1 Powering Electric Aircraft at O’Hare Airport: A Case Study

Boya Hou Subhonmesh Bose Kiruba Haran

Abstract—Electrification of aviation can potentially mitigate ratio of the peak power delivered by electric propulsion to that carbon emissions of air travel. Battery-electric and hybrid electric by jet fuel. Retrofitted conventional regional jets and single- aircraft will impose charging requirements on airports. In this aisle (narrow-body) airplanes has the potential to serve as paper, we study the impacts of electrification on electricity an incipient hybrid electric passenger jets, as [4] envisions. demands at O’Hare International (ORD) airport. We consider Switching long-range airplanes to an electric configuration will the scenario where domestic US carriers partially adopt hybrid likely take much longer. electric aircraft (HEA), but operate their airplanes based on cur- rent flight schedules. Under various battery technology evolution scenarios, we then compute the increase in electricity demand Research group Architecture BSED (Wh/kg) Ref. (both daily average and peak) to operate flights at ORD. The data Boeing-GE SUGAR Volt Parallel hybrid 750 [5], [6] analysis reveals substantial increase in airport energy demands Bauhaus Parallel hybrid 1,000 - 1,500 [4] that demands grid upgrades at airports. UTRC Parallel hybrid Not specified [7] Airbus Series hybrid 800 [8] Cambridge Parallel hybrid 750 [9] I.INTRODUCTION Georgia Tech Parallel hybrid 750 [7] In 2018, jet fuel for aviation produced 251 million metric TABLE I: Summary of regional and single-aisle hybrid electric tons of CO2 in the US, accounting for 4.8% of all CO2 emis- aircraft concepts and research sions from energy consumption, as per US Energy Information Administration. Aircraft-generated emissions are predicted to triple in volume by 2050, says the International Civil Aviation If commercial airline carriers adopt electric aircraft within Organization. Electric aircraft can reduce the dependency of air their fleets, charging events at airports would increase the travel on jet fuel and utilize electricity instead. If the trend of airports’ electricity demands. In this paper, we seek to quantify renewable integration in the power grid continues, transition to that increase. Such an analysis will aid grid infrastructure electric aircraft will reduce the carbon footprint and the climate investments required to support electrification of aviation. In impact of this industry. SectionII, we consider flights that landed in or departed from Various configurations such as turbo-electric, hybrid- O’Hare International Airport (ORD) in 2018. By systemat- electric, and all-electric aircraft have been proposed. For ically switching a portion of the flights to HEAs under a example, authors of [1] analyze flight performance of parallel variety of BSED and MF values, we compute their charging turbofan systems, and that of [2] and [3] demonstrate benefits requirements in Section III. The analysis suggests significant of parallel hybrid propulsion system utilized to boost power increase in electricity demands–both in annual aggregate en- during takeoff and climb. The future of commercial airline ergy and daily peak power requirements. We also analyze the operation with electric aircraft crucially depends on projected energy demands from limited growth in battery capabilities as advances in battery technology that enables electric propulsion. predicted in [3] for the near future. The results indicate that While jet fuel has a specific energy density of ∼13000 Wh/kg, BSED greater than 500 Wh/kg will require significant upgrades current batteries offer a mere 200-250 Wh/kg. To derive the to the grid to fully leverage HEAs within commercial airline same amount of energy, batteries therefore weigh considerably operations. more than its aviation fuel counterpart. Moreover, fuel depletes ORD is one of the largest airports in the US, operating over the course of the flight. Thus, an aircraft loses weight north of 900K flights annually. Charging requirements will during travel. The higher the weight of an aircraft, the greater be naturally higher at larger airports. A switch to HEAs is the energy required to propel it. Battery weight poses a will rely on capital-intensive and time-consuming upgrades fundamental challenge to operation of electric aircraft. Weight of the grid infrastructure at such airports. Our framework to considerations limit the range of distances such aircraft can calculate electricity demands, however, applies more generally. travel. We present preliminary results with data from Los Angeles In- The aviation industry anticipates a rapid rise in battery spe- ternational Airport (LAX) in SectionIV. The paper concludes cific energy density (BSED), as evidenced by TableI. There is with remarks and future directions in SectionV. a growing consensus in the industry that commercial operation of short-range parallel hybrid aircraft will become viable in II.METHODOLOGYTO COMPUTE CHARGING the near future. Energy requirements of hybrid electric aircraft REQUIREMENTSFOR REGIONAL FLIGHTS (HEA) also depend on the motor factor (MF) that equals the Electricity demands of HEAs will depend on how many flights switch over to HEAs as well as on their flight and All authors are affiliated with the Department of Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign, Urbana, IL charging schedules. Number of HEA-operated flight legs will 61801. Emails: {boyahou2, boses, kharan}@illinois.edu. The authors thank depend on the battery technology that determines the maxi- Lavanya Marla from Illinois for helpful discussions. mum range of such aircraft. We make several assumptions to 2

# Tail Number Departed from Next Destination Time of Arrival Time of Departure Distance (mi) Duration (hr:min)

1 N804UA ORD PHL 09:30 10:54 426 01:24 2 N804UA PHL ORD 13:54 16:15 426 02:21 3 N804UA ORD LGA 17:36 19:15 455 01:39

TABLE II: Example rows of processed dataframe obtained from BTS dataset.

BSED (Wh/kg) determine ranges. Then, we utilize flight schedules of domestic Type MF (%) commercial flights from the US Department of Transporta- 500 700 1000 1250 1500 tion’s Bureau of Transportation Statistics (BTS) to compute 12.5 31.1 30.7 30.4 30.2 30.0 charging requirements. We assume that flight schedules do not ERJ-175 25 64.8 63.6 62.6 62.1 61.6 change, even though airlines adopt HEAs to replace conven- 50 - 133.3 130.6 129.1 127.9 tional aircraft in their fleet. Modification of flight schedules 12.5 26.5 26.3 25.7 25.4 25.2 to accommodate both charging requirements and passenger Boeing 737-700 25 57.3 55.5 54.0 53.2 52.6 demands remains an interesting direction for future work. 50 - 119.8 115.1 112.8 111.2 We use the BTS Airline On-Time Performance data for flights in 2018 to compute energy demands. The dataset TABLE III: Battery energy usage γ in Wh per passenger-mile describes all non-stop domestic flights reported by certified of ERJ-175 and Boeing 737-700. US carriers. Relevant details for each flight includes its tail BSED (Wh/kg) number, scheduled and actual departure and arrival times, Type MF (%) origin and destination airport, and non-stop distance. We 500 700 1000 1250 1500 process the data to construct a dataframe shown in TableII as 12.5 683.1 914.2 1170.7 1339.7 1476.6 follows: (i) filter flights with arrival or departure to be ORD, ERJ-175 25 305.9 484.1 714.1 877.45 1020.1 (ii) group the data by date, (iii) group by tail number on a 50 - 170.2 331.2 458.8 579.6 given day, and sort by departure time for a given tail number. 12.5 1110.9 1466.2 1876.8 2141.3 2357.5 This newly constructed data-frame allows us to conclude how Boeing 737-700 25 558.9 836.1 1193.7 1452.4 1680.1 long each airplane stays at ORD that proves useful to compute 50 - 376.1 637.1 842.9 1038.4 its charging requirement at ORD. TABLE IV: Maximum range of HEAs in miles In computing the charging profiles, we only consider do- mestic commercial flights. Thus, our analysis does not reflect charging requirements of long haul, wide body international flights, freight and military aviation. Given serious range battery energy consumed during taxi, takeoff, cruise, approach, limitations of HEAs, international flights will likely utilize and landing. To estimate γ for a flight in the BTS dataset, conventional aircraft in the near future. Our analysis can be we first identify whether it is a regional jet or a single-aisle repeated to cover freight and other sectors to more accurately (narrow-body) aircraft. For regional jets, we utilize γ for ERJ- estimate the energy needs of airports. 175 and for that of Boeing 737-700 for single-aisle aircraft. We only consider HEAs obtained from retrofitting conven- Values are given in Table III. Notice that battery sizes required tional aircraft with electric propulsion systems. Novel HEA for BSED below 700 Wh/kg and MF above 50% are deemed designs optimized for electric propulsion, e.g., in [10], will no impractical for any useful flight range. doubt alter the conclusions drawn from this study, and are left for future efforts. B. Flight Legs that Operate HEAs In the BTS dataset, regional jets and narrow body aircraft A. Energy Requirement of a Single Flight each accounted for nearly half of the 300K+ short-range Battery energy needed to travel d miles with p passengers domestic flights that had a dwell time at ORD in 2018. We on an HEA is given by assume that all domestic commercial flights operate an HEA E = p × d × γ, (1) to serve a trip if it becomes feasible to do so, given the flight range of the aircraft with an MF and BSED combination. where γ is the battery energy usage per passenger-mile. For Adoption of HEAs by airline companies in practice will likely each flight in the BTS database, we use the tail-number to deviate from such operation. A similar study with possible identify the aircraft type from airplane manufacturer’s web- HEA adoption paths is relegated to future work. sites, that in turn, gives us the total number of seats on the Range capabilities of retrofit hybrid electric regional jets and plane. We assume a uniform load factor of 85% to estimate narrow body aircraft are adopted from [11] in TableIV. Figure p for each flight, consistent with yearly average load factor 1 provides the distribution of flight distances with regional jets estimates by BTS. and single aisle aircraft from the BTS dataset. Flight range The values of γ for HEAs are adopted from [11], assuming together with the distance distribution determine the number a battery-pack voltage of 128V. Their analysis accounts for of flights operated with HEAs. 3

For different MF and BSED configurations, we compute E from (1) and report in TableV. Dividing E by 60 minutes yields the power requirements of N33289 from 10:41 to 11:41am on 03/20/2018.

BSED (Wh/kg) MF (%) 500 700

12.5 4.36 MW 4.25 MW 25 9.28 MW 8.99 MW

(a) Regional Jets TABLE V: Power requirements of N33289 for flight to TPA.

III.ANALYSIS OF CHARGING REQUIREMENTS AT ORD Having described our data and method to compute charging requirements of HEAs, we now present the results from our analysis. In the sequel, compare the demand increase from HEAs to the average power consumption of 50.3MW at ORD in 2002, computed from the aggregate annual electricity de- mand of 440.7 GWh, according to the O’Hare Modernization (b) Single Aisle Aircraft Final Environmental Impact Statement, published in 2005. Fig. 1: Histogram and cumulative distribution function (CDF) of flight distances in BTS database.

C. Power Demand for Each Flight Leg Assume that each HEA arrives at ORD with a depleted battery. It immediately starts charging at a constant rate that allows it to charge the battery up to the energy needs of the next flight leg, by the time it leaves the gate for its next destination. Thus, power required to serve an outgoing flight is given by P = E/T, Fig. 2: Aggregate annual electricity demand of HEAs at ORD under various battery MF and BSED settings. where E is given by (1) and T is the dwell time at ORD. Divide each day into 1440 one-minute intervals. Adding the charging requirements of all flights in a given 1-min interval Figure2 illustrates that moderate MF and BSED values lead yields the charging requirements for that interval. We omit to annual battery energy consumption comparable to aggregate flights with connection times shorter than 15 minutes at ORD. demand of ORD in 2002. This plot confirms that electrification Our approach to compute charging requirements ignores of commercial airlines can considerably amplify electricity several practical considerations. First, maximum charging rates demand requirements at airports. typically depends on the ratings of the battery and its charging For a given MF, one might expect total energy consumption equipment. We ignore such upper bounds on this rate. Second, from batteries on HEAs to decrease with BSED, as one HEAs with stops at ORD may vary in their charging require- requires lighter batteries to deliver the same amount of power. ments. Residual charge in the battery can cover portion of the However, that is not always the case as Figure2 reveals. To energy requirements of the next flight. It may be necessary to explain this apparent paradox, notice that lighter batteries with charge an HEA for a round-trip journey from ORD and back improved BSED allows HEAs to operate more flights over to compensate for possible lack of charging infrastructure at longer distances, increasing aggregate charging requirements. the destination airport. Third, batteries often have roundtrip Grid infrastructure to deliver power at airports must be efficiency losses, that we ignore. able to cover daily peak power demands from HEAs. The To illustrate the power demand requirements through an histograms in Figure3 suggest a substantial increase in peak example, consider a single hybrid electric retrofit of Boeing power demands. For MF = 12.5%, BSED of 500,700 and 1000 737-800 aircraft with tail number N33289 operating a 1012- Wh/kg yield median peak power requirements of 70.6, 101.2 miles flight to TPA (Tampa, FL) on 03/20/2018. N33289 and 169.4 MW, respectively. Highest daily peak demands can arrives at ORD at 10:41 AM and leaves for TPA at 11:41AM. be ∼250MW, indicating the need for significant grid upgrades. 4

