TP 3 – Experimental investigation of the influence of mixture preparation on cyclic variations in direct-injection spark-ignition engines

Compared with homogeneously premixed combustion processes, stratified combustion processes achieve higher velocities and turbulence in the combustion chamber due to the impulse input of the high-pressure gasoline injection. Despite an overall lean fuel-air ratio, high conversion rates are thus possible. Investigations in optically accessible engines have shown that in general not only injection and ignition influence the cyclical variations in DI operations, but also the local fuel-air ratio at the ignition point. Between liquid spray and ignitable mixture are evaporation and mixing. The turbulent internal cylinder flow results in a fuel-air ratio that varies in time and space. These processes are much less well studied in direct-injection engines regarding to the resulting cyclic variations.

Stratified combustion processes offer great potential for increasing efficiency, but on the other hand they are also subject to cyclic variations. It is of great practical and fundamental interest to understand how mixture inhomogeneity influences cyclic variations, especially in less extreme cases.

Figure 1: Mixture inhomogeneity in the gas phase due to differential evaporation of a multi-component fuel, visualized by two-tracer-LIF. (a-c) Three exemplary single images show the ratio of the local concentration of the highly volatile and low volatile component. A ratio of 1 means that the components are in the same ratio as in liquid fuel. (d) Average of 200 cycles for an inlet valve lift of 4 mm and (e) corresponding average value at 9 mm valve lift. Time of recording at 37°CA, injection ends 44°CA before ignition OT, 200 bar injection pressure, 8-hole “ECN Spray G” injector.

This sub-project experimentally investigates the influence of spray, ignition location, local mixture composition, and turbulent flow on each other and on the cyclic variations in an optically accessible direct-injection spark-ignition engine. Closely collaborating with sub-projects 4 and 5 of the research unit, the goal is to understand that part of the cyclic variabilities’ causal chain that comprises injection – evaporation – mixing – ignition – inflammation. Here, we consider a cylinder charge that is homogenous before injection, such that the injected spray is the only source of mixture inhomogeneity. The 8-hole injector for „Spray G“ of the international research collaboration „Engine Combustion Network“ (ECN) is used. From the ECN’s previous work, extensive data on the exact internal nozzle geometry, statistics of droplet size and velocity, macroscopic spray parameters, etc. are available. After establishing a small set of stably-running reference cases, we then vary the location of ignition (via laser ignition) and the in-cylinder flow (via the intake-valve lift and tumble-enhancing intake inserts). In particular the concentration of the evaporated fuel is quantitatively imaged by laser-induced fluorescence. These and other imaging data such as overall spray structure, morphology and movement of the ignition’s spark, flame propagation, and flow field are statistically correlated with pressure-trace analysis. To this end, conventional evaluation of conditional statistical moments is employed, as well as methods that are novel in this research area, like independent component analysis (ICA).

Prof. Dr. Sebastian Kaiser sebastian.kaiser(at)
Prof. Dr. Christof Schulz christof.schulz(at)
Judith Laichter judith.laichter(at)