This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE SYSTEMS JOURNAL 1

Optimization-Based Power and Energy Management

System in Shipboard Microgrid: A Review
Peilin Xie , Student Member, IEEE, Josep M. Guerrero , Fellow, IEEE, Sen Tan , Student Member, IEEE,
Najmeh Bazmohammadi , Juan C. Vasquez , Senior Member, IEEE, Mojtaba Mehrzadi,
and Yusuf Al-Turki , Senior Member, IEEE

Abstract—The increasing demands for reducing greenhouse sonars, and electromagnetic (EM) weapons may exceed the
emissions and improving fuel efficiency of marine transportation ship’s rated generation capacity, leading to unstable operation.
have presented opportunities for electric ships. Due to the complex-
The use of an energy storage system (ESS) can increase the
ity of multiple power resources coordination, varied propulsion
loads, changeable economical, and environmental requirements, flexibility in power allocation among the hybrid power sources,
power/energy management system (PMS/EMS) becomes essential offering the potential to improve efficiency and reduce emission
in both designing and operational processes. The existing literature during ship operations [2], [3]. Therefore, future ship power
on PMS/EMS can be categorized into rule-based and optimization- systems have a tendency to hybrid power ships, which include
based approaches. Compared to the rule-based PMS/EMS, which traditional gensets and various new equipment, such as, alterna-
relies heavily on human expertise, as well as predefined strategies
and priorities, the optimization-based approaches can offer more tive energy sources, fuel cell, gas capture system, ESSs, and
efficient solutions and are more widely used nowadays. This article so on. However, issues arise when different types of energy
provides a comprehensive review of the marine optimization-based sources work together. The complex power flow condition, the
power/energy management system and discusses the future trends requirement of coordination between multiple energy resources,
of PMS/EMS in ship power systems. and the potential in improving fuel efficiency and reducing
Index Terms—Energy management, optimization, power total costs make the PMS a necessity. PMS refers to a group
management, review, shipboard power system. of functions, scheduling algorithms, and control methods. It
determines the distribution of the power demanded between
different energy sources to promise continuous power supply
I. INTRODUCTION
for complex load conditions and to ensure a cost-effective,
HE transportation industry is currently the foundation of
T the national economy. Marine transportation takes 80%
of the world’s trade. Currently, diesel generators are still the
environmentally friendly, and reliable integrated energy system
with high efficiency. Furthermore, motivated by the need for
achieving a flexible shipboard arrangement and meeting the
major power source for all maritime applications. Due to the future on-board power demand, future ships will trend towards
widespread use of fossil fuels, marine fleet becomes a large all-electric ships. The electrification of ships, especially for the
contributor to greenhouse gasses (GHGs) and other emissions. propulsion electrification, has enlarged the total capacity of
As a result, there is a growing interest towards improving fuel the hybrid shipboard power system (SPS) and, thus, provides
efficiency and reducing the environmental footprint of the ma- a foundation for PMS to determine the ship economic and
rine vessels [1]. The use of high fuel-efficiency power sources environmental behaviors, which further complicate PMS but
such as fuel cells, and renewable energy sources (RESs) such also facilitate the study of it.
as wind, and solar energy would be opportunities to solve that Fig. 1 gives the overall structure of the SPS integrated with
problem. However, the presence of pulse loads, such as radars, a PMS/EMS. An SPS is typically powered by hybrid power
sources, which consist of traditional gensets (diesel generators),
RESs (PV panel, wind turbine, sea-wave energy), fuel cells,
Manuscript received June 28, 2020; revised October 14, 2020 and November and ESSs (batteries, fly wheels, UCs). And the electrical loads
30, 2020; accepted December 13, 2020. This work was supported in part by
VILLUM FONDEN, Center for Research on Microgrids, Aalborg University, of it have the characteristics of high dynamics, periodicity,
Denmark, under the VILLUM Investigator Grant 25920, and in part by the uncertainty, and high dependence on the marine environment,
Deanship of Scientific Research, King Abdulaziz University, Jeddah, Saudi which consists of propulsion load, ship service load, and pulsed
Arabia, under Grant 22-135-35-HiCi. (Corresponding author: Peilin Xie.)
Peilin Xie, Josep M. Guerrero, Sen Tan, Najmeh Bazmohammadi, Juan C. load. The PMS/EMS acts as a coordinator between the ship loads
Vasquez, and Mojtaba Mehrzadi are with the Center for Research on Microgrids, and power sources. Besides, An SPS also includes electronic
Department of Energy, Aalborg University, 9220 Aalborg, Denmark (e-mail: converters, transmission network, communication lines, and
[email protected]; [email protected]; [email protected]; [email protected]; [email protected];
[email protected]). so on.
Yusuf Al-Turki is with the Center of Research Excellence in Renewable En- An SPS can be considered as a typical mobile microgrid that
ergy and Power Systems, Department of Electrical and Computer Engineering, usually operates in islanded mode when the ship is at sea and in
Faculty of Engineering, King Abdulaziz University, Jeddah 21589, Saudi Arabia
(e-mail: [email protected]). grid-connected mode when it arrives to the seaport. Therefore,
Digital Object Identifier 10.1109/JSYST.2020.3047673 ship microgrids show some resemblances to terrestrial

1937-9234 © 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE SYSTEMS JOURNAL

However, it may not achieve an optimal solution. Furthermore,

they may require significant tuning efforts and may change
significantly for each topology [26]–[28]. On the contrary, the
optimization-based strategies rely on analytical or numerical
optimization algorithms, which can give optimal or subopti-
mal solution [29], and have drawn much attention from re-
searchers. Currently, a lot of research works on optimization-
based PMS/EMS have appeared. In this article, we provide an ex-
tensive review of PMS/EMS in SPSs, focusing on optimization-
based strategies especially.
This article is organized as follows. In Section II, the overall
targets and constraints used in the shipboard power management
strategies are summarized. In Sections III and IV, we discuss the
global planning based and real-time optimization based PMS
Fig. 1. Overall structure of PMS/EMS integrated SPS.
approaches, respectively, and several significant methods are
briefly studied. Conclusion and future trends are presented in
Section V.
microgrids such as the network architecture and increased use of
power electronic converters. Strategies developed for terrestrial II. OPTIMIZATION-BASED PMS APPROACHES
microgrids can be extended for maritime microgrids as well [4],
such as droop control, virtual synchronous generators [5], or An optimization problem refers to finding the minimum of a
other controller developed for electrical converters [6]–[9]. cost function with consideration of several constraints. Hence,
However, specific characteristics and challenges exist for it requires computational resources and data gathered from the
SPSs, such as the presence of high dynamic loads (planned whole power system to have a global view of the entire process.
or unexpected pulse loads, large share of propulsion loads, In this section, the optimization problem formulation will be
fluctuated loads from the sea), the requirement of economical discussed by introducing its two essential parts: objectives and
and environmental power sharing between the onboard gensets constraints.
and ESSs, the various operating scenarios (regular cruising,
full-speed sailing, docking, loading/unloading, anchoring, etc.), A. Objectives
and limited ship space for energy facilities installation. For any optimization methods, an objective function must
All these require further studies of PMS/EMS for maritime be defined, which usually takes into account environmental,
microgrids. technical, and economic aspects. The existing literature for the
Although a large number of reviews on SPSs can be found in PMS/EMS problem of SPSs are mainly focused on the optimal
the literature, the existing articles mostly emphasize on hybrid design of power plants and optimal power/energy scheduling
renewable energy systems [10], electric propulsion systems [2], and can be classified into the following problems.
SPS architectures [4], [11], [12], SPS stability and power qual- o1. Fuel Consumption Minimization: Due to the widespread
ity [12], [13], control technologies [4], [13], as well as the use of fossil fuels, the efficiency of diesel marine vessels has
coordination between SPS and seaport microgrids [14]. Rare of been one of the major concerns. In ship power generation, the
them focuses on a detailed overview of the PMS strategies [4], fuel consumption typically includes two parts, for generation
[11]–[13]. It is worth mentioning that, apart from PMS, many (F Cdg ) and for start-up (F Cst ) [30]
research works focus on EMS. The main difference between
a PMS and an EMS is that a PMS deals with the instanta- F C = F Cdg + F Cst . (1)
neous power flow, while an EMS focuses on the utilization and
planning of energy sources during a certain time period and is Start-up cost is generally considered as a constant value and only
always integrated with the future predictions and estimations dependent on the on/off status of the generator, and sometimes
systems. The targets of PMS are mainly about the enhancement ignored in some literature. While generation cost is affected by
of electrical reliability or availability, and EMS is more related a number of factors [31], [32]
to cost saving and energy efficiency. However, usually there is
not a clear line between PMS and EMS. In some cases, different 
T 
Ndg
F Cdg = (stij · SFOC(Pij , nij ) · Pij · ΔTj ) (2)
terminologies are used for the same control system. Therefore, j=0 i=1
this review does not make a detailed distinction between EMS
and PMS. where T is the generator operating time; Ndg is the number of
The existing PMS strategies of SPSs can be generally classi- diesel generators; stij is the switching status of the generator,
fied into optimization-based [15]–[18] and rule-based [19]–[24] using two-bit representation, 0 (switch-off) and 1 (switch-on);
methods. The rule-based strategies rely on human expertise, SFOC is the specific fuel consumption (grFuel/M W h), decided
predefined strategies, and priorities[25], which are easier for by engine output power Pij , and for variable-speed diesel gen-
implementation, and do not require high computation efforts. erator, it is also affected by engine speed nij .

