Mercurial > repos > public > sbplib
view +scheme/Heat2dCurvilinear.m @ 1327:7ab7d42a5b24 feature/D2_boundary_opt
Fix typo in comment
author | Vidar Stiernström <vidar.stiernstrom@it.uu.se> |
---|---|
date | Mon, 14 Feb 2022 10:49:49 +0100 |
parents | 8d73fcdb07a5 |
children |
line wrap: on
line source
classdef Heat2dCurvilinear < scheme.Scheme % Discretizes the Laplacian with variable coefficent, curvilinear, % in the Heat equation way (i.e., the discretization matrix is not necessarily % symmetric) % u_t = div * (kappa * grad u ) % opSet should be cell array of opSets, one per dimension. This % is useful if we have periodic BC in one direction. properties m % Number of points in each direction, possibly a vector h % Grid spacing grid dim order % Order of accuracy for the approximation % Diagonal matrix for variable coefficients KAPPA % Variable coefficient D % Total operator D1 % First derivatives % Second derivatives D2_kappa H, Hi % Inner products e_l, e_r d1_l, d1_r % Normal derivatives at the boundary alpha % Vector of borrowing constants % Boundary inner products H_boundary_l, H_boundary_r % Metric coefficients b % Cell matrix of size dim x dim J, Ji beta % Cell array of scale factors % Numerical boundary flux operators flux_l, flux_r end methods function obj = Heat2dCurvilinear(g ,order, kappa_fun, opSet) default_arg('opSet',{@sbp.D2Variable, @sbp.D2Variable}); default_arg('kappa_fun', @(x,y) 0*x+1); dim = 2; kappa = grid.evalOn(g, kappa_fun); m = g.size(); m_tot = g.N(); % 1D operators ops = cell(dim,1); for i = 1:dim ops{i} = opSet{i}(m(i), {0, 1}, order); end I = cell(dim,1); D1 = cell(dim,1); D2 = cell(dim,1); H = cell(dim,1); Hi = cell(dim,1); e_l = cell(dim,1); e_r = cell(dim,1); d1_l = cell(dim,1); d1_r = cell(dim,1); for i = 1:dim I{i} = speye(m(i)); D1{i} = ops{i}.D1; D2{i} = ops{i}.D2; H{i} = ops{i}.H; Hi{i} = ops{i}.HI; e_l{i} = ops{i}.e_l; e_r{i} = ops{i}.e_r; d1_l{i} = ops{i}.d1_l; d1_r{i} = ops{i}.d1_r; end %====== Assemble full operators ======== KAPPA = spdiag(kappa); obj.KAPPA = KAPPA; % Allocate obj.D1 = cell(dim,1); obj.D2_kappa = cell(dim,1); obj.e_l = cell(dim,1); obj.e_r = cell(dim,1); obj.d1_l = cell(dim,1); obj.d1_r = cell(dim,1); % D1 obj.D1{1} = kron(D1{1},I{2}); obj.D1{2} = kron(I{1},D1{2}); % -- Metric coefficients ---- coords = g.points(); x = coords(:,1); y = coords(:,2); % Use non-periodic difference operators for metric even if opSet is periodic. xmax = max(ops{1}.x); ymax = max(ops{2}.x); opSetMetric{1} = sbp.D2Variable(m(1), {0, xmax}, order); opSetMetric{2} = sbp.D2Variable(m(2), {0, ymax}, order); D1Metric{1} = kron(opSetMetric{1}.D1, I{2}); D1Metric{2} = kron(I{1}, opSetMetric{2}.D1); x_xi = D1Metric{1}*x; x_eta = D1Metric{2}*x; y_xi = D1Metric{1}*y; y_eta = D1Metric{2}*y; J = x_xi.*y_eta - x_eta.*y_xi; b = cell(dim,dim); b{1,1} = y_eta./J; b{1,2} = -x_eta./J; b{2,1} = -y_xi./J; b{2,2} = x_xi./J; % Scale factors for boundary integrals beta = cell(dim,1); beta{1} = sqrt(x_eta.^2 + y_eta.^2); beta{2} = sqrt(x_xi.^2 + y_xi.