result = fmiModelExchange(fmu, tStop)
result = fmiModelExchange(fmu, tStop, dt)
| Parameter | Description |
|---|---|
| fmu | a string scalar or character row vector: the path to a .fmu archive, or to an already-extracted FMU directory. The FMU must provide the Model Exchange interface. |
| tStop | a positive real scalar: the stop time. The simulation starts at time 0. |
| dt | an optional positive real scalar: the fixed integration step. When omitted it defaults to tStop / 1000. Model Exchange usually needs a finer step than Co-Simulation because Nelson integrates the states itself. |
| Parameter | Description |
|---|---|
| result | a scalar structure with the fields time (N x 1), outputNames (1 x nOut) and outputs (N x nOut). |
fmiModelExchange imports a Functional Mock-up Unit (FMU) that follows the FMI 2.0 or 3.0 Model Exchange interface and integrates it with Nelson's own solver.
The key difference with fmiCoSimulate is who owns the solver. A Co-Simulation FMU contains its own solver and is advanced with doStep. A Model Exchange FMU exposes only the model equations (state derivatives, outputs and event indicators); the importing tool provides the solver. fmiModelExchange integrates the FMU's continuous states with a fixed-step fourth-order Runge-Kutta method and handles state events detected at step boundaries (entering event mode, running the discrete-update fixed point, and re-reading the continuous states).
No external inputs are applied: the parameters and inputs keep their start values. An error is raised when the FMU does not provide the Model Exchange interface.
fmu = [modulepath('nflow_fmi', 'root'), '/examples/VanDerPol.fmu'];
r = fmiModelExchange(fmu, 20, 0.01);
plot(r.time, r.outputs); legend(r.outputNames);| Version | Description |
|---|---|
| 2.0.0 | initial version |