Re: RE: OT: Three-Mile Island Events

>It amazes me that educated people continue to think that "engineers"
>should be able to design and install any system (simple or complex) to
>be completely fail-proof. Everything becomes obvious in hind-site.
I don't think the issue is fail-proof design--the issue is three fold. You design it fail safe where you can and impose either redundancy or constraints where failsafe design isn't possible. As a final resort you use procedural controls. That's the philosophy of prudent design, but you really need proper, experienced technical leadership in both the design process and for operations.

When you compare the success of the US nuclear navy and the failure of the US nuclear power program, you really see the difference that technical leadership makes. Rickover may have been an arrogant horse's ass, but he was a competent, dedicated, incorruptible arrogant horse's ass who was committed to excellence from everyone associated with the program.

One of the reasons the US nuclear power program is such a shambles (as compared with, say, the French program) is the lack of leadership. One of the important conclusions of the Kemeny report on TMI identified among other things a lack in the NRC of quantified safety goals, inability to set priorities, inexperienced staff, arbitrary requirements and lack of a comprehensive system approach, among other things. These are all leadership issues. Another symptom is an overwhelming emphasis on conservative models and assumptions--concentration on major disaster scenarios to the detriment of service conditions. The fact that the PORV indicator light indicated only that the valve had been commanded to close and not the actual close is an example of a critical design error, and resulted in a small non-catastrophic loss-of-coolant-accident turning into a big one.

Bruce Babbit's supplemental view is very useful, addressing the issue of who should be allowed to run nuke plants. The US nuclear program as basically been left up to individual utilities, in Babbitt's words the 'unquestioned assumption by the NRC, until now, that any utility that wanted to produce nuclear power could do so -- a policy the no matter how small or unsophisticated the utility, it was eventually entitled to wrap its arms around a nuclear reactor.'

I think the roll-your-own approach has effectively destroyed public confidence in nuclear power because it's inherently leaderless. Every plant is different and problems and procedural matters are addressed one utility at a time. Some utilities have paid their technical dues and operate well, others haven't. There's a lot to be said for a free-enterprise approach like this when you're talking about selling toasters or garden hoses. When a schlockmeister starts selling crappy toasters the marketplace sorts him out pretty quick because the worst thing that can happen is you have to go out and buy another toaster or garden hose. Nuke plants aren't like that. A decision to go nuclear is too big a decision to leave to people who only cares about his stock options and the next quarterly report. Market forces don't come into play until after a product is in service. That's way too long with respect to a design cycle to be effective in safety-critical situations. It's also way too far down the line with respect to personnel issues like training and experience. You can't afford to train your designers on the first system or train your operators during the first operating cycle.

It's pretty obvious that we're seeing the same lack of intelligent control internationally, with nations using civilian nuclear generating power to screen weapons development. There are a lot of nations who have absolutely no business putzing with civilian nuclear power, let alone throwing away badly needed money on bombs. My favorite approach would be to have Michael Rennie fly down in a robot-piloted space ship and force humankind to get our shit together or face annihilation. Too late for that, just like it's probably too late to get the nuclear genie back into his jug.

Christopher Wright P.E.    |"They couldn't hit an elephant at
chrisw@skypoint.com        | this distance"   (last words of Gen.
___________________________| John Sedgwick, Spotsylvania 1864)