Maritime fuel cell applications: A tool for conceptual decision making

3.1.1.Physical based hydrogen storage

Physical based hydrogen is a type of storage which stores hydrogen in its pure form. Hydrogen is the lightest and most abundant chemical in the universe. It has the second lowest boiling and melting point of all substances, helium is the only element which has a lower boiling and melting point [5].

Hydrogen is a colour- and odourless gas at ambient conditions with a very low flash point temperature of −253°C [10]. Hydrogen scores high in terms of gravimetric density because of its low weight (33.3 kWh/kg [3]). On the other hand, the volumetric density of hydrogen is very low, the density of hydrogen at ambient conditions is 0.08 kg/m³ [10] compared to 1.2 kg/m³ [31] for air. The low volumetric density makes it impractical to store hydrogen at ambient condition. Therefore, pure hydrogen is typically stored in a compressed or liquid state to increase the volumetric density.

3.1.2.Fuel based hydrogen storage

Storing hydrogen in its pure form is not a very energy dense solution with respect to the volume, even if it is stored at high pressure or cryogenic temperature. Hydrocarbon fuels provide a more dense alternative that can be used as fuel either directly in some types of fuel cells, or in most of the cases by implementing a fuel reforming process in advance of the fuel cell to purify the fuel to a hydrogen rich gas. Different hydrogen containing fuels can be used, such as: Methanol, Ethanol, DME, LNG and Diesel.

For all fuel based solutions some kind of fuel reforming is required, either internal (in case of high temperature fuel cells) or external (for low temperature fuel cells) and sometimes a combination of both for difficult re-formable fuels. The goal of reforming is to process a fuel containing hydrogen into a hydrogen rich gas which can be used in a fuel cell.

Fig. 4.

Fuel processing steps.

A general overview of the fuel processing steps needed to process various types of fuels in combination with different types of fuel cells is given in Fig. 4. In this diagram the fuel cells on the right are ranked from high temperature (top) to low temperature (bottom), which simultaneously orders them according to the required purity of the fuel. It is important to keep in mind that every step adds to the complexity of the system and will reduce the overall efficiency since almost all reforming steps require energy. In general, it can be said that the lower a solution is located in the diagram the lower the efficiency will be with simultaneously increasing complexity [7,27,28].

3.1.3.Material based hydrogen storage

The last type of storage makes use of a carrier material which can be used for bonding hydrogen atoms. The carrier can be ‘charged’ and ‘discharged’ of hydrogen with this type of storage method. When the hydrogen is released from the carrying material it offers pure hydrogen which can directly be used in almost any fuel cell type. Examples of material based hydrogen storage are: Liquid Organic Hydrogen Carriers (LOHC) and Metal Hydrides (such as Sodium Borohydride and AB-5-type alloys).

4.1.1.Overall density

One of the most important factors is the density of the system. The type of fuel in combination with the type of fuel cell determines the necessity and quantity of any type of fuel reforming and thereby influences the overall density of the system. Therefore, it is key to not only look at the fuel and fuel cell separately but as a whole.

The densities of the combined fuel cell and fuel storage system are used to calculate the effective density which is used to draw a Ragone chart for comparing the different combined solutions.

The market study (Fig. 3) showed a large spread in the density between models and/or manufacturers of the same fuel cell technique, this could cause differences in the results of the Ragone charts. The most power dense solutions currently available on the market were used to calculate the overall densities.

Fig. 5.

Energy storage solution, volumetric density comparison [4,8–10,18–21,23,24,33].

Fig. 6.

Energy storage solutions, gravimetric density comparison [4,8–10,18–21,23,24,33].

All mentioned fuels have been compared on both volumetric (Fig. 5) and gravimetric density (Fig. 6) and are plotted together in a graph to easily compare the density of the different fuels. For both densities55 the pure fuel without a storage system and the fuel including a storage system are shown in the two graphs. A distinction is made between fuels which can be stored in a conventional steel tank66 and fuels which need to be stored in a special tank for the fuel including storage system. Special tanks are needed for the storage of pressurised or cryogenic fuels. For the fuels which can be stored in conventional tanks, a factor to estimate the weight and volume of a tank according to S.C. Misra [22] is used to determine the density including storage system. MGO, Methanol, Ethanol, Formic Acid and LOHC are liquid fuels under ambient conditions and can be stored in approximately the same kind of tanks as used for diesel, while all other fuels need a special storage system. The density of these storage systems are based on actual commercially available storage systems, the impact of these tanks on the overall density of the fuel is shown in Figs 5 and 6.

Effective density. The effective power (Peffective) is the power density corrected for the energy density of the fuel storage and efficiency of the energy converter. The formulas for the effective power and energy are formulated according to the method used in the paper written by van Biert et al. [34] as shown in Figs 4 and 5.

Since this overview should give an indication of the density of the total system the power density (P) and efficiency (η) is the combination of both the fuel cell and fuel processing system (Eq. (6) and (7)).

To check this formulation the limits of the Peffective and Weffective can be calculated, which give the limits Peffective≈P for t=0 and Weffective≈ηW for t=∞ [34].

Ragone chart. A Ragone chart is a commonly used type of plot to make a performance comparison of different energy storing devices. With this type of chart the effective energy density (Wh/kg or Wh/L) is plotted versus the effective power density (W/kg or W/L) on a logarithmic scale. In this case the effective densities are used to generate the Ragone chart. A Ragone chart for both the volumetric and gravimetric density are shown in Figs 7 and 8 for several fuel cell and fuel combinations in reference to a diesel engine running on MGO.

Fig. 7.

Ragone chart, volumetric density.

Fig. 8.

Ragone chart, gravimetric density.

The two values Weff and Peff are plotted for a certain time range, since time is included in the formula of the effective density. The mentioned time (t) indicates the continuous amount of time a ship could produce (rated) power without having to fuel in between. This time range is illustrated within the graph by the parallel dashed lines varying from 1 to 1000 hours. An intersection of a Ragone curve with a diagonal timeline will give the effective density and power for that specific solution and time frame. The higher a curve will intersect with a time line the denser this solution will be. In other words, if a graph is diagonally above another combination, this solution is more dense.

From the Ragone chart it can be concluded that there is no single solution that is optimal for a wide range of applications, as is the case with diesel combustion engines. For fuel cells, there are different optimal technological solutions for different operating times. From the two Ragone example diagrams it’s visible that the MGO powered diesel engine (reference) is in nearly all the cases the most energy dense solution. For projects which do not require to be operational for more than a few hours without having to fuel in between, the low temperature fuel cells can offer a more energy dense solution then a diesel engine. When energy converters are compared on the energy density, fuel cells at the moment only seems to be a viable solutions for ships which require a short range and can be refuelled easily, for example: ferries, inland ships or yacht which are used as day boats.

Both the effective volumetric and effective gravimetric densities are taken into account in the tools as decision criteria. As the effective density varies with the expected operation time of the system, the time – factor (t) is implemented as a variable in the tool.

4.1.2.Storage type

The type of storage is seen as an important criteria factor for selecting the type of fuel, therefore it was decided to have separate decision criteria for the type of storage, despite the fact that storage is already included within the density.

Diesel fuel is generally stored inside the double bottom for the current yachts of Oceanco, therefore the amount of space and weight which needs to be added for the construction of tanks is kept to a minimum. Fuel that can be stored in this same location would have the same benefits as well. Fuels that cannot be placed in the double bottom tank have to be placed in another location which is used for other purposes, typically luxury space. For this reason a distinction is made between a fuel which could be stored in the conventional double bottom tank and a fuel where this is not possible. It is assumed that all fuels which are in liquid or powder state under ambient conditions can be stored in the double bottom. These fuels are: MGO, Ethanol, Methanol, Metal Hydrides (Sodium Borohydride), Formic acid and LOHC.

4.1.3.Maturity

The maturity of a fuel and/or fuel cell technology is regarded as an important criteria, since this influences to what extent such a technique can or cannot be readily applied. The technological readiness level (TRL) is used to determine the maturity of the different technologies. The TRL system was developed by NASA in the 1970’s and has since been used by various institutes. In total there are nine different levels used in the TRL system: the first three levels are used for the discovery phase, level 4–6 represent the development phase, level 7–8 are used for projects that have made it to the demonstration phase and lastly level 9 is used for projects that are actually deployed. The TRL status is determined for all different fuel and fuel cell combinations [25].

Next to the TRLs, the existence of relevant companies is taken into account as an additional factor to assess the maturity status of the different fuel and fuel cell combinations. This is done by introducing four different distinction levels: active company for the marine sector (1), active manufacturer in another sector (0.75), active company in the past/ very limited activity (0.25) and lastly no active company (0). For the maturity level, the tool uses the TRL status, which is corrected for the level of company activity with the values mentioned in parentheses.

4.1.4.Safety

As mentioned before, the SOLAS convention addresses the safety requirements assuming that fuels are to have a flash point temperature of at least 60°C, for fuels with lower flash point temperatures the IGF code has to be followed. Due to these regulations the flash point temperature is one of the safety factors which needs to be considered. Besides the flash point temperature, which covers the fire hazards, other factors should be considered concerning safety: health hazards, environmental hazards, instability hazards and chemical compatibility with materials.

The basic health, fire and instability hazards are all covered in a systematic way by the NFPA 704. This is a standard system of the National Fire Protection Association of the U.S. where hazards are indicated with a number from 0-4, where 0 is a minimal hazard and 4 is a severe hazard [12,26]. This hazard indication system is used as a base in the tool to score the importance of health, fire and instability hazards for the different fuels.

4.1.5.Emissions

Emission reduction in shipping is one of the main drivers why fuel cells are considered, therefore the emissions of different fuels are an important decision criteria.

Carbon dioxide emissions are defined to be the most harmful for the environment with respect to greenhouse gasses. Next to the carbon emissions there are also other harmful emissions (e.g. NO_x and SO_x) and for which regulations are becoming increasingly strict.

When a fuel cell is fuelled with hydrogen the fuel cell operates completely emission free, as a fuel cell then only emits heat and water. Other fuels like Ammonia (NH₃) or LNG can contain sulphur or nitrogen which can result in NO_x or SO_x emissions. Sulphur is poisonous for all types of fuel cells, therefore any sulphur traces should be removed in advance. Nitrogen oxides (NO_x) are typically formed by the reaction of nitrogen and oxygen under high temperatures in the form of NO. All types of fuel cells and fuel reformers practically operate below the temperature where NO_x or only limited NO_x will be formed. Therefore, the emissions of NO_x and SO_x by a fuel cell are not taken into account in the tool.

The amount of CO₂ emissions is considered to be of influence when selecting a solution. In this respect only the operational emissions are taken into account in the tool. The mass ratio of carbon dioxide can be obtained using Eq. (8) [29].

This mass ratio gives the amount of CO₂ per kilogram fuel. Together with the energy density of the fuel, the amount of CO₂ emission per kilowatt hour of fuel can be calculated. An overview of the CO₂ emissions for the different discussed fuels are shown in Table 3.

For hydrogen carriers and fuels that do not contain any carbon no CO₂ will be emitted. These fuels or carriers are: Hydrogen (compressed or liquefied), Ammonia (NH₃), LOHC, Metal Hydride (Sodium Borohydride and AB-5-type alloy).

Next to the emissions on board the ship itself, the production and supply of the fuel requires energy as well and therefore contributes to the life cycle emissions of a yacht. This aspect is not covered in the emission criteria nor is it related to a specific amount of CO₂ emissions in the tool. It is however, incorporated in the defined fuel options: i.e. grey hydrogen vs green hydrogen.

For various designs different criteria can be more or less important, therefore a weight factor is added to all decision criteria. These weight factors are implemented as variables within the decision matrix so that the weight factors can be entered by the preference of the user, by changing the weight factors certain important characteristics for the design can be emphasised.

4.2.1.Examples for different application

Case 1: Fast passenger ferry (short range). Distance: 25 nm | Speed: 20 kn | Continuous sailing time: 1.2 h

Explanation of weight factors: For a fast sailing vessel weight is an important factor, volume however is in this case of less importance because this passenger ferry only has to transport people for a short time so a seat per person should do the job.

