\label{sec:trade_power} This section elaborates on two subsequent trade-offs regarding the power subsystem. Section \ref{subsec:trade_power} discusses the initial trade-off resulting in either a fuel cell, redox flow battery (RFB) or battery subsystem. Section \ref{subsec:trade_battery} will go more in-depth and a trade-off of different solutions within the previously chosen subsystem will be made.
\label{subsec:trade_power}
The first trade-off that needs to be performed is between the fuel cell, battery and RFB. First, the chosen main criteria and sub-criteria will be discussed including the weights per criteria.
The trade-off of the different power sources is based on three main criteria, being performance, complexity/cost and sustainability. In order to clearly define which aspects belong to each criteria another set of sub-criteria was defined. Both the criteria as well as the sub-criteria were also assigned weights to indicate their relative importance. All (sub-)criteria and the corresponding weights are summarized in Table \ref{tab:criteria_det_weights}.
The performance criterion is subdivided in the weight, power and volume. Due to the snowball effect an increase in weight of the power source results in a bigger increase in the take-off weight, which jeopardizes the maneuverability of E-SPARC. Therefore the power source weight is the main performance parameter. Secondly the source’s ability to deliver sufficiently high power to the electric motor(s) is weighted slightly more important than the volume of the subsystem. This is mainly based on the fact that E-SPARC mission only takes a maximum of 33 minutes, resulting in minor energy capacity compared to long-range missions.
The complexity/cost criterion is defined by its four sub-criteria, being designability, system complexity, readiness of technology and cost. Due to the steep requirement on total cost XXXX REF TO REQ XXXX the corresponding sub-criteria has the highest weight within the complexity/cost criteria. The readiness of technology is weighted similar to the system complexity. System complexity includes aspects as the number of moving parts as well as unpredictability of the system. Designability describes to what level the DSE group will be able to design the power source. Since the main focus of E-SPARC is its competitiveness with respect to the current aerobatic racers, the designability only has a minor influence on the trade-off.
The sustainability criterion for the trade-off of power source is divided in scarcity of resources, well-to-wheel efficiency, lifespan and recyclability. E-SPARC should prove the competitiveness of an eco-friendly aerobatic racer, therefore the well-to-wheel (or actually well-to-propeller) efficiency of the power source is of paramount importance, as can be observed in Table \ref{tab:criteria_det_weights}. The lifespan of the power source also scores high, as it has a big influence on the ecological footprint of E-SPARC. Since E-SPARC is mainly designed as a prototype to prove the concept of an ecofriendly aerobatic racer and not for mass production (yet), scarcity of resources and recyclability have lower weights. Additionally the sub-criteria safety is added for completeness, which has a low weight because E-SPARC is primarily designed for racing.
Finally the weights of the three main criteria are all set to \(\frac{1}{3}\). In addition a sensitivity analysis with respect to the performance criteria weight is performed at the end of this subsection. For this analysis the complexity/cost and sustainability weights remain equal while the performance criteria weight is varied.
| Main Criteria | Weight | Sub-criteria | Weight |
|---|---|---|---|
| A1. Weight | 0.539 | ||
| A2. Power | 0.297 | ||
| A3. Volume | 0.164 | ||
| B1. Designability | 0.083 | ||
| B2. System Complexity | 0.249 | ||
| B3. Readiness of Technology | 0.284 | ||
| B4. Cost | 0.384 | ||
| C1. Scarcity of Resources | 0.095 | ||
| C2. Well-to-Wheel Efficiency | 0.550 | ||
| C3. Lifespan | 0.221 | ||
| C5. Recyclability | 0.088 | ||
| C4. Safety | 0.046 |
\label{tab:criteria_det_weights}
The trade-off was performed by qualitatively determining the relative performance of each subsystem for each single criterion. The results of this trade-off is given by Table \ref{tab:TOfinalresultpower}. Overall the battery scores highest but it is clear that the fuel cell has an obvious advantage in the performance criteria. The relative performance of the subsystems will be elaborated on in the following paragraphs.
