Modeling the blanket module Multiphysical computations and rapid

Modeling the blanket module: Multiphysical computations and rapid turnaround capabilities Ramakanth Munipalli, C. M. Rowell, K. -Y. Szema, P. -Y. Huang Hy. Per. Comp Inc. , 2629 Townsgate Rd. , #105, Westlake Village, CA 91361 Alice Ying, Neil Morley, Mohamed Abdou, Sergey Smolentsev, Manmeet Narula UCLA FNST Meeting, UCLA, August 14, 2008

Phase-I SBIR: CAD-Centric Management System for a virtual blanket module Period of performance: July 2008 – March 2009 Objective: To create a simulation management software with integrated prediction capability for blanket modules (extendable to the divertor and other PFCs) operating in a fusion environment • Conventional modeling involves using various codes to individually model physics • This is a CAD-centric multiphysical software integration project – no new physical modeling ability is sought in individual disciplines

Outline Ø Scope of the VTBM project Ø A description of the functionality and uniqueness of the VTBM environment Ø Some details on implementation Ø Phase-I tasks

Blanket Module: Location and attributes

Blanket Module: Details and flow features Key Physical Phenomena: Fluid flow, MHD (we will focus on liquid breeder) Heat transfer Structural mechanics Neutronics Tritium transport

Typical Input and Output data Inputs: Blanket geometry including front and return ducts: Toroidal width, radial depth, poloidal height, FCI dimensions, gap thickness Helium duct dimensions Inlet manifold geometry Surface Heating (function of poloidal coordinate) Volumetric heating as a function of (r, p) coordinates Temperature distribution in cooling He channels Heat transfer coefficient in He flows Thermophysical properties of Pb. Li, He, Fe, Si. C Inlet/Outflow Delta T in Pb. Li and in He Inlet Pb. Li velocity for each duct Magnetic field distribution Outputs: Velocity field in the bulk flow and in all sections of the gap MHD pressure drop Temperature distribution in Pb. Li, ferritic structure and FCI Temperature drop across the FCI Interface temperature between FCI and Pb. Li Calculated inlet/outflow temperature in Pb. Li for each duct Heat losses from Pb. Li into He flows Structural deformation Tritium concentration

VTBM: A description based on functionality Major VTBM attributes: · Project management – manage the entire simulation process from one convenient interface · · · Assistance in problem setup across multiple software platforms Time and Space coupling – “loosely” or “tightly” coupled simulation process Intelligent problem setup, maintain I/O schedules for each solver Conservative and accurate interpolation techniques – Fast octree approaches CAD-based data transfer: Redo CAD when geometry deforms Visualization of results

Some commercial MDA implementations AML (Technosoft Inc. ): Adaptive Modeling Language. Product, process development cycle integration, multidisciplinary modeling. Knowledge based engineering (KBE) framework that captures knowledge from the modeled domain and creates parametric models. ISIGHT (Engineous Software): Rapid integration of commercial and in-house simulation programs. Automates code executions. Optimization, design of experiments, quality engineering, visualization. Model. Center (Phoenix Integration): Visual environment for process integration. “Adaptable”. Design, archive, update the design process all in a visual environment. Process Data Management (PDM) tools help store information about process and design data. ANSYS Multiphysics: The capability is available. However, the basic survey shows that the usage in industry is virtually nonexistant. MDOPT (Boeing): CORBA based interdomain communication facility creates workflow criteria for multiphysical coupling and optimization.

User concerns · “Every business needs a great deal of customization” · “How does one troubleshoot a multiphysical solution: Who is the culprit? ” · “I would like to be able to go under the hood and perform diagnostics. The dash-board type control is insufficient” · “Existing commercial Multi Disciplinary Analysis (MDA) environments require a lot of customization before they can be integrated into a development cycle” · “A truly adaptable MDA environment does not currently exist”

Uniqueness of the VTBM approach · Physical phenomena encountered are extreme: Strong sensitivity to geometry changes, high EM interaction, large gradients in material properties, intense material interface effects in heat, current and mass transfer, multiscale coupling across physics · Our emphasis will be on the accurate and robust coupling of physics relevant to the fusion environment. · The graphical interfaces, as well as geometry generation and post-processing utilities will be customized to modeling blanket/heat-shield physics · Problem setup, and troubleshooting using an “intelligent” front-end Change in cross sectional velocity profile due to change in Si. C conductivity (left – 5 /ohm/m , right – 500 /ohm/m

Computational analysis tools of interest to VTBM

VTBM Work-flow diagram: problem setup

VTBM Work-flow diagram: Field solution and postprocessing

VTBM: Integrated physics modeling schematic (G)

Neutronics Treatment In the phase-I project neutronics data will be assumed to be given, computed from prior studies. We will use onedimensional distributions of power density as a source term in thermal and flow modeling.