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 3

From 2, it can be learned that for the purpose of minimizing For the purpose of evaluating the system cost, most literature
the fuel consumption, diesel generators are encouraged to work simply uses net present cost, which is the total sum of the
at their optimal fuel efficiency point, which can be achieved capital cost including all the operation and investment costs.
by an appropriate adjustment of generator output power, engine Such method can effectively evaluate the economic cost during
speed, and switching status. Besides, due to the utilization of a certain period of time in a simple way. However, we remark that
ESSs, it provides more opportunities, and challenges as well, for in a long time horizon, the annualized cost of the system (repre-
the flexible operation and optimal control of diesel generators senting the summation of the system capital cost per annum) and
in improving fuel efficiency. life cycle cost (representing the summation of all one-time and
It is also worth to be noted that, the fuel cell is another fuel con- recurring costs during the useful lifespan) are necessary [36].
sumer, attracting growing interests because of its advantages in And considering different types of ESS technologies, it is par-
high efficiency, small in size, and low environmental impact [33] ticularly important to assess the cost of the storage subsystem
and is usually used in hybrid with ESSs. To enhance the fuel cell per unit energy stored over the lifetime of the storage. In this
efficiency, optimal power split between the hybrid power sources way, it provides a fair comparison for the different capital costs
is the key point and has been studied a lot in vehicle area but and lifetimes of various ESS technologies [37].
few on marine applications [33], [34]. o4. Ship Equipment Weight and Size Optimization: The
o2. Environmental Footprint Reduction: Reducing environ- optimal equipment (ESS, DGs, engines, carbon capture sys-
mental footprint generally refers to reduce GHG emission [31], tem, etc.) sizes help to reduce footprint and improve SPS’s
which is commonly assumed to be proportional to the fuel survivability, and quality of service with a possible minimum
consumption and usually measured in grams of CO2 per kWh investment cost. The problem of finding the optimal size of an
of the consumed electricity. The conversion factor between fuel energy equipment requires optimal scheduling and dispatch for
consumption and CO2 emission can be found in the guidebook the whole power generation system and is always formulated
given by International Maritime Organization (IMO) [35]. And as a cost minimization problem. It involves a large number of
according to the IMO policy, energy efficiency design indicator variables, parameters, and specific information, for example, the
(EEDI) and energy efficiency operation indicator (EEOI) are energy supplied, the on/off time period of gensets, the number of
two key indicators that are commonly used to define ship GHG charging and discharging cycles, the SOC status of ESS, the ship
emissions during its lifetime operation. It should be noted that, efficiency factor EEOI, the foot space, the fuel oil consumption
due to the drawback of considering only one operation point, costs, as well as the input information such as load information,
EEDI can not accurately account the GHG emissions during ship operating conditions, ship cruising, and voyage times.
ship lifetime, which makes EEOI more widely applied in ship It should be mentioned that, typically, ship sizing problem
PMS/EMS designing. is formulated in conjunction with other objectives, such as
EEOI is the ratio of CO2 mass emitted per unit of transport minimization of GHG emission, total cost or footprints, etc. Due
work, representing ship operational efficiency indirectly, and it to the inherent conflict between them, the tradeoff between the
is defined as multiobjective has to be addressed carefully with respect to ship
 load conditions, electrical, and mechanical constraints [38].
(F Ck · CFk ) o5. Maximization of the Endurance of Navigation: It refers to
EEOI = k (3)
mcargo · D increase the cruising capacity of the ship, which mainly depends
on the battery state of charge and the remaining fuel volume of
where k is the fuel type, F Ck is the mass of consumed fuel, the diesel engines [39]. However, not much literature has been
CFk is the conversion factor transferring fuel mass to CO2 mass, found in solving this problem.
mcargo is cargo tonnes, and D is the distance in nautical miles. It is worth to be noted that, instead of considering only
Equation (3) maintains a good balance between ship opera- one objective, multiobjective optimization has been drawing
tional efficiency and GHG emission. To improve the ship opera- much attention by considering many conflicting objectives while
tional efficiency, several factors can be optimized, for example, taking into account their priorities. For example, o2 and o3, o3
ship routing and scheduling, vessel speed, generator operation and o4, and o1 and o5 are commonly considered as conflicting
time, and the tradeoff between carbon emission and investment and have been investigated in many studies.
cost.
o3. Economic Investment Minimization: It refers to the finan-
cial expenditures associated with the management of energy, B. Constraints
which mainly includes operation costs and investment costs. In many practical problems, design variables must satisfy
Ship operation costs are the expenses for running gensets, ESSs, certain specified electrical or physical requirements and so may
or other assets, consisting of fuel cost, maintenance cost, start- not take arbitrary values. These restrictions are referred to as
up/shut-down cost, etc. Generally, it is estimated as functions of design constraints and are key to ensure system stability and
the produced power (electrical, thermal, or cooling), running safety. The most common constraints are listed as follows.
time, online gensets numbers, and capacity. Ship investment c1. Power and Energy Balance: Generally, ship electric
costs are composed of the installation and replacement of the loads comprise of propulsion loads Lprop , service loads Lservice ,
ESS, typically decided by economic specific values, installed auxiliary load Lauxi , and pulse loads Lpulse from high power
size, capacity, and life-span of gensets. mission. The total output power of ship power resources (DGs,

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE SYSTEMS JOURNAL

TABLE I
OBJECTIVES CONSIDERED IN SHIPBOARD PMS/EMS LITERATURE

ESS, fuel cell, RESs, etc.) should meet the electrical loads at Fig. 2. Ratios of various constraints.
any time interval. So it is formulated as an equality constraint

Ndg priority regardless of the optimization objectives, followed by
stij · Pij + PESS,j = Lservice + Lprop + Lpulse + Lauxi . restraints for onboard gensets and ESSs. These ensure the stable
i=1 operation of the whole power system. Although there is no
(4) mandatory policy requirement for the GHG emissions, we can
c2. Restraints for Power Quality: It refers to the maintenance see that a large number of papers voluntarily take into account the
of the voltage and frequency stability of the electrical power sys- constraint of ship efficiency factors in the context of maintaining
tem. To ensure stable operation, voltage and frequency variations normal marine navigation. These aforementioned constraints
should be limited. Commonly, the voltage variation is limited broadly cover several aspects of the study and provide a general
under ±5% and the frequency variation is no higher than ±3%. direction to guide future research. The specific constraints used
c3. Restraints of Power Plants: To avoid unsafety and me- in each study depend on the researchers’ goals and working
chanical damages, restraints on generators must be addressed, conditions.
such as generator loading constraints, generator ramp rate con- This section reviewed the general optimization objectives
straint, generator operation, and out of operation time con- and constraints used in the previous studies. To solve the
straints. To avoid system blackout, the maximum allowable power/energy management problem, optimization techniques
continuous loading of the generators should be defined. Similar are reviewed in the two following sections. The general clas-
constraints also hold for other types of power resources (RESs, sification principle in this article is whether the method requires
fuel cell, etc.). In addition, to compensate for power shortages entire system information, has the ability to acquire the optimal
or frequency drops within a given period of time, one necessary solution, and can be applied in real-time or not. From this
way is to keep active spinning reserves capacity at a certain level, perspective, the optimization-basde PMS/EMS are categorized
mainly by maintaining the total unused capacities of DGs and into global planning methods and real-time methods.
the ESS.
c4. Restraints of ESS: It mainly refers to the limits of ESS III. GLOBAL PLANNING
energy capacity, charging and discharging power, current state of
charge (SOC), and the maximum depth of discharge (DoDmax ), Global planning strategies minimize the cost functions using
formulated as inequality constraints. the knowledge of future and past information while considering
c5. Environmental Constraints: GHG emissions should be the aforementioned objectives. Table I gives a summary of
monitored online and kept below a certain upper limit. Ac- recent studies categorized by PMS/EMS objective functions
cording to IMO, EEDI, and EEOI require a minimum energy while Table II presents an overview of the commonly used
efficiency level (CO2 emissions) for different ship type, size optimization methods. Nowadays, there exist many optimization
segments, voyage, and transport mode. Specific data can be methodologies falling into this category. They can be generally
found in [40]. classified into the classical optimization methods and heuristic
c6. Ship Voyage Constraints: Including ship speed constraint, algorithms-based methods.
which should be bounded in a specific region; limits of total
A. Classical Optimization Methods
traveled distance, which should be almost equal to the total
route distance at the end of optimization process and is always Classical optimization methods take advantage of the analyti-
represented by the deviation of the actual traveled distance from cal properties of the problem to generate a sequence of points that
the scheduled one; limits on the quantity carried by the ships, converge to a global optimal solution. Dynamic programming,
and so on. linear programming, and mixed-integer programming fall into
c7. Constraints for the Auxiliary System: For example, heat this category.
balance, limitation of heat losses, and temperature. 1) Dynamic Programming: Dynamic programming (DP) is
After searching the relevant literature on shipboard one of the most efficient mathematical techniques, capable of
optimization-based PMS/EMS since 2016, the ratios of these solving multiple optimization problems, and can always gen-
constraints and research objectives are summarized in Fig. 2 erate the most fuel-efficient results while dealing with other
and Table I, respectively. Concluded from Fig. 2 that, the power energy management problems. It systematically evaluates a large
balance constraint is the most important one with the highest number of possible decisions in a multistep problem, and make

