Application of Level 3 Probabilistic Safety Analysis in UK HPR1000

Abstract

A Level 3 probabilistic safety assessment (L3 PSA) is required in United Kingdom (UK) generic design assessment (GDA) to demonstrate that a new nuclear power plant is suitable to be built in UK. L3 PSA is used to assess the individual and societal risk and compare the results against the offsite radiation protection targets (RPTs) for fault and accident conditions. Its results can be used for risk evaluation and environmental impact assessment and provide useful information for alternative design features, rulemaking, and regulatory procedures. It takes into account atmospheric dispersion, demography, dosimetry, pathways to man, and plant/site characteristics. The radioactive source terms and their frequencies often are passed on from Level 1/2 PSA analyses. The framework for demonstrating the acceptability of the new plant design at GDA is presented for the assessments against RPTs 7, 8, and 9. There is little relevant good practice and mature standard for L3 PSAs that have recently been implemented worldwide. In this study, a pilot L3 PSA is performed for UK Hua-long Pressurized Water Reactor (UK HPR1000) to reflect the UK context and relevant good practices. It introduces the methodology and the processes to be followed to perform conditional consequence calculations for the faults and accident scenarios. All radiation sources are considered and analyzed. The radiological risks to a potential UK site are analyzed and compared against RPTs. A widely used code - PC COSYMA, is selected as a primary tool for the accident consequence calculations. The strengths and limitations of the code are identified based on the project situation, and either qualitative arguments or supplementary analyses are subsequently proposed to overcome the limitations. The final L3 PSA results are derived to support the demonstration that the offsite radiological risks for UK HPR1000 have been achieved as low as reasonably practicable (ALARP) and have met the UK regulatory expectation.