# stability and nonlinear equation.

 Register Blogs Members List Search Today's Posts Mark Forums Read

 February 28, 2004, 10:56 stability and nonlinear equation. #1 bnedse Guest   Posts: n/a Hello, I want to know how I can check the stability of nonlinear system of equation. For example, if I have the following where everything is partial derivative: du_dt + f(u) du_dx = 0 then, how can I use Neumann stability analysis to analyize a particular discretization of this equation. Any help or references to textbook type information would be most kind.

 February 28, 2004, 11:26 Re: stability and nonlinear equation. #2 rvndr Guest   Posts: n/a Linearity of the equation is a general requirement for the application of the von neumann stability analysis. But in this case your equation is non-linear so , locally linearize the non-linear equation and subsequently apply the von Neumann stability analysis. I hope this will help. rvndr

 February 29, 2004, 18:37 Re: stability and nonlinear equation. #3 bnedse Guest   Posts: n/a O.k. Thank you for the help. How would you suggest I locally linearise it? Also, how can I show that if the equation is locally lineraised, that it reflects the non-linear version well? In other words, how do I see if linearising it destroys any relation to the orgiinal non linear equation.

 March 2, 2004, 12:23 Re: stability and nonlinear equation. #4 Jane Hosky Guest   Posts: n/a when dealing with linear equations you often only need to examine the behaviour of a single component, as all other will behanve similarly. linearisation of your equation depends on the function of u, i.e. 1D linearlised Burgers equation: du_dt + c du_dx = 0 where c is constant and positive. this is a hyperbolic equation and leads to the CFL condition. if you assume an explicit scheme in your original equation: du_dt + f(u) du_dx = 0 linearisation is trivial, by assuming old values. for an implicit scheme a linearisation method is needed, or alternatively an iteration within a time step can be used. look at the work by beam and warming where it is suggested to expand the flux function using a taylor series. hope this helps