UC Berkeley Cristhian Galvez, Nicolas Zweibaum, Per Peterson Thermal Hydraulics Laboratory Department of Nuclear Engineering University of California, Berkeley 2010 RELAP 5 International Users Seminar Design and Analysis of the PB-AHTR using RELAP5

UC Berkeley Outline Introduction Overview of Plant Design Modeling needs Plant system->process modeling breakdown Solution methodology Results Conclusion + Future Work

UC Berkeley Introduction The Pebble Bed Advanced High Temperature Reactor (PB-AHTR) is a pebble fueled, fluoride- salt cooled, 900-MWt reactor under development at UC Berkeley. Design features large thermal margins to fuel damage. Thermal limits are imposed by metallic primary loop structures. Peak core outlet temperature is the parameter of interest 1600°C Fuel failure fraction vs. temperature max. PB-AHTR temp

UC Berkeley Overview of Plant Design: Diagram

UC Berkeley Overview of Plant Design: 3D render Reactor Primary Pumps Recuperator Turbines Compressors Generators Intercoolers Precoolers Helium heaters Intermediate pumps Intermediate heat exchangers Intermediate drain tank

UC Berkeley Coolant Flow diagram Primary heat removal system composed of 4 Intermediate Heat Exchangers (IHX). Passive decay heat removal mechanism accomplished through 8 Direct Reactor Auxiliary Cooling System (DRACS). Heat is absorbed by the Direct Heat Exchanger (DHX), which is similar in design to the IHX and rejects heat to the environment through air- cooled Natural Draft Heat Exchangers (NDHX)

UC Berkeley Annular-type Core Annular core and components diagramLateral cross section

UC Berkeley Annular Pebble Bed core design Radially-zoned injection of buoyant pebbles Alternative injection of seed and blanket pebbles (axial zoning) Pebble Recirculation Experiment (PREX-2), 42% actual core size, high density polyethylene spheres, dry Radially and axially zoned pebble bed corePREX-2 filled with 129,840 pebbles

UC Berkeley Channel-type core Pebble channel assembly core and components Elevation view and lateral cross section

UC Berkeley Channel Pebble Bed core design Baseline design for lower half of PCA showing configuration of pebble channels

UC Berkeley Fuel and Coolant Flibe Primary Coolant (Li 2 BeF 2 ) Excellent heat transfer Transparent, clean fluoride salt Boiling point ~1400ºC Reacts very slowly in air No energy source to pressurize containment RELAP5-3D pebble fuel model description from pebble center (left) to pebble surface (right) for the annular pebble design

UC Berkeley Active cooling system: Intermediate Heat Exchanger (IHX) and pumps IHX Tube and shell, disk and doughnut baffled heat exchanger Primary coolant (Flibe) on tube side, Intermediate coolant (Flinabe) on shell side Forced convection on external and internal side driven by active centrifugal conventional pumps Derived from MSBR heat exchanger design

UC Berkeley Passive cooling system (DRACS): DHX, NDHX, Fluidic diode NDHX Tube and shell helical heat exchanger Natural circulation coolant (Flinabe) on tube side, Natural draft coolant (Air) on shell side Radiation heat transfer important Fluidic Diode Low resistance during forward flow, high resistance during reverse flow Passive operation DHX Tube and shell heat exchanger Forced primary coolant (Flibe) on shell side, Natural circulation coolant (Flinabe) on tube side Radiation heat transfer possibly important

UC Berkeley Reactivity control: Feedback mechanisms Fuel and Moderator Temperature Feedback Monte Carlo studies performed to determine fuel and moderator temperature reactivity feedback coefficients Study was done for various fuel-burn up levels, however RELAP5 analysis assumes average burn- up

UC Berkeley Reactivity control: Shutdown rod Shutdown-rod design Neutrally buoyant rod remains above the core during normal operation at typical coolant temperatures, but looses buoyancy and sinks into rod channel during above-normal coolant temperatures during transients Analytical and experimental work to determine rod insertion speed and rod worth

UC Berkeley Analysis Objectives Steady state: Mass, Pressure and Temperature distribution Transient: Peak core outlet temperature Safety system performance Decay heat removal system performance Experiment design analysis Design and analysis of the PB-AHTR requires investigation employing analytical, computational and experimental tools In order to obtain variables of interest and capture important phenomena, a methodology to breakdown the system and model it is used

UC Berkeley System-process breakdown: Core Core CoolantReflectorFuel 1- ϕ liquid pebble bed void volume Solid sphericalSolid cylindrical COEnergyCOMassCOMomCOEnergy -Energy generation -Conduction -Continuity-Convection-Form loss -Friction loss -Convection -Conduction Not available in current version of RELAP

