A flexible multi-agent system for managing demand and variability in hybrid energy systems for rural communities

Abstract

Access to reliable, economical, and sustainable energy is a critical challenge in remote communities where infrastructure constraints and unreliability of renewable energy sources (RESs) complicate the possibility of having a stable supply. This study is motivated by the urgent need for intelligent, adaptive energy management systems that can ensure the reliability of the supply while maximizing the use of RESs. To meet this need, an adaptive and scalable multi-agent system (MAS) framework for hybrid energy systems can be employed. The system includes electric vehicle batteries (EVBs), hydrogen energy storage systems (HESSs), and battery energy storage systems (BESSs) and wind turbines (WTs) and PV. A hybrid backup architecture for energy supply continuity in low availability of RESs, in addition to vehicle-to-grid (V2G) functionality enabling EVBs to support grid stability. The MAS is evaluated under four scenarios: PV–WTs–BESSs, PV–WTs–BESSs–EVBs, PV–WTs–BESSs–HESSs, and PV–WTs–BESSs–EVBs–HESSs. Scenario 4 attains the lowest operating cost of $10,688.06, a reduction of 0.91% from scenario 1, in a 25 kW peak load microgrid. The artificial gorilla troops optimizer optimizes the real-time energy dispatch by learning to adjust to changing system conditions. Simulation results confirm that the proposed MAS improves cost-effectiveness, energy stability, and sustainability in constrained settings.

Introduction

The transition to renewable energy systems has become a global imperative as nations seek to reduce carbon emissions, enhance energy security, and provide sustainable energy solutions to remote communities^1,2,3,4. Remote areas, particularly in developing countries like Egypt, have unique challenges in achieving these goals. Limited infrastructure, dilapidated power supply, and non-uniform energy demand necessitate innovative approaches for energy management^5,6,7. Hybrid energy systems integrating a number of renewable energy sources such as PV and wind with advanced energy storage technologies are a suitable solution for off-grid locations. Hybrid systems can be used to maximize the utilization of renewable energy, reduce the dependence on fossil fuels, and stabilize the grid^8,9,10,11. Sophisticated management strategies are, however, required to handle the inherent variability of renewable energy and the advanced dynamics of hybrid systems. This work presents a dynamic MAS structure that meets such challenges. The MAS framework includes renewable sources and hybrid energy storage devices comprising batteries, hydrogen storage devices, and EVBs. Using real-time data and improved optimization algorithms, the MAS provides maximum energy dispatch, minimizes cost of operation, and enhances system reliability.

Literature review

The coordination of renewable energy sources with novel energy storage methods has been an area of emphasis over the last few years given the global push toward sustainable and reliable energy systems^12,13,14. Previous research indicates towards the prospect of hybrid energy systems as an answer to reducing renewable energy variability and peak load fluctuations, particularly for off-grid regions^15,16,17. MAS has also proven to be a good framework for managing these complex systems since it is flexible, scalable, and able to make decisions in real-time. However, the majority of classical approaches utilize fixed optimization or forecast-based techniques, which are likely to struggle to adapt under changing and uncertain conditions^18,19,20,21. In²², grid-integrated renewable energy system PV, wind, and fuel cells (FCs) was optimized by implementing Walrus Optimization Algorithm (WaOA), Coati Optimization Algorithm (COA), and Osprey Optimization Algorithm (OOA). The system achieved a cost of energy (COE) of $0.517/kWh with negligible loss of power supply probability (LPSP), with the focus on feasibility and reliability. In²³, they used the Java Agent Development Framework (JADE) to implement a MAS for distributed hybrid renewable energy management. Five agents and a three-layer framework make up the system, and it was validated using scenario testing, demonstrating the feasibility and usability of the system in the way of energy management optimization. In²⁴, MAS has been applied for the optimization of the energy management of BTSs in Pakistan based on RET and DSM. Energy trading prices between the grid were 0.08 USD/kWh and between BTSs were 0.04 USD/kWh, and 44.3% reduction for grid reliance and 71.4% reduction for carbon emissions were obtained, which shows cost-saving and improvement of sustainability. In their pursuit to ensure efficient energy management in hybrid renewable systems, other frameworks developed include hybrid energy management systems (EMS) that enhance MAS by focusing on optimizing the utilization of diverse energy sources, efficiency, and minimization of operational expenses in merged dynamic power configurations.

In²⁵, an integrated EMS for renewable energy sources with PV, PEMFC, SC, and batteries was proposed. By integrating fuzzy logic and frequency decoupling methods, it reduced hydrogen usage by 19.6%, expanded battery state-of-charge (SOC) by 5.4%, and improved cost-effectiveness and storage duration. In¹³, an FPGA- and AVOA-based EMS was optimized for microgrid operation to reduce operational cost by 3.77%, prolong BESS lifetime, and suppress peak load caused by EV charging. In²⁶, a slime mould algorithm (SMA) is used for energy management in a grid-connected microgrid with RERs, EVs, diesel generators, and batteries. The approach improves reliability, voltage stability, and cost-efficiency under deterministic and probabilistic conditions in the IEEE-33 bus system. In²⁷, a hierarchical decentralized EMS was developed for intercoupled MGs using the marine predator algorithm (MPA). The EMS optimized the utilization of power, filtered power quality problems, and facilitated energy sharing, allowing for reliability, stability, and operational efficiency improvement in remote clusters. In²⁸, a cost-effective microgrid that integrated solar, wind, diesel, and battery storage was envisioned for northern Kandahar, Afghanistan. Utilizing the HOMER software tool, the most efficient design had a COE value of 0.19$/kWh, proving itself to be a viable sustainable energy option. In²⁹, a stochastic energy management system for hybrid microgrids operating in islanded mode was presented. The system used transient search optimization and Monte Carlo simulation to reduce operation costs, improve reliability, and optimize power dispatch with consideration of renewable energy uncertainty and battery and EV integration. In³⁰, an FPGA-based AI-aided intelligent EMS optimized microgrid autonomous mode operation. With the inclusion of optimization techniques, it increased reliability, reduced costs by 26.43%, and provided quick response to load or power generation changes, thus making energy management economical and effective. In³¹, a hybrid renewable energy system (HRES) consisting of horizontal (HAWT) and vertical axis WTs (VAWT) was simulated using the multi-objective genetic algorithm (MOGA). Outcomes revealed that the HAWT-based systems were more effective and cost-efficient ($0.02/kWh) with an 80.5% renewable share and 73.2% less CO₂ emission. In³², a slime mould algorithm is applied to optimize autonomous microgrid operation with demand response programs. Results show it outperforms PSO and GA in reducing emissions and operational costs under curtailable and interruptible load scenarios. In³³, an enhanced slime mold multi-objective algorithm (MOISMA) is employed for the optimal energy management of a multi-renewable microgrid in an uncertain environment. In addition to time-of-use tariffs and demand response, the algorithm significantly reduces the generation cost and emissions while outperforming MOGA and MOPSO to deal with unstable wind and sun patterns. In³⁴, the hybrid pumped hydro storage (PHS) and floating photovoltaic (FPV) system was optimized through the MOGA. The system achieved an energy cost of $40/MWh, a decrease in CO₂ emissions by 581,830 tons annually, and reduction in evaporation of water by 17.28 million m³, proving robustness and environmental superiority. Egypt’s extensive solar and wind resources, alongside concerns like increasing energy demand and reliance on fossil fuels, accentuate the value of hybrid power systems^35,36,37.

The incorporation of cutting-edge energy management solutions can improve reliability and reduce costs while accelerating the country’s renewable energy revolution^38,39,40,41. In⁴², the Improved Archimedes optimization algorithm (IAOA) was utilized to optimize a HRES for Egypt’s Farafra microgrid, with 90% renewable share, cost of energy around 0.213 $/kWh, and high reliability. The results demonstrated IAOA to be more effective compared to other competing algorithms in developing cost-effective, reliable HRES configurations. In⁴³, a fuzzy-PID controller and an improved slime mould algorithm (MSMA) were utilized to maximize battery management in a PV–wind microgrid. The technique improved state-of-charge forecasting and battery efficiency. Simulation under different load conditions showed improved performance over PSO-based techniques, confirming its validity. In⁴⁴, a hybrid power system model in Abu-Monqar village, Egypt, comprising biomass, PV, WTs, and batteries was designed. Employing real meteorological data and Slime Mould Algorithm (SMA), the system performed the best at a net present cost of $3,476,372, energy cost of $0.118/kWh, and an LPSP of 0.032, which were higher compared to other algorithms both in cost and efficiency. Two FL based EMS and HHO were employed in⁴⁵ to optimize an Egyptian solar-powered seawater desalination plant. With dynamic energy pricing and uncertainty management, the EMS balanced battery storage, renewable energy, and the utility grid, generating profits of $10.28 and $10.11, respectively, with excellent dynamic performance. A profit-maximization model of virtual power plants (VPPs) in regulated markets was introduced in⁴⁶ with solar PV, combined cooling, heating, and power (CCHP), and storage systems. Optimized by GA, the model reduced CO₂ emissions by 47%, improved cost of energy by 36%, and improved energy efficiency. In⁴⁷, a model predictive controller and the P&O algorithm are used to reduce THD in PV systems. The technique has improved MPPT performance as well as significantly eliminates harmonic distortion over the incremental conductance technique. A modified slime mould algorithm (MSMA)-optimized fuzzy PID controller is proposed for effective battery management of a PV–wind microgrid in⁴⁸. The technique stabilizes voltage, manages uncertainty, and is more effective under varying loads in MATLAB simulations.

