Historical review of fire safety at npp and application of fire psa to westinghouse pwr npp in the frame of risk-informed decision making
- Autores: Matthew Asamoah
- Directores de la Tesis: Javier Dies Llovera (dir. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2018
- Idioma: español
- Tribunal Calificador de la Tesis: Guillem Cortes Rossell (presid.), Carolina Ahnert Iglesias (secret.), Edward Horga Kordzo Akaho (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería Nuclear y de las Radiaciones Ionizantes por la Universidad Politécnica de Catalunya
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
The importance of fire as a potential initiator of multiple-system failures took on a new perspective after the cable-tray fire at Browns Ferry in 1975 The review have shown that the first generation Nuclear Power Plant (NPP) fire safety was not factored as high risk area that needed to be effectively assessed and quantified. This resulted in development of peculiar fire safety regulations, standards and expensive backfits. Lack of appropriate regulations and effective methods of fire risk assessment, prescriptive, difficult and expensive retrofit regulations were instituted in USA. The alternative risk-informed performance based regulation was established in USA to resolve the challenges of the prescriptive rules. The review have revealed that both the prescriptive and risk-informed performance based approaches will not represent adequate design basis for new Nuclear Power Plants. The Japanese were pulled in the path of renew fire safety regulations and risk quantification after the Fukushima accident. It has been recognized that effective fire safety assessment, and culture, in concert with countermeasures to prevent, detect, suppress, and mitigate the effect of fires if they occur, will minimized NPP fire risk. Among the numerous recommendation the fire safety at NPP must be planned and engineered before construction begin using the state-of-the-arts technology. Also, the methods of fire risk assessment must integrate the state-of-the-arts deterministic and probabilistic approaches. Two methods are presented which serve to incorporate the fire-related risk into the current practices in nuclear power plants with respect to the assessment of configurations. The first method is a fire protection systems and key safety functions Unavailability Matrix (UM) which is developed to identify structures, systems, and components significant for fire-related risk. The second method is a fire zones and key safety functions (KSFs) fire risk matrix which is useful to identify fire zones which are candidates for risk management actions. The UM is an innovative tool to communicate fire risk. The Monte Carlo method has been used to assess the uncertainty of the UM. The analysis shows that the uncertainty is sufficiently bounded. The significant fire-related risk is localized in six KSF representative components and one fire protection system which should be included in the maintenance rule.
The unavailability of fire protection systems does not significantly affect the risk. The fire risk matrix identifies the fire zones that contribute the most to the fire-related risk. These zones belong to the control building and electric penetrations building. The aggregation of Internal Events PSA model and Fire PSA model have shown that the Fire PSA contributes 38.4% to the Risk increase.
The feasibility of developing Fire-related Risk Monitor from the FIRE PSA for the Spanish NPP was carried out. One of the main challenges is that RiskSpectrum® fire PSA has 384 fire cases and 384 CDF but in Risk Monitor one CDF is required. However, CAFTA is unable to convert a Sequential Fault Tree structure of the internal Event tree in the Fire PSA. The conversion fails to implement neither all of the sequences leading to core damage nor the Fault Tree selection of the frequency of fire. The proposal is to suppress exchange events and introduce the alignment of the consequences so that a unique result of core damage can be quantified. The detection and fire suppression Event Trees in the reference model were replaced by detection and fire extinction Fault trees. The frequency of each Fire Case of the conversion model and the reference model are quantified and the frequencies compared. The results shows that 90% of the cases are valid, however, the rest have challenges with MCS. A unique CDF of 7.65x10-7 is quantified compared with 9.83×10-6 of the reference. The conversion of the new model in CAFTA was not successful due to software incompatibility.
