Search Results

Earlier work has shown that the in-cylinder flow in internal combustion engines can be modelled, with reasonable accuracy, as the sum of an ensemble averaged mean component, a non-stationary ‘turbulence’ component and a ‘cycle-to-cycle’ variation component, the latter being phase-locked to the engine cycles. The development of the LDA technique has enabled direct measurements of the in-cylinder velocity field to be taken, either at a single position in space over the engine cycle, or over a range of spatial positions, at effectively one point in the engine cycle (scanning LDA). Previously, different approaches have been developed for separating the various flow components in the model described above, dependent on the type of data acquired. In this paper a single ‘unified’ method is presented, based on the computation of autocorrelation functions and a completely parametric representation of the various components in the flow model.

An investigation has been carried out to compare the ability of swirling and tumbling flow regimes to enhance the turbulence in a disc-shaped gasoline combustion chamber. Scanning LDA measurements have been made of spatial velocity fluctuations in a high swirl, a tumble and a baseline low swirl build. All of the testwork was carried out under motored conditions at an engine speed of 1200 rev/min. A parametric model has been developed to account for the effects of mean flow cyclic variation and system noise. It is shown that the model fits very well to the experimental data, enabling unbiased estimates of turbulence intensity and turbulence length scale to be made. In the region measured around TDC the high swirl build achieves a uniform increase in turbulence Intensity of about 55% over the baseline build. The tumble build however achieves a peak in turbulence intensity of more than twice the baseline build at 30° BTDC.