Parameter Identification of a Quasi-Dimensional Spark-Ignition Engine Combustion Model 2014-01-0385

Parameter identification of a math-based spark-ignition engine model is studied in this paper. Differential-algebraic equations governing the dynamic behavior of the engine combustion model are derived using a quasi-dimensional modelling scheme. The model is developed based on the two-zone combustion theory with turbulent flame propagation through the combustion chamber [1]. The system of equations includes physics-based equations combined with the semi-empirical Wiebe function.

The GT-Power engine simulator software [2], a powerful tool for design and development of engines, is used to extract the reference data for the engine parameter identification. The models is GT-Power are calibrated and validated with experimental results; thus, acquired data from the software can be a reliable reference for engine validation purposes. Homotopy optimization procedure, in which the original differential equations are modified by coupling the experimental data to the mathematical model using a homotopy parameter and gains, has been utilized in this work to obtain a global minimum for the parameters giving the best match to experiments [3]. Algebraic equations in the mathematical model of the engine make the process of optimization more complicated, as the algebraic relations should be satisfied in updating the initial conditions at every step of the applied optimization procedure.

The parameters chosen are difficult to estimate due to the lack of a tangible physical significance, e.g. the coefficients used in empirical relations in the two-zone combustion model. The primary results obtained from parameter identification of the dynamic model show the efficacy of the proposed optimization procedure and computation algorithm.