Journal of Energy Storage 78 (2024) 109888

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

Research papers

Capabilities of compressed air energy storage in the economic design of

renewable off-grid system to supply electricity and heat costumers and
smart charging-based electric vehicles
Farshad Khalafian a, Nahal Iliaee b, Ekaterina Diakina c, d, Peyman Parsa e,
Mohammed M. Alhaider f, Milad Hadizadeh Masali g, Sasan Pirouzi h, Min Zhu i, *
a
Department of Electrical Engineering, Ahvaz Branch, Islamic Azad University, Ahvaz, Iran
b
Department of Electronics, Carleton University, Ottawa, Canada
c
Department of Mechanical Engineering, Bauman Moscow State Technical University, Moscow, Russia
d
Department of Mathematics and Natural Sciences, Gulf University for Science and Technology, Mishref Campus, Kuwait
e
Faculty of Electrical Engineering, Shahid Beheshti University, Tehran, Iran
f
Department of Electrical Engineering, College of Engineering in Wadi Alddawasir, Prince Sattam bin Abdulaziz University, 11991, Saudi Arabia
g
Department of Electrical Engineering Lamar University, Beaumont, TX, USA
h
Department of Engineering, Semirom Branch, Islamic Azad University, Semirom, Iran
i
College of Information Science and Technology, Zhejiang Shuren University, Hangzhou, China

A R T I C L E I N F O A B S T R A C T

Keywords: The study presents the sizing of a hybrid renewable islanded system that can feed both electrical and thermal
Bio-waste power generation energy simultaneously. The aforementioned energies are generated through the utilization of wind systems and
Compressed air energy storage bio-waste units. The integration of combined heat and power technology in the bio-waste unit enables the
Electricity and thermal consumers
concurrent generation of electricity and heat power. Electric vehicles consume electric energy, but function
Electric vehicles parking lot
Point estimate method
based on a smart charging. The study employs compressed air energy storage as a means to bridge the disparity
Renewable off-grid system between the patterns of electric power generation and consumption, with the aim of enhancing energy efficiency
and reducing planning expenses. Thermal energy storage serves as an intermediary between renewable power
and load profiles within the thermal sector. Objective here is find the minimum annual capital and maintenance
costs of sources, storages, and power electronic converters. Problem is constrained to operational model of
renewable sources, storages, and electric vehicles. The present model prioritizes the utilization of renewable
sources for the purpose of supplying electrical and thermal loads. Electric vehicles regulate their active power to
minimize the disparity between load and electric production profiles. Storages perform charging and discharging
functions to bridge the disparity between power generation and consumption. Strategy incorporates uncertainty
related to load, wind speed, bio-waste-derived gas production, and electric vehicles parameters. The study
employs a point estimate method to attain a dependable solution. The numerical results indicate that incorpo­
ration of compressed air storage in the hybrid system results in a decrease of 7.7 % (12.9 %) in the planning costs
relative to system with batteries (hydrogen storage). The proposed design's concurrent provision of electrical and
thermal energy leads to an approximately 57 % increasing in planning costs when compared to a system that
solely provides electrical energy. The implementation of a smart charging strategy for electric vehicles drop the
hybrid system's overall scheduling cost by 3.9 %.

* Corresponding author.
E-mail address: [email protected] (M. Zhu).

https://doi.org/10.1016/j.est.2023.109888
Received 12 July 2023; Received in revised form 8 November 2023; Accepted 23 November 2023
Available online 14 December 2023
2352-152X/© 2023 Elsevier Ltd. All rights reserved.
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

1. Introduction approaches for mitigating surplus electricity in stand-alone hybrid RESs.

This research focuses on the performance concepts associated with the
1.1. Motivation reduction of excess power. Thus, this study introduces four distinct
categories pertaining to the utilization of surplus electricity, namely:
Deprived of energy distribution networks, consumers in remote areas direct utilization, storage, indirect utilization, and reduction of excess
are supplied by different sources and storage equipment by establishing electricity production. The most prevalent methods identified include
an islanded system [1]. This system consists of renewable energy sources shiftable demand, power to heat, storage elements, power to hydrogen,
(RESs) to reach clean energy supply conditions [2]. Among these sour­ power to gas cycles, several generators, and loss of generation. Ref. [9]
ces, wind turbines (WT) and photovoltaics (PVs) produce energy based focuses on the performance of a hybrid system that utilizes green
on environmentally related parameters, like wind speed and irradiance hydrogen generated from solar and wind sources. The energy system
[2]. Also, bio-waste units provide consumers with energy by converting under consideration comprises PV panels, WTs, and a Proton Exchange
environmental waste into methane gas and then converting it into en­ Membrane water electrolyzer. The meteorological data, including solar
ergy [3]. WT and PVs are low in emissions due to their independence irradiance, temperature, and wind speed, were acquired from the
from fossil fuels. Bio-waste units play an effective role in reducing Meteorology Station located at Atilim University Incek Campus in
environmental pollution thanks to the removal or consumption of Ankara, Turkey. The study conducted in reference [10] investigates a
environmental waste. Therefore, according to the mentioned cases, the variety of off-grid hybrid RESs configurations to provide electricity to a
use of the mentioned resources in the island system is significant. In the remote island community. A total of six distinct configurations under­
following, the demand profile is not the same as the production power went comprehensive evaluations encompassing technical, economic,
generated by renewables. Hence, to make the load and production environmental, and social aspects, to determine the most optimal
profiles almost the same, energy storage needs to be incorporated in design. A sensitivity analysis has also been conducted on the system.
addition to RESs [1]. The storage device works in charging mode when Moreover, various machine learning models were incorporated to
the generation level surpasses the demand and stores the excess energy. forecast the system's functionality.
It then operates in the discharge mode in the case the energy con­ Ref. [11] discusses the economic scheduling of an environmentally
sumption level surpasses the energy generation and supplies the con­ friendly islanded system. This system comprises a WT, battery, CHP
sumption energy shortage [1]. As another issue, electricity and heat system, and heat storage elements. The objective is to simultaneously
demand are generally considerable in the consumption areas, so the fulfill the electricity and heat demand. The proposed methodology aims
mentioned system should be able to supply these energies simulta­ to achieve the minimum capital, repair and maintenance, and operation
neously [4]. In this situation and to improve energy efficiency, com­ cost, while also seeking to minimize the associated pollution levels. The
bined heat and power (CHP) units can play an effective role [4]. Of operation-scheduling model of these generation units and storage im­
course, it is possible to extract this unit from renewable equipment. For poses constraints on the system. Specifically, the WT and CHP system is
instance, the bio-waste unit turns environmental waste into methane. A preferred when it comes to supplying electricity, while the CHP units
CHP then converts this gas into electricity and heat. Thus, by using 100 take precedence in supplying heat. Additionally, batteries and storage
% RESs, it is possible to simultaneously provide electricity and heat to elements are responsible for bridging the disparity between generation
the islanded system. In addition, to extract an optimal structure for the and consumption levels. The scheme's uncertain quantities encompass
island system, the scheduling of the system should be optimized. In this electricity and heat demand, as well as WTs' power generation. The
problem, the optimal number and size of generation units and storage proposed design utilizes the unscented transformation technique to
equipment are specified [5]. Nonetheless, it should be said, that un­ accurately model the uncertainties that have been mentioned. The study
certain parameters such as load, renewable resources, and some storage described in [12] outlines the design of a hybrid RES incorporating WTs
elements need to be applied, and to provide a modeling of the afore­ and bio-waste energy units, as well as stationary energy storage (e.g.
mentioned uncertainties, a safe optimal structure is derived. This is batteries) and mobile energy storage (e.g. electric vehicle parking lot,
because it is possible that by taking into account the uncertainties, in a EVPL). The proposed model aims to reduce the expenses associated with
scenario, energy supply (demand) by sources (consumers) is lower construction, maintenance, and storage deterioration. RESs are given
(higher) than the predicted level (deterministic case). Thus, to extract a the highest priority in terms of energy supply allocation in this particular
reliable solution, a large number of storage elements need to be applied model. The storage elements mentioned above are utilized to mitigate
compared to that of the deterministic modeling in the islanded system so the discrepancy between demand and RESs' supply. Ref. [13] introduces
that a dependable solution is achieved. a scholarly examination of the economic, dependable, and sustainable
design of a hybrid PV-biowaste-fuel cell system. This design relies on
1.2. Literature review hydrogen energy storage and incorporates the utilization of the whale
optimization algorithm (WOA). The analysis takes into account the
The topic of planning off-grid hybrid systems has been extensively availability of components over a projected useful lifespan of 20 years
discussed in academic literature. Ref. [6] provides a thorough assess­ for the project. In a study conducted by researchers, a hybrid RES that
ment and juxtaposition of various combined systems involving Proton combines PV, WT, and battery technologies has been suggested as a
Exchange Membrane Fuel Cell (PEMFC) and Solid Oxide Fuel Cell potential solution for powering off-grid telecom towers located in rural
(SOFC) both with batteries. The evaluation is conducted with a focus on regions [14]. Due to the considerable capital investment required and
their suitability for mobility and other grid-disconnected applications. the intermittent availability of RESs, it is imperative to derive the
The analysis encompasses multiple aspects, including system arrange­ optimal sizing of hybrid RESs when attempting to achieve both eco­
ments, technical characteristics, energy management techniques, and nomic and reliable power supply. A metric referred to as the excess
practical tests. The project discussed in [7] pertains to the electrification energy generation (EEG) index has been devised to mitigate the surplus
of three villages in India. This will be achieved through the imple­ renewable energy that is unable to be sold to the grid. Formulations of
mentation of an islanded hybrid RESs. The objective of optimizing the three objective functions, namely, probability of power loss, electricity
control, sizing, and component selection of hybrid energy systems is to cost, and EEG are provided. These functions' nature is one of inherent
ensure the provision of a cost-effective power solution for society. incompatibility and contradiction. Therefore, a discrete multi-objective
Ref. [7] employs the Genetic Algorithm (GA) and HOMER program to Grey Wolf Optimization (GWO) algorithm was adopted to reach the
find the minimum amounts of the overall net preset cost, energy cost, optimal sizing of hybrid RESs. Optimally designing off-grid hybrid RESs
demand not supplied, and carbon dioxide level. The objective of the poses a significant challenge, often requiring the resolution of conflict­
study conducted by researchers [8] is to identify and categorize effective ing objectives [15]. The simultaneous consideration of the levelized cost

