The MELTSPREAD Code for Modeling of Ex-Vessel Core Debris Spreading Behavior, Code Manual – Version3-beta

Abstract

MELTSPREAD3 is a transient one-dimensional computer code that has been developed to predict the gravity-driven flow and freezing behavior of molten reactor core materials (corium) in containment geometries. Predictions can be made for corium flowing across surfaces under either dry or wet cavity conditions. The spreading surfaces that can be selected are steel, concrete, a user-specified material (e.g., a ceramic), or an arbitrary combination thereof. The corium can have a wide range of compositions of reactor core materials that includes distinct oxide phases (predominantly 2 , 2 , and steel oxides) plus metallic phases (predominantly Zr and steel). The code requires input that describes the containment geometry, melt "pour" conditions, and cavity atmospheric conditions (i.e., pressure, temperature, and cavity flooding information). For cases in which the cavity contains a preexisting water layer at the time of RPV failure, melt jet breakup and particle bed formation can be calculated mechanistically given the time-dependent melt pour conditions (input data) as well as the heatup and boiloff of water in the melt impingement zone (calculated). For core debris impacting either the containment floor or previously spread material, the code calculates the transient hydrodynamics and heat transfer which determine the spreading and freezing behavior of the melt. The code predicts conditions at the end of the spreading stage, including melt relocation distance, depth and material composition profiles, substrate ablation profile, and wall heatup. Code output can be used as input to other models such as CORQUENCH that evaluate long term core-concrete interaction behavior following the transient spreading stage.MELTSPREAD3 was originally developed to investigate BWR Mark I liner vulnerability, but has been substantially upgraded and applied to other reactor designs (e.g., the EPR), and more recently to the plant accidents at Fukushima Daiichi. The most recent round of improvements that are documented in this report have been specifically implemented to support industry in developing Severe Accident Water Management (SAWM) strategies for Boiling Water Reactors. This document is a Code Manual which contains i) a technology review, ii) descriptions of models and correlations, iii) descriptions of the implicit, finite difference numerical solution scheme, iv) a user's guide, and v) a summary of code validation calculations that have been performed to date.

A Case Study on Severe Accident Water Addition and Water Management for a MARK I Containment

“…One of the basic modelling assumptions within MELTSPREAD (and CORQUENCH) is that the fission products predominately reside in the oxide phase. The modelling approach [6,7] is thus to partition the decay heat in the core debris in direct proportion to the amount of fuel (i.e. UO2) that is locally present.…”

“…With this background, the specific objective of this study is to apply these upgraded modeling tools (i.e., MELTSPREAD [6] and CORQUENCH [7]) to perform a parametric case study on SAWA/SAWM using the well-studied Peach Bottom nuclear power plant as the basis (see Figure 1-2) in order to provide insights related to cavity flooding strategies. To this end, Section 2 summarizes the ex-vessel melt pour conditions predicted by MAAP and MELCOR for an SBO sequence at Peach Bottom.…”

“…Inlets to the 'four' (i.e., two symmetrically combined) downcomers are uniformly distributed at discrete nodes within the drywell in order to capture water overflow to the wetwell during the spreading phase for those cases in which water is initially present on the drywell floor. The new detailed water inventory model [6] was used for all cases with water initially present.…”

“…The heat transfer coefficient from the melt to the surface of the core debris is calculated as the larger of the following two correlations: i) the well-known Dittus-Boelter forced convection correlation with the hydraulic diameter based on local melt depth, and ii) Bradley's bubble agitation-driven heat transfer model that is applicable to MCCI conditions [16]. This is the same modelling option used in the MELTSPREAD validation exercises [6] for cases involving melt spreading over concrete surfaces. With this assumption, the code is expected to provide a best estimate of the melt penetration distance in containment.…”

“…Thus, the current work has further aimed to automate the multi-nodal CORQUENCH analysis capability, and to interface this capability with a realistic water inventory model to provide an integrated modelling framework for assessing longterm water throttling strategies for BWRs. The status of these model improvement and validation efforts have been documented in several technical reports [5][6][7].…”