Lessons learnt from PSA for new and advanced reactors in Russia

Abstract

Customer requirements to probabilistic safety targets are usually stronger than existing Regulatory or IAEA ones. It appears that industry takes the lead over regulation in this case and forces the designer to find and implement appropriate means to enhance safety, which sometimes have no reference to practical experience. On the other hand, regulatory documents and the existing PSA methodology are mainly oriented to operating plants. This creates problems when developing a PSA as well as performing regulatory reviews. The scope of the PSA may be different depending on a design stage such as the development conceptual, basic or detailed design. In addition, the base case PSA is usually performed for NPP in design. However, a customer may require additional PSA applications to consider, for instance, risk monitoring. In this case the scope of the PSA should be extended to implement special attributes of the application needed that often requires specific information not available at the design stage. Lack of design information affecting PSA development may be associated with incompleteness of the design that is typical for interim design stages and communication problems caused by the involvement of many different companies in the deign activity. To deal with this issue bounding technologies and the iterative PSA development are used. However this sometimes contradicts to the “best estimate” approach recommended by regulatory guides. PSA development for advanced NPPs has raised some issues originated from unknown new components, processes and technologies incorporated into the design of an advanced plant. The paper addresses some issues resolved while carrying out PSAs for advanced NPPs. Some PSA results for new advanced VVER plants under construction and the first lessons learnt from the Fukushima accident are also discussed.