2
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

of energy and CO2 pollutant gas has been explored in [15] through the analytic hierarchy approach is employed to determine the optimal lo­
utilization of the ε-constraint method in conjunction with the particle cations for district heating in Gaziantep, Turkey. In [22], a proficient
swarm optimization (PSO) algorithm. energy management system has been introduced, utilizing a MAS to
Various studies were carried out in the area of application and ca­ regulate energy sources inside a local grid. This system is implemented
pabilities of the sources and energy storage devices. Ref. [16] presents a in conjunction with many environmentally-friendly residential build­
versatile multi-objective optimization methodology for assessing and ings. The primary aim of the study conducted in [23] is to evaluate and
implementing vehicle-to-grid and grid-to-vehicle technologies, taking utilize the capabilities of expansive industrial energy hubs in enhancing
into account technological, economic, and environmental consider­ the resilience of power systems. Mobile energy storage and demand
ations. In addition, the life cycle of electric vehicle (EV) batteries, the response programs are employed as strategies to mitigate unintentional
charging and discharging patterns, and the driving behaviors of EV demand shedding within a system and improve system resilience.
owners are taken into account. In [17], a robust multi-objective opti­ Table 1 presents a compilation of various research articles pertaining
mization approach is presented, aiming to determine the optimal plan to the subject matter being investigated.
for the retailer by considering uncertainties in RES and pricing. A multi-
objective model is utilized to evaluate the integration of various demand
response programs, taking into account the cost to merchants and the 1.3. Research gaps
advantages for customers. This model enables the examination of how
actions made by different stakeholders affect one another. The authors The realm of islanded system scheduling exhibits significant study
of [18] put out a proposition for a cooperative multi-agent system (MAS) gaps, as indicated previously and Table 1. The specifics of these gaps are
designed specifically for home microgrids, with the aim of achieving outlined below:
continuous real-time operation. The real-time MAS engages with agents
in home microgrids to function as autonomous learners by employing a - In consumption areas, electrical and thermal energy is generally
distributed and cooperative reinforcement learning approach. This in­ consumed simultaneously. However, in most studies such as
volves the sharing of state variables to effectively coordinate their real- [6–10,12–15], the design of the island system structure has been
time actions. The authors in [19] employed an enhanced Markov Chain based on the supply of electrical energy, and the concurrent supply of
Artificial Neural network methodology for the purpose of load fore­ various types of energies in the mentioned system was not
casting. The design proposal entails a two-step implementation pro­ completely addressed by scholars.
cedure, which takes into account several statistical elements, including - The bio-waste power generation (BPG) unit is capable of simulta­
daily and weekly load, date and time of the year, environmental vari­ neously generating electricity and heat power based on environ­
ables, and user behavior during weekdays and weekends. The study mental waste [3]. So, combining BPG within the islanded system will
referenced in [20] presents a mixed-integer linear programming (MILP) impact providing the mentioned energies at the same time with the
model designed to enhance the load factor by employing bulk sched­ aim of high energy efficiency. Nonetheless, the presence of this type
uling of residential, commercial, and industrial electrical appliances of renewable source in the system has been included in few pieces of
throughout the day. This model takes into account a baseline demand research such as [12,13], which has considered only the electric
and an hourly rate. The objective of the study conducted in [21] is to energy generation model for BPG.
develop and execute a model for selecting suitable sites for district - Batteries have been employed in multiple research studies to bridge
heating using a geographic information system (GIS). The methodology the disparity between supplied power by RESs and the demand
utilized in this study incorporates the utilization of Boolean and fuzzy profile. One notable attribute of the battery is its high efficiency,
toolboxes to conduct an analysis of many layers. Subsequently, the typically reaching approximately 90 %, along with its remarkable
power density. However, the installation cost of this storage system is

Table 1
A review of the literature stated in this paper.
Ref. Energy Investigate CAES capability in CHP-based BPG modeling in IHS Considering energy consumption of Uncertainty
consumption IHS EVPL modeling

[6] Electricity No No No No
[7] Electricity No No No No
[8] Electricity No No No No
[9] Electricity No No No No
[10] Electricity No No No No
[11] Electricity and No No No Stochastic
heat optimization
[12] Electricity No Considering only electricity management Yes Robust optimization
[13] Electricity No by BPG No No
[14] Electricity No No No No
[15] Electricity No No No No
[16] Electricity No No Yes Stochastic
optimization
[17] Electricity No No No Robust optimization
[18] Electricity No No Yes Stochastic
optimization
[19] Electricity No No No No
[20] Electricity No No No Stochastic
[21] Electricity and No No No optimization
heat
[22] Electricity No No No No
[23] Electricity and No No No Stochastic
heat optimization
Current Electricity and Yes Yes Yes PEM
paper heat

IHS: Islanded hybrid system, CAES: compressed air energy storage, BPG: bio-waste power generation, PEM: point estimate method.

3
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

relatively high, its useful life is limited, there are size constraints, and devices, and power electronic converters. This issue is governed by the
there is also the additional cost associated with battery wear operational framework encompassing RESs [51,52], storage elements,
resulting from the charge/discharge function [12]. In order to EVPL, and converters. Providing thermal (electricity) energy using the
address this concern, certain researchers (e.g., [13]) have employed RESs such as BPG (BPG and WT) is of high importance concerning the
hydrogen storage as a means of mitigation. Special precautions must operational model of the discussed equipment. The active power con­
be taken in order to store hydrogen in this particular storage device. sumption of EVs is also a decision variable, the amount of which is
Furthermore, it is worth noting that the discharge capacity of this determined as per the goals of the system operator (objective function)
particular storage system is relatively low, estimated to be approxi­ and the owners of EVs (providing the required energy for travel in the
mately 50 %. Consequently, the energy loss experienced in this operating horizon). Then, if the produced energy of RESs is more (less)
storage system is significantly high, as indicated by previous research than the demand, storage elements (CAES and TES) operate in charge
[13]. In order to address these instances, it is anticipated that com­ (discharge) mode. The mentioned problem has uncertain parameters
pressed air energy storage (CAES) can offer advantageous potential related to the load, renewable power, and EVPL. To model these un­
in the planning of island systems. Due to its relatively higher effi­ certainties, stochastic optimization based on point estimate method
ciency of approximately 80 %, which surpasses that of hydrogen (PEM) is adopted. In this method, the reliable solution can be accessed in
storage. Additionally, it should be noted that this particular storage the least number of scenarios. In the following, Ant Lion Optimizer
solution offers the advantage of a relatively low installation cost and (ALO) is used to extract a reliable optimal solution. Upon careful ex­
a long useful life. Furthermore, it has the capability to provide a amination and analysis of the suggested scheme in relation to the
larger storage capacity compared to traditional battery systems, as existing research background, the following improvements can be
highlighted in a previous study [24]. attributed to this plan:
- Due to the reduction of environmental emissions, the use of EVs may
grow dramatically in the future [25]. Therefore, in the consumption - Planning a 100 % renewable islanded hybrid system to optimize the
area, there is also the presence of EVs, because batteries of the ve­ size and number of sources and storage elements so as to simulta­
hicles must be charged by connecting them to an electric source. A neously provide electricity and heat power.
smart charging model exists for EVs wherein the EVs regulate their - Providing a modeling of the energy consumption of EVPLs in the
active power consumption based on the requirements of the con­ hybrid island system based on smart charging,
nected electric system and the preferences of their owners. This - Formulating the performance of the BPG unit by taking into account
matter can be employed to address the disparity between the power its ability to generate electricity and heat concurrently,
characteristics of RESs and the power requirements of EVs. There­ - Using compressed air energy storage in off-grid system to reduce
fore, the availability of EVs in the islanded system and their energy planning cost and energy efficiency,
management can be effective in reducing the number of stationary - Deriving a reliable optimal solution to deal with the prediction errors
storage generators, which will reduce the investment cost of the of the load, renewable power and EVPL parameters using PEM.
islanded system. However, it is worth mentioning that the presence - Objectives of the proposed scheme in this paper are as follows:
of EVs in the island system has been considered by only few studies, - Simultaneous supply of electrical and thermal energy, taking into
including ref. [12]. account the hybrid renewable island system,
- In the islanded system's planning problem, there are uncertain - Supplying the energy consumption of electric vehicles, taking into
quantities related to load, renewable power and parameters of EVs. account their smart charging model,
The prediction error of the mentioned parameters may result in a - CHP-based BPG modeling,
different structure than that of the deterministic modeling of the - The use of compressed-air storage in the combined system to reduce
scheme. Therefore, to reach a dependable optimal solution, the un­ the planning cost.
certainties are modeled. This topic has been less researched such as - Modeling various uncertainties with the PEM method
[11,12].
1.5. Paper organization
1.4. Contributions
Different sections of the paper are structured as described here.
This paper aims to address the existing research gaps by presenting a Section 2 expresses the mathematical expressions related to the sug­
model for the scheduling of an islanded hybrid system. The system in­ gested hybrid system planning by considering its uncertainty modeling.
corporates WT, BPG, and stationary storage [26,28] elements to ensure Section 3 explains the process of solving the problem using the ALO.
the concurrent provision of electricity and heat power, as well as EVPL. Section 4 discusses numerical reports from various study cases. And,
The stationary storage in the electrical sector is of the CAES type, which Section 5 gives the conclusions.
is a combination of motor, compressed air tank (CAT) and a generator. In
the charging mode of this storage, motor converts electricity into com­ 2. Planning of the electrical and thermal hybrid system
pressed air and stores it in the CAT. In the discharge mode, the generator
delivers the compressed air stored in the CAT to the island system by 2.1. Suggested scheduling
converting it into electrical energy. The stationary storage in the thermal
sector includes thermal energy storage (TES). This storage device stores Fig. 1 shows the suggested schematic for the hybrid island system.
thermal energy (delivers to the island system) in charging (discharging) The electrical sector includes both AC and DC buses. Considering the
mode. In the given plan, EVs [41–45] based on the smart charging frequency-variable output voltage of WT and BPG [12], these generation
strategy receive their energy consumption from the hybrid system. units are firstly linked to the DC bus via an AC/DC converter to create a
Simply put, they control their active power in the charging mode as per constant frequency corresponding to the operating frequency of the AC
the request of the hybrid system operator (minimizing the planning cost) load. Then, a DC/AC converter connects this bus to the AC bus. The
and EV owners (charging the EV battery at a certain rate for a certain converter creates a constant frequency in the AC bus. Moreover, EVs are
period of time). In this article, the discharge mode of EVs [46,47,50] is equipped with two distinct charger types: one that incorporates an AC/
omitted due to the reduction of the useful life of their batteries in these DC converter and another that relies on a DC/DC converter [67]. The
conditions. The scheme is introduced in the form of an optimization study assumes that EV owners incorporate an AC/DC model. So, as
problem that seeks to minimize the aggregate annual expenses related to shown in Fig. 1, EVPL is only linked to the AC bus. Moreover, EVs are
the construction, maintenance, and operation of sources, storage assumed to be the only energy consumers and operate only in charging

4
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Fig. 1. Structure of the suggested islanded hybrid system.

