Presentation of a new software for the analysis of the Stirling cycle: GGSISM

In 2005 I was hired as a researcher at the Engeneering School of Seville (Spain), being responsible for the operation and maintenance of a dish-Stirling system. Then I began to study the Stirling cycle and to write a code for its simulation. This work has taken years, and the result, GGSISM (García-Granados SImulation for Stirling Machines), is now optimized and available.

GGSISM is a 3rd order model based program created for a main purpose: approaching the thermodynamic operation of Stirling cycle machines as much as possible without the use of complex CFD (Computational Fluid Dynamics) software, while keeping the simulation time within reasonable limits.

How does it work?

• The gas circuit is divided in 19 gas control volumes. Mass, pressure, temperature, and mass flow rate for each control volume are variable with time.
• Regenerator matrix is divided in 10 parts. The temperature of each part is time dependent too.
• Equations to solve: for each gas control volume we have the state equation, the mass balance, and the energy balance. Then, between continuous volumes, we have the momentum equations to take into account the pressure drop. On the other hand, for the regenerator matrix parts, there are the energy balances.
• Resolution: it is an initial value problem solved with Euler method. Energy balance for the parts of the regenerator matrix are solved separately. The other equations are ordered in linear form to be solved faster for each integration point. This linearization implies an “iterative” treatment for pressure drop. Note that the inclusion of pressure drop in the equations is closer to reality than estimating the work losses after the resolution, with no effect in absorbed or rejected heat.
• In order to compute energy balances, it is assumed a linear distribution of the temperature in each gas control volume of the regenerator and the expansion/compression exchangers.
• Friction factors and heat transfer coefficients are calculated for each integrating point according to the instantaneous values of Reynolds number and other parameters.
• At the end of each cycle the program evaluates energetic magnitudes and recalculates variables to avoid numerical errors.
• The program stops according to user's instructions (criteria for convergence or maximum number of cycles).

Some characteristics of the software:

• The program can accelerate the convergence, which makes it much faster. Although the computation time depends on many parameters, it usually takes less than half an hour for a normal PC.
• The user can choose from different correlations for friction factors and heat transfer coefficients.
• For alpha machines, it can work with senoidal or crank drives. For beta type machines, it can work with rhombic drive too.
• It allows different temperatures for different zones of the heat exchanger tubes.
• It works both for engines and refrigerators.
• You need MATLAB to run the program, but you may use the MATLAB Compiler to compile the “.m” files to a standalone application.
• Data and results files can be MATLAB files, or you can import/export from/to Excel files.
• Results: It produces a file with the evolution of the variables, energetic magnitudes and other parameters. It also asks the user to obtain plots interactively.
• There is a demo of the program available. It has limited capabilities.

Is the program validated?
Previous versions of the software have been used to simulate several Stirling engines and the results have been published:
a) García-Granados, Francisco. J., Silva-Pérez, M. A., Ruiz-Hernández, V., 2008. Thermal Model of the Eurodish Solar Stirling Engine, Journal of Solar Energy Engineering, 130, pp. 011014-1 – 011014-8.
The whole system is analysed and the Stirling engine as a part of it. Simulation for Solo 161 engine fits well with experimental data.
b) García-Granados, Francisco J., Silva-Pérez Manuel A., Prieto, Jesús-Ignacio, and García, David, 2009, Validation of a Stirling engine thermodynamic simulation program, 14th International Stirling Engine Conference and Exhibition, Groningen, The Netherlands.
For two Stirling engines: Vølund SM-1 and United Stirling P-40, simulations results are compared with experimental data and with simulations results from other three programs. In general terms, the present program produces lower errors than the others.

Acknowledgment: to the Stirling International Association, for offering me the possibility to release my work.

For more information:
Francisco J. García-Granados (Mechanical Engineer)
fjggranados@gmail.com