Research gaps and key contributions

Despite significant advances in energy management systems for hybrid renewable energy systems, fundamental challenges persist that limit their efficiency and broader applications, especially under dynamic and limited-resource conditions. Existing energy management paradigms have no means to cope with real-time variability and uncertainties of renewable energy systems, leading to inefficiencies and added operational expenses. The second significant shortcoming is the relative insensitivity and flexibility of existing hybrid energy management systems. The majority of systems do not have dynamic integration of different storage technologies such as battery storage, HESSs, and EVBs in a bid to offer optimal performance under different energy conditions. Besides, most of the current energy management systems don’t efficiently exploit synergies between different energy storage technologies, which results in higher degradation rates, power loss waste, and inept energy distribution policies. Remote and rural areas’ problems, particularly in nations like Egypt, are yet to be tackled on a large scale. These areas, with poor infrastructure and irregular power supply, require scalable, decentralized, and real-time energy solutions. Even though hybrid energy systems can provide such areas with energy, the current frameworks lack the flexibility to utilize their potential to the fullest. These loopholes indicate the necessity of revolutionary, real-time energy management solutions that ensure scalability, economic viability, and optimal utilization of energy in underdeveloped and geographically challenging areas. The contributions of this research can be encapsulated as follows:

• 1. MAS structures that facilitate successful coordination between hybrid storage devices and renewable energy sources as well as decentralized electricity access, particularly in rural areas, are recommended.

• 2. Integrates battery storage, HESS, and V2G-enabled EVBs with optimization of their coordination for enhanced efficiency, grid stability support, and reduced utilization of a single storage medium.

• 3. Tunes the operation of energy storage to lower costs, maximize the efficiency of energy utilization, and decrease operating costs by ensuring effective energy distribution policies, prevention of wasteful power curtailments, and extended system life.

• 4. Decreases costs and reliance on fossil fuels, making clean energy more accessible to remote regions by leveraging the use of decentralized energy solutions and advanced storage coordination.

• 5. Conducts rigorous simulation-based verification to validate the proposed MAS framework for scalability, flexibility, and economic feasibility, and demonstrates its usefulness in real-time applications across various hybrid energy systems under dynamic operation.

The remainder of this paper is organized as follows: In “Literature review” presents an overview of Egypt’s sustainable energy, commenting on Egypt’s energy challenges and why decentralized renewable energy systems are important. In “System architecture” explains the system architecture, commenting on the integration of renewable energy sources, hybrid storage systems, and the V2G concept. In “Multi-agent system coordination” section elaborates on MAS coordination, observing real-time decision-making and energy distribution optimization methods. In “Scenario analysis” section provides the scenario analysis, explaining the simulation environment, major assumptions, and the evaluation criteria. In “Results and discussion” section provides the findings and discussion, comparing the performance of the proposed strategy to conventional energy management methods. In “Conclusions” section summarizes the paper by outlining the primary findings and observing potential challenges.

System architecture

The architecture of MAS is tuned for efficient and dynamic energy control of hybrid systems. MAS consists of task-specialized agents that are responsible for carrying out a specific function in the energy management system. The agents operate independently but still communicate with a centralized multi-agent controller (MAC) to accomplish maximum energy dispatch in real-time as depicted in Fig. 1. These key agents are the Battery Agent, Hydrogen Storage Agent, EVB Agent, Renewable Energy Agent, and Load Agent. The Battery Agent takes charge of charging and discharging batteries. The agent maximizes the utilization to enhance the battery’s lifespan along with minimizing operational cost. It also conveys real-time SOC and efficiency levels to the MAC for making informed decisions. The Hydrogen Storage Agent manages hydrogen production and consumption in the fuel cell system. It manages long-term energy storage as a reserve during low renewable energy production hours. It also provides hydrogen levels and operation efficiency to the MAC for easy incorporation into energy dispatch. The EVB Agent manages the EV battery interaction with the grid via V2G technology. It gives flexible storage and peak load demand response. It also includes SOC and availability to be provided to the MAC for effective energy management. The Renewable Energy Agent gets current energy output from PV and wind sources. It regulates renewable inputs based on environmental conditions to maintain maximum efficiency. The Load Agent gets current energy consumption from loads. It provides priority to loads of high criticality during energy shortage and communicates load profiles to the MAC for ensuring guarantee of distribution of energy in accordance with system priorities. The MAC is the MAS decision-making hub, collecting information from all agents and calculating the optimal energy dispatch strategy dynamically. It collects information on energy availability, storage capacity, and load demand from all the agents to offer synchronization. It attains optimal short-term operational efficiency at the expense of long-term sustainability through the GTO algorithm. It reduces operational expenses, LPSP, and excess power. It provides optimized energy schedules to independent agents to implement in real time.

Fig. 1

System architecture for multi-agent energy management.

The MAS follows a structured work flow for energy management. The renewable energy generation data and load demand profiles are obtained in real time, whereas storage system states, i.e., SOC and efficiency, are conveyed by individual agents. The MAC executes the GTO algorithm to calculate the optimal energy dispatch strategy with an objective to reduce the cost, reduce LPSP, and avoid excessive power wastage. Optimized dispatch schedules are communicated to agents, who modify their operations accordingly. Agents provide feedback on execution results, enabling the MAC to modify its optimization policy based on new information. The system architecture as a whole can be described in four main steps, as illustrated in Fig. 2. The first phase, MAS Framework, includes energy demand profiling, resource availability analysis, and limitation of the system and capability of agents analysis. Scalability alternatives, performance criteria, and economic and environmental aspects are also assessed at this point to determine the system’s intrinsic requirements.

