In fact I hope to use a single Jacobian for both variables which is a submatrix originally formed from the full matrix of k & omega model. Since the time derivatives, convective terms and diffusion coefficients are identical for both k & omega, then the jacobian of k should be the same as the one of omega.
As suggested above, if I define a linearImplicitSystem, the default jacobian includes both the above submatrices. But in my case, I think a submatrix should do the work and save the memory costs as well. But How? Thanks. Zhenyu -----原始邮件----- 发件人: "Kirk, Benjamin (JSC-EG311)" <benjamin.k...@nasa.gov> 发送时间: 2015-10-03 21:38:35 (星期六) 收件人: Zhang <zyzh...@nuaa.edu.cn> 抄送: 主题: Re: [Libmesh-users] How to share the same jacobian matrix for different scalar unkown Sure - you'll just declare an additional linear implicit system (or derived nonlinear one) with two coupled variables - that will allocate a angle Jacobian that coupled he two variables. This assumes you want that decoupled from your principal Navier-Stokes system. -Ben > On Oct 3, 2015, at 3:53 AM, Zhang <zyzh...@nuaa.edu.cn> wrote: > > Hi, > > > I am using libMesh to write a RANS code with two-equation turbulence model. > > > Since both turbulence variables are introduced into an advection-diffusion > equation, may I > use a single Jacobian matrix shared by both of them? > > > If yes, how? > > > Thank you very much for your attention. > > > Zhenyu Zhang > > > > ------------------------------------------------------------------------------ > _______________________________________________ > Libmesh-users mailing list > Libmesh-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/libmesh-users ------------------------------------------------------------------------------ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
