Sonic Boom Over-land Avoidance and the Impact on Economic Feasibility Holger Pfaender 1 Georgia Institute of Technology 270 Ferst Dr Atlanta, GA 30332 United States Jiajie (Terry) Wen 2 Georgia Institute of Technology 270 Ferst Dr Atlanta, GA 30332 United States Dimitri Mavris 3 Georgia Institute of Technology 270 Ferst Dr Atlanta, GA 30332 United States

ABSTRACT There has been renewed interest in commercial supersonic air travel in recent years with sev- eral new efforts being pursued by several companies. They currently advertise entry into ser- vice of these new concepts within the next decade. Current regulations do not permit overland operations for supersonic vehicles due to the sonic booms produced during flight. As a result, these vehicles will need to abide by the over land boom restrictions. This study presents a com- prehensive approach to overland-prohibited supersonic flight routing and route-specific de- mand forecasting for commercial supersonic air travel, in order to assess the expected market in 2050. This also includes the need for the seasonally-dependent need to avoid secondary sonic boom exposure of land. 1. INTRODUCTION

In recent years, a number of start-up companies have started development on new commercial supersonic transports (SSTs). These aircraft represent a potential second generation of supersonic aircraft, which would not include low boom design such as explored in the experimental X-59. The companies pursuing this type of aircraft design include Boom Supersonic, Exosonic, and others.

A key consideration for these supersonic concepts is not only the technical feasibility to meet all requirements, but also economic viability. One important limitation for these aircraft is the essential 1 Holger.pfaender@asdl.gatech.edu

2 twen@gatech.edu

3 dimitri.mavris@aerospace.gatech.edu

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

complete ban of supersonic booms generated reaching land. This has been a key consideration in exploring the economic viability because if these aircraft, which are designed to fly primarily at su- personic speeds, are forced to spend long parts of actual operations missions flying at subsonic speeds, the time savings can be severely reduced while the operating cost is increased. This study is intended to investigate the impact of different operating scenarios (related to sonic boom ground exposure) on the feasibility and viability of commercial supersonic operations. 2. MODELING AND SIMULATION METHODS

Analyzing the feasibility and viability of commercial supersonic flight is a multifaceted problem that requires a wide variety of domain knowledge. It is impossible to cover all the relevant topics in detail, but the following subsections will provide an overview on sonic boom, flight path planning, aircraft mission analysis, aircraft modeling, and demand estimation. 2.1. Sonic Boom Considerations

There are many different ways for sonic boom to propagate through the atmosphere and reach land. When the sonic boom disturbance emanating from the aircraft reaches ground directly, the area of contact is known as the primary boom carpet. Primary boom’s signature (overpressure) on ground is usually “N”-shaped, after all the smaller shocks and expansion waves generated by the aircraft coalesce as the disturbance travel through the atmosphere. Figure 1 below is a schematic of the two types of sonic boom and their boom signatures. It should be noted that when the aircraft is accelerating beyond Mach 1, the shocks from the aircraft can coalesce and form a stronger boom known as tran- sition focus boom. Boom focusing can also occur when the aircraft is turning or diving. As a result, extra care is needed when maneuvering an SST near land at supersonic speeds.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 1. Types of sonic booms and their signatures, from [1]

Secondary boom’s propagation paths are more complex than those followed by the primary boom’s rays. Both the rays emanating above the aircraft and reflected rays after primary boom’s incidence with ground can bend downward again by means of diffraction, and eventually reach ground at some distance farther away than the primary boom carpet. The secondary booms are not as

loud as the primary booms, and they tend to sounds like rumbles instead of bangs. Because of the longer and more complex propagation paths, these secondary booms typically reach ground at over 100 nmi away from the instantaneous aircraft location in which they are generated. Figure 2 shows the maximum extent of the secondary boom’s ground interception, modeled using an arctic tempera- ture profile that is known to favor secondary booms under no-wind conditions. The specific condi- tions under which these secondary booms reach ground is subject to upper stratosphere wind and temperature gradients. The details of this are described in [2] with several follow-on investigations being conducted in the years since.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 2. Maximum extent of secondary boom ground interception, from [2] 2.2. Flight Routing Algorithm

Two different types of path planning methods have been successfully implemented by researchers to find the optimal trajectory under some pre-defined criteria. The first is an analytical approach based on optimal control theory, which solves a set of ordinary differential equations (ODEs) that govern the aircraft’s flight heading. Even though this type of method is widely used to calculate wind-optimal subsonic flight trajectories, they have some limitations when analyzing supersonic missions. The pri- mary challenge is that these methods use penalty functions to model less favorable regions (land areas in the case of supersonic flights). However, imposing penalty functions that follow the coastline of the continents could be non-trivial.