Fig. 2

Proposed multi-agent-based energy management methodology.

The second stage, MAC, is aimed at energy management optimization by addressing some of the most critical operation and environmental problems. These include minimizing operational costs, mitigating capacity bottlenecks, and decreasing greenhouse gas emissions. Optimization is subject to several system constraints to achieve energy efficiency and reliability. These include maintaining power balance, managing the EV and battery SOC, regulating the amount of stored hydrogen, and guaranteeing the maximum proportion of renewables in the energy mix. A key activity in this step is the definition and evaluation of configuration scenarios that contrast energy system configurations for uncovering the most promising solutions. The configurations of the energy system considered are the integration of WTs, PV systems, BESS, FCs, and EVs. Each of the alternatives is comprehensively analyzed for its capability to offer energy reliability, affordability, and environmental sustainability. The outcome of this phase serves as a foundation for further analysis in the next energy management workflow stage.

The third step, energy management workflow, is a general review of the energy system on technical, economic, and environmental bases. This is where the optimum strategies for energy efficiency, robustness of the system, and sustainability are arrived at. On the technical basis, the energy system is looked at in terms of how suitably it is capable of providing grid stability and efficiency during changing operating conditions. Finally, in the fourth stage, power quality analysis, optimized backup system configurations and enhanced energy management strategies are implemented. GTO and other similar algorithms are used to give effective energy supply and system stability. The critical system components like BESS, EVB, and HESSs function in order to achieve the primary objectives: finding optimal backup system settings, maximizing utilization of resources, maximizing the utilization of renewables, and minimizing energy wastage.

Multi-agent system coordination

MAS coordination refers to persistent communication and collaboration between agents within the MAS in order to achieve shared objectives in hybrid energy management. The structure of MAS provides unified functioning of the agents through real-time sharing of information, reacting to altered situations, and maximizing decision-making. This coordination facilitates the proper regulation of the renewable energy generation, storage devices, and load so that the performance and reliability of hybrid energy systems are enhanced.

To enable seamless real-time coordination among diverse energy storage systems including EVBs, HESSs, and BESSs, and RESs such as PV and WTs, the proposed MAS framework is built on a hierarchical control structure integrated with a high-speed field-programmable gate array (FPGA) based MAC. Every energy system has a local sensing, control, and communication-intelligent agent responsible for managing it. The agents monitor key parameters such as SOC, charge/discharge efficiency, availability for dispatch, and health metrics. The agents report in real time to the MAC via an embedded low-latency communication bus with minimal data packets designed for high-frequency updates. The MAC operates on the data concurrently with its FPGA logic and can react to grid condition changes in under a sub-millisecond.

The MAC is supplemented by the GTO, which is continuously monitoring the system state and computing optimal charging and discharging schedules for each storage unit. The predicted and actual load demand, real-time available renewable sources, operational cost, SOC of the storage, and unit availability are considered by the optimization algorithm in order to make efficient and balanced energy management. For example, when there is excess renewable generation, the power can be shifted to BESS, EVBs, or hydrogen generation based on the actual system capacity, cost, and efficiency at any given time. During periods of power deficit, the GTO dynamically determines the optimal order of discharging among available storages in an effort to maintain system reliability and minimize cost, independent of a rigid dispatching sequence. These decisions are executed immediately by the FPGA-based MAC for high-speed energy balancing with minimal delay.

If real-time fluctuations in load or renewable output are detected, the GTO dynamically redistributes the optimal dispatch plan in real time based on updated agent inputs. Demand management procedures to shed unessential loads during periods of restricted supply can likewise be implemented by the MAC. The FPGA-based implementation adds increased system robustness by incorporating self-monitoring, redundancy, and watchdog timers in the design for safe operation despite partial system malfunction. This is a scalable and modular architecture that allows for new energy units to be added without interrupting ongoing operations. The integration of high-speed FPGA control, real-time agent communication, and dynamic optimization provides a robust and intelligent solution for controlling hybrid energy systems in remote or resource-poor environments.

The MAS applies advanced optimization algorithms to enable cooperative decision-making for the agents. The agents advance their contribution to the aggregate-level decision in that they provide the localized information and knowledge. MAC pools the same and uses for policy dispatch optimization. For example, MAC can prepay the storage facility from the periods of renewable energy overproduction to adaptively adjusting schedule loads in real time based on the amount of available energy. Agents in the MAS are programmed to adapt dynamically based on changing conditions. For instance, the renewable energy agent varies operation based on solar irradiance or wind speed fluctuations, while the storage agent scales its charge–discharge cycle with load demand changes. Dynamic role adaptation therefore ensures that the MAS will be adaptive and responsive to intrinsic and extrinsic properties affecting the hybrid energy system.

Optimization framework for MAS coordination

The MAS coordination optimization model endeavors to ensure seamless and efficient energy management by striking a balance between crucial goals such as cost minimization, renewable energy maximization, and system reliability. Its basis is a set of well-defined objective functions that guide decision-making and carefully thought-out constraints that impose the feasibility, efficiency, and stability of the system.

Objective function

The model of optimization attempts to address critical issues in hybrid energy systems with two main criteria, which are aimed at enhancing the overall performance, sustainability, and flexibility of energy systems. The criteria are crucial in making sure that hybrid energy systems meet demands of existing energy infrastructure without addressing environmental problems and economic viability. The framework is structured around two fundamental optimization goals: cost minimization, which ensures economic feasibility by optimizing energy resource allocation and reducing operational expenses, and reliability maximization, which guarantees system stability and resilience against fluctuations in energy supply and demand.

Cost minimization

The economic efficiency of hybrid energy systems is key to their sustainability in the long run, particularly in resource-constrained and isolated regions. Energy systems in these regions typically do not carry gigantic fiscal burdens with them in terms of huge upfront costs, complexity of operation, and limited access to infrastructure. The economic efficiency standard tries to give a system that is financially sustainable and feasible, particularly in underdeveloped or underserved regions, where energy affordability and reliability are crucial to maintaining economic growth and standard of living.

• Operational and maintenance costs These include cost considerations related to day-to-day maintenance and operation of energy generation and storage devices. The aim is to maintain such costs as minimal as possible without compromising system performance in order to attain financial sustainability. The formula can be employed to estimate the total overall operating and maintenance (O&M) costs of the proposed system⁴⁹.

  $$C_{O\& M} = \left( {C_{O\& M}^{PV} + C_{O\& M}^{WECS} + C_{O\& M}^{BESS} + C_{O\& M}^{EV} } \right) \times \delta$$

  (1)
  $$\delta = \frac{{\left( {1 + k} \right)^{\tau } - 1 }}{{k \left( {1 + k} \right)^{\tau } }}$$

  (2)

• Excess power utilization Excess power occurs when more renewable energy is generated than demand and cannot be consumed. Lowering excess power ensures the renewable energy is utilized to its optimum, with minimal wastage and system inefficiency. It also minimizes fossil fuel consumption by maximizing the use of renewables. The amount of system surplus power provided can be calculated by using this formula^50,51.

  $$P_{surplus} = \mathop \sum \limits_{t = 1}^{N} \frac{{P_{dummy} \left( t \right)}}{{P_{L} \left( t \right)}}$$

  (3)

Reliability maximization

Reliability is a critical aspect of energy systems, particularly for regions where energy access is intermittent or where renewable resources are variable. This criterion emphasizes ensuring a consistent and dependable energy supply. LPSP measures the probability of the system being unable to satisfy energy demand at any moment. A lower LPSP indicates a more dependable system capable of managing variations in renewable energy generation and changing demand patterns. This formula can be used to calculate the LPSP of the proposed system⁵²:

$$LPSP \left( \% \right) = \frac{{\mathop \sum \nolimits_{t = 1}^{T} P_{deficit } \left( t \right)}}{{\mathop \sum \nolimits_{t = 1}^{T} P_{L} \left( t \right)}}$$

(4)

where ${\uptau }$ represents microgrid life period, k is annual interest rate (8%), ${\text{C}}_{{{\text{O}}\& {\text{M}}}}^{{{\text{PV}}}}$,${\text{C}}_{{{\text{O}}\& {\text{M}}}}^{{{\text{WT}}}}$, ${\text{C}}_{{{\text{O}}\& {\text{M}}}}^{{{\text{HESS}}}} ,{\text{ C}}_{{{\text{O}}\& {\text{M}}}}^{{{\text{BESS}}}}$, and ${\text{C}}_{{{\text{O}}\& {\text{M}}}}^{{{\text{EV}}}}$ represent O&M costs of the PV, WT, HESS, BESS, and EV, respectively, ${\text{P}}_{{{\text{dummy}}}}$ symbolizes power lost in the dummy load, and $P_{L} \left( t \right)$ represents load power.

Agent-specific constraints

To ensure operational feasibility and system efficiency, the optimization framework adheres to a set of critical constraints.