The performance advantage of fuel cells can be explained by the high criteria weight of the sub-criterion weight. The fuel cells system and hydrogen altogether are considered to be moderately lighter than batteries, even though the fuel cells add a lot of mass [source about fuel cell mass] to the subsystem. The main reason is the much higher gravimetric energy density of hydrogen, equal to [<VALUE>][<Source here>]. Also, the energy density of current RFBs is low even when comparing it to Lithium-ion batteries \cite{FB1,FB3, FB2_5}. Comparing it to forecast battery specific energy, the difference becomes even larger \cite{marketanal:luxresearchinc,e_bat}. RFBs also deal - similar to fuel cells - with extra weight due to the need for pumps and other complexities. Given the current application of hydrogen and batteries in vehicles, it is assumed that both systems can deliver plenty of power which results in a higher emphasis on the weight and volume. The lesser known RFBs are also capable of delivering the required power of several hundreds of kilowatts \cite{FB1,FB2}. In terms of volume, hydrogen performs much worse than batteries. RFBs have an extremely low volumetric energy density of about \(20-35\,Wh/\)L \cite{FB3}.
Batteries perform better at the complexity/cost criterion, which is best explained by the relatively low complexity of the battery subsystem with very few components compared to RFBs and fuel cells \cite{FB1}. The criterion with the largest weight - namely cost - has no effect on the trade-off as it was assumed equal. Batteries and redox flow batteries have similar cost per kWh storage \cite{FB1, marketanal:luxresearchinc,e_bat}. Hydrogen itself is cheaper, but the required fuel cell is considered to add similar cost. The readiness of technology, another important criterion, is similar for fuel cells and batteries. They have both been applied in (H)EVs and should be ready for usage. Flow batteries are currently suited for storage of more than 500kWh - in mainly stationary applications - and not yet commercialised \cite{FB1,FB2, FB2_5}. Altogether, this means a dominance of the battery and fuel cells, with a clear complexity advantage for batteries.
The sustainability criterion was subdivided into five criterion with again a clear advantage for the battery. Important in this explanation is the well-to-wheel (or well-to-propeller) efficiency of the subsystems. This is where especially hydrogen scores low, mostly due to the processes from the source to the hydrogen tank. Important considerations here are the required compression of hydrogen for transportation, storage in stationary tanks and the storage in the FCEVs. Furthermore, hydrogen needs to be produced by electrolysis, which again requires energy. Concluded can be that hydrogen is basically an extra - although with high gravimetric energy density - storage step which requires a lot of energy from well-to-wheel. This can also be seen from the CO2 equivalent emissions of electricity-based vehicle fueling pathways \cite{FB8}. Furthermore, the efficiency of batteries is generally higher than the other subsystems \cite{FB3}. Another criterion that plays a major role is lifespan. For batteries, this ranges from about 500-2,000 cycles \cite{marketanal:panasonic18650, e_bat}. For RFBs, this is 5,000-14,000 at a DoD of 80%. Fuel cells also have a much larger lifespan than batteries, with up to 10,000 hours of run-time.
In terms of scarcity of resources, fuel cells perform well although most of the energy for hydrogen is currently from fossil fuels. Both RFBs and batteries perform worse as often used components such as Vanadium in RFBs and Lithium - as a cathode in batteries - have limited reserves. However, they ar still widely available at the moment \cite{FB6,FB7}. A rapidly growing demand will clearly result in shrinking reserves. This may not be problematic for the coming centuries.
| Options | Performance | Complexity/Cost | Sustainability | Total |
|---|---|---|---|---|
| Battery | 0.35 | 0.51 | 0.44 | 0.44 |
| Fuel cell | 0.47 | 0.31 | 0.30 | 0.36 |
| Flow battery | 0.18 | 0.18 | 0.26 | 0.20 |
\label{tab:TOfinalresultpower}
The results in Table \ref{tab:TOfinalresultpower} are based on the equal weights for the performance, complexity/cost and sustainability criteria. However, since the hydrogen fuel-cell scores significantly higher on the performance criterion compared to the battery, it is also interesting to find out what happens to the results if the performance criterion has more weight than the other main criteria. The results of this sensitivity analysis are plotted in Figure \ref{fig:sensitivity_power_tradeoff}. The extreme cases are a performance weight of 0.00, in which case both the other criteria have a weight of 0.50, and a performance weight of 1.00, in which case the other criteria are completely disregarded. The score of the battery option decreases for an increase of performance weight while the opposite holds for the fuel-cell. The fuel-cell scores better than the battery option for a performance criterion weight higher than 0.59 or complexity/cost and sustainability criterion weight smaller than 0.21. Since the complexity/cost and sustainability criterion weights should definitely be higher than 0.21 it can be concluded the performed trade-off is robust enough to accept the results from Table \ref{tab:TOfinalresultpower}.
