2.1 Top-level aircraft requirements

The top-level aircraft requirements (TLAR) are selected according to the authors’ previous work described in [Reference Nasoulis, Gkoutzamanis and Kalfas11] and the mission profile used for the design evaluation is a simple mission profile including a diversion mission. The design range is 400nm with an additional 100nm of reserves, cruising altitude is at 10,000ft for both all engines operating (AEO) and one engine inoperative (OEI) cases, cruise speed is 235 KTAS (0.35 M), and 30 min of loiter time are considered. Both conventional and hybrid aircraft follow the same TLAR, are sized for the same mission profile, and the powertrain units are sized for the OEI case. The maximum passenger number for a Level 4, CS-23 certified, normal-category aircraft is 19, [17] while the maximum payload capacity is selected to be $1,900kg$ , considering $100\textrm{kg}$ per passenger ( $87\textrm{kg}$ per person, and $13\textrm{kg}$ carry-on luggage) [Reference Nasoulis, Gkoutzamanis and Kalfas11].

The maximum range selection is a result of an airport network study, considering possible operating markets in European countries. Three countries were selected to consider several country shapes, i.e. Greece, Sweden and Switzerland. Greece has large island complexes with small airports and the mainland is at the tip of the Balkan peninsula. Sweden has a longitudinal axis almost three times the latitudinal, and Switzerland is small, almost square, but with the Alps posing flight altitude restrictions. With the global coordinates of the aerodromes placed on the map for all three countries, their mutual distance is calculated, following a point-to-point mission strategy, using the haversine formula. The mutual airport distance distribution for all three countries is shown in Fig. 1, with the average and median distance marked. By observing Table 1, the maximum average airport distance appears in Sweden, as it’s the largest country of the three. However, when compared to the average airport mutual distance in Greece, it is only 20% greater, considering that Greece has less than one-third of Sweden’s area, but many airports are scattered in island complexes. Furthermore, the coverage rate for various mission ranges is calculated and shown in Table 1, namely for the 300, 400 and 600 nm cases. It appears that for a mission range of 600 nm, the coverage rate reaches 100% for Greece and Switzerland and 97.6% for Sweden, whereas when a range of 400 nm is selected, the coverage rate remains almost unaffected for Greece and Switzerland but is reduced to 85.3% for the case of Sweden. However, since the majority of Sweden’s airports are located in the southern half of the country, the 85% coverage rate is considered acceptable, and the mission range of 400 nm is selected. Additionally, an extra 100 nm of reserves range is considered for redundancy purposes in the design, which can be utilised in the case of Sweden to increase the coverage rate.

Figure 1.

Mutual airport distance distribution in Switzerland, Greece, and Sweden for a point-to-point network.

Table 1.

Coverage rates for different mission ranges in Greece, Sweden, and Switzerland

2.2 Aircraft powertrain requirements

The Parallel hybrid configuration consists of two electrically boosted turboprop engines, whereas the Series hybrid consists of four propellers connected to four electric motors, as shown in Fig. 2. A turboshaft engine is located aft of the Series aircraft configuration, which is coupled to an electric generator able to either charge the batteries or provide power directly to the electric motors. The power output of the gas generator is selected according to the aircraft operating modes, which will be described in the following paragraph.

Figure 2.

Parallel (left) and Series (right) hybrid-electric propulsion configurations.

The Parallel configuration operates as conventional during ground operations, descend and approach phases, whereas it operates in hybrid mode in all other phases, as observed in Fig. 3. Additionally, onboard charging is not possible, due to the powertrain architecture (Fig. 2). Furthermore, the Series hybrid configuration operates in electric mode during ground operations, in conventional mode during descend, approach and loiter phases, and in hybrid mode during the rest. The Series architecture enables onboard charging during all conventional operating modes in which the gas generator’s available power exceeds the mission segment power requirement. The power output of the gas generator is defined by the loiter mission segment, as it is the most demanding phase of which the aircraft operates in conventional mode. Moreover, to prolong the battery life, the charging current is limited to $C/3$ of the selected batteries’ C-rating, and the maximum charging power is calculated using that current limit. The power management strategy selected for each propulsion architecture aims to maximise the contribution of the electric motors and batteries to the overall thrust production while maintaining a battery state of charge of at least 20% at the end of the mission, for battery preservation and safety reasons.

