Quote: (Originally Posted by
Tino de Rijk)
There is only one real proof it will in fact work, and that is stress-testing in the flesh. <snip>
Parting note: APD realised also that there is a limit to the amount of paper and computer design & testing you can do. So a few years ago they invested for a fortune in what is still the world's biggest and best live CCR testing rig: a custom-made ANSTI rig where a complete working CCR unit can be taken to 200 meters, in real water, cold & salt, in any position, at any breathing rate, with measurement at the mouthpiece. I've seen you do some "proof" on an Inspiration scrubber canister, about breaktrough, on a workbench, with some analysers attached. It made we very sad, and shows exactly my problem with your sort of "proof".
First point you make: the units are being tested very thoroughly in the flesh right now. We have been testing parts for a considerable time (not CAD, actual parts). .
On the APD rig, it is neither the world's largest nor has the greatest capability. Our own rigs, which are by no means the largest or the deepest either, go to 1400msw, any position, using sea water at breathing rates to 90lpm. The figures I published are from actual tests, not CAD. What is your issue?
Units were at DEMA 2006. Between 1st Jan and DEMA 2007, there are 17 verification reports to publish (most in Q1 2007), giving detailed results of testing every subsystems, and their most recent changes.
On the unexpected failure issue raised by Genesis, I agree that is the most important issue here. It is why we publish our Test Plans and do such thorough testing of every part, as well as formal verification of the design. On the faulty valve type accidents, that is exactly the sort of thing that formal verification picks up. Reactor designs do not have 100% formal verification.
Alex