[Date Index] [Thread Index] [Author Index]

Initial conditions for DAE

• To: mathgroup at smc.vnet.net
• Subject: [mg65788] Initial conditions for DAE
• From: Maarten <maarten at stack.nl>
• Date: Mon, 17 Apr 2006 02:28:45 -0400 (EDT)
• Sender: owner-wri-mathgroup at wolfram.com

Dear fellow users,

I would like to use NDSolve to solve a DAE with a single algebraic variable. The algebraic variable has to be solved numerically, so SolveDelayed->True. The problem is that it has a root of multiplicity >1. When starting the integration I know the correct value for the algebraic variable and I would like to use this value (eg. 0.9) as a starting point for the numerical root finding that occurs during initialisation. However, when I specify an additional equality of the form:

AEvar[0]==0.9

in addition to the algebraic equation

f(AEvar[t])==0

NDSolve reports the initial conditions are inconsistent (they are overdetermined probably are initial conditions for all the differential variables are also defined). As far as I can see there are no options to pass NDSolve my initial guess, while this is a common feature in other DAE solving programs I have worked with? Is there a way of accomplishing this without resorting to FindRoot functions in my DAE problem formulation?

Best Regards,

Maarten Nauta