Figure 3.

Design mission profile definition (left) and hybrid-electric operational modes (right).

Considering the state of the art of electrical components, three different EIS dates are explored and presented in Table 2. A different battery type is considered for each EIS date, based on their current TRL [Reference Gkoutzamanis, Kavvalos, Srinivas, Mavroudi, Korbetis, Kyprianidis and Kalfas9], namely Solid State LI-ion batteries, Li-Sulfur and Li-Air, for 2027, 2030 and 2040, respectively. Each battery technology promises a different gravimetric specific energy at cell level that ranges from $350 - 1,050\textrm{Wh/kg}$ . However, the cell-level specific energy must be translated into pack-level specific energy, prior to estimating the total amount of batteries required in a hybrid-electric aircraft. In that direction, the mass of the accumulator container, cooling system and electronics must be accounted for, in the overall battery system mass calculation. The electronics and the electronic auxiliaries masses are estimated empirically, whereas the container mass is calculated as a function of battery cells dimensions, and finally, the cooling system dimensions and mass are calculated analytically [Reference Gkoutzamanis, Tsentis, Valsamis-Mylonas and Kalfas18]. Moreover, a remaining state of charge (SoC) of 20% at the end of each mission is considered for safety reasons and to prolong battery life. Finally, the battery end-of-life threshold is set at 80% of its initial capacity.

Table 2.

Electric powertrain system characteristics for the different entry into service dates

The specific power of electric motors and power electronics is selected in relevance to the work of Nasoulis et al. [Reference Nasoulis, Gkoutzamanis and Kalfas11] and Jansen et al. [Reference Jansen, Bowman, Jankovsky, Dyson and Felder19], respectively, and is adjusted for the different EIS dates. The motors specific power examined in the timescale of 2027–2040 ranges from 6 to 16 kW/kg, considering that today’s electric motors have a specific power of 5 kW/kgFootnote ¹ (calculated for the maximum continuous power), whereas research suggests that this metric can reach up to 16 kW/kg for a partially superconducting wound field synchronous motor [Reference Jansen, Bowman, Jankovsky, Dyson and Felder19, Reference Scheidler, Tallerico, Miller and Torres20]. In order to achieve motors with high specific power, research indicates that a higher number of pole counts, moderate shear stress and higher rotational speed are some of the advancements required, as specific power has a declining trend with power increase [Reference Zhang and Haran21–Reference Van Der Geest, Polinder, Ferreira and Christmann23]. A more detailed analysis and review of the state-of-the-art, TRL and scalability challenges on electric propulsion components can be found in [Reference Gkoutzamanis, Kavvalos, Srinivas, Mavroudi, Korbetis, Kyprianidis and Kalfas9, Reference Nasoulis, Gkoutzamanis and Kalfas11, 24]. Finally, the gas turbine used in the reference aircraft shows similar efficiency and fuel consumption to PT6A-67D turboprop engine [25], whereas the gas turbines used in the hybrid-electric configurations are based on the reference, assuming a specific fuel consumption (SFC) reduction ranging from 8% to 27% for the various EIS dates. The power management strategy and the hybrid-electric operational modes presented in Fig. 3 are maintained for the examined EIS dates, to focus on the impact of technological improvements on hybrid aircraft design.

2.3 Aircraft sizing

For aircraft sizing, an in-house tool is used, developed to size novel aircraft designs, using unconventional propulsion units, such as hybrid-electric propulsion [Reference Nasoulis, Gkoutzamanis and Kalfas11]. This model follows a component build-up method to apply a weight-based aircraft sizing approach, that captures the impact of aircraft dimensions on the maximum take-off mass, and vice-versa. This implicit connection allows capturing the impact of the additional propulsive components mass, i.e. electric motors, batteries, etc., on aircraft mass and dimensions, allowing for a clean sheet design, rather than a retrofit approach. Then, the actual block fuel requirement for the mission is calculated, using a modified Breguet equation that integrates the electric propulsion system through the definition of the power-level degree of hybridisation (DoH) for each mission phase. The power-level DoH is defined as the ratio of the electric power (EL) to the summation of the thermal power (TH) and electric power (EL), at any given mission phase, as shown in Equation (1). The aircraft’s energy-level DoH is calculated as the ratio of battery energy (Bat) to the total onboard energy (fuel and batteries). Finally, the aircraft sizing tool is linked to the DOC and LCAevaluation tools.

(1) \begin{align}{H_P} = \frac{{{P_{EL,i}}}}{{{P_{EL,i}} + {P_{TH,i}}}},{H_E} = \frac{{{E_{Bat}}}}{{{E_{Fuel}} + {E_{Bat}}}}\end{align}

2.4 Direct operating cost

The present method for the calculation of aircraft DOC follows the guidelines set by the work of Hoelzen et al. [Reference Hoelzen, Liu, Bensmann, Winnefeld, Elham, Friedrichs and Hanke-Rauschenbach26] where the DOC and capital costs for a 70-passenger hybrid-electric regional aircraft are calculated, and it is modified accordingly to fit smaller aircraft class like the commuter. The modifications mainly focus on the tuning of the equations’ constants, to scale the formulae for smaller aircraft. The present model calculates the DOC requiring data, such as aircraft sizing parameters, and energy and aircraft manufacturing pricing, as shown in Fig. 4. The aircraft sizing parameters are classified into airframe parameters, namely maximum take-off, operating empty mass, wing span, etc., propulsion parameters, like maximum power output and propulsion system mass, etc., and mission parameters, such as block fuel, batteries energy, mission range, etc. The DOC calculation includes key cost indicators, such as energy cost per flight (both fuel and batteries), maintenance expenses, operational fees (crew, airport, and traffic control), as well as capital costs.

Figure 4.

Direct operating costs and capital costs computational pipeline overview.

The total DOC is a function of energy, maintenance, crew and fees costs, whereas the annual capital cost is calculated separately. That is because the annual capital cost is higher during the first years of the depreciation period, whereas all other costs remain constant through that period. The capital costs include the aircraft and onboard batteries costs, and for the aircraft capital costs, the gas turbine, electric motors, motor controllers and airframe costs are considered. The gas turbine cost is calculated to be 543 €/kW for a turboprop engine similar to PT6A-67DFootnote ^{2 ,} Footnote ³, therefore a range of 500-600 €/kW is selected for the analysis. For the airframe cost, the Beechcraft 1900D is used as a reference, considering its price when sold new in 1991 [Reference Aarons27], which is then realised to include USD to euro exchange rates, inflation and increased material costs, due to the integration of composite materials in the structure, especially for the years 2030 and 2040. As a result, a range between 1,400 and 1,600 €/kg is selected in the present analysis. The electric motor and motor controller costs are derived from Hoelzen [Reference Hoelzen, Liu, Bensmann, Winnefeld, Elham, Friedrichs and Hanke-Rauschenbach26], and a range of 120–180 €/kW for the motors and 50–100 €/kW for the inverters is selected herein respectively. On the other hand, for the batteries’ capital costs, three battery packs are considered per aircraft, with a lifetime of 1,500 cycles per pack. The battery cost per kWh is selected to be in the range of 130–170 €/kWhFootnote ⁴, taking into account that the batteries price may rise in the future, in case of increased demand. The depreciation cost of the initial investment is calculated using the annuity rate, which requires depreciation rate, interest rate and residual value factors as inputs. These values were also selected to be in harmony with the work of Hoelzen et al. [Reference Hoelzen, Liu, Bensmann, Winnefeld, Elham, Friedrichs and Hanke-Rauschenbach26] A different depreciation period is selected for batteries; therefore, a second annuity rate is calculated, and the batteries’ depreciation rate is considered a function of their lifetime.