Case 2: Super yacht (medium range). Distance: 5000 nm | Speed: 14 kn | Continuous sailing time: 357 h

Explanation of weight factors: For a yacht the biggest selling point is the luxury space, this should be maximised therefore the volume density is very important. Weight is also quite important since yachts are generally designed for high maximum speeds. The fuel is always stored in the double bottom, so a fuel which could be stored in the double bottom is preferable.

Case 3: Bulk carrier (long range). Distance: 12300 nm | Speed: 12 kn | Continuous sailing time: 1025 h

Explanation of weight factors: Volume density is important since the payload needs to be maximised as the money will be earned here. As the speed is relatively low and the ship is transporting heavy goods the weight is not seen as an important factor. Storage type is important from the volume point of view because this can save space for cargo.

4.2.2.Input variables

The input variables are coupled to the requirements/targets of the project and are therefore influenced by the user’s assessment. The input variables used for the example cases (shown in Table 4) are used only for demonstration. The reasoning behind the input values of weight, volume, and storage type has been explained in the example cases.

The emission weight factor is the same for all cases and is set to a relatively high value, since the reduction of emission is the main reason to look into the possibilities of fuel cells. The internal weight factors of the different safety categories vary from 1 to 3 indicating the importance of the different type of hazards, however the overall weight factor is determined by the safety factor and is set the same for all cases. Maturity is set to one (1) for all cases, indicating the importance of the technology being reasonably developed, while not adding extra complexity for comparing between the cases. Lowering the weight factor of maturity gives the possibility to compare ‘future’ solutions with currently available solutions.

4.2.3.Results

The top ten results for all three cases are shown in Fig. 10, the most important conclusions are mentioned below.

Example case 1: only LT-PEM fuel cells with different types of fuels are been brought forward in the top ten results. For this type of ship the following fuels are the most promising options in combination with a LT-PEMFC: Liquid hydrogen, Compressed hydrogen, LOHC, Formic Acid (renewable) and metal hydride (AB-5-type alloy).

For the other two examples the HT-PEMFC and SOFC become increasingly attractive, since the sailing time increases. For longer continuous operating times these types of fuel cells in combination with energy dense hydrocarbon fuels will be more energy dense.

For example case 2: all different types of fuels cells end up within a relatively small difference in score. The methanol HT-PEMFC’s and SOFC in combination with methanol or diesel are very close and are the most promising. Further a LT-PEMFC could be interesting in combination with methanol. For this specific example case a snapshot of the decision tool including the normalised decision criteria is shown in Fig. 9 for the top ten results.

Fig. 9.

Decision making tool, top 10 results (diesel engines are mentioned in the tool for reference) example case 2.

The best options for the last case only consist of methanol HT-PEMFC and SOFC’s in combination with diesel, methanol or ethanol.

Next to the defined weight factors, the results were checked with a sensitivity analysis. During this sensitivity analysis the criteria weight factors are varied with 700 random values between 0 and 1 while keeping a constant time factor. These random values are then used to calculate the total scoring, with this approach the result should be independent of any personal or project orientated preference. The 700 different scores are compared based on the: geometric mean, average and median scores. These three scores all showed the same result, therefore only an overview of the median is shown in Fig. 11. Comparing the calculated scores with the sensitivity analyses (Fig. 10 and Fig. 11) shows that the weight factors have an impact on the results. However, this only has a limited impact on the top ten results. In each case, the same types of fuels and fuel cells appear in the top ten results. There are some small changes in the order of the results and in some cases changes in the type of technology. Since the impact of the weight factors is relatively limited in these example cases, it can be concluded that the results offer a robust solution since largely the same is found for a more average approach. However, the results of the sensitivity analysis cannot be used as optimal solution for all applications. When there are large outliers in the chosen weighing factors, this will have a large impact on the results and will result in bigger deviations.

Fig. 10.

Decision making tool, top 10 results example cases.

Fig. 11.

Decision making tool, sensitivity analysis.