Nuclear fission

11.2ASETC Calculations OF SFP ACCIDENTS IN L2 PSA (IRSN)

11.2.1CONTEXT IN FRANCE

The French operating PWRs (Gen II) includes SFP which are located outside the reactor building. An additional common SFP is planned to be created in near future to limit the quantity of fuel assemblies in existing SFPs.

The safety of SFP is based on design features that allow prevention of fuel melt accident. Accident situations initiated by a long term total loss of cooling would be managed thanks to additional water make-up. These additional water make up are reinforced by EDF after the Fukushima. Accidents situations initiated by SFP draining have to be excluded thanks to specific features like siphon breakers, pipe elevations, system design (isolation), quality of the SFP steel liner, etc. SFP L1 PSAs were developed by EDF for operating NPPs and lead to improvement of the SFP circuits and transfer channel from reactor. These improvements by EDF are implemented during PSR.

Concerning “criticality”, the safety demonstration is provided by the geometry of the SFP (minimal distance between the assemblies) for the 900 MWe PWR series, by the geometry of SFP and the presence of neutron absorber plates between the assemblies (high density racks) for the 1300 and 1450 MWe PWRs.

In PSAs fuel melt accidents in a SFP are always associated to “large release” accidents. No mitigation measures are credited, until now, because the SFP buildings are not designed for that purpose.

11.2.2900 AND 1300 MWE PWR SFP CONFIGURATIONS STUDIED WITH ASTEC

To prepare a future extension of IRSN L1-L2 PSA to SFP accidents, ASTEC simulations have been performed for the 900 and 1300 MWe PWRs. Such simulations need number of data like:

geometric or material data: for the fuel bundles, racks, SFP, fuel building

data on the source of radioactivity: assembly number stored in the fuel, residual power;

data on the reactor and SFP operating states and fuel management (charge/discharge) process: power states of reactor, working shutdown, refuelling shutdown, and reactor fully discharged.

The reactor building is separated from the fuel building and to enable the transfer of the fuel, a transfer tube links the reactor and the fuel pools. The pools in reactor and fuel buildings are divided in two and three compartments. Each compartment is separated from another by a gate. The Fuel Pool Cooling System is connected to the Reactor Heat Removal System and to the Reactor Water Safety Tank. So, in addition to the cooling of the spent fuel, this system may be used to back up the reactor cooling system or to fill and to drain the pools.

All the possible states of spent fuel pool have been studied

Reactor core fully discharged in the SFP (RP-API in Figure 10.2.): the IRSN calculations consider that the SFP accident occurs at the beginning (point 2 of the Figure 10.2.) and at the end (point 3 of the Figure 10.2.) of this SFP state. At this moment, the whole reactor assemblies have been discharged in the pool, therefore the residual power is maximal,

Reactor core fully charged (APR in Figure 10.2.): the SFP state is defined by the end of refuelling (point 4 of the Figure 10.2.) and the beginning of the unloading (point 5 of the Figure 10.2.). The time of that SFP state is generally very long because the reactor produces electricity during this period. The assemblies stored in the pool are cooled and have lost most of their residual power. This is the reason why the assessment distinguishes the point 4 and point 5.

Refuelling shutdown (APR in Figure 10.2.): the IRSN assessment considers that an accident occurs during handling of an assembly. During this operation, the water height above handling assembly is minimal. The variation of residual power contained in the pool can be important: at the beginning of the refuelling shutdown, only “cold” assemblies are in the pool but at the end of refuelling shutdown, reactor assemblies are in the pool.

Two types of initiating events are addressed in the ASTEC calculations: the total loss of cooling and the rapid drainage accidents. This is a parametric study and realism of such situation has not been considered. A loss of SFP cooling accident could occur if the cooling system of the pool is failed and if the water injections in the pool are not available. A draindown flow rate could be induced by a little leak of the pool but an error on circuit configurations (that would not be detected and stopped by automatism).

For each state of spent fuel pool except refuelling shutdown, IRSN has studied with ASTEC:

the uncovering of fuel assemblies stored in the pool due to a loss of cooling

the uncovering of fuel assemblies stored in the pool due to a rapid draindown.

For the particular case of refuelling shutdown, IRSN has assessed with ASTEC:

the uncovering of a fuel assembly during handling due to a loss of cooling

the uncovering of a fuel assembly during handling due to a rapid draindown.

Sensibility assessments have been performed on the residual power, the draindown flow rate (3000 and 6000 m3/h) and the initial pressure in the gap of assemblies.

