|
This is a question for Brent and Hamid, but it also addresses the recent thread on reliability and redundancy.
I have been recently educated on the FAA approach to reliability for certified airplanes, which can be simply summarized as a connection between the probability of failure and the resulting consequences. For the aircraft I was working with, a "catastrophic" failure (results in loss of aircraft, loss of all life) needed to have a probability of 10e-9 (that's one in a billion). A "major hazard" (damage to aircraft, loss of a passenger) needed to have a probability of 10e-8. Similarly, lower consequences were allowed to happen more frequently. The key point here is that there was no failure that was prohibited -- even a wing falling off. We just had to make it so unlikely that the risks were acceptably small. Acceptable risks are lower for airliners, and higher for two-person propeller airplanes.
FWIW, this goal was accomplished by making systems -- not components -- more reliable by strategically adding redundancy where it improved systm reliability.
My question for Brent and Hamid is this -- given your comments that an EFIS has a finite probability of failure, what is that number? What would it be for a certified EFIS (like a Chelton) and what would it be for a non-certified system (like a Dynon)?
What I am getting at here is a quantitative way of ensuring that our proposed flight display systems have a reliability equal to that on a certified aircraft. For example, if a Chelton failure probability is one in a million, and a Dynon failure probability is one in a thousand, and the power supplies are separate (Dynon internal battery backup), wouldn't a Chelton + Dynon give you a one in a billion probability of total failure?
By the way, I personally don't believe the failure probability numbers that the systems safety engineers come up with. I believe that the actual rate of failure is significantly higher. However, I *do* believe that using this methodology has given us a fleet of aircraft whose reliability we are comfortable with.
- Rob Wolf
|
|