Fire Risk Assess. all Types of Reactors

The overall objective of this project is to provide the Beneficiary (Rosenergoatom) with a methodology for fire risk assessment which could be applied to all types of Russian designed reactors.

The short term objectives are to further develop the methodology for fire risk analysis carried out under the TACIS contract R3.06/91, to apply this methodology to the reference plant Novovoronezh Unit 3 using the event tree/fault tree models already developed within the plant specific PSA and to make a set recommendations specific to each type of reactor.

The project was designed to address:

1. The undertaken improvements to fire protection at Novovoronezh Nuclear Power Plant, Unit 3, and to review the extent to which fire hazard analyses were carried out.
2. The modifications which were carried out to the plant since the completion of the predecessor project R3.06/91.
3. The finalisation and extension of the methodology developed in the predecessor project R3.06/91 in order to carry out a fire risk assessment, incorporating the analysis of the effects of fire on the nuclear safety related systems.
4. The application of this methodology to a fire risk assessment to the reference plant.
5. The assistance to the Beneficiary in providing recommendations for performing a fire risk analysis for all reactor types (excluding the VVER-440/V230).

The project consisted of six tasks, which were completed successfully in accordance with the project programme. The updating of the fire hazard database at Novororonezh, prepared during the predecessor project, was followed by a deterministic fire hazard analysis using a computer code (FIREPRO) which has been developed and previously verified by ANSALDO.

The fire hazard analysis for the predecessor project and the recommended plant fire modifications were reviewed and this showed that further improvements were still needed to meet modern Western safety standards. However, the fire hazard analysis performed for new rooms, where safety related systems were recently introduced in the plant, established that no further measures were needed for these areas because the new fire protection measures applied comply with Western safety protection criteria and Western practice.

In this latest assessment the methodology was extended into a complete fire risk assessment, incorporating the analysis of the effects of fire on nuclear safety-related systems into a Probabilistic Safety Assessment (PSA). This methodology uses the results of the updated fire hazard analysis at Novororonezh and applies these to event tree/fault tree models. The fire-induced Core Damage Frequency (CDF) is estimated for the power operation of the Novovoronezh Unit 3 to be 1.29E-4 per year, which increases the overall CDF from 1.88E-3 to 2.01E-3 per year.

The final task was the assembly of data for the main differences of the Russian designed reactor types and to identify the differences specific to a fire PSA and to describe the modifications to the methodology developed for the reference plant or particular issues which needed to be considered.

Technical Recommendations

1. Western Project Management support should be maintained for future plant analyses to ensure that the technical approach developed is applied consistently to Western standards.
2. The PSA model should be updated to take into account the new systems which were being installed into NVV3, in particular mobile pump units and the new super emergency feedwater system (affecting the turbine hall fire sequence).
3. From the existing analysis it can be expected that the Main Control Room and the cable basement would be potentially significant contributors to the risk from fires. This is partially due to conservative modelling assumptions necessary because of lack of data. More detailed information on cable routing may allow reducing this conservatism. It is recommended to carry out this in conjunction with the update of the fire model.
4. Additional analysis should be carried out for fires during shutdown conditions. This could be done after completion of TACIS Project R2.01/96.
5. Carry out additional analysis on the reliability of operator actions following a fire, including the reliability which can be claimed from operating staff for early fire detection, late fire detection, early fire suppression and late fire suppression; this could reduce the conservatism in the assumptions made in this area.
6. Attention should be paid to the fact that for some nuclear power plant units the analyses may require a larger amount of information to be collected and stored. The Reference Unit VVER-440 requires approximately an order of magnitude less information to be collected/stored than some other NPPs. That in turn may require database modifications with the objective of reducing computing inputs.
7. The database could also include modifications to allow the introduction of universal queries for each specific room covering available equipment, cable routing, construction materials, etc. Several queries should be developed on cable connections of equipment in different rooms.

Major Problems Encountered

The site visit to the reference plant Novovoronezh (Task 2 activity) should have commenced on 24 May 1999. The local subcontractor could not finance his experts to travel before receiving an advance for his fees and direct expenses. As NNC did not receive the first payment from the EC before this date, it could not pay the agreed advance of 10% to the subcontractor. It was agreed that NNC would finance the trip of the subcontractor experts to NVV before the first payment from the EC but requiring that the subcontract was signed before paying this advance. The site visit could be planned after this agreement.

This did not delay the commencement of Task 3 (extend and finalise the methodology) the task was completed early November with the training seminar (3 weeks later than originally programmed). Task 4 also started slightly late because it required the completion of Task 2 and the training seminar at the end of Task 3. The update of the database introduced some additional delays and the problem with unqualified fire barriers caused further delays. The project programme was therefore delayed by around two months.

At the project meeting in April 2000 it was established that it was not possible to complete the project by the middle of June 2000. There were several reasons for this delay in addition to the small delays which had occurred in the past. The main reasons were:

1. GAN, the Russian Regulator, issued Directive No. 7 on 25th April 2000 which states that new fire protection requirements have to be taken into account in all projects. Assessment of these requirements and potential incorporation of the effects into TACIS R2.08/96 needed time not foreseen before the issued of this Directive.
2. At the time of agreeing the scope of the project, the Fire PSA was to be carried out with RISKSPECTRUM DOS version. ROSENERGOATOM (the Beneficiary) and Novovoronezh NPP then required that all projects used RISKSPECTRUM Windows version because the reports were going to be part of the NVV3 In Depth Safety Analysis Report. All tools used in support of this safety report had to be consistent. It required time to transfer the NVV3 plant specific PSA from one version to another and to verify the results.
3. Russian export regulations were changed and then the Ministry of Justice and Customs required clearance for all goods exported including reports. The payment had to be received by the Russian sub-contractor within three months of custom clearance; otherwise heavy penalties could be imposed.

A formal request to extend the contract until 15th December 2000 was sent to European Commission, the contract amendment was received on 20th June 2000.