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 5

TABLE II
SUMMARY OF GLOBAL PLANNING BASED PMS/EMS REFERENCES IN SPSS

a wise decision in each step, minimizing the total cost for all problems [90]. For example in [32], a two-step multiobjective
decisions made. DP requires a mathematical model where deci- optimization method for hybrid ESS management is established
sion variables, parameters, and constraints are clearly defined, with the aim of minimizing battery life cycle degradation. And
and the whole power demand is known as a prior. by applying the piecewise linearization, the NLP problem is
With the assumption that the ship load forecasting is available, solved by Gurobi optimizer. However, linearization might result
Bukar and Tan [30] used DP to solve the optimization problem in in increasing computational complexity or failure in capturing
its demand-side optimal power management system. By doing dynamic responses, especially when the NLP problems are
that, reduction in the operation cost and GHG emission can be sequential, parallel, or multiobjective. Due to the quadratic
achieved. However, ESSs and investment capital are not consid- relationships between system states, for example, the fuel con-
ered. Considering the computing complexity induced by energy sumption of diesel generators can be approximately represented
storage and propulsion power adjustment means, Tjandra et al. by a quadratic function of produced power [78], [91], and
[41] separated the optimization problem into three stages and the quadratic cost functions in multiobjective optimization,
uses DP to deal with the first two stages. The same assumption quadratic programming (QP), a particular and one of the simple
about the ship load is made in [41]. It should be noted that types of NLP, is the most commonly used algorithm in real-
although DP can handle these complicated problems in some application optimization problems. In addition, as a particular
cases, it may cost huge computation efforts and may fail to deal case of convex optimization, algorithms such as the alternating
with the situation when there are time coupling constraints, such direction method of multipliers (ADMM) can solve the QP
as the coupling startup state and running state of devices [86]. problem with contradicting objectives [92].
It also requires prior knowledge of the ship power demand, 3) Mixed Integer Programming: It is worth mentioning that,
which is not always possible in real applications. Thus, it is in real-world applications, some or all of the variables are
not normally suitable for real-time control but can be used as a constrained to be binaries or integers. Such problems are called
benchmark for improving other strategies [87]. integer programming problems that can be categorized into
2) Linear and Nonlinear Programming: Linear programming mixed-integer linear programming (MILP) [65]–[67], [81] and
(LP) is the most simple form of classical optimization methods, mixed-integer nonlinear programming (MINLP) [56], [66], [78].
which captures the first-order effects of various system parame- The main reasons that result in a mix-integer problem are
ters that should be optimized. LP has been widely used in various the quantities that can only be integers, e.g., number of DGs,
engineering disciplines, such as flow control, power manage- and batteries [93], or the quantities that represent decisions,
ment, and so on [88], [89]. However, due to the complexity of e.g., switching status of the diesel generators, charging and
the SPS and the nonlinear inherent of it, LP is rarely used in the discharging of ESSs, or ship speed [66], [67], [69], [81], [82],
marine field and most of the works are based on its derivative [84]. The objective function can be generally classified into a
methods or nonlinear programming (NLP) approaches. single objective of minimizing the fuel cost [69], [93], overall
To solve a nonlinear programming problem, one efficient operation cost [66], [68], or finding the optimal selection, sizing,
way is using linearization to transform it into a series of linear and management of the ESS [67], [81], [82], [84], and the

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE SYSTEMS JOURNAL

multiobjective of optimizing the ship operating cost and gas 2) Particle Swarm Optimization Algorithm: Particle swarm
emissions simultaneously [32], [53], [55], [56], [64]. optimization (PSO) algorithm’s precursor was a simulator of
The most common ways to solve the NLP and MINLP prob- social behavior, that was used to visualize the movement of a
lems are either converting the problem into an LP problem by birds’ flock, first introduced in 1995. Since then, several vari-
using linearization methods or incorporating nonlinear solvers ants of PSO have been developed while incorporating concepts
such as GUROBI, CPLEX, and normal boundary intersection such as nearest-neighbor velocity matching and acceleration by
method as used in [53]. distance. Comparing to other heuristic algorithms, PSO has the
4) Other Methodologies: Apart from the aforementioned opti- benefits of fast convergence, good computational efficiency, and
mization methods, there are many other methods that can be used relatively robust operation in locating global optima, while it can
in the power management system of SPSs, such as interval opti- solve highly nonlinear and complex problems.
mization methodology [70]. The classical algorithms can obtain Traditional PSO has been widely used to solve the optimiza-
the optimal solution within the desired tolerance [94]. However, tion problem in SPS. A PSO-based EMS is developed in [48]
owing to the inherent complexity, they may result in large to reduce fuel consumption by optimizing ship voyage based on
computational efforts and is considered to be NP-Hard in nature, real operation profile. The performance of PSO is affected by
especially for large-scale systems. As a result, researchers have a set of parameters described as inertia weight and acceleration
to seek alternative algorithms to make a satisfactory compromise factors. For the purpose of achieving a higher optimization per-
between the optimality of the solution and the computational formance, researchers have suggested to develop a self-adaptive
burden. In [68], for example, a metaheuristics method is utilized mechanism to tune these parameters along with the algorithm
to solve the MINLP problem. iterations. For example in [61], a fuzzy-based PSO is proposed
to improve the computational efficiency of the algorithm and
tackle with antagonistic optimization objectives. The proposed
B. Heuristic Algorithm method can significantly reduce the operation cost and footprint
The complexity of some optimization problems that cannot while satisfying the technical and operational constraints of
be tackled by exact mathematical methods has suggested using the ship. However, in some cases, PSO suffers from immature
heuristic algorithms to explore more quickly the solution space, convergence or entrapment to a local solution. To deal with this
even finding a suboptimal solution. In heuristic optimization issue, an advanced version of PSO named PSO of composite
strategies, logical rules determine the operating mode of the particle (PSO-CP) was developed initially in [97] and used
plant and the setting for the battery charging and discharging in [50] and [68]. Although PSO and its variants have been
system. However, it has to be noted that the probability of finding proven to be effective and computationally efficient in many
the global solution decreases when the problem size increase. applications. These methods might not be very efficient when
1) Genetic Algorithm: Genetic algorithm (GA) is a search applied to a dynamic system. Thus, adaptive PSO, which allows
algorithm based on the mechanism of natural selection and randomizing particles and dynamically changing parameters
natural genetics, which dates back to earlier than 1975. GA has been introduced to compensate for this issue. In [74], an
comprises three processes: selection, crossover, and mutation. adaptive multicontext cooperative coevolving PSO is employed
During these processes, parents with higher fitness values are to dispatch power flows with great electricity cost savings. PSO
more likely to be chosen and the offsprings are more likely to can also be used to deal with the multiobjective optimization
be similar to these parents. Hence, GA tends to converge to problems along with sorting technique based on NSGA-II [15],
suboptimal solutions and cannot produce all potential solutions. [47]. Optimal results can be picked up depending on the user’s
However, due to the advantages in solving complex nonlinear preferences.
optimization problems and promising more accurate exploration 3) Other Methodologies: Apart from the above-mentioned
of the solution space than other gradient-based heuristic al- most commonly used algorithms, other methods relying on
gorithms, GA has been widely used in SPSs. Some examples heuristic algorithms have been used to solve the power man-
are optimal sizing of generation system [83], power demand agement optimization problems such as evolutionary algo-
prediction and allocation [43], [44], [72]. rithm [98], differential evolution algorithm [77], Pareto concav-
In other examples to solve multiobjective PMS problems, ity elimination transformation [99], and Grey Wolf optimiza-
Nondominated Sorting Genetic Algorithm II (NSGA-II) has tion [51].
been utilized because of its ability to maintain a good spread
of solutions and converge near the true Pareto-optimal set [95].
As in [15], [47], [58], and [59], conflicting objectives of C. Discussion
reducing GHG emission and overall operational cost are con- In general, classical optimization methods are suitable for
sidered. Some additional objectives, which are included in the problems with small numbers of variables. The results obtained
cost function are optimizing ESS size, minimization of fuel can guarantee to be optimal and can be used as a benchmark
consumption [39], [45], [46], and minimization of the overall for other methods. Compared to classic optimization methods,
mechanical losses [96]. In addition, to enhance the performance heuristic algorithms are most often employed when approximate
of the local optimization of NSGA-II in the population diversity solutions are sufficient and exact solutions are computationally
preservation, an improved NSGA-II algorithm was adopted expensive [100]. Generally, to ensure the optimum and com-
in [39] by replacing the polynomial variation with difference putational efficiency at the same time, heuristic algorithms are
mutation operators. usually supplemented with classic algorithms.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 7

TABLE III ECMS has been recently proposed to automatically tune the ef
SUMMARY OF REAL-TIME OPTIMIZATION-BASED PMS/EMS
REFERENCES IN SPS
according to the current or predictive information, which in par-
ticular can be categorized into two groups including adaptation
based on driving cycle prediction [17], [103] and adaptation
based on driving pattern recognition. The former predicts the
future information to calculate the most appropriate ef , while
the latter assumes that ef are similar for cycles with similar
statistical properties and selects the most suitable ef from the
predefined set after defining the driving pattern. For ships with
fixed-cycle operational modes and on a fixed sailing course, the
latter will largely simplify the optimization process and, thus,
require less computational space. However, most literature on
driving pattern recognition remaining concentrate on electric
IV. REAL-TIME OPTIMIZATION vehicles, relevant works on SPS has yet to be published.
It should be noted that more accurate prior knowledge of the
Although global planning strategies are capable of sorting marine load profile can guarantee better final results. Given the
global optimum solutions, they may cost too much computa- full ship routine, ECMS has been shown effective in achieving
tional effort and require prior knowledge of entire ship routine a good fuel efficiency that is close to the global optimal solution
information, which makes it capable of designing, sizing, voyage from dynamic programming [104]. However, in cases that the
scheduling, and energy dispatch in early-stage management, load profile is not a priori known, an accurate load prediction
but unsuitable for real-time power management. Real-time op- system will be required to be integrated with ECMS to achieve
timization (RTO) refers to the techniques, which allow the better results. Hence, there would be a tradeoff between the
continuous evaluation and manipulation of the process operating computational effort and the optimality of the final solution,
conditions considering the most recent information to minimize which makes ECMS a method that deserves further research.
the desired cost function. The direct benefit of RTO is providing
the system with real-time optimized commands to reach the B. Model Predictive Control
desired operation following the currently available information.
RTO strategies consist of equivalent consumption minimiza- MPC is a promising optimal control, which has been proven to
tion strategy (ECMS), model predictive control (MPC), se- be efficient and robust for dynamic systems. It enables the PMS
quential function method (SFM), and so on. Table III gives a to look ahead as far as the established prediction horizon and gen-
summary of the literature on real-time PMS/EMS optimization erate a future sequence of control inputs (u(k), u(k + 1), u(k +
methodologies. 2), . . . , u(k + Nc )), which optimizes a predefined cost function
L, while meeting all the electrical and physical constraints. Basic
MPC scheme takes the following procedure:
A. Equivalent Consumption Minimization Strategy
1) measure the state x(k) of the system;
ECMS is the most well-known technique for instantaneous 2) solve the optimization problem
optimization, which has been widely researched for land-based
vehicles and has proven to be effective in improving fuel econ- 
Np
min J(u(·)) = L(x(k), u(k)) (7)
omy. Motivated by that, ECMS is investigated for electrical ships
k=1
recently. ECMS formulates a cost function as a sum of the real
fuel consumption and equivalent fuel consumption related to the with respect to
energy storage SOC variation. The instantaneous minimization u(·) ∈ U N (8)
problem will be solved at each instant only using arguments
based on actual energy flow x(k + Np ) = f (x(k), u(k), u(k + 1), . . . , u(k + Nc ))
(9)
min C = Cf (t) + Cess (t) (5)
subject to
Pess (t)
Cess (t) = ef (t) · (6) h(x(k)) = 0
QLHV
g(x(k)) ≥ 0 (10)
where Cf (t) represents the real fuel consumption.
QLHV is the fuel lower hearing value (energy content per unit where Np is the prediction horizon and Nc is the control
of mass) and considered as a constant. From 6, the equivalent fuel horizon.
consumption of ESS Cess (t) depends on its output power Pess and 3) Compute the control input value in the next sampling
equivalence factor ef . ef converts electrical power consumption period. MPC takes the form of receding horizon control
into fuel consumption and needs to be chosen carefully. procedure, which means although the optimal trajectory
In conventional ECMS, ef is assumed to be constant [87], of the future control signal is completely described within
[101], [102]. Although this assumption will ease the real-time the moving horizon window, only the first sample of the
implementation, it might fail to capture the real transformation control inputs are utilized while neglecting the rest of the
relationship and result in unsatisfactory performance. Adaptive samples.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

8 IEEE SYSTEMS JOURNAL

The main advantage of MPC is that it optimizes the cur- in large SPSs. In a distributed control scheme, each energy
rent timeslot while taking future timeslots into account, which source sends signals to the local controller. Local controllers
makes it adapt well to the high dynamic system. To solve communicate with each other to make appropriate decisions for
the optimization problem at each control step, it is essential global optimizing. In that case, it can significantly reduce the
to choose a proper method that balances the needs for fast computational burden on each local controller without any risk
convergence and optimality of the results. Efficient numerical of single point failure.
algorithms have been proposed to address challenges in the Although there has been a lot of research in both energy
real-time implementation of MPC, such as NLP [105], and management and distributed control, the union of these two
LP [106]. Heuristic algorithms mentioned in Section III can ideas has not reached too much fruition so far. The prior-art
be also used to solve optimization problems in MPC, such as in the distributed ship PMS mainly focuses on MAS, distributed
PSO [107], [108]. Since most optimization problems of MPC MPC (DMPC), and ADMM.
are in quadratic form, quadratic programming becomes the most MAS is inspired by the biological phenomena, which aims
commonly used method [18], [109]–[113]. Typically, sequential to achieve system objectives cooperatively that are difficult to
quadratic programming (SQP), an iterative procedure, is the reach by a single agent or a centralized controller. By dividing
most well-known gradient-based method that has been found a single optimization problem into several sub-problems (i.e.,
wide applications in MPC. SQP solves the QP subproblems being solved individually by every single agent), MAS tech-
and uses the solution to construct a new iterate for every it- nologies have great potential for real-time power management
eration [114]. To improve computational efficiency, integrated in large-scale systems. In [91], a distributed MAS-based PMS
perturbation analysis and SQP (IPA-SQP) approach has been is proposed for the optimal power-sharing with minimum dis-
utilized in [115]–[118]. IPA-SQP achieves optimal solutions for tribution losses and economic dispatch of resources. Multiple
each MPC sampling instant by combining perturbation anal- agents take part in the process of solving the MIQP problem
ysis (i.e., providing closed-form solutions when some of the by solving a subset of the resulting search tree. In [126], the
parameters are changed) together with SPQ (i.e., achieving local optimization problem is solved by using a real-time distributed
optimality). PSO methodology for achieving the optimal distributed power
Apart from the traditional MPC method, advanced method- controllers’ parameters. In [127], a real-time heterogeneous
ologies can be integrated with MPC for more precise load MAS-based load management strategy is proposed to achieve
prediction. To efficiently handle the uncertain pulse-power load dynamic generation and load balancing for dc zonal SPSs. Ar-
condition, integration of MPC with the auto regressive integrated tificial potential function is integrated into the MAS framework
moving average model is one of the effective methods for to coordinate various electrical elements to achieve group goals
uncertain pulse load forecasting, which largely improves the and improve the dynamic behavior of the system.
system overall performance under high power pulsed load condi- Apart from the MAS, ADMM has been widely used in other
tion [119]. In order to estimate the propulsion-load torque, [120] power system applications as it can break convex optimization
integrates adaptive parameter identification approach with MPC problems into smaller pieces, each of which is easier to be
to formulate an AMPC and uses linear prediction for future handled. In the SPS, Lai and Illindala [128] developed a modified
load information prediction, which achieves much better per- nested EMS based on the ADMM to obtain the solution strategy
formance in terms of improved system efficiency, enhanced with contradicting objectives. Edrington et al. [129] and Vu et al.
reliability, improved thrust production, and reduced mechanical [130] combined the ADMM with MPC to formulate a DMPC
wear and tear. Besides, to improve the system stability in case of for real-time power management. DMPC can not only inherit
fault occurrence, fault scenario-based MPC is developed in [121] the advantage of explicit accommodating of constraints and
to predict the system states for fault-free operation. good optimization performance but also has the advantages of
It is worth to be noticed that, while MPC is suitable for a distributed framework, namely flexibility and error tolerance,
real-time control, a key drawback is the difficulty in making which can support plug-and-play operation.
a satisfactory balance between real-time computing efficiency Although there has been a small amount of research working
and ensuring long-term optimality. This motivates to seek for al- on the distributed PMS for the SPSs, especially for ships on
ternative approaches, such as hierarchical control strategy [122], zonal electrical distribution (ZED), most studies remain focused
[123], multiagent system (MAS), artificial intelligence algo- on centralized power management strategies. Current studies on
rithm [76], [124], sensitivity function method [125], recursive distributed ship PMS mainly concentrate on the maximum load
searching algorithm [98], deploying improved solver, and mul- supplement with limited energy capacity [127], [130], [131],
ticore hardware [116]. maintaining a high SOC of ESSs [129] and reducing the system
operational cost [91], [92], [130]. Little of them considers the
factors that have been studied a lot in the centralized SPS, such
C. Distributed Real-Time PMS for Large-Scale SPS as the GHG emission, the uncertainties brought by sensor failure
The aforementioned PMS strategy is in the centralized ar- or environmental variation, and the negative effects from pulsed
rangement. Thanks to the knowledge of the entire system, it loads and fluctuated loads. However, with the emerging of ZED
can converge to the global optimum solutions. However, for systems, ZED-based SPSs are drawing more interest because of
large-scale ship power management, the centralized PMS might their advantages in lower acquisition costs, lower weight, and
impose a high computational burden and suffer from single point better operational flexibility. This will motivate further research
failures, which allows the distributed PMS to be further studied and application of distributed PMSs.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 9

complexity of the energy management systems. As a

result, a robust energy management method combining
the real-time control strategy with the forecasting method
is necessary to provide good performance even when
the forecast data deviates significantly from the real val-
ues [141].
3) Only a few studies have been conducted on the shipboard
distributed PMS in previous literature. Most studies focus
on the centralized arrangement. Therefore, many chal-
lenges should be resolved, such as system uncertainties,
communication delays, and the tradeoff between compu-
tational efficiency and system complexity.
4) Future EMS/PMS should have the capability of tackling
Fig. 3. Recommendations to future PMS/EMS.
not only the ship operational issues but also offering
comfortable service to customers. Human intention feed-
forward control is recommended to be integrated into the
V. CONCLUSION AND FUTURE TRENDS PMS [142], so that on-demand monitoring and customers
A. Conclusion commands could be allowed to more actively interact with
power/energy management.
The increasing of global environmental concerns has boosted
the development of the electrical ships. In that case, there is an
REFERENCES
urgent to improve the power/energy management system per-
formance in terms of energy saving, environmentally friendly, [1] P. Nema, R. Nema, and S. Rangnekar, “A current and future state of
art development of hybrid energy system using wind and pv-solar: A.
safe, and economic operation. review,” Renewable Sustain. Energy Rev., vol. 13, no. 8, pp. 2096–2103,
In this article, methods employed in the literature for 2009.
optimization-based shipboard power/energy management have [2] R. Geertsma, R. Negenborn, K. Visser, and J. Hopman, “Design and
control of hybrid power and propulsion systems for smart ships: A. review
been reviewed. The main strategies of global planning based of developments,” Appl. Energy, vol. 194, pp. 30–54, 2017.
PMS and real-time optimization-based PMS for SPSs have been [3] S. Tan, Y. Wu, P. Xie, J. M. Guerrero, J. C. Vasquez, and A. Abusorrah,
discussed. It is clear that the former requires full knowledge of “New challenges in the design of microgrid systems: Communication
networks, cyberattacks, and resilience,” IEEE Electrific. Mag., vol. 8,
the system information or accurate load prediction to achieve the no. 4, pp. 98–106, Dec. 2020.
global optimum solution. It is more suitable for early-stage de- [4] M. D. Al-Falahi, T. Tarasiuk, S. G. Jayasinghe, Z. Jin, H. Enshaei, and
signing, ship routine scheduling, and energy dispatch problems J. M. Guerrero, “AC ship microgrids: Control and power management
optimization,” Energies, vol. 11, no. 6, p. 1458, 2018.
with considerations of designer’s control targets and constraints. [5] P. Xie et al., “Stability analysis considering dual physical constraints of
While the latter uses the instantaneous measured data or short- parallel-connected virtual synchronous generators forming microgrids,”
term load forecasting results and provides real-time guidance for in Proc. IEEE Energy Convers. Congr. Expo., 2019, pp. 2092–2098.
[6] B. Wei, A. Marzbal, J. Perez, R. Pinyol, J. M. Guerrero, and J. C.
enhancing system performance. Thus, it is capable of dealing Vasquez, “Overload and short-circuit protection strategy for voltage
with uncertainty, pulsed loads, and long-term disturbances, and source inverter-based ups,” IEEE Trans. Power Electron., vol. 34, no. 11,
mainly used for real-time power splitting, generator start/stop pp. 11 371–11 382, Nov. 2019.
[7] B. Wei, Y. Gui, S. Trujillo, J. M. Guerrero, J. C. Vasquez, and A. Marzbal,
controlling. protection, and etc. Apart from centralized PMS, “Distributed average integral secondary control for modular ups systems-
distributed PMS is gaining more interest because of the safety based microgrids,” IEEE Trans. Power Electron., vol. 34, no. 7, pp. 6922–
and computational efficiency requirements for large-scale SPSs. 6936, Jul. 2019.
[8] B. Wei, A. Marzbal, R. Ruiz, J. M. Guerrero, and J. C. Vasquez, “Davic: A.
new distributed adaptive virtual impedance control for parallel-connected
B. Potential Research Trends voltage source inverters in modular ups system,” IEEE Trans. Power
Electron., vol. 34, no. 6, pp. 5953–5968, Jun. 2019.
In the future, works recommended to improve the perfor- [9] B. Wei, J. M. Guerrero, J. C. Vasquez, and X. Guo, “A circulating-current
mance of PMS/EMS in SPSs are listed in the following. The suppression method for parallel-connected voltage-source inverters with
common dc and ac buses,” IEEE Trans. Ind Appl., vol. 53, no. 4, pp. 3758–
suggested studies are exhibited in Fig. 3 to build a future 3769, Jul./Aug. 2017.
EMS/PMS structure as follows. [10] P. Nema, R. Nema, and S. Rangnekar, “A current and future state of
1) Given the limited computational capacities and the re- art development of hybrid energy system using wind and PV-solar: A.
review,” Renewable Sustain. Energy Rev., vol. 13, no. 8, pp. 2096–2103,
quirement of real-time scheduling for the shipboard 2009.
PMS/EMS, it is recommended that the measures of [11] J. Thongam, M. Tarbouchi, A. Okou, D. Bouchard, and R. Beguenane,
complexity and flexibility should be included in the “All-electric ships-a review of the present state of the art,” in Proc. IEEE,
8th Int. Conf. Exhib. Ecological Veh. Renewable Energies, 2013, pp. 1–8.
PMS/EMS. Thus, it can provide a fair comparison of the [12] S. G. Jayasinghe, L. Meegahapola, N. Fernando, Z. Jin, and J. M.
tradeoff between optimal scheduling and computational Guerrero, “Review of ship microgrids: System architectures, storage
considerations. technologies and power quality aspects,” Inventions, vol. 2, no. 1, p. 4,
2017.
2) The uncertainties imposed by RESs, propulsion loads, [13] K. Ni, Y. Hu, and X. Li, “An overview of design, control, power manage-
variable sea conditions, and communication failures might ment, system stability and reliability in electric ships,” Power Electron.
result in large deviation of the forecast data and increase Drives, vol. 2, no. 2, pp. 5–29, 2017.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

