An Investigation of coupled atmospheric turbulence and ship airwakes for helicopter-ship dynamic interface simulations
Open Access
- Author:
- Santos Thedin, Regis
- Graduate Program:
- Aerospace Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 23, 2019
- Committee Members:
- Sven Schmitz, Dissertation Advisor/Co-Advisor
Sven Schmitz, Committee Chair/Co-Chair
Joseph Francis Horn, Committee Member
Xiaofeng Liu, Outside Member
Michael P Kinzel, Committee Member
Amy Ruth Pritchett, Program Head/Chair - Keywords:
- CFD
Atmospheric Boundary Layer
Ship Airwakes
Flight Dynamics
Rotorcraft - Abstract:
- The dynamic interface between a ship and a helicopter is a complex, hazardous environment, demanding high levels of pilot workload. In modeling & simulation of such environment for pilot training purposes, high levels of fidelity are required on the airwake module. The objectives of this research effort are two-fold. The first one is to analyze in details the effects of the turbulence present in the atmospheric boundary layer (ABL) on ships and the resulting airwake. The other objective is to use airwake data saved from the numerical simulations as external disturbances to a helicopter model in order to quantify an increase in pilot workload. Two different types of inflow are investigated: unsteady ABL and steady ABL. Unsteady cases are executed in OpenFOAM, are representatives of an actual stability state and include realistic features such as coherent structures. Steady ABL cases are executed in OVERFLOW and represent an appropriate velocity profile, but do not include any freestream turbulence. Uniform inflow cases are also executed on both codes as baseline cases. The SFS2 ship geometry is used and it is modeled by the immersed boundary method within OpenFOAM, while body-fitted overset grids are used in OVERFLOW. Pilot workload is quantified by a frequency-domain analysis of the energy associated with the usage of the input sticks. Initially, neutral cases with two levels of shear are investigated and compared to uniform inflow solutions. Analysis of velocity distributions along probe lines at the deck revealed that the presence of an ABL modifies the recirculation region and delays the reattachment point. Different levels of shear yield different characteristics. For airwakes modified by unsteady ABL inflow, spectral analysis at locations near the ship's flight deck indicated that higher energy content at frequencies above 3 Hz, with better agreement to Kolmogorov's -5/3 cascade, have been captured. Increased content has also been observed in the 0.1--0.3 Hz range. This energy cascade matches in situ experiments. Spectral content on the uniform inflow cases fails to match content above 3 Hz, which are also not usually captured in standard CFD simulations. The airwakes related to ABL inflows were not related to each other by a common factor, indicating that these solutions are not scalable, differently than what is usually observed for uniform inflows. Next, two hover locations outside of the airwake are considered, subject only to the turbulence present in the inflow -- in at an altitude of 20 ft and another at 80 ft. The ABL turbulence had a substantial effect on the vehicle, resulting in significantly more disturbances, considerably more power fluctuations, and fluctuations on the vehicle's attitudes. While this was observed on both cases, it was much more prominent in the high altitude case. The fluctuations were reflected on the stick usage, and thus pilot workload. The uniform inflow case barely exerted any effect and had comparable results to a scenario where only Pitt-Peters inflow model is used (no external disturbances). Analysis of the energy related to the stick usage showed that substantially more energy was found for the ABL case across all of the frequency range investigated for high altitude case, and a lower increase for the lower altitude case, found in the range of approximately 0.1--0.6 Hz. These results suggested that the large length-scale eddies that are present in the atmosphere seems to affect the vehicle and the pilot workload. Lastly, two hover locations at the flight deck have been investigated. The vehicle was subject to the highly turbulent air shedding off the superstructure and chimney. Comparisons between steady ABL, unsteady ABL, and uniform inflow are made. One hover location is within the highly separated region, and another is slightly higher. The second location represents a mix of the pure unsteady ABL flow and airwake turbulence. These locations were selected in order to check whether or not the effects from the atmospheric turbulence observed previously would apply here. For the unsteady ABL case, the results indicated that when the airwake turbulence dominates, an increase in the energy associated with frequencies in the range of 0.1--0.3 Hz has been observed. For frequencies above 0.5 Hz, not many differences are observed at the energy associated with the stick use. However, one of the main findings of this work is that when the aircraft was hovering only 10 ft higher, in a flowfield that was a mix of airwake and atmospheric turbulence, the energy associated with the unsteady ABL was higher than that associated with the uniform inflow for all of the 0.2--2 Hz spectrum. Now, with respect to the OVERFLOW's steady ABL case, no appreciable differences have been captured, in neither of the hover locations. The steady ABL approach did not affect the vehicle nearly as much as the unsteady ABL did. In fact, the uniform inflow consistently exhibited higher energy (although very small) than that seen under the steady ABL. The steady ABL did not add any relevant information. The results indicate that when the aircraft is flying at a location that is subject to more of the atmospheric eddies, the vehicle tends to react to the unsteadiness present, which represents additional pilot workload. This is especially relevant for a ship with a flat deck (similar to the LHA class). If the vehicle is solely in the wake of the superstructure, no relevant differences were observed. The lower fidelity approach of modeling the ABL as a steady ABL did not add any relevant pilot workload for the SFS2 with zero wind-over-deck case investigate.