Numerical difficulty arises when DAEs are solved with inconsistent initial
values of dependent variables. II can cause the solution failures of many p
opular DAEs solvers. An initialization subroutine, DAEIS (Differential Alge
braic Equations Initialization Subroutine), has been developed to handle th
is issue. In DAEIS, consistent initial values of dependent variables for in
dex-1 DAEs are obtained through solving a set of nonlinear equations based
on the initialization settings. To achieve good efficiency and robustness o
f the initialization computation, a nonlinear equation solver, GNES, has be
en specially built for DAEIS. The usage of DAEIS is demonstrated through ha
ndling inconsistent initializations for combined continuous/discrete simula
tions. (C) 2001 Elsevier Science Ltd. All rights reserved.