(a) MF=12.5% (b) MF=25% (c) MF=50% Fig. 3: Histogram of daily peak demands with different MFs. Here, ( ) plots the histogram for BSED = 500 Wh/kg, ( ) plots with BSED = 700 Wh/kg, and ( ) plots with BSED = 1000 Wh/kg.

(a) MF=12.5%, BSED=500Wh/kg (b) MF=12.5%, BSED=700Wh/kg (c) MF=12.5%, BSED=1000Wh/kg Fig. 4: Average daily power consumption profiles for HEAs with various MF and BSED configurations. Here, ( ) plots power requirements of single-aise aircraft, ( ) denotes the same for regional jets, ( ) denotes the sum-total of the above two, and ( ) denotes the average consumption over the day.

Compare these numbers to 50.3 MW, the average power Daily charging profiles averaged across 2018 in Figure4 in- demand on ORD in 2002. dicate that energy requirements of regional jets are significantly smaller than that of single-aisle aircraft. Figure1 suggests that single-aisle aircraft serve more long-distance flights than regional jets. For regional jets, the median distance of flights is 409 mi. The same number for single-aisle aircraft is 867 mi. Longer distances translate to higher energy requirements. The gap between the two different aircraft classes increases as larger ranges are made possible with higher BSED at a constant MF.

A. Estimates of Charging Requirements in the Near Future Despite ambitious targets for battery technology from avi- ation industry in TableI, battery experts project more con- Fig. 5: Profile of charging requirements on 08/10/2018 with servative growth in battery capabilities in the next decade. MF = 12.5%, BSED = 1000 Wh/kg. Recent study in [12] anticipates that current BSED of 200- 250 Wh/kg will grow to 300 Wh/kg at pack level by 2025 and A representative charging profile of ORD airport for HEAs 400-500 Wh/kg by 2030. Densities beyond 700 Wh/kg are in Figure5 has a peak demand of 187 MW, while the average more than a decade away. In light of this study, we compare demand is 74 MW, that is lesser than half the peak. This our earlier results to one with a restricted battery performance analysis suggests that peak demands can be considerably with BSED = 355 Wh/kg. With such a battery configuration, reduced by optimizing the flight and its charging schedule the authors of [3] advocate to use batteries as the energy source instead of operating HEAs with current schedules. The de- only during takeoff and climb. Following [3], we restrict the sign of efficient flight and charging schedules accounting for range of such flights to 287 mi and take γ = 3.0589 Wh per airlines operating HEAs within their fleet defines an interesting passenger-mile. Figure6 provides the histogram of daily peak research question. electricity demands for charging HEAs under this restricted 5 setting. The median charging requirement is 6.88MW, an order answering the larger question: is the grid infrastructure at the of magnitude smaller than with higher BSED values. And, the airports, and the power system more broadly, ready for aviation annual energy demand of HEAs is 18.4 GWh. Such increases electrification? We make several assumptions about range ca- can likely be accommodated through selective updates to pabilities and calculate charging requirements of HEAs using the distribution grid or through on-site generation at ORD, existing flight schedules out of ORD under various BSED e.g., from the envisioned solar farm. As battery technology and MF projections. Our results indicate that while near-term breaks the 500 Wh/kg barrier, HEAs can exploit the on- capabilities of HEAs will only marginally affect airport elec- board batteries for taxi, takeoff, cruise, approach, and landing, tricity demands, the increase will be substantially higher with requiring a commensurate increase in charging infrastructure the anticipated growth in battery technology. Supporting grid at the airports. infrastructure will play a key role in successfully electrifying the aviation industry. We aim to utilize our framework for a comprehensive study of power requirements from HEAs at all major airports in the US. We also aim to study how airlines with partially electrified fleets would design their flight schedules and charging plans to account for energy requirements of electric aircraft.