^2); J = spdiag(J); Ji = inv(J); for i = 1:dim beta{i} = spdiag(beta{i}); for j = 1:dim b{i,j} = spdiag(b{i,j}); end end obj.J = J; obj.Ji = Ji; obj.b = b; obj.beta = beta; %---------------------------- % Boundary operators obj.e_l{1} = kron(e_l{1},I{2}); obj.e_l{2} = kron(I{1},e_l{2}); obj.e_r{1} = kron(e_r{1},I{2}); obj.e_r{2} = kron(I{1},e_r{2}); obj.d1_l{1} = kron(d1_l{1},I{2}); obj.d1_l{2} = kron(I{1},d1_l{2}); obj.d1_r{1} = kron(d1_r{1},I{2}); obj.d1_r{2} = kron(I{1},d1_r{2}); % D2 coefficients kappa_coeff = cell(dim,dim); for j = 1:dim obj.D2_kappa{j} = sparse(m_tot,m_tot); kappa_coeff{j} = sparse(m_tot,1); for i = 1:dim kappa_coeff{j} = kappa_coeff{j} + b{i,j}*J*b{i,j}*kappa; end end ind = grid.funcToMatrix(g, 1:m_tot); % x-dir j = 1; for col = 1:m(2) D_kappa = D2{1}(kappa_coeff{j}(ind(:,col))); p = ind(:,col); obj.D2_kappa{j}(p,p) = D_kappa; end % y-dir j = 2; for row = 1:m(1) D_kappa = D2{2}(kappa_coeff{j}(ind(row,:))); p = ind(row,:); obj.D2_kappa{j}(p,p) = D_kappa; end % Quadratures obj.H = kron(H{1},H{2}); obj.Hi = inv(obj.H); obj.H_boundary_l = cell(dim,1); obj.H_boundary_l{1} = obj.e_l{1}'*beta{1}*obj.e_l{1}*H{2}; obj.H_boundary_l{2} = obj.e_l{2}'*beta{2}*obj.e_l{2}*H{1}; obj.H_boundary_r = cell(dim,1); obj.H_boundary_r{1} = obj.e_r{1}'*beta{1}*obj.e_r{1}*H{2}; obj.H_boundary_r{2} = obj.e_r{2}'*beta{2}*obj.e_r{2}*H{1}; %=== Differentiation matrix D (without SAT) === D2_kappa = obj.D2_kappa; D1 = obj.D1; D = sparse(m_tot,m_tot); d = @kroneckerDelta; % Kronecker delta db = @(i,j) 1-d(i,j); % Logical not of Kronecker delta % 2nd derivatives for j = 1:dim D = D + Ji*D2_kappa{j}; end % Mixed terms for i = 1:dim for j = 1:dim for k = 1:dim D = D + db(i,j)*Ji*D1{j}*b{i,j}*J*KAPPA*b{i,k}*D1{k}; end end end obj.D = D; %=========================================% % Normal flux operators for BC. flux_l = cell(dim,1); flux_r = cell(dim,1); d1_l = obj.d1_l; d1_r = obj.d1_r; e_l = obj.e_l; e_r = obj.e_r; % Loop over boundaries for j = 1:dim flux_l{j} = sparse(m_tot,m_tot); flux_r{j} = sparse(m_tot,m_tot); % Loop over dummy index for i = 1:dim % Loop over dummy index for k = 1:dim flux_l{j} = flux_l{j} ... - beta{j}\b{i,j}*J*KAPPA*b{i,k}*( d(j,k)*e_l{k}*d1_l{k}' + db(j,k)*D1{k} ); flux_r{j} = flux_r{j} ... + beta{j}\b{i,j}*J*KAPPA*b{i,k}*( d(j,k)*e_r{k}*d1_r{k}' + db(j,k)*D1{k} ); end end end obj.flux_l = flux_l; obj.flux_r = flux_r; % Misc. obj.m = m; obj.h = g.scaling(); obj.order = order; obj.grid = g; obj.dim = dim; obj.alpha = [ops{1}.borrowing.M.d1, ops{2}.borrowing.M.d1]; end % Closure functions return the operators applied to the own domain to close the boundary % Penalty functions return the operators to force the solution. In the case of an interface it returns the operator applied to the other doamin. % boundary is a string specifying the boundary e.g. 