Person:

Mavris, Dimitri N.

Permanent Link

https://hdl.handle.net/1853/71501

Associated Organization(s)

Organizational Unit

Daniel Guggenheim School of Aerospace Engineering

Organizational Unit

Aerospace Systems Design Laboratory (ASDL)

Full item page

Publication Search Results

Now showing 1 - 10 of 53

Machine Learning Enabled Turbulence Prediction Using Flight Data for Safety Analysis

(International Council of the Aeronautical Sciences (ICAS), 2021-09) Emara, Mariam ; dos Santos, Marcos ; Chartier, Noah ; Ackley, Jamey ; Puranik, Tejas G. ; Payan, Alexia P. ; Kirby, Michelle R. ; Pinon, Olivia J. ; Mavris, Dimitri N.

The hazards posed by turbulence remain an important issue in commercial aviation safety analysis. Turbulence is among the leading cause of in-flight injury to passengers and flight attendants. Current methods of turbulence detection may suffer from sparse or inaccurate forecast data sets, low spatial and temporal resolution , and lack of in-situ reports. The increased availability of flight data records offers an opportunity to improve the state-of-the-art in turbulence detection. The Eddy Dissipation Rate (EDR) is consistently recognized as a reliable measure of turbulence and is widely used in the aviation industry. In this paper, both classification and regression supervised machine learning models are used in conjunction with flight operations quality assurance (FOQA) data collected from 6,000 routine flights to estimate the EDR (and thereby turbulence severity) in future time horizons. Data from routine airline operations that encountered different levels of turbulence is collected and analyzed for this purpose. Results indicate that the models are able to perform reasonably well in predicting the EDR and turbulence severity around 10 seconds prior to encountering a turbulence event. Continuous deployment of the model enables obtaining a near-continuous prediction of possible future turbulence events and builds the capability towards an early warning system for pilots and flight attendants.
Impact of Adverse Weather on Commercial Helicopter Pilot Decision-Making and Standard Operating Procedures

(Georgia Institute of Technology, 2021-08) Speirs, Andrew H. ; Ramee, Coline ; Payan, Alexia P. ; Mavris, Dimitri N. ; Feigh, Karen M.

Helicopter pilots face unique challenges with regard to adverse weather when compared to fixed-wing pilots. Rotorcraft typically operate at lower altitudes in off-field areas that are not always well covered by weather reporting stations. Although recent technological advances have increased the amount of weather data that pilots can access in the cockpit, weather remains a factor in 28% of fatal helicopter accidents. In this work, commercial helicopter pilots were surveyed and interviewed to better understand how they gather and process weather information, what the perceived limitations of current weather tools are, and how their decision-making process is affected by the information they gather and/or receive. Pilots were found to use a wide variety of weather sources for their initial go or no-go decision during the preflight phase, but use fewer weather sources in the cockpit while in-flight. Pilots highlighted the sparsity and sometimes inaccuracy of the weather information available to them in their prototypical operational domain. To compensate, they are forced to rely on local and experiential weather knowledge to supplement weather reports while still working to mitigate other external pressures. Based on the literature and on results from this work, recommendations are made to address the weather-related gaps faced by the rotorcraft community. This includes the installation of additional weather reporting stations outside of airports and densely populated areas, the further promotion of the HEMS tool to helicopter pilots in all industries, the development of weather tools capable of visualizing light precipitation such as fog, and the development of in-flight graphical displays that can help reduce the cognitive workload of interpreting weather information.
Performance Assessment of a Distributed Electric Propulsion System for a Medium Altitude Long Endurance Unmanned Aerial Vehicle

(Georgia Institute of Technology, 2021-08) Markov, Alexander A. ; Cinar, Gokcin ; Gladin, Jonathan C. ; Garcia, Elena ; Denney, Russell K. ; Mavris, Dimitri N. ; Patnaik, Sounya S.

Distributed propulsion systems are enabled by electrified aircraft and can provide aero-propulsive benefits. The magnitude and impact of these benefits rely on the location of propulsors on the vehicle, the amount of propulsive force generated by those propulsors, vehicle geometry, and the extent of hybridization of the propulsion system. With an increased number of degrees of freedom over conventionally electrified aircraft, the full extent of the impacts of this technology have not yet been explored, especially for military applications. This study builds on a previous one that implemented a series hybrid and turboeletric propulsion architecture on a turboprop UAV, by introducing a distributed electric propulsion system on the same vehicle. The previous study showed that with a hybrid architecture, the same performance, in terms of range and endurance, could not be achieved for a fixed gross take-off weight. This study investigates the impact of the distributed propulsion system with the goal of identifying the benefits over the previous vehicle and determining the level of technology required to break even with the conventional propulsion UAV. In incorporating the new propulsion system, the engine and main motor are resized, leading edge wing mounted propellers and motors are added to the configuration, and a new battery sizing strategy is implemented. Preliminary results show that, although this new system shows increased range and endurance over the series hybrid vehicle, it still falls short compared to the conventional vehicle with current levels of technology. Although improvements are needed to the electrical system technology to reduce the weight enough to break even with the conventional system, the new vehicle shows increased performance during climb, and has the capability to store energy during the mission. With the proper power management and battery utilization strategies, this system can provide reduction in fuel burn and improved performance during certain phases of the mission which could be beneficial for military applications.
Analysis of Weather-Related Helicopter Accidents and Incidents in the United States

(Georgia Institute of Technology, 2021-08) Ramee, Coline ; Speirs, Andrew H. ; Payan, Alexia P. ; Mavris, Dimitri N.

