:
Main features of IDA Solver include:
|
| |
Modeling is input/output free, i.e. variables have no irrevocable
roles as given or calculated. Input/output free modeling
naturally leads to models described by equations rather
than the traditional calculation procedures, thus getting
closer to the physical relationships known to the modeler. |
| |
|
| |
The
system can handle algebraic as well as differential equations,
including algebraic loops. A range of powerful methods are
available for solution of the algebraic part of the problem,
making IDA suitable also for demanding static problems. |
| |
|
| |
Analytical
Jacobians are utilized when available. They are normally
automatically generated from the NMF model description. |
| |
|
| |
The
integration of dynamical systems uses variable time step
and order, for efficiency and for consistent, easy to use,
accuracy control. IDA Solver is based on MOLCOL methods
(Multistep One Leg COLlocation). |
| |
|
| |
Sparseness
in the system of equations is utilized effectively using
a variety of algorithms, including tearing methods. |
| |
|
| |
Models
can be precompiled and distributed as ready building blocks.
|
| |
|
| |
Models
may contain vectors and matrices, sizes of which may be
altered without component re-compilation. |
| |
|
| |
Discontinuities
in driving functions and in model equations are handled
properly. The time step is adjusted to hit points of state
and time events. Special solution methods are used to cross
discontinuities. |
| |
|
| |
Extensions
to the basic equation modeling allow handling of discrete
system states, as required by e.g. hysteresis. |
| |
|
| |
Discrete
time models with separate time step (often much shorter
than for continuous system). |
| |
|
| |
Sequentially
solved model partitions. Weakly coupled sub problems may
be solved in sequence, e.g., fluid concentrations are solved
after a central massflow-pressure system has been solved. |
| |
|
