Results 1 to 2 of 2

Thread: DAESL implementation problem

  1. #1
    Junior Member
    Join Date
    Sep 2014
    Posts
    1

    DAESL implementation problem

    Dear all,

    I am trying to change the solver for our system of ordinary differential equations from IVPAG to DAESL. My code is running flawlessly with the IVPAG solver, which requires the form y' = f(t,y). DAESL on the other hand requires an implicit formulation of the form 0 = g(t,y',y). Rewriting the system is thus easily done.

    DAESL is correctly initialized, and I call it with the following arguments:

    Code:
     call D_DAESL(t, tend, ido, Y, YPRIME, GCN, IYPR = 0, IUJAC = 0, MATSTR = 0, ISOLVE = 0, ML = neq - 1, MU = neq - 1, &
        RTOL = 1.d-4, H0=0.d0, HMAX=0.d-6, MAXORD = 5, MAXSTEPS = 10000)
    Note that IYPR = 0 because I am certain that YPRIME is initialized such that 0 = g(t,y',y).


    In the first time step I receive the error:

    Code:
    ***FATAL       ERROR 10 from DDAESL. The iteration matrix is singular
    I do not understand what the iteration matrix is. Does it refer to the Jacobian (termed partial derivative matrix in the documentation) of g?

    Thx for any suggestions.

    max

  2. #2
    Senior Member mecej4's Avatar
    Join Date
    Dec 2009
    Posts
    136
    It is not clear to me why you wish to use a DAE solver when your problem is just an ODE IVP. DAESL has a more complex calling sequence and many more usage options, and these features create more opportunities for introducing errors.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •