Energy Transition and Resilient Control for Enhancing Power Availability in Microgrids Based on North African Countries: A Review

Featured Application

This review subscribes to the MiDiNA project as a part of the European Union-African Union partnership for renewable energy, ‘LEAP-RE’. This review will shed light on the energy transition in North African countries and the resilient controls for improving power availability in microgrids. It seeks to draw attention to the advantages and drawbacks of each type of control as well as discuss the specifications for resilient controls that are necessary for microgrids in the North African environment.

Abstract

The ambition of making North Africa a hub for renewable energies and green hydrogen has prompted local governments and the private sector to work together towards boosting the growth of locally available, sustainable energy resources. Numerous climate and energy challenges can be addressed by microgrid technologies, which enable cost-effective incorporation of renewable energy resources and energy storage systems through smart management and control infrastructures. This paper discusses the ongoing energy transition in the countries of North Africa, highlighting the potential for renewable energy sources as well as regional obstacles and challenges. Additionally, it explores how robust and stable controls and advanced management strategies can improve microgrids’ performances. Special attention is given to assessing the advantages and disadvantages of conventional and advanced controllers, with an emphasis on resilience needed within the harsh North African environment.

1. Introduction

The quality of electricity may be described as a set of parameters such as supply reliability, service quality, voltage, and current quality [1]. In 2020, in North Africa, the highest quality of electricity was reported for Morocco (80% followed by Tunisia (73%), Egypt (71%), and Algeria (60%) [2]. The percentage of people without access to electricity in North Africa is less than 2%, while in West and East Africa, the figures are 50% and 75%, respectively [3]. Despite North Africa’s heavy reliance on fossil fuels, the region is making significant efforts to switch to renewable energy sources. The degree of reliance differs between countries that export natural gas and oil, such as Egypt, Algeria, and Libya, and those that import them, such as Morocco and Tunisia. North Africa is a special, endowed region all its own. Blessed with natural resources (oil, coal, and natural gas) and unique geographic features (large deserts and abundance of sunshine), it is perfect for integrating renewable energies and producing green hydrogen on a large scale, as shown in Figure 1. In the Sahara Desert, solar insolation levels range from 2500 to 3000 kWh per square meter annually [3]. North African countries also stand out for their pre-existing pipeline networks and offshore wind potential. In May 2022, six African countries—including three from North Africa: Morocco, Egypt, and Mauritania— signed the “Africa Green Hydrogen Alliance” with the aim of pooling their efforts and adding value to Africa through sustainable energy and extended an invitation to other African countries [4].

North Africa is an important exporter of fertilizers, and it is utilizing green hydrogen to produce zero-carbon ammonia by the Haber-Bosch process by combining nitrogen and hydrogen [12]. Thereupon, the idea of a microgrid is particularly appealing in North African nations, and governments support it through public-private partnerships. In addition, through the use of energy storage systems, reliable controllers, electrical vehicles, and renewable energy sources, microgrids in North Africa can reduce their carbon footprint while supplying local loads. In addition to offering financial advantages, microgrids can guarantee a steady supply of high-quality power, support the grid when it is connected, and boost resilience [13], as shown in Figure 2.

Depending on the operation mode and the objective control, different controllers can be used at different levels. Microgrids can be controlled using centralized, decentralized, distributed, or hierarchical control. Hierarchical control is categorized into three levels: primary, secondary, and tertiary. In the primary control, the inverter can operate as a grid feeder, grid form, or grid support. Different types of controllers can be used, conventional ones as proportional (P), proportional-integral-derivative (PID), and proportional resonant (PR), and advanced like neural network (NN), practical swarm optimization (PSO), genetic algorithm (GA), fuzzy logic (FLC), H-infinity, adaptive control, sliding mode control (SMC), backstepping control (BC), model predictive control (MPC), and disturbance estimation (Kalman filter). Advanced controllers have the benefit of being able to manage complicated dynamics and uncertainties. Nevertheless, they have the following noteworthy shortcomings: for example, in MPC, the computational cost is higher. SMC has chattering problems, and H-infinity control has slow system dynamics and responses for more details see, [14].

On the other side, microgrid stability can be impacted by operational and infrastructural damages such as hackers, false data injection attacks, denial-of-service (DoS) attacks, and uncertainties. Therefore, microgrids need to be prepared to cover the attacks and guarantee the restoration of stable operation. The types of cyberattacks on microgrids are presented in Figure 3. A false data injection attack, for example, which refers to the situation in which an attacker deceives sensor measurements to introduce unnoticed flaws into the state variable and value calculations [15], can have disastrous effects, particularly in emergency situations and situations involving critical loads. Therefore, one effective way to increase microgrid resilience against cyberattacks is through resilient control.

Resilience can be defined as the ability to sufficiently prepare for and withstand effectively uncommon, severe abuse events in order to successfully reduce their damaging effects and/or shorten the period of the disruption [16].

This review subscribes to the MiDiNA project as a part of the European Union-African Union partnership for renewable energy, ‘LEAP-RE’. This review will shed light on the energy transition in North African countries and the resilient controls for improving power availability in microgrids. It seeks to draw attention to the advantages and drawbacks of each type of control as well as discuss the specifications for resilient controls that are necessary for microgrids in the North African environment.

This work is organized as follows: Section 2 introduces the energy transition in North Africa. Section 3 presents the structures of microgrids. Section 4 focuses on the review of microgrid designs and controls. Section 5 discusses the control methods applied to microgrids developed in the North African region. A conclusion summarizes this study.

2. Energy transition in North Africa

2.1. Introduction to North Africa’s Energy Landscape

While efforts are being made to expedite the energy transition in North Africa, the continent’s main energy sources remain dependent on natural gas, coal, and oil. In 2019, 95% of the world’s primary energy supply was derived from fossil fuels, as mentioned in Figure 4. Oil dominates the economies of Algeria, Egypt, Libya, Morocco, and Tunisia, making up around 45% to 85% of all of North Africa’s final consumption [17]. The two largest producers of natural gas in Africa are Algeria and Egypt, which helps to explain why these two countries are so reliant on non-natural resources. Gas and oil make about 40% of all Egyptian product exports [18] and provide 95% of Algeria’s hard currency earnings [19]. Regarding proven crude oil reserves, Libya owns the highest amount in Africa.

Even though North African nations rely heavily on fossil fuels, their carbon dioxide (CO₂) emissions are still very low when measured against developed nations. In 2021, North African countries accounted for just 1.59% of global carbon dioxide (CO₂) emissions, while China, the US, India, Russia, and Japan accounted for 60.48%. As mentioned in Figure 5 and Figure 6.

2.2. Renewable Energy Potential

The deployment of large-scale solar, wind, and green hydrogen projects is accelerating the penetration of renewable energy sources in North African nations (Figure 7).

With four solar plants totaling 3000 hectares in size and a 580 MW capacity, considered the biggest concentrated power plant in the world, the enormous “Noor Ouarzazate” public-private solar energy project in southeast Morocco has the potential to supply roughly 6% of the nation’s electricity needs. Figure 8 [23]. Morocco’s wind capacity is also significant (Figure 9).

In 2023, Morocco demonstrated a strong performance in North Africa among leading nations such as Denmark and Sweden (ranked 4 and 5), coming in at number 7 on the Climate Change Performance Index (CCPI) (Figure 10).

2.3. Government Initiatives and Policies

Every country in North Africa is undergoing an energy transition with the goal of achieving national plans and commitments for installed renewable energy capacity, energy consumption, and greenhouse gas emissions. According to the Moroccan energy strategy 2030, 52% of power capacity will come from solar, onshore wind, biomass, and hydroelectric, and energy consumption will be reduced by 20% according to the National Energy Efficiency Strategy, with the Green House Gas (GHG) emissions aimed at being reduced by 25%. As for Tunisia, the country’s targets seek to reach 30% renewables in the mix of energy sources and reduce energy consumption by 30%, and the GHG is planned to be reduced by 41% by 2030 [27].

Regarding Egypt, the Integrated Sustainable Energy Strategy projects that by 2035, 42% of the electricity mix will come from renewable sources, as mentioned in Figure 11.

Algeria has committed to installing approximately 22 GW of renewable energy capacity by 2030. The Renewable Energy and Energy Efficiency Development Plan 2016–2030 established a conditional target of 37% of electricity generation coming from renewable sources, as mentioned in Figure 12.

By 2035, Libya wants to achieve its ambitious goal of obtaining up to 20% of its electricity from renewable sources [29].

Mauritania wants to reach 60% of its energy coming from renewable sources by 2030 [30].

2.4. Challenges and Barriers

Building infrastructure, in addition to water scarcity, is one of the hardest and most time-consuming tasks. Installing renewable energy resources requires a lot of land, and it also involves economic challenges to set up the infrastructure needed to build electrolysis plants, solar and wind farms, hydrogen pipelines, and green hydrogen production, storage, and transportation. Because of this, nations attempt to get around this issue by enticing foreign investment through the implementation of policies and facilities such as Morocco’s “Moroccan offer” iniative, for more details see [31]. Another challenge is the workforce and education development. It will be crucial to train and develop a skilled workforce capable of operating and maintaining green hydrogen infrastructure and renewable energy sources in order to attract private investment and maintain a stable business climate.

North Africa has abundant resources for mining, including heavy mineral sands, iron ore, copper, zinc, and gold. According to Africa Mining IQ [32], there are fifteen mining ventures in Egypt, eight in Morocco, and one in Algeria. The switch to renewable energy opens the way to decarbonizing some heavy and mining industries, which have traditionally been difficult to clean up. This includes long-distance transportation as well as the steel, iron, and chemical industries.

2.5. Role of International Collaborations

An important factor in speeding up and facilitating the region’s transition to renewable energy is collaboration with other stakeholders, international organizations, and other nations. This collaboration could involve money and technology transfers, knowledge exchanges, and cooperative projects. Through regional and international electrical connections, low-carbon energy produced in North Africa will benefit not only the countries of the region but also the European Union, North Africa, and Middle East countries (Figure 13).

Recently, during the Convention on Climate Change (COP28) organized in Dubai in 2023, Morocco and Portugal signed an agreement on electrical interconnection [34].

Hydrogen, the flag shift of the energy transition in recent years, is anticipated to be the bridge between Europe and North Africa. The natural gas infrastructure of the Morocco-Nigeria gas pipeline could be exploited to transport hydrogen from North Africa to Europe.

With regard to technology transfer, the German-Tunisian Energy Partnership and the German-Moroccan Energy Partnership (PAREMA), which were established in 2012 with the goal of promoting renewable energy projects and energy efficiency in North Africa, are two notable examples of the International Cooperation’s (GIZ) active involvement in the energy transition of North Africa [35]. Morocco and Russia signed a Memorandum of Understanding (MoU) on 27 July 2023. The goal of the agreement is to look into the construction of water desalination plants in Morocco that will use nuclear power technology from Rosatom to deliver water for use within industry, agriculture, and domestic use [36]. Within the same framework, up to half of Egypt’s electricity production capacity is anticipated to come from the 4800 GW El Dabaa Nuclear Power Plant (NPP), a nuclear power plant being developed by the Russian State Atomic Energy Corporation (ROSATOM) [37]. The Xlinks Morocco-UK Power Project will connect the Guelmim Oued Noun region to Britain via four 3800 km of high-voltage DC sub-sea cables. This project will reach 8% of Britain’s electricity needs; around 7 million houses will be powered by solar and wind farms installed on a surface area of 1500 km² in Morocco [38].

3. Structures of Microgrids

3.1. Introduction to Microgrids

Microgrids promote energy decentralization, independence, resilience, and efficiency. A microgrid is a local grid with the ability to function independently of the main grid and in island mode. It is a combination of local energy production sources and distribution and sufficient capabilities to support local loads. Microgrids can be categorized as DC, AC, or hybrid according to the kind of bus they use. Combining features from both DC and AC microgrids builds a hybrid microgrid. A shared DC bus is used by DC microgrids, whereas an AC bus is used by AC microgrids [39,40]. A DC microgrid can be configured in various topologies to ensure resilience and remove malfunctioning parts without compromising the integrity of the entire microgrid in the case of ring, ladder, and zonal structures (Figure 14).

3.2. Operation Modes of a Microgrid

A microgrid enters islanding mode when it cuts off from the grid. In this scenario, a voltage control mode is applied by the Distributed Generator (DG) sources in order to maintain the voltage of the adjacent loads constant [42]. A point of common coupling (PCC) connects the microgrid to the grid. It acts as an interface for the exchange of electrical energy [43]. At this stage, the microgrid needs to satisfy the set interface specifications, like those found in the IEEE Standard 1547 series [44]. In grid-connected mode, the microgrid operates in accordance with distribution network regulations independently of the main power system [45]. Whether the microgrid is in connected or grid-islanded mode, it is still subject to uncertainties and imbalances that need to be managed in order to maintain the microgrid’s proper operation (Figure 15).

3.3. Integration of Renewable Energy Sources

The incorporation of multiple sustainable energy sources, including solar, wind, bioenergy, hydropower, marine energy, and combined heat and power (CHP), can result in a microgrid that can offer clean energy that emits no carbon dioxide or other pollutants, diversifies energy sources, lessens reliance on imported fuels, boosts energy independence, generates new job opportunities, and is more affordable. Integrating renewable energy sources, operating with some autonomy, and using contemporary controls can all help the microgrid function better in terms of quality, dependability, and resilience [47].

3.4. Energy Storage System

Energy storage system (ESS) application increases power availability, resilience, and stability for load demand. In the event that the primary power supply source fails or is insufficient, ESS functions as a backup power system. Load shedding, load leveling via energy storage system charging or discharging, power fluctuation smoothing, voltage and frequency fluctuation issues, and power quality improvement are additional areas where ESS can be useful, as shown in Figure 16. A classification of energy storage systems is presented in

Table 1. Energy storage systems.

Technology	Type	Invention Date [49]	Response Time [50]	Advantages [51]	Drawbacks [51]
Compressed Air Energy Storage	Mechanical	In 1978, Germany developed the world’s first compressed air energy storage plant.	1–15 min	High capacity	Need for fuel and underground cavities.
Pumped Hydro Storage	Mechanical	1907 was the first use of pumped storage	s–min	High capacity	Geographical restrictions
Thermal Energy Storage: Hot-water storage; Molten-salt energy storage; Phase change material storage (PCM); and Thermochemical Energy Storage (TCES).	Thermal		-	High energy density	Lower efficiency
Hydrogen storage and synthetic natural gas (SNG)	Chemical	The history of hydrogen started in 1766.	-	Minor environmental issues.	High investment costs.
Hydrogen Fuel cell	Chemical	1839, by Sir William Robert Grove	<1 s	Long term storage	expensive catalyst
flyweel	mechanically store energy as kinetic energy	1883, by John A. Howell	<4 ms–s	High power and efficiency	Lower energy density
Second-life batteries Lead-acid batteries Nickel-Cadmium batteries Sodium-Sulfur batteries Lithium-ion batteries Vanadium redox-flow batteries Flow batteries	Electrochemical	Lead acid battery: 1859 by Gaston Planté. Nickel-cadmium battery: 1899 by Waldemar Jungner. 1960 by Ford Motor Company. Lithium-based batteries has been invented by Michael Stanley Whittingham in 1976. The graphite anode (negative pole) for lithium-ion batteries was developed by Rachid Yazami. The first commercial lithium-ion battery was released by Sony in 1991.	- 5–10 ms ms 1 ms 20 ms–s - -	Higher power and energy density	Cycle life. Temperature Dependent
Capacitor	Electrical		ms	Fast reponse	Lower efficiency
Supercapacitor	Electrical	1982 by the Pinnacle Research Institute	8 ms	High efficiency	Low energy density
Superconducting magnetic energy storage	Electrical		<100 ms	High effiicency	Impact on health

.

3.5. Electric Vehicles (EVs) and Green Hydrogen Integration

Over 2021 and 2022, the transportation sector’s global CO₂ emissions increased by 3% [52]. The transportation sector’s carbon dioxide emissions in North Africa are depicted in

Table 2. CO₂ emissions by the transport sector in North Africa.

Country	Transport Sector Emissions Pollutants	Reference
Morocco	30%	[53]
Algeria	33.33%	[54]
Tunisia	25.8%	[55]
Libya	30.7%	[56]
Egypt	32%	[57]

.

Electric cars can be utilized as Vehicule to Grid (V2G) or Vehicule to Home (V2H) devices to discharge stored energy back into the grid or into residential buildings. Designed for use in both the home and workplace, iSmart is a new line of intelligent charging outlets that are exclusively Moroccan. An entire car charge can be accomplished in thirty minutes with the borne, which has power ranges of 7.5 kW, 22 kW, and 50 kW, as mentioned in Figure 17 [58]. Therefore, the number of charging points increases by putting up electric charging infrastructure, as shown in Figure 18.

As for hydrogen, which is considered the flag shift of the energy transition, it can decarbonize a wide range of industries, be a powerful alternative to pollutant-rich sectors, and accelerate the energy transition. For hydrogen purity issues, hydrogen requires potable water to be produced, and because of the scarcity problem in the North African region, water desalination must be a solution to produce hydrogen from sea water, as shown in Figure 19.

3.6. Role of Fuel Cells

As a dispatchable energy source, fuel cells can improve the quality and reliability of power while preserving the balance between supply and demand. Fuel cell technology (FC) integration into microgrids is a promising development because fuel can be used to continuously produce electricity. Furthermore, the electrolyzer and FC, working together, can provide a long-term energy storage solution to support the load [62]. A polymer electrolyte membrane (PEM) fuel cell, for example, emits no CO₂ because its only byproducts are heat and water. Hydrogen, which is produced by water electrolyzers using renewable resources, serves as the fuel source for fuel cells. Fuel cells are becoming more and more popular in microgrids as a power source, much like in the transportation sector, and as a buck-up system for the re-electrification of hydrogen.

4. Review of Microgrid Control and Design

4.1. Microgrid Control Methods

The challenge of classifying and identifying control techniques in distributed energy resources (DERs) based on microgrids remains difficult due to the fact that each microgrid has unique characteristics [13]. Microgrid control strategies aim to maintain power balance, regulate voltage and frequency, facilitate seamless islanding transitions, and optimize economic dispatch. These strategies fall into three categories, which are represented in Figure 20a–c, respectively, for centralized, decentralized, and hierarchical approaches.

Some researchers have favored centralized control strategies, where a single entity makes control decisions for the entire system. Centralized control employs a central controller, typically a supervisory control and data acquisition (SCADA) system, to manage all microgrid components. This approach offers precise control and coordination but is vulnerable to communication disruptions and single points of failure [64,65]. Others have explored decentralized approaches, where different entities make control decisions independently. Decentralized control distributes control responsibilities among individual DERs, enabling autonomous responses to local conditions and peer-to-peer communication. This approach enhances resilience and scalability but may pose challenges in maintaining overall system stability [66,67,68,69]. Some other researchers have proposed another approach that combines centralized and decentralized elements to optimize the operation of the microgrid and address the complex challenges associated with DGs. Hierarchical control combines the benefits of centralized and decentralized approaches by employing a multi-layered structure. A central controller oversees the overall microgrid operation, while local controllers manage individual DERs or groups of DERs. This approach provides flexibility and scalability while maintaining system-wide coordination [70,71,72,73].

Each one of the above control structures have some advantages and drawbacks. For instance, centralized control offers optimal coordination but may be vulnerable to failures in a central entity and can result in communication delays, while decentralized approaches enhance resilience and reduce dependence on a single entity but may lack overall coordination and efficiency. Hierarchical control combines the advantages of both, but its implementation can be complex, requiring effective coordination between levels.

Table 3. General comparison of three microgrid control strategies.

Feature	Centralized	Decentralized	Hierarchical
High resilience	×	✓	✓
Central controller	✓	×	✓
Low cost	×	✓	✓
High flexibility	×	✓	✓
High coordination	✓	×	✓
Easy implementation	✓	✓	✓

× means No. ✓ means yes.

provides an overview of key differences between different control strategies [74].

4.2. Hierarchical Control Structure in Microgrids

The increasing penetration of DERs into microgrids (MGs) has introduced new challenges in maintaining stability, efficiency, and resilience [75]. Hierarchical control strategies have emerged as a promising approach to addressing these challenges by decomposing the complex control problem into multiple levels with specific functionalities. Three levels usually make up the hierarchical control system, which is depicted in Figure 21: primary, secondary, and tertiary control [76,77].

The three levels of control are interconnected and interdependent with specific functionalities, as shown in

Table 4. Functionalities of hierarchical control levels.

	Primary	Secondary	Tertiary
Objective	Maintain voltage and frequency stability at the DER level	Restore voltage and frequency deviations caused by primary control actions	Handle supervisory tasks, such as economic optimization, energy management, and demand response programs
Responsiveness	Fastest response time (milliseconds)	Slower response time (seconds to minutes)	Slowest response time (minutes to hours)
Mechanisms	Droop control, inverter control	Communication-based control and adaptive control	Optimization algorithms, decision-making tools
Role	Ensures fast power sharing among DERs to respond to sudden changes in generation or load	Coordinates power dispatch among DERs to optimize energy consumption and minimize operational costs.	Interacts with the main grid for energy exchange and backup power, ensuring long-term economic and operational efficiency

. Primary control provides the foundation for stable operation, while secondary and tertiary control layers refine and optimize the microgrid’s performance. Information flows between the levels to ensure coordinated control and overall system stability [78].

4.3. Primary Control Level

Primary control levels in the literature are generally classified into two or three classes based on their functions. Consequently, grid forming, grid following, and grid-supporting are the three control mechanisms that can be applied, as explained in [79,80]. However, because the grid-supporting approach falls within the grid-forming category because it aids in network voltage regulation, some research [72,81] ha simplified this classification by restricting the number of principal control techniques to grid-forming and grid-following.

4.3.1. Grid following Control

DG voltage and frequency are controlled by the main grid while operating in a grid-connected mode. In this case, the generation systems typically operate in current control mode to extract the maximum possible energy, for example, maximum power point tracking (MPPT) for solar and wind systems [72]. Furthermore, if the active and reactive power references are established by superior control levels, this kind of control may function at a non-optimal point beyond the maximum power band [82]. In this context, the work of Blaabjerg et al. [83] reviews the grid following control schemes and covers grid synchronization techniques, including phase-locked loop, zero crossing, and grid voltage filtering. These techniques are made up of two nested loops: an external voltage loop and a fast-dynamic internal current loop [84]. Following this work, a categorization of primary control methods was recently discussed in the work of Costa, Israel D et al. [85], namely, the dq reference frame, the alpha beta reference frame, and the abc reference frame. Similarly, some of the most relevant control techniques for the integration of MGs into the electrical grid are detailed in the works of Sachidananda Sen et al. [86] and Rodriguez et al. [87].

4.3.2. Grid Forming Control

When MGs operate in autonomous mode, the renewable energy sources (RESs) and storage systems must be responsible for the stability of voltage and frequency, which therefore requires optimal control of active and reactive powers to guarantee that the devices share power in an appropriate manner. Guerrero et al. [88] have already handled this kind of control in their work for islanded MGs. Depending on the needs, a portion or all of the RES inverters will operate in parallel to ensure microgrid voltage control in grid-forming mode and in grid-following mode for the remaining ones [89]. The following are the two primary configurations in which this control approach has been used:

-

Single grid forming unit:

In this scenario, one sustainable energy source is interfaced to a grid-forming power converter to generate the voltage and frequency references, while all other units operate in grid-following mode and are controlled by the current to extract the maximum possible energy from the RESs [90,91].

-

Multiple grids forming units:

This approach involves the use of several renewable energy generation units operating in grid-forming mode. Consequently, synchronization is required to ensure the voltage and frequency stability of the microgrid while meeting power demands. Depending on whether or not a communication network links the interface converters, the techniques in this field can be classified into several categories [92,93,94].

4.4. Secondary Control Level

While primary control provides a rapid response, it may not be sufficient to fully restore voltage and frequency stability, especially in large or complex MGs. Secondary control, operating on a slower timescale, addresses the limitations of primary control. It uses communication and coordination among DERs and central controllers to achieve precise voltage and frequency regulation. For this purpose, several methods have been suggested in research publications, as Li et al. [95] propose a two-layer control architecture for effective secondary control in microgrids. The top layer, a communication-based layer, facilitates the exchange of information between the central controller and DER agents. The bottom layer comprises the DERs themselves, which respond to control signals from the top layer. The authors in [96,97] have pointed out that communication plays a pivotal role in secondary control. DERs can exchange information about their local voltage and frequency measurements with their neighbors and the central controller. This information is essential for coordinated control actions that restore and maintain voltage and frequency stability across the microgrid. Also, some researchers have investigated the optimization techniques for secondary control that can be implemented using various optimization techniques. Shan-Yinghao et al. [98] focus on optimization methods driven by economic considerations, while Simpson-Porco et al. [99] emphasize power quality optimization. Another important aspect to consider in secondary control is voltage and reactive power sharing, where maintaining voltage stability and ensuring equitable reactive power sharing among DERs are key challenges in secondary control. A. Ait Ben Hassi et al. [100] introduced a method to get voltage and frequency back to their nominal settings without appealing to power sharing between DG systems. Practical MG implementations must consider the possibility of communication failures. Effective secondary control schemes should be capable of maintaining stability and power quality even under communication disruptions [99]. There are two possible implementation modes for secondary control: distributed and centralized control. A central controller that manages all DERs is required for centralized control, while distributed control utilizes decentralized algorithms that enable self-coordination among DERs. A typical hierarchical control technique used to regulate frequency and voltage for a voltage source inverter is shown in Figure 22.

4.4.1. Centralized Secondary Control

In centralized secondary control, loads are assigned to a common bus and controlled by a central controller. In this context, Peyghami et al. [101] suggested a controller that keeps the voltage near a reference value by controlling the DERs’ output and continuously tracking the voltage at the common bus. Because the communication lines between the MG converters and the central controller are not always dependable, distributed approaches are utilized by Nasirian et al. [102] for each converter. Subsequently, the voltages at specific buses are transmitted to the converters, ensuring the regulation of the average [103,104]. A frequency-based droop control is proposed by Peyghami et al. [105].

4.4.2. Distributed Secondary Control

Distributed secondary control approaches are often employed due to the inherent limitations of centralized control, such as the vulnerability to communication failures. In distributed secondary control, each DER has its own local controller that communicates with neighboring DERs to coordinate voltage and frequency regulation. In this context, Shafiee et al. [106] proposed an approach that involves the implementation of a distributed control strategy, where necessary control signals are locally transmitted to the primary controllers. In this configuration, every DG unit comes with a local controller. Consequently, in the event of a failure in one unit, only that specific unit experiences a breakdown without impacting the functionality of other DG units [106]. Distributed control methods manage both voltage regulation and power sharing, and numerous methods have been presented in the literature [107]. For instance, the use of input-output feedback linearization for voltage control at the secondary level of MGs has been demonstrated by Bidram et al. [108]. This method requires a distributed communication network with unidirectional communication, proving more efficient than centralized controllers at the secondary level.

4.5. Tertiary Control Level

The higher level of MGs control is called the tertiary control level and focuses on tasks such as economic optimization, energy management, and demand response programs. It operates on longer timescales (from minutes to hours) and interacts with the main grid for energy exchange information. Its objectives include minimizing energy costs, ensuring energy balance, and coordinating demand response [109]. Using optimization algorithms, decision-making tools, and communication protocols, it interacts with secondary and primary control layers to optimize system performance. The process involves data collection, processing, decision-making, command generation, and continuous monitoring for adaptation. The benefits include economic efficiency, improved energy management, optimized demand response, and resilient operation. Challenges include computational complexity, dealing with uncertainties, and effective communication. Future efforts aim to improve the level of tertiary control by developing more efficient algorithms, incorporating advanced forecasting techniques, and implementing distributed control strategies for scalability and resilience in environments with limited communications [110].

4.6. Conventional and Advanced Control Techniques in Microgrids

In essence, MG applications have employed a range of traditional control strategies, including PI, PID, and droop control strategies with set parameters for a specific operating point. The control system parameters can be set to the recommended values to achieve proper steady-state performance. But when operating conditions change highly, these approaches are limited in their ability to ensure a satisfactory compromise of dynamic performance. Therefore, a readjustment of the control parameters is required [111]. Furthermore, because of the RESs’ low inertia, uncertainties, and fluctuating nature, one of the primary issues is maintaining the stability and control of autonomous MGs [112]. Therefore, more sophisticated and advanced control structures must be developed in order to maintain the intended performance in the face of unanticipated disturbances and model uncertainties. A classification of MG control techniques is depicted in Figure 23.