The second type of algorithm used for flight trajectory optimization is raster-based. Unlike the ODE-based analytical approach, these algorithms discretize physical space into graphs (or more simply, grids). The routing algorithm used in this research is based on the Theta* algorithm [3], which is a modification of the well-known A* search algorithm. The main difference is that the Theta* algorithm performs a “line of sight” check after every iteration to find out if there is a clear path from the parent node of the current node to the chosen neighbor node.

DISTANCE AHEAD OF AIRCRAFT (N.M.) 250: 200: 150. 100. 50 7 . a y “, 7\ 24 am H GMG Rays MG Rays Aircraft at origin Flight along vertical axis 50 100 150 200 250 LATERAL DISTANCE (N.M.) a MG Rays are initially upward GMG Rays are initially downward, reflected from primary carpet

Figure 3. Example implementation of the Theta* algorithm 2.3. Mission Analysis

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

After the routing algorithm optimizes the trajectory for a given route, the trajectory data is then fed into FLOPS, an aircraft mission analysis tool [4]. FLOPS will determine whether the trajectory is feasible – i.e., complete the flight with enough reserve fuel. The routing algorithm’s default setting will result in a more time-optimal solution. If a route is not feasible, then the routing algorithm is rerun with a more fuel-optimal setting. A refuel stop will be added if this is still not adequate, which leads to an estimated time penalty of 90 minutes (that accounts for taxi-in, refuel, and taxi-out time). 2.4 Commercial Supersonic Demand Estimation

The demand forecasting process starts with existing routes in the current commercial flight net- work and the market forecasts. The approach relies on calculating a “switching percentage” of pre- mium passengers who would choose to fly supersonically if enough value is provided (in terms of time savings). The impact of induced demand (the additional demand that could occur purely due to the availability of supersonic service) is neglected. Even though multiple mechanisms could poten- tially lead to induced demand, it is unclear whether this would constitute a significant number of additional passengers. Figure 4 summarizes the overall approach implemented:

Line-grid intersection Start node (1 Goal node MM Obstacle —— > A* shortest path —— > Theta* shortest path

Figure 4. Commercial supersonic demand assessment process 3. EXPERIMENTS

This section first introduces the aircraft model and the process to obtain the list of potential com- mercial supersonic routes in 2050. Then, the land buffer variation is discussed in detail, and the set of filters to ensure route viability is presented. The routes that remain after all the filtering process are considered to be feasible and viable commercial supersonic routes.

[51 Input [4 Routing algorithm [= Calculation [5 output Potent Supersoni Routes Aircraft Characteristics irline Cost. Structure Income Distribution | __ Compute VTTS and Potential Demand Compute Time Savings | | H Compute SST Fare per Unit Time Saved | Compute Switching Rees Percentage

3.1 Aircraft Model

For this study, the 55-passenger Mach 1.8 supersonic transport (SST) vehicle model is used for flight path planning and mission analysis. The vehicle was developed by ASCENT 10 Project team in the Aerospace Systems Design Laboratory at Georgia Tech [5]. Even though the team has devel- oped a variety of vehicles with different seating capacities and Mach numbers, the authors choose the 55-pax Mach 1.8 variant because this is the vehicle that is closest to the Boom Overture aircraft.

Figure 5. ASDL 55-pax Mach 1.8 SST 3.2. Routes Considered

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Since an extensive commercial flight network already exists, this research assumes that airlines will not introduce supersonic-specific routes. Even though such possibility cannot be excluded, it is likely going to occur on a case-by-case basis and would require much more in-depth market analyses. The database used for global commercial aviation routes is the Federal Aviation Administration (FAA) Global Inventory of 2015. This data includes the total number of operations and estimated total number of seats [6]. Additionally, airport data (such as location, runway length, etc.) is retrieved from the FAA Aviation Environmental Design Tool (AEDT) [7]. These two inventories provided the essential information to filter routes based on great circle distance and passenger demand.