Energy costs can fluctuate during the depreciation period, according to the energy market price, both for fuel and electricity production; therefore, a sensitivity analysis for fuel and electricity prices fluctuation will be presented in the results section. Maintenance costs include airframe materials, propulsive system, technology and labor costs. Airframe materials are related to the aircraft’s block time and a fixed repair cost per flight [28]. The labor cost is assumed constant for all examined configurations and engine maintenance costs are expressed using the maximum power requirement and ${V_1}$ velocity at take-off condition, as it is the most critical engine loading condition. Finally, the technology cost is considered a function of the aircraft’s basic dimensions, i.e. wingspan and fuselage length, and it is compensated by the airlines for fleet maintenance. The annual fees costs include air traffic management and airport costs and can be categorised into ground operation, navigation and landing fees.

2.5 Environmental impact assessment

In this section, the methodology for adapting the classical LCA to account for novel propulsion systems, such as hybrid-electric propulsion, is discussed as means to answer the research questions posed in the Introduction section. The model is based on the work of Johanning et al. [Reference Johanning30] and is adapted accordingly to evaluate the environmental impact of hybrid-electric aircraft that is still in the conceptual design phase.

The approach is based on ISO 14040:2006, where the inventory must be described and then its impact can be calculated [31]. The inventory analysis assessment is the first step and includes all aircraft life phases, namely the aircraft design and development, production, operation of aircraft, and finally the end-of-life. However, material information, engine emissions, information about different process types, etc., are difficult to be known a priori. Commercial and non-open access databases exist, such as Ecoinvent 3 [Reference Wernet, Bauer, Steubing, Reinhard, Moreno-Ruiz and Weidema32], that can be used to assess the impact of the above; however, for the scope of this study, a more conservative approach is followed. The level of detail for each life phase is limited to the extent of databases found in the open literature, for aircraft production and operation [Reference Johanning30].

(2) \begin{align}{x_{PKM,C{O_{2,cruise}}}} = {m_{fuel}} \cdot IO{F_{C{O_2}}} \cdot PK{M_f} \cdot Pal{o_{pax}}\end{align}

An example of the inventory analysis can be found in Equation (2), concerning the evaluation of the $C{O_2}$ emission caused by the fuel consumption during the cruise phase. In Equation (2) the ${x_{PKM,C{O_{2,cruise}}}}$ $[k{g_{C{O_2}}}/PKM]$ is the emissions of $C{O_2}$ per passenger per kilometer (PKM), ${m_{fuel}}$ $[k{g_{fuel}}]$ is the consumed fuel during the cruise flight phase, $PK{M_f}$ $[PKM/flight]$ is the number of kilometer-passenger per flight, $Pal{o_{pax}}$ $[k{g_{passenger}}/k{g_{payload}}]$ is the percentage of passenger mass on total payload mass and $IO{F_{C{O_2}}}$ $[k{g_{C{O_2}}}/k{g_{fuel}}]$ is the emission conversion factor. The latter can be calculated with a higher fidelity model or taken from literature through lower fidelity studies, considering however parameters such as the type of aircraft, type of engine, type of fuel, etc. Similarly, all other phases and emissions can be calculated [Reference Johanning30]. As it can be understood, the inventory analysis is mainly based on statistics that define the emission conversion factor. For that reason, its assessment can be greatly improved if information coming directly from manufacturers and operators can be included in the computational pipeline.

The next step is to translate the total gaseous emissions and material consumption into more comprenhensive indicators; therefore, the ReCiPe method is employed [Reference Huijbregts, Steinmann and Elshout33]. This method exploits characterisation factors that are able to indicate the environmental impact per unit of the stressor (e.g. per released emission or kg of extracted resource). There are two different categories of impact–the midpoint and the endpoint. The former refers to climate change, ozone depletion, toxicity and more. For example, it calculates the impact of 1kg of emitted $C{O_2}$ on climate change. The latter has three categories, namely human health, ecosystem quality and resource scarcity, that can be combined to form a single final score with respect to the input. The selected non-dimensional parameter selected to define the final score is the unit of impact per passenger per kilometer.

