Dynamics of Exhaust Valve Flows and Confined Bluff Body Vortex Shedding

Sammanfattning: This thesis can be divided into two interconnected topics; engine exhaust-valve flows and confined bluff-body vortex shedding. When optimising engine flow systems it is common to use low order simulation tools that require empirical inputs, for instance with respect to flow losses across the exhaust valves. These are typically obtained from experiments at low pressure ratios and for steady flow, assuming the flow to be insensitive to the pressure ratio and that it can be considered as quasi-steady. Here these two assumptions are challenged by comparing measurements of mass-flow rates under steady and dynamic conditions at realistic pressure ratios. The experiments with a static valve were carried out using a high-pressure flow bench at cylinder pressures up to 500 kPa. For the dynamic-valve experiments the transient flow rate during the blowdown phase of an initially pressurised cylinder was determined. Here a linear motor actuated the valve to obtain equivalent engine speeds in the range 800–1350 rpm. It was shown that neither of the above mentioned assumptions are valid and a new non-dimensional quantification of the steadiness of the process was formulated. Furthermore it was shown through Schlieren visualisation that the shock structures in the exhaust port differ depending on if the system dynamics are included or not. The study shows that reliable results of flow losses past exhaust valves can only be obtained in dynamic experiments at representative pressure ratios. The second topic arose from the need to monitor time-resolved mass-flow rates in conduits. A mass-flow meter based on vortex shedding from bluff bodies was designed where microphones are used to detect the shedding frequency. It consists of a forebody and a downstream mounted tail and the system was shown to be capable of measuring pulsating flow rates. Furthermore, the flow topology associated with different forebody and splitter plates has been characterised, through visualisation of the flow behind the shedder and on the splitter plate. It has been shown that for long splitter plates a “horse shoe” like vortex, which attaches to the tail, is formed. It has also been shown that another energetic mode (denoted mode-II) can interact with and disrupt the primary vortex formation. A hypothesis for the appearance of mode-II has been formulated, linking it to the periodic separation of the boundary layer at the conduit wall.