Finite Element Clifford Algebra: A New Toolkit for Evolution Problems

Similar documents
Geometric Decomposition of Serendipity Finite Element Spaces

Serendipity Finite Element Methods for Cardiac Electrophysiology Models

New Model Stability Criteria for Mixed Finite Elements

A posteriori error estimates in FEEC for the de Rham complex

A Generalization for Stable Mixed Finite Elements

AposteriorierrorestimatesinFEEC for the de Rham complex

Introduction to finite element exterior calculus

SECOND ORDER TIME DISCONTINUOUS GALERKIN METHOD FOR NONLINEAR CONVECTION-DIFFUSION PROBLEMS

An abstract Hodge Dirac operator and its stable discretization

A space-time Trefftz method for the second order wave equation

Finite element exterior calculus: A new approach to the stability of finite elements

Smoothed projections in finite element exterior calculus

Finite Differences for Differential Equations 28 PART II. Finite Difference Methods for Differential Equations

Finite element approximation of the stochastic heat equation with additive noise

Finite Element Multigrid Framework for Mimetic Finite Difference Discretizations

Numerical Analysis and Methods for PDE I

Une méthode de pénalisation par face pour l approximation des équations de Navier-Stokes à nombre de Reynolds élevé

A primer on Numerical methods for elasticity

Preconditioned space-time boundary element methods for the heat equation

Lecture Note III: Least-Squares Method

Mixed Finite Element Methods. Douglas N. Arnold, University of Minnesota The 41st Woudschoten Conference 5 October 2016

A space-time Trefftz method for the second order wave equation

Basic Principles of Weak Galerkin Finite Element Methods for PDEs

An a posteriori error estimate and a Comparison Theorem for the nonconforming P 1 element

Motivations. Outline. Finite element exterior calculus and the geometrical basis of numerical stability. References. Douglas N.

INTRODUCTION TO PDEs

Invariant differential operators on the sphere

Past, present and space-time

ENERGY NORM A POSTERIORI ERROR ESTIMATES FOR MIXED FINITE ELEMENT METHODS

Finite Element Exterior Calculus. Douglas N. Arnold, University of Minnesota The 41st Woudschoten Conference 5 7 October 2016

arxiv: v1 [math.na] 29 Feb 2016

Implementation of Sparse Wavelet-Galerkin FEM for Stochastic PDEs

ELLIPTIC RECONSTRUCTION AND A POSTERIORI ERROR ESTIMATES FOR PARABOLIC PROBLEMS

SUPERCONVERGENCE PROPERTIES FOR OPTIMAL CONTROL PROBLEMS DISCRETIZED BY PIECEWISE LINEAR AND DISCONTINUOUS FUNCTIONS

A local-structure-preserving local discontinuous Galerkin method for the Laplace equation

A simplicial gauge theory on spacetime

Iterative methods for positive definite linear systems with a complex shift

Tutorial 2. Introduction to numerical schemes

Space-time sparse discretization of linear parabolic equations

A Primal-Dual Weak Galerkin Finite Element Method for Second Order Elliptic Equations in Non-Divergence Form

FINITE ELEMENT EXTERIOR CALCULUS FOR PARABOLIC EVOLUTION PROBLEMS ON RIEMANNIAN HYPERSURFACES

Numerical Methods for PDEs

Juan Vicente Gutiérrez Santacreu Rafael Rodríguez Galván. Departamento de Matemática Aplicada I Universidad de Sevilla

Yunrong Zhu: Curriculum Vitae

Sharp Sobolev Strichartz estimates for the free Schrödinger propagator

From Completing the Squares and Orthogonal Projection to Finite Element Methods

Finite Element Exterior Calculus and Applications Part V

A spacetime DPG method for acoustic waves

Adaptive methods for control problems with finite-dimensional control space

Analysis and Applications of Polygonal and Serendipity Finite Element Methods

Outline. 1 Motivating examples. 2 Hilbert complexes and their discretization. 3 Finite element differential forms. 4 The elasticity complex 1 / 25

A Mixed Nonconforming Finite Element for Linear Elasticity

Simple Examples on Rectangular Domains

Chapter 12. Partial di erential equations Di erential operators in R n. The gradient and Jacobian. Divergence and rotation

[2] (a) Develop and describe the piecewise linear Galerkin finite element approximation of,

AMS subject classifications. Primary, 65N15, 65N30, 76D07; Secondary, 35B45, 35J50

APPLIED PARTIAL DIFFERENTIAL EQUATIONS

Scientific Computing: An Introductory Survey

On Pressure Stabilization Method and Projection Method for Unsteady Navier-Stokes Equations 1

Chapter Two: Numerical Methods for Elliptic PDEs. 1 Finite Difference Methods for Elliptic PDEs

We denote the space of distributions on Ω by D ( Ω) 2.

