Hidemasa Kosaka

A fractal-based flame propagation model for large eddy simulation

A novel combustion model for large-eddy simulation (LES) for gasoline engines has been developed. Unlike conventional models based on Reynolds-averaged Navier–Stokes (RANS) models, the new model takes a unique approach; it is described by the fractal characteristics of flame front and a universal expression for the subgrid scale (SGS) flame speed. The present fractal combustion model was applied to calculations of a spark ignition engine. Both the 0–10 per cent and 10–90 per cent combustion periods agree well with the experimental data. Because the modelling of the SGS turbulent speed is based on fractal analysis with experimental observations, the SGS combustion model is able to apply a wide range of engine operating conditions. The present model was applied to a multi-cycle simulation of a single-cylinder engine. The fluctuations at the instant when the heat release rate peaked were compared with data that was obtained experimentally. The calculated magnitude of the fluctuations was found to be close to the experimental values. It is thought that the flow variation generated during the intake stroke significantly influences the cyclic variations.

State Estimation Algorithm Based on a Disturbance Observer for a Free-Piston Engine Linear Generator

In order to realize a thin, compact, efficient, and fuel-flexible power-generation system, we have been developing and investigating a free-piston engine linear generator (FPEG), which consists of a two-stroke combustion chamber and an adjustable gas spring chamber. This paper proposes a novel observer algorithm that has two advantages. First, the piston position can be interpolated at high accuracy beyond the resolution of the piston position sensor, which improves the accuracy and stability in controlling the combustion parameters, such as ignition timing and compression ratio. Second, the pressure in the gas spring chamber and the coefficient of damping (lubrication condition) are estimated simultaneously and independently. In this paper, the principle of the developed observer algorithm is described and tested through simulation and experimentation using the developed FPEG system. Based on the simulation, the output of the position sensor (resolution: 0.55 mm) was interpolated with an accuracy of ±0.1 mm. Based on the experimental results for the estimation of pressure in the gas spring chamber and the coefficient of damping, the feasibility of sensorless monitoring of the pressure in the gas spring chamber and the coefficient of damping has been verified.