[ RadSafe ] Oyster Creek, et al.

George J. Vargo vargo at physicist.net
Thu Feb 16 22:27:48 CST 2006

Oyster Creek's design is the GE Mark I containment system, with the reactor
pressure vessel and recirculation system contained in a drywell (resembling
an inverted light bulb) and connected to a toroidal-shaped suppression pool
(commonly referred to as the "torus") connected by downcomer pipes
approximately 7 feet in diameter.  Any steam released into the drywell flows
through the downcomer pipes and is vented below the waterline in the torus
and quenched.  The Mark I does have its vulnerabilities, as do all designs.

"Failure" is a relative term.  Anything in excess of the Technical
Specification leakage rate (typically 0.5 to 1% volume per day) can be
considered to be a failure.  The scenarios that usually drive toward
containment failure come from overheating (maximum design temperature
~310F), not overpressure, and usually involve an extended anticipated
transient without scram (ATWS), which is a pretty far-fetched scenario,
given about 3 layers of ATWS prevention and mitigation in the GE design AND
the standby liquid control system, which can inject a high density sodium
pentaborate solution enriched in B-10 at about 1600 psig (normal operating
pressure ~1005).

Oyster Creek and Nine Mile Point Unit 1 are the only two domestic BWR-2
design plants. This is a 5 loop design that does not incorporate the jet
pumps used in BWR-3 and later.  The jet pump produces a standpipe effect
that can maintain about 2/3 core coverage with low-pressure coolant
injection (LPCI) in a large-break LOCA.  Core reflood in the BWR-2 works a
bit differently.  In the worst case, a fire truck can actually be tied into
the feedwater system for raw water injection.

Dresden 1 was a ~200 MWe BWR, that was a combined direct and indirect cycle
demonstration plant.  It employed a dry containment sphere (a-la Yankee Rowe
and many earlier PWR designs).  D1 was a complicated design and small
capacity.  Ultimately, it was deemed economically unfeasible to upgrade it
to meet the 1976 ECCS Rule (10 CFR 50 Appendix E), and to meet seismic
standards.  Humboldt Bay and  Indian Point 1 also met the same fate.  None
could be reasonably expected to recover the costs of necessary upgrades.

George J. Vargo, Ph.D., CHP
Senior Scientist
MJW Corporation
610-925-5545 (fax)
vargo at physicist.net

More information about the RadSafe mailing list