Safety analyses with uncertainty quantification for fusion and fission nuclear power plants. Applications to EU DEMO fusion reactor and BWRs

The EUropean DEMOnstrating fusion power reactor (DEMO) Water Cooled Lithium Lead (WCLL) breeding blanket concept is currently in its pre-conceptual design phase by the EUROfusion consortium members. It aims to be the first tokamak fusion reactor to demonstrate the capability for net electricity production, tritium-self-sufficiency, and a lifetime plasma operation of several full-power years. The WCLL breeding blanket is one of the two concepts being studied for implementation in the EU DEMO reactor. This concept relies on the separate-cooled architecture, where the liquid metal is utilized exclusively as tritium breeder and neutron multiplier, whereas the role of coolant is fulfilled by pressurized water. Following the previous experience of the experimental International Thermonuclear Experimental Reactor (ITER), the design project of the DEMO reactor is constantly supported by safety studies for all the different breeding blanket concepts under investigation, so that best performance within safety requirements are achieved. The basic goal of safety is to ensure that a nuclear reactor will not contribute significantly to individual and societal health risks. These risks stem mainly from the radioactive inventory inside the reactor, so this basic goal translates into the prevention of radioactive material releases toward the environment. A secondary, but fundamental, objective is to prevent damage of fusion power plant main components. To deal with safety requirements within the DEMO project, a widely use of passive safety systems (a smart mix with active safety systems) will be made together with established safety principles, such as defense in depth and maintaining doses as low as reasonably achievable (ALARA). The research activity described in this Ph.D. dissertation has the main objective to characterize and quantify the safety and environmental aspects of the EU DEMO WCLL concept design, studying the reactor response to some of the most severe possible accident scenarios. Safety analyses are performed with the fusion adapted versions of MELCOR code, to investigate the thermal hydraulic behavior of DEMO main components and radioactive source term mobilization. Moreover, the performed safety analyses, supported by sensitivity studies, could be useful to provide insight into physics and technology issues that need addressing to develop fusion as an optimal electricity generation alternative in the near future. In this early development phase of the DEMO design, in the frame of the EUROfusion safety working project (WPSAE), a list of initiating events, which could start an accident sequence, has been identified through a Functional Failure Mode and Effect Analysis (FFMEA). Many accident sequences enveloping from these initiating events have been investigated comprehensively. The dissertation is divided in three main parts. The first part concerns a general description of the principal safety issues associated with fusion reactors and provides an overview description of main EU DEMO components from a safety perspective. The second part contains preliminary safety studies relating to design basis accidents, beyond design basis situations and hydrogen mitigation systems. The progression of design basis accidents has been simulated following a conservative approach taking into consideration the passive and active accident mitigation capabilities of the plant. Four different loss of coolant accident (LOCA) scenarios have been studied: in-vessel LOCA; ex-vessel LOCA; single in-vessel LOCA; and an in-box LOCA. Concerning the multiple in-vessel LOCA, parametric analyses have been performed to determine the minimum flow area required by the suppression system pipework to limit the vacuum vessel (VV) pressure below the limit of 2 bar imposed as requirement by safety. Moreover, because limiters could be introduced in future design of the EU DEMO reactor to prevent the plasma to touch the breeding blankets plasma facing components (PFC), the same parametric study has been performed to evaluate their accident mitigation effects. In this framework, a new vacuum vessel suppression system (VVPSS) concept has been proposed following the ITER experience. It is based on six separated suppression tanks located in the containment basement, one of them is dedicated to retaining small leakages. The pipework consists mainly of six bleed lines connecting the VV to the small leakage tank, and five rupture disks line one for each suppression tank. To avoid steam and radioactive flows inside neutral beam ports, pipework connecting the vacuum vessel to the suppression system has been attached to the upper port. This last choice caused the necessity of a detailed nodalization of in-vessel volumes to model correctly steam flow path from VV to VVPSS and relative convective heat transfer effects between the modules’ back supporting structure (BSS) and the steam flowing at