Thermal analysis Traditional thermal analysis for non-conducting flows such as Helium and water will be performed using off-the-shelf third part software – motivated by their speed MHD flows with heat transfer and natural convection, including heat transfer in conducting solid walls will be computed using HIMAG. While numerous commercial codes are able to compute flow and heat transfer in complex structures, we will focus on the use of SC/Tetra in Phase-I. Future extensions will include FLUENT and CFdesign. Radiative heat input MHD (Pb. Li) Helium Thermal and electrically conducting wall

The exacting needs of numerical MHD FLOW 1 / sqrt(Ha) B U(z) j(y, z) 2 h d//≈Ha-1/2 1 / Ha Side layer d. Ha≈Ha-1 Hartmann layer

Hy. Per. Comp Incompressible MHD solver for Arbitrary Geometry Benchmarked against expt. data at fusion relevant conditions Second order convergence for high Hartmann no. flows Complex geometries, non-orthogonal, hybrid meshes, parallel computing

Development of an effective user interface - 1 Physics model editor (left), Graphical BC selection (below)

Development of an effective user interface - 2 Customization of the view pane based on simulation stage

VTBM : Use of CGNS as a common data file format In general, simulation codes use native grid/data formats. e. g. : TEMPUS-G IGES, STEP CAD files, and native TGP format HIMAG UX Unstructured Grid UGM Boundary patches and BC info SC/Tetra PRE FLD ANSYS CDB Computational Mesh including material regions Field data, including computational mesh and other physical information Common database format including mesh and field solution We seek to provide a common base-format to maintain simulation data throughout. CGNS (CFD General Notation System) is a likely candidate. It is open-source, well documented and used by various commercial software already

Template-based approach: Geometry parameterization nofelmts : Number of FCI channels noftube_l: Number of He channels along width

Template-based approach: Geometry creation

Template-based approach: Grid generation

Mesh generation by segregation of geometrical features Helium Ferritic Steel Pb 17 Li Si. C Helium Pb. Li

Interpolation techniques E Interpolation techniques: Point-element relations for standard interpolation Element-element relations based on intersection Point-point relations for matching grids/nearest neighbor P E 2 E 1 P Octree search procedures CAD based surface data interpolation is being developed

Coupling across physical disciplines Two types of data communication: All-to-all: n 2 -n interactions CAD Based: 2 n interactions Technical challenge: Conservation of forces, moments, etc. CAD Model MHD mesh Stress analysis mesh

VTBM : Data interpolation

VTBM : Data interpolation – contd. (a) (c) (b) (d)

NURBS (Non. Uniform Rational B-Spline) procedure for structural deformation CAD Surfaces are represented parametrically with NURBS as: u, v are parameters, Wij are weights, Bip is a B-spline of degree p at control point i Pij are locations of NURBS control points The field solver computes loads, deflections, etc. at discrete points rm These are projected onto the NURBS surfaces and corresponding um, vm are found (This is done at the grid generation stage itself) If the NURBS expression above is rewritten as: A new NURBS surface is fitted using least squares after deformation, using:

VTBM : “Appropriate” visualization of physics 2 -D dominant 3 -D dominant

Integrated TECPLOT for complex visualizations Customized post-processing: TECPLOT EDGE® Traditional TECPLOT layout can be customized to suit the application

VTBM – Software Development Issues Ø Dealing with third party software and APIs Ø Communication issues, resource management Ø Compliance with industry standards for I/O data Ø Interaction with the TBM community and timeline for development Ø Licensing, Documentation and Software dissemination Ø Using open source modules: CGM (Sandia, Argonne), CGNS (NASA)

VTBM – Object Oriented Software Development

Phase-I Project Objectives Task-1: TBM model assessment, redefinition of the VTBM concept in light of current developments in neutronics, existing template-based tools and advancements in CAD-coupling. Task-2: Development of a unified data flow system which will enable storage and transfer of simulation data across heterogeneous software relevant to TBM using CGNS and native data. Task-3: Physics-dependent accurate interpolation technique across computational meshes Task-4: Perform essential visualization procedures and plan automation Task-5: Development of a preliminary geometry deformation scheme for CAD/parametric model. Task-6: Verification and validation of the managed simulation technique on test problems. Task-7: Assessment of project needs and scope of a full scale implementation of the VTBM

Phase-I Timeline