Prev: xPC Target boot disk not being recognized by Target PC
Next: Scaling of the spatial variable in function radon
From: Chan Chun K Chan on 7 Apr 2010 14:35 Hi! What about programming instead of using the solver? how to specify t=tstart? %this one apparently doesn't work. any suggestions? for k=1:M for i=1:N u(i,k)=u(i,1); end end Torsten Hennig <Torsten.Hennig(a)umsicht.fhg.de> wrote in message <1325632387.514251.1270622312919.JavaMail.root(a)gallium.mathforum.org>... > > How to specific boundary condition for heat equation > > that at one side the B.C. takes the form of dU/dt=0? > > thanks! > > This is a Dirichlet boundary condition: > U(t) = U(t=tstart) for all t. > Every solver should include an option to define > a Dirichlet condition as a boundary condition. > > Best wishes > Torsten.
From: Torsten Hennig on 7 Apr 2010 21:31
> Torsten Hennig <Torsten.Hennig(a)umsicht.fhg.de> wrote > in message > <1325632387.514251.1270622312919.JavaMail.root(a)gallium > .mathforum.org>... > > > How to specific boundary condition for heat > equation > > > that at one side the B.C. takes the form of > dU/dt=0? > > > thanks! > > > > This is a Dirichlet boundary condition: > > U(t) = U(t=tstart) for all t. > > Every solver should include an option to define > > a Dirichlet condition as a boundary condition. > > > > Best wishes > > Torsten. > > > Hi! What about programming instead of using the > solver? > how to specify t=tstart? > > %this one apparently doesn't work. any suggestions? > for k=1:M > for i=1:N > u(i,k)=u(i,1); > end > end Assuming your problem is one-dimenional and the boundary condition du/dt = 0 is specified at the right end (x_n) of the interval of integration, you have u(i,n) = u(1,n) for all times t_i, and u(1,n) is given by the initial condition for u. Does that help ? Best wishes Torsten. |