The following subsections are devoted to the traditional and advanced control methods used in MGs.

4.6.1. Conventional Control Method Using PI/PID

Because of its straightforward design, the PI/PID controller has been an essential component for the control of power generation systems and the industrial sector for many years. With the right gain tuning, this controller provides almost ideal control system performance. Various strategies have been employed in the literature to tune PID gains, such as Ziegler-Nichols, Cohen Coon, Chien, Hrones, and Reswick, and the empirical rule [111]. A frequency control by the PI controller of an island microgrid, including energy storage using soft computing has been proposed by M. Dashtdar et al. [113] to overcome these limitations. This technique automatically modifies and optimizes the coefficients of the PI-controller.

4.6.2. Linear Quadratic Control

In this method, the optimal control law is determined by minimizing or maximizing the performance criterion, which is the basic principle of linear quadratic control. Choosing the right weighted matrices Q and R as required by the Riccati equation to guarantee a suitable response is one of the primary difficulties presented by linear quadratic regulation. This kind of control has been used to address a number of microgrid research issues. For instance, H. Keshtkar [114] introduced an enhanced optimal LQR controller for the purpose of regulating frequency in a resilient microgrid subjected to cybersecurity intrusions. In [115], the authors have presented an output-feedback-based robust LQR Vf controller for PV-battery MG with a decentralized structure and consideration of generation uncertainties.

4.6.3. Sliding Mode Control

Sliding mode control (SMC) has undergone extensive exploration as a robust nonlinear control method aimed at ensuring stability in the face of varying parameter constraints. The application of SMC to systems with variable structures may encounter the chattering problem. Common issues associated with SMC include undesirable transient errors and non-zero steady-state errors. However, the direct-action controller’s smart design and sliding surface reduce chattering. Mishra et al., in their study [116], designed a PV-diesel microgrid with a multiple feedback loop topology that can function in both isolated and grid-connected modes. Within this framework, three setpoints, based on the PI controller, were formulated to generate the necessary reference values for the internal loops of the PV system utilizing second-order sliding mode control. A comparable strategy is evident in the research by Cucuzzella et al. [117], where higher-order sliding mode control techniques were advocated for a microgrid featuring multiple energy resources interfaced with a voltage source converter. Specifically, a second-order sliding mode control algorithm was deployed to guarantee efficient chattering attenuation, a third-order sliding mode controller was used for the isolated mode, and a second-order sliding mode control technique was used for both grid-connected and isolated modes.

4.6.4. Model Predictive Control (MPC)

MPC algorithms are improved to anticipate reference signals. One of the key benefits of these control methods is their capacity to lower tracking errors. In their study [118], Wang et al. introduced a voltage predictive control strategy for autonomous MGs, utilizing an estimator for each energy resource to acquire a voltage response independently of a communication link. The proposed control strategy involves both an estimator and a predictive controller for voltage control. Employing small-signal analysis, this approach can effectively execute offset voltage control without relying on any communication link. Furthermore, it demonstrates resilience against parametric uncertainties such as disturbances in line parameters, variations in load parameters, disruptions in output impedance, and faults in the energy resources. Another notable contribution in the field of autonomous MGs is the optimal predictive control scheme presented by Minchala-Avila et al. [119]. This scheme employs a nonlinear predictive control algorithm to manage a dataset encompassing the state of charge of batteries, energy resource production, and predicted load. The methodology preserves an appropriate equilibrium between power generation and consumption and maintains the voltage amplitude within the allowable deviation range.

4.6.5. AI-Based Control Techniques

A wide variety of heuristic algorithms have been utilized in the literature to improve MG’s energy management and optimization. These include methodologies such as particle swarm optimization, fuzzy logic, neural networks, and genetic algorithms. The successful application of these techniques extends to both the connected-to-the-grid and standalone modes of MGs. Additionally, the literature introduces robust methodologies aimed at finely tuning controller parameters. Among them are the differential evolution algorithm [120], the bacterial foraging optimization technique [121], the colonial competitive algorithm [122], and the cuckoo search technique [123]. Below is a quick analysis of particle swarm optimization, fuzzy logic, and neural network optimization.

(a)

Particle Swarm Optimization (PSO)

The PSO technique has garnered significant research efforts due to its effectiveness in optimizing uncertain parameters for a wide range of complex optimization problems. In ref [124], the authors developed an intelligent online technique combining the PSO algorithm and fuzzy logic to optimize PI gains. The proposed technique achieves better stability performance compared to fuzzy logic and Ziegler-Nichols and is robust against environmental and dynamic variations. Another work is presented in [125] by Hassan et al., using the PSO technique to optimally design the microgrid in every operation mode. It is employed to look for the best power sharing coefficient, filter, and PI controller parameter values.

(b)

Fuzzy Logic Control (FL)

The FL control technique is extensively used in various domains. It is considered one of the most significant intelligent tools for resolving power optimization problems related to renewable energy sources. In such a scenario, this technique is often used to optimize energy resources such as wind and solar power. In islanded mode, the authors in [126] suggested three methods for smoothing wind power. Fuzzy logic is used to construct a pitch angle controller that helps to partially smooth changes in output wind power. Another application of fuzzy logic is introduced in [127], where an automatic adjustment structure of PI regulator parameters is used for active and reactive power control in MGs.

