2. Release notes for JModelica.org version 1.6

2.1. Highlights

  • A new derivative free parameter optimization algorithm for FMUs

  • A new pseudo spectral optimization algorithm

  • Index reduction to handle high-index DAEs

  • A new graphical user interface for plotting of simulation and optimization results

  • Icon rendering and many improvements in the Eclipse Modelica plug-in

2.2. Compilers

2.2.1. Index reduction

High-index systems, commonly occurring in mechanical systems, are supported in JModelica.org 1.6. The implementation relies on Pantelides' algorithm and the dummy derivative selection algorithm.

2.2.2. Modelica compliance

The following improvements to the Modelica compliance of the editors has been made:

  • Partial support for the smooth() operator (not used in event handling, otherwise supported).

  • Support for global name lookup (i.e. names starting with a dot are looked up from the top scope).

2.3. Python

2.3.1. Graphical user interface for visualization of simulation and optimization results

A new graphical interface for displaying simulation and / or optimization results have been implemented. The interface also supports results generated from Dymola, both binary and textual.

2.3.2. Simulation with function inputs

The Python simulation interface has been improved so that top level inputs in FMUs can be driven by Python functions in addition to tables.

2.3.3. Compilation of XML models

A new convenience function for compilation of Modelica and Optimica models into XML, including equations, has been added.

2.3.4. Python version upgrade

The Python package has been updated to Python 2.7.

2.4. Optimization

2.4.1. Derivative- free optimization of FMUs

The derivative-free optimization algorithm in JModelica.org enables users to calibrate dynamic models compliant with the Functional Mock-up Interface standard (FMUs) using measurement data. The new functionality offers flexible and easy to use Python functions for model calibration and relies on the FMU simulation capabilities of JModelica.org. FMU models generated by JModelica.org or other FMI-compliant tools such as AMESim, Dymola, or SimulationX can be calibrated.

2.4.2. Pseudo spectral methods for dynamic optimization

Pseudo spectral optimization methods, based on collocation, are now available. The algorithms relies on CasADi for evaluation of derivatives, first and second order, and IPOPT is used to solve the resulting non-linear program. Optimization of ordinary differential equations and multi-phase problems are supported. The algorithm has been developed in collaboration with Mitsubishi Electric Research Lab, Boston, USA, where it has been used to solve satellite navigation problems.

2.5. Eclipse Modelica plugin

The JModelica.org Eclipse plugin has improved to the point where we are ready to do a release. Version 0.4.0 is now available from the JModelica.org website.

Changes from the versions that has been available from the SVN repository are mainly stability and performance improvements. To this end, some features have been disabled (auto-complete and format file/region). There are also a few new features, most notably support for rendering of class icons.

2.6. Contributors

Christian Andersson

Tove Bergdahl

Sofia Gedda

Magnus Gäfvert

Petter Lindgren

Fredrik Magnusson

Jesper Mattsson

Patrik Meijer

Lennart Moraeus

Kristina Olsson

Johan Ylikiiskilä

Johan Åkesson

2.6.1. Previous contributors

Philip Nilsson

Roberto Parrotto

Jens Rantil

Philip Reuterswärd