To assess the future market of supersonic transport, current subsonic routes with potential for su- personic operations is first identified. Such routes have to exceed a certain minimum distance to guarantee value in time savings. They also need to be popular routes for premium-class passengers, so that enough of them are likely to switch to supersonic service. Generally, any long-distance route with high demand would be considered a potential supersonic route. 3.3. Aviation Growth Forecast and Preliminary Route Filtering

Since this study accesses the expected commercial supersonic operations in 2050, existing aviation traffic need to be forecasted out to future years. To accomplish that, this study uses the 2019 Boeing Commercial Market Outlook (CMO) [8]. The Boeing CMO divides the world into 12 different re- gions and includes the forecasted traffic growth between them up to 2038 (growth rates were extrap- olated beyond 2038). Potential supersonic routes are filtered from the forecasted 2050 subsonic de- mand based on distance and seating capacity. Routes with daily passengers less than the SST’s seating capacity are disregarded. Furthermore, routes with great circle distance below 1500 nmi are excluded, since an SST will not be able to leverage its speed advantage on short flights.

3.4. Land Buffer Variation

When optimizing the trajectories for potential supersonic routes, a fixed land buffer is imposed globally to ensure that the SST is at least a certain distance away from land. In this study, two land buffer distances (27 nmi and 150 nmi) are imposed. The Concorde flight manual [9] gives guidelines for the protection distances for level flight and various turn angles. The routing algorithm described above allows aircraft movement in 8 directions, which limits the largest turn angle to 45⁰. The 27 nmi buffer was recommended by the Concorde flight manual for 45⁰ turns. Since the M1.8 SST is lighter and slower than the Concorde, this can be considered as a conservative estimate.

The larger buffer that includes standoff distances including secondary booms was based on the data shown in Figure 2 that includes lateral and forward ray cone distances for a cruise Mach of 1.8 to ensure even ground reflected booms will not reach land.

The buffer is constructed from the Natural Earth land vector [10]. To add the buffers, the land vector shapes were expanded via a simple polygon set operation. However, these operations assume cartesian coordinates, while the Earth’s land data is provided in WGS84 latitudes and longitudes. The distance between two points separated by a fixed number of degrees in the longitudinal direction shrinks to zero the closer they get near the poles. Therefore, simply using a fixed conversion of the distance from nautical miles at the equator to degrees will result in much too small buffer sizes in the longitudinal direction near the poles.

To generate proper buffers, it would be ideal to convert land polygons to a distance preserving map projection. However, no such projections work simultaneously for the entire globe. The area preserving projection “Mollweide” was used instead. Most issues that arose were fixed manually, but some obscure artifacts remained. The final buffers around the Earth’s land masses are shown in Figure 6 (land in brown, 27 nmi buffer in dark blue, and 150 nmi buffer in light purple).

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 6. Land areas and buffers

The greater land buffer tested in this study could have a huge impact on supersonic flight opera- tions, since it renders some regions infeasible for supersonic flight (such as most of Southeast Asia and the Mediterranean Sea). For flights over other regions, the aircraft would need to either reduce the supersonic flight segment or take a distance penalty due to farther detours.

3.5. Viability Filtering

For a route to be deemed viable, it had to meet the following criteria:

1. 20% time savings or more relative to the reference subsonic aircraft 2. 2 hours of time savings or more relative to the reference subsonic aircraft 3. at least one flight per day in 2050 4. increase in ticket fare per hour saved is less than $1,000 The aforementioned land buffer variation can have a direct impact on route viability, since an increase in land buffer leads to additional detours and reduces supersonic flight segment. In some cases, it can also prevent the aircraft from threading between islands. 4. RESULTS

Here are some relevant assumptions used in this study:

1. Jet fuel density is 0.8 kg/L 2. Jet fuel to CO 2 conversion is 3.156 kg fuel /kg CO2 3. Passenger load factor is 70%

4. ASK is calculated using GCD and not rerouted distance 5. Refuel takes 90 minutes 6. Forecast year is 2050 The effect of costal buffer variation on the feasible and viable routes is shown in Figure 7:

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 7. Comparison of the 27nmi-buffer and 150nmi-buffer networks Since the extended land buffer will cover most of the Mediterranean Sea and the Black Sea en- tirely (seen in Figure 6), many supersonic routes that pass through the Middle East are no longer viable. Similar situation can be observed in Southeast Asia, where the extended buffer makes island threading much more difficult, removing most supersonic flights in that region.

The results are presented in Table 1 through Table 3. As mentioned previously, the 27 nmi buffer corresponds to the buffer distance needed to avoid primary boom from reaching land in most cases, while the 150 nmi buffer should be adequate to keep the secondary booms away from land.

Table 1. Number of routes and refuel percentage

Coastal Buffer

Feasible and Viable Routes

% Routes with Refuel

27 nmi 752 33.8% 150 nmi 371 27.8%

The number of feasible and viable routes reduced by roughly 51% when the coastal buffer limit increases from 27 nmi to 150 nmi. Following the trend seen in Table 1, the reduction in the number of routes leads to 57-62% decrease in the key transportation metrics presented in Table 2.

Table 2. Key transportation metrics

Total Annual Flights (Thousands)

Total Annual Passengers (Millions)

Total Annual Flight Distance (Billion km)

Total Annual Flight Hours (Millions)

Total Annual ASK (Billions) 27 nmi 1295 49.9 8.534 6.849 469.358 150 nmi 490 18.9 3.467 2.968 190.690

Coastal Buffer

Table 3. Fuel efficiency and emissions

Total Annual CO 2 (Megatonne)

Fuel Efficiency (RPK/L) 27 nmi 49.98 157.75 814.66 0.106 5.26 150 nmi 20.93 66.05 313.44 0.110 5.10

Total Annual Fuel Burn (Megatonne)

Total Annual NO x (Megatonne)

Fuel Intensity (kg/ASK)

Coastal Buffer

For the 150 nmi buffer case, the total fuel burn and emissions also decrease in a similar fashion. However, it is interesting to note that in terms of fuel intensity/efficiency, the difference is not very significant. The authors suspect that this is because the viability filters imposed are removing the routes that make the SST inefficient to fly under the extended costal buffer scenario. 5. CONCLUSIONS

Potential commercial supersonic aircraft will face many challenges in the attempt to successfully re-enter this currently no-longer-serviced part of the market. One of these challenges will be the sonic boom over-land avoidance, at least until future generations of supersonic aircraft can be designed with low boom characteristics and potentially make such exposure permissible. Current or near-future technology SST designs will be less economically viable due to sonic boom over-land constraints. This study shows that including secondary sonic boom restrictions significantly reduces viable routes compared to analyses that do not include such constraints. Using simplified constraints can serve as a first pass in identifying potentially affected routes. Any future manufacturer or operator would have to perform a more detailed route and weather-specific analysis to confirm viability.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

6. ACKNOWLEDGEMENTS

This research was funded by the U.S. Federal Aviation Administration Office of Environment and Energy through ASCENT, the FAA Center of Excellence for Alternative Jet Fuels and the Environ- ment, project 10 through FAA Award Number 13-C-AJFE-GIT-006 under the supervision of Ranga- sayi Halthore, Maryalice Locke, and Laszlo Windhoffer. Any opinions, findings, conclusions or rec- ommendations expressed in this this material are those of the authors and do not necessarily reflect the views of the FAA. 7. REFERENCES 1. D. Maglieri, et al., Sonic Boom: Six Decades of Research, NASA/SP-2014-622 (2014). 2. Baize, Daniel G., ed. The 1995 NASA High-Speed Research Program Sonic Boom Workshop. Vol. 1. NASA, Langley Research Center, 1996. 3. K. Daniel, A. Nash, S. Koenig, and A. Felner, “Theta*: Any-angle path planning on grids,” Journal of Artificial Intelligence Research, vol. 39, pp. 533–579, 2010 4. McCullers, L. A., “Aircraft Configuration Optimization Including Optimized Flight Profiles,” in Proceedings of the Symposium on Recent Experiences in Multidisciplinary Analysis and Optimization, Part I, NASA Langley Research Center, 1984. 5. D. Mavris, W. Crossley, J. Tai, and D. Delaurentis, “Project 010 aircraft technology modeling and assessment,” FAA Center of Excellence for Alternative Jet Fuels & Environment, Tech. Rep., 2020 6. FAA, 2015 inventory as modeled in Aviation Environmental Design Tool (AEDT), 2016. 7. FAA, Aviation Environmental Design Tool (AEDT) version 3c – user manual, 2020. 8. Boeing, Commercial market outlook 2019–2038, 2019. 9. Air France, Manuel d’utilisation Concorde, 2003. 10. Made with Natural Earth. Free vector and raster map data @ naturalearthdata.com.[Accessed April 25 2022]

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW