2.1 Approach

Developing alternative propulsion systems like turbo-electric and hybrid-electric engines implies an increase of complexity compared to the conventional GT cycle. The increase in number of components results in an enhanced total weight and thrust requirements. The complex system of parameters influencing the aircraft’s mission performance is thereby brought out of balance and is required to be redesigned from scratch. All engine components must be matched with one another and the aircraft in total, so that the entire system can operate efficiently.

The design of the propulsion systems is based on the individual components technology. Table 1 presents the assumptions for power or energy densities and efficiencies.

Table 1 Technology assumptions at present state (2017) for design of alternative propulsion systems

The assumptions correspond to state-of-the-art technology levels and are based on a literature review and expert surveys. According to van der Geerst^{(Reference van der Geest6)} the power density depends on the surface speed or rotational frequency of the electric motor. In each concept, the power density will be adapted to the appropriate level of rotational speed. The electric motor’s power densities in this study are related to Martini^{(Reference Martini4)} and Compact Dynamics GmbH⁽⁷⁾.

The battery is chosen as a high energy cell. Therefore, its power density is limited to approximately 0.7 kW/kg^{(Reference Wagner8)}. The power density of power electronics is based on the manufacturer’s data. The presented magnitude is confirmed by Erickson et al.^{(Reference Erickson, Maksimovic and Afridi9)} and Casady^{(Reference Casady10)}. Data for the DC-DC converter is taken from Matlok^{(Reference Matlok11)}.

2.2 Case of application

This study uses the ATR 72 aircraft as a reference case to determine the potential for electrification. Hence, the propulsion system configurations are designed to meet the requirements for this regional aircraft, which is propelled by two PW127 turboprop engines. Table 2 presents a selection of characteristic aircraft properties for the ATR 72 aircraft model. Typical flight missions for regional aircrafts are around 200 - 300 nautic miles (370 – 556 km). This is much shorter than the maximum range (r), see Table 2.

Table 2 Characteristic properties of ATR 72 aircraft⁽¹⁴⁾

2.3 Reference case

The different configurations are assessed in comparison to the reference case, which is described in this section. According to the PW127 engine, a DLR127 model is set up within the GTlab performance synthesis tool. This engine operates with three spools, two compressors, a combustion chamber, three turbines, and a secondary air system. Thrust is generated by a propeller. The core engine is used for power generation providing just a small fraction of thrust. The propeller is linked through a gearbox to the power turbine. The layout and station numbering of the model is presented in Fig. 1.

Figure 1. Schematic of the thermodynamic model of DLR127 engine.

For the modeling of the thermodynamic cycle, data from literature^{(Reference Daly and Gunston12,Reference Hosking, Kenny, McCormick, Moustapha, Sampath and Smailys13)} , data sheets^(14,15) and the emissions database⁽¹⁶⁾ are taken into account. Data could be found for power, rotational frequencies, temperature limits, fuel flow rates, and pressure ratios. For engine modeling, the cruise operating point is chosen as the design point. A selection of performance parameters of the design point is shown in Table 3.

Table 3 Parameters of engine components in design point

For all off-design points, the compliance of calculated values to temperature limits and plausibility of all other engine data is verified.

3.1 Turbo-electric

In the turbo-electric concept, mechanical energy from the GT is converted into electric energy by the generator. Since larger GTs generally work more efficiently than smaller ones, a single GT for both propelling units is conceivable. The disadvantage, in this case, is the loss of redundancy for the case of one engine inoperative (OEI). Relying purely on electric propulsion as a backup system would require large batteries. With current technology, the weight of such a battery system would by far outweigh the efficiency gain of a single, larger gas turbine. Thus, it is not regarded here. In this study, two variations of turbo-electric propulsion systems are investigated.

Turbo-electric power controlled (tu-el pc)

Figure 2(a) shows the principal setup. Between generator and motor, two power electronics are arranged. The first one converts the voltage to a high DC level (1000V), thereby reducing conduction losses in the wires. The second one creates the voltage and frequency of AC (580V), which is necessary to drive the motor. The component’s properties are chosen in such a way that a gearbox is not necessary for this concept. Applying the component efficiency values from Table 3, the efficiency of the electric system yields

(1) $$\begin{equation} {\eta }_{{\rm el-sys}}={\eta }_{{\rm mot}}*{\eta }_{PE}*{\eta }_{{\rm cond}}*{\eta }_{{\rm PE}}*{\eta }_{{\rm gen}}=0.8562 \label{eqn1} \end{equation}$$

Due to the number of additional components, the resulting engine is quite heavy. It is more than 1000 kg heavier compared to the reference engine.