mode when connected to the electrical system. EVs can also operate in proposed in consideration of the elimination of harmful environmental
discharge mode by connecting to the electrical system. However, this pollutants:
issue results in an increase in the number of charge/discharge cycles of
the EV battery during the day, which leads to a decrease in the useful life 1) In the electrical sector, WT and BPG are the main sources to supply
of the EV battery [12]. Since the EV battery cost is high [12], it is the consumption energy of AC load and EVPL.
assumed in this paper that the EV owner is not inclined to operate EVs in 2) In the electrical sector, provided that the overall energy production
the discharge mode when connected to the electrical system. CAES has by WT and BPG exceeds the consumed energy, CAES will function in
an engine, generator, and CAT. The engine operates in CAES charging charging mode, if not, it will be active in discharging mode.
mode receives active power from the electrical system and then stores 3) In the thermal sector, BPG is the main responsible for feeding energy
that compressed air in the CAT. In the discharge function of CAES, a consumption.
generator is used, which receives compressed air from the CAT and 4) In the thermal sector, provided that the energy production by BPG
produces active power at its output. Generally, to have high efficiency in exceeds the energy consumed, TES functions in charging mode, if
CAES, the motor and generator are of AC type [24]. Therefore, in Fig. 1, not, it is active in discharging mode.
CAES is linked to the AC bus. In the electrical sector, AC load and EVPL
are supplied by WT and BPG as renewable sources. Then, CAES makes
2.2. Problem formulation
the power profile of load and EVPL and power profile of resources as
close as possible to each other. In other words, in the hours when the
This part of the paper deals with the planning formulation of the
energy produced by WT and BPG is higher than that demanded by the
suggested hybrid system as per Fig. 1. The objective function attempts to
AC load and EVPL, the surplus energy is stored in CAES. Additionally,
reach the minimum overall yearly cost of construction and maintenance.
when the energy consumption exceeds the energy generation, the lack of
The operation model of generation units, storage elements, and power
consumption energy is supplied by the CAES. The thermal sector has just
electronic converter are the constraints. The problem is described in the
one node. In this node, BPG sends its thermal power to feed the thermal
following paragraphs.
load. Then, TES as a storage [68,69] covers the difference between the
generated and consumed thermal power profiles, which performs the
2.2.1. Objective functions
same as CAES. The following is a list of the performance parameters that
In this proposed scheme, the goal is to achieve the minimum annual
must be met by resources and storage devices to achieve the highest
overall net present cost (ATNPC) of sources, power electronic con­
possible level of efficiency in the hybrid island system that has been
verters, and storage elements in the form of a hybrid system like Eq. (1).

5
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

ATNPC equals the sum of annually capital net present cost (ACNPC) and and scenario w (HBPG
t,w ) (kW) is a multiplication of its active power, as
operation and maintenance annually net present cost (O&M-ANPC) for expressed in Eq. (7) [4]. In this equation, parameters ηhBPG and ηlBPG
the mentioned elements [11,12]. The formulation of these two costs is display heat conversion efficiency and loss efficiency, respectively,
presented in rows 1 to 2 of Eq. (1), respectively. The mentioned costs are which are expressed in %. Finally, the number of BPGs to be allocated to
in $. In this equation, the variables NWT, NBPG, NM, NCAT, NG, NT, and NI the hybrid system is limited by Eq. (8), in which NBPG represents the
respectively represent the number of WT, BPG, motor (M), CAT, maximum number of BPGs.
generator (G), TES, and power electronic converters to be allocated to
the hybrid system. C and M parameters with indices WT, BPG, M, CAT, PBPG
t,w = N
BPG BPG BPG
η Gt,w LHV BPG ∀t, w (5)
G, T, and I represent annual construction cost and maintenance cost for
WT, BPG, motor, CAT, generator, TES, and power electronic converter. ρCH4
LHV BPG = LHV CH4 (6)
These parameters are expressed in terms of $. Coefficient λ is also 100
adopted to turn the construction cost into the net present cost, whose
formulation can be written in Eq. (2) [2]. In these equations, r is the (1 − ηBPG − ηlBPG )ηhBPG
BPG
Ht,w = PBPG
t,w ∀t, w (7)
interest rate expressed in % and n denotes the number of scheduling ηBPG
years.

CNPC
⏞̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( ̅⏟⏟̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅)⏞
min ATNPC = λ N WT CWT + N BPG CBPG + N CAT CCAT + N M CM + N G CG + N T CT + N I CI
O&M− NPC
(1)
(⏞̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅⏟⏟̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅)⏞
+ N WT M WT + N BPG M BPG + N CAT M CAT + N M M M + N G M G + N T M T + N I M I

where:
{ }
(1 + r)n − 1 N BPG ∈ 1, 2, ..., N BPG (8)
λ= (2)
r(1 + r)n
2.2.4. EVs parking lot model
2.2.2. Wind turbine constraints Based on Section 2.1, this paper assumes that EVs connected to the
The total active power produced by WTs at hour t and scenario w electrical system only operate in charging mode, so as not to decrease
(PWT the real life of the EV battery. The active power of EVPL in charge mode,
t,w ) in the proposed hybrid system can be extracted based on Eq. (3)
that is Pc,EV, is in the form of decision variables, whose performance
[2]. Within the context of this equation, WT operates in four distinct
range is proportional to Eq. (9). This constraint represents the charging
domains. In the case the wind speed parameter (v in m/s) is less than the
rate limit for EV batteries in EVPL [25]. CR parameters also indicate the
threshold speed (vci in m/s), the WT is disconnected. But if it is between
charge rate of total EVs in EVPL. In Eq. (10), energy stored in EV bat­
vci and rated speed (vr in m/s), PWT increases linearly with NWTpr coef­
teries (EEV in kWh) is calculated. EEV at hour t and scenario w based on
ficient. The pr parameter denotes the nominal power output of WT in
Eq. (10) depends on EEV in the previous hour (EEV t− 1,w ), the initial energy
kW. In the third operating area, if the wind speed varies between vr and
cut-off speed (vco in m/s), WT produces a constant power of NWTpr. And, of EVs in the hour of connection to the system (EAt,w in kWh), the
when the wind speed exceeds vco, WT is off. The installable number of consumed energy required for traveling EVs in the hour of disconnection
WT in the hybrid system is also given in Eq. (4) in the form of an integer from the system (EDt,w in kWh), and energy stored in charge mode
variable, which has a maximum value of NWT [2]. (ηc,EV Pc,EV
t,w ) [12,25]. Parameter η
c,EV
is the charging efficiency of EVs.
⎧
0 Based on Eq. (11), the energy stored in EVPL must always have a positive
⎪
⎪
⎪
⎨ ci vt,w ≤ vci and vt,w ≥ vco value. The remarkable point about the performance of EVs is that there
WT r vt,w − v
WT
Pt,w = N p r vci ≤ vt,w ≤ vr ∀t, w (3) various EVs are available in the hybrid system every hour. Also, the
⎪ v − vci
⎪
⎪
⎩ v r
≤ vt,w ≤ vco values of CR, EA, and ED will vary over time. The amount of CR at hour t
equals the overall charge rate of all EVs linked to the system. EA will also
WT r
N p
{ } be equal to the total initial energy of EVs newly linked to the system at
N WT ∈ 1, 2, ..., N WT (4) hour t. The initial energy of a single EV equals the product of the battery
capacity (in kWh) and the state of charge (SOC in %). ED is also equal to
2.2.3. Bio-waste power generation constraints the overall demanded energy for the travel of EVs recently disconnected
The main source of biowaste is replaceable plants and animals; thus, from the system [12,25]. The paper assumes that each EV is fully
plants that can be replaced over and over are an invaluable source of charged, that is, the energy required for the travel of each EV equals the
energy for bio-waste power generation. The waste is used in the bio­ capacity of its battery.
waste process and turned into a source of energy. By doing this, the
waste disposal and uselessness are terminated or at least reduced and the 0 ≤ Pc,EV
t,w ≤ CRt,w ∀t, w (9)
amount of pollutant emission is significantly decreased [3]. The overall
active power generation by BPGs at hour t and scenario w (PBPG
EV A
+ Et−EV1,w − Et,w
D
+ ηc,EV Pc,EV (10)
t,w ) (kW), as
Et,w = Et,w t,w ∀t, w
depicted in Eq. (5), equals the product of the number of BPGs, electrical EV
(11)
conversion efficiency (ηBPG in %), gas flow at hour t and scenario w
Et,w ≥ 0 ∀t, w
3 BPG
(GBPG
t,w in m /h), and lower heating value (LHV in kWh/m3) [12]. Eq. It is noteworthy that there are generally two charging strategies for
(6) expresses the amount of biogas heating that depends on methane EVs, the details of which are as follows:
ρCH4 as the methane percentage and LHV CH4 as the lower heating value
of methane [13]. The total thermal power produced by BPGs at hour t

