Discussion of Contingencies

DOE-STD-1134-99

The final method to consider is computer code reactivity calculations. Typically these will

involve KENO and/or MCNP. The code and cross section set (e.g., Hansen-Roach 16 Group,

ENDFB-v, etc.) used should be stated.

Key Review Issues (As Applicable)

Clear statement of method used to derive subcritical values.

Referenced handbook or ANSI/ANS standard data.

Description any hand-calculations relied upon.

Listing of code and cross-section set utilized.

Referenced code validation calculations.

5.0 Discussion of Contingencies

This is perhaps the most important section of the entire evaluation. This discussion of

contingencies should provide a stand-alone summary of double-contingency regardless of the

underlying references. As stated in Appendix A of ANSI/ANS-8.1:

The few criticality accidents that have occurred in industrial operations have resulted

from failure to anticipate conditions that might arise; none has resulted from a faulty

calculation of keff.

Appendix A of ANSI/ANS-8.1 contains a list of typical scenarios to consider when reviewing

the contingency analysis for a particular evaluation. The reviewer must be familiar with the

operation being evaluated to perform an adequate review. Without such knowledge, no decision

can be made relative to whether postulated abnormal events are "anticipated", "unlikely", or

"incredible," or if any credible contingencies have been omitted.

Each contingency that could lead to criticality should be shown to be unlikely, independent and

non-concurrent with other contingencies that could lead to criticality. If this can not be done,

then the contingency under evaluation becomes part of the normal operating conditions, i.e., an

anticipated event.

Two simplifications are possible here. Criticality scenarios that are deemed "incredible" and

those that are not physically possible need no contingency analysis. The reviewer's task in each

of these two cases is to decide whether the arguments against criticality occurring are sound. The

reviewer should take care to note the definition of "contingency" and "credible" in DOE-STD-

3007-93. If a quantitative probabilistic risk assessment is utilized in the CSE despite the

qualitative definition of "credible", the reviewer should rigorously scrutinize any calculations

and assumptions leading to extremely small probabilities. In the more likely situation, the

experienced, professional judgement of engineering and operational personnel will be the basis

for this argument. The bases for these judgements should be carefully documented in the NCS

Evaluation. The reviewer's task is to verify that the bases (assumptions and conditions) for the

experts' judgement are reflected, as appropriate, in the NCS limits and requirements for NCS

controls.