QUENCH-06 Experiment Post-Test Calculations and Integrated Uncertainty Analysis With RELAP/SCDAPSIM/MOD3.4 and MOD3.5

Abstract

The Quench-06 experiment (ISP-45) has been used as a benchmark and training aid for Innovative Systems Software (ISS) and our users/students since it was completed in the early 2000s. The experiment was first analyzed by several international organizations using RELAP/SCDAPSIM/MOD3.2. These results were submitted to the “blind” and “open” phases of the ISP. The experiment was subsequently used for basic user training for experimental analysis by our RELAP/SCDAPSIM/MOD3.4 and MOD3.5 users. It is also used extensively in our university support and training internships. This paper describes an integrated uncertainty analysis of the QUENCH-06 electrically heated experiment, looking at the influence of uncertainties in experimental conditions and important models/correlations. The QUENCH calculations demonstrated the use of the new IUA, “Integrated Uncertainty Analysis”, option introduced into RELAP/SCDAPSIM/MOD3.4 in the summer of 2017 and MOD3.5 in the fall of 2017. The input models and results from both versions are discussed in the paper. The MOD3.4 results are based on the original input model developed for MOD3.2 and refined in the open phase of the ISP. The MOD3.5 results are based upon two base input models. The first was developed specifically to test the impact of MOD3.5 modeling improvements for the Quench electrically heated fuel rod simulator and was used in an early paper presented at this meeting in 2014. The second input model has been refined as part of the university support and training internship program and was used originally in 2016 to look at the influence of different approaches in modeling the insulated shroud used to minimize radial heat losses. The uncertainty analysis provided in this paper looks at the influence of uncertainties in (a) the parabolic equations for Zircaloy oxidation, (b) the tungsten heater element resistances, (c) the convective heat transfer coefficients, (d) the contact resistance of the heater elements, and (e) the thermal conductivity of the porous zirconia used in the shroud. The uncertainty analysis demonstrated very clearly that a bias was introduced into the 2014 MOD3.5 input model. This bias was subsequently determined to be primarily associated to the modeling of the shroud and associated radial heat losses. This bias was reduced in the 2016 version of the MOD3.5 input model and will be further refined as our training activities on the modeling of experiments continue.

Modeling and simulation of the QUENCH 12 experiment with the RELAP/SCDAPSIM/MOD3.5 code

“…A good agreement with the measured data is displayed. Similar, to the observations made in [37], there is an under-prediction of radial heat losses in the upper test section. As a result of the computed higher cladding temperature during the heat-up and pre-oxidation phases, the material phase transition temperature is reached more rapidly (which for the Sokolov's correlation is 1773 K).…”

Analysis of loss of coolant accident without ECCS and DHRS in an integral pressurized water reactor using RELAP/SCDAPSIM

“…The The specific features of IUA option have been described in a series of publicly published papers over the past several years including [17,18,23]. In very general terms, it allows RELAP/SCDAPSIM users to easily propagate the influence of both input and "source code" parameter uncertainties into user selected output quantities using a Monte Carlo based approach.…”

“…The SCDAP and RELAP5 based models used in RELAP/SCDAPSIM have been validated against a wide range of separate effects and integral experiments starting in the mid-1970s with some of the first fuel rod and bundle heating and melting experiments performed in Germany, and the subsequent electrically-heated and nuclear heated experiments performed in the United States, Europe, Russia, and Asia. The references [17][18][19][20][21][22][23][24][25][62][63][64][65][66] provide examples of validation activities specifically for RELAP/SCDAPSIM for these models and correlations.…”

“…As such, RELAP/SCDAPSIM was designed to minimize the influence of user effects by using detailed best estimate models, minimizing the ability of users to modify critical modeling parameters. In addition, as noted in Section 4, since the initial development and release of the code in the late 1990s, RELAP/SCDAPSIM has been used to (a) support the design and analysis of integral thermal hydraulic experiments performed in the French PHEBUS and German QUENCH facilities [17][18][19][20][21][22][23][24][25], (b) analyze a wide range of reactor designs including passive cooled designs like the AP1000 and smaller research reactors [26][27][28],…”