6
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

- Non-smart charging strategy: In this strategy, EVs perform charging ⎛ ( )⎞

operations once linked to the system. They are fed from the system max PGt,w
by active power equal to their charging rate. Therefore, after a N = round⎝
G ⎠ (18)
PG
certain charging period, their batteries are fully charged and they do
not receive power from the system. In the described method, Eq. (9)
2.2.6. Thermal energy storage constraints
will be as Pc,EV
t,w = CRt,w [70].
As was discussed in Section 2.1, if the heat produced by BPGs exceeds
- Smart charging strategy: In this strategy, the charging operation of EVs
the heat consumed (HL in kWh), that is Ht,w BPG
≥ HLt,w , the excess thermal
follows the hybrid system operator's and EV owners' goals. The
formulation of EVPL performance in this strategy is given by Eqs. power produced (HBPG L
t,w − Ht,w ) is stored in TESs. So, based on Eq. (19),
(9)–(11) [25]. the energy stored in TESs at hour t and scenario w (ETt,w in kWh) will be
the sum of ET at the last hour (ETt−
1,w ) and the surplus of thermal power
2.2.5. Compressed air energy storage (CAES) formulation ( )
produced from the perspective of TESs (ηc,T HBPG L
t,w − Ht,w ) [4]. The
The performance model of CAESs is written in Eqs. (12)–(18). As was
stated in Section 2.1, if the sum of the active power produced by WTs parameter ηc,T (in %) is the charging efficiency of TES. However, pro­
and BPGs from the point of view of the AC bus at hour t and scenario w, i. vided that the thermal load exceeds the heat power generation of BPGs,
( )
that is HBPG L
t,w < Ht,w , TESs are in discharge mode and meet the heat
e., PWT BPG I I
t,w + Pt,w η η , in the electrical sector exceeds the sum of the active
deficiency. In this situation, according to Eq. (20), ETt,w equals the dif­
power of the AC load (PL in kW) and EVPL active power from the point of
ference between ETt− 1,w and the demand deficiency from TESs' perspective
view of the AC bus, i.e. PLt,w + Pc,EV
t,w , then the motors in CAESs will receive ( )
active power from the AC bus. So, based on Eq. (12), the active power of (η1c,T HLt,w − Ht,w
BPG
) [4]. The parameter ηd,T (in %) is the discharge effi­
( )
motors (PM in kW) equals the difference between PWT BPG I I ciency of TES. In constraint (Eq. (21)), the limit of stored energy in TESs
t,w + Pt,w η η and
I is formulated. The parameters ET and ET (in kWh) represent the mini­
PLt,w
+ Pc,EV
t,w .
In this equation, η represents the efficiency of the electronic
mum and maximum energy stored in a TES, respectively. Finally, the
power converter in %. In this condition, the generators' active power in
limit of the number of TESs is presented in Eq. (22), where NT is the
CAES (PG in kW) is zero. However, if the overall active production power
( ) maximum number of TESs allocated to the hybrid system.
of renewable resources in the AC bus ( PWT BPG I I
t,w + Pt,w η η ) is less than ( )
G
T
Et,w = Et−T 1,w + ηc,T Ht,w
BPG L
− Ht,w BPG
if Ht,w L
≥ Ht,w ∀t, w (19)
PLt,w + Pc,EV
t,w , based on Eq. (13), the P will be equal to the difference of
( )
Pt,w + Pt,w and Pt,w + Pt,w η η . In this situation, PM is equal to zero. In
L c,EV WT BPG I I
1 ( L )
T
Et,w = Et−T − BPG
Ht,w − Ht,w BPG
if Ht,w L
< Ht,w ∀t, w (20)
Eq. (14), the energy stored in CATs is found for hour t and scenario w 1,w
ηc,T
CAT
(ECAT
t,w in kWh). It depends on E in the last hour (ECAT
t− 1,w ), the energy fed
N T ET ≤ Et,w
T
≤ N T ET ∀t, w (21)
by motors (ηM PM
t,w ), and the energy injected into the system by genera­
M G
tors (ηG PGt,w )
1
[16]. Also, η and η are the efficiency of the motor and { }
N T ∈ 1, 2, ..., N T (22)
generator in %, respectively. In Eq. (15), the limit of energy stored in
CATs is included, where ECAT andECAT (in kWh) state the boundaries of 2.2.7. Converter constraints
energy stored in a CAT, respectively. The number of CATs in the hybrid Based on Fig. 1, the number of power electronic converters (NI)
system (NCAT), based on Eq. (16), is an integer variable that has a equals the sum of the number of converters in the paths of WTs (NI-W),
maximum value ofNCAT . In Eqs. (17) and (18), respectively, the number BPGs (NI-B), and AC bus (NI-A). This is presented in Eq. (23). NI-W, NI-B
of motors (NM) and generators (NG) that can be installed in the hybrid and NI-A are found by dividing the maximum power flow through the
system are calculated. NM (NG) equals the ratio of the maximum active converters by the converter's rated power (PI in kWh) based on Eqs.
power of motors (generators) to the nominal power of the motor (24)–(26). In the converters placed on the path of WTs (BPGs), the
(generator). Rated powers of motor and generator in Eqs. (17) and (18) maximum active power of WTs (BPGs) flows. In the converters in the AC
are represented by parametersPM and PG (in kW). bus path, the power PM or PG flows.
(( ) ( ))
PM
t,w = PWT BPG
t,w + Pt,w ηI ηI − PLt,w + Pc,EV
t,w , PGt,w N I = N I− W
+ N I− B + N I− A
(23)
( ) ( )
= 0 if PWT BPG
t,w + Pt,w ηI ηI ≥ PLt,w + Pc,EV ∀t, w (12) ⎛ ( )⎞
t,w
max PWT
t,w
(( ) ( ) ) N I− W
= round⎝ ⎠ (24)
PI
PGt,w = PLt,w + Pc,EV
t,w − PWT BPG
t,w + Pt,w ηI ηI , PM t,w
( ) ( ) ⎛ ( )⎞
WT BPG I I L
= 0 if Pt,w + Pt,w η η < Pt,w + Pt,w c,EV
∀t, w (13) max PBPG
t,w
N I− B
= round ⎝ ⎠ (25)
PI
PGt,w
CAT
Et,w = Et−CAT M M
1,w + η Pt,w − ∀t, w (14)
ηG ⎛ ( )⎞
max PM G
t,w + Pt,w
N CAT ECAT ≤ Et,w
CAT
≤ N CAT ECAT ∀t, w (15) N I− A
= round⎝ ⎠ (26)
PI
{ }
N CAT ∈ 1, 2, ..., N CAT (16) The scheme under consideration incorporates a mathematical model
( )⎞ [71–75]. The mathematical model in this paper is based on the opti­
⎛
max PM mization formulation [76–80]. Optimization model comprises an
t,w
N M = round⎝ ⎠ (17) objective function. Objective function includes min (max) term to
PM
calculate minimum (maximum) value of this function. It expresses as
single or multi-objective function. Optimization problem includes the
different constraints [101–104]. Constraints are as equality and

7
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

inequality formulation. Constraints are as non-linear, linear, mixed / ( )

integer non-linear or mixed integer linear models. In order to implement ωl,k = ( − 1)3− k
ξl,k ξl,1 − ξl,2 , k = 1, 2 (34)
the optimization model on a network or hybrid system, it is necessary to
have intelligent devices [122–126]. Smart systems are based on tele­
communications equipment. These systems use to coordinate of the • Step 9: Update E(Sa ):
different power elements. In this condition, the management of network ∑
2
is easy and it includes high speed in system processing. E(Sa ) = E(Sa ) + ωl,k .(S(l, k) )a , a = 1, 2 (35)
k=1

2.3. Modeling uncertainties

• Step 10: Repeat steps 3 to 9, until all uncertainty variables are
involved.
In the problem (1)–(26), load (PL and HL), gas flow in BPGs (GBPG),
• Step 11: Calculate the impact coefficient of the mean point:
wind speed (v), and parameters of the EVPL (CR, EA, and ED) are un­
certainties. In this section, PEM provides a modeling of the aforemen­ ∑
m
/( ( )2 )
tioned uncertain quantities as stated in the following sentences. This ωμ = 1 − 1 λl,4 − λl,3 (36)
particular method is characterized by its ability to extract the minimum
l=1

number of scenarios in comparison to alternative stochastic program­

ming methods. Therefore, the optimal solution with high confidence and • Step 12: Update E(Sa ):
minimal computational time is achieved. This article employs the PEM ( )a
E(Sa ) = E(Sa ) + ωμ . Sμ , a = 1, 2 (37)
(Parameter Estimation Method) approach, utilizing the 2n + 1 model. In
this model, the variables n and 2n + 1 represent the number of uncertain
quantities and derived scenarios. The specifics of the approach are • Step 13: Find the mean and standard deviation of the output un­
outlined below [136]. certain parameter as follows:
( )
• Step 1: Find the number of uncertainties (n). μS = E S 1 (38)
• Step 2: Put E(Sa ) = 0 (a = 1, 2), where a represents the output torque √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( ) ( ( ) )2̅
index, and S as an optimization problem. σS = E S2 − E S1 (39)
• Step 3: Select the uncertainty parameter (Z1).
• Step 4: Calculate skewness (λzl ,3 ) and kurtosis (λzl ,4 ) of Z1 according
3. Solution process
to Eqs. (27) and (28), respectively:
[( )3 ]
E z l − μz l Problem (1)–(26) is in the format of mixed-integer non-linear pro­
λzl ,3 = ( )3 (27) gramming. In this section, ALO [137] helps find a solution to the
σ zl problem. According to [137], ALO can derive the optimal solution to
[( )4 ] complex engineering problems. In this solver, N (representing the pop­
E z l − μz l ulation size) random values are produced for decision variables
λzl ,4 = ( )4 (28) including NWT, NBPG, Pc,EV, NCAT, and NT based on their allowed limits
σ zl
using Eqs. (4), (8)–(9), (16) and (22), respectively. Then, N values are
found for dependent variables, like PWT, PBPG, HBPG, PM, PG, EEV, ECAT, ET,
where, μzl and σzl are the mean and standard deviation of Z1, respec­
NM, NG, NI, NI-W, NI-B, NI-A, and ATNPC based on Eqs. (1), (3), (5)–(7),
tively. E(f) is the mean value (expected value) of f, calculated for the
(10), (12)–(14), (17)–(20), and (23)–(26) and the values of the decision
term within Eqs. (27) and (28):
variables. In this section, to comply with constraints (11), (15), and (21),
[( )3 ] ∑N
( )3 ( ) the penalty function technique can be adopted [138]. The fitness func­
E zl − μzl = zl,j − μzl × Prob zl,j (29) tion (FF) equals the sum of the objective function (1) and the penalty
functions resulting from constraints (11), (15), and (21). The penalty
j=1

[( function for the condition a ≤ b is μ × max(0, a – b) [138]. Note that μ ≥

)4 ] ∑N
( )4 ( )
E zl − μzl = zl,j − μzl × Prob zl,j (30) 0 is the Lagrangian multiplier whose value is found in a similar way the
j=1 decision-making variables were obtained. The formulation of the FF
( ) based on penalty functions is presented in Eq. (40). Next, the updating of
N and Prob zl,j denote the number of scenario samples and the
the decision variables based on their allowed limits is done by ALO.
probability of individual samples, respectively.
Here, ALO derives N new values for decision variables and μ according
to FF's best value in the previous step. Then N values of dependent
• Step 5: Find two standard locations (ξ) using Eq. (31).
variables and FF are found using new values of decision variables and μ.
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
λz ,3 3( )2 The process will be terminated once the convergence point is reached.
ξl,k = l + ( − 1)3− k λzl ,4 − λz ,3 , k = 1, 2 (31) The convergence point can be reached after the maximum convergence
2 4 l
iteration (I). Fig. 2 illustrates the steps of problem solution using the
ALO.
• Step 6: Calculate two approximated locations as per Eq. (32).
FF = ATNPC
zl,k = μzl + ξzl ,k .σzl , k = 1, 2 (32) ∑ ( )
+ μEVt,w
EV
max 0, − Et,w
t,w
∑( ( ) ( ))
• Step 7: Solve the deterministic problem by taking into account esti­ μCAT CAT CAT CAT
+ μCAT max 0, N CAT ECAT − Et,w
CAT
+ t,w max 0, Et,w − N E
mated locations: t,w
t,w

( ) ∑( ( ) ( ))
Sl,k = f μz1 , μz2 , ..., zl,k , ..., μzn k = 1, 2 (33) + μTt,w max 0, Et,w
T
− N T ET + μTt,w max 0, Et,w
T
− N T ET
t,w

(40)
• Step 8: Express the impact factor (ω):

8
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Fig. 2. Flowchart of the solution process.