Towards parametric model order reduction for nonlinear PDE systems in networks

Switching, sparse and averaged control

Efficient Solvers for Stochastic Finite Element Saddle Point Problems

MIXED FINITE ELEMENT APPROXIMATION OF THE VECTOR LAPLACIAN WITH DIRICHLET BOUNDARY CONDITIONS

Optimal control in fluid mechanics by finite elements with symmetric stabilization

Greedy Control. Enrique Zuazua 1 2

Basic Concepts of Adaptive Finite Element Methods for Elliptic Boundary Value Problems

Numerical Solutions to Partial Differential Equations

An introduction to Isogeometric Analysis

MULTIGRID PRECONDITIONING IN H(div) ON NON-CONVEX POLYGONS* Dedicated to Professor Jim Douglas, Jr. on the occasion of his seventieth birthday.

A Very General Hamilton Jacobi Theorem

(1) u (t) = f(t, u(t)), 0 t a.

A u + b u + cu = f in Ω, (1.1)

A POSTERIORI ERROR ESTIMATES FOR FINITE ELEMENT EXTERIOR CALCULUS: THE DE RHAM COMPLEX

Partial Differential Equations

Isogeometric Analysis:

1. Differential Equations (ODE and PDE)

Scientific Computing WS 2018/2019. Lecture 15. Jürgen Fuhrmann Lecture 15 Slide 1

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES)

New DPG techniques for designing numerical schemes

Is a dual mesh really necessary?

Weak Galerkin Finite Element Scheme and Its Applications

MEASURE VALUED DIRECTIONAL SPARSITY FOR PARABOLIC OPTIMAL CONTROL PROBLEMS

MIXED FINITE ELEMENT METHODS FOR LINEAR VISCOELASTICITY USING WEAK SYMMETRY

Space-time Finite Element Methods for Parabolic Evolution Problems

Preconditioning in H(div) and Applications

Weierstraß-Institut. für Angewandte Analysis und Stochastik. Leibniz-Institut im Forschungsverbund Berlin e. V. Preprint ISSN

Applied Math Qualifying Exam 11 October Instructions: Work 2 out of 3 problems in each of the 3 parts for a total of 6 problems.

Remarks on the analysis of finite element methods on a Shishkin mesh: are Scott-Zhang interpolants applicable?

Proper Orthogonal Decomposition. POD for PDE Constrained Optimization. Stefan Volkwein

Introduction. J.M. Burgers Center Graduate Course CFD I January Least-Squares Spectral Element Methods

Local flux mimetic finite difference methods

DS-GA 1002 Lecture notes 0 Fall Linear Algebra. These notes provide a review of basic concepts in linear algebra.

Yunrong Zhu: Curriculum Vitae

A Finite Element Method for the Surface Stokes Problem

Zdzislaw Brzeźniak. Department of Mathematics University of York. joint works with Mauro Mariani (Roma 1) and Gaurav Dhariwal (York)

Space-time XFEM for two-phase mass transport

Multiscale Diffusion Modeling in Charged and Crowded Biological Environments

Laplace s Equation FEM Methods. Jacob White. Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy, Jaime Peraire and Tony Patera

Transcription:

Finite Element Clifford Algebra: A New Toolkit for Evolution Problems Andrew Gillette joint work with Michael Holst Department of Mathematics University of California, San Diego http://ccom.ucsd.edu/ agillette/ Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 1 / 18

Motivation Poisson s equation: Given f find u(x) such that { 0 = u + f in Ω R n u = 0 on Ω a Heat equation: Given f and g, find u(x, t) such that u t = u + f in Ω R n, for t > 0, u = 0 on Ω, for t > 0, u t=0 = g in Ω Finite element exterior calculus (FEEC) provides: abstract framework for analyzing numerical approximation of elliptic PDEs classification of stable finite element methods with optimal convergence rates How can the FEEC framework be expanded to classify stable finite element methods for evolutionary PDEs? Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 2 / 18

Outline of approach Two possible methods for extending Finite Element Exterior Calculus: Semi-discrete: Finite element method in space, ODE in time t Ω domain: Ω [0, T ] R n R solution basis: φ h t=t0 : Ω R, for each t 0 [0, T ] error analysis: FEEC + Bochner space theory Fully discrete: Finite element method in space and time Ω domain: solution basis: error analysis: Ω [0, T ] R n R φ h : Ω [0, T ] R Finite Element Clifford Algebra This talk: Initial results on semi-discrete approach + a preview of FECA Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 3 / 18

Table of Contents 1 Motivation 2 Background: FEEC, Bochner Spaces, Semi-Discrete methods 3 Results: New error estimates in Bochner norms 4 Preview: Why FECA is needed Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 4 / 18

Outline 1 Motivation 2 Background: FEEC, Bochner Spaces, Semi-Discrete methods 3 Results: New error estimates in Bochner norms 4 Preview: Why FECA is needed Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 5 / 18

Finite Element Exterior Calculus in context Consider a mixed method for Poisson s problem on a domain Ω R n : continuous u + f = 0, u H 2 mixed weak (div σ, φ) + (f, φ) = 0, φ L 2 = Λ n (σ, ω) + (u, div ω) = 0, ω H(div) = Λ n 1 mixed FEM (div σ h, φ h ) + (f, φ h ) = 0, φ h Λ n h L 2 (σ h, ω h ) + (u h, div ω h ) = 0, ω h Λ n 1 h H(div) Major Conclusions from FEEC The finite elements spaces Λ n 1 h and Λ n 1 h should be chosen from two classes of piecewise polynomial spaces, denoted P r Λ k h and Pr If this choice is made in a compatible manner implied by the exterior calculus structure, then optimal a priori error estimates are guaranteed Λ k h Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 6 / 18

Finite Element Exterior Calculus in context Theorem [Arnold, Falk, Winther; Bulletin of AMS, 2010] Assume the elliptic regularity estimate u H s+2 + u H s+1 + σ H s+1 + div σ H s c f H s holds for 0 s s max. Choose finite element spaces Λ n 1 h =, Λn h = P r+1 Λ n 1 (T ) or P r+1 Λn 1 (T ) P r+1 Λn (T ) or P r Λ n (T ) Then for 0 s s max, the following error estimates hold { ch f L 2 if Λ n h = P 1 u u h L 2 Λn (T ), ch 2+s f H s otherwise, if s r 1 σ h σ L 2 ch s+1 f H s if { Λ n 1 h = P r+1 Λ n 1, s r + 1, Λ n 1 h = P r+1 Λn 1, s r, div (σ h σ) L 2 ch s f H s, if s r + 1. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 7 / 18

Semi-discrete Mixed Formulation Consider a mixed method for the heat equation on Ω R n for t I := [0, T ]. continuous u t u = f, u t=0 = g. mixed weak (u t, φ) (div σ, φ) = (f, φ), φ Λ n, t I, (σ, ω) + (u, div ω) = 0, ω Λ n 1, t I, u t=0 = g. mixed FEM (u h,t, φ h ) (div σ h, φ h ) = (f, φ h ), φ h Λ n h, t I, (σ h, ω h ) + (u h, div ω h ) = 0, ω h Λ n 1 h, t I, u h t=0 = g h. linear system AU t BΣ = F B T U + DΣ = 0 AU t + BD 1 B T U = F Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 8 / 18

Semi-discrete Error Bounds Theorem [Thomée; Galerkin FEM for Parabolic Problems, 1997] Fix n = 2 and set Λ 2 h := discontinuous linear, Λ 1 h := Raviart-Thomas elements. Let g h be the solution to the elliptic problem with f = g. Then for t 0: t ) u h (t) u(t) L 2 ch ( u(t) 2 H 2 + u t H 2 ds, 0 ( t ) 1/2 ) σ h (t) σ(t) L 2 ch ( u(t) 2 H 3 + u t 2 H ds. 2 0 Homogeneous case (f = 0), g h as above, t 0: u h (t) u(t) L 2 ch 2 g H 2, if g Ḣ2, σ h (t) σ(t) L 2 ch 3 g H 3, if g Ḣ3. Homogeneous case (f = 0), g h := orthogonal projection of g on to Λ 2 h, t > 0: u h (t) u(t) L 2 ch 2 t 1 g L 2 σ h (t) σ(t) L 2 ch 2 t 3/2 g L 2 Note: These bounds are space-only and restricted to the case n = 2. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 9 / 18

Bochner spaces and norms Our new error bounds will employ the theory of Bochner spaces Definition Let X be a Banach space and I = (0, T ). Define Equip this space with the norm C(I, X) := {u : I X u bounded and continuous} u C(I,X) := sup u(t) X. t I The Bochner space L P (I, X) is defined to be the completion of C(I, X) with respect to the norm: ( ) 1/p u L p (I,X) := u(t) p X dt. We combine notations to get Bochner differential form spaces: I L 2 X k := L 2 (I, L 2 Λ k (Ω)) These are parametrized differential form spaces. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 10 / 18

Outline 1 Motivation 2 Background: FEEC, Bochner Spaces, Semi-Discrete methods 3 Results: New error estimates in Bochner norms 4 Preview: Why FECA is needed Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 11 / 18

Bochner-FEEC Parabolic Error Estimates We combine the FEEC and parabolic error estimates to derive the following. Theorem [G, Holst, 2011] Let n 2 and fix I := [0, T ]. Suppose regularity estimate u(t) H s+2 + u(t) H s+1 + σ(t) H s+1 + div σ(t) H s c f (t) H s holds for 0 s s max and t I. Choose finite element spaces P r+1 Λ n 1 (T ) P Λ n 1 h = or P, r+1 Λn (T ) Λn h = or r+1 Λn 1 (T ) P r Λ n (T ) Then for 0 s s max and g h the solution to the elliptic problem we have ( ch f L 2 (I,L 2 ) + ) T f t L 1 (I,L 2 ) if Λ n h = P 1 Λn (T ) u h u L 2 X n ( ch 2+s f L 2 (I,H s ) + ) T f t L 1 (I,H s ) otherwise, if s r 1 and... Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 12 / 18

Bochner-FEEC Paraoblic Error Estimates Theorem [G, Holst, 2011] Let n 2 and fix I := [0, T ]. Suppose regularity estimate u(t) H s+2 + u(t) H s+1 + σ(t) H s+1 + div σ(t) H s c f (t) H s holds for 0 s s max and t I. Choose finite element spaces P r+1 Λ n 1 (T ) P Λ n 1 h = or P, r+1 Λn (T ) Λn h = or r+1 Λn 1 (T ) P r Λ n (T ) Then for 0 s s max and g h the solution to the elliptic problem we have: Λ n 1 h = P 1 Λ n 1 (T ), s 1 If or Λ n 1 h = P and Λn h = P 1 Λn (T ), then 1 Λn 1 (T ), s = 0 ( ) σ h σ L 2 X c h 1+s f n 1 L 2 (I,H s ) + h T f t L 2 (I,L 2 ) For any other choice of spaces, if s r 1, ( ) σ h σ L 2 X c h 1+s f n 1 L 2 (I,H s ) + h2+s T f t L 2 (I,L 2 ) Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 13 / 18

Proof and Significance Key idea of the proof: u(t) u h (t) L 2 }{{} error between weak and semi-discrete u(t) ũ h (t) L 2 }{{} error between weak and time-ignorant elliptic + ũ h (t) u h (t) L 2 }{{} error between time-ignorant elliptic and semi-discrete Significance of the error estimates These results give a priori estimates of convergence rates for the semi-discrete Galerkin FEM for the heat equation. By using the FEEC framework, we have classified choices of semi-discrete finite element spaces that guarantee optimal convergence rates. The results hold for arbitrary spatial dimension n, not just n = 2. For the homogeneous case (f = 0) with sufficiently regular g, we expect to find stronger error estimates akin to Thomée s. G, HOLST, Finite Element Exterior Calculus for Evolution Problems, in preparation. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 14 / 18

Outline 1 Motivation 2 Background: FEEC, Bochner Spaces, Semi-Discrete methods 3 Results: New error estimates in Bochner norms 4 Preview: Why FECA is needed Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 15 / 18

The Bochner Complex FEEC theory studies discretizations of the L 2 derham complex: 0 HΛ 0 d Ω (grad) HΛ 1 d Ω d Ω (div) HΛ n d Ω 0 We can define a parametrized exterior derivative operator on Bochner spaces: d : HX k HX k+1 where (dµ)(t) := d Ω (µ(t)). This gives rise to a Bochner domain complex: 0 HX 0 d HX 1 d d HX n d 0 Ω For a fully discrete method, we need an exterior derivative operator on spacetime elements which can distinguish spacelike and timelike dimensions. Such an operator needs the Lorentzian signature of basis elements - a tool available in Clifford Algebra (or Geometric Calculus) but not exterior calculus. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 16 / 18

Beyond the derham Complex... The derivative operator in Clifford algebra is a formal sum of d and its adjoint: := d + δ [0, T ] Λ k+2 d The derham complex appears as diagonals in a full Clifford complex The Bochner complex appears as parametrizations of these diagonals [0, T ] Λ k d δ Λ k+1 Λ k 1 δ d Λ k δ Λ k 2 Finite Element Clifford Algebra will study discretizations of this larger complex. Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 17 / 18

Questions? Slides and pre-prints available at http://ccom.ucsd.edu/ agillette Andrew Gillette - UCSD Finite ( ) Element Clifford Algebra SIAM PD11 - Nov 2011 18 / 18