2.1. Containment pressure suppression systems analysis for boiling water reactors (COPSAR)

The BWR containment is a complex system that includes many elements, which affect each other’s operation. The elements are e.g. the pressure suppression pool, spray and containment venting systems for containment pressure control, blowdown pipes for rapid steam condensation in case of a LOCA (Loss-of-Coolant Accident), in which the core is insufficiently cooled. There are also spargers for the pressure vessel relief valves, strainers for water supply to emergency core cooling and spray systems, nozzles and strainers of the residual heat removal (RHR) system, vacuum breakers, etc. There are a number of safety important scenarios, where containment pressure suppression function operation can be affected by (i) stratification and mixing phenomena, (ii) interactions with emergency core cooling system (ECCS), spray, RHR system, filtered containment venting system (FCVS), and (iii) overall water distribution between containment compartments. Such scenarios include (i) interplaybetween pool behaviour, diagnostics and operator procedures that can affect activation and performance of ECCS and containment spray systems; (ii) small LOCA; (iii) station blackout; (iv) leaking safety relief valve; (v) LOCA with broken blowdown pipe; (vi) severe accidents; and (vii) steam line breaks inside the radiation shield.

In the COPSAR project, the need is addressed for a validated modelling tool for simulation of realistic accident scenarios with interplay between phenomena, safety systems, operational procedures, and overall containment performance. Investigations of the phenomena that can affect pressure suppression function due to the operation of the other equipment and systems in the BWR containment are made. The experimental work is performed in the PPOOLEX facility at Lappeenranta University of Technology (LUT). PPOOLEX is a downscaled model of a BWR containment, which consists of a closed stainless steel vessel divided into two compartments, drywell and wetwell, that can be pressurized and equipped with different spray installations. Pre-test analysis and simulations for selection of operational regimes and test procedures is performed at KTH. Post-test analyses are made both at KTH and at VTT.

In a BWR, steam released from primary coolant system is condensed in the pressure suppression pool. Thermal stratification in the pool affects pressure suppression capacity of the pool. Heat and momentum sources generated by the steam condensation define pool behaviour. Direct Contact Condensation (DCC) of steam presents a challenge for contemporary modelling tools. In previous work, the Effective Heat Source (EHS) and Effective Momentum Source (EMS) models were proposed to simulate development of thermal stratification or mixing induced by steam injection into a large pool of water.

The experimental results from several tests in the PPOOLEX facility have been reported in the NKS report series; mixing tests with a residual heat remover (RHR) nozzle [2], sparger tests [3], spray tests [4], single spray nozzle tests [5]and sparger tests with reduced number of injection holes [6]. For the RHR nozzle tests particularly the effects of nozzle orientation, ∆T in the pool, injection water temperature and injection water mass flow rate were studied. The tests verified that orientation of an RHR nozzle plays an important role in the success of the mixing process of a thermally stratified pool. The injection water temperature and flow rate at the nozzle and ∆T in the pool have an effect on the mixing process but it is not as dominant as the nozzle orientation.

The analytical work by VTT and KTH is reported in several reports; e.g. pre-calculation of a PPOOLEX spray experiment [7] and CFD calculations on the thermal stratification and mixing in the sparger tests [8], modelling of pool behaviour and the Direct Contact Condensation (DCC) phenomena [9], [10].

LUT, KTH and VTT have studied suppression pool phenomena also in the ENPOOL project (2011–2014), see e.g. Refs. [11], [12].

High pressure in the containment due to reduced pressure suppression function is a safety concern for the containment integrity since it is the last barrier to the surrounding environment. The COPSAR and ENPOOL projects have provided important experimental data on pool behaviour addressing stratification and mixing issues that enable validation of computer models and realistic evaluation of safety margins in BWRs.

3.1. Debris coolability and steam explosions (DECOSE)

Lower drywell flooding is adopted in some Nordic BWRs as a means of mitigation of core melt accident consequences. It is assumed that upon reactor pressure vessel failure the melt will fragment, quench and form a debris bed coolable by natural circulation of water in a deep pool under the vessel. However, if the debris bed is not coolable, or if there is an energetic steam explosion when molten material meets water as it is released from the vessel, containment integrity can be threatened, potentially leading to release of radioactive materials into the environment.

VTT and KTH have collaborated in the DECOSE project to develop experimental facilities at both sites in order to produce data necessary for development of new models and codes for severe accident analysis. These codes and models can be used to address the long-standing technical issues of ex-vessel debris coolability and steam explosions in Nordic BWRs. The goal of the project was to reduce uncertainties in debris coolability and steam explosion impact. The project had significant co-funding, VTT through the SAFIR2018 program (The Finnish Research Program on Nuclear Power Plant Safety 2015–2018) and KTH through the APRI-MSWI program (Accident Phenomena of Risk Importance – Melt-Structure-Water Interactions).

A summary of the results of the experimental studies of debris bed coolability in the COOLOCE program at VTT is presented in Ref. [13]. The experiments addressed the effects of the debris bed geometrical shape, which is a result of the melt jet fragmentation and solidification in a water pool. Six variations of the debris bed geometry with different flooding modes were examined in the experiments, including a top-flooded cylinder and five beds with more complex, heap-like geometries. The experimental results in Ref. [14] suggest that the heap-like shape of the debris bed is favourable to coolability.

