Example Margin of Safety Statement

DOE-STD-1134-99

All anticipated or unlikely abnormal events (i.e., contingencies) must result in a subcritical

system. The evaluation must contain a summary statement concluding that the system will

remain subcritical after the unlikely event occurs. This typically takes the form of a "margin of

safety" statement.

EXAMPLE MARGIN OF SAFETY STATEMENT

"If the system is double batched it will remain subcritical as shown by Monte-Carlo calculations. A

quadruple batch with full water reflection is required before criticality can be achieved."

Key Review Issues

Reasonable, credible failures are bounded and/or controlled.

Contingencies are unlikely, independent events.

Abnormal, unlikely events considered are documented clearly and specifically.

Statements of subcriticality or safety margin for the contingent events are provided.

Abnormal conditions listed in any available process description are dispositioned

appropriately.

6.0 Evaluation & Results

The calculational results, if applicable, are summarized in section. The "normal" and "credible

abnormal" cases should be documented in this section and the associated calculations must

demonstrate that these are subcritical. The calculations may be incorporated explicitly and

completely, as in a completely new evaluation, or by reference to existing evaluations (e.g.,

CSAs). The discussion in this section should clearly define the connection between the

calculational models and the process descriptions, as applicable.

In the case of a new CSE, the reviewer should compare the models described here to the

description of the system contained in Section 2 to ensure they are the same. Geometries and

dimensions should be spot checked for accuracy and consistency.

It is good practice to perform a "sanity check" on the reported results by comparing to accepted

industry (i.e., handbook) data or by performing appropriate hand calculations. The reviewer

should have a good idea of what configurations are critical and which ones are clearly

subcritical. A quick review of applicable handbook data and/or hand calculations will serve to

independently confirm the results of this section. If any calculation does not seem consistent

with this "sanity check" or analysis from previous work, further investigation of the calculational

models is warranted.

Another good "sanity check" is to look at fission densities. In a large array of fissile material,

incorrect neutron start types can be used. The easiest way to check this is to look for similar

fissile items in an array and assure that the fission densities are similar when reflector and edge

effects are taken into consideration. If similar units have dissimilar fission densities, something

is wrong and the results should be questioned.