10 IEEE SYSTEMS JOURNAL

[14] S. Fang, Y. Wang, B. Gou, and Y. Xu, “Toward future green maritime [36] A. L. Bukar and C. W. Tan, “A review on stand-alone photovoltaic-wind
transportation: An overview of seaport microgrids and all-electric ships,” energy system with fuel cell: System optimization and energy manage-
IEEE Trans. Veh. Technol., vol. 69, no. 1, pp. 207–219, Jan. 2019. ment strategy,” J. Cleaner Prod., vol. 221, pp. 73–88, 2019.
[15] Y. Huang, H. Lan, Y.-Y. Hong, S. Wen, and S. Fang, “Joint voyage [37] E. Eriksson and E. M. Gray, “Optimization and integration of hybrid
scheduling and economic dispatch for all-electric ships with virtual renewable energy hydrogen fuel cell energy systems-a critical review,”
energy storage systems,” Energy, vol. 190, 2020, Art. no. 116268. Appl. Energy, vol. 202, pp. 348–364, 2017.
[16] R. Derollepot and E. Vinot, “Sizing of a combined series-parallel hybrid [38] C. Yan, G. K. Venayagamoorthy, and K. A. Corzine, “Optimal location
architecture for river ship application using genetic algorithm and optimal and sizing of energy storage modules for a smart electric ship power
energy management,” Math. Comput. Simul., vol. 158, pp. 248–263, system,” in Proc. IEEE Symp. Comput. Intell. Appl. Smart Grid, 2011,
2019. pp. 1–8.
[17] M. Kalikatzarakis, R. Geertsma, E. Boonen, K. Visser, and R. Negenborn, [39] D. Gao, X. Wang, T. Wang, Y. Wang, and X. Xu, “An energy optimization
“Ship energy management for hybrid propulsion and power supply with strategy for hybrid power ships under load uncertainty based on load
shore charging,” Control Eng. Pract., vol. 76, pp. 133–154, 2018. power prediction and improved NSGA-II algorithm,” Energies, vol. 11,
[18] T. Van Vu, D. Gonsoulin, F. Diaz, C. S. Edrington, and T. El-Mezyani, no. 7, p. 1699, 2018.
“Predictive control for energy management in ship power systems under [40] C. S. Index, “Clean Shipping Index Methodology and Reporting Guide-
high-power ramp rate loads,” IEEE Trans. Energy Convers., vol. 32, no. 2, lines.,” Gothenburg, Sweden: Clean Shipping index, 2018.
pp. 788–797, Jun. 2017. [41] F. Kanellos, “Optimal power management with GHG emissions lim-
[19] Y. Guo, M. Khan, M. Faruque, and K. Sun, “Fuzzy logic based energy itation in all-electric ship power systems comprising energy storage
storage supervision and control strategy for MVdc power system of all systems,” IEEE Trans. Power Syst., vol. 29, no. 1, pp. 330–339, Jan. 2013.
electric ship,” in Proc. IEEE Pow. and Energy Soc. Gen. Meeting, 2016, [42] F. Baldi, F. Ahlgren, F. Melino, C. Gabrielii, and K. Andersson, “Optimal
pp. 1–5. load allocation of complex ship power plants,” Energy Convers. Manage.,
[20] M. M. S. Khan, M. O. Faruque, and A. Newaz, “Fuzzy logic based energy vol. 124, pp. 344–356, 2016.
storage management system for MVdc power system of all electric ship,” [43] H.-M. Chin, C.-L. Su, and C.-H. Liao, “Estimating power pump loads
IEEE Trans. Energy Convers., vol. 32, no. 2, pp. 798–809, Jun. 2017. and sizing generators for ship electrical load analysis,” IEEE Trans. Ind.
[21] Y. Yuan, T. Zhang, B. Shen, X. Yan, and T. Long, “A fuzzy logic energy Appl., vol. 52, no. 6, pp. 4619–4627, Dec. 2016.
management strategy for a photovoltaic/diesel/battery hybrid ship based [44] M. A. Ancona et al., “Efficiency improvement on a cruise ship: Load
on experimental database,” Energies, vol. 11, no. 9, p. 2211, 2018. allocation optimization,” Energy Convers. Manage., vol. 164, pp. 42–58,
[22] S. Faddel, A. A. Saad, M. El Hariri, and O. A. Mohammed, “Coordination 2018.
of hybrid energy storage for ship power systems with pulsed loads,” IEEE [45] J. J. Valera-García and I. Atutxa-Lekue, “On the optimal design of hybrid-
Trans. Ind. Appl., vol. 56, no. 2, pp. 1136–1145, Mar./Apr. 2020. electric power systems for offshore vessels,” IEEE Trans. Transport.
[23] E. B. Beşikçi, T. Kececi, O. Arslan, and O. Turan, “An application of Electrific., vol. 5, no. 1, pp. 324–334, Mar. 2019.
fuzzy-AHP to ship operational energy efficiency measures,” Ocean Eng., [46] C. Zhang and B. Jia, “Research of power allocation in diesel-electric
vol. 121, pp. 392–402, 2016. hybrid propulsion system,” in Proc. IEEE Chinese Autom. Congr., 2019,
[24] N. Karim, R. Lisner, H. Kazemi, and F. Annaz, “Rule-based power pp. 3664–3668.
management for the all-electric ship,” in Proc. Australian Univ. Power [47] R. Tjandra, S. Wen, D. Zhou, and Y. Tang, “Optimal sizing of BESS for
Eng. Conf., Citeseer, 2002. hybrid electric ship using multi-objective particle swarm optimization,” in
[25] S. F. Tie and C. W. Tan, “A review of energy sources and energy Proc. IEEE 10th Int. Conf. Power Electron. ECCE Asia, 2019, pp. 1460–
management system in electric vehicles,” Renewable Sustain. Energy 1466.
Rev., vol. 20, pp. 82–102, 2013. [48] E. A. Sciberras, B. Zahawi, D. J. Atkinson, A. Breijs, and J. H. van
[26] E. Silvas, T. Hofman, N. Murgovski, L. P. Etman, and M. Steinbuch, Vugt, “Managing shipboard energy: A stochastic approach special issue
“Review of optimization strategies for system-level design in hybrid on marine systems electrification,” IEEE Trans. Transport. Electrific.,
electric vehicles,” IEEE Trans. Veh. Technol., vol. 66, no. 1, pp. 57–70, vol. 2, no. 4, pp. 538–546, Dec. 2016.
Jan. 2017. [49] F. D. Kanellos, A. Anvari-Moghaddam, and J. M. Guerrero, “Smart
[27] S. F. Tie and C. W. Tan, “A review of energy sources and energy shipboard power system operation and management,” Inventions, vol. 1,
management system in electric vehicles,” Renewable Sustain. Energy no. 4, p. 22, 2016.
Rev., vol. 20, pp. 82–102, 2013. [50] A. De, S. K. Kumar, A. Gunasekaran, and M. K. Tiwari, “Sustainable
[28] N. Sulaiman, M. Hannan, A. Mohamed, E. Majlan, and W. W. Daud, maritime inventory routing problem with time window constraints,” Eng.
“A review on energy management system for fuel cell hybrid electric Appl. Artif. Intell., vol. 61, pp. 77–95, 2017.
vehicle: Issues and challenges,” Renewable Sustain. Energy Rev., vol. 52, [51] M. D. Al-Falahi, K. S. Nimma, S. D. Jayasinghe, H. Enshaei, and J. M.
pp. 802–814, 2015. Guerrero, “Power management optimization of hybrid power systems in
[29] Y. Huang, H. Wang, A. Khajepour, H. He, and J. Ji, “Model predictive electric ferries,” Energy Convers. Manage., vol. 172, pp. 50–66, 2018.
control power management strategies for hevs: A review,” J. Power [52] S. Fang, Y. Xu, and Z. Li, “Joint generation and demand-side management
Sources, vol. 341, pp. 91–106, 2017. for shipboard carbon capture and storage system,” in Proc. IEEE/IAS 55th
[30] F. D. Kanellos, G. J. Tsekouras, and N. D. Hatziargyriou, “Optimal Ind. Commercial Power Syst. Tech. Conf., 2019, pp. 1–8.
demand-side management and power generation scheduling in an all- [53] S. Fang and Y. Xu, “Multi-objective robust energy management for all-
electric ship,” IEEE Trans. Sustain. Energy, vol. 5, no. 4, pp. 1166–1175, electric shipboard microgrid under uncertain wind and wave,” Int. J. Elect.
Oct. 2014. Power Energy Syst., vol. 117, 2020, Art. no. 105600.
[31] F. Kanellos, “Optimal power management with GHG emissions lim- [54] P. Michalopoulos, F. D. Kanellos, G. J. Tsekouras, and J. M. Prousa-
itation in all-electric ship power systems comprising energy storage lidis, “A method for optimal operation of complex ship power systems
systems,” IEEE Trans. Power Syst., vol. 29, no. 1, pp. 330–339, Jan. 2014. employing shaft electric machines,” IEEE Trans. Transport. Electrific.,
[32] S. Fang, Y. Xu, Z. Li, T. Zhao, and H. Wang, “Two-step multi-objective vol. 2, no. 4, pp. 547–557, Dec. 2016.
management of hybrid energy storage system in all-electric ship mi- [55] Z. Li, Y. Xu, S. Fang, Y. Wang, and X. Zheng, “Multi-objective co-
crogrids,” IEEE Trans. Veh. Technol., vol. 68, no. 4, pp. 3361–3373, ordinated energy dispatch and voyage scheduling for a multi-energy
Apr. 2019. ship microgrid,” IEEE Trans. Ind. Appl., vol. 56, no. 2, pp. 989–999,
[33] A. Fathy, H. Rezk, and A. M. Nassef, “Robust hydrogen-consumption- Mar./Apr. 2019.
minimization strategy based salp swarm algorithm for energy man- [56] A. De, A. Choudhary, and M. K. Tiwari, “Multiobjective approach for
agement of fuel cell/supercapacitor/batteries in highly fluctuated load sustainable ship routing and scheduling with draft restrictions,” IEEE
condition,” Renewable Energy, vol. 139, pp. 147–160, 2019. Trans. Eng. Manag., vol. 66, no. 1, pp. 35–51, Feb. 2019.
[34] A. M. Bassam, A. B. Phillips, S. R. Turnock, and P. A. Wilson, “An [57] C. Yao, M. Chen, and Y.-Y. Hong, “Novel adaptive multi-clustering
improved energy management strategy for a hybrid fuel cell/battery algorithm-based optimal ess sizing in ship power system considering
passenger vessel,” Int. J. Hydrogen Energy, vol. 41, no. 47, pp. 22 453–22 uncertainty,” IEEE Tran. Power Syst., vol. 33, no. 1, pp. 307–316,
464, 2016. Jan. 2018.
[35] M. E. P. Committee et al., “Guideline for Voluntary Use of the Ship [58] C. Shang, D. Srinivasan, and T. Reindl, “Economic and environmental
Energy Efficiency Operational Indicator (EEOI), London, U.K.: Int. generation and voyage scheduling of all-electric ships,” IEEE Tran.
Maritime Org.,” 2009. Power Syst., vol. 31, no. 5, pp. 4087–4096, Sep. 2015.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 11

