Large break LOCA uncertainty evaluation and comparison with conservative calculation

The first formulation of the USA Code of Federal Regulations (CFR) 10CFR50 with applicable sections specific to NPP licensing requirements was released 1976. Over a decade later 10CFR 50.46 allowed the use of BE codes instead of conservative code models but uncertainties have to be identified and quantified. Guidelines were released that described interpretations developed over the intervening years that are applicable. Other countries established similar conservative procedures and acceptance criteria. Because conservative methods were used to calculate the peak values of key parameters, such as peak clad temperature (PCT), it was always acknowledged that a large margin, between the 'conservative' calculated value and the 'true' value, existed. Beside USA, regulation in other countries, like Germany, for example, allowed that the state of science and technology is applied in licensing. I.e. the increase of experimental evidence and progress in code development during time could be used. There was no requirement to apply a pure evaluation methodology with licensed assumptions and frozen codes. The thermal-hydraulic system codes became more and more best-estimate codes based on comprehensive validation. This development was and is possible because the rules and guidelines provide the necessary latitude to consider further development of safety technology. Best estimate codes are allowed to be used in licensing in combination with conservative initial and boundary conditions. However, uncertainty quantification is not required. Since some of the initial and boundary conditions are more conservative compared with those internationally used (e.g. 106% reactor power instead 102%, a single failure plus a non-availability due to preventive maintenance is assumed, etc.) it is claimed that the uncertainties of code models are covered. Since many utilities apply for power increase, calculation results come closer to some licensing criteria. The situation in German licensing is different to the USA. Significant differences of results are presented between conservative calculations according to the USA Code of Federal Regulation which requires to apply conservative models in conformance with the required and acceptable features of ECCS Evaluation Models, and best estimate plus uncertainty evaluation. Consequently, additional margin to licensing criteria is available by changing from conservative evaluation to best estimate calculations plus uncertainty analysis in the USA. This is not the case in other countries where the use of best estimate computer codes is already a common practice for 'conservative' calculations. However, uncertainty of calculation results is especially important when approaching licensing limits, e.g. due to power u prates. This is the reason why a sub-committee of the German Reactor Safety Commission recently recommended the assessment of uncertainty in calculated results in licensing