The typical equation for the assessment of the value of the impact of a certain emission to the midpoint category, $M{d_Y}$ , is presented in Equation (3), where $X$ indicates an emission, ${x_{PK{M_X}}}$ $[{g_X}/PKM]$ the quantity of emission $X$ in grams per passenger per kilometer, and $CFmidpoin{t_{Y,X}}$ the coefficient that indicates the impact of emission $X$ on the midepoind category $Y$ .

(3) \begin{align}M{d_Y} = \sum\limits_X ({x_{PK{M_X}}} \cdot CFmid\;poin{t_{Y,X}})\end{align}

In the ReCiPe method, depending on the importance that is given to each of the midpoint categories, there are three different ranks that can be utilised: (i) the Individualist, which is short-term and assumes that technology can avoid many future problems; (ii) the Hierarchist, that is considered a consensus model; and (iii) the Egalitarian, which is a long-term model based on preventive principle thinking. For this work, the Hierarchist version is used, a balance between the Individualist and Egalitarian methods, including global normalisation and average weighting set. Depending on the rank, the $CFmidpoin{t_X}$ of each midpoint category is changed accordinglyFootnote ⁵ . Specifically for climate change, as Johanning is proposing, the Schwartz, E. and Kroo approach is used for assessing the impact of the cruise emissions depending also on the altitude [Reference Schwartz and Kroo34]. Note that in the work of Johanning (2016) the coefficients used in order to perform the assessment were based on a ReCiPe version of 2008 [Reference Goedkoop, Heijungs, Huijbregts, De Schryver, Struijs and Van Zelm35]. In this work, the coefficients were updated with the newer available version of 2016 [Reference Huijbregts, Steinmann and Elshout33].

Once all midpoint categories are evaluated, the three endpoints can be calculated. The equation to advance from midpoint to endpoint categories is very similar as can be seen in Equation (4), where $CFendpoin{t_{E,Y}}$ is the coefficient to transform the value of the midpoint category $Y$ , into an endpoint division $E$ , that can be human health, ecosystem quality and resource scarcity. In addition, $Nor{m_Y}$ represents a normalisation factor that is used to normalise the results of each midpoint category, $Y$ , with respect to an endpoint category, $E$ . This is used since a midpoint class can affect more than one endpoint tier.

(4) \begin{align}E{d_E} = \sum\limits_Y\!\left(M{d_Y} \cdot \frac{{CFendpoin{t_{E,Y}}}}{{Nor{m_Y}}}\right)\end{align}

In order to compute the final single score, that can give the overall environmental impact of the aircraft, Equation (5) is used. A weighting factor can be used, ${W_E}$ , in order to define the importance of each of the three endpoint categories to the study. This parameter also depends on the rank introduced earlier.

(5) \begin{align}SS = \sum\limits_E (E{d_E} \cdot {W_E})\end{align}

The work of Johanning et al. considers conventional or fully electric engines only. Thus, modifications have been made to explore hybrid-electric aircraft architectures. Towards this direction, the model is expanded to include the possibility to have both electricity and kerosene consumption during a flight phase, using the energy-level DoH as described in the Aircraft Sizing section. The modification is presented in Fig. 5, where the original streamlines are presented, i.e. for the conventional and fully electric configurations. In addition, the hybrid configuration is presented that combines data from both streamlines and merges the two, upstream of the midpoint categories block. Finally, for the hybrid configuration, both electric and conventional propulsion systems must be considered, including the batteries.

Figure 5.

Flowchart of life cycle assessment evaluation.

For that reason, some of the equations computing the overall inventory emissions have been modified to include hybrid configurations. For instance, the equation computing the natural gas consumption has been modified as shown in Fig. 6, where ${X_{PKM,naturalgas}}$ is the natural gas consumed, $Ba{t_{prod}},E{l_{flight}}$ , $Ke{r_{prod}}$ and $Ke{r_{flight}}$ are the battery production, electricity consumption, kerosene production and kerosene consumption, respectively. Prior to these modifications, the tool allowed the selection of one component at a time, making it impossible to study hybrid configurations.

Figure 6.

Example of life cycle assessment updated equations.

Having explained the modules of the proposed approach in the previous sections, the overall computational pipeline is presented in Fig. 7.

Figure 7.

Environmental and techno-economic evaluation overall flowchart.