[59] V. Bolbot, N. L. Trivyza, G. Theotokatos, E. Boulougouris, A. Rentizelas, [82] A. Dolatabadi, R. Ebadi, and B. Mohammadi-Ivatloo, “A two-stage
and D. Vassalos, “Cruise ships power plant optimisation and comparative stochastic programming model for the optimal sizing of hybrid
analysis,” Energy, vol. 196, 2020, Art. no. 117061. PV/diesel/battery in hybrid electric ship system,” J. Oper. Autom. Power
[60] Z. Jianyun, C. Li, X. Lijuan, and W. Bin, “Bi-objective optimal design Eng., vol. 7, no. 1, pp. 16–26, 2019.
of plug-in hybrid electric propulsion system for ships,” Energy, vol. 177, [83] A. Boveri, F. Silvestro, and P. Gualeni, “Ship electrical load analysis and
pp. 247–261, 2019. power generation optimisation to reduce operational costs,” in Proc. Int.
[61] F. D. Kanellos, A. Anvari-Moghaddam, and J. M. Guerrero, “A cost- Conf. Elect. Syst. Aircr., Railway, Ship Propulsion Road Veh. Int. Transp.
effective and emission-aware power management system for ships with Electrific. Conf., 2016, pp. 1–6.
integrated full electric propulsion,” Electric Power Syst. Res., vol. 150, [84] A. Boveri, F. Silvestro, M. Molinas, and E. Skjong, “Optimal sizing of
pp. 63–75, 2017. energy storage systems for shipboard applications,” IEEE Trans. Energy
[62] R. Kumar et al., “Optimal sizing of hybrid ship power system using Convers., vol. 34, no. 2, pp. 801–811, Jun. 2019.
variants of particle swarm optimization,” in Proc. IEEE Recent Develop. [85] S. Wen et al., “Optimal sizing of hybrid energy storage sub-systems in
Control, Autom. Power Eng., 2017, pp. 527–532. PV/diesel ship power system using frequency analysis,” Energy, vol. 140,
[63] F. D. Kanellos, A. Anvari-Moghaddam, and J. M. Guerrero, “Smart pp. 198–208, 2017.
shipboard power system operation and management,” Inventions, vol. 1, [86] S. Cao, Z. Yang, H. Cheng, Y. Zheng, and J. Geng, “Time coupling
no. 4, p. 22, 2016. constraints modeling and analysis in unit commitment,” in Proc. Int. Conf.
[64] Z. Li, Y. Xu, S. Fang, Y. Wang, and X. Zheng, “Multiobjective coordinated Elect., Mech. Ind. Eng., Apr. 2016, pp. 176–180. [Online]. Available:
energy dispatch and voyage scheduling for a multienergy ship microgrid,” https://doi.org/10.2991/icemie-16.2016.44
IEEE Trans. Ind Appl., vol. 56, no. 2, pp. 989–999, Mar./Apr. 2020. [87] M. Jaurola, A. Hedin, S. Tikkanen, and K. Huhtala, “Topti: A flexible
[65] E. Skjong, T. A. Johansen, M. Molinas, and A. J. Sørensen, “Approaches framework for optimising energy management for various ship machinery
to economic energy management in diesel-electric marine vessels,” IEEE topologies,” J. Mar. Sci. Technol., vol. 24, no. 4, pp. 1183–1196, 2019.
Trans. Transport. Electrific., vol. 3, no. 1, pp. 22–35, Mar. 2017. [88] C. Wang, K. Ren, and J. Wang, “Secure optimization computation out-
[66] Z. Li, Y. Xu, S. Fang, X. Zheng, and X. Feng, “Robust coordination sourcing in cloud computing: A case study of linear programming,” IEEE
of a hybrid ac/dc multi-energy ship microgrid with flexible voyage and Tran. Comput., vol. 65, no. 1, pp. 216–229, Jan. 2016.
thermal loads,” IEEE Trans. Smart Grid, vol. 11, no. 4, pp. 2782–2793, [89] D. G. Luenberger et al., Linear and Nonlinear Programming, vol. 2.
Jul. 2020. Berlin, Germany: Springer, 1984.
[67] A. Ritari, J. Huotari, J. Halme, and K. Tammi, “Hybrid electric topology [90] M. Carrión and J. M. Arroyo, “A computationally efficient mixed-integer
for short sea ships with high auxiliary power availability requirement,” linear formulation for the thermal unit commitment problem,” IEEE
Energy, vol. 190, 2020, Art. no. 116359. Trans. Power Syst., vol. 21, no. 3, pp. 1371–1378, Aug. 2006.
[68] A. De, V. K. R. Mamanduru, A. Gunasekaran, N. Subramanian, and M. K. [91] M. R. Hossain and H. L. Ginn, “Real-time distributed coordination of
Tiwari, “Composite particle algorithm for sustainable integrated dynamic power electronic converters in a dc shipboard distribution system,” IEEE
ship routing and scheduling optimization,” Comput. Ind. Eng., vol. 96, Trans. Energy Convers., vol. 32, no. 2, pp. 770–778, Jun. 2017.
pp. 201–215, 2016. [92] K. Lai and M. S. Illindala, “A distributed energy management strategy for
[69] H.-M. Chin, C.-L. Su, and C.-H. Liao, “Estimating power pump loads resilient shipboard power system,” Appl. Energy, vol. 228, pp. 821–832,
and sizing generators for ship electrical load analysis,” IEEE Trans. Ind. 2018.
Appl., vol. 52, no. 6, pp. 4619–4627, Nov./Dec. 2016. [93] F. Baldi, F. Ahlgren, F. Melino, C. Gabrielii, and K. Andersson, “Optimal
[70] S. Wen, H. Lan, Y.-Y. Hong, C. Y. David, L. Zhang, and P. Cheng, load allocation of complex ship power plants,” Energy Convers. Manage.,
“Allocation of ESS by interval optimization method considering impact vol. 124, pp. 344–356, 2016.
of ship swinging on hybrid PV/diesel ship power system,” Appl. Energy, [94] A. Miró, C. Pozo, G. Guillén-Gosálbez, J. A. Egea, and L. Jiménez,
vol. 175, pp. 158–167, 2016. “Deterministic global optimization algorithm based on outer approx-
[71] Y. Yan et al., “Multi-objective design optimization of combined cooling, imation for the parameter estimation of nonlinear dynamic biological
heating and power system for cruise ship application,” J. Cleaner Prod., systems,” BMC Bioinf., vol. 13, no. 1, p. 90, 2012.
vol. 233, pp. 264–279, 2019. [95] M. Basu, “Dynamic economic emission dispatch using nondominated
[72] A. Boveri, P. Gualeni, D. Neroni, and F. Silvestro, “Stochastic approach sorting genetic algorithm-ii,” Int. J. Elect. Power Energy Syst., vol. 30,
for power generation optimal design and scheduling on ships,” in Proc. no. 2, pp. 140–149, 2008.
IEEE PES Innov. Smart Grid Technol. Conf. Eur. (ISGT-Europe), 2017, [96] D. Gao, X. Wang, T. Wang, Y. Wang, and X. Xu, “Optimal thrust
pp. 1–6. allocation strategy of electric propulsion ship based on improved non-
[73] S. Mashayekh and K. L. Butler-Purry, “An integrated security- dominated sorting genetic algorithm ii,” IEEE Access, vol. 7, pp. 135 247–
constrained model-based dynamic power management approach for iso- 135255, 2019.
lated microgrids in all-electric ships,” IEEE Trans. Power Syst., vol. 30, [97] L. Liu, S. Yang, and D. Wang, “Particle swarm optimization with compos-
no. 6, pp. 2934–2945, Nov. 2015. ite particles in dynamic environments,” IEEE Trans. Syst., Man, Cybern.,
[74] R. Tang, X. Li, and J. Lai, “A novel optimal energy-management strat- B. (Cybernetics), vol. 40, no. 6, pp. 1634–1648, Dec. 2010.
egy for a maritime hybrid energy system based on large-scale global [98] A. Armellini, S. Daniotti, P. Pinamonti, and M. Reini, “Evaluation of gas
optimization,” Appl. Energy, vol. 228, pp. 254–264, 2018. turbines as alternative energy production systems for a large cruise ship
[75] A. Letafat et al., “Simultaneous energy management and optimal com- to meet new maritime regulations,” Appl. Energy, vol. 211, pp. 306–317,
ponents sizing of a zero-emission ferry boat,” J. Energy Storage, vol. 28, 2018.
2020, Art. no. 101215. [99] V. Sarfi and H. Livani, “A novel multi-objective security-constrained
[76] H. Chen, Z. Zhang, C. Guan, and H. Gao, “Optimization of sizing and power management for isolated microgrids in all-electric ships,” in Proc.
frequency control in battery/supercapacitor hybrid energy storage system IEEE Elect. Ship Technol. Symp., 2017, pp. 148–155.
for fuel cell ship,” Energy, vol. 197, 2020, Art. no. 117285. [100] S. A. Cook, “An overview of computational complexity,” Commun. ACM,
[77] B. Manasa and K. Vaisakh, “Optimal shipboard power management by vol. 26, no. 6, pp. 400–408, 1983.
classical and differential evolution methods,” Int. Res. J. Eng. Technol., [101] L. C. W. Yuan, T. Tjahjowidodo, G. S. G. Lee, R. Chan, and A. K.
vol. 6, pp. 269–279, 2019. Ådnanes, “Equivalent consumption minimization strategy for hybrid all-
[78] Q. Xu, B. Yang, Q. Han, Y. Yuan, C. Chen, and X. Guan, “Optimal power electric tugboats to optimize fuel savings,” in Proc. Amer. Control Conf.,
management for failure mode of MVdc microgrids in all-electric ships,” 2016, pp. 6803–6808.
IEEE Trans. Power Syst., vol. 34, no. 2, pp. 1054–1067, Mar. 2019. [102] L. W. Chua, T. Tjahjowidodo, G. G. Seet, and R. Chan, “Implementation
[79] G. La Tona, M. Luna, M. di Piazza, and A. Pietra, “Energy management of optimization-based power management for all-electric hybrid vessels,”
system for efficiency increase in cruise ship microgrids,” in Proc. 45th IEEE Access, vol. 6, pp. 74 339–74354, 2018.
Annu. Conf. IEEE Ind. Electron. Soc., 2019, vol. 1, pp. 4056–4062. [103] J. Zhu, L. Chen, X. Wang, and L. Yu, “Bi-level optimal sizing and energy
[80] S. Fang, Y. Xu, Z. Li, Z. Ding, L. Liu, and H. Wang, “Optimal sizing management of hybrid electric propulsion systems,” Appl. Energy,
of shipboard carbon capture system for maritime greenhouse emis- vol. 260, 2020, Art. no. 114134.
sion control,” IEEE Trans. Ind Appl., vol. 55, no. 6, pp. 5543–5553, [104] C.-C. Lin, H. Peng, J. W. Grizzle, and J.-M. Kang, “Power management
Nov./Dec. 2019. strategy for a parallel hybrid electric truck,” IEEE Trans. Control Syst.
[81] A. Anvari-Moghaddam, T. Dragicevic, L. Meng, B. Sun, and J. M. Technol., vol. 11, no. 6, pp. 839–849, Nov. 2003.
Guerrero, “Optimal planning and operation management of a ship elec- [105] F. Balsamo, C. Capasso, G. Miccione, and O. Veneri, “Hybrid storage
trical power system with energy storage system,” in Proc. 42nd Annu. system control strategy for all-electric powered ships,” Energy Procedia,
Conf. IEEE Ind. Electron. Soc., 2016, pp. 2095–2099. vol. 126, pp. 1083–1090, 2017.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