'l','r' or 'e','w','n','s'. % type is a string specifying the type of boundary condition. % data is a function returning the data that should be applied at the boundary. % neighbour_scheme is an instance of Scheme that should be interfaced to. % neighbour_boundary is a string specifying which boundary to interface to. function [closure, penalty] = boundary_condition(obj, boundary, type, symmetric, tuning) default_arg('type','Neumann'); default_arg('symmetric', false); default_arg('tuning',1.2); % nj: outward unit normal component. % nj = -1 for west, south, bottom boundaries % nj = 1 for east, north, top boundaries nj = obj.getBoundarySign(boundary); Hi = obj.Hi; [e, flux] = obj.getBoundaryOperator({'e', 'flux'}, boundary); H_gamma = obj.getBoundaryQuadrature(boundary); alpha = obj.getBoundaryBorrowing(boundary); Hi = obj.Hi; Ji = obj.Ji; KAPPA = obj.KAPPA; kappa_gamma = e'*KAPPA*e; switch type % Dirichlet boundary condition case {'D','d','dirichlet','Dirichlet'} if ~symmetric closure = -Ji*Hi*flux'*e*H_gamma*(e' ); penalty = Ji*Hi*flux'*e*H_gamma; else closure = Ji*Hi*flux'*e*H_gamma*(e' )... -tuning*2/alpha*Ji*Hi*e*kappa_gamma*H_gamma*(e' ) ; penalty = -Ji*Hi*flux'*e*H_gamma ... +tuning*2/alpha*Ji*Hi*e*kappa_gamma*H_gamma; end % Normal flux boundary condition case {'N','n','neumann','Neumann'} closure = -Ji*Hi*e*H_gamma*(e'*flux ); penalty = Ji*Hi*e*H_gamma; % Unknown boundary condition otherwise error('No such boundary condition: type = %s',type); end end function [closure, penalty] = interface(obj,boundary,neighbour_scheme,neighbour_boundary) % u denotes the solution in the own domain % v denotes the solution in the neighbour domain error('Interface not implemented'); end % Returns the boundary operator op for the boundary specified by the string boundary. % op -- string or a cell array of strings % boundary -- string function varargout = getBoundaryOperator(obj, op, boundary) assertIsMember(boundary, {'w', 'e', 's', 'n'}) if ~iscell(op) op = {op}; end for i = 1:numel(op) switch op{i} case 'e' switch boundary case 'w' e = obj.e_l{1}; case 'e' e = obj.e_r{1}; case 's' e = obj.e_l{2}; case 'n' e = obj.e_r{2}; end varargout{i} = e; case 'd' switch boundary case 'w' d = obj.d1_l{1}; case 'e' d = obj.d1_r{1}; case 's' d = obj.d1_l{2}; case 'n' d = obj.d1_r{2}; end varargout{i} = d; case 'flux' switch boundary case 'w' flux = obj.flux_l{1}; case 'e' flux = obj.flux_r{1}; case 's' flux = obj.flux_l{2}; case 'n' flux = obj.flux_r{2}; end varargout{i} = flux; end end end % Returns square boundary quadrature matrix, of dimension % corresponding to the number of boundary points % % boundary -- string function H_b = getBoundaryQuadrature(obj, boundary) assertIsMember(boundary, {'w', 'e', 's', 'n'}) switch boundary case 'w' H_b = obj.H_boundary_l{1}; case 'e' H_b = obj.H_boundary_r{1}; case 's' H_b = obj.H_boundary_l{2}; case 'n' H_b = obj.H_boundary_r{2}; end end % Returns the boundary sign. The right boundary is considered the positive boundary % boundary -- string function s = getBoundarySign(obj, boundary) assertIsMember(boundary, {'w', 'e', 's', 'n'}) switch boundary case {'e','n'} s = 1; case {'w','s'} s = -1; end end % Returns borrowing constant gamma*h % boundary -- string function gamm = getBoundaryBorrowing(obj, boundary) assertIsMember(boundary, {'w', 'e', 's', 'n'}) switch boundary case {'w','e'} gamm = obj.h(1)*obj.alpha(1); case {'s','n'} gamm = obj.h(2)*obj.alpha(2); end end function N = size(obj) N = prod(obj.m); end end end