TACIS pr RF/TS/15 TSO supp to Gosatomnad

In 1998, Rosenergoatom concern issued the final report Analysis of Influence of Fires and their Consequences to the Safe Shut-down of VVER-1000 (W-320) reactor plant for the 4th unit of Balakovo NPP.

The Balakovo NPP decided to analyse the safe shut-down procedures in fire conditions using the so-called Fire PSA methodology.

SEC NRS of Russian Gosatomnadzor carried out the review of the Balakovo NPP with the assistance of Western partners within the framework of the TACIS R/TSO/15 project. The object of this project is to perform the Balakovo NPP review and the development of a Guide on the review of a Fire Risk Analysis (FRA).

These different tasks of the project were:

• Review of the collected data and information;
• Review of critical zones selection;
• Review of deterministic analysis and fire event trees;
• Review of event trees quantification and core damage assessment;
• Review of uncertainty analysis;
• Guide to perform a review of a FRA;
• Quality assurance procedure for performing review of FRA.
   

The objective of this project is:

• To provide GAN RF with a technical assessment evaluation of a fire risk analysis evaluation performed for one VVER 1000-320 reactor by the Russian industry;
• To take benefit from the lessons learnt from this assessment evaluation in order to develop guidance documents for the fire PSA review of fire risk evaluation, taking into account the Western practice.

The main results of the project are:

1. Review of collected data and information

1.1 Review of selection of initiating events caused by fire. 
The information on selection and grouping of IEs is presented in the Balakovo NPP, rather completely and allows to perform a review with a high level of confidence. The work has been performed according to the state of the art.

1.2 Review of selection of compartments, systems and equipment.
The information on selection of systems is partially presented in the Balakovo NPP and does not allow to perform a detailed review with high level of confidence.

Concerning the selection of systems that could lead to an initiating event or are necessary to reach or maintain the safe shutdown, the information presented does not indicate, in a detailed way, the reasons that permit to exclude from consideration the systems screened out.
Concerning the selection of equipment, the work is performed rather completely; however there is no complete list of equipment selected.

Concerning the selection of compartments, the information on methods implemented is submitted restrictedly and does not allow to perform a complete review of the work performed with a high level of confidence. There is no link between the list of compartment selected and the list of equipment selected.

Concerning the presence of equipment and electrical cables in each compartment, essential deficiencies of the Balakovo NPP are the following:

• Absence of a data base on the equipment and cable information;
• Absence of the information on cable routing;
• Absence of results of data verification of data by a sample verification of compartments;
• Absence of results of walkdown of compartments.

1.3. Review of fire load determination for all types of the considered equipment, electrical cables and transient fire sources.
The information on the methods of fire load analysis of the equipment, cables and transient sources is submitted rather completely and allows to perform the review with a high level of confidence. Nevertheless the assumptions related to the estimation of amount and types of the equipment and also cable mass in compartments are not sufficiently detailed. Moreover, the estimation is not based on data providing from factories and/or manufacturers. 
The information on fire load of transient sources was not determined and used.

1.4. Review of fire frequencies.
The information on the ignition frequency for equipment and cables was not determined on the basis of operating experience of Russian NPP. It is the same thing for the transient combustible for which generic data ignition frequencies from USA NPP were used. In this case it could lead to incorrect estimation of the fire frequency for the Balakovo NPP.

1.5. Review of data on fire barriers and openings between compartments.
The detailed information related to constructive and layout and to the resistance rating of fire barriers is submitted. 
An essential deficiency is the absence of verification of the data with walkdown, notably to verify the absence of openings not indicated in the available documentation.

2. Review of critical zones selection

2.1. Scope of the review.
To select critical zones, the following points have been analysed:

• Resistance analysis of the fire barriers;
• Analysis of fire propagation;
• Estimation of fire frequencies;
• Analysis of initiating events caused by fire;
• Estimation of safe shutdown equipment damage;
• Impact of fire on the probability of human errors;
• Estimation of conditional probability of core damage.

2.2. Conclusions. 
The review of fire propagation zones study cannot lead to clear conclusions due to the imprecision of the following data, assumptions and approach:

• The approach retained for the critical zones selection is not sufficiently detailed and the input and output of each step of the screening are not clearly indicated;
• It is considered that the fire cannot propagate through the fire barriers. This assumption could be incorrect in case of lack of justification, indeed, it limits the fire development scenarios;
• The absence of estimation of fire duration in the compartment;
• The absence of justification of fire rating of fire barriers;
• The non-consideration of failure of fire extinguishing systems;
• Concerning the Human Factor, the analysis was carried out only for personnel actions in rooms adjacent to the MCR.