12 IEEE SYSTEMS JOURNAL

[106] R. Tang, Z. Wu, and X. Li, “Optimal operation of [127] X. Feng, K. L. Butler-Purry, and T. Zourntos, “Real-time electric load
photovoltaic/battery/diesel/cold-ironing hybrid energy system for management for DC zonal all-electric ship power systems,” Elect. Power
maritime application,” Energy, vol. 162, pp. 697–714, 2018. Syst. Res., vol. 154, pp. 503–514, 2018.
[107] S. Paran, T. Vu, T. El Mezyani, and C. Edrington, “MPC-based power [128] K. Lai and M. S. Illindala, “A distributed energy management strategy for
management in the shipboard power system,” in Proc. IEEE Elect. Ship resilient shipboard power system,” Appl. Energy, vol. 228, pp. 821–832,
Technol. Symp., 2015, pp. 14–18. 2018.
[108] K. Wang, X. Yan, Y. Yuan, X. Jiang, X. Lin, and R. R. Negenborn, “Dy- [129] C. S. Edrington et al., “Distributed energy management for ship power
namic optimization of ship energy efficiency considering time-varying systems with distributed energy storage,” J. Mar. Eng. Technol., vol. 19,
environmental factors,” Transp. Res. D, Transport Environ., vol. 62, pp. 31–44, 2020.
pp. 685–698, 2018. [130] T. V. Vu et al., “Large-scale distributed control for MVdc ship power
[109] L. C. W. Yuan, T. Tjahjowidodo, G. S. G. Lee, and R. Chan, “Optimizing systems,” in Proc. 44th Annu. Conf. IEEE Ind. Electron. Soc., 2018,
fuel savings and power system reliability for all-electric hybrid vessels pp. 3431–3436.
using model predictive control,” in Proc. IEEE Int. Conf. Adv. Intell. [131] X. Feng, K. L. Butler-Purry, and T. Zourntos, “Multi-agent system-based
Mechatronics, 2017, pp. 1532–1537. real-time load management for all-electric ship power systems in DC zone
[110] G. Papalambrou, S. Samokhin, S. Topaloglou, N. Planakis, N. Kyrtatos, level,” IEEE Trans. Power Syst., vol. 27, no. 4, pp. 1719–1728, Nov. 2012.
and K. Zenger, “Model predictive control for hybrid diesel-electric ma- [132] Z. Hong, Q. Li, Y. Han, W. Shang, Y. Zhu, and W. Chen, “An energy
rine propulsion,” IFAC-PapersOnLine, vol. 50, no. 1, pp. 11 064–11069, management strategy based on dynamic power factor for fuel cell/battery
2017. hybrid locomotive,” Int. J. Hydrogen Energy, vol. 43, no. 6, pp. 3261–
[111] D. E. Gonsoulin, T. V. Vu, F. Diaz, H. Vahedi, D. Perkins, and C. 3272, 2018.
S. Edrington, “Coordinating multiple energy storages using MPC for [133] D. Gonsoulin, T. Vu, F. Diaz, H. Vahedi, D. Perkins, and C. Edrington,
ship power systems,” in Proc. IEEE Elect. Ship Technol. Symp., 2017, “Centralized MPC for multiple energy storages in ship power systems,”
pp. 551–556. in Proc. 43rd Annu. Conf. IEEE Ind. Electron. Soc., 2017, pp. 6777–6782.
[112] Z. Wu and X. Xia, “Tariff-driven demand side management of green [134] A. R. Dahl, L. Thorat, and R. Skjetne, “Model predictive control of
ship,” Sol. Energy, vol. 170, pp. 991–1000, 2018. marine vessel power system by use of structure preserving model,”
[113] A. Haseltalab and R. R. Negenborn, “Model predictive maneuvering IFAC-PapersOnLine, vol. 51, no. 29, pp. 335–340, 2018.
control and energy management for all-electric autonomous ships,” Appl. [135] P. Stone et al., “Shipboard power management using constrained nonlin-
Energy, vol. 251, 2019, Art. no. 113308. ear model predictive control,” in Proc. IEEE Elect. Ship Technol. Symp.,
[114] J. Hou, J. Sun, and H. F. Hofmann, “Mitigating power fluctuations 2015, pp. 1–7.
in electric ship propulsion with hybrid energy storage system: De- [136] T. I. Bø and T. A. Johansen, “Battery power smoothing control in a marine
sign and analysis,” IEEE J. Ocean. Eng., vol. 43, no. 1, pp. 93–107, electric power plant using nonlinear model predictive control,” IEEE
Jan. 2018. Trans. Control Syst. Technol., vol. 25, no. 4, pp. 1449–1456, Jul. 2017.
[115] H. Park et al., “Real-time model predictive control for shipboard power [137] T. V. Vu, D. Gonsoulin, D. Perkins, F. Diaz, H. Vahedi, and C. S.
management using the IPA-SQP approach,” IEEE Trans. Control Syst. Edrington, “Predictive energy management for MVdc all-electric ships,”
Technol., vol. 23, no. 6, pp. 2129–2143, Nov. 2015. in Proc. IEEE Elect. Ship Technol. Symp., 2017, pp. 327–331.
[116] J. Hou, Z. Song, H. Park, H. Hofmann, and J. Sun, “Implementation and [138] J. Hou, J. Sun, and H. Hofmann, “Integrated control of power generation,
evaluation of real-time model predictive control for load fluctuations mit- electric motor and hybrid energy storage for all-electric ships,” in Proc.
igation in all-electric ship propulsion systems,” Appl. Energy, vol. 230, Amer. Control Conf., 2016, pp. 6797–6802.
pp. 62–77, 2018. [139] A. Haseltalab and R. R. Negenborn, “Predictive on-board power man-
[117] J. Hou, Z. Song, H. Hofmann, and J. Sun, “Adaptive model pre- agement for all-electric ahips with DC distribution architecture,” in Proc.
dictive control for hybrid energy storage energy management in all- Aberdeen, 2017, pp. 1–8.
electric ship microgrids,” Energy Convers. Manage., vol. 198, 2019, [140] A. Haseltalab, R. R. Negenborn, and G. Lodewijks, “Multi-level pre-
Art. no. 111929. dictive control for energy management of hybrid ships in the presence
[118] T. V. KISHOR, J. HPawar, and S. MODUGU, “Analysis: Hybrid RES of uncertainty and environmental disturbances,” IFAC-PapersOnLine,
fuzzy based predictive control for shipboard using the IPA-SQP,” Int. J. vol. 49, no. 3, pp. 90–95, 2016.
Sci. Eng. Technol. Res., vol. 5, pp. 7492–7499, 2016. [141] S. Tan, J. M. Guerrero, P. Xie, R. Han, and J. C. Vasquez, “Brief survey
[119] N. Zohrabi, H. Zakeri, and S. Abdelwahed, “Efficient load management on attack detection methods for cyber-physical systems,” IEEE Syst. J.,
in electric ships: A model predictive control approach,” in Proc. IEEE vol. 14, no. 4, pp. 5329–5339, Dec. 2020.
Appl. Power Electron. Conf. Expo., 2019, pp. 3000–3006. [142] D. Lee and C.-C. Cheng, “Energy savings by energy management sys-
[120] J. Hou, J. Sun, and H. Hofmann, “Adaptive model predictive control tems: A review,” Renewable Sustain. Energy Rev., vol. 56, pp. 760–777,
with propulsion load estimation and prediction for all-electric ship energy 2016.
management,” Energy, vol. 150, pp. 877–889, 2018.
[121] T. I. Bø and T. A. Johansen, “Dynamic safety constraints by scenario-
based economic model predictive control of marine electric power
plants,” IEEE Trans. Transport. Electrific., vol. 3, no. 1, pp. 13–21,
Mar. 2017.
[122] G. Seenumani, H. Peng, and J. Sun, “A reference governor-based hi-
erarchical control for failure mode power management of hybrid power
systems for all-electric ships,” J. Power Sources, vol. 196, no. 3, pp. 1599–
1607, 2011.
[123] G. Seenumani, J. Sun, and H. Peng, “A hierarchical optimal control
strategy for power management of hybrid power systems in all electric
ships applications,” in Proc. 49th IEEE Conf. Decis. Control, 2010,
pp. 3972–3977. Peilin Xie (Student Member, IEEE) received the
[124] S. Hasanvand, M. Rafiei, M. Gheisarnejad, and M.-H. Khooban, B.S. degree in electrical engineering from Beijing
“Reliable power scheduling of an emission-free ship: Multi-objective Jiaotong University, Beijing, China, in 2015, and the
deep reinforcement learning,” IEEE Trans. Transport. Electrific., vol. 6, M.S. degree in electrical engineering and automation
no. 2, pp. 832–843, Jun. 2020. from North China Electric Power University, Beijing,
[125] G. Seenumani, J. Sun, and H. Peng, “Real-time power management of China, in 2018. She is currently working toward the
integrated power systems in all electric ships leveraging multi time scale Ph.D. degree with the Department of Energy Tech-
property,” IEEE Trans. Control Syst. Technol., vol. 20, no. 1, pp. 232–240, nology, Aalborg University, Aalborg, Denmark.
Jan. 2012. Her research interests include virtual synchronous
[126] T. Vu, S. Paran, T. El Mezyani, and C. Edrington, “Real-time distributed generator technology, the power management control
power optimization in the dc microgrids of shipboard power systems,” in for shipboard microgrids, and distributed PMS for
Proc. IEEE Elect. Ship Technol. Symp., 2015, pp. 118–122. large-scale shipboard microgrids.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