Turbo-electric direct (tu-el dir)

Without power electronics, the generated electricity is conducted without any conversion to the motor (see Fig. 2(b)). That causes a direct link of both rotational speeds. Because of the different rotational speeds of propeller and power turbine, there has to be an adaption. This can be reached by a gearbox, a frequency converter, different numbers of poles in motor and generator or any combination of these aspects. Due to skin- and proximity-effects direct conduction with the power turbines frequency is not possible without further changes to the electrical system, e.g. voltage. In order to limit complexity, a gearbox is chosen to be placed in front of the generator. For this purpose, the gearbox from the reference engine is assumed to be suitable. That setup results in an electric efficiency of

(2) $$\begin{equation} {\eta }_{el-sys}={\eta }_{mot}*{\eta }_{cond}*{\eta }_{gen} =0.8994 \label{eqn2} \end{equation}$$

Although at first glance this configuration seems to be simpler, it finally causes a similar additional weight like the tu-el dir concept. Reasons are the heavy wires as there is no conversion to higher voltages and the necessity of the gearbox.

3.2 Hybrid-electric

The hybrid-electric concepts use energy from two carriers. Besides the GT, a battery supplies additional energy. During the different flight phases, the energy carriers can be balanced in a way that the total system can optimally perform. For the configuration of the battery, energy density and power density are the major limitations. The electric support is contributed during takeoff and climb. That leads to a smaller GT, as the operating points relevant for sizing of the GT are electrically supported. Two different types of hybrid-electric propulsion systems are investigated.

Hybrid-electric parallel (hy-el pa)

In the hybrid-electric parallel setup, the electric motor is assembled on the power shaft of the reference engine. The setup is presented in Fig. 2(c). The components are located close to each other. Therefore, the available design space and temperature limitations might be critical constraints for that arrangement. The electric motor should not get any extra heat from the GT unit. Thus, it seems reasonable to place the motor in front of the compressors or at the gearbox. The electric efficiency for the hy-el pa concept is

(3) $$\begin{equation} {\eta }_{el-sys}={\eta }_{mot}*{\eta }_{PE}*{\eta }_{cond}*{\eta }_{bat}=0.9152 \label{eqn3} \end{equation}$$

The smaller scaled GT causes the engine to operate with higher temperatures in cruise conditions. This behaviour is also reproduced by the thermodynamic model. Generating energy during descent is also an option to get enough energy reserves for a potential go around scenario. In the flight-idle state, the gas turbine must maintain a certain rotational speed to ensure quick acceleration in case of an emergency. With the fan or propeller windmilling, this residual power is available for generating electricity. The battery is designed for the energy, which is required for takeoff and climb. The scaling of the GT is realised by the selection of the engine design point. It is chosen with lower thrust and power, resulting in a smaller mass flow rate. That corresponds to the smaller scaled engine. Parameters like component efficiencies, pressure ratios or fuel to air ratio are kept unchanged to obtain comparability. However, the power coefficient of the propeller has to be adapted, because of the reduced power requirement at the design point. This must be considered during the propeller design of propulsion systems in that functional context. The amount of additional power from the battery can be chosen freely. For the case of application which is treated here, it is limited by the MTOW of the aircraft. In this study, two variants of a parallel hybrid-electric concept are proposed. In both versions, the power density is close to today’s power density limit of 0.7 kW/kg.

Hy-el pa 10 supplies 10% of the takeoff power electrically from the battery. This power stays constant during takeoff and climb. The relatively small share of electric power leads to a moderate battery weight. So the additional weight per propulsion unit is around 300 kg (see Table 4).

Table 4 Mass effects of alternative propulsion systems per propulsion unit

Hy-el pa 40 has a larger battery and supports 40% of the takeoff power. The support also stops after the climb segment. The battery becomes so heavy, that the limitations of MTOW are reached with this configuration.

