunnel, and they would save development time and money. Somewhere along the line they made an assumption that didn’t match reality. The result was a launch failure.
Biological systems are complicated – they are far more complicated and less understood than the aerodynamics of a rocket. If engineers who are intimately familiar with their launch vehicle can make an error in a computer model that causes an accident, how much more likely is it that an error in a biological model will mask problems with a new drug? Animal testing is required because a human’s safety is too important to risk on an inadequate computer model.
Stumble it!