XIE et al.: OPTIMIZATION-BASED POWER AND ENERGY MANAGEMENT SYSTEM 13

Josep M. Guerrero (Fellow, IEEE) received the B.S. Juan C. Vasquez (Senior Member, IEEE ) received
degree in telecommunications engineering, the M.S. the B.S. degree in electronics engineering from the
degree in electronics engineering, and the Ph.D. de- Autonomous University of Manizales, Manizales,
gree in power electronics from the Technical Univer- Colombia, in 2004, and the Ph.D. degree in auto-
sity of Catalonia, Barcelona, Spain, in 1997, 2000, matic control, robotics, and computer vision from
and 2003, respectively. BarcelonaTech-UPC, Barcelona, Spain, in 2009.
Since 2011, he has been a Full Professor with In 2011, he joined as an Assistant Professor with
the Department of Energy Technology, Aalborg Uni- the Department of Energy Technology, Aalborg Uni-
versity, Denmark, where he is responsible for the versity, Denmark, where he became an Associate
Microgrid Research Program. He has authored and Professor in 2014. In 2019, he became Professor
coauthored more than 500 journal papers in the fields in Energy Internet and Microgrids and he is cur-
of microgrids and renewable energy systems, which are cited more than 50 000 rently the Co-Director of the Villum Center for Research on Microgrids (see
times. His research specially focused on microgrid technologies applied to crom.et.aau.dk). He was a Visiting Scholar with the Center of Power Electronics
offshore wind, maritime microgrids for electrical ships, vessels, ferries and Systems, Virginia Tech, USA and a Visiting Professor with Ritsumeikan Univer-
seaports, and space microgrids applied to nanosatellites and spacecrafts. His sity, Japan. He has authored and coauthored more than 450 journal papers in the
research interests include oriented to different microgrid aspects, including field of microgrids, which in total are cited more than 19 000 times. His current
power electronics, distributed energy-storage systems, hierarchical and coop- research interests include operation, advanced hierarchical and cooperative
erative control, energy management systems, smart metering, and the Internet control, optimization and energy management applied to distributed generation
of Things for ac–dc microgrid clusters and islanded minigrids. in ac–dc microgrids, maritime microgrids, advanced metering infrastructures,
Prof. Guerrero is an Associate Editor for a number of IEEE transactions. and the integration of Internet of Things and energy Internet into the smartgrid.
Since 2014 he is the Chair Professor at Shandong University; Since 2015 he is Prof. Vasquez is an Associate Editor of IET Power Electronics and a Guest Ed-
a Distinguished Guest Professor with Hunan University; and since 2016 he is a itor for the IEEE TRANSACTIONS ON INDUSTRIMAL INFORMATICS Special Issue
Visiting Professor Fellow at Aston University, U.K., and a Guest Professor with on Energy Internet. He was awarded as Highly Cited Researcher by Thomson
the Nanjing University of Posts and Telecommunications, Nanjing, China. In Reuters from 2017 to 2019 and He was the recipient of the Young Investigator
2019, he became a Villum Investigator by The Villum Fonden, which supports the Award 2019. He is currently a member of the IEC System Evaluation Group
Center for Research on Microgrids, Aalborg University, as he is the founder and SEG4 on LVdc Distribution and Safety for use in Developed and Developing
Director of the same centre. He was the recipient of the Best Paper Award of the Economies, the Renewable Energy Systems Technical Committee TC-RES in
IEEE TRANSACTIONS ON ENERGY CONVERSION for the period 2014 and 2015, IEEE Industrial Electronics, PELS, IAS, and PES Societies.
and the Best Paper Prize of IEEE-PES in 2015. He has also received the Best
Paper Award of the Journal of Power Electronics in 2016. During six consecutive
years, from 2014 to 2019, he was awarded by Clarivate Analytics (former
Thomson Reuters) as Highly Cited Researcher. In 2015, he was elevated as
IEEE Fellow for his contributions on distributed power systems and microgrids.

Mojtaba Mehrzadi received the B.Sc. degree in

electrical engineering and the M.Sc. degree in control
theory and control engineering from Islamic Azad
University (IAU), Tehran, Iran, in 2007 and 2012,
respectively, He is currently working toward the Ph.D.
Sen Tan (Student Member, IEEE) received the B.S. degree with the Department of Energy Technology,
degree in automation and the M.S. degree in control Aalborg University, Aalborg, Denmark.
engineering from Northeastern University, Liaoning, His research interests include the power manage-
China, in 2014 and 2017, respectively. He is currently ment system in the hybrid maritime microgrid.
working toward the Ph.D. degree with the Department
of Energy Technology, Aalborg University, Aalborg,
Denmark.
His research interests include distributed control
and power management strategy design for microgrid,
fault detection, and motor drive technologies.

Yusuf Al-Turki (Senior Member, IEEE) received the

Ph.D. degree in power systems from the University
of Manchester, Manchester, U.K., in 1985.
Since 1999, he has been a Professor with the De-
partment of Electrical and Computer Engineering,
King Abdulaziz University, Jeddah, Saudi Arabia,
Najmeh Bazmohammadi received the bachelor’s where he is currently the Vice-President for Graduate
degree in electrical engineering and the master’s de- Studies and Scientific Research. His research interests
gree in electrical engineering-control from the Fer- include system modeling, power system dynamics,
dowsi University of Mashhad, Mashhad, Iran, in renewable energy, and microgrids.
2009 and 2012, respectively, and the Ph.D. degree in
electrical engineering-control from the K. N. Toosi
University of Technology, Tehran, Iran, in 2019.
She is currently a Postdoctoral Research Fellow
with the Center for Research on Microgrids, De-
partment of Energy Technology, Aalborg University,
Denmark. Her current research interests include mod-
eling and control of dynamic systems, decision-making under uncertainty, model
predictive control and its application in energy management of hybrid and
renewable-based power systems and life support systems.

Authorized licensed use limited to: Carleton University. Downloaded on June 02,2021 at 10:52:18 UTC from IEEE Xplore. Restrictions apply.