In a hybrid-electric serial propulsion system, the GT and the electric propulsion are decoupled. Figure 2(d) shows that the arrangement is similar to the tu-el pc concept. In the hy-el se concept, the generator does not provide all the power. During takeoff and climb the power requirement is partly contributed by the battery. Due to MTOW restrictions, this cannot exceed about 10% of the takeoff power. The electric system efficiency during pure generator operation is the same as tu-el pc concept, ${\eta }_{el-sys}=0.8562.$ If additional power is provided by the battery, a mixed value will result as the total electric efficiency. This value will be slightly higher and depend upon the power ratio of the battery and the gas turbine. Since the system consists of many components, it is complex and heavy resulting in major disadvantages comparable to turbo-electric concepts.

3.3 Pure electric (p-e)

The principal setup of the pure electric concept is quite simple. It is presented in Fig. 2(e). The battery supplies electric power, which drives the motor. In between, power electronics are necessary. Hence the electric system efficiency is analog to hy-el pa concept

(4) $$\begin{equation} {\eta }_{el-sys}={\eta }_{mot}*{\eta }_{PE}*{\eta }_{cond}*{\eta }_{bat}=0.9152 \label{eqn4} \end{equation}$$

Regarding the component masses, the electric motor including all necessary peripheries is heavier than the reference engine (see Table 4). The battery weight can be seen as the limiting parameter for this configuration, which is not feasible assuming today’s technology levels. To enable an all-electric flight for the considered flight length, the energy density of the battery has to be approximately 800 Wh/kg. That is four times more than the assumed specification for today’s batteries. Therein, safety reserves are still not considered yet. At the same time, the battery’s power density has to be improved by at least the same factor.

3.4 Mass breakdown results

For all concepts, the masses of propulsion units are calculated which are shown in Table 4 (effect per propulsion unit). The component masses are based on technology assumptions shown in Table 1. Fuel savings are still roughly approximated here. Results from the flight mission calculations could be used in the next design cycle.

3.5 Performance results

As the performance characteristics of the different propulsion systems have a significant impact on the mission fuel consumption, the different behaviours are illustrated in Fig. 3. The curves do not include energy taken from batteries in hybrid-electric propulsion systems.

Figure 3. Thrust specific fuel consumption versus thrust of GT in different propulsion systems; dotted lines symbolise that one temperature limit is exceeded (T4 or T49).

The plots confirm that turbo-electric configurations work inefficiently because of multiple converting mechanisms. The similarity between tu-el pc and hy-el se can easily be seen. Tu-el dir causes fewer losses, as there are no power electronics. Hy-el pa configurations are mostly similar to the reference engine’s behaviour. Due to the smaller scaled GT, the TSFC-optimum stays on the same level but moves to a smaller thrust. The dotted lines indicate the region where GT temperature exceeds the limits. The closer the operating point approaches to the temperature limit, the more efficiently the process works. The temperature level in the scaled GTs is higher than in the reference. That indicates the necessity of an adaption of GT cycle for that application, e.g. optimization of the cooling system.

4.1 Flight mission simulation

To determine the fuel consumption on aircraft level the engine data is coupled with an aircraft model via GTlab Flight. Therein, the aircraft is assumed as a point mass model and defined by a characteristic mass. Drag polars are based on BADA data. Engine performance data are calculated for the entire flight envelope and different thrust settings are stored within a performance data table. During flight missions, altitude, thrust, lift, drag and weight are calculated for each step. The corresponding fuel flow is interpolated from the engine performance data table. For the flight mission simulations in this study, a constant trajectory is used for all alternative propulsion systems. The flight profile corresponds to the ATR72 data in the Eurocontrol Aircraft Reference Database. This implies that the assumed flight velocities and rates of climb are kept constant and the required power is calculated for each step. This is a simplification, which is assumed to be not significant in the comparison of relative results for the first instance. The aircraft itself is kept untouched. As there are significant differences in the weight of the propulsion systems, it is necessary to reduce the payload in order to enable constant transport capacities through all concepts for the compared flight mission. The resulting payload is 4,974 kg in comparison to the maximum payload of the ATR 72 which is 7,000 kg. Thus, the reference and hy-el pa 10 configurations transport less payload than practically feasible.

