eVTOL arrival sequencing and scheduling in on-demand urban air mobility

eVTOL arrival sequencing and scheduling in on-demand urban air mobility

Kleinbekman, Imke (TU Delft Aerospace Engineering)

Mitici, Mihaela (mentor)
Wei, Peng (mentor)

Delft University of Technology

Aerospace Engineering

2019-01-18

Urban Air Mobility (UAM) has the ability to reduce ground traffic congestion by enabling rapid on-demand flight through three-dimensional airspace. Due to the use of electric Vertical Take-Off and Landing (eVTOL) vehicles with distributed electric propulsion, UAM operates under zero operational emissions and cheaper and at a lower noise level compared to helicopters. In the long term with more UAM flights, air traffic control is expected to limit further growth of such operations. Therefore, first research has been performed on energy-efficient trajectory optimisation for a given required time of arrival, as the arrival phase is the most safety-critical flight phase with higher air traffic density and limited battery energy. However, research on the separation between eVTOL aircraft by computing their optimal required time of arrival (RTA) is limited. Besides, the available research has not considered limited battery power of the eVTOL aircraft or a limited vertiport landing pad capacity, neither discusses the ATC procedures for eVTOL flight. Unlike fixed-wing aircraft or helicopters in commercial aviation, eVTOL aircraft fly on-demand and have different flight dynamics, limited battery energy supply and a limited number of landing spots at a vertiport such as on top of high-rise buildings. Therefore, this research has aimed to take the first step in the development of Urban Air Mobility Traffic Management for autonomous on-demand eVTOL operations by developing the arrival procedures and arrival sequencing and scheduling tool searching for minimum total delay considering limited battery power and vertiport landing pad capacity. A concept of operations for vertiport terminal area airspace design has been proposed for a vertiport with one or two landing pads while making iterative use of the existing energy-efficient trajectory optimisation tool. A final approach area with a standard arrival route and two approach fixes is used to organise the flow inside the high-density area. Any scheduled delay is absorbed outside this area by flying a shallow descent, an approach fix detour or, only if necessary, in hover. The characteristics of the EHANG-184 multi-rotor eVTOL have been used throughout the research, such that a time separation requirement of 90s is determined. This work is the first to compute the optimal RTAs for eVTOLs to safely separate them for minimum delay based on remaining battery state of charge and vertiport capacity in three modules. Module 1 computes the most energy-optimal arrival trajectory for a set of different RTAs at the vertiport. The state and control vectors corresponding to this set of trajectories are fed intoModule 2. Also, the earliest and latest feasible arrival time based on flight dynamics, most energy-optimal arrival trajectory and the flight time between the approach fix and vertiport landing pad is obtained from this optimisation and fed intoModule 3. Module 2 is used to relate the initial eVTOL battery status to the scheduler in Module 3. It first computes the power required to perform each of the RTA trajectories using the flight dynamics. Afterwards, the power demand and the required SOC to perform each RTA trajectory are determined using a simplified battery model. A regression is then created between the RTAs and the required SOC to compute the latest possible landing time based on the initial SOC of each arriving eVTOL. This so-called RTA constraint is an input for Module 3. Module 3 is a mixed-integer linear programwhich ensures eVTOL separation and selects the arrival route and corresponding landing pad for minimum total delay. A column generation algorithm has been applied to enable delay absorption in hover. Besides, it contains a position shifting constraint with respect to the first-come-first-serve sequence and a rolling horizon algorithmto reduce the computational time required to solve themodel. The concept of operations and eVTOL arrival sequencing and scheduling tool have been tested for a proof of concept and afterwards validated using a hexagonal vertiport network and eVTOL arrival demand model for Houston, TX, USA. The number of eVTOLs expected to arrive has been obtained from the demand model, after which the expected time of arrival for each eVTOL has been modelled as a Poisson process. The initial state of charge of the arriving eVTOLs has been assumed to be normally distributed.

http://resolver.tudelft.nl/uuid:0d0a18da-de1f-4cfa-afe8-b03314b0db49

2019-01-17

Student theses

master thesis

© 2019 Imke Kleinbekman