I see. But is there a way to remove the nulspace explicitly in libmesh other than forcing the presure value at one node, such as substraction the average pressure value?
Currently I am facing a problem shown in http://sourceforge.net/p/libmesh/mailman/libmesh-users/thread/tencent_672F5BDA5899183B587C02E3%40qq.com/#msg33737791 In this case, when I make a parallel run, the pinned value added in compute_residual(...) can not be successfully updated by the TransientNonlinearImplicitsystem. Then all boundary pressures are all zero ,which causes a whole zero pressure field inside. There 's only one example of TransientNonlinearImplicitsystem, say, misc_ex3 for reference. So I am still trying to deal with it. Could you show any hints about it? Zhenyu ------------------ Original ------------------ From: "Roy Stogner";<[email protected]>; Date: Apr 7, 2015 To: "grandrabbit"<[email protected]>; Cc: "libmesh-users"<[email protected]>; Subject: Re: [Libmesh-users] Pin presure in lid-cavity case On Tue, 7 Apr 2015, grandrabbit wrote: > Later I found in systems_of_equations_ex2, even I comment out the part of > pinning pressure, > the code still can give a reasonable cavity flow, any body can show me why? > Thanks a lot. Without pinning the pressure, you have a non-invertible matrix. But it's not an overconstrained system with no solutions, it's an underconstrained system with an infinity of solutions, and in that case many iterative linear solvers will still find some solution to the system, one in which the final values of any underconstrained modes are just indirectly determined by the initial iterate. --- Roy ------------------------------------------------------------------------------ BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your own process in accordance with the BPMN 2 standard Learn Process modeling best practices with Bonita BPM through live exercises http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
