The QSS Solver is a modeling and simulation environment for continuous and hybrid systems and it's optimized for large scale model simulation. The solver is implemented as a set of modules coded in plain C language, simulations can be performed using one of the following integration methods:
- QSS methods (the entire family is supported)
- DASSL
- DOPRI
- CVODE
- IDA
Models are described using a subset of the standard Modelica language called μ–Modelica (the language specification can be found here), the compiler extracts all the structural information needed and automatically generates the plain C code that contains an appropiate model definition needed by the QSS Solver engine together with all the configuration files (makefiles, config, etc.).
The QSS Solver has a simple GUI interface that allows the end-user to access all the developed tools in a unified way.
The GUI has the following features:
- It has a text editor, where models in μ-Modelica can be defined.
- It invokes the corresponding tools to compile and run simulations.
- It provides debug information in case of errors during the model generation.
- It invokes GnuPlot to plot the simulation output trajectories.
- It shows statistics about simulations (number of steps, simulation time, etc.).
[8] Ernesto Kofman, Joaquin Fernandez, Denise Marzorati. Compact sparse symbolic Jacobian computation in large systems of ODEs Applied Mathematics and Computation
[7] Franco Di Pietro, Joaquin Fernandez, Gustavo Migoni, Ernesto Kofman. Mixed-mode state–time discretization in ODE numerical integration Journal of Computational and Applied Mathematics
[6] Gustavo Migoni, Ernesto Kofman, Federico Bergero, Joaquin Fernandez. Quantization-based simulation of switched mode power supplies Simulation
[5] Federico Bergero, Joaquin Fernandez, Ernesto Kofman, Margarita Portapila. Time discretization versus state quantization in the simulation of a one-dimensional advection–diffusion–reaction equation Simulation
[4] Federico Bergero, Francesco Casella, Ernesto Kofman, Joaquin Fernandez. On the efficiency of quantization-based integration methods for building simulation Building Simulation
[3] Joaquin Fernandez, Ernesto Kofman, Federico Bergero. A parallel Quantized State System Solver for ODEs Journal of Parallel and Distributed Computing
[2] Federico Bergero, Xenofon Floros, Joaquin Fernandez, Ernesto Kofman, Francois Cellier. Simulating Modelica models with a Stand–Alone Quantized State Systems Solver Proceedings of the 9th International Modelica Conference
[1] Joaquin Fernandez, Ernesto Kofman. A stand-alone quantized state system solver for continuous system simulation Simulation
The MicroModelica compiler allows the user to describe models using a subset of the standard Modelica language called μ–Modelica (the language specification can be found here), the compiler extracts all the structural information needed and automatically generates the plain C code that contains an appropiate model definition needed by the QSS Solver engine together with all the configuration files (makefiles, config, etc.). The compiler documentation can be found here
The QSS Engine is a library that allows the simulation of continuous and hybrid systems and it's optimized for large scale model simulation. The solver is implemented as a set of modules coded in plain C language, where the simulated models can contain time and state discontinuities. Simulations can be performed using one of the following integration methods:
- QSS methods (the entire family is supported)
- DASSL
- DOPRI
- CVODE
- IDA
In order to obtain an executable model, a plain C file that contains the following functions must be provided:
- Model definition.
- Model settings.
- Model dependencies (for QSS methods).
- Model Jacobian.
- Model zero-crossing function.
- Model event handlers.
- A function in charge of the simulator data structure initialization.
The signature of these functions and the data structures that are used depends on the integration method used and they can ve viewed here
Finally, the file containing the model has to be compiled and linked with the engine library.
The QSS Solver GUI is a simple interface that allows the end-user of the QSS Solver to access all the developed tools in a unified way.
The GUI has the following features:
- It has a text editor, where models in μ-Modelica can be defined.
- It invokes the corresponding tools to compile and run simulations.
- It provides debug information in case of errors during the model generation.
- It invokes GnuPlot to plot the simulation output trajectories.
- It shows statistics about simulations (number of steps, simulation time, etc.).
The GUI documentation can be found here
MicroModelica packages with math helper functions and file manipulation functions that uses the C user libraries described bellow.
These packages are included in qss-solver, so that in order to use them no aditional annotation has to be used (such as package location, include paths, etc), just import the package into the corresponding model.
Available packages:
- math.mo: That contains math utility functions.
- file.mo: That contains file manipulation functions.
Both files are located in the packages folder in the root install location.
C libraries that contains helper functions used in MicroModelica packages. The Makefile included compiles the source files and generates the static library in the lib folder of the qss-solver. To compile a new user library, follow these steps:
-
Put the header file in the include directory.
-
Put the source file in the src directory.
-
Add the source file and the target library name in the Makefile.
-
Finally, run make to generate the corresponding static lib.
The generated lib is stored in the default path that the MMOC looks for, so the lib can be directly accesed from MicroModelica functions.
These are generic installation instructions.
In order to be able to install and compile the QSS Solver, the following dependencies must be installed (Ubuntu 22.04):
- bison++
- libboost-dev (boost1.71 in Ubuntu 20.04)
- cmake
- g++
- gfortran
- gnuplot
- libatlas-base-dev
- libcln-dev
- libconfig-dev
- libginac-dev
- libgsl-dev
- libsbml5-dev
- libscotch-dev
- libsuitesparse-dev
- pkgconf
- qtbase5-dev qt5-qmake (qt5-default in Ubuntu 20.04)
The simplest way to compile this package is:
-
cdto the directory containing the package's source code (src) and typemaketo compile the binaries and libraries. -
Type
make installto install all the binary files and the libraries. The binaries are located in the bin folder and the libraries are located in the lib folder. -
You can remove the program binaries and object files from the source code directory by typing
make clean.
The makefile script accepts the following options:
- MODE = <Debug|Release> When set to Debug (default), adds the compiler's debug flags.
The makefile script accepts the following targets:
- qss-engine: Builds the QSS solver libraries.
- mmo-compiler: Builds the MicroModelica compiler.
- qss-solver-gui: Builds the graphic interface.
- qss-user-libs: Builds the user packages included in this distribution.
- mmo-interfaces: Builds the SBML-MicroModelica translator.
- qss-solver-doc: Builds the documentation for the QSS solver libraries, the MicroModelica compiler and the QSS solver GUI.
- Added mLIQSS article models to distribution.
- Fix mLIQSS recompute next time to only update the Jacobian approximation when the influencer is a state variable.
- Allow quoted Q-IDENT variable definition in MicroModelica.
- Adapt reinit statements and for definition in algorithm sections to new Modelica specification.
- Added code generation for pure algebraic recursive dependencies.
- Added mLIQSS methods as experimental, they can be anebled from extended settings on the GUI.
- Reinit state assignments are now computed using the current state value instead of the quantized value.
- Fix terminate command code generation.
Please see the file called LICENSE.
Report bugs to: fernandez@cifasis-conicet.gov.ar