of BPG depends on the maximum gas flow (0.4 m3). The expected curve
4. Numerical results and discussion
of wind speed rate and passing gas BPG is presented in [12 and 139]. The
data of WT, BPG, CAES, EVPL, TES, and converter can be found in
4.1. Case study
Table 2. This paper considers two smart and non-smart charging stra­
tegies for EVs, which are detailed in Section 2.2.5.
The system shown in Fig. 1 is adopted to test the proposed scheme
As in [25,70], EVs will be connected to the network after their last
using the data of Espoo (60◦ 12′20″N, 24◦ 39′20″E), a city in Finland
travel. So, the expected daily curve of the penetration rate of EVs in
[11,12,139]. The peak values of electrical and thermal demand in this
EVPL for the two mentioned charging strategies is based on [70]. The
city are 21 kW and 10 kW, respectively. The amount of load in each hour
number of grid-connected EVs in EVPL in one hour can be found by
is the multiplication of the peak load and load factor in this hour
multiplying the penetration rate of EVs in this hour and the number of
[140–149]. The expected daily curve of the electrical load factor can be
EVs. According to [12], considering that EVs are connected to the system
found in [12,139], and it is plotted for thermal load in [11]. The
after their last trip of the day, all EVs are present in EVPL at 1:00–4:00
maximum wind speed of the area is 9.3 m/s, and the daily wind speed
and 23:00–00:00. They depart from EVPL between 5:00 and 9:00, with
profile is a factor of the maximum wind speed. The daily gas flow profile
the least number of departures at 5:00 and 9:00, and the most departures

9
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Table 2 all other hours is zero. Eventually, the overall proposed scheme can be
Specifications of the study case. implemented on various data of renewable resources and storage as well
r (%) [6] 10 NCAT 150 as different economic data with no limits on its implementation.
n (year) 20 CM ($) 700
Characteristics of WT [12]: MM ($/year) 1
pr (kW) 1 CCAT ($) 500
4.2. Case studies
vci (m/s) 2.5 MCAT ($/year) 0.40
vr (m/s) 10 CG ($) 750 In this section, five different study cases are analyzed to assess the
vco (m/s) 13 MG ($/year) 1.2 described scheme, which include:
Life span (year) 20 Characteristics of TES [11]:
NWT 50 ET (kWh) 1.35
CWT ($) 3200 0.10
- Case I: Suggested scheme by taking into account only the electrical
ET (kWh)
MWT ($/year) 5 ηc,T (%) 80 energy model without EVPL
Characteristics of BPG ηd,T (%) 80 - Case II: Case I with non-smart charging strategy
[4,12,13]: - Case III: Case II with smart charging strategy
ηBPG (%) 35 Life span (year) 20 - Case IV: Suggested scheme by modeling thermal and electrical en­
ηlBPG (%) 15 NT 40
ergy and using a non-smart charging strategy
ηhBPG (%) 50 CT ($) 2000
ρCH4 (%) 65 MT ($/year) 1.5 - Case V: Case IV with smart charging strategy.
LHVCH4 (kWh/m3) 10 Characteristics of converter
[2]:
4.3. Results
Life span (year) 20 ηI (%) 90
N BPG 30 PI (kW) 1
CBPG ($) 6500 Life span (year) 15 The design under consideration has been implemented and simu­
MBPG ($/year) 16.25 CI ($) 800 lated as per the data provided in Section 4.1 along with the problem-
Characteristics of CAES MI ($/year) 0.40 solving process depicted in Fig. 2 in the MATLAB software environ­
[24]:
ment. In the following, the numerical reports related to the study cases
ηM (%) 81 Characteristics of EVPL
[150,151]: are presented.
ηG (%) 79 EVs number 10
PEL (kW) 1 EV charge rate (kW) 3 4.3.1. Evaluation of convergence of the hybrid system planning problem
PFC (kW) 1 EV battery capacity (kWh) 12 Table 3 reports the state of convergence of solving the proposed
ECAT (kWh) 1.35 ηc,EV (%) 90
problem using ALO, hybrid solver by honey bee mating optimization
ECAT (kWh) 0.10 SOC (%) 20
(HBMO) and artificial bee colony (ABC) [12], hybrid algorithm of
Life span (year) 20
HBMO and Grey Wolf optimization (GWO) [11], whale optimization
algorithm (WOA) [13], and GWO [14]. The population size is set at 80,
at 6:00–8:00. EVs also arrive at EVPL at 12:00, with the number of ar­ and the maximum convergence iteration is set to 4000. The rest of the
rivals at 12:00–13:00 and 21:00–22:00. In other hours, the number of parameters in the solvers were chosen using refs. [11–14, 137]. More­
EVs entering the EVPL is almost the same. The non-smart charging over, the solver finds solutions to the mentioned problem 30 times to
strategy for EVs entails their immediate transition into charging mode provide statistical metrics, such as the standard deviation of the final
upon connection to the system, consistently receiving power at a rate response. Further, as per Table 3, the ALO algorithm has been able to
equivalent to their charging capacity from the system. Subsequently, obtain the minimum point in comparison to other mentioned solvers, in
subsequent to the completion of its designated charging duration, the which the ATNPC is equal to $25,503.65, but other algorithms have
system will cease to supply power to it. In this particular scenario, it can earned more than $26,500 for ATNPC. Also, ALO reached the optimal
be conceptualized as being detached from the overarching system. Based point in the least convergence iteration and computing time in com­
on the data presented in Table 2, the charging rate and battery capacity parison to HBMO+ABC, HBMO+GWO, WOA, and GWO. It finds this
of an electric vehicle (EV) are reported as 3 kW (kW) and 12 kW-hours point in 1472 iterations and 4.7 min, but the calculation time of other
(kWh), respectively. The charging duration for each EV is equivalent to solvers exceeds 6.8 min. As another point, ALO provides the minimum
4 h, calculated by dividing 12 h by three. Consequently, each EV that is standard deviation, which is 0.93 %. It can be inferred that the
linked to the system will undergo disconnection from the network mentioned algorithm reaches a much lower dispersion than other
following a duration of 4 h. Given the aforementioned matter and the mentioned algorithms for the proposed problem, or can obtain a more
procedure of incorporating them into the system during the time frame nearly unique solution. Therefore, according to the presented results,
of 12:00–23:00, as indicated by the EVs penetration rate curve in The ALO algorithm has demonstrated its capability to extract the most
reference [12], it can be asserted that the quantity of EVs adhering to precise solution with a notable rate of convergence in comparison to
their non-smart charging strategy from 14:00 to 21:00 is equivalent to 1, alternative solvers.
2, 3, 6, 7, 7, 4, and 1 respectively. Conversely, the number of EVs during
4.3.2. Deriving the economic structure of the hybrid system
Table 4 tabulates the optimal size (number) of sources, storage ele­
Table 3 ments, and power electronic converters in the hybrid system of Fig. 1 for
Convergence state of the hybrid system planning based on the different algo­ study cases I to V. In cases I to III, the thermal energy consumption
rithms in Case V. model is not considered. EVPL is absent in Case I either. Therefore, in
Algorithm ATNPC Convergence Convergence Standard
($/year) iteration time (min) deviation Table 4
(%)
Size of sources, converters, and storages in the hybrid system.
ALO 25,503.65 1472 4.7 0.93
Case NWT NBPG NT NM NCAT NG NI-W NI-B NI-A
HBMO+ABC 26,614.43 1721 6.9 0.98
[12] I 50 3 0 19 106 16 46 3 19
HBMO+GWO 26,925.25 1902 7.4 1.02 II 50 15 0 22 140 31 46 14 31
[11] III 50 2 0 17 101 17 46 2 17
WOA [13] 27,967.31 2131 8.5 1.46 IV 37 27 36 20 116 28 34 25 28
GWO [14] 28,548.46 2654 9.5 1.87 V 37 27 36 20 51 7 34 25 20

10
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Fig. 3. Expected daily active power curve of AC load, non-smart charging-based EVs, and total electrical consumption.

this study, 50 WTs and 3 BPGs are installed in the system. As per Table 2, In total, Case II requires 91 power electronic converters. In Case III, a
the capital and maintenance costs of WT are lower than that of BPG, so smart charging strategy is used for EVPL. In the smart charging strategy,
in Case I, the maximum number of WTs to be established (according to EVs carry out charging according to the demands of the system operator
Table 2 is equal to 50) is selected for supplying electrical energy. Next, to and their owners. This has made the number of BPGs in Case III decrease
cover the difference between the active power profile of the mentioned to 2 compared to Case II. Therefore, the number of converters has been
RESs and the active power of the AC load, the CAES needs 19 motors, 16 reduced to 2. The number of WTs is still 50, and the number of WT
generators, and 106 CATs. To pass the maximum active power of WTs converters is also 46. Number of converters between DC and AC busses is
through the power electronic converter, 46 converters need to be placed 17. Consequently, this scheme needs 65 power electronic converters. In
in the path of WTs. Also, 3 converters are placed in the path of BPGs until this condition, CAES has 17 motors, 17 generators, and 101 CATs. By
they pass the maximum active power of BPGs. Nineteen converters are comparing study cases II and III, if the smart charging strategy of EVs is
also required to pass the maximum active power between DC and AC adopted by the hybrid system operator, a smaller number of converters,
buses. Therefore, Case I requires a total of 68 power electronic con­ BPG, motors, generators, and CATs will be required in the system. In
verters. Case II is the same as Case I, except that EVPL is added in Case II cases IV and V, the thermal energy consumption model is included in
considering the non-smart charging strategy. Thus, the EVPL con­ addition to the electrical energy consumption. Case IV has the same non-
sumption load is added to the AC load based on Fig. 3 from 14:00 to smart charging strategy for EVPL as Case II. The addition of the thermal
21:00, which is the peak EVPL consumption overlapping with the AC load model has resulted in increasing the number of BPGs up to 12 more
load peak hours, i.e. 17:00–21:00, according to Fig. 3. This significantly compared to Case II (the number of BPGs was 27 in Case IV), increasing
increases the electrical energy consumption in the hours 14:00–21:00 in the number of BPG converters up to 11 more (25 converters), and 36
Case II than in Case I. Accordingly, the number of BPGs in Case II in­ TESs are required to cover the gap between the thermal power profile of
creases to 15, and the number of WTs equals its maximum value of 50. In BPG production and the consumed thermal power. However, compared
this situation, to cover the gap between the profile of active power to Case II, the number of WT and converters has reduced by 13 and 12,
produced by RESs and the total active power consumed by AC and EVPL and the number of motors and generators in CAES has reduced to 20 and
load, it is necessary to install 22 motors, 31 generators, and 140 CATs 28. The number of CATs has decreased to 116. It should be noted that in
within the CAES. In Case II, due to the increase in the number of BPGs the proposed plan, only BPG produces heat. Therefore, to meet the heat
(CAES capacity) compared to Case I, the number of converters in the demand, their number was increased than that of Case II. Nonetheless, as
path of BPGs (between DC and AC busses) also increases to 14 (31). the electrical energy consumption is the same in cases II and IV, as the
Nonetheless, since the number of WTs is the same in cases II and I, the number of BPGs increases, the number of WTs should decrease, which
number of converters for it in Case II is identical to the results of Case I. can be seen in Table 4. Similar outcomes to those in Case IV are seen in

Table 5
Annually planning cost of sources, converters, and storage in the hybrid system.
Case ACNPC ($/year) of

