16.2 - Severe Accident Analysis (RRC-B) - EDF Hinkley Point

SUB-CHAPTER : 16.2PRE-CONSTRUCTION SAFETY REPORTCHAPTER 16: RISK REDUCTION AND SEVEREACCIDENT ANALYSESPAGE : 145 / 295Document ID.No.UKEPR-0002-162 Issue 04Despite the increase in the duration of the MCCI, the failure of the RPV-bottom in the “mixedmode” cases takes place at about the same erosion depth as in the corresponding layeredcases. This again confirms that, because the RPV-bottom and the MCCI pool form a coherentsystem, achieving the failure criterion of the RPV-bottom is more a function of erosion depththan interaction time. As Sub-section 16.2.2.4 - Figure 4 shows, the criterion for the failure of theRPV-bottom is achieved at an axial ablation depth of about 20 cm for all investigated cases.Moreover, it shows that, for a given initial melt mass, the influence of the actual decay heat levelon the ablation depth at the time of RPV-bottom failure is not significant.This feature of the chosen retention concept is attributed to the fact that the failure of theRPV-bottom takes place after it has absorbed a certain amount of energy, which - in the appliedmodel - is mainly emitted from the melt by thermal radiation. The time needed to absorb thisenergy is governed by the radiant heat flux at the melt surface. For a low heat flux, because of alow decay heat level in the melt, the time to achieve the failure temperature of the RPV-bottomis correspondingly increased. This is compensated by the fact that, in this case, the heat fluxinto the concrete, which determines the ablation rate, is also lower. As heat fluxes to the top andbottom in the gas-mixed molten pool are at least equal (may be higher to the top), both heat fluxvalues change in roughly the same proportion. Consequently, the failure of the RPV-bottomalways takes place at a similar axial ablation depth, independent of the actual level of decayheat. If the heat flux distribution within the MCCI pool is anisotropic, the energy transport to thetop would be proportionally increased, which would lead to an even earlier failure of theRPV-bottom.Considering that the calculated residual thickness of concrete at the position of the gate at thetime of the failure of the RPV-bottom is approximately 55% of the initial thickness, the marginswith respect to melt accumulation are high.Sub-section 16.2.2.4 - Figure 5 compares the results obtained for the generalised melt releasesequence with specific scenario-dependent release sequences, obtained with MAAP-4. Thefigure shows the bounding character of the generalised melt release sequences, in particular forLB(LOCA) and LOOP scenarios. The melt release sequences obtained with MAAP-4 for thesescenarios are covered by each of the generalised sequences analysed in the parametric studyindependently of whether the metallic and oxidic phases in the MCCI pool are assumed asmixed or layered.The melt release sequence for a SB(LOCA) scenario, which is characterised by a late first meltrelease 29 hours after scram, is consistent with the generalised melt release sequences thatassume a late first melt release after 1 day. Among the cases that assume stratified conditions,the case 60%L-L is most representative for the SB(LOCA) scenario and in good agreement withthe MAAP-4 prediction. The cases 40%L-L and 80%L-L bound the duration of melt releasecompared by MAAP-4. All cases that assume mixing of the metallic and oxidic melt phases yieldlonger release times and are thus bounding for the MAAP-4 SB(LOCA)-case. This verifies theassumption that the generalised melt release sequences lead to bounding melt release curvesas compared to those calculated with the integral code MAAP-4.First large pour consisting of metallic melt onlyThe results given in Sub-section 16.2.2.4 - Figure 6 show that for this case also the melttemperature is predicted to rise initially due to the oxidation of the Zr. Due to this, thetemperature increases with increasing Zr concentration in the initial melt. After the Zr content isdepleted, temperatures decrease to the solidifying temperature, where they remain for the restof the MCCI. Concurrent with this transient cooldown, the melt erodes into the concrete at a rateof about ~2.5x10 -4 m/s, almost independently of the melt initial temperature and Zr-content.Following the transient phase, the ablation front progression is quasi-steady and predominantlydriven by the decay heat, see Sub-section 16.2.2.4 - Figure 7.