Surplus power constraints

Effective handling of surplus energy is necessary to optimize the cost of the overall system and achieve effective use of resources. The constraint for minimizing the operating cost by reducing wasted energy in a dummy load is given as:

$$P_{surplus} \le P_{surplus}^{max}$$

(5)

Reliability constraints

One of the most significant factors for successful system operation is the reliability of an energy system. To quantify reliability, the LPSP indicator is employed in this research, which quantifies the probability that the system will not be capable of meeting energy demand. Following is how this constraint is implemented⁵³:

$$LPSP \le \beta$$

(6)

Power balance constraints

Balance of power constraint is applied to make generated power always equal to the energy demand. It is a requirement for system stability and energy deficit prevention. This constraint can be expressed mathematically as follows⁵⁴:

$$P_{L} \left( t \right) = P_{Gen} \left( t \right),\quad \forall t \in \tau$$

(7)

$$P_{PV} \left( t \right) + P_{WT} \left( t \right) + P_{FCs} \left( t \right) + k \times P_{BESS} \left( t \right) + k \times P_{EV} \left( t \right) = P_{L} \left( t \right)$$

(8)

$${\text{k}} = \left\{ {\begin{array}{*{20}l} {k = 1} \hfill & {\quad discharge} \hfill \\ {k = 0} \hfill & {\quad charge\;stays\;constant} \hfill \\ {k = - 1} \hfill & {\quad charge} \hfill \\ \end{array} } \right.$$

(9)

BESS charging and discharging constraints

To ensure BESS longevity and performance, redundant charge or discharge cycles should be avoided. SOC of the BESS is restricted within a range of safe operation, usually between 20 and 80%. BESS charging below 20% leads to deep discharge, which degrades the battery health in the long term. In the same context, overcharging above 80% results in a shorter lifespan for the battery. This limitation can be expressed mathematically as follows:

$$20\% \le {\text{SOC}}_{BESS} \le 80\%$$

(10)

EV charging and discharging constraints

EVs play a dual role in hybrid energy systems as they provide mobility and at the same time act as an energy storage device using the V2G technology. Both roles, however, demand maintaining the EV battery SOC in its optimal range. SOC limitations ensure that EVs contain enough charge for daily commuting needs, which ranges from 70 to 80%. This constraint can be mathematically expressed as follows:

$$P_{{r_{EV,max} }} \ge P_{{r_{dis} }} , P_{{r_{ch} }} \ge P_{{r_{EV,min} }}$$

(11)

$$70\% \le {\text{SOC}}_{EV} \le 80\%$$

(12)

Optimization and monitoring of efficiency are also very important in achieving maximum range and overall performance of EVs and minimizing energy wastage. One of the most significant steps of this process is testing the charging and discharging efficiency of the EV battery, and this can be precisely measured by the following equation:

$$\eta_{EV} = \frac{{C_{EV,Pr} \times R_{EV} }}{{D_{EV}^{Home} \times CAP_{EV}^{total} \times C_{pr} \left( t \right)}}$$

(13)

The operational status of the EV is represented as follows:

$$LOC_{EV} = \left\{ {\begin{array}{*{20}l} 1 \hfill & {\quad EV\; available} \hfill \\ 0 \hfill & {\quad EV \;not\;available} \hfill \\ \end{array} } \right.$$

(14)

where ${\text{C}}_{{{\text{EV}},{\text{Pr}}}}$ refers to the cost incurred due to power losses, ${\text{R}}_{{{\text{EV}}}}$ denotes the range of the EV, ${\text{D}}_{{{\text{EV}}}}^{{{\text{Home}}}}$ represents the distance between the EV and home, ${\text{CAP}}_{{{\text{EV}}}}^{{{\text{total}}}}$ is the total battery capacity of the EV, ${\text{C}}_{{{\text{pr}}}} \left( {\text{t}} \right)$ indicates the EV power price, $P_{{r_{EV,min} }}$ and $P_{{r_{EV,max} }}$ define the minimum and maximum operational power of the EV, respectively. Additionally, ${\text{SOC}}_{BESS}$ signifies the state of charge of the BESS, while $P_{PV} \left( t \right)$, $P_{WT} \left( t \right),{ }P_{FCs} \left( t \right),{ }P_{BESS} \left( t \right),{\text{ and }}P_{EV} \left( t \right)$ represent the power generated by PV, WT, FCs, BESS, and EV, respectively. Lastly, $LOC_{EV}$ refers to the location of the EV.

GTO-enhanced MAS optimization

Artificial gorilla troops optimizer

The inclusion of the GTO in the MAS framework significantly enhances its optimization for hybrid power systems on various renewable energy resources. GTO is inspired by the cooperation and foraging of gorilla groups in which each member of the group collectively works towards achieving a common target, such as food or protection. This group behavior is best suited to address difficult optimization problems within energy systems comprising a vast number of agents—including renewable resources, energy storage devices, and loads—that coordinate among themselves to achieve optimal performance. With GTO and MAS integrated, the framework can effectively handle tasks such as dynamic resource allocation, energy dispatching, and balancing variable energy loads. The GTO is a new metaheuristic (MH) algorithm introduced in 2021 by⁵⁵. It mimics the social intelligence and dynamic behavior of gorillas in the wild. The GTO algorithm has two major phases: exploration and exploitation, both of which have a fundamental role to play in system optimization and decision-making.