WTs BPGs TESs Motors CATs Generators Converters Total

I 8000 975 0 665 2650 600 3626.44 16,516.44

II 8000 4875 0 770 3500 1162.5 4853 23,160.5
III 8000 650 0 595 2525 637.5 3466.64 15,874.14
IV 5920 8775 3600 700 2900 1050 4640 27,585
V 5920 8775 3600 700 1275 262.5 4213 24,745.5

Case O&M-ANPC ($/year) of

WTs BPGs TESs Motors CATs Generators Converters Total

I 250 48.75 0 19 42.4 19.2 27.2 406.55

II 250 243.75 0 22 56 37.2 36.4 636.35
III 250 32.5 0 17 40.4 20.4 26 386.3
IV 185 438.75 54 20 46.4 33.6 34.8 812.55
V 185 438.75 54 20 20.4 8.4 31.6 758.15

11
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Table 6
Annually cost ($/year) of hybrid system.
Uncertain modeling Deterministic Stochastic (based on PEM) Cost of uncertain model

Case ACNPC O&M-ANPC ATNPC ACNPC O&M-ANPC ATNPC

I 16,012.2 396.3 16,408.5 16,516.44 406.55 16,923 514.5

II 22,125 615.4 22,740.4 23,160.5 636.35 23,796.85 1056.45
III 15,379.1 376.1 15,755.2 15,874.14 386.3 16,260.44 505.24
IV 26,356 789.5 27,145.5 27,585 812.55 28,397.55 1252.05
V 24,123.4 744.3 24,867.7 24,745.5 758.15 25,503.65 635.95

Case V, where a smart charging strategy is adopted for EVPL, where, the the deterministic and stochastic models of uncertainties. In the sto­
number of BPGs and converters increase to 27 and 25, respectively. Yet, chastic model, the demand (renewable output) can exceed or fall short
in this situation, by adopting the smart charging strategy for EVPL, the of the corresponding amount in the deterministic model. Consequently,
number of WTs is 37, so the number of converters is also 34. Also, in this in such cases, the stochastic model necessitates a greater number of
condition, CAES includes 20 motors, 7 generators, and 51 CATs. storage devices in comparison to the deterministic model. The afore­
Table 5 reports the ACNPC and O&M-ANPC for sources, storage mentioned section highlights the efficacy of addressing the problem
devices, and power electronic converters according to the numerical through the incorporation of the fourth novelty outlined in Section 1.4.
findings provided in Table 4. The annual construction cost for an
element based on the data in Table 2 is equal to dividing C by Lifespan. 4.4. Investigating the performance of the elements of the hybrid system
As presented in Table 5, the lowest ACNPC and O&M-ANPC exist for
WTs, CATs, and generators in Case V, but they exist for BPGs, TESs, Figs. 4–5 illustrate the expected daily curve of active and thermal
motors, and converters in Case III. In general, the lowest amount of these power of sources and storage devices for cases I to V. Fig. 4 depicts the
costs is for CAES in Case V. These results correspond to the number of results of cases I to V, in which the thermal energy model is not
elements mentioned in Table 4. Finally, the lowest value of ACNPC and considered. According to Fig. 4, WTs produce high active power during
O&M-ANPC is obtained for the hybrid system of Fig. 1 in Case III. This is 1:00–11:00 and 21:00–00:00, according to the wind speed data in [12],
because of the management of EVs charging in Case III and represents but their production power in other hours is low. Also, based on the data
the second novelty as given in Section 1.4. In Case III, there is no thermal of bio-waste gas in [12], BPGs produce high active power in the hours
energy consumption model. This is the case in cases IV and V. By 7:00 to 20:00, while their active power level is low in other hours. Next,
comparing the values of ACNPC and O&M-ANPC in Table 5, the case considering the performance and number of RESs given in Table 4,
with the lowest cost value among cases IV and V is Case V. Moreover, by motors are active in the hours 1:00–11:00 (1:00–13:00) and
comparing cases III and V, the proposed scheme (V) has increased 21:00–00:00 in cases I (II) and receive electricity from the sources and
ACNPC and O&M-ANPC by 56 % ((24,745.5–15,874.14)/15,784.14) then store compressed air in the CATs. This is because the sources' active
and 96 % compared to Case III and supplies the thermal load with TESs power is higher than the AC and EVPL load within these intervals. In
and BPGs. other words, CAES is in charging mode during these hours. In other
Table 6 shows the values of ACNPC, O&M-ANPC, and ATNPC for the hours, generators are active and they take compressed air from CATs and
hybrid system in different study cases for various uncertainty models. give it as active power to AC load and EVPL. In other words, in this
Also is observed that the EVPL's smart charging strategy decreases the situation, CAES is in discharge mode. In Case III, EVs are in the charging
planning cost in cases III than that of Case I. It provides a 3.9 % reduction position from 1:00 to 8:00 and receive all their energy consumption
in planning cost in the stochastic uncertainty model based on PEM in from the hybrid system because, in this interval, the active power pro­
Case III compared to Case I. Thus, by adopting the EVPL smart charging duction of RESs exceeds the AC load and EVs can be in charging mode.
strategy in Case III, compared to Case I (it is not an EVPL case and there Considering the performance of RESs and EVPL in Case III as shown in
is no thermal load model), the system operator can fully charge the EVs; Fig. 4(c), motors are active only in the hours 6:00–11:00, and generators
nonetheless, its planning cost is lower than that of Case I. This refers to inject active power into the hybrid system during 12:00–21:00. The
the second contribution of the paper presented in Section 1.4. However, operation of generation units and storage devices in the electrical sector
if the non-smart charging strategy is adopted for EVPL (Case II), the for case IV/V based on Fig. 4(d)/(e) is close to their operation in Case II/
planning cost will increase compared to Case I so the cost increase in III in Fig. 4(b)/(c), but their active power values are different.
Case II compared to Case I is about 40.6 %. Adding the thermal load to The daily thermal curve of sources and storage devices in cases IV
the islanded system in cases IV and V, as reported in Table 6, escalates and V are drawn in Fig. 5. The performance results of sources and
the planning cost because BPG is the only provider of thermal energy. storage in the thermal sector are the same for cases IV and V because,
Therefore, compared to cases I to III, the number of BPGs is increased according to Table 4, the number of BPGs and TESs was the same in
based on Table 4 to feed the thermal load. Considering that the number these two cases. Based on Fig. 5, the process of changes in the thermal
of WTs decreases in cases IV and V, but because of the high construction power of BPGs is the same as its active power, which is also evident in
cost of BPGs, the cost of planning in cases IV and V is higher than in cases Eq. (7). Also, the thermal power of BPGs exceeds the heat demand in the
I to III. However, the proposed plan (V) can obtain the lowest planning hours 8:00–18:00, and the opposite is true in other hours. So, TESs will
cost for the islanded system with thermal and electric load and EVs. In operate in charge mode during 8:00–18:00 (their power is positive) and
this case, ATNPC in PEM has an increase of about 57 % compared to Case they operate in discharge mode (their power is negative) during the rest
III. In other words, feeding the thermal load by the hybrid system results of the intervals.
in a 57 % increase in the planning cost. In addition, it can be seen that
there is a difference in the values of uncertainties in terms of ATNPC
4.5. Evaluation of CAES capability
between the deterministic and stochastic models. This difference is
introduced as the uncertainty modeling cost. This cost is equal to
To investigate the capability of CAES in islanded system planning,
$635.95 for the proposed case (V) in the stochastic model based on PEM,
the planning status of the said system for the use of CAES, battery, and
which is due to the consideration of uncertain parameters of load,
hydrogen storage has been examined. In [12], instead of using CAES, a
renewable power, and EVPL parameters. The disparity in costs can be
DC-bus-connected battery is adopted. The mathematical model and
attributed to the varying quantities of hybrid system elements present in
characteristics of the battery are described in [12]. In [13], hydrogen

12
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

storage is also used as a static storage. This storage receives electrical

energy from the DC bus by the electrolyzer and stores it as hydrogen
energy in the hydrogen tank. Also, the fuel cell receives hydrogen energy
from the mentioned tank and delivers it to the DC bus in the form of
electrical energy. The mathematical model and its specifications are
presented in [13]. Comparing the data of these storage devices in
[12,13,24], it can be stated that the cost of building an engine, gener­
ator, and CAT in CAES is lower than that of electrolysis, hydrogen tank,
and fuel cell. Also, the efficiency of CAES elements is higher than
hydrogen storage elements. The useful life of CAES elements is about 20
years. However, the useful life of the fuel cell is about 15 years, but other
(a)
hydrogen storage elements have a useful life of about 20 years. Batteries
have higher efficiency than other storage devices, but their useful life is
around 5 years. Also, their cost construction is higher than that of other
storage facilities. The planning results of the islanded system using
CAES, battery, and hydrogen storage according to the uncertainty model
based on PEM are reported in Table 7. As is provided in [12], the
maintenance cost of batteries is insignificant, therefore according to
Table 7, the amount of O&M-ANPC based on the islanded system with
batteries is the lowest. In the context of ACNPC, CAES can obtain a lower
construction cost than other storages for the islanded system due to
having a suitable efficiency (around 80 %), lower construction cost, and
higher useful life. Comparing the battery and the hydrogen storage, the
(b)
battery creates a lower construction cost due to its high efficiency
(around 90 %). Yet, because the fuel cell efficiency in hydrogen storage
is low (around 50 %) [13], the cost of constructing an islanded system in
hydrogen storage is the highest. Finally, referring to Table 7, the ATNPC
for the system with CAES is lower by about 7.7 % (12.9 %) compared to
the battery (hydrogen storage).

5. Conclusions

In this paper, optimal scheduling of a full renewable hybrid system

combined with a wind turbine, bio-waste energy unit, and stationary
(c) storage such as compressed air energy storage (with a motor, generator
and compressed air tank) and heat storage was provided to concurrently
supply electricity and heat and EVPL consumption energy. The bio-
waste unit can simultaneously generate active and thermal power. The
proposed scheme included minimization of the yearly cost of construc­
tion and maintenance of sources, storage devices and power electronic
converters. It was dependent on the operation model of these elements.
The model is designed such that renewable resources are given priority
in meeting the demand. Subsequently, static storage elements are
employed to bridge the discrepancy between the power profile of these
sources and the energy demand. Stochastic programming based on point
estimate method was adopted to introduce a modeling for uncertain
(d) quantities related to the load, renewable power, and EVPL parameters.
Then, the ALO find a solution to the problem. According to the results,
the algorithm in question achieves the optimal solution with a faster
convergence rate than other evolutionary algorithms. Additionally, the
final response exhibits a relatively low standard deviation of approxi­
mately 0.93 %, which shows the low dispersion of the final response
compared to its counterpart solvers. Then, the proposed scheme obtains
an optimal number of sources, storage devices and power electronic
converters. Thus, considering the smart charging strategy of electric
vehicles reduces in the number of bio-waste units, converters and as
compressed air energy storage compared to the case without electric ve­
hicles. Also, considering the thermal load model increases the number of
(e) bio-waste units and thermal storage, but in return, the number of wind
turbines, compressed air storage and converters decreases. Nevertheless,
Fig. 4. Expected daily active power curve of sources and storages in, a) Case I, the use of a non-mart charging strategy leads to the presence of more
b) Case II, c) Case III, d) Case IV, e) Case V. elements. Accordingly, the presence of EVPL in the islanded system
based on non-smart (smart) charging management results in an increase
of 40.6 % (decrease of 3.9 %) in planning cost. The addition of thermal
consumers to the islanded system leads to a 57 % increase in the plan­
ning cost. Finally, considering the uncertainties model has led to an

13
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

Fig. 5. Expected daily curve of thermal power for the different sources and storages in Cases IV and V.

