Many of the principles for design basis assessment have been established for present-day reactors over many years. These include the ‘defence-in-depth’ principle, needs for diversity and redundancy, Safety Analysis Report (SAR) assessments and so on. In this section, the principal design basis approaches that are likely to apply to reactor licensing in the future are reviewed.

Most of the licensing submittals for present-day plant have been submitted using the conservative evaluation model (EM) methodology (Table 8.3). Conservative modelling was required to overcome lack of detailed knowledge of the phenomena. This methodology is commonly referred to as ‘Appendix K’ referring to the relevant appendix in the US CFRs (10 CFR 50).

BE methods are likely to be a common goal for licensing in the future, including those for future reactors. BE methods have been accepted by the USNRC (and other regulators) c. f. Appendix K Revision in 1988.

The acceptance criteria for fault studies are established by good understanding of the physics of present designs. However, as designs evolve, these criteria may need to be re-evaluated. The additional changes may also be required to accommodate extensions in, e. g. mode of operation.

There are likely to be increased requirements for Probabilistic Safety Assessment (PSA) studies (Table 8.4) to underpin deterministic studies and to help estimate doses to the population, source term of release, etc. Probabilistic targets are likely to become more

Table 8.4. Safety approach for severe accidents

PSA approach Selection of most probable sequences leading to a core melt

Provision of preventative or mitigative measures Wide coverage of possible sequences

Good quantification of the benefits from proposed measures Depends on the status of PSA accident analysis

Deterministic safety Definition of containment challenges from core melt behaviour

analysis approach Assurance of containment integrity by design measures

stringent, e. g. on core damage frequency or on containment limits. The current trend is to use best estimate methods for the frequencies and probabilities in PSAs.

The move towards BE methods is being supported by regulators and utilities because more realistic margin estimates enable a better quantification of actual risk to be obtained and enable a wider operating window.

However, there are developments required before the methodology is likely to be regarded as a mature engineering tool. It is necessary to be able to quantify unbiased uncertainty limits on key parameters (e. g. peak clad temperatures) and as yet the methodologies are not yet very practicable for licensing studies.

Another factor is that there is generally reluctance to change from an established methodology that is accepted by all parties.

Looking further to the future, risk informed methods (Wahlstrom, 2003) are being put forward by the USNRC but these methods require further development. Traditionally, the safety of NPPs has been justified by a deterministic approach based on the defence — in-depth principle and single failure criterion for design basis accidents, etc. Probabilistic approaches have further developed and now the probabilistic safety analysis methodology is becoming well established. These provide a means of taking a systematic approach to determining the probability and therefore risk of various failure sequences.

The idea of risk-based or risk-informed approaches is to focus on the most important issues in terms of risk. If PSAs are used to determine the risk then clearly it is important that there is confidence in the PSA methodology used. Risk-informed approaches can be applied to new reactor design or indeed to assist modifications of old plants, to target maintenance actions and inspections. PSA methodology can also be used to identify the safety categorisation of components.

The USNRC has made a commitment to move towards a risk-informed regulatory regime. Other regulators are considering the development of the approach.

The notions of design basis and defence-in-depth have been well established in the licensing of present generation reactors. For some future systems, these notions may need to be revised in the light of newer technologies with very different designs, materials and fuel cycles.