Salta al contenuto principale
Passa alla visualizzazione normale.

EMILIANO PIPITONE

Calibration of a knock prediction model for the combustion of gasoline-LPG mixtures in spark ignition engines

  • Autori: Beccari, S.; Pipitone, E.; Genchi, G.
  • Anno di pubblicazione: 2015
  • Tipologia: Articolo in rivista (Articolo in rivista)
  • OA Link: http://hdl.handle.net/10447/134434

Abstract

Gaseous fuels, such as liquefied petroleum gas (LPG) and natural gas (NG), thanks to their good mixing capabilities, allow complete and cleaner combustion than gasoline in spark ignition (SI) engines, resulting in lower pollutant emissions and particulate matter. In a previous work the authors showed that the simultaneous combustion of gasoline and LPG improves an SI engine efficiency with respect to pure gasoline operation with any significant power loss. The addition of LPG to the gasoline-air mixture produces an increase in knock resistance that allows running the engine at full load with overall stoichiometric mixture and better spark advance. In order to predict both performance and efficiency of engines fed by LPG-gasoline mixtures, a specific combustion model and in particular a knock prediction sub-model is required. Due to the lack of literature works about this matter, the authors investigated the knock resistance of LPG-gasoline mixtures. As a result, a reliable knock prediction sub-model has been obtained. The model can be easily implemented in thermodynamic simulations for a knock-safe engine performance optimization. The authors recorded light knocking in-cylinder pressure cycles on a cooperative fuel research (CFR) engine fueled by LPG-gasoline mixtures in different proportions. The tests were performed varying the compression ratio, the spark advance, and the inlet mixture temperature. The collected data have been used to calibrate and then compare two classical knock-prediction models. The models have been calibrated with a heterogeneous set of experimental data in order to predict knock occurrence in SI engines of different kinds. The results show that the models predict the knock onset position with a maximum error of around 6 crank angle degrees (CAD).