(c)

Neural Networks

The concept of neural networks is fundamentally inspired by the workings of the human brain, and these networks have found applications across various research domains. Neural networks prove to be highly effective in tasks such as system parameter identification, control, and optimization, whether applied in offline or real-time scenarios. In the research conducted by Mohamed et al. [128], they utilized neural network identification and current regulation to devise an adaptive interfacing scheme for MGs functioning in a grid-connected mode. Additionally, another study by the authors in [129] introduced an intelligent online management approach for active and reactive power in a synchronous static compensator. This method, driven by neural networks, aims to prevent instability and failures in autonomous MG systems.

4.6.6. Adaptive Control Techniques

Adaptive control techniques can be used to successfully achieve robust convergence, stability, and dynamic system tracking. Adaptive solutions are a desirable way to address inadequate controller performance when a system’s operating conditions change. A number of studies, including adaptive PI/PID control [130], adaptive droop control [131], adaptive sliding mode control [132], and reinforcement learning [133], have been reported in the literature on adaptive control in MGs.

4.6.7. Robust H-Infinity Control

Numerous academics are interested in the application of the robust H-infinity control concept to microgrid frequency and voltage regulation for a number of reasons. Firstly, H-infinity techniques and µ-synthesis can address control objectives such as disturbance attenuation, robust stabilization of uncertain systems, and shaping the open-loop response [134]. The founded solution can be considered optimal concerning the defined criterion. Nevertheless, there would not be a solution if the control objectives cannot be met. Second, a strong relationship exists between control design and the required dynamic performance. Consequently, appropriate implementation of control system design is necessary to achieve control objectives. Finally, µ-synthesis and H_infinity approaches may be related to sensitivity and robustness analysis for model system uncertainties. Overall, linear matrix inequality (LMI) techniques have been used to handle H-infinity controller synthesis challenges in order to guarantee system stability in the presence of model uncertainties and external disturbances [135].

4.6.8. Active Disturbance Rejection Control (ADRC)

The ADRC provides a novel approach to address model uncertainties and external disturbances in in microgrids. This algorithm, initially proposed by Professor Jingqing Han [136], has gradually found applications in various industrial domains such as aerospace, robotics, and renewable energy production [137]. Y. Oubail et al. [138] have proposed an active disturbance rejection control technique that is applied to an islanded PV/wind/battery MG, where the objective is to address the frequency and voltage variations and also to enhance the power quality of the MG by employing a shunt active power filter (SAPF). In [139], the authors have presented a dynamic modeling and a robust control by ADRC of a grid-connected hybrid PV-Wind microgrid and tested its robustness to internal parameter variations.

4.7. Comparison Analysis of the Control Techniques

An overview and analysis of the explained control techniques is provided in

Table 5. Summary analysis of microgrid control techniques.

Control Technique	Level	Advantages	Drawbacks	References
PID control	Primary	Simple to implement, effective for maintaining voltage and frequency stability	May not handle complex dynamics and uncertainties	[111,113]
Droop control	Primary	Simple, robust, and balanced power sharing among DERs	May not provide precise voltage and frequency regulation	[105,112,140]
LQR	Primary	Optimal performance, can handle complex dynamics and uncertainties	Requires an accurate system model and knowledge of system parameters	[114,115]
Sliding mode control	Primary	Robust to disturbances and uncertainties	Complex to implement, May require a high switching frequency	[116,117]
Backstepping	Primary	Can handle complex nonlinear dynamics	Complex to design and requires an accurate system model	[141,142]
ADRC	Primary	Robust to disturbances and uncertainties and can handle complex dynamics	Requires careful selection of parameters	[137,138,139]
Consensus dispatch optimization	Secondary	Enables decentralized and coordinated power dispatch among DERs	May require extensive communication between DERs	[143,144]
Virtual impedance	Secondary	Robust to disturbances and uncertainties, Enables decentralized voltage regulation	Requires communication between DERs	[140,145]
VSG (Virtual Synchronous Generator)	Secondary	It emulates the behavior of a synchronous generator, providing inertia and damping	Requires an accurate system model and knowledge of system parameters	[146,147]
PSO	Secondary	Optimizes power dispatch and can handle complex dynamics	Computationally demanding, you may get stuck in local optima	[124,125]
GA (Genetic Algortithms)	Secondary	Optimizes power dispatch, can handle complex dynamics	Computationally demanding, may not converge to the global optimum	[148,149]
MPC	Primary and Secondary	Optimal performance, Can handle complex dynamics and uncertainties	Computationally demanding, requires an accurate system model	[118,119]
Adaptive control	Primary and Secondary	Adapts to changing operating conditions	Complex to implement, and may require extensive training data	[130,131,132,133]
Fuzzy logic control	Primary and Secondary	Can handle nonlinear dynamics and uncertainties	Requires careful design of fuzzy rules	[126,127]
Neural network control	Secondary	Can learn complex relationships from data	Requires large amounts of training data and may be difficult to interpret	[128,129]
H-infinity control	Primary and Secondary	Robust to disturbances and uncertainties	Complex to design, requires accurate system model	[134,135]
Decentralized economic dispatch	Secondary and Tertiary	Simple to implement, scalable	May not achieve optimal performance	[150,151]
Multi-agent systems	Tertiary	Can handle complex interactions among DERs	Complex to design and requires coordination between agents	[152,153]
Event-triggered control	Secondary and Tertiary	Reduces communication overhead	May not be suitable for systems with strict latency requirements	[154,155]
Game theory-based approaches	Tertiary	Model strategic interactions among DERs for resource allocation	May be challenging to implement in real-time applications	[156,157]
Demand response management	Secondary and Tertiary	Coordinates energy consumption patterns to balance supply and demand	May require behavioral changes and incentives for consumers	[158]
Deep reinforcement learning	Tertiary	Can learn optimal control strategies from data	Requires large amounts of training data and may be difficult to interpret	[159,160]
Decentralized consensus control	Secondary and Tertiary	Enables resilient operation in the presence of communication failures	May not achieve optimal performance due to limited information exchange	[161,162]
Distributed optimization algorithms	Secondary and Tertiary	Enables collaborative decision-making among DERs	May require extensive computational resources	[163]

. The table compares the performance, robustness, control level, and complexity of each technique by giving the advantages and drawbacks of each control strategy. Other applied techniques for both secondary and tertiary control were also given. This table shows that advanced control techniques offer better performance and robustness than classical techniques, but they are generally more complex to implement. The control approach is chosen based on the specific requirements of MG usage.