REFERENCES [1] J. C. Gladin, D. Trawick, D. Mavris, M. Armstrong, D. Bevis, and K. Klein, “Fundamentals of Parallel Hybrid Turbofan Mission Analysis with Application to the Electrically Variable Engine,” 2018 AIAA/IEEE Electric Aircraft Technologies Symposium, EATS 2018, pp. 1–16, 2018. Fig. 6: Histogram of daily peak demand with HEAs capable [2] C. Lents, L. Hardin, J. Rheaume, and L. Kohlman, “Parallel hybrid gas- of flying up to 287 mi on BSED = 355 Wh/kg. electric geared turbofan engine conceptual design and benefits analysis,” 52nd AIAA/SAE/ASEE Joint Propulsion Conference, 2016, pp. 1–13, 2016. [3] P. Bertrand, T. Spierling, and C. E. Lents, “Correction: Parallel Hybrid Propulsion System for a Regional Turboprop: Conceptual Design and IV. PRELIMINARY RESULTS FROM Benefits Analysis,” no. August, pp. 1–7, 2019. CASE STUDYFOR LAX [4] C. Pornet, C. Gologan, P. C. Vratny, A. Seitz, O. Schmitz, A. T. Our framework can be applied to analyze the charging Isikveren, and M. Hornung, “Methodology for Sizing and Performance requirements of HEAs at any major airport in the US. Figure Assessment of Hybrid Energy Aircraft,” Journal of Aircraft, vol. 52, 7 depicts the annual demands from HEAs with BTS data for no. 1, pp. 341–352, 2014. LAX. Median daily peak power demands become 31.0 MW, [5] M. K. Bradley and C. K. Droney, “Subsonic ultra green aircraft research: Phase I,” 2011. 55.9 MW, 90.6 MW, respectively with BSED = 500, 700, 1000 Wh/kg and MF = 12.5%. Compare these numbers to the total [6] ——, “Subsonic ultra green aircraft research: Phases II,” 2015. [7] C. Propulsion, E. Systems, C. Aviation, C. Emissions, S. E. Board, electricity use of 184 GWh in 2015 according to [13], that D. Engineering, P. Sciences, and N. Academies, Commercial Aircraft amounts to 21.0 MW of average power. Propulsion and Energy Systems Research. National Academies Press, 2016. [Online]. Available: http://www.nap.edu/catalog/23490 [8] J. Delhaye and P. Rostek, “Hybrid Electric Propulsion Europe-Japan Symposium Electrical Technologies for the Aviation of the Future,” Europe-Japan Symposium Electrical Technologies for the Aviation of the Future, no. March, 2015. [9] C. Friedrich and P. Robertson, “Hybrid-Electric Propulsion for Aircraft,” Journal of Aircraft, vol. 52, no. 1, pp. 176–189, 2014. [10] M. K. Bradley and C. K. Droney, “Subsonic Ultra Green Aircraft Research: Phase II – Volume II – Hybrid Electric Design Exploration,” Nasa/Cr–2015-218704, vol. II, p. 207, 2015. [11] G. E. Wroblewski and P. J. Ansell, “Mission Analysis and Emissions for Conventional and Hybrid-Electric Commercial Transport Aircraft,” Journal of Aircraft, no. January, pp. 1–14, 2019. [12] A. Misra, “Summary of 2017 NASA Workshop on Assessment of Fig. 7: Aggregate annual electricity demand of HEAs at LAX Advanced Battery Technologies for Aerospace Applications,” pp. 1–18, under various battery MF and BSED settings. 2018. [13] “LAWA: Sustainability Elements Energy Management,” https: //www.lawa.org/en/lawa-sustainability/sustainability-elements-energy, accessed: 11/07/2019. V. CONCLUDING REMARKS In this paper, we estimate the demand requirements from HEAs that replace a portion of short-range domestic flights that pass through ORD. This work defines the first step towards