UC Berkeley Active cooling Secondary PumpIntermediate Heat Exchanger IHX 1- ϕ liquid tube side volume Solid cylindrical 1- ϕ liquid pump volume COEnergy COMassCOMom COEnergy -Conduction-Continuity-Convection-Form loss -Friction loss -Momentum addition 1- ϕ liquid shell side volume Primary Pump 1- ϕ liquid pump volume COMom -Momentum addition System-process breakdown: Active Cooling

UC Berkeley Passive cooling Direct Heat Exchanger DHX 1- ϕ liquid tube side volume COMassCOMomCOEnergy -Continuity-Convection-Form loss -Friction loss 1- ϕ liquid shell side volume Natural Draft Heat Exchanger NDHX 1- ϕ gas shell side volume 1- ϕ liquid diode volume COMom System-process breakdown: Passive Cooling Fluidic diode -Form loss -Friction loss 1- ϕ liquid tube side volume

UC Berkeley RELAP5-3D Model Evolutionary steps taken to deal with modeling ‘gaps’ 1st: Input heat and flow loss coefficients manually –Only valid for steady state calculations 2nd: Input heat and flow loss coefficients manually as a function of time with self-consistent heat / flow loss coeficients and mass flow history –Approximation for transient 3rd: Manipulate existing LWR options in RELAP5-3D to add user-input factors to replicate correlation using multipliers (fouling factor for h and internal junction form loss for f) –Better approximation but still incomplete since power exponents of Re and Pr do not exactly match with available correlations coded in RELAP5-3D (Shah & ESDU cross flow) 4th: Implement pebble bed correlations into source code

UC Berkeley Annular Core RELAP5-3D Model 1/8 symmetric core modeled 3 multi-dimensional axial zones: inlet, mid-section and outlet Active mesh: inlet: 47, mid-section:81, outlet:32 Fixed T,P at coolant sources and fixed P at coolant sinks Power distribution resulting from coupling studies with MCNP5 Geometrical Configuration of the Core and the RELAP5-3D Model

UC Berkeley Annular core flow distrubition Core Diagram COMSOL FEM Multiphysics Model RELAP5-3D Model

UC Berkeley Outlet Temperature Parametric Analysis Inlets and Outlets Distributions in the Bottom, Mid-Section and Upper Core (a) (c) (b) (d)

UC Berkeley Outlet Temperature Distributions Temperature distributions of the outlets in different model variations

UC Berkeley Best Model Variation ΔT=97K, optimal difference Best model variation sketch and simulation result

UC Berkeley Channel Core RELAP5-3D Model

UC Berkeley Transient Description Several transients are analyzed, but focus of this study is Loss of Forced Circulation (LOFC) and Loss of Heat Sink (LOHS) LOFC involves the trip of the primary pumps, LOHS involves the trip of the intermediate pumps Both transients are evaluated under a different assumed safety system response 1.Normal scram immediately after primary or intermediate pumps. Shutdown rod bank inserted. 2.Failure to actively scram reactor with shutdown rods. Passive, buoyancy driven shutdown rod insertion occurs. Scram accomplished after a delay 3.Failure to scram reactor with either system. Power reactivity coefficient is the only mechanism present to shutdown the reactor

UC Berkeley Transient Results: Loss of Forced Circulation Fast loss of primary flow at t = 1000 s. Passive shutdown rod insert ~32 s after transient initiation. Average fuel and core outlet coolant temperatures rise to acceptable levels

UC Berkeley Transient Results: Loss of Forced Circulation Fast loss of primary flow at t = 1000 s. Flow within the Direct Heat Exchanger passively inverts shortly after the transient initiation. Steady state natural circulation for decay heat removal is rapidly obtained. Temperatures in metallic heat exchanger remain acceptable during severe transient

UC Berkeley Transient Results: Loss of Heat Sink Fast loss of intermediate flow at t = 1000 s. Passive shutdown rod insert ~32 s after transient initiation. Coolant temperatures rise to acceptable levels

UC Berkeley Transient Results: Loss of Heat Sink Fast loss of intermediate flow at t = 1000 s. Intermediate coolant flow is quickly reduced to negligible amounts. Thermal reactivity feedback shuts down the reactor quicker in the case of LOHS transients vs. LOFC transients.

UC Berkeley Conclusions RELAP5-3D model matches well with its analytical results, confidence in model exists for steady and transient conditions Passive and inherent reactor control mechanism perform well under postulated transients and maintain temperatures well below thermal damage limits for fuel (~1600 o C) and metallic structures (~765 o C) for Hastelloy 800 H Model provides preliminary insights on passive safety performance of the PB-AHTR. Additional work is necessary in order to consider other limiting cases such as 1) partial loss of flow 2) partial core flow blockage 3) partial heat exchanger flow blockage Need to configure the annular core model for transient simulations with optimized coolant outlet geometric distribution