| Achieving effective control of the in-cylinder flow is a key issue for the optimizationof the combustion and emissions of internal combustion (IC) engines. However, theviolently unsteady characteristics of the in-cylinder fluid motion causes greatdifficulties for the control of the in-cylinder flow. To understand the turbulent flow ofIC engines and achieve effective control of the in-cylinder flow, studies wereconducted from the aspects of the test and evaluation of the tumble flow, the unsteadycharacteristics of in-cylinder flow, the boundary layer flow, and the accurateprediction of the complex flow in IC engines. The main contents and conclusions aredescribed as follows:Focusing on the problem that there is no unified standard for the tumble flow testmethods, from the aspect of intake flow motion evaluation, theoretical analysis basedon the law of conservation of angular momentum, as well as the experiments usingsteady flow test rig and particle image velocimetry (PIV) methods were carried out toanalyze the influence of different parameters on the measurement results. The resultsshowed that the changes in the test apparatus for the tumble flow leads to significantdifferences in tumble intensity. Too small diameter of outlet causes the distortion ofthe in-cylinder flow and substantial increase of tumble intensity as high as four times.By optimizing the structure of the measurement apparatus, the overestimation of thetumble intensity can be eliminated, thereby a reasonable comparison of the differentmeasurement results can be achieved.From the aspect of in-cylinder unsteady flow motion, PIV experiments wereconducted to investigate the effect of variable valve lift (VVL) on the unsteadycharacteristics of in-cylinder flow as well as cycle-to-cycle variations (CCV). Theresults showed that, as the valve lift profile is changed, the in-cylinder flow patternsignificantly varies. Under the higher maximum valve lift (MVL) conditions, the largescale tumble flow can be formed and maintained until the end of the compressionstroke (around80°CA BTDC). Whereas, under the lower MVL conditions, thetumble flow cannot be formed, and the mean tumble intensity is nearly zero at80°CA BTDC. It is also found that the CCV of the in-cylinder flow is significantlyaffected by the variation of valve lift profile. Under the lower MVL conditions, theCCV of the tumble center is as high as1.4times of that under higher MVL conditions,and the CCV of the low-frequency fluctuating flow is higher than that under higherMVL conditions by28percent. The weak tumble motion and the enhancement of thelow-frequency fluctuating flow by the intake jet are the main reasons for the strongCCV of the in-cylinder flow under the lower MVL conditions. By strengthening thetumble motion the CCV of the in-cylinder flow can be effectively inhabited.From the aspect of the boundary-layer flow, micro particle image velocimetry(microPIV) was used to investigate the flow characteristics near the cylinder wall under the complex flow conditions of gasoline engines. The experimental resultsshowed that there is an obvious viscous wall region in the in-cylinder flow due to themolecular viscosity of air. In the flow separation zone near bottom of cylinder wall,there are a large number of small-scale vortex structures with the length scale of1~2mm in the instantaneous flow fields. The thickness of the viscous sublayer decreaseswith the increasing valve lift and the pressure drop (P) between the intake port andexit. For the case with higher valve lift (VL) and pressure drop (VL=7.975mm, P=1kPa), the viscous sublayer thickness is only about0.3mm. In the viscous sublayer,the dimensionless velocity distribution is linear, which is in good agreement with thelaw of wall. However, in the logarithmic layer, no obvious logarithmic lawdistribution region is observed in in-cylinder flow, and the slope of the velocitydistribution increases with the enlarged P. The distribution of Reynolds stressnormalized by friction velocity of in-cylinder flow is similar to that of the channelflow, but its value is five times higher than that of channel flow for the componentalong the flow direction because of the lower frequency fluctuation from the unsteadyin-cylinder flow, not only turbulence.From the aspects of precisely predicting the complex in-cylinder flow motion, it isfound that the differences in dimensionless velocity distribution between thein-cylinder flow and the channel flow results in the near-wall treatments (e.g., thestandard wall function) is no longer applicable for the in-cylinder flow simulation.The boundary-layer flow and separation flow caused by the intake jet in IC enginescan be satisfactory predicted by using improved delayed detached eddy simulation(IDDES). The IDDES method captures the quasi-periodic fluctuations of the intake jet.It is revealed that the boundary-layer separation occurs at the valve edge, and then thevortex sheds with a quasi-periodic frequency, which disturbs the intake jet flow. Inaddition, there are obvious vortices with quasi-periodical reversion in the rotationdirection in the coherent-structure flow fields of the intake jet. The quasi-periodicfluctuations in the intake jet directly affect the boundary-layer flow in the downstreamand lead to frequently detaching and re-attaching flow, which results in the differencebetween the in-cylinder flow and the channel flow. By using the IDDES method thesimulation of the in-cylinder flow in a whole cycle, it was found that the predictedmean flow fields are in good agreement with the PIV experimental results. Theviscous sublayer expands to y+=15~20at the end of the compression stroke, since theincreasing temperature and density of the in-cylinder air substantially reduce thekinematic viscosity of the fluid. The distribution of the coherent structure andturbulent kinetic energy of the turbulent flow in the cylinder revealed that the intakejet and the shear of the wall are the main source of the coherent structure. Moreover,the spatial distribution and the intensity of the coherent structure directly determinethe turbulence fluctuations. |
PDF Full Text Request