Experimental investigation of the phenomenon of oil break-up in an engine crankcase

A phenomenological study of the characteristics of a two-phase, oil and air mixture was undertaken in a combustion engine crankcase model at speeds of up to 6000 rpm. The idealized model comprised production components and tolerances. Nonintrusive measurement techniques, such as PIV and PDA and high-speed photography, were applied to the study of the oil aerosol and liquid films. Agitated oil pools and impingement on the crankcase surfaces contributed to the aeration of oil in the sump due to windage effects. A mean, wakelike, flow pattern was observed that lagged the crankshaft rotation. Harmonic frequencies were recorded that could be identified with the passage of the position of the maximum crank profile radii and a delay time related to the involute-shaped cord and droplet trajectories induced by drag forces. Highly unsteady flow conditions were measured close to the chamber walls. Three classical breakup modes were identified over three speed ranges. Sheet, ligament, and droplet breakup were observed in the ranges of idle to 1800, 4200, and 6000 rpm. Classification of the particle size distribution for the break-up modes was carried out using the PDA technique. Droplet diameter and velocity data varied in the ranges of 2 μm (dmin) to 130 μm (dmax) and 0−40 ms−1 at 6000 rpm. The ratio of the droplet diameter, dmin;max to a characteristic diameter D at which the oil was observed to break away was compared to the disk Bond number, BoD. For the largest measured droplets, the ratio was approximated by dmax/D = 1.30BoD−0.33 . The disk Bond number was 2.21, while the empirical constant was in the range of 1.82−2.39. The result showed that an appropriate correction must be applied to the universal power law, proposed by Simpkins in 1997 for the estimate of the primary droplet size generated by a spinning disk homogenous atomizer, which had overpredicted the droplet sizes in the case of the crankshaft.