Applying optical and acoustic diagnostic tools to probe turbomachinery wakes, cardiac flows, and multi-phase media

Embargo until
Date
2017-08-30
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
This dissertation has three parts. In the first part, we study the complex structure of turbulent flow past an automotive cooling fan. A 360 gallon optically index-matched test facility is specially built for performing impeller phase-locked particle image velocimetry (PIV) measurements. A full-scale fan (30 cm diameter) and its shroud (70 cm outer diameter) designed by Robert Bosch LLC are installed in an acrylic test section. The fan and shroud are also made out of acrylic, and an aqueous solution of NaI (62% by weight), with the same refractive index as acrylic (n=1.49), is used as the working fluid, providing unobstructed optical access to the fan for flow measurements. A turbulence grid is placed two fan diameters upstream to simulate the effects of a radiator wake. Phase-locked 2D PIV measurements are performed focusing at the inlet, near wake and tip gap of the axial fan for multiple blade phases. To achieve high spatial resolution over large fields of view, nine sample areas cover the entire wake, three sample areas cover the inlet and a higher magnification region covers the tip gap. To sufficiently resolve the evolution of the wake and tip vortex system, data are acquired for 10 and 31 blade phases respectively. Inflow measurements quantify the turbulence intensity distribution generated by the grid upstream and the acceleration of the flow by the fan. Wake measurements reveal the formation and evolution of the tip vortex and three generations of the fan wake. The results show the entrainment of the blade wake by the tip vortex, as well as the decay in vortex strength, entrainment speed and wake velocity deficit with axial distance, as well as their associated turbulence characteristics. Higher magnification measurements focusing on the tip gap reveal the complex formation of the tip vortex due to the merging of multiple vortex filaments during the blade passage. In the second part, we study the complex blood flow characteristics that may be linked to the formation and cure of left ventricular thrombus (LVT) in cardiomyopathy patients. Echocardiographic PIV-PTV analysis is performed on routine in-vivo contrast ultrasound images acquired from four patients with LVT to obtain time-resolved velocity distributions. The contrast agent comprises of gas filled microbubbles that trace the blood flow and enhance the acoustic backscatter of the ultrasound signal. The concentration of the contrast agent is adjusted in each case to obtain time-resolved images with sufficiently discernable bubble motion. Due to constraints on the maximum possible spatial and temporal resolution, the data quality is not comparable to standard optical PIV. Hence, to obtain reliable velocity data, optimized procedures that integrate image enhancement, PIV and particle tracking velocimetry (PTV) are introduced. Initial steps involve performing cross-correlation based PIV analysis over several image enhancement and cross-correlation parameters to obtain multiple velocity vectors at each grid point. Optimization is subsequently performed using outlier removal and smoothing to select the correct vector. These vectors are then used as part of a multi-parameter PTV procedure to further refine the data. Phase averaged velocity and vorticity distributions reveal the LV vortex formation and evolution as it fragments and decays during the cardiac cycle. The formation of LVT predominantly occurs near the left ventricular (LV) apex. The efficacy of apical washing is estimated directly from the phase-time history of apical velocity as well as from the vortex-induced apical velocity, providing preliminary LVT risk quantification criteria. In the third part, we aim to characterize the aerosolization of crude oil-dispersant contaminated slicks due to bubble bursting. Bubble bursting observed in oceanic whitecaps is a well-known mechanism of marine aerosol generation. Although oil-spills occur frequently in the ocean, the emissions of oily marine aerosols from bubble bursting are not well characterized. Recent spills have witnessed the unprecedented use of chemical dispersants as a strategy for spill-response. They significantly modify several properties of crude oil, potentially altering the size, concentration, and composition of aerosolized particles. Hence, in this study, bubble plumes with controlled size distributions are injected into a vertical seawater column. They rise to the surface contaminated with slicks of crude oil-dispersant mixtures and burst. The distributions of aerosolized particles (10-380nm and 0.5-20μm) above the slicks are monitored before, during and after bubble injection. Measurements are performed at the same air injection rate for varying bubble plumes (Φ 86μm, 178μm and 595μm), slick thicknesses (50 and 500 μm) as well as interfacial mixtures (pure crude oil, pure dispersant Corexit 9500A and dispersant premixed with crude oil at a ratio (DOR) of 1:25). Results show an order of magnitude increase in the nano-size particle concentrations only when the largest bubble plumes burst on slicks of 500μm DOR-1:25 oil or 50μm pure dispersant. Small and medium-sized bubble plumes generate micron–size aerosols for thin crude oil slicks as well as those containing dispersants. Preliminary chemical analysis confirms the presence of crude oil in both micro- and nano-aerosols generated from slicks with DOR 1:25 oil. Potential operating mechanisms are discussed and aerosol emission factors per bubble are provided, enabling the risk-assessment associated with the inhalation of oily aerosols.
Description
Keywords
bubble bursting, LV flow, automotive fan, PIV, echo-PIV, aerosols
Citation