In order to test the robustness of the trade-off some extreme cases could also be considered. For instance, would the battery still be the best option if the structural weight of E-SPARC will be very low compared to current reference aircraft due to advancements in composite technology? In that case the weight of the power source could increase to still sum up to an acceptable take-off weight. Since the battery actually scores worse than the fuel cell on the weight criterion, decreasing the importance of weight would only increase the score of the battery with respect to the fuel cell. Even if an acceptable MTOW is not a major issue, a large decrease of the structural weight - and thus a lower power requirement - is not necessarily advantageous for fuel cells. The absolute weight decrease of the hydrogen is close to negligible, whereas the absolute decrease of batteries is a lot larger. Another extreme case could be that the cost is chosen as a driven requirement because the costumer wants E-SPARC to be as cheap as possible. Since current knowledge indicates little difference in cost for the three options this would result in a draw (which means that a battery is still the ‘best’ option, together with the alternatives). It would also be possible to focus mainly on the sustainability aspect of the power source as the mission of E-SPARC is proving the concept of an ecofriendly aerobatic racer. In that case the battery would also clearly score the highest, mainly due to the high well-to-wheel efficiency. To conclude, the most applicable ‘extreme’ trade-off cases result in a win for the battery option or a draw, therefore confirming the results from Table \ref{tab:TOfinalresultpower}.
Subsection \ref{subsec:trade_battery} elaborates on the trade-off of the different battery technologies.
Scores of subsystems as a function of the performance weight
\label{fig:sensitivity_power_tradeoff}
\label{subsec:trade_battery}
Now that batteries are chosen as the most adequate subsystem, the type of batteries has to be determined next. Due to the large number of important specifications for battery cells, the criteria need to be chosen with care. Furthermore, the criteria will be - if possible - scored quantitatively, which further reduces the probability of a biased result. Table \ref{tab:bat_criteria} shows the determined criteria and assigned weights of each of these criteria. Four out of seven criteria can be determined quantitatively.
The weights assigned for most criteria were done quantitatively to increase objectiveness. An extensive literature study has been conducted to perform a proper weight determination. During the distribution of the weights for all criteria it was agreed that the readiness of the technology has the largest influence. The goal of the project is to have the first flight of the airplane in 2025. During the literature study it was noticed that the prediction of the performance is not consistent in literature. This induces a large risk as an overestimation of the specifications can result in an un-ready aircraft. The specific energy (or gravimetric energy density) of the battery was also deemed of significant importance since one the design goals is to have a light weight airplane for better performance. This is also closely related to the first criterion. The next criterion was the energy density, which is important since the fuselage only has limited volume. A low volume would also allow for more flexibility of the battery lay-out in the fuselage. The cycle life was also considered very important, but less than the first three criteria. However, the cycle life has an influence on sustainability and marketability (market analysis). The remaining criteria are cost, sustainability and complexity which are ranked in this order respectively. Cost was considered more important, since there is a limitation on the budget. Complexity was chosen to be ranked as the last criteria below sustainability. Due to the fact that it is preferred to have a more sustainable system over a simple system. The sustainability criterion considers the materials used in the batteries, simplified to solely the anode and cathode configuration.
An initial feasibility check was performed to reduce the number of battery technologies that had to be analysed. A large number of battery types - such as Lead-Acid, NiCd and NiMH - were left out due to their weak performance in terms of specific energy. Much weaker than the leading technology currently available on the market. The battery types that were considered were Lithium-ion, Lithium-Polymer, Lithium-Sulfur, Lithium-Air and Zinc-Air. Battery types that are currently being produced or of which the existence of prototypes is known are listed twice in table \ref{tab:bat_chosen_last}, once as the current state (2015) and once as a future prediction (2020). There exist many other types of batteries but they had a low specific energy or were not feasible at all for the next years.
Table \ref{tab:bat_chosen_last} does not include the current values (2015) of Lithium-Air and Zinc-Air, that is because these batteries are currently tested and predicted to be released to the market in the coming years. Furthermore, a forecast of Lithium-Polymer was not found, but is seemingly similar to that of Lithium-ion. Lithium-Polymer is a variant of Lithium-ion which could explain that.
Relationship (hyperbole) between battery gravimetric density and the MTOW.
\label{MTOW_bat}
| Criteria | Unit | Method of determination | Assigned weight |
|---|---|---|---|
| A. Specific Energy | [Wh/kg] | Quantitatively | 0.239 |
| B. Energy Density | [Wh/L] | Quantitatively | 0.156 |
| C. Cycle Life | [C-D cycles] | Quantitatively | 0.114 |
| D. Cost | [$/kWh] | Quantitatively | 0.091 |
| E. Sustainability | [-] | Qualitatively | 0.04 |
| F. Readiness of Technology | [-] | Qualitatively | 0.313 |
| G. Complexity (effect on subsystems etc) | [-] | Qualitatively | 0.0379 |
\label{tab:bat_criteria}
It is possible to discard the majority of the given technologies in Table \ref{tab:bat_chosen_last} based on the Class I analysis. Figure \ref{MTOW_bat} shows the relationship between the battery gravimetric density and the MTOW of the airplane. It was decided that the MTOW should be lower than approximately 700kg which means that the specific energy should be higher than ca. 340Wh/kg. The reason is that the relationship is hyperbolic and the MTOW increases rapidly when the gravimetric energy density is lower than this value. Following from this, another major consideration is risk mitigation. Due to the hyperbolic nature, a change in specific energy results in smaller MTOW changes when the battery technology turns out to perform differently. Furthermore, a MTOW of 700kg or lower is competitive compared to current aerobatic racing aircraft. With this information all the non-feasible options were discarded and removed for further research and the trade-off. The battery types that are feasible are the Lithium Sulfur, Lithium-Air and Zinc-Air. This overview is given in Table \ref{tab:bat_criteria}.
| Battery technology | Year | Wh/kg | Feasible |
|---|---|---|---|
| Li-ion | 2015 | 233 \cite{marketanal:teslabatteryreport}, 243 \cite{marketanal:panasonic18650} | No |
| Li-ion | 2020 | 300\cite{marketanal:luxresearchinc} | No |
| Li-Po | 2015 | 100-158 \cite{lipoly}, 200 \cite{clark2007evaluation} | No |
| Li-Po | 2020 | No | |
| Li-S | 2015 | 300 \cite{OXIS_ULTRA_LIGHT} | No |
| Li-S | 2020 | 500 (2019), 450 (2018) \cite{OXIS_ULTRA_LIGHT}, 400 \cite{litsulf12} | Yes |
| Li-Air | 2020 | 200-800 \cite{marketanal:luxresearchinc},362 \cite{stanfordliair}, 300 \cite{litair123}, 700 \cite{polyplus1} | Yes |
| Zn-Air | 2020 | 320-650 \cite{marketanal:luxresearchinc}, 370 \cite{lithiumair}, 400 \cite{EnZinc} | Yes |
\label{tab:bat_chosen_last}
After discarding a large number of options in the initial feasibility check, the last three feasible options can be further analysed with data from research for the other six criteria. Table \ref{tab:lit_study_bat} summarises the finding of the extensive literature study.
The quantitative criteria specific energy, volumetric energy density and cycle life (C-D cycles) provide several values for the different battery technologies. Two values were also found for the cost of Zinc-Air. For Li-S the 2019 values provided by Oxis Energy - a pioneer in Lithium Sulfur battery technology - were decisive. The 2018 and current (2015) values were added for comparison. Products or even cell prototypes of Lithium-Air or Zinc-Air did not seem available and most results are from academic or institutional research. For the trade-off the cost of Li-Air was assumed to be slightly higher than the Zn-Air batteries due to the higher cost of lithium. Actual values could not be found. In all the other cases the average of the sources was taken. The quantitative values used for the trade-off are summarised in Table \ref{tab:final_val_TO_bat}. The results of the trade-off in the same table show a clear advantage for Li-S and the worst performance for Li-Air.
The explanation of this result lies in the determined weights. Li-S performs equal or better than the Metal-air technologies at most criteria that have been assigned a high weight. Zinc-Air performs extremely well at the quantitative criteria. The impact of readiness of technology is clearly visible as the result is far less optimistic about Zn-Air batteries. The issues that both Metal-Air batteries have are rechargeability and since they are air as a cathode they are influenced by properties of the atmosphere. One of the challenges is moisture that affects the performance of Metal-Air technology, in particular Li-Air \cite{metal_zinc}. Sustainability is where Zinc-Air performs best, as the Zinc reserves - which are mostly found in the earth’s crest - are large. Lithium reserves will also be significant for the coming century. Sulfur quantities are abundant but clearly perform worse than the air cathode of the metal-air batteries \cite{future_sulfur, future_lithium}.
Metal-air battery technology is dependent on an air inflow as oxygen forms the cathode of these batteries \cite{lithiumair}. This increases complexity as sufficient oxygen levels and a large exposed surface of the cells is required. The impact of the designability of these cells can therefore be large, however, cells are likely to be bought off-the-shelf. The main complexity therefore lies in maintaining the cathode during flight. Furthermore, the attraction of \(O_2\) - which chemically bonds with lithium in the process - increases the weight of the battery pack during flight as \(LiO_2\) is formed during discharge. This exact weight change and the effect on the centre of gravity position would also have to be investigated. To conclude: Li-S is the most adequate option to be used in the battery pack. The next paragraph will briefly analyse the sensitivity of these trade-off results. Weights will be shifted in favour of the Zn-Air technology to see how acceptable the current result is.
| Battery Type | Wh/kg | Wh/L | C-D Cycles | $/kWh | Sustainability | ROT | Complexity |
|---|---|---|---|---|---|---|---|
| Li-S 2020 | 500 (2019), 450 (2018) \cite{OXIS_ULTRA_LIGHT}, 400 \cite{litsulf12} | 550 \cite{OXIS_ULTRA_LIGHT}, 425 \cite{litsulf12} | 1000 (2018), 1500 (2019), 100 (2015) \cite{OXIS_ULTRA_LIGHT} | 250 \cite{OXIS_ULTRA_LIGHT} | Lithium is more scarce than Zn, Sulfur cathode performs worse | Not ready yet, but promising prospects \cite{e_bat} | Lower complexity than Metal-Air |
| Li-Air 2020 | 362 \cite{stanfordliair} ,200-800\cite{marketanal:luxresearchinc}, 300 \cite{litair123}, 700 \cite{polyplus1} | 628 \cite{zheng2008theoretical}, 300-400 \cite{litair123} | 50, 50, 10-100 \cite{marketanal:luxresearchinc} | 250 | Lithium is more scarce than Zn, air cathode | Rechargeability and atmospheric issues | Air inlet, large surface area battery |
| Zinc-Air 2020 | 320-650 \cite{marketanal:luxresearchinc}, 370 \cite{lithiumair}, 400 \cite{EnZinc} | 1000 \cite{EnZinc} | 800-1500 \cite{ZiAir1}, 150-4800 \cite{marketanal:luxresearchinc} | 235 \cite{ZiAir2}, 160 \cite{marketanal:luxresearchinc} | Zinc is abundant, air cathode | Rechargeability, affected by atmospheric properties | Air inlet, large cell surface area |
\label{tab:lit_study_bat}
| Battery Tech | Wh/kg | Wh/L | C-D Cycles | $/kWh | Trade-off Result |
|---|---|---|---|---|---|
| Li-S 2020 | 500 | 550 | 1500 | 250 | 0.443 |
| Li-Air 2020 | 532 | 628 | 50 | 250 | 0.239 |
| Zinc-Air 2020 | 418 | 1000 | 1800 | 198 | 0.318 |
\label{tab:final_val_TO_bat}
The sensitivity of the results will be investigated by shifting weights in favour of the second best battery technology, Zn-Air. Criteria at which these cells excel compared to Li-S are especially volumetric energy density and sustainability. A shift of these criteria was investigated resulting in the sensitivity analysis shown in Figure \ref{fig:sens_bat}. The legend shows the battery technology and the criterion that is adjusted per graph. The x-axis indicates the values of that criterion and the y-axis the score of each subsystem. It can be seen that the weight of volumetric energy density needs to be doubled in order to have a better score for Zn-Air. For sustainability, the score should be approximately three times higher than it currently is. Reducing the importance of complexity even further has a negligible effect on the results and was therefore not included in Figure \ref{fig:sens_bat}. This means the sensitivity is low and the analysis is thus robust.
Sensitivity analysis of Wh/L and sustainability
\label{fig:sens_bat}