Exploration phase

The exploration step employs three mechanisms to enhance the ability of the algorithm to fully explore the search space: migrating to an unseen area, approaching other gorillas, and moving to a familiar area. Such mechanisms enable the algorithm to balance exploration and avoid getting trapped in local optima. The position of the gorilla in iteration (t + 1) is determined by the following equations:

$$GX\left( {t + 1} \right) = \left( {UB - LB} \right) \times r_{1} + LB, \quad rand < p,$$

(15)

$$GX\left( {t + 1} \right) = \left( {r_{2} - C} \right) \times X_{r} \left( t \right) + L \times H, \quad rand \ge 0.5,$$

(16)

$$GX\left( {t + 1} \right) = X \left( i \right) - L \times \left( {L \times \left( {X\left( t \right) - GX_{r} \left( t \right)} \right) + r_{3} \times \left( {X\left( t \right) - GX_{r} \left( t \right)} \right)} \right)\quad rand < 0.5.$$

(17)

The supporting parameters are calculated as follows:

$$C = F \times \left( {1 - \frac{It}{{MaxIt}}} \right),$$

(18)

$$F = cos \left( {2 \times r_{4} } \right) + 1$$

(19)

$$L = C \times l$$

(20)

$$H = Z \times X\left( t \right)$$

(21)

$$Z = \left[ { - C, C} \right]$$

(22)

Exploitation phase

The exploitation phase enhances the search by aiming at areas close to good-quality solutions. It imitates such behaviors as silverback imitation (leader) and dominance competition with the aim of enhancing solution quality.

Following the silverback

The silverback represents the best solution found in each iteration. Gorillas adjust their positions based on the silverback’s guidance, simulated by:

$$GX\left( {t + 1} \right) = L \times M \times \left( {X\left( t \right) - X_{silverback} } \right) + X\left( t \right)$$

(23)

where,

$$M = \left( {\left| {\frac{1}{N}\mathop \sum \limits_{i = 1}^{N} GX_{i} \left( t \right)} \right|^{g} } \right)^{\frac{1}{g}}$$

(24)

$$g = 2^{L}$$

(25)

Competition for adult females

This strategy introduces competition among gorillas, driving them to explore other potential solutions. It is mathematically represented as:

$$GX\left( i \right) = X_{silverback} - \left( {X_{silverback} \times Q - X\left( t \right) \times Q} \right) \times A$$

(26)

where:

$$Q = 2 \times r_{5} - 1$$

(27)

$$A = \beta \times E$$

(28)

$$E = \left\{ {\begin{array}{*{20}l} {N_{1} ,} \hfill & {\quad r \ge 0.5,} \hfill \\ {N_{2} ,} \hfill & {\quad r < 0.5,} \hfill \\ \end{array} } \right.$$

(29)

The two method exploitations are chosen based on the value of C. Whenever C is equal or greater than W, the silverback method is pursued in which positions are filled based on the direction of the silverback. Whenever at any point C is less than W, the algorithm is using the method of competition for adult females using competitive interactions for an effort towards further better solutions.

Role of GTO in MAS decision-making

In the proposed system, the GTO is employed as a crucial component in the MAS decision-making process. GTO decides on the optimal charging and discharging schedules between BESS, HESS, and EVBs, based on real-time information including load demand, renewable energy availability, SOC of storage units, and operational constraints. GTO balances MAS coordination through renewable energy real-time distribution among varying storage devices in a nonsequential manner, illustrated in Fig. 3. In conditions where there’s redundant renewable power, the system will not remain steadfast to some priority charging rule. Instead, it uses an analysis of near-instant system health, such as state-of-charge status, availability of different units, and operational limits to determine which to charge most suited. Depending on these conditions, energy may be directed to BESS, EVBs, or HESS. If all storage options are fully charged, the remaining surplus is safely diverted to a dummy load to maintain system stability.

Fig. 3

GTO-enhanced MAS optimization for efficient renewable energy distribution and storage.

When there is an energy shortage period, the system also suppresses a pre-set discharge priority usage. The GTO algorithm is used to take into account the cost-effectiveness, responsiveness, and availability of every storage unit and thus select the optimal source to supply the required energy. This dynamic and adaptive decision-making enables the MAS to control responses to varying load demands and levels of renewable generation effectively while guaranteeing reliability as well as cost-effectiveness in energy management. The internal process of GTO is operated in an orderly iterative process as in Fig. 3. It initiates by specifying a population of candidate solutions, and in each solution, there would be a possible dispatch plan. Populations, iterations, and variable limits are established as parameters. The candidate solutions are then ranked using an operational cost, energy efficiency, and constraint satisfaction fitness function. The algorithm continues in two major phases: exploration and exploitation. During the exploration phase, candidate solutions are explored extensively in the solution space to avoid becoming stuck in local optima. During the exploitation phase, candidate solutions are refined to enhance precision. The algorithm toggles between these stages based on control parameters in order to balance global exploration and local refinement.