Helicopters typically operate at lower altitudes than fixed-wing aircraft and can take-off and land away from airports. Thus, helicopter pilots have decreased access to weather information due to connectivity issues or sparsity of weather coverage in those areas and at those altitudes. Moreover, regulations allow most rotorcraft to operate in marginal weather conditions. Therefore, weather is a challenge to rotorcraft operations. In this study, rotorcraft events in the United States between 2008 and 2018 in which weather was determined to be a factor are analyzed using the National Transportation Safety Board aviation database. Results show that weather was a factor in 28% of rotorcraft fatal accidents. Wind was involved in most incidents but more rarely involved in fatalities. Bad visibility conditions due to a combination of low illumination and clouds were responsible for most fatal weather-related accidents. Personal flights had the highest accident and incident rates. Finally, the Helicopter Air Ambulance industry had the largest number of incidents and accidents related to visibility conditions out of all other industries. The authors recommend improving awareness of the conditions in which weather events occur and improving training to maintain control of the aircraft in windy conditions or during inadvertent instrument meteorological conditions.
Modeling and Simulation of Novel Electric/Hybrid Electric Multicopter Architectures for Urban Air Mobility

(Georgia Institute of Technology, 2021-08) Demers Bouchard, Etienne ; Verberne, Johannes ; D’Arpino, Matilde ; Ozcan, Metin ; Porpora, Francesco ; Gladin, Jonathan C. ; Patel, Srujal ; Justin, Cedric Y. ; Mavris, Dimitri N.

This paper introduces a dynamic simulation environment developed for novel multi-copter aircraft architectures. The development is motivated by the need to better understand the safety implications of architectural design choices and to provide a formal reliability assessment framework for new Vertical Take-Off and Landing (VTOL) concepts able to consider various airframe and subsystems dynamic behavior. The concepts of interests are different multi-copters configurations investigated by NASA and featuring either electric, hybrid electric, or turboshaft driven powertrains. The simulation environment is a timemarching dynamic simulator formulated using physics-based subsystem models for the batteries, electric motors, turboshaft engines and electric generators. Identified fault modes are integrated into the subsystem models for subsequent use during reliability assessments. The impacts of subsystem faults are propagated to the vehicle flight dynamic response for analysis of their impact on the ability of the vehicle to sustain safe operations. Detailed features of the electric quadrotor model are provided to illustrate the simulation capabilities. Some faults are inserted on the different aircraft in hover and the subsystems behavior is successfully propagated at the vehicle level.
Modeling Framework for Identification and Analysis of Key Metrics for Trajectory Energy Management of Electric Aircraft

(Georgia Institute of Technology, 2021-08) Beedie, Seumas M. ; Harris, Caleb ; Verberne, Johannes ; Justin, Cedric Y. ; Mavris, Dimitri N.

To prepare for the upcoming entry into service of electric and hybrid-electric aircraft, regulators may have to update or develop new regulations and standards to ensure safe operations of these new vehicles. To ensure public acceptance, these vehicles need to demonstrate an equivalent level of safety consistent with existing regulations. However, the ability to fly in different modes (forward flight, vertical flight) and the different powertrain elements may require significant changes to regulations to ensure that an insightful representation of the usable energy is provided to flight crews. This requires an understanding of the major operational differences between conventional and electric aircraft, and how these differences impact the trajectories a vehicle can fly. For instance, there is no simple analog to fuel gauges for measuring the extractable energy available on board electric aircraft, as energy related metrics can vary with a range of variables, such as component temperatures, battery health, and environmental conditions. It is thus more complex for flight crews to gauge in real-time how much usable energy is available and to figure out which trajectories are feasible with respect to both energy and power. To assess the feasibility of trajectories and quantify the adequacy of novel energy tracking metrics and methodologies, a trajectory energy management simulation environment is implemented allowing the simulation of various energy metrics across a range of vehicles and missions. This allows decision makers and regulators to assess the importance of these metrics for safe operation across a wide variety of missions. The impact of ambient air temperature, battery state of health, and initial battery, motor, and inverter temperatures are assessed for a typical flight mission. It is concluded that state of health, ambient temperature, and initial battery temperature all had significant impacts on the final state of charge and amount of extractable energy. Additionally, at high ambient temperatures and in aggressive climbs, motor temperature limits and inverter temperature limits can sometimes be reached, further complicating the assessment of what can be done with the amount of energy stored on board. Proper management of these constraints is therefore crucial for optimizing trajectories with respect to energy metrics. Future work is proposed regarding further expansion of the framework simulating aircraft with vertical takeoff and landing capability, and flight-dynamics algorithms that will enable simulation of optimal energy mission profiles.
Helicopter Operations Weather Information Pilot Interviews

(Georgia Institute of Technology, 2021-01-29) Speirs, Andrew ; Ramee, Coline ; Alexia, Payan ; Mavris, Dimitri N. ; Feigh, Karen M.

