This paper presents an experimental approach to identify the sources of instabilities in arc plasma devices. The phenomena of demixing in arcs have been utilized to explore the characteristics of different instabilities. Problems in explaining the observed behavior with our current understanding of the phenomena are discussed. Hydrogen is used as a secondary gas with argon as the primary plasma gas for this study. Results indicate that the observed behavior such as steady, takeover, and restrike modes of instabilities in arcs may essentially originate from the thin boundary layer over the anode wall primarily at the location of the anodic arc root. The bulk core flow apparently does not play any significant role in such instabilities. Arc currents rather than flow rates control the behavior of the instabilities in frequency space. Bifurcation of the system behavior and evidence for the existence of quadratic zones in flow space of binary gas mixtures separating steady and unsteady behavior are discussed.

The large eddy simulation methodology is used to predict and understand the unsteady flow around a marine propeller in crashback operation. A non-dissipative, robust numerical algorithm developed by Mahesh et al. (2004, J. Comput. Phys., 197: 215-240) for unstructured grids is extended to include the effect of rotating frame of reference. Flow around Propeller 4381 (for propeller specification see e.g. Jessup et al.: 2004, 25 th Symposium on Naval Hydrodynamics, 270-292) at advance ratio J = −0.7, Reynolds number Re = 480, 000 (based on propeller diameter and relative speed between freestream flow and the propeller) is computed for a period of 300 revolutions. The simulation shows the presence of a highly unsteady ring-vortex, and irregular low frequency unsteady loads on the propeller. The spectra also show distinct peak at higher frequency of 5 rev −1 , corresponding to passage of individual blades of the fivebladed propeller. Mean values, root mean square (RMS) fluctuations and spectra of computed thrust, torque and side-forces show good agreement with experiment. Circumferentially averaged mean velocity and RMS fluctuation of velocity obtained from the simulation are compared to experimental data and good agreement is observed. The cross-flow aft of the propeller, which represents inflow for a propeller in crashback, was investigated. The cross-flow shows low frequency fluctuations similar to spectra of side-forces, but without the peak at blade frequency of 5 rev −1. An unsteady actuator disk model is constructed in order to understand unsteadiness in propeller crashback. The model considers crashback as a competition between two flows of opposite directions: reversed flow through the propeller and the ambient flow due to motion of the vessel. Visualization of the flow around the actuator disk in crashback shows creation, asymmetric growth, tilting, stretching and shedding of unsteady ring vortices which are correlated to the fluctuation of the thrust of the actuator disk.

Fan-wake/outlet-guide-vane interaction broadband noise in turbofan jet engines is studied. The mechanism and some issues are first discussed using a two-dimensional gust-prediction model. An oblique gust-prediction model is then developed. Quasi-three-dimensional unsteady lift is calculated using a two-dimensional equivalence method. It is coupled with annular duct modes to obtain the sound power spectrum density. Spanwise turbulence integral length scales and their impact on power spectrum density predictions are investigated. A spanwise integration limit suitable for the complete frequency range is proposed. The model is validated using the NASA Source Diagnostic Test data. Sound power scaling with vane count B is examined. If solidity is maintained, the cascade response does not converge on the single-airfoil response, even for low vane counts. The sound power varies inversely with B at low frequency; it scales with B at very high frequency. The power spectrum density trend with the fan tip Mach number M T is also identified. It scales with M 5 T if turbulence intensity in the fan wake scales "ideally" with M T. At offdesign conditions, fan wakes are not ideal; therefore, different speed trends apply. M 3.3 T scaling is found to best fit the Source Diagnostic Test data and the prediction.

An unsteady actuator disk model is constructed in order to understand unsteadiness in crashback operation of a marine propeller as a competition between two flows of opposite direction: reversed flow through the propeller and the ambient flow due to motion of the vessel. The large eddy simulation methodology is applied to predict the flow corresponding to forward and crashback modes of operation using a non-dissipative, robust numerical algorithm developed by Mahesh et al. (2004, J. Comput. Phys., 197: 215-240) for unstructured grids. Flow visualization shows creation and shedding of unsteady ring vortices which are correlated to the fluctuation of the thrust coefficient. Results of the model are compared to the LES of the full propeller geometry as well as to experimental data.

No abstract