Databases and their analysis for applications in quantitative risk analysis of NPP

There are several equipment failure databases originating in the nuclear power plant that are successfully used in quantitative risk analysis applications Le, probabilistic safety assessment (PSA). Regulatory bodies (like AERB, India) now accept the importance and use of databases for safety and reliability studies in the nuclear industry as industry practice. Databases containing human error data originated from studies carried out in the nuclear industry and are used, in conjunction with equipment failure models to assess overall levels of safety to operation of a NPP. Based on these databases, very powerful risk analysis and reliability engineering methods such as Fault and Event tree methods are used to determine quantitatively what levels of safety, reliability and performance exist at one's plant operation and where improvement resources should be applied. Also, the use of risk informed approaches has been both identified and encouraged as an option to improve management of station resources more effectively and this technique has seen notable successes in the areas of in-service inspection and testing However, for any analysis a large database is required. Hence, compilation of data is key element and must for PSAto get the real picture and genuine results. This paper describes about the station practices that are followed at Rajasthan Atomic Power Station, Unit-3&4 regarding data collection from computerized maintenance management system (CMMS), which is a deficiency report based system in which deficiencies are raised and their subsequent follow up is done till their rectification. After reporting of a deficiency, work permits are applied by maintenance agency and these permits are subsequently issued by operation section and ultimately surrendered by maintenance agency after rectification of deficiency. Appropriate cause codes, failure codes and history of work being done is maintained in CMMS databases, which are referred to for their inclusion in PSA model.