high velocity in the interspace volume between the BSS and the VV. Relatively to a simple in-vessel event involving the rupture of 10 first wall cooling pipes, two different simulations have been performed to evaluate downstream isolation valves' effects in terms of radioactive releases and thermal hydraulic behavior of main DEMO components. In fact, the large number of downstream valves (isolation and Safety Relief Valves (SRV) to be installed, could give rise to safety and reliability constraints. Concerning ex-vessel LOCA events, a very unlikely double-ended pipe rupture is postulated in a coolant distributor ring of the EU DEMO reactor. The fusion power termination system is assumed to terminate the plasma burn with a mitigated disruption. Two different simulations have been performed related to failure in FW-PHTS and BZ-PHTS, respectively. However, due to its similarity, only the results of the former are described. The objective of these analyses is to show that the accident consequences are within the safety requirements for tokamak building structures which must withstand large internal pressures as well as avoiding significant leak rate into the environment. Because the tokamak building layout of the EU DEMO is currently in a preliminary design phase, parametric studies have been performed to support design activities. A preliminary, but quite detailed, model of the TCR was made to take into account steam condensation phenomena on TCR walls, being the only available effect for mitigating the overpressure in the TCR. In fact, no active systems for containment cooling are currently foreseen for DEMO. To complete the wide range of DBA performed for the EU DEMO WCLL concept, a preliminary analysis of an in-box LOCA has been carried out. This kind of accident has not been yet deeply investigated for fusion reactors because of the lack of multi-phase safety-related system codes able to deal with water and liquid metals. To overcome this code limitation, a Python script has been developed for an external coupling of two MELCOR input decks working with different fluids. At each user-imposed time step the information of one MELCOR run are extracted and used as feedback for the other input deck. In the framework of BDBA an ex-vessel LOCA and a Loss of Flow Accident (LOFA) have been studied with the objective to show the robustness of the defense in depth approach and demonstrating that no cliff edge effects occur in the safety analysis. In both the simulations the failure of active plasma shutdown system has been assumed as aggravating event. Differently from DBA these accident analyses should be performed using best estimate assumptions and not conservative ones. At this purpose, the failure temperature of FW structure is increased from 1273 K to 1598 K. However, this parameter results very correlated, in particular for the ex-vessel LOCA simulation, with the amount of tritiated water and other radioactive aerosols that could be released toward the external environment. Instead, preliminary safety analyses for the LOFA beyond event, highlighted that the major safety concern is not related to radiological releases, but to the huge pressurization of in-vessel components, for such a reason this accident has been simulated using a lower pressure setpoint for safety relief valves. Three different simulations have been performed, by changing the number of FW channels affected by the rupture. In-box LOCAs, as well as other accidents involving a chemical reaction between hot steam and lead lithium, could led to the production of large amounts of hydrogen inside the tokamak vacuum chamber. In order to avoid that flammable concentrations could be achieved, the production of hydrogen must be limited and properly monitored. In particular, the simultaneous presence of hydrogen and dust in the VV volume enhances the risk of explosion. After a short description of possible technical solutions suitable for EU DEMO to mitigate hydrogen concentration, preliminary accident study involving the use of passive autocatalytic recombiners (PARs) are reported. The MELCOR ESF package has been activated to simulate the presence of PARs directly installed in the atmosphere of the VVPSS suppression tanks. Successively, in the third part, sensitivity and uncertainty analyses are reported. Because severe accidents in both fission and fusion power plants involve a wide range of uncertain phenomena and parameters, sensitivity and uncertainty analysis have to be performed to evaluate the influence of input parameters on selected figures-of-merit (FoMs). At this purpose a Python interface has been developed to allow the interaction between RAVEN and MELCOR. The Python interface allows to perturb all the parameters accessible through the MELCOR input deck. In such a way RAVEN is capable to investigate the system response as well as the input space using sampling schemes. Two sensitivity and uncertainty analyses have been performed, with applications to EU DEMO reactor and to the unit 3 of the Fukushima Daiichi power plant.