The error probabilities of the personnel actions in other rooms, used in the PSA level 1 were not corrected. This could lead to non-conservative Human Factor assessment and, therefore, to an underestimation of the core melt frequency and to the possible excessive screening of some fire zones. 

However, some conservative assumptions have been used, as the consideration of the damage of all equipment inside the fire zone.

3. Review of Fire event trees and deterministic analysis

An approach consisting to elaborate to study and define fire scenarios has not been used. It has been conservatively assumed that all the equipment has been damaged by fire. Moreover, for the most significant fire zones, to confirm the non-fire propagation, fire simulation has been performed in the BR. So, to verify numerical fire-related deterministic results obtained, fire modelling calculations have been performed with the FLAMME-S code during the mission of Russian experts from SEC NRS GAN (Moscow) in Cadarache, France.

It can be concluded that the results of FLAMME-S simulations in general have confirmed the results obtained in the Balakovo Study. In particular, a good agreement in the results is observed in terms of the maximum temperatures and pressure values averaged by the volume of the rooms analysed when the room is closed. 

However, some differences in the results for the fire duration time could be observed. In general, the tendency is that FLAMME-S provides more lower durations than the code used in the Balakovo Study.
A general conclusion is that in whole fire-deterministic results obtained in Balakovo Study are confirmed by FLAMME-S simulations, but to make the results and conclusions more specific, more detail information on the data used in Balakovo Study and more analyses are required.
A walkdown has been performed on Balakovo unit 4 by RF Gosatomnadzor in July of 1999. This walkdown was carried out for two compartments considered as critical compartments in the Balakovo NPP. During the walkdown, the procedure limiting the entrance (or storage) of transient combustible has not been examined. So, it would be useful, during a future walkdown, to get from the utility, this kind of information. All participants in this contract agreed that this first walkdown performed by GAN is very useful to increase the level of confidence of the information included in the Balakovo Report. The way to perform a walkdown and the way to take into account the results has been included in the review on fire event trees.

4. Review of event tree quantification and assessment of core damage frequency

The supporting documentation of BR is incomplete. Not all of the results obtained are available for the review. Because of this, for a considerable number of cases, the reviewers were not able to come up with a certain judgement on the adequacy of the quantification approaches and results. 
There is a concern about adequacy of approaches and assumptions implemented in the study under review for event trees (ETs) modelling and quantification, which could introduce non-justified optimism in the quantitative estimations.

The most important are:

• Human error probabilities from Level-1 Internal Event PSA were not modified to take into account for fire effects;
• number of accidental sequences were excluded from the core damage end state;
• Hot shutdown is modelled for most of the cases without taking into account the requirement of technical specifications to perform cold shutdown.

The ETs from Level 1 Internal Event PSA were modified to reflect the fire-induced equipment failures. However, there is no comprehensive information/listing about components and electrical cables located in affected fire zone. This was also a source of reviewers concern.

In general, the study under review demonstrated a comprehensive work of identification of plant vulnerabilities towards internal fires and identification of most hazardous fire scenarios. 
However, to make the results consistent:

• The omissions and mistakes should be corrected;
• More justifications are required for confirmation of optimistic assumptions and approaches implemented in the study (or elimination of unconfirmed optimistic assumptions);
• The documentation on the work performed should be essentially improved.

5. Review of uncertainty analysis

The objective of uncertainty analysis is to provide both a qualitative discussion and quantitative measures of the uncertainties in the results of FRA. The uncertainty analysis has not been performed in the Balakovo NPP. So, the review of Balakovo NPP consisted in the list of the main sources of uncertainties with indication of their potential effect on the results.

Based on the review performed, a general conclusion can be made that the uncertainties related to the Balakovo study are considerable. In order to make the fire risk results more consistent with the state of the art, additional analyses would be helpful.

6. Guide for reviewing a Fire Risk Assessment

The recommendations in this Guide are designed for the use by experts engaged for performing a review of FRA. This Guide is developed in accordance with 'Statement on the organisation and performance of review of nuclear plant safety, radiation sources and storage places'.
The recommendations of this Guide should to be taken into account by central machinery administrations, SEC NRS and district administrations of Gosatomnadzor of Russian Federation at conclusion of contract with strange expert companies recruited to carrying out the review.