Helicopter pilots face unique challenges with regard to adverse weather when compared to fixed-wing pilots. Rotorcraft typically operate at lower altitudes in off-field areas that are not always well covered by weather reporting stations. Although recent technological advances have increased the amount of weather data that pilots can access in the cockpit, weather remains a factor in 28% of fatal helicopter accidents. In this work, commercial helicopter pilots were interviewed to better understand how they gather and process weather information, what the perceived limitations of current weather tools are, and how their decision-making process is affected by the information they gather and/or receive.
Imitation Learning for UAS Navigation in Cluttered Environments

(Georgia Institute of Technology, 2021-01-04) Harris, Caleb M. ; Choi, Youngjun ; Mavris, Dimitri N.

Autonomous navigation is a critical component for the use of unmanned aerial systems (UAS) in complex tasks such as package delivery and disaster response. In recent years, these systems have seen increased usage in harsh tasks such as search and rescue and disaster relief, however there remains challenges for efficient and safe operation in a fully autonomous mission. This work seeks to provide a data-driven, vision-based method to navigating and searching through a clustered environment, which is high-speed, low-cost and vehicle-agnostic. This is done by first assuming obstacle avoidance as a two-dimensional navigation task that can be solved by knowing the relative location of the goal and the 2D image of the obstacle in the camera frame. Imitation learning is used to train a deep neural network from an expert planning policy, while a model predictive controller tracks the target. All the processing is capable onboard the vehicle, with the assumption that the general target direction is forward of the camera-frame, and that the global state estimation error is low. The framework and trained model are tested in simulation, with a quadcopter conducting search scenarios in different environments. The resulting framework is quicker to avoid obstacles and can be applied on small, low-cost systems with a single monocular camera.
Use of Machine Learning to Create a Database of Wires for Helicopter Wire Strike Prevention

(Georgia Institute of Technology, 2021-01-04) Harris, Caleb M. ; Achour, Gabriel ; Payan, Alexia P. ; Mavris, Dimitri N.

Rotorcraft collisions with wires and power lines have been a major cause of accidents over the past decades. They are rather difficult to predict and often result in fatalities. For this reason, there is a push to provide pilots with additional information regarding wires in the surrounding environment of the helicopter. However, the precise locations of power lines and other aerial wires are not available in any centralized database. This work proposes the development of a wire database in two phases. First, power line structures are detected from aerial imagery using deep learning techniques. Second, the complete power grid network is predicted using a centralized many-to-many graph search. The two-step framework produces an approximate medium-voltage grid stored as a set of connected line segments in GPS coordinates. Experiments are conducted in Washington D.C. using openly available datasets. Results show that utility pole locations can be predicted from satellite imagery using deep learning methods and a full grid network can be generated to a level of detail depending on computational power and available data for inference in the graph search. Even with limited computational resources and a noisy dataset, over a a fourth of the grid network is directly predicted within a range of seven meters, and the majority of the network is visually inferred from nearby detections. Moving forward, the goal is to apply the proposed framework to larger regions of the U.S., with rural and urban environments, to map all wires and cables that are a threat to rotorcraft safety.
Parametric Design, Manufacturing and Simulation of On-Demand Fixed Wing UAVs

(Georgia Institute of Technology, 2021-01-04) Chitale,Yash H. ; Justin, Cedric Y. ; Mavris, Dimitri N.

As the market for Unmanned Aerial Vehicles (UAVs) continues to expand, an unfulfilled need has been identified for tailor-made solutions leveraging an end-to-end process for the design and manufacture of the vehicle. The use of computer aided design combined with new manufacturing techniques allows small UAVs to be parametrically sized and quickly prototyped and deployed. This parametrization technique can be used throughout the entire design process to create optimized, attritable, on-demand solutions that can be adapted to evolving customer requirements. High-level requirements are mapped to quantitative design constraints and an automated process uses these constraints to design and manufacture a vehicle within a specified amount of time. The proposed framework is demonstrated with the generation of a fixed wing UAV solution for the detection and tracking of wildlife in remote areas. National Parks seek to prevent illegal poaching but often lack either the resources to monitor endangered animals, or the budget to purchase UAVs specially designed for wildlife tracking. First, mission requirements are identified and define a design space from which an optimal design point is selected. This design point sizes a UAV model, which is then optimized to minimize manufacturing time with the objective to yield a ready-to-fly solution within 48 hours. A flight simulation of the mission is then performed to ensure that the vehicle will fly as designed. Structural limitations of the UAV are accounted for and linked to parameters of the flight control algorithm to ensure that the UAV can safely fly its mission.