The gap defines the volume between the fuel and its cladding. During fuel irradiation, gaseous fission products are created which brings an increase of the pressure in the gap. The cladding rupture time is shorter as the gap pressure is high and the irradiation period is important. IRSN has tried to study the behavior of “hot” and “cold” fuel bundles.

Time (day)

Residual power (MW)

Figure 10.2. – Example of residual power evolution in a 900 MWe PWR SFP

11.2.3PHYSICAL PHENOMENA

A loss of cooling accident in a pool can be divided into several phases:

Phase 1: heat-up of the SFP to the saturation conditions: no radiological danger for the operators and the population near the plant.

Phase 2: vaporization of water until it reaches the top of the fuel assemblies: the progressive loss of coolant above the stored assemblies generates a radiological danger for the operators of the plant that could limit the interventions in the fuel building.

Phase 3: uncovering/degradation of fuel assemblies contained in the SFP: the temperature of the assemblies increases when uncovering. Rupture cladding occurs when the temperature reaches 1000 K. Cladding oxidation appears at 1200 K with air and 1400 K with steam. The air oxidation is much more exothermic than the steam oxidation.

Phase 4: after progressive fuel melt, MCCI can occur at the bottom of the SFP.

After the cladding rupture, the volatile fission products that have accumulated are released. The release continues with the release of other fission products presented in the fuel, because the temperature of release mechanisms is reached.

In the case of a slow uncovering due to a loss of cooling, the predominant phenomenon is rather steam oxidation, although a mixed atmosphere could be present above the uncovering front. This atmosphere influences the degradation and the H2 production.

Once uncovering, the top of the racks undergoes degradation (at 1700 K). Shortly after, when the melting point of zircaloy is reached (2500 K), the top of fuel rods melts. Because of racks design, the molten metals relocate just below. The degradation process is defined by the successive relocation of molten mixture.

11.2.4SUMMARY OF ASTEC RESULTS AND LIMITATIONS OF THE CODE

The main outcomes are listed below and the Figure 10.2. presents the synthesis of degradation calculations performed with ASTEC.

These calculations provide first order of magnitude results with rather pessimistic assumptions (e.g. no thermal loss by pool walls or atmosphere of the building). Some conclusions are proposed here:

concerning loss of cooling situations,

• 
  the top fuel rod dewatering starts between 53 and 841 hours, the cladding rupture time of fuel assemblies stored in the pool occurs when the water level in the SFP is below 2 or 3 meters of the top of stored assemblies: clad ruptures occur between 18 and 184 hours after top fuel rod dewatering,
  
• 
  times are shorter during fuel handling (dewatering can start after 36 hours),
  
• 
  there is an important hydrogen production (650 to 1140 kg) but the steam released from the SFP may inert the SFP building atmosphere. One concern for late accident management (to avoid explosion) would be to control air ingress …
  

concerning loss of coolant situations,

• 
  the delay before uncovering of course depends on the drain down flow rate, then the delay before cladding rupture depends on the “hot assemblies” power (roughly from 1 h to 10 h),
  
• 
  there is limited steam in the SFP building so the building could not be inerted.
  

Concerning the impact of gap pressure and the respective behavior of “hot” fuel assemblies on “cold” fuel assemblies, the following conclusions are obtained from these ASTEC calculations (limitations are explained below):

the degradation of “cold assemblies” begins immediately after “hot assemblies” when “cold assemblies” are close to “hot assemblies”;

if “cold assemblies” are completely isolated from “hot assemblies”, the time before cladding rupture is not significantly impacted;

the impact of the gap pressure is not significant.
Limitations of the code

The geometrical modelling (axisymmetric assumption, suitable for core modelling) of fuel bundles in ASTEC is not appropriate to SFP geometry and approximations have to be done. Nevertheless, the most important parameter is the cooling capacity of fluid channels, therefore the fluid sections in the model. These approximations create uncertainties on the calculated cladding rupture time. Sensitivity studies have shown that it seems to be limited in comparison with the time between top fuel uncovering and the cladding rupture.

Figure 10.2. – Synthesis of ASTEC fuel degradation results

11.2.5CONCLUSIONS

This type of ASTEC calculations provides preliminary insights on the accident progression in case of SFP accident. They (of course) confirm that such accidents must be associated to very large release accidents in L2 PSA.

For the future, there is still some open area for research to identify some mitigation strategies:

effect of cooling by natural convection with air, by spray …

limitation of heat transfer from hot bundles to cold bundles if this can preserve the cold ones from melting,

feasibility to inert the fuel building against hydrogen explosion,

radioactive release filtration, …

controlling the criticality even if the geometry is modified.