Table 7 Conceptualization, Methodology, Software, Validation, Formal analysis,

Compare CAES capability on IHS planning with respect to battery and hydrogen Investigation, Resources, Data curation, Writing – original draft.
storage in Case V.
Storage CAES Battery [12] Hydrogen storage [13] Declaration of competing interest
ACNPC ($/year) 24,745.5 27,021 28,211
O&M-ANPC ($/year) 758.15 621.4 1075.6 The authors declare that they have no known competing financial
ATNPC ($/year) 25,503.65 27,642.4 29,286.6 interests or personal relationships that could have appeared to influence
the work reported in this paper.

increase in the planning cost than that of the deterministic model. Data availability
Therefore, as per the obtained numerical results, the merits of this
scheme include: The data that has been used is confidential.

- Simultaneous supply of electrical and thermal energy in the hybrid Acknowledgment

islanded system by using renewables and storage facilities to provide
suitable environmental conditions, This study is supported via funding from Prince Sattam bin Abdulaziz
- Supply of demanded energy of EVs in the hybrid islanded system University project number (PSAU/2023/R/1444).
based on a smart charging strategy as they will be present as a new
consumer of energy in the power system within upcoming years, References
- Reducing the planning cost using the compressed-air storage, and
- Simultaneous extraction of electrical and thermal energy from the [1] A. Askarzadeh, Distribution generation by photovoltaic and diesel generator
bio-waste unit. systems: energy management and size optimization by a new approach for a stand-
alone application, Energy 122 (2017) 542–551.
[2] A. Maleki, A. Askarzadeh, Optimal sizing of a PV/wind/diesel system with battery
One of solutions to minimize cost is the use of demand side man­ storage for electrification to an off-grid remote region: a case study of Rafsanjan,
agement or load response. This was not considered in the proposed Iran, Sustain. Energy Technol. Assess. 7 (2014) 147–153.
[3] M.A. Voelklein, D. Rusmanis, J.D. Murphy, Biological methanation: strategies for
scheme but the proposed scheme model is considered based on load in-situ and ex-situ upgrading in anaerobic digestion, Appl. Energy 235 (2019)
response in the electrical and thermal sectors as the future work. In the 1061–1072.
suggested scheme, the reliability scheme was not analyzed. However, in [4] M.R. Akbari-Zadeh, T. Niknam, A. Kavousi-Fard, Adaptive robust optimization for
the energy management of the grid-connected energy hubs based on hybrid meta-
some case, some of the sources and/or power converters may not be heuristic algorithm, Energy 235 (2021), 121171.
accessible due to the failure. In these conditions, outage occurs for a [5] R. Dufo-López, et al., Multi-objective optimization minimizing cost and life cycle
percentage of consumers. To deal with this issue, the suggested scheme emissions of stand-alone PV–wind–diesel systems withbatteries storage, Appl.
Energy 88 (2011) 4033–4041.
needs to be constrained to the reliability, and this is considered as the
[6] S. Ma, et al., Fuel cell-battery hybrid systems for mobility and off-grid applications:
future work. a review, Renew. Sust. Energ. Rev. 135 (2021), 110119.
[7] V. Suresh, M.M. Muralidhar, R. Kiranmayi, Modelling and optimization of an off-
CRediT authorship contribution statement grid hybrid renewable energy system for electrification in a rural areas, Energy
Rep. 6 (2020) 594–604.
[8] M.A. Vaziri-Rad, A. Kasaeian, X. Niu, K. Zhang, O. Mahian, Excess electricity
Farshad Khalafian: Conceptualization, Methodology, Software, problem in off-grid hybrid renewable energy systems: a comprehensive review
Validation, Formal analysis, Investigation, Resources, Data curation, from challenges to prevalent solutions, Renew. Energy 212 (2023) 538–560.
[9] C. Ceylan, Y. Devrim, Green hydrogen based off-grid and on-grid hybrid energy
Writing – original draft. Nahal Iliaee: Conceptualization, Methodology, systems, Int. J. Hydrog. Energy 48 (2023) 39084–39096.
Software, Validation, Formal analysis, Investigation, Resources, Data [10] D. Roy, Modelling an off-grid hybrid renewable energy system to deliver electricity
curation, Writing – original draft. Ekaterina Diakina: Writing – original to a remote Indian island, Energy Convers. Manag. 281 (2023), 116839.
[11] S. Arabi-Nowdeh, A. Naderipour, I. Faraji-Davoudkhani, J.M. Guerrero, Stochastic
draft. Peyman Parsa: Writing – original draft. Mohammed M. optimization – based economic design for a hybrid sustainable system of wind
Alhaider: Writing – original draft. Milad Hadizadeh Masali: Writing – turbine, combined heat, and power generation, and electric and thermal storages
original draft. Sasan Pirouzi: Supervision, Conceptualization, Meth­ considering uncertainty: a case study of Espoo, Finland, Renew. Sust. Energ. Rev.
183 (2023), 113440.
odology, Software, Validation, Formal analysis, Investigation, Re­
[12] M.R. Jokar, et al., Stationary and mobile storages-based renewable off-grid system
sources, Data curation, Writing – original draft. Min Zhu: Supervision, planning considering storage degradation cost based on information-gap decision
theory optimization, J. Energy Storage 58 (2023), 106389.

