Characterisation of diesel fuel injection rate profile by means of a metaheuristic approach

Abstract

The fuel injection rate is notably one of the most significant parameters in compression ignition engine operation both for diesel only and dual fuel combustion. For engine combustion simulation, the injection profile needs to be well defined. The relevance of the injection rate is due to its influence on fuel delivery, spray breakup, atomization and the injector influence on the actual start (SOI) and duration (DOI) of fuel delivery. Unfortunately, the injection rate is difficult to measure experimentally. The modelling of injectors using hydromechanical sub-models within 1D engine modeling is possible but is dependent on accurate physical measurements. This work presents an approach of determining the injection rate using a metaheuristic methodology. An algorithm implemented in MATLAB controls a sequence of simulations in ANSYS Forte to determine the rate profile which creates the closest match between simulation and reference heat release data. A decision tree is used to generate candidate profiles and may be adjusted to control the variation of candidates. A variable neighborhood search is applied by testing the candidate profiles, while subsequently optimizing each profile SOI and DOI to determine local minima. The residual sum of squares of the heat release rate is analyzed to identify a global minimum which leads to the rate profile with the best match between simulation and reference data.