4.8. Resilient Control Strategies for Microgrids

Resilient control of microgrids refers to the design and implementation of control strategies that enable microgrids to withstand and recover from disturbances, uncertainties, and cyberattacks. Microgrids, as distributed energy systems, are inherently vulnerable to various disruptions, including natural disasters, equipment failures, and intentional cyberattacks. Resilient control aims to enhance the stability, reliability, and security of microgrids by incorporating mechanisms to detect, mitigate, and adapt to these challenges. Resilient control is an active area of research, with ongoing efforts to develop advanced control strategies and technologies that can further enhance the resilience of MGs in the face of increasing complexity and interconnectedness.

Key characteristics of resilient control for MGs include:

-

Disturbance tolerance: The control system should be able to maintain stability and performance when perturbations are present, like sudden variations in load or generation, faults, and weather events, by using control techniques such as ADRC, droop control, H-infinity, and nonlinear controls [134,135,137,138,164].

-

Adaptability: The control system should be able to adapt to changing operating conditions and uncertainties, such as variations in renewable energy generation and load patterns, where adaptive control algorithms are suitable because they are continuously adjusting their parameters based on real-time system data to maintain stability and performance despite changing conditions [132,133].

-

Cybersecurity: The control system must be designed to withstand cyberattacks, such as unauthorized access, data manipulation, and malware infections. Some researchers have proposed a cybersecurity protocol to protect the control system against cyberattacks by implementing measures such as encryption, access control, and intrusion detection [165,166].

-

Fault-tolerant and self-healing: The control system should have mechanisms to detect and isolate faults and to restore normal operation following disruptions. Various techniques were used in this context. For instance, in the research conducted by T. Wang et al. [167], they developed a model-based fault detection and isolation algorithm in DC microgrids in order to identify and isolate faults in the MGs, preventing them from propagating and causing widespread disruption.

-

Decision-making under uncertainty: The control system must be able to make informed decisions under conditions of uncertainty when complete information about the system’s state or future events is not available. The authors in [167] have proposed a three-stage decision-making approach that considers the spatial uncertainty of failure probabilities, which are represented by probability distributions, and weights the loss of performance due to preventive and post-event measures.

5. Control Methods Applied to Microgrids Developed in the North African Region

One of the methods inspired by the natural variation of African countries is the African vulture optimization algorithm (AVOA), which is a new methaheuristic algorithm that mimics African vulture behavior and competition for food. This algorithm was proposed in 2021 by Benyamin Abdollahzadeh et al. in ref [168], and it is used for global optimization problems and has been applied in different studies. For instance, an autonomous residential microgrid was studied experimentally and through simulation by the authors of Ref [169]. The suggested method relies on the African Vulture Optimization algorithm (AVOA) to minimize operation costs and demand-side management to overcome the stochastic and interrupted character of renewable energy sources (RES), lower energy costs, and maintain customer comfort and lifestyle. Given that the building is situated in a region with appropriate wind speed and radiation, such as countries in North Africa, the suggested strategy performed well in terms of meeting customer demands and covering load needs. In reference [170], a cascade double-loop control scheme for controlling the DC bus voltage of an islanded DC microgrid is examined. It is based on a PI controller tuned with the AVOA algorithm. The outcomes demonstrate that the AVOA-based PI controller outperforms the PSO-based PI controller in terms of performance.

Numerous difficulties arise as a result of the growing usage of renewable energy sources, including frequency deviation brought on by fluctuations, uncertainty, and the unpredictability of RESs. A (1+PD) and PID combined controller-based African vulture optimization algorithm (AVOA) was proposed by the authors in Ref [171]. Wind speed variations and solar radiation uncertainties are taken into consideration when evaluating the system. The study shows promising findings for frequency deviation compensation. From the same angle and in line with the roadmap that North African countries have put forward to develop green hydrogen facilities, a solar-islanded microgrid with a renewable hydrogen conversion unit is proposed in reference [164]. The system is controlled with a primary control type of P/F-Q/V with a virtual output impedance and a secondary control-based PI controller. The system is simulated under step load changes, taking into consideration the penetration of renewable energy sources. The results are satisfying regarding the voltage and frequency restoration.

6. Conclusions

The idea of this review is to describe the status of microgrid development in Africa per participant country, communication network, and cybersecurity considerations. The North African region has potential for hydrogen production and renewable energy. Countries need to increase their investments and cooperative initiatives in order to meet the 2030 targets and expedite the energy transition. Microgrids reduce reliance on fossil fuels and are crucial, especially for isolated communities in North African nations. Instead of using predefined microgrid solutions adapted for European or North American scenarios, the objective is to demonstrate that taking into consideration local African particularities, including weather conditions and local population needs, would lead to more cost-effective solutions. Resilient control is essential in microgrids in order to enhance stability and ensure the ability to provide power for consumers. Depending on the main objectives of the control strategy, resilient control can be applied in the three layers of microgrids: primary, secondary, and tertiary. Additionally, it is used to counteract threats that are frequently reported, such as cyberattacks, disturbances and faults, communication errors, and natural disasters.

In our future work, we will be conducting research on microgrid monitoring, practical demonstration and integration studies, and microgrid cooperation and coordination for off-grid operation to increase the amount of energy flexibility exchanged among microgrids for the delivery of energy services.