14
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

[13] H. Sun, et al., Designing framework of hybrid photovoltaic-biowaste energy system [69] J. Jiang, L. Zhang, X. Wen, E. Valipour, S. Nojavan, Risk-based performance of
with hydrogen storage considering economic and technical indices using whale power-to-gas storage technology integrated with energy hub system regarding
optimization algorithm, Energy 238 (2022), 121555. downside risk constrained approach, Int. J. Hydrog. Energy 47 (93) (2022)
[14] R. Kaur, V. Krishnasamy, N.K. Kandasamy, S. Kumar, Discrete multiobjective grey 39429–39442, https://doi.org/10.1016/j.ijhydene.2022.09.115.
wolf algorithm based optimal sizing and sensitivity analysis of PV-wind-battery [70] S. Pirouzi, J. Aghaei, Mathematical modeling of electric vehicles contributions in
system for rural telecom towers, IEEE Syst. J. 14 (1) (March 2020) 729–737. voltage security of smart distribution networks, Simulation 95 (5) (2019) 429–439.
[15] P. Marocco, D. Ferrero, A. Lanzini, M. Santarelli, The role of hydrogen in the [71] L. Lin, et al., Non-intrusive residential electricity load decomposition via low-
optimal design of off-grid hybrid renewable energy systems, J. Energy Storage 46 resource model transferring, J. Build. Eng. 73 (2023), 106799, https://doi.org/
(2022), 103893. 10.1016/j.jobe.2023.106799.
[16] S.E. Ahmadi, S.M. Kazemi-Razi, M. Marzband, A. Ikpehai, A. Abusorrah, Multi- [72] S. Zhou, et al., Stability effect of different modulation parameters in voltage-mode
objective stochastic techno-economic-environmental optimization of distribution PWM control for CCM switching DC-DC converter, IEEE Trans. Transp.
networks with G2V and V2G systems, Electr. Power Syst. Res. 218 (2023), 109195. Electrification (2023), https://doi.org/10.1109/TTE.2023.3293811.
[17] R. Khalili, et al., Robust multi-objective optimization for the Iranian electricity [73] X. Zheng, et al., Combustion characteristics and thermal decomposition mechanism
market considering green hydrogen and analyzing the performance of different of the flame-retardant cable in urban utility tunnel, Case Stud. Therm. Eng. 44
demand response programs, Appl. Energy 334 (2023), 120737. (2023), 102887, https://doi.org/10.1016/j.csite.2023.102887.
[18] A. Salari, S.E. Ahmadi, M. Marzband, M. Zeinali, Fuzzy Q-learning-based approach [74] Z. Wu, et al., Effect of dielectric relaxation of epoxy resin on dielectric loss of
for real-time energy management of home microgrids using cooperative multi- medium-frequency transformer, IEEE Trans. Dielectr. Electr. Insul. 29 (5) (2022)
agent system, Sustain. Cities Soc. 95 (2023), 104528. 1651–1658, https://doi.org/10.1109/TDEI.2022.3193652.
[19] A. Alhendi, A.S. Al-Sumaiti, M. Marzband, R. Kumar, A.A. Zaki Diab, Short-term [75] L. Zhu, Z. Li, K. Hou, Effect of radical scavenger on electrical tree in cross-linked
load and price forecasting using artificial neural network with enhanced Markov polyethylene with large harmonic superimposed DC voltage, High Volt. (2022),
chain for ISO New England, Energy Rep. 9 (2023) 4799–4815. https://doi.org/10.1049/hve2.12302.
[20] F.V. Cerna, J.K. Coêlho, M.P. Fantesia, E. Naderi, M. Marzband, J. Contreras, Load [76] S. Zhou, G. Zhou, X. Liu, H. Zhao, Dynamic freewheeling control for SIDO Buck
factor improvement of the electricity grid considering distributed energy resources converter with fast transient performance, minimized cross-regulation, and high
operation and regulation of peak load, Sustain. Cities Soc. 98 (2023), 104802. efficiency, IEEE Trans. Ind. Electron. 70 (2) (2023) 1467–1477, https://doi.org/
[21] S. Eslami, Y. Noorollahi, M. Marzband, A. Anvari-Moghaddam, Integrating heat 10.1109/TIE.2022.3156169.
pumps into district heating systems: a multi-criteria decision analysis framework [77] Y. Duan, Y. Zhao, J. Hu, An initialization-free distributed algorithm for dynamic
incorporating heat density and renewable energy mapping, Sustain. Cities Soc. 98 economic dispatch problems in microgrid: modeling, optimization and analysis,
(2023), 104785. Sustain. Energy Grids Netw. 34 (2023), 101004, https://doi.org/10.1016/j.
[22] S.S. Ghazimirsaeid, M.S. Jonban, M.W. Mudiyanselage, M. Marzband, J.S. segan.2023.101004.
R. Martinez, A. Abusorrah, Multi-agent-based energy management of multiple grid- [78] P. Li, J. Hu, L. Qiu, Y. Zhao, B.K. Ghosh, A distributed economic dispatch strategy
connected green buildings, J. Build. Eng. 74 (2023), 106866. for power–water networks, IEEE Trans. Control Netw. Syst. 9 (1) (2022) 356–366,
[23] A. Rezaee Jordehi, et al., Industrial energy hubs with electric, thermal and https://doi.org/10.1109/TCNS.2021.3104103.
hydrogen demands for resilience enhancement of mobile storage-integrated power [79] H. Tian, R. Li, B. Salah, P. Thinh, Bi-objective optimization and environmental
systems, Int. J. Hydrog. Energy (2023), https://doi.org/10.1016/j. assessment of SOFC-based cogeneration system: performance evaluation with
ijhydene.2023.07.205. various organic fluids, Process. Saf. Environ. Prot. 178 (2023) 311–330, https://
[24] A.R. Daneshvar-Garmroodi, F. Nasiri, F. Haghighat, Optimal dispatch of an energy doi.org/10.1016/j.psep.2023.07.040.
hub with compressed air energy storage: a safe reinforcement learning approach, [80] Z. Zhang, F.M.A. Altalbawy, M. Al-Bahrani, Y. Riadi, Regret-based multi-objective
J. Energy Storage 57 (2023), 106147. optimization of carbon capture facility in CHP-based microgrid with carbon
[25] M.A. Norouzi, J. Aghaei, S. Pirouzi, T. Niknam, M. Fotuhi-Firuzabad, Flexibility dioxide cycling, J. Clean. Prod. 384 (2023), 135632, https://doi.org/10.1016/j.
pricing of integrated unit of electric spring and EVs parking in microgrids, Energy jclepro.2022.135632.
239 (2022), 122080. [101] H. Yuan, B. Yang, System dynamics approach for evaluating the interconnection
[26] S. Pirouzi, Network-constrained unit commitment-based virtual power plant model performance of cross-border transport infrastructure, J. Manag. Eng. 38 (3)
in the day-ahead market according to energy management strategy, IET Gener. (2022), https://doi.org/10.1061/(ASCE)ME.1943-5479.0001015.
Transm. Distrib. 17 (2023) 4958–4974. [102] Y. Wang, Q. Wang, Y. Li, H. Wang, Y. Gao, Y. Sun, M. Zhan, Impact of incineration
[28] X.W. Zhang, X. Yu, X. Ye, S. Pirouzi, Economic energy managementof networked slag co-disposed with municipal solid waste on methane production and
flexi-renewable energy hubs according to uncertainty modeling by the unscented methanogens ecology in landfills, Bioresour. Technol. 377 (2023), 128978,
transformation method, Energy 278 (2023), 128054. https://doi.org/10.1016/j.biortech.2023.128978.
[41] Y. Chen, Research on collaborative innovation of key common technologies in new [103] Z. Luo, et al., Extraordinary Role of Zn in Enhancing Thermoelectric Performance
energy vehicle industry based on digital twin technology, Energy Rep. 8 (2022) of Ga-doped n-type PbTe, 2021, p. 15, https://doi.org/10.1039/d1ee02986j.
15399–15407, https://doi.org/10.1016/j.egyr.2022.11.120. [104] C. Yuan, et al., Effect of charge transport on electrical degradation in
[42] C. Min, et al., Trajectory optimization of an electric vehicle with minimum energy polypropylene/organic molecular semiconductor composites for HVDC cable
consumption using inverse dynamics model and servo constraints, Mech. Mach. insulation, Appl. Phys. Lett. 122 (11) (2023), https://doi.org/10.1063/
Theory 181 (2023), 105185, https://doi.org/10.1016/j. 5.0133417.
mechmachtheory.2022.105185. [122] C. Lu, et al., Split-core magnetoelectric current sensor and wireless current
[43] X. Zhang, Z. Wang, Z. Lu, Multi-objective load dispatch for microgrid with electric measurement application, Measurement: J. Int. Meas. Confederation 188 (2022),
vehicles using modified gravitational search and particle swarm optimization 110527, https://doi.org/10.1016/j.measurement.2021.110527.
algorithm, Appl. Energy 306 (2022), 118018, https://doi.org/10.1016/j. [123] G. Liu, Data collection in MI-assisted wireless powered underground sensor
apenergy.2021.118018. networks: directions, recent advances, and challenges, IEEE Commun. Mag. 59 (4)
[44] X. Zhang, Z. Lu, X. Yuan, Y. Wang, X. Shen, L2-gain adaptive robust control for (April 2021) 132–138, https://doi.org/10.1109/MCOM.001.2000921.
hybrid energy storage system in electric vehicles, IEEE Trans. Power Electron. 36 [124] W. Zou, et al., Limited sensing and deep data mining: a new exploration of
(6) (2021) 7319–7332, https://doi.org/10.1109/TPEL.2020.3041653. developing city-wide parking guidance systems, IEEE Intell. Transp. Syst. Mag. 14
[45] Z. Liu, Y. Wu, J. Feng, Competition between battery switching and charging in (1) (Jan.-Feb. 2022) 198–215.
electric vehicle: considering anticipated regret, Environ. Dev. Sustain. (2023), [125] Z. Yan, H. Wen, Electricity theft detection base on extreme gradient boosting in
https://doi.org/10.1007/s10668-023-03592-4. AMI, IEEE Trans. Instrum. Meas. 70 (2021), https://doi.org/10.1109/
[46] L. Zhang, et al., Research on the orderly charging and discharging mechanism of TIM.2020.3048784.
electric vehicles considering travel characteristics and carbon quota, IEEE Trans. [126] X. Ma, Z. Liao, Y. Wang, J. Zhao, Fast dynamic phasor estimation algorithm
Transp. Electrification (2023), https://doi.org/10.1109/TTE.2023.3296964. considering DC offset for PMU applications, IEEE Trans. Power Deliv. (2023),
[47] L. Zhang, C. Sun, G. Cai, L.H. Koh, Charging and discharging optimization strategy https://doi.org/10.1109/TPWRD.2023.3285949.
for electric vehicles considering elasticity demand response, eTransportation 18 [136] R. Homayoun, B. Bahmani-Firouzi, T. Niknam, Multi-objective operation of
(2023), 100262, https://doi.org/10.1016/j.etran.2023.100262. distributed generations and thermal blocks in microgrids based on energy
[50] Y. Xiao, Y. Zhang, I. Kaku, R. Kang, X. Pan, Electric vehicle routing problem: a management system, IET Gener. Transm. Distrib. 15 (9) (2021) 1451–1462.
systematic review and a new comprehensive model with nonlinear energy [137] S. Mirjalili, The ant lion optimizer, Adv. Eng. Softw. 83 (2015) 80–98.
recharging and consumption, Renew. Sust. Energ. Rev. 151 (2021), 111567, [138] W.K.A. Najy, H.H. Zeineldin, W.L. Woon, Optimal protection coordination for
https://doi.org/10.1016/j.rser.2021.111567. microgrids with grid-connected and islanded capability, IEEE Trans. Ind.
[51] H. Chen, H. Wu, T. Kan, J. Zhang, H. Li, Low-carbon economic dispatch of Electron. 60 (4) (April 2013) 1668–1677.
integrated energy system containing electric hydrogen production based on VMD- [139] Nord Pool. https://www.nordpoolgroup.com/historical-market-data/.
GRU short-term wind power prediction, Int. J. Electr. Power Energy Syst. 154 [140] E. Akbari, S.F. Mousavi-Shabestari, S. Pirouzi, M. Jadidoleslam, Network
(2023), 109420, https://doi.org/10.1016/j.ijepes.2023.109420. flexibility regulation by renewable energy hubs using flexibility pricing-based
[52] K. Liao, D. Lu, M. Wang, J. Yang, A low-pass virtual filter for output power energy management, Renew. Energy 206 (2023) 295–308.
smoothing of wind energy conversion systems, IEEE Trans. Ind. Electron. 69 (12) [141] Z. Qu, C. Xu, F. Yang, F. Ling, S. Pirouzi, Market clearing price-based energy
(2022) 12874–12885, https://doi.org/10.1109/TIE.2021.3139177. management of grid-connected renewable energy hubs including flexible sources
[67] M.C. Kisacikoglu, Vehicle-to-grid Reactive Power Operation Analysis of the EV/ according to thermal, hydrogen, and compressed air storage systems, J. Energy
PHEV Bidirectional Battery Charger, Tennessee University, 2013. Storage 69 (2023), 107981.
[68] S. Lu, et al., An asymmetric encoder–decoder model for Zn-ion battery lifetime [142] S. Pirouzi, J. Aghaei, M. Shafie-Khah, G.J. Osório, J.P.S. Catalão, Evaluating the
prediction, Energy Rep. 8 (2022) 33–50, https://doi.org/10.1016/j. security of electrical energy distribution networks in the presence of electric
egyr.2022.09.211. vehicles, in: Conf. of IEEE Manchester PowerTech, 2017, pp. 1–6.

15
F. Khalafian et al. Journal of Energy Storage 78 (2024) 109888

[143] S. Pirouzi, M.A. Latify, G.R. Yousefi, Investigation on reactive power support agent system-based restoration scheme in active distribution network, IET Renew.
capability of PEVs in distribution network operation, in: 23rd Iranian Conference Power Gener. 15 (8) (2021) 1765–1777.
on Electrical Engineering, 2015, pp. 1591–1596. [148] M.A. Norouzi, J. Aghaei, S. Pirouzi, Enhancing distribution network indices using
[144] A. Shahbazi, et al., Holistic approach to resilient electrical energy distribution electric spring under renewable generation permission, in: International
network planning, Int. J. Electr. Power Energy Syst. 132 (2021), 107212. Conference on Smart Energy Systems and Technologies (SEST), 2019, pp. 1–6.
[145] H.R. Hamidpour, et al., Coordinated expansion planning problem considering [149] M.A. Norouzi, J. Aghaei, T. Niknam, S. Pirouzi, M. Lehtonen, Bi-level fuzzy
wind farms, energy storage systems and demand response, Energy 239 (2022), stochastic-robust model for flexibility valorizing of renewable networked
122321. microgrids, Sustain. Energy Grids Netw. 31 (2022), 100684.
[146] G. Piltan, S. Pirouzi, A.R. Azarhooshang, A. Rezaee-Jordehi, A. Paeizi, [150] H.R. Zafarani, S.A. Taher, M. Shahidehpour, Robust operation of a multicarrier
M. Ghadamyari, Storage-integrated virtual power plants for resiliency energy system considering EVs and CHP units, Energy 192 (2020), 116703.
enhancement of smart distribution systems, J. Energy Storage 55 (2022), 105563. [151] A.R. Azarhooshang, D. Sedighizadeh, M. Sedighizadeh, Two-stage stochastic
[147] L. Bagherzadeh, H. Shayeghi, S. Pirouzi, M.R. Shafie-khah, J.P.S. Catalão, operation considering day-ahead and real-time scheduling of microgrids with
Coordinated flexible energy and self-healing management according to the multi- high renewable energy sources and electric vehicles based on multi-layer energy
management system, Electr. Power Syst. Res. 201 (2021), 107527.

16