A study and comparison of data on coolability of particulate beds packed with irregular multi-size particles is presented in Refs. [15], [16]. The data from COOLOCE are compared with that of POMECO-FL and POMECO-HT facilities at KTH. The effective particle diameters obtained from the experiments are discussed, which is followed by the dryout heat flux comparison. It is suggested that the observed coolability variations are due to variations in porosity of the particulate beds resulting from the filling processes that were used in the experiments. It also includes the discussion over the possible factors (heaters’ geometry and porosity) affecting the coolability of the bed.

The DECOSIM code has been developed for analysis of porous debris coolability and further validated against COOLOCE data [17]. Debris bed cooling in post-dryout regime was addressed. Analytical models for prediction of the maximum temperature of the debris and relative size of the dry zone as a function of overheating parameter was proposed and validated against DECOSIM simulations. In the study, the DECOSIM code was extended to in-vessel coolability analysis.

In case of a core melt accident where the reactor pressure vessel fails, the melt ejected from the lower head into the water pool may be in the form of a continuous jet. A series of DEFOR-A (Debris Bed Formation – Agglomeration) tests was carried out in order to clarify the effect of the melt jet velocity on the particle size distribution and fraction of agglomeration [18], [19]. In these works phenomena relevant to the debris bed formation and coolability were also addressed experimentally and analytically.

A steam explosion is a fast fuel-coolant interaction that might occur as molten core material is released into the flooded lower drywell. Steam explosions have three distinct stages: premixing, triggering and propagation. In the premixing stage, the molten corium is fragmented into the coolant due to thermohydraulic forces. A large portion of the corium forms molten drops suspended in the coolant by vapour film. A triggering pulse locally collapses this instable liquid-vapour-liquid system. If the mixture properties are favourable, the trigger propagates in the mixture collapsing all the melt drops so that the thermal energy of the melt is almost instantly transferred to the coolant causing instantaneous high-pressure increase.

The effect of an ex-vessel steam explosion was analysed at VTT via computational models using the MC3D code, a multiphase CFD (computational fluid dynamics) code for fuel coolant interactions [20]. The focus of the analysis lies on the dynamic loads on the cavity wall imposed by the explosion. Simulations were made to analyse the effect of different triggering times on a standard case with central break at the lower head of the reactor vessel. The results showed that as long as the mixture properties allow the trigger to propagate then the resulting explosion behaves in a similar way. Different side breaks scenarios were also tested but here the mixture did not trigger. The sensitivity analysis was done for melt temperature, coolant subcooling, cavity water level and melt drop size. The results show that the parameter with the strongest effect is the drop size formed in the premixing phase, which is largely tied to the physical properties of the melt.

In an earlier sensitivity analysis [21], the three input parameters, melt temperature, triggering time and ambient pressure, were varied to see how changes in these affect the results. Among them, changing the triggering time had the largest impact on the steam explosion occurrence and strength. The melt temperature is not a limiting factor in this case. The ambient temperature had no effect on the explosion strength but instead on the probability of the explosion to occur.

A study performed at KTH deals with the premixing and explosion phase calculations of a Nordic BWR flooded drywell cavity, using MC3D [22]. The main goal of the study was the assessment of pressure build-up in the cavity and the impact loading on the sidewalls. The study included a sensitivity analysis of the parameters in modelling of fuel coolant interactions, to reduce uncertainty in assessment of steam explosion energetics. Moreover, the effects of water pool depth and jet velocity on the energetics of steam explosion were investigated. In an earlier study [23], it was shown that the amount of liquid melt droplets in the water is maximum even before the jet reaches the bottom of the pool. In the explosion phase, maximum pressure is attained at the bottom and the maximum impulse on the wall is at the bottom of the wall. The pressure attained and impulses on the wall are higher for bigger jet diameters, bigger melt droplets and higher subcoolings.

An approach for analysis of steam explosion in Nordic BWR and sensitivity of steam explosion impulse to the uncertain modelling and scenario parameters has been developed using the TEXAS-V code [17]. The Texas-V is a code for analysis of gas, liquid and fuel particles. It comprises two modules: one for calculation of premixing and another one for calculation of steam explosion. The first results indicated that the most influential parameters are water level and water temperature [24]. Extensive simulations using the TEXAS-V revealed that explosion impulse is a chaotic function of the triggering time [17]. The obtained database of impulse and pressure is used in the development of a computationally efficient tool for assessment of ex-vessel steam explosion risk in Nordic BWRs. This modelling approach is further developed in the ongoing SPARC project (see below).

3.2. The effect of aerosols and air radiolysis products on the transport of ruthenium (ATR)

The behaviour of the fission product ruthenium in a model primary circuit during a severe accident has been studied by VTT and Chalmers University of Technology in Sweden (CTH) in the ATR project [25]. The radiotoxicity of ruthenium oxides is similar to that of iodine in the short term and similar to that of caesium in the long term. All experiments were conducted using VTT’s Ru transport facility. The results show that the impact of additional NO2 gas feed (75 ppm volume) to the flow of ruthenium oxides (in humid air) was significant both on the transport of ruthenium through the facility and on the speciation of the transported ruthenium. Transport of gaseous RuO4 was increased significantly, whereas at the same time the amount of aerosols was decreased.

The following year the effect of additional air radiolysis products N2O, NO2, HNO3 [26] and CsI aerosol [27] on the transport of gaseous and particulate ruthenium species through a model primary circuit was studied. In the experiments, nitrogen oxides as well as nitric acid originating from air radiolysis had a significant effect on the ruthenium chemistry in the model primary circuit. The obtained results indicate a strong effect of air radiolysis products on the quantity of transported ruthenium and its partition to gaseous and aerosol compounds.