A probabilistic safety assessment of the Ignalina NPP was performed in conjunction with the Barselina project . The project is a multilateral co-operative study conducted by Lithuanian, Russian and Swedish experts. The Barselina project, four phases of which have been completed, was initiated in the summer of 1991. Its long term objective is to establish common perspectives and unified bases for assessing severe accident risk and establishing requirements for remedial measures for RBMK reactors. In this project the Swedish BWR Barseback is being used as a reference plant and the RBMK-1500 at the Ignalina NPP is being used as the applicant plant.
The Barselina project has been split into four phases. Phase 1 included familiarization with and analysis of a limiting number of safety systems and one single initiating event. It ran from October 1991 to the end of March 1992. Phase 2 included analysis of the principal components for all important safety systems and extension to several initiating events, but excluding external events and with limited treatment of human factors. This phase ran from April 1992 to February 1993. During phase 3, from March, 1993 to June, 1994, a full scope Probabilistic Safety Assessment (PSA) model of the Ignalina unit 2 was developed in order to identify the reduction of risk that can be achieved with possible safety improvements. The probabilistic methodology was applied on a plant specific basis for a channel type reactor of RBMK design. To increase the realism of the risk model a set of deterministic analyses were performed and plant-specific data base were developed and used. A general concept for analyzing this type of reactors was developed. During phase 4, July 1994 to September 1996, the Ignalina PSA model was further developed, taking into account plant changes, improved modeling methods and extended plant information concerning dependencies (area events, dynamic effects, electrical and signal dependencies). The PSA model is also updated to reflect the “as built” plant. The phase 4 PSA work used insights from the peer review performed by Battelle Pacific Northwest Laboratories on the phase 3 work. Another review is planned for phase 4.
The scope of the PSA study in the Barselina project is as follows. The source of radioactivity is the reactor core. The PSA also is based only on full power operation. Internal initiating events such as transients, LOCAs and Common Cause Initiators as well as internal hazards, such as fire, flooding and missiles are taken into consideration. Final consequence of the accident is core damage, equal to level 1 PSA. During the work, however the core damage states have been defined in such a way, that the results can be used partly as level 2 results - the damage stages represent 4 classes of environmental impact.
The hazard states in the core are evaluated on the basis of the development of accident event sequences resulting in conditions of either “safe conditions”, “violation”, “reactor core damage” and “severe accident”. The plant is considered to have met the “safe condition” requirements when temperature limits are not exceeded or exceeded in no more then 3 fuel channels, but cladding temperature of 800 oC are not exceeded in any channels. Safe operation limits are listed in Table 6.2. If the fuel cladding integrity is breached in more than three channels due to cladding defects and damages or because the cladding temperature limit of 800 oC is exceeded, the state is classified as “violation”. The “ violation” category can be regarded as belonging to relatively mild consequences. The reactor core damage category is characterized by severe accidental conditions caused by significant deviation from the design scenario which lead to cladding temperatures above 800 oC in no less than 3 and no more than 90 fuel channels of the reactor. Such accidents do not lead to loss of core structural integrity and this category can been looked upon as resulting in medium severity consequences. The “severe” accident category is characterized by severe accidental conditions caused by significant deviation from the design scenario and accompanied by the rupture at high pressure of more than 3 and less then 9 pressure tubes before the reconstruction of reactor cavity over-pressure protection system and 9 pressure tubes after reconstruction. Such an event can be accompanied by fuel melting or fuel damage in more than 90 fuel channels. This is the most severe consequence.
The accident sequence model for reactor cooling is a phased mission model divided into three time period:
The phase 4 results indicate that the overall core damage frequency is lower than the phase 3 results. The reason for this is the implementation of plant safety improvement features, and improved analytical procedures which eliminated unnecessary conservatism’s. The new results are also balanced by the improvements in the modeling of the CPS and ACS systems. The quantitative results obtained are based partly on plant specific data and partly on generic data. The results are not intended to show absolute risk levels, but to give a risk topography and to serve as a basis for identifying risk dominant features and systems design aspects and hence serve as a basis for safety improvement.
The general results show a probability of the “violation” end state to be in the order of 10-2 per reactor year. This probability is dominated by single channel blockage events. The assessment of probability value is based on operational data. To date 3 such cases have occurred in the RBMK reactors. However, the design of control isolation valves has been changed, which should have a positive impact on the initiating event probability. The “ damage” and “accident” end states show probabilities together on the order of 10-5 per reactor year, the same range as is expected for “core damage” as defined for Western reactors.
The risk typography is shown schematically in Fig. 11.1. The characteristic of the risk topography is that for “damage” and “accident” end states transients dominate the risk rather than loss of coolant accidents. Transients contribute more than half of the total frequency. Furthermore it is the long term failure to cool the core that produces the dominating contributions, Fig. 11.2. The distribution of risk between short term, intermediate term and long term contribution shows that most of the sequences lead to damage or accident only in the long term. Only the core blockage sequences lead to damage in the short term. This demonstrates both the high redundancy of the front line engineered safety systems and the “forgiving” features of the reactor. Low power density and a high heat capacity enables the reactor to survive at least a one hour total loss of electrical power without core damage. In the long term, support functions
Fig. 11.1 Damage and accident contributors in different initiating event classes 
become more important and their failures become thedominating contributions. The results indicate that a long term lack of coolant leads to severe environmental consequences because the core damage is assumed to occur at high reactor pressure. Human factors also contribute significantly to the core damage frequency. However, the development and introduction of event-based Emergency Operating Procedures is still not accounted for in the phase 4 results.
Fig. 11.2 Damage and accident contributors in short, intermediate and long term cooling 
Since January 1996 a newly formed internal PSA group at Ignalina NPP is responsible for the probabilistic safety assessment. The experience and information from the Barselina PSA phases provides valuable information to other projects, e.g., the in In-Depth Safety Assessment of the Ignalina NPP project, for development of the event-based Emergency Operating Procedures and Reliability and Maintenance Management System .