Nükleer Güç Santrallerinde Ağır Kazalar İçin 3+ Seviye Olasılıklı Güvenlik Analizi Yönteminin Geliştirilmesi Ve Akkuyu Nükleer Güç Santrali İçin Uygulanması
Duman Kantarcıoğlu, Veda
xmlui.mirage2.itemSummaryView.MetaDataShow full item record
The purpose of this thesis is to develop 3+ level probabilistic safety analysis (PSA) method for severe accidents in nuclear power plants and to apply the developed method to Akkuyu Nuclear Power Plant. For this purpose, first of all, safety principles in nuclear power plants were investigated. Off-site emergency management is defined as the ultimate safety barrier for severe accidents. For this reason, off-site emergency management approaches in Turkey and around the world have been examined. International standards for emergency and early protective measures have been explored. Mass evacuation practices were also investigated to detail the evacuation procedures which is one of the protective measures applied during general emergencies. In this study, 2nd and 3rd Level PSA applications were carried out in order to comprehend PSA methods in an integrated manner. For this purpose, AES 2006 reactor design and containment safety systems of Mersin Akkuyu NPP were examined. In order to study the behavior of containment safety systems in a selected severe accident scenario, a simplified Level 2 PSA application was performed. In this application, containment event tree (CAT) and fault trees for passive safety systems of containment were developed. Analysis were performed using SAPHIRE code. Moreover, calculations were performed to understand the logic behind atmospheric dispersion of radioactive materials in case of severe accidents. The calculation of the air concentrations at different distances were performed using the PAVAN code as a fundamental application. In addition, as an application related to Level 3 PSA, radiation exposures that may arise from atmospheric dispersion were calculated for emergency planning zones. For the severe accident case, RASCAL code was used to estimate the dose distribution. Also, with NRC DOSE code, dose distributions were estimated using the expected releases during normal operation. All data obtained were used to verify the data presented by Akkuyu Project Company in the Environmental Impact Assessment (EIA) Report. According to the relevant regulations, size of Urgent Protective Action Zone (UPZ) is a circle with 20 km diameter. Today, approximately 15,000 people live within UPZ. It is assumed that Akkuyu NPP will be online in 2025 and the population estimated for the year 2025 is 26,000 within UPZ. In the light of obtained data, the area subject to the evacuation plan was examined in detail and a probabilistic approach to the delay of evacuation, due to the failures in the off-site emergency management process, was developed. Based on this approach the evacuation model was generated. In the analysis, fault tree method was used. The uncertainties in the results were analyzed and the confidence intervals were determined. In addition, different combinations of failures that may arise independently from each other were studied and their possible consequences were predicted. A risk matrix was constructed to illustrate the probability-consequence analysis. Combinations with high probability and large negative impacts on evacuation and with low probability but large negative impacts were stated as having high risk profile because they may cause a serious break down in evacuation procedures. The results show that the probability of accidents during mass evacuation practices is extremely high. The confidence intervals of the calculations are wide. The generated risk matrix shows that disruptions can significantly affect the evacuation processes and in some combinations the process can be seriously interrupted. As a result, it is concluded that the developed method is a method that can be used to base plans on numerical data and analysis, to make the necessary arrangements in the plans by anticipating possible problems and to develop realistic emergency management approaches in this way.