One of the greatest advantages of GTO is the capability of responding dynamically to real-time system fluctuations. On receiving new information, i.e., load demand, renewable generation, or storage level changes, the optimizer re-calculates and updates the energy dispatch plan. The ability to respond dynamically allows the MAS to guarantee optimal operation even under rapidly changing conditions. The flexibility of GTO is particularly valuable in systems that have a high level of penetration of intermittent renewables or in environments of high uncertainty. GTO is put in comparison to some of the well-known existing optimization algorithms including Particle Swarm Optimization (PSO), Gray Wolf Optimization (GWO), Whale Optimization Algorithm (WOA), Moth–Flame Optimization (MFO), Multi-Verse Optimizer (MVO), Sine–Cosine Algorithm (SCA), and Gravitational Search Algorithm (GSA). According to literature and early simulations conducted as a part of this work, GTO was found to have higher convergence rate, solution precision, and global exploration capability. Its robustness against premature convergence makes it more suitable to address complicated, nonlinear, and constrained optimization problems that are commonly found in multi-agent energy systems.

Overall, the addition of GTO to the MAS significantly enhances the hybrid system’s ability to operate in an efficient, cost-saving, and dependable manner. With the facilitation of real-time optimal energy dispatch, GTO improves energy consumption, reduces operational costs, and increases the adaptability of system response. These advantages render the proposed hybrid energy system ideal for application in rural or remote areas, where robust and adaptive energy management is paramount.

Scenario analysis

The proposed MAS framework was simulated under four scenarios with varying combinations of renewable energy sources and hybrid storage devices. The purpose of the analysis was to ascertain the applicability, affordability, and consistency of the framework under varying circumstances, particularly those environments characterized by scarce resources. Each scenario was designed to yield specific strengths and weaknesses and offer a comprehensive description of the system’s operation.

Scenario 1: PV–wind–battery

The BESS is employed as the principal storage system to serve as a buffer for variability of renewable energy supply over short-term scales. Surplus solar or wind generation is reserved in batteries whenever the generated energy exceeds demand. In contrast, during low renewable energy generation, the load demand is fulfilled by the batteries and system stability is ensured by discharging it. The arrangement is unique because it is simple and cost-effective, hence very appropriate for sites with relatively moderate and constant renewable energy uncertainty. Batteries use is promising quick reaction to alter demand and supply of energy in that it promises efficient and effective energy management. However, the absence of long-duration energy storage technologies, such as hydrogen storage, may limit its capacity to address extended periods of low renewable power generation. It is for this reason that, in the scenario, there is a focus on short-term balancing and not energy storage over the long term. Scenario 1 is most appropriate for contexts in which patterns of renewable energy generation are fairly stable and short-term energy fluctuations are the chief concern. Its minimal design and battery reliance on balancing the power make it a prudent option for low-infrastructure remote areas with moderate energy needs.

Scenario 2: PV–wind–battery–EVB

The WTs and PV panels are paired with both EVBs and BESS. This arrangement builds upon the capability of scenario 1 by introducing V2G technology, wherein EVBs are utilized as a source of energy as well as an energy storage device. The utilization of EVBs introduces flexibility in the system so that surplus renewable power is stored in EVBs and fed to the grid during peak demand or when there is shortage of energy. EVBs introduce a further dynamic level of demand response to stabilize moment-to-moment load fluctuation. This flexibility not only reduces wastage of surplus power but also brings about the activation of EV owners’ engagement in trading energy. The V2G feature enables two-way energy transfer, and EVs can serve as rolling batteries that can balance the grid in peak demand periods.

This configuration is particularly beneficial in high EV take-up urban or semi-urban regions. With an opportunity to leverage the installed EV base, the system can offer improved energy usage and cost savings without making huge new investments. Its efficiency, however, depends on coordination and availability of EVs and thus is not best suited to low EV penetration regions. Scenario 2 demonstrates the ability to enhance the utilization of both system reliability and renewable energy through the strategic use of EVBs. This does not only ensure the optimum utilization of renewable energy but also benefits the overall sustainability and resilience of the hybrid energy system.

Scenario 3: PV–wind–battery–fuel cells

PV panels and WTs are integrated with both BESS and hydrogen-based FCs. The system is designed to address both short-term and long-term energy storage needs by leveraging the complementary strengths of batteries and FCs. Batteries provide fast response to short-term fluctuation in energy supply and demand, while FCs are utilized for long-term energy storage and supply energy stably over extended periods of low renewable energy output.

Excess renewable energy is employed to produce hydrogen by electrolysis, and it is stored for future use. In the event of excess energy requirements or deficient renewable energy generation, stored hydrogen is passed through FCs to once more generate electricity. The process ensures a reliable source of energy and reduces reliance on fossil fuels. The two-layered storage system optimizes system stability as a whole and optimizes utilization of renewable energy sources. This scenario is particularly suitable in regions of high variability in renewable sources and long periods of low production, such as those with seasonal variations in solar and wind resources. While the incorporation of FCs increases the initial infrastructure costs, the long-term benefits of enhanced reliability and reduced cost of operation weigh higher over these difficulties. Furthermore, the hydrogen storage center provides a scