1132 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO.

4, OCTOBER 2021

A Strategy for Sizing and Optimizing the Energy

System on Long-Range AUVs
Ariel Chiche , Göran Lindbergh , Ivan Stenius , and Carina Lagergren

Abstract—Autonomous underwater vehicles (AUVs) are utilized

in a variety of unmanned missions, such as environmental surveys
or guarding coastal waters. Currently, the most commonly used
energy storage is a secondary battery pack. A potential solution
for increasing the range even further on larger AUVs is to uti-
lize hybrid fuel cell/battery systems. Fuel cell systems can have a
significantly increased specific energy compared to batteries and
are now starting to become well-developed technologies. However,
in an underwater environment, both hydrogen and oxygen need
to be stored for the fuel cell. This study considers the sizing of a
hybrid fuel cell/battery system with respect to operational power
profiles, i.e., how to select the best combination of fuel cell stack,
battery pack, and reactant storage, and compares its volume and
weight with conventional battery systems. This article presents
a sizing strategy developed for this purpose, where the optimum
combination of fuel cell output power and battery size is analyzed.
The strategy is implemented on power profiles from real AUV
missions. The sizing strategy will outline the techniques to reach
the optimum hybrid configuration capable of meeting the power
demand at any time of the mission. Results show that for longer
missions, the volume of the hybrid system becomes significantly Fig. 1. Ragone plot (based on [14]) comparing the power density and energy
lower than for the battery system, meaning that the endurance can density of different types of energy-storage systems.
be increased thanks to the use of fuel cells.
Index Terms—Autonomous underwater vehicle (AUV), battery,
fuel cell (FC), hydrogen storage, optimization. The Swedish Maritime Robotics Centre (SMaRC) aims to
design and develop AUVs to achieve missions in unknown
waters and perform environmental analyses. These AUVs will
I. INTRODUCTION
need not only highly advanced sensors and sonars for the studies,
UTONOMOUS underwater vehicles (AUVs) are typically
A used for missions, such as ocean monitoring and sam-
pling of water or oceanographic data collection. Well-developed
but also an efficient power system. The research in SMaRC
focuses on three target application areas: 1) ocean production of
algae; 2) safeguarding society; and 3) environmental monitoring,
sensors, sonars, and underwater communication systems make including missions in unexplored polar marine environments.
these AUVs very efficient and capable of ambitious missions. The general objective is to increase the autonomy and endurance
However, all the onboard equipment requires power, and that is of marine robots to reduce the reliance on ships and crews,
achieved at the expense of endurance of the AUV. Therefore, resulting in a reduction of the costs of these operations. Today,
the power required for increasing range and endurance is a key AUVs are usually powered by secondary (rechargeable) batter-
performance parameter when evaluating an AUV. ies, such as Li-ion batteries [1]–[4]. These have a high efficiency
but both weight and volume increase rapidly with the energy
Manuscript received December 13, 2019; revised November 13, 2020 and needs. Moreover, working in a cold environment has a negative
February 12, 2021; accepted February 17, 2021. Date of publication April
2, 2021; date of current version October 13, 2021. This work was supported influence on the battery performance, since that decreases when
by the Swedish Foundation for Strategic Research (SSF) through the Swedish the temperature drops, which is likely to occur during many
Maritime Robotic Centre (SMaRC IRC 15-0046). This work was presented in underwater missions [5]. Compared to batteries, fuel cell (FC)
part at the IEEE OES-AUV 2018 Symposium, Porto, Portugal, Nov. 6–9, 2018.
(Corresponding author: Ariel Chiche.) systems can offer a higher energy density in watthour per liter
Associate Editor: N. Cruz. but have a lower power density in watt per liter, as shown in the
Ariel Chiche, Göran Lindbergh, and Carina Lagergren are with the Division Ragone plot; see Fig. 1. Such a plot is often used to compare the
of Applied Electrochemistry, School of Engineering Sciences in Chemistry,
Biotechnology and Health, KTH Royal Institute of Technology, 114 28 Stock- power density and energy density of different types of energy
holm, Sweden (e-mail: chiche@kth.se; gnli@kth.se; carinal@kth.se). storage systems and it highlights the potential of FC systems in
Ivan Stenius is with the Division of Naval Architecture, School of Engineering terms of energy density. This means that FCs can provide larger
Sciences, KTH Royal Institute of Technology, 114 28 Stockholm, Sweden
(e-mail: stenius@kth.se). endurance for the AUV although the maximum power is lower.
Digital Object Identifier 10.1109/JOE.2021.3062047 Several studies in which AUVs powered by a pure FC system
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1133

type, and the technology is relatively mature and robust. Interest

in the use of PEMFC, for example, propulsion in passenger
cars, has led to strong growth in the FC industry. A PEMFC
consumes oxygen and hydrogen to produce electricity, water,
and heat according to the following electrochemical reactions:
Anode : H2 = 2H+ + 2e− (1)
1
Cathode : O2 + 2H+ + 2e− = H2 O (2)
2
1
Total : O2 + H2 = H2 O. (3)
2
The FC system usually requires storage of hydrogen, for
Fig. 2. AUV “HUGIN” used to run test missions within SMaRC.
which several alternatives exist. Given that this study concerns
operations underwater, oxygen storage is also needed, and this is
or a hybrid one (FC/batteries) are presented in the literature [1], less commonly explored than the hydrogen storage. The choice
[6]–[8]. These studies show the feasibility of and interest in of the hydrogen and oxygen storage for the FC system has a very
investigating such systems for AUVs. Hybridization aims to use strong impact on the total weight and volume of the system.
FCs for the basic power demands of the AUV. When power
demands are high, the FCs are supplemented by batteries, which B. FC-Powered AUVs
can then be recharged when power demands are low [9], [10]. Since endurance is one of the main challenges for the per-
Although there are several established examples of hybrid formance of AUVs, primary and secondary batteries , in some
FC/battery vehicles such as cars [11] and buses [12], there is as projects over the past two decades, have been replaced by
yet no standard drive cycle for AUVs. Additionally, the inherent FCs [1], [6], [8]. The projects have had different goals and
underwater nature of AUVs generates additional constraints, ranges, but they all use FC-based systems to power the AUV
e.g., neutral buoyancy [13], not faced by land-based vehicles. and have been successfully conducted. Even if there is no
Furthermore, there is no regenerative braking that can be used commercial FC-powered AUV as yet, these examples show that
to charge the battery on an AUV. The battery can only be charged this solution is credible and likely to be developed in the coming
by the FC stack. years. The Japanese URASHIMA project [1], in which a PEM
This study presents an analysis to select the best combination FC and Li-ion batteries power the AUV, and a metal hydride is
of FC, battery, and reactants storage (hereafter: hybrid system) used as fuel storage, started in 1998. In 2005, a 56-h mission was
with regard to volume and weight for use in AUVs. The study performed, travelling 317 km and reaching a depth of 800 m, and
considers aspects such as FC stack, hydrogen storage, oxygen in 2009, a 1000-h test was performed with an energy efficiency
storage, battery pack, battery state-of-charge (SoC), compensa- of 60%. Other projects such as IDEFIX [1] (France) and [7],
tion needed to reach neutral buoyancy, and operational power which use compressed hydrogen and compressed oxygen for its
profile. A sizing strategy for designing hybrid FC/battery sys- hybrid system, and DEEPC (Germany) [8], which achieved a
tems and comparing volume and weight of a full battery system 60-h endurance mission in 2004, point to strong and continuing
and a hybrid system is developed and used on real power profiles interest and success in using FCs for AUVs. Table I provides
from AUV missions. In a preliminary study [15], only the weight an overview of projects found in the open literature using this
and volume of the battery pack, the FC, and the reactants were kind of technology to power AUVs. All the presented projects
estimated. This article extends the study by also considering showed promising capabilities. However, there is a need to define
the battery SoC during the mission, the total buoyancy of the a methodology to design efficient AUVs and take advantage of
energy system, and the compensation needed to reach neutral both the FC and the battery [17].
buoyancy. This article also introduces an optimization strategy
to choose the output FC power and keep the system as compact III. DESIGN PHILOSOPHY
as possible. By using this sizing strategy, the main goal is to
identify the energy system that is most volumetric and weight A. Methodology
efficient for a given mission. The methodology was applied Several methodologies for optimizing the energy system on
both to missions undertaken by an existing AUV, Kongsberg’s AUVs have been developed in recent years. One important
HUGIN (see Fig. 2), and to hypothetical missions. step is to define the boundary conditions [20]. Such conditions
highlight the key parameters that influence the design of the
II. FCS FOR AUVS AUV and what needs to be optimized. In space exploration, it
is mass [21]. For underwater vehicles, it is typically volume
A. FC System and neutral buoyancy [17]. In this article, the results will,
There are several types of FCs, but the proton exchange mem- therefore, primarily be compared regarding volume. Since the
brane FC (PEMFC) is the most suitable for powering AUVs, key is to minimize both volume and weight, where volume is
since the operating temperature is low (15 ◦ C–80 ◦ C) compared more important than weight, each component of the system
to other FCs [16]. The PEMFC is currently the most common FC will be evaluated according to its technical performance and
1134 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO. 4, OCTOBER 2021

TABLE I will be chosen to optimize for regenerative braking that charges

EXAMPLES AND CHARACTERISTICS OF FC-POWERED
(FROM AUVS [1], [3], [6], [8], [18], AND [19])
the battery [26]. For underwater vehicles, the battery can only be
charged by the FC stack, and hence, the degree of hybridization
is not the same as for wheel-based land vehicles. Moreover,
the power profile of a mission is also an important input when
designing the energy storage system. It dictates the energy
consumption and the dimensionless parameters, e.g., the power
ratio P and the time ratio T
Pmax
P = (5)
Pbase
t1
T = . (6)
t2
For periodic behavior, the power profile can be divided into
cycles, where each of them corresponds to a battery discharging
time t1 and charging time t2 . The average power Paverage can
also be estimated

Power(t)dt
Paverage = (7)
Time
thanks to the mission’s power profile where “Power” is the power
demand and “Time” is duration of the mission.
Examples of power profiles are shown in Figs. 3, 6, and 10.
Since the FC is supposed to charge the battery during base
power demand, it is important to highlight which phase of the
operation is limited: the charging or discharging phase of the
to its volume. The remaining available space of the AUV can battery. Parameters T and P determine the criterion for which
be used for the implementation of other systems such as new this phenomenon is limited. The charging is limited if
sensors and sonars, but only if the weight is reduced as well. It
is important to note that the size of the tubing is not considered 2 × T × (P − 1)
DoHPower < . (8)
in the study, although it is mentioned in Section V. The general P × (T + 1)2
sizing procedure described in this section is summarized in When the charge is limited, the battery cannot be fully charged
Algorithm 1. by the FC during the base power demand of the AUV. If the
The hybrid system is composed of the FC system and the discharge is limited, the battery is not fully discharged after peak
battery pack. It is assumed that the FC is maintained at constant power utilization. Using these parameters, the sizing strategy
power during the entire mission [22]. The battery pack is being has been applied to different cases and power profiles. The main
charged by the FC when the power demand of the AUV is idea is to estimate the total energy need of the mission and, thus,
below the FC output power and discharged when the power to determine the size of the FC, the amount of hydrogen and
demand is higher than the FC power. The assumption of constant oxygen needed, and the size of the battery. This gives the total
FC power is simplistic but is commonly used when sizing volume and weight of the hybrid system for three oxygen and
hybrid systems [9], [22]–[25]. Furthermore, using dimensionless nine hydrogen storage alternatives using their energy density and
parametric characterization of the mission is beneficial in the specific energy and compares it with full primary and secondary
design of hybrid AUV systems, since the hybridization of an battery systems. For the batteries, an efficiency of 95% and a
FC and electrical energy storage unit depends on the system utilization of 80% of the total SoC window are considered. The
power requirements. Several dimensionless parameters can be batteries are assumed to be fully charged at the beginning of the
used to characterize the system [9]. The degree of hybridization mission. The overall FC system efficiency ηfc is assumed to be
with respect to power (DoHPower ) is the most important one; it 60% in all the following estimations, as in [25].
compares the power provided by the FC, Pfc and the maximum
power demand Pmax
B. Hydrogen and Oxygen Storage
Pmax − Pfc Storage of hydrogen and oxygen represents a real issue for
DoHPower = (4) underwater applications, especially the storage of oxidant. In
Pmax
most other FC applications, the oxidant is the ambient air. For
If DoHPower is high (close to 1), it indicates a high power commercial PEMFC cars, the hydrogen is stored as compressed
output from the batteries compared to the FCs, meaning that the gas at 350 or 700 bar. Table II compares several solutions
batteries will provide a greater part of power during power peak for storage of both hydrogen and oxidant, whereas Table III
demand. For land-based vehicles, the degree of hybridization shows the gravimetric and volumetric densities. The content of
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1135

TABLE II it possible to check if the proposed design is suitable and to study

GAS STORAGE CHARACTERISTICS (FROM [1], [8], AND [20])
how it behaves.
As a starting point, the battery is sized to fulfill one cycle of
peak power demand Escycle . The estimation takes the battery
efficiency Beff and its utilization BU into account
(Pmax − Pfc ) × t1
Escycle = (12)
Beff × BU
The battery SoC

(Power(t) − Pfc )dt
SoC(t + 1) = SoC(t) − (13)
Escycle
is also estimated for each time step of the mission (each second):
The battery is charged when the power demand P ower is below
the FC power. If that is the case and if the battery SoC reaches
a set maximum value (around 1), the battery is no longer being
charged.
The battery sizing is included in a recursive loop: If the SoC
reaches a minimum value (0 or 0.1 for example) at some point,
the size of battery pack Escycle is multiplied by a given coefficient
(here 1.01) and the SoC as function of time is estimated again.
This goes on until the SoC never reaches the set low limit. This
procedure ensures a full energy system able to fulfill the mission
with the smallest possible battery pack coupled with the FC. The
final battery energy content is then Escyclefinal . This procedure is
hydrogen per volume and weight of storage are compared. The also presented in Algorithm 1.
storage includes the tank, its wall, and insulation. Furthermore, Crate of the battery has to be considered. Crate
To estimate the necessary amounts of hydrogen and oxygen, of 1 C means 1 h for a full charge (or discharge), 0.5 C or C/2
the FC is assumed to run at constant power during the entire is a 2-h charge, whereas 2 C means 30 min. The Crate can have
mission. The electric energy delivered by the FC (Efc ) is then an important impact on the battery’s aging process [29]. If Crate
is higher than 1, the battery should be designed for power; if
Efc = Pfc × Time (9) it is lower than 1, it should be designed for energy. This rate
with “Time” being the mission duration. The amount of hydro- is estimated using the maximum battery power PmaxBat and its
gen needed mH2 is estimated using the low heating value of energy content
hydrogen LHVH2 and the overall FC system efficiency ηfc PmaxBatt
Crate = . (14)
Efc Escyclefinal
mH2 = . (10)
LHVH2 × ηfc On a practical level, it is better to have Crate close to, or below, 1
so that the battery can deliver all its energy and limit the ageing
LHVH2 is used since the produced water is not entirely in the
process [29]–[31].
liquid form. The amount of oxygen consumed is estimated using
Eventually, when the battery is correctly dimensioned, a new
stoichiometric (3) and (10)
dimensionless parameter can define the entire energy system:
mH2 × MO2 the degree of hybridization with respect to energy (DoHEnergy )
mO 2 = (11)
MH 2 × 2 Efc − Escyclefinal
DoHEnergy = . (15)
where MH2 and MO2 are the molar masses of hydrogen and Efc
oxygen, respectively. The weight and volume of the storage can If DoHEnergy is close to 1, it means that most of the energy
then be estimated by dividing the gas weights simply calculated comes from the FC, i.e., from the hydrogen and oxygen storage.
by the gravimetric and the volumetric density of each storage In contrast, if it is close to 0, it means that the major part of the
solution (see Table III). energy comes from the battery pack.

C. Battery Design D. Optimization Strategy

Since hybrid systems are considered in this study, sizing the One important step is to find the optimal FC output power,
battery stack is part of the methodology. The battery SoC during when the volume of the entire system is limited. The idea is to
the mission is a key parameter that needs to be taken into account estimate the volume and weight of the hybrid system for several
when sizing the battery pack. Several articles about strategies configurations (FC output power within a certain range) and find
for energy management explain how to estimate the SoC as a a local minimum. By changing the FC output power, the amount
function of time [27], [28]. Being able to follow the SoC makes of hydrogen and oxygen to store is changing, as is the size of the
1136 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO. 4, OCTOBER 2021

TABLE III
FUEL AND OXYGEN STORAGE CHARACTERISTICS (FROM [9])

battery pack. The result is a combination of an FC output power

and a battery size that can fulfill the mission while minimizing
volume and weight.
The FC size is estimated using linear interpolation for weight
and volume from several commercial PEMFC stacks [32], [33].
The type of hydrogen and oxygen storage does not change the
optimum point found. Indeed, changing the hydrogen and oxy-
gen storage influences the weight and volume but the minimum
will appear for the same FCs output power and, therefore, the
same amount of hydrogen and oxygen needed. On a practical
level, the optimization is made using a loop, which makes the
full weight and volume estimation for each FCs power.

E. Neutral Buoyancy of the Energy System

Since this study concerns an underwater application, it is, thus,
important to take into account the buoyancy of the system [13],
[34]. Submarine subsystems typically aim to have a volumetric
mass of the energy system ρ close to 1 kg/L, which means neutral
buoyancy. If it is below the density of water, approximately
1 kg/L, some weight has to be added to increase the volumetric
mass. Commonly, lead or steel with high density is used. If
the volumetric mass on the other hand is above 1 kg/L, the IV. RESULTS
buoyancy needs to be increased by adding some volume with
Several missions were conducted using HUGIN, an AUV
a lighter material than water (most efficient gain in buoyancy
equipped with Li-ion batteries, pictured in Fig. 2. In this article,
is achieved by only filling that volume with air; however, typi-
two representative missions from the tests with HUGIN are used,
cally this is achieved with some lightweight foam). Considering
one short (5-h mission), case 1, and one longer mission (40-h
such compensation to ensure neutral buoyancy is crucial when
mission), case 2. All parameters such as t1 , t2 , Pmax , Pbase , Pfc ,
designing an AUV and the volumetric mass ratio of each energy
and mission duration mentioned earlier are estimated using these
system is, therefore, estimated. In this article, it will be assumed
real power profiles. For nonperiodic profiles, t1 and t2 were
that steel is used to increase the mass of the system and empty
estimated by taking the longest peak power demand and the
volume with only air inside will be used to decrease the mass of
longest low power time. Finally, a further case, case 3, is studied,
the system (referred to as steel or air compensation). In addition,
exemplifying how arbitrary user-defined power profiles can be
the compensation (air or steel) needed to reach neutral buoyancy
analyzed.
can be estimated by solving the following system of equations:
 Whybrid-system +Wcomp A. Case 1: Short-Range Mission
Vhybrid-system +Vcomp =1
Wcomp (16) Case 1, which corresponds to the power profile from the short-
Vcomp = ρcomp mission test (see Fig. 3), was implemented in the sizing and
optimization process (see Algorithm 1), and the optimal power
where Whybrid-system and Vhybrid-system are the weight and volume output of 900 W was found minimum in the blue curve in Fig. 4.
of the hybrid system without compensation, respectively, Wcomp The average power demand is 929 W, as presented in Fig. 3.
and Vcomp are the weight and volume of the compensation In this case, using an FC power of 900 W, the battery charging
needed, respectively, and ρcomp is the density of the considered and discharging times are found to be comparable, 20.2 and
compensation. If the buoyancy is above 1, the compensation 29.6 min, respectively. DoHPower is low [0.31; see (4)], which
is air, with a density of 1.23 g/L. If the buoyancy is below 1, means that most of the power is taken from the FC, whereas
the compensation is steel with a density of 8 kg/L. The neutral the obtained power ratio P is high [2.08; see (5)], showing an
buoyancy procedure is also presented in Algorithm 1. important difference between the base power and the maximum
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1137

Fig. 3. Power profile for a short-range HUGIN mission with a 900-W FC stack
and estimation of the battery SoC. Fig. 5. Volume distribution for several hydrogen storage solutions and liquid
oxygen storage using the power profile from the short HUGIN mission.

The short-mission power profile was used to simulate and

compare the size of different energy systems; see Fig. 5. While
varying the hydrogen storage for the hybrid systems, liquid
oxygen storage was used in all cases. The results show that
the hybrid system is more volume efficient than the secondary
battery one, although the final volume is highly dependent on
the method used to store hydrogen. In addition, the type of
compensation, i.e., air or steel, and its size varies with the type of
storage. It is also interesting to note that, for the two battery cases
and the liquid oxygen combined with metal hydride system,
the contribution of the air compensation to the total volume is
very high. When a storage system for hydrogen with higher
volumetric density is chosen, the volume of the hybrid system
comes closer to the full primary battery. The FC itself contributes
to a substantial part of the total volume in the hybrid systems. For
Fig. 4. Volume of the entire hybrid system (on the left axis) of the battery pack a short mission, the difference between hybrid and pure battery
and the FC stack (on the right axis) as a function of the FC output power for the
short HUGIN mission (liquid oxygen and metal hydride storage).
systems is not significant but an interesting trend in favor of
optimized hybrid systems appears.
TABLE IV
BATTERY CHARACTERISTICS (FROM [35] AND [36]) B. Case 2: Long-Range Mission
A long-range 40-h mission (see Fig. 6), conducted by
HUGIN was also implemented in the optimization process (see
Algorithm 1). Fig. 7 shows that the best power output of the FC
is 980 W, whereas Fig. 6 presents the average power demand of
the mission 923 W. With an FC power of 980 W, the analysis of
one. Crate is equal to 1.07 [see (14)] meaning that the battery the power profile gave a low DoHPower [0.2; see (4)], which was
does not need to be designed for a specific purpose (energy or expected since the power profile is relatively steady and peak
power) and can be implemented with this design. According to power demands rarely occur.
the criterion defined in (8), this means that battery charging is The low DoHPower shows a high utilization of the FC in
limited, not discharging. Furthermore, the battery SoC during this case. Moreover, for this longer mission, the charging is
the mission is estimated and it never goes below the minimum limited, but it is interesting to note that the ratio T is close to
limit (0.1). DoHEnergy is 0.82 [see (15)], which means that the 1, again, due to the stable power profile. In addition, the end
energy primarily comes through the FC from the hydrogen and of the mission requires less power than the rest of it; therefore,
oxygen storage. It is interesting to note that there are two peaks the battery SoC reaches 100%, meaning that the battery is no
of power demand at the end of the mission. The battery energy longer being charged when the power demand is low. It should
is, therefore, depleted down to 10% of SoC. The characteristics lead to lower reactants consumption. The high Crate [5.10; see
of the batteries considered in this article are listed in Table IV. (14)] highlights the need for power-designed batteries. Since
1138 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO. 4, OCTOBER 2021

Fig. 6. Power profile for a long-range HUGIN mission with a 980-W FC stack
and estimation of the battery SoC.

Fig. 8. Weight distribution for several hydrogen storage solutions and liquid
oxygen storage using the power profile from the long HUGIN mission with
neutral buoyancy.

Fig. 7. Volume of the entire hybrid system (on the left axis) and of the battery
pack and the FC stack (on the right axis) as a function of the FC output power
for the long HUGIN mission with liquid oxygen and metal hydride storage.

Fig. 9. Volume distribution for several hydrogen storage solutions and liquid
the longest discharging time is not the interval that draws the oxygen storage using the power profile from the long HUGIN mission with
most power, the battery SoC never reaches 0.1, meaning that neutral buoyancy.
the battery stack is slightly oversized. The very high DoHEnergy
[0.99; see (15)] is due to the very short peaks of high power maintain a neutral buoyancy in the AUV. In contrast, the metal
demand and highlights an uneven distribution of energy storage hydride and 200-bar hydrogen storage options have a high volu-
in the system (liquid oxygen and metal hydride storage) between metric mass, which means that some air is needed in the system
the FC system and the battery pack. to ensure neutral buoyancy. When comparing the compensations
The weight and volume estimation for the 40-h mission, needed for the short mission with the ones needed for the long
presented in Figs. 8 and 9, shows that the hybrid system is mission, it appears that the compensations needed increase with
more weight and volume efficient than the secondary battery the duration of the mission. The addition of air as compensation
solution. In addition, for some combinations of hydrogen storage highly impacts upon the total volume of the system, whereas the
and liquid oxygen, e.g., 700 bar or liquid hydrogen, the hybrid addition of steel mostly increases the total weight. Once again,
system can also compete with the primary battery system in the compensations needed for the two battery cases represent
terms of weight and volume. In the case of an AUV, the metal an important part of the total volume of the system. When
hydride alternative can be a very interesting hydrogen storage comparing the two HUGIN missions (long and short), it seems
solution. In that case, the oxygen storage becomes the most that when the mission length increases, the performance of the
volume-demanding part. It is important to estimate the com- hybrid system, in terms of volume, comes closer to the full
pensations needed. Storing hydrogen in liquid form or as a gas primary battery system. The two studied missions have similar
at 700 bar, combined with the liquid oxygen storage, has very power demands, but different lengths and behavior. The first
low volumetric mass, resulting in the necessary steel addition to analysis showed that when the mission duration increases, the
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1139

TABLE V
MISSION SPECIFICATIONS FOR THE HYPOTHETICAL POWER PROFILE

Fig. 11. Volume of the entire hybrid system (on the left axis) and of the battery
pack and the FC stack (on the right axis) as a function of the FC output power
for the hypothetical power profile with liquid oxygen and metal hydride storage.

Fig. 10. Hypothetical power profile with a 1260-W FC stack and estimation
of the battery SoC.

hybrid system becomes more relevant. The choice of hydrogen

storage is shown to have a significant impact on the final weight
and volume efficiency of the hybrid system. The battery in the
hybrid system seems to be negligible compared to the rest of the
system, meaning that an oversized battery pack is not a serious
issue in that case.

C. Case 3: User-Defined Mission

Generated simple and cyclic power profiles can be used to
conduct the same weight and volume analysis as presented ear-
lier. Instead of reading a power profile, the user can implement
the key parameters t1 , t2 , Pmax , Pbase , Pfc , and duration of the Fig. 12. Volume distribution for the hypothetical 20-h mission when using
mission into the sizing strategy. Such a hypothetical profile for liquid oxygen and different hydrogen storage solutions with neutral buoyancy.
a 20-h mission was generated to match the parameters from the
short mission with HUGIN and its specifications are listed in
Table V. An optimization (see Algorithm 1) was also made for distribution in terms of energy storage between the battery pack
this power profile (see Fig. 10), and led to an FC output power and the FC system (hydrogen and oxygen stored).
equal to 1260 W, as shown in Fig. 11. The power profile with By using the sizing strategy (see Algorithm 1) with input
the optimized FCs output power and the corresponding battery from Table V, it is possible to size and compare different energy
behavior is presented in Fig. 10. storage solutions. Fig. 12 shows the volume distribution of the
The resulting power profile is highly cyclical and reaches the energy system for several alternatives for hydrogen storage and
maximum power at every cycle, giving a more energy-intense with oxygen stored as liquid. The battery stack is, in this case,
system than for the previously studied power profiles. The anal- very small and its volume is negligible when compared to the
ysis of the dimensionless parameters gives a time ratio T of 0.67 other parts. It is clear from the graph that the choice of solution
[see (6)], i.e., times for charging and discharging are comparable. for gas storage is essential. However, storing hydrogen with
According to the criterion in (8), the time for charging the battery lower energy density, such as compressed gas at 200 bar, does not
from the FC system is limited. The power ratio P is 2 [see (5)], clearly favor the hybrid system over the full secondary battery
showing that the difference between the base and peak power system.
demand is significant. Furthermore, in this hypothetical case, Fig. 13 compares the volume of three different energy systems
C rate is high, 2.7 [see (14)], meaning that the battery pack has as a function of the energy needs, i.e., as a function of the
to be power optimized. This is due to the numerous cycles in mission duration. In the hybrid system, liquid oxygen and liquid
the power profile and the relatively long duration of the mission. hydrogen are used. The simulations provide one intersection
DoHEnergy , in this case, is 0.99 [see (15)], which shows an uneven point: between the hybrid system and the rechargeable battery
1140 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO. 4, OCTOBER 2021

hydride (red) curves are above 1. Since volumetric mass of each

system show an asymptotic behavior when the duration of the
mission increases, the compensation needed will eventually be
proportional to the time of the mission.

D. Generalization
It is interesting to compare the specific energy and the energy
density of all the energy system solutions, as shown in Fig. 15,
and for a long time simulation. To run this simulation, the
hypothetical power profile was extended to up to a 1000-h
mission. The total weight and volume were estimated for 1100,
500, and 1000 h. The hypothetical power profile can be compared
to the short HUGIN mission in terms of t1 , t2 , Pbase , and peak
power demand. The primary battery shows good characteristics
in terms of weight and volume, but it needs to be changed
Fig. 13. Volume as a function of energy need for the hypothetical 20-h mission for each mission, which makes this solution less sustainable.
at 1260-W average power when using liquid hydrogen and oxygen storage for Fig. 15 shows that when the hydrogen and the oxygen storage
the hybrid system without neutral buoyancy.
are well selected, the hybrid system’s specific energy and energy
density are higher than for the secondary battery system. The
combination of liquid hydrogen and oxygen (yellow symbols)
shows the best results in terms of specific energy but not in terms
of energy density; indeed, the combination of liquid oxygen and
metal hydride (green symbols) reaches a higher energy density.
The combination of liquid oxygen and compressed hydrogen at
700 bar (purple symbols) also seems to be a promising solution
since it follows the trend of the liquid oxygen and hydrogen
combination.
Fig. 15 also shows how the specific energy and the energy
density of the hybrid solution dramatically increase with the
mission duration. In addition, when considering neutral buoy-
ancy (squares), the same trend appears. The final specific energy
and energy density are, therefore, not as high as without neutral
buoyancy considered; however, they are still higher than the
performance of the secondary battery. There is a clear influence
Fig. 14. Volumetric mass as a function of energy need for the extended
of the compensation to reach neutral buoyancy on the final
hypothetical 1000-h mission for several hydrogen storage solutions when using energy storage capacity of the different systems, especially
liquid oxygen storage without neutral buoyancy. for the metal hydride case. This trend further highlights the
potential for using FC systems in AUVs. Finally, the simulation
highlights an asymptote in the energy density and specific energy
system. This point occurs around 3 kWh of energy demand (see
for a hybrid system when the duration becomes very long for
Fig. 13), and the difference in volume between these two system
all the combinations. This behavior was expected since when
solutions keeps increasing with increasing energy demand. The
the energy storage increases, the FCs size and weight become
volume of the primary battery system is the highest in these
negligible in the total system.
simulations, but the trend of the black (hybrid system) and blue
(primary battery) curves is that they also will cross each other
when more energy is needed, i.e., when longer missions are V. DISCUSSION
considered. The sizing strategy developed in this article shows a clear
The designed mission is extended up to a 1000-h mission, trend in favor of hybrid systems. However, some assumptions
leading to an energy consumption going from 0 to 1275 kWh. can be discussed. First of all, the different parameters of each
The volumetric mass of hybrid systems with different types energy storage solution and the overall efficiency of the battery
of hydrogen storage and liquid oxygen, without the neutral and the FC system are considered. Additionally, the tubing part
buoyancy, is calculated and presented in Fig. 14. The 700-bar for the hybrid FC/battery system has not been considered in
compressed hydrogen (blue) and the liquid hydrogen storage this study. For this reason, the real buoyancy is likely to change
(green) follow exactly the same pattern; the volumetric mass first since free space is not considered when the volumetric mass is
decreases rapidly when the energy need increases and then seems estimated and will, therefore, be lower than the calculated one.
to reach an asymptote close to 0.66 kg/L. All four solutions To consider the buoyancy as a major parameter has not been
reach an asymptote and both the 200 bar (black) and metal the scope of this study, but, of course, this might decrease the
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1141

Fig. 15. Energy density and specific energy for the short HUGIN mission, the extended hypothetical power profile simulations, and the full battery solutions.
For the hybrid system, oxygen is stored as liquid and several solutions for hydrogen storage are considered with an optimized FC output power without and with
neutral buoyancy. The two battery systems correspond to the batteries presented in Table IV.

volume efficiency of the hybrid solution. The shifting point in additional heavy sensors or more energy. On the other hand,
Fig. 13 is expected to move; however, the general trend in the when the energy system needs to be lighter, there is less design
results shown in this article should be similar. An important part freedom in utilizing the volume to add payload capacity.
of future research should be to further investigate the scalability Moreover, the power split between the FC and the battery
and energy density of complete systems. needs to be carefully considered in the future, since this analysis
Furthermore, it is important to put into perspective the good does not take into consideration any advanced energy manage-
performance of the primary battery. This solution is worthwhile ment strategy including the AUV’s dynamic behavior during a
for a very long mission to achieve a specific goal, but it is mission (speed, sensors ON/OFF, etc.). Many studies concerning
not suitable for repeated missions since the battery needs to be energy management strategy for hybrid cars and buses have been
changed for every mission. Primary batteries also typically have performed in the last few years [11], [27], [28], [37]–[42] and
issues under high power demand, which can lead to overheating it will be the next step for this project. In addition, the energy
of the system [36]. The compensation needed to reach neutral management strategy is crucial to increase the lifetime of the
buoyancy is also very important when considering the two battery pack [43]. A dynamic model of the entire system will
battery cases. be created within the SMaRC project to simulate its behavior
Although not presented in Section IV, other oxygen storage and highlight its capabilities. It is also very important to note
systems (compressed at 153 bar and H2 O2 ) can also be consid- that the flexibility in terms of total power output (i.e., speed
ered. They have lower performance in terms of gravimetric and of the AUV) is highly dependent on the battery since the FC
volumetric density than the liquid solution, but they also show is assumed to work at constant power. The optimization tends
that when the mission becomes long and energy intense, they to minimize the battery size to the greatest extent possible, so
are volume and weight efficient. it might be important to slightly oversize the battery pack to
Furthermore, as seen in Figs. 5, 9, and 12, when the mission guarantee more freedom for the AUV.
duration increases and the FC output power is optimized, the In addition, when designing a real AUV, the ratio size/energy
battery pack does not increase that much in terms of size and of the entire AUV can be interesting to estimate and can then
the hybrid solution becomes very attractive compared to the be used to predesign underwater vehicles. The main idea is to
full-battery solution. This also highlights the importance of determine the minimum size of a vehicle for a certain mission
optimizing the FCs output power, shown in Figs. 3, 7, and 11. and endurance. Results show that a full secondary battery system
If the FC is wrongly sized, the battery pack can become bulky for small AUV performing short-range missions is preferred,
and the benefits of the hybrid system will be reduced. whereas an optimized hybrid system is the best solution for long-
Even though it is important to reach neutral buoyancy in the range AUVs.
final design, a low volumetric mass of the energy system is Safety is also an important issue in regard to batteries
beneficial because this gives more design freedom. For example, and hydrogen, especially in a closed environment. Large
the added weight can be used for trimming purposes, adding batteries, both primary and secondary, could be hazardous.
1142 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 46, NO. 4, OCTOBER 2021

Hydrogen can also be a problem and is highly explosive in some be more volume efficient than the full-battery solution. For a
concentration ranges. Pressure around the energy storage hull small AUV performing short missions, rechargeable batteries
can also become a problem for the vessel and all these safety are more suitable than the hybrid system, considering the higher
issues will need to be carefully studied in the future to prove complexity of the hybrid system and the small difference in
the practical relevance of a hybrid FC/battery system. This terms of volume between these two solutions. Hybrid FC/battery
developed strategy does not investigate the difference between power systems are more interesting than battery systems for
pressure-tolerant hybrid systems and enclosed atmospheric AUVs when the mission duration increases. In addition, for long
hybrid systems. Other practical aspects must be considered, for and cyclical missions, the optimization tends to give FC power
example, the condensation due to the FC has to be handled [44]. close to the average power demand. However, when the mission
One solution is to include water reservoir in the system. These is not very cyclical, there is a difference between the optimized
issues will be investigated in the future. FC power and the average power demand.
When building a real FC-powered AUV, heat management The strategy can be updated with the newest technology and
has to be considered given that the AUV may be working in cold used as a first step to estimate the relevance of FC imple-
environments, where both FC and battery need some heat to start mentation for several types of AUVs and underwater missions.
and perform correctly. Hydrogen and oxygen tanks might also Published data on FC powered AUVs show promising results.
need to be maintained at certain temperatures. Since the AUV However, FCs are not yet found in commercial AUVs. The
also produces heat (electric motor for propulsion, FC, sensors, SMaRC project will provide further opportunities to analyze
etc.) a heat management strategy will be investigated in the future and implement this technology.
to ensure all the AUVs systems work properly. Studying the
heat properties of the different hydrogen and oxygen storage
solutions will also help the decision process. As with weight and REFERENCES
volume, heat is a key parameter in selecting the best hydrogen
[1] A. Mendez, T. J. Leo, and M. A. Herreros, “Current state of technology of
and oxygen storage solutions. fuel cell power systems for autonomous underwater vehicles,” Energies,
Last but not least, even if the Ragone plot in Fig. 1 shows vol. 7, no. 7, pp. 4676–4693, 2014.
that an FC system has lower power density (watt per liter) than [2] G. Schubak and D. Scott, “A techno-economic comparison of power
systems for autonomous underwater vehicles,” IEEE J. Ocean. Eng.,
secondary batteries, the proposed combination of FC stack and vol. 20, no. 1, pp. 94–100, Jan. 1995.
battery pack always meet the power demand of the vehicle, as [3] Ø. Hasvold, N. J. Størkersen, S. Forseth, and T. Lian, “Power sources
shown in Figs. 3, 6, and 10. In other words, such hybrid systems for autonomous underwater vehicles,” J. Power Sources, vol. 162, no. 2,
pp. 935–942, 2006.
have potential to meet both the power and energy requirements [4] X. Wang, J. Shang, Z. Luo, L. Tang, X. Zhang, and J. Li, “Reviews of power
of AUVs. systems and environmental energy conversion for unmanned underwater
vehicles,” Renewable Sustain. Energy Rev., vol. 16, no. 4, pp. 1958–1970,
2012.
VI. CONCLUSION [5] A. M. Aris and B. Shabani, “An experimental study of a lithium ion
cell operation at low temperature conditions,” Energy Procedia, vol. 110,
In this article, a strategy for designing compact hybrid pp. 128–135, 2017.
[6] N.-C. Shih, B.-J. Weng, J.-Y. Lee, and Y.-C. Hsiao, “Development of a
FC/battery systems for AUVs has been developed. Both real small fuel cell underwater vehicle,” Int. J. Hydrogen Energy, vol. 38, no. 25,
and user-designed power profiles are utilized to compare hybrid pp. 11138–11143, 2013.
systems with secondary and primary battery systems in terms of [7] E. Raugel, V. Rigaud, and C. Lakeman, “Sea experiment of a survey AUV
powered by a fuel cell system,” in Proc. IEEE/OES Auton. Underwater
weight and volume. Crucial aspects of a hybrid energy storage Veh., 2010, pp. 1–3.
system for AUVs, i.e., FC stack, battery pack, battery SoC, [8] W. Hornfeld, “DeepC: The German AUV development project,” STN
hydrogen storage, oxygen storage, and compensation to reach ATLAS Elektronik GmbH, Bremen, Germany, Tech. Rep., 2002.
[9] Q. Cai, D. Brett, D. Browning, and N. Brandon, “A sizing-design method-
neutral buoyancy, are considered. Results show that a well- ology for hybrid fuel cell power systems and its application to an unmanned
designed hybrid system can be more compact in terms of stored underwater vehicle,” J. Power Sources, vol. 195, no. 19, pp. 6559–6569,
energy per volume than the battery solutions and, consequently, 2010.
[10] P. Thounthong, V. Chunkag, P. Sethakul, B. Davat, and M. Hinaje,
increase the endurance of the vehicle. “Comparative study of fuel-cell vehicle hybridization with battery or
The choices of fuel and oxygen storage for the FC are the supercapacitor storage device,” IEEE Trans. Veh. Technol., vol. 58, no. 8,
main parameters influencing the final weight and volume of the pp. 3892–3904, Oct. 2009.
[11] A. De Risi, V. Paladini, T. Donateo, and D. Laforggia, “Control strategy
vessel’s energy system. Volume is the most limiting parameter in optimization of a fuel-cell electric vehicle,” in Proc. ASME EFC05 Tech.
underwater applications. The optimization, in terms of volume, Conf., 2005, pp. 14–16.
is a crucial procedure to ensure that the system can benefit [12] D. Gao, Z. Jin, and Q. Lu, “Energy management strategy based on fuzzy
logic for a fuel cell hybrid bus,” J. Power Sources, vol. 185, no. 1,
from the advantages of the hybrid system. The optimized energy pp. 311–317, 2008.
systems can meet the power requirement of the AUV at any time, [13] D. F. Waters and C. P. Cadou, “Estimating the neutrally buoyant energy
and were tested for real power profiles. density of a Rankine-cycle/fuel-cell underwater propulsion system,” J.
Power Sources, vol. 248, pp. 714–720, 2014.
The optimization shows promising results to select the best [14] M. Winter and R. J. Brodd, “What are batteries, fuel cells, and superca-
combination of FC stack, battery pack, and reactants storage pacitors?” Chem. Rev., vol. 104, pp. 4245–4269, 2004.
to design the most compact hybrid system for AUVs. For [15] A. Chiche, C. Lagergren, G. Lindbergh, and I. Stenius, “Sizing the energy
system on long-range AUV,” in Proc. IEEE/OES Auton. Underwater Veh.
long-range missions requiring larger amounts of energy, hybrid Workshop, 2018, pp. 1–6.
systems with well-selected hydrogen and oxygen storage so- [16] C. Spiegel, PEM Fuel Cell Modeling and Simulation Using MATLAB.
lutions and a well-designed battery pack have the potential to Amsterdam, The Netherlands: Elsevier, 2011.
CHICHE et al.: STRATEGY FOR SIZING AND OPTIMIZING THE ENERGY SYSTEM ON LONG-RANGE AUVs 1143

