Author/Authors :
Akbari, H Vehicle, fuel and Environment Research Institute - University of Tehran, Tehran, Iran , Salavati-Zadeh, A Vehicle, fuel and Environment Research Institute - University of Tehran, Tehran, Iran , Javaheri, A Vehicle, fuel and Environment Research Institute - University of Tehran, Tehran, Iran , Esfahanian, V School of Mechanical Engineering - University of Tehran, Tehran, Iran , Ghavami, S.V Vehicle, fuel and Environment Research Institute - University of Tehran, Tehran, Iran , Ghomashi, H Faculty of Engineering - Islamic Azad University, South Tehran Branch, Tehran, Iran
Abstract :
The present study aims to reduce the computational cost of in-cylinder phenomena simulation under the light
of employing Proper Orthogonal Decomposition (POD) technique. The equivalence ratio as the main
identifier for soot formation tendency along with temperature and nitrogen oxide fields, are studied inside a
gas-fuelled engine. The required correlation matrix is built based on ten snapshots obtained from the results
of engine three-dimensional simulation, which are verified based on experimental data. The AVL-FIRE
v.2013 software is used to carry out the three dimensional simulations. The flow field at 3250 rpm is then
estimated by POD coefficients and subsequent curve fittings. To validate the reduced order results, this
condition is simulated by the software. For instance, temperature and equivalence ratio fields at top dead
center and 5 degree after top dead center are compared. The relevance index for equivalence ratio indicates
about 96% consistency between reduced order and 3D simulation results. On the other hand, this index is
found to be about 99% at both crank angels for temperature, which proves a more coherent structure in the
temperature field than that of equivalence ratio. Meanwhile, the analysis of 3D simulation results by POD
demonstrates a more coherent structure for the in-cylinder flow regime at top dead center. This consistency is
obtained in spite of computation time of POD being approximately 1% of 3D simulation time.