Search Results

Accurate modeling of the internal flow and spray characteristics in fuel injectors is a critical aspect of direct injection engine design. However, such high-fidelity computational fluid dynamics (CFD) models are often computationally expensive due to the requirement of resolving fine temporal and spatial scales. This paper addresses the computational bottleneck issue by proposing a machine learning-based emulator framework, which learns efficient surrogate models for spatiotemporal flow distributions relevant for static coupling injection maps, namely total void fraction, velocity, and mass, within a design space of interest. Different design points involving variations of needle lift, fuel viscosity, and level of non-condensable gas in the fuel were explored in this study. An interpretable Bayesian learning strategy was employed to understand the effect of the design parameters on the void fraction fields at the exit of the injector orifice.

Vertical takeoff and landing vehicle platforms with many small rotors are gaining importance for small UAVs as well as distributed electric propulsion for larger vehicles. To predict vehicle performance, it must be possible to gauge interaction effects. These rotors operate in the less-known regime of low Reynolds number, with different blade geometry. As a first step, two identical commercial UAV rotors from a flight test program are studied in isolation, experimentally and computationally. Load measurements were performed in Georgia Tech’s 2.13 m × 2.74 m wind tunnel. Simulations were done using the RotCFD solver which uses a Navier-Stokes wake computation along with rotor-disc loads calculation using low-Reynolds number blade section data. It is found that in hover, small rotors available in the market vary noticeably in performance at low rotor speeds, the data converging at higher RPM and Reynolds number.

Combat aircraft maneuvering at high angles of attack or in landing approach are likely to encounter conditions where the flow over the swept wings is yawed. This paper examines the effect of yaw on the spectra of turbulence above and aft of the wing, in the region where fins and control surfaces are located. Prior work has shown the occurrence of narrowband velocity fluctuations in this region for most combat aircraft models, including those with twin fins. Fin vibration and damage has been traced to excitation by such narrowband fluctuations. The narrowband fluctuations themselves have been traced to the wing surface. The issue in this paper is the effect of yaw on these fluctuations, as well as on the aerodynamic loads on a wing, without including the perturbations due to the airframe.

At high angles of attack, the flow over a swept wing generates counter-rotating vortical features. These features can amplify into a nearly sinusoidal fluctuation of velocity components. The result is excitation of twin-fin buffeting, driven at clearly predictable frequencies, or at nearby lock-in frequencies of the fin structure. This is distinct from the traditional model of fin buffeting as a structural resonant response to broadband, large-amplitude excitation from vortex core bursting. Hot-film anemometry was conducted ahead of the vertical fins of a 1:48 scale model of the F-35B aircraft, in the angle of attack range between 18 and 30 degrees. Auto spectral density functions from these data showed a sharp spectral peak in the flow ahead of the fins for angles of attack between 20 and 28 degrees. Small fences placed on the top surface of the wing eliminated the spectral peak, leaving only a broadband turbulent spectrum.

Heterogeneous physical systems can often be considered as highly complex, consisting of a large number of subsystems and components, along with the associated interactions and hierarchies amongst them. The simulation of a large-scale, complex system can be computationally expensive and the dynamic interactions may be highly nonlinear. One approach to address these challenges is to increase the computing power or resort to a distributed computing environment. An alternative to improve the simulation computational performance and efficiency is to reduce CPU required time through the application of surrogate models. Surrogate modeling techniques for dynamic simulation models can be developed based on Recurrent Neural Networks (RNN).This study will present a method to improve the overall speed of a multi-physics time-domain simulation of a complex naval system using a surrogate modeling technique.

Test methods and data acquisition system specifications are described for measurements of the energy consumption of the control system of a servo-pump continuously variable transmission (CVT). Dynamic measurements of the power consumption of the servo-pump CVT control system show that the control system draws approximately 18.9 W-hrs of electrical energy over the HWFET cycle and 13.6 W-hrs over the 505 cycle. Sample results are presented of the dynamic power consumption of the servo-pump system under drive cycle conditions. Steady state measurements of the control power draw of the servo-pump CVT show a peak power consumption of 271 W, including lubrication power. The drive-cycle averaged and steady state energy consumption of the servo-pump CVT are compared to conventional CVT pump technologies.

This paper critically evaluates the use of Neural Networks (NNs) as metamodels for design applications. The specifics of implementing a NN approach are researched and discussed, including the type and architecture appropriate for design-related tasks, the processes of collecting training and validation data, and training the network, resulting in a sound process, which is described. This approach is then contrasted to the Response Surface Methodology (RSM). As illustrative problems, two equations to be approximated and a real-world problem from a Stability and Controls scenario, where it is desirable to predict the static longitudinal stability for a High Speed Civil Transport (HSCT) at takeoff, are presented. This research examines Response Surface Equations (RSEs) as Taylor series approximations, and explains their high performance as a proven approach to approximate functions that are known to be quadratic or near quadratic in nature.

Neural network augmented model inversion control is used to provide a civilian tilt-rotor aircraft with consistent response characteristics throughout its operating envelope, including conversion flight. The implemented response types are Attitude Command Attitude Hold in the longitudinal channel, and Rate Command Attitude Hold about the roll and yaw axes. This article describes the augmentation in the roll channel and the augmentation for the yaw motion including Heading Hold at low airspeeds and automatic Turn Coordination at cruise flight. Conventional methods require extensive gain scheduling with tilt-rotor nacelle angle and airspeed. A control architecture is developed that can alleviate this requirement and thus has the potential to reduce development time. It also facilitates the implementation of desired handling qualities, and permits compensation for partial failures.

A Stagnation Point Actuator is used to control the lateral dynamics of vortices generated over a sharp-pointed forebody, at high angles of attack, and the resulting rolling moment is studied. Effective roll control is demonstrated, including the ability to suppress the wing rock phenomenon. Piecewise-linear transfer functions are developed from experimental data for the changes in roll moment and pressure difference with actuator frequency content. These transfer functions are reduced to compact form in the frequency domain, and then to a time-domain model using 2 gains and 2 time scales. The roll response is classified according to angle of attack range. Some long time scales are observed in the surface pressure, velocity field and rolling moment, making the response relatively insensitive to speed. Thus over substantial speed ranges, linear transfer functions are shown to effectively describe the roll response to motion of the Stagnation Point Actuator.

The wind-driven dynamic manipulator is a device which uses the wind tunnel freestream energy to drive multi-axis maneuvers of test models. This paper summarizes work performed using the device in several applications and discusses current work on characterizing the aerodynamics of an X-38 vehicle model in pitch-yaw maneuvers. Previous applications in flow visualization, adaptive control and linear-domain parameter identification are now extended to multi-axis inverse force and moment measurement over large ranges of attitude. A pitch-yaw-roll version is operated with active roll to measure forces and moments during maneuvers. A 3-D look-up table generated from direct force calibration allows operation of the manipulator through nonlinear regimes where control wing stall and boom wake-wing interactions are allowed to occur. Hybrid designs combining conventional and wind-driven degrees of freedom are discussed.