Hazard Analysis (HA) cont'd

DOE-DP-STD-3016-99

are already in place to prevent or mitigate accidents and the need for any additional safety controls.

The HA shall provide assurance that the safety features of the nuclear explosive are not compromised.

Acceptable methods of HA include combinations of what-if/checklist analysis, process HA, hazard

and operability analysis, failure modes and effects analysis, fault tree and event tree analysis, and

human reliability analysis. In many cases, a combination of these techniques should be employed for

the NEO HA.

The HA evaluates potential hazards associated with specific NEOs and associated activities. The HA

shall consider hazards specific to the operation, including failure of equipment, tooling, support

systems, and human actions. The HA shall also address hazards external to the facility where the

operations are conducted (e.g., natural phenomena, transportation accidents, explosions or accidents at

neighboring facilities), and hazards internal to the facility but external to the operation (e.g., fires and

floods). Based on these analyses, the HA identifies and evaluates the effectiveness of relevant safety

controls including systems, structures, components, tooling, and equipment that are necessary to

prevent accidents or mitigate consequences. Due in part to the large uncertainty associated with

accident scenario likelihood, selection of controls should use safety conservative expert judgment.

Human reliability is an important element in nuclear explosive safety and needs to be considered in

both the HA and controls development. NEOs primarily entail hands-on tasks involving direct work

with the nuclear explosive; as a result, the potential for human error must be considered in the

development of hazard scenarios.

Generally, rare events have large uncertainties in accident frequency estimates due to a lack of data

concerning occurrence frequency, accident progression, and phenomenology; therefore, qualitative

estimates are sufficient for hazard scenario likelihoods. While quantitative uncertainty analysis is not

required, the magnitude of the uncertainties and the potential impact of large uncertainties on the

results should be discussed and documented in qualitative terms. Additional deterministic or

probabilistic calculations should be employed to provide a better understanding of the largest

consequence accident scenarios.