Despite significant differences in the total weights of alternative propulsion units, all concepts are to be integrated into the ATR 72. Due to the same payload, the takeoff weights of the concepts changed according to the weight modifications of the propulsion systems and the required mission fuel burn. All this has to be respected in the design process leading to an iterative procedure. Based on the data presented in Table 4, the pure electric propulsion system clearly exceeds the MTOW restrictions of ATR 72. The resulting takeoff weights for the residual concepts under consideration are 21,500 kg for hy-el se, 21,100 kg for hy-el pa 40, 21,010 kg for tu-el dir, 20,960 kg for tu-el pc, 19,160 kg for hy-el pa 10 and 18,500 kg for the reference.

4.2 Fuel consumption

Figure 4 presents the resulting fuel consumption on the flight missions with all propulsion systems. As expected, tu-el pc, tu-el dir, and hy-el se concepts cause increased fuel consumption on the flight mission. The difference between power controlled and direct turbo-electric variants can be seen. Additional conversion implies increased losses and fuel consumption. Hy-el pa configurations show the potential to save fuel compared to the reference engine. During the climb the GT is supported by the electric motor. Despite the additional weight of the battery, this saving can be kept over the whole mission. The partially parallel curves in the cruise segment point out that behaviour. The saving is 5.21% for hy-el pa 10 and 21.6% for hy-el pa 40.

Figure 4. Fuel consumption with alternative propulsion systems and reference engine at 300 nm mission; electric energy from batteries is not presented.

To analyze the effect of smaller-scaled GTs, Fig. 5 shows the fuel mass flow rates of the alternative propulsion systems during the compared mission. In the climb segment, the mass flow rates of hy-el pa configurations are clearly reduced. In the case of the hy-el pa 40, it is even lower than the level of cruise segment. Here, a design with a constant level of fuel mass flow or load in the climb and cruise segments seems to be reasonable.

Figure 5. Fuel mass flow rate of different propulsion systems during flight mission.

During the cruise, the hy-el pa concepts and the reference configuration are on the same level. For this it is crucial that hy-el pa has a smaller scaled GT. It works at an operating point with lower TSFC which is compensated by higher thrust requirements because of the battery weight.

The lower fuel mass flow rates during descent can be explained by enhanced efficiency at lower load.

4.3 Energy consumption

The electric energy obtained from the batteries has not been considered so far. To discuss the total amount of energy necessary to cover the flight mission, the sum of electric energy and fuel-bound energy is shown in Fig. 6.

A fuel heating value for jet fuel of 43.25 MJ/kg⁽¹⁷⁾ is assumed. The shares for fuel and electric energy are each multiplied by a primary energy factor, to consider the efforts to generate and transport energy. The specific energy input for electricity is 2.13⁽¹⁸⁾, which is more than the primary energy factor of kerosene with 1.1^{(Reference Novelli19)}. Thus, the share of electricity at the total energy demand increases. The optimisation of primary energy demands with the help of hy-el pa propulsion systems is correspondingly lower. The main influence of the share of electricity is the energy mix in the production of electricity.

Figure 6. Primary energy inputs for compared mission covered by alternative propulsion systems.

4.4 Emissions

For a holistic view, the behaviour of emissions is shown besides fuel consumption and energy demands. The direct CO ${}_{2}$ emissions from the combustion of Jet-A1 are 3.156 kg/kg^{(Reference Sutkus20)}. The equivalent CO ${}_{2}$ emissions are 0.6055 kg/kg^{(Reference Wolters, Becker and Schaefer21)}. Based on Germany’s average energy mix in 2015, electricity causes 534 g/kWh of CO ${}_{2}$ emissions^{(Reference Icha and Kuhs22)}. Figure 7 presents the calculated emissions split into direct and indirect. The potential for saving emissions is in the field of direct emissions. These are proportional to the fuel consumption described in section 4.2. In terms of indirect emissions, there is just a negligible change for hy-el pa 10 and for hy-el pa 40, there is a slight increase. In general, indirect emissions are advantageous as they arise mostly at power plants which are stationary. This offers possibilities of filtering and treating the emitted gases and particles. Generally, the share of life cycle equivalent emissions caused by electricity can be reduced by an increased amount of regenerative energy in the electricity mix.

Figure 7. Direct and equivalent emissions for the compared flight mission covered by alternative propulsion systems.