[17] R. d’Amore Domenech, M. A. Raso, A. Villalba-Herreros, Ó. Santiago, [42] J. Garus and J. Małecki, “Simulation model of fuel cell energy systems
E. Navarro, and T. J. Leo, “Autonomous underwater vehicles powered by for AUV,” in Proc. IEEE 27th Conv. Elect. Electron. Eng., 2012, pp. 1–4.
fuel cells: Design guidelines,” Ocean Eng., vol. 153, pp. 387–398, 2018. [43] E. Schaltz, A. Khaligh, and P. O. Rasmussen, “Influence of bat-
[18] C.-H. Lee and J.-T. Yang, “Modeling of the Ballard-Mark-V proton tery/ultracapacitor energy-storage sizing on battery lifetime in a fuel
exchange membrane fuel cell with power converters for applications in cell hybrid electric vehicle,” IEEE Trans. Veh. Technol., vol. 58, no. 8,
autonomous underwater vehicles,” J. Power Sources, vol. 196, no. 8, pp. 3882–3891, Oct. 2009.
pp. 3810–3823, 2011. [44] H. Weydahl, M. Gilljam, T. Lian, T. C. Johannessen, S. I. Holm, and J. Ø.
[19] I. Yamamoto et al., “Fuel cell system of AUV ‘URASHIMA’,” in Proc. Hasvold, “Fuel cell systems for long-endurance autonomous underwater
MTS/IEEE Techno-OCEANS Conf.’, vol. 3, 2004, pp. 1732–1737. vehicles — challenges and benefits,” Int. J. Hydrogen Energy, vol. 45,
[20] M. Hitscherich, C. Cremers, D. Stolten, K. Pinkwart, and J. Tübke, “De- no. 8, pp. 5543–5553, 2020.
velopment study of an air independent fuel cell system for an autonomous
underwater vehicle (AUV),” ECS Trans., vol. 75, no. 14, 2016, Art. no. 491.
[21] Y. Sone, M. Ueno, and S. Kuwajima, “Fuel cell development for space
applications: Fuel cell system in a closed environment,” J. Power Sources, Ariel Chiche was born in Paris, France, in 1994. He
vol. 137, no. 2, pp. 269–276, 2004. received the M.Sc. degree in engineering in physics,
[22] O. Siddiqui and I. Dincer, “A review on fuel cell-based locomotive pow- electronics, and materials sciences in 2017 from
ering options for sustainable transportation,” Arabian J. Sci. Eng., vol. 44, Grenoble INP – PHELMA, Grenoble, France, where
no. 2, pp. 677–693, 2019. he specialized in electrochemistry and processes for
[23] M. Nadal and F. Barbir, “Development of a hybrid fuel cell/battery powered the energy and the environment. He is currently work-
electric vehicle,” Int. J. Hydrogen Energy, vol. 21, no. 6, pp. 497–505, ing toward the Ph.D. degree with the Division of
1996. Applied Electrochemistry, KTH Royal Institute of
[24] H. S. Das, C. W. Tan, and A. Yatim, “Fuel cell hybrid electric vehicles: A Technology, Stockholm, Sweden.
review on power conditioning units and topologies,” Renewable Sustain. His work is on the implementation of fuel cell
Energy Rev., vol. 76, pp. 268–291, 2017. systems for underwater and maritime applications for
[25] T. Hyakudome, T. Tani, H. Ito, and K. Sugihara, “Concept of long cruising which he investigates both the design of such systems and the behavior of fuel
range autonomous underwater vehicle powered by fuel cell,” in Proc. cells in such an environment.
MTS/IEEE OCEANS Conf. Kobe, Japan, 2018, pp. 1–5.
[26] A. Rousseau, P. Sharer, and R. Ahluwalia, “Energy storage requirements
for fuel cell vehicles,” SAE, Warrendale, PA, USA, Tech. Rep. 2004-01-
1302, 2004. Göran Lindbergh received the M.Sc. and Ph.D. de-
[27] P. Bubna, D. Brunner, S. G. Advani, and A. K. Prasad, “Prediction-based grees in 1985 and 1991, respectively.
optimal power management in a fuel cell/battery plug-in hybrid vehicle,” Since 2003, he has been a Professor of Chem-
J. Power Sources, vol. 195, no. 19, pp. 6699–6708, 2010. ical Engineering with the KTH Royal Institute of
[28] T. Fletcher, R. Thring, and M. Watkinson, “An energy management Technology, Stockholm, Sweden. He is heading the
strategy to concurrently optimise fuel consumption & PEM fuel cell Applied Electrochemistry group working with elec-
lifetime in a hybrid vehicle,” Int. J. Hydrogen Energy, vol. 41, no. 46, trochemical power sources and electrolytic processes
pp. 21503–21515, 2016. within the field of electrochemical engineering. A
[29] A. S. Mussa, M. Klett, M. Behm, G. Lindbergh, and R. W. Lindström, common theme in the ongoing research projects is the
“Fast-charging to a partial state of charge in lithium-ion batteries: A mathematical modeling and electrochemical charac-
comparative ageing study,” J. Energy Storage, vol. 13, pp. 325–333, 2017. terization of electrochemical systems.
[30] S. Saxena, Y. Xing, D. Kwon, and M. Pecht, “Accelerated degradation
model for C-rate loading of lithium-ion batteries,” Int. J. Elect. Power
Energy Syst., vol. 107, pp. 438–445, 2019.
[31] F. Benavente-Araoz, M. Varini, A. Lundblad, S. Cabrera, and G. Lind-
bergh, “Effect of partial cycling of NCA/graphite cylindrical cells in Ivan Stenius received the M.Sc. degree in lightweight
different SoC intervals,” J. Electrochem. Soc., vol. 167, no. 4, 2020, structures from the KTH Royal Institute of Technol-
Art. no. 040529. ogy (KTH), Stockholm, Sweden, in 2003, and the
[32] Horizonfuelcell. Accessed: Jan. 2018. [Online]. Available: https://www. Doctor in Technology degree in lightweight structures
horizonfuelcell.com/h-series-stacks. on hydroelasticity and fluid structure interactions on
[33] Powercell. Accessed: Jan. 2018. [Online]. Available: https://www. high-speed craft from KTH in 2009.
powercell.se/en/products-and-services/fuel-cell-stacks/ He is currently a full-time Associate Professor with
[34] K. L. Davies and R. M. Moore, “UUV FCEPS technology assessment the Department of Aeronautics and Vehicle Engi-
and design process,” Hawaii Natural Energy Inst. (HNEI), School Ocean neering, KTH. He has expertise in underwater ve-
Earth Sci. Technol. (SOEST), Univ. Hawaii Manoa, HI, USA, Tech. Rep. hicle modeling design and construction, as well as
061027, 2006. maneuvering and navigation. He has been the lead
[35] em3ev. Accessed: Feb. 2018. [Online]. Available: https://em3ev.com/ developer of design and analysis software tools in collaboration with FMV and
product-category/ebike-battery-parts/ebike-battery-in-case/ The Swedish Coast Guard. In the recent years, he has been heavily involved in
[36] T. Reddy and D. Linden, Linden’s Handbook of Batteries. New York, NY, building up a research group in underwater technology and maritime robotics
USA: McGraw-Hill, 2010. with KTH and is currently a PI of the Swedish Maritime Robotics Centre hosted
[37] K. Ettihir, L. Boulon, and K. Agbossou, “Optimization-based energy by KTH that involves research on underwater robotics.
management strategy for a fuel cell/battery hybrid power system,” Appl.
Energy, vol. 163, pp. 142–153, 2016.
[38] L. Xu, J. Li, J. Hua, X. Li, and M. Ouyang, “Adaptive supervisory control
strategy of a fuel cell/battery-powered city bus,” J. Power Sources, vol. 194, Carina Lagergren received the Ph.D. degree in 1994.
no. 1, pp. 360–368, 2009. She is currently a Professor with the Department of
[39] G. Li, J. Zhang, and H. He, “Battery SoC constraint comparison for pre- Chemical Engineering, KTH Royal Institute of Tech-
dictive energy management of plug-in hybrid electric bus,” Appl. Energy, nology, Stockholm, Sweden. Her research focus is in
vol. 194, pp. 578–587, 2017. electrochemical systems and has mainly been within
[40] Z. Hong, Q. Li, Y. Han, W. Shang, Y. Zhu, and W. Chen, “An energy the fuel cell area, in which both high-temperature
management strategy based on dynamic power factor for fuel cell/battery molten carbonate fuel cells and polymer electrolyte
hybrid locomotive,” Int. J. Hydrogen Energy, vol. 43, no. 6, pp. 3261–3272, fuel cells, acid as well as alkaline, have been studied.
2018. Her research interest also includes electrolysis of,
[41] J. Han, J.-F. Charpentier, and T. Tang, “An energy management system e.g., water at high temperature for production of fuels.
of a fuel cell/battery hybrid boat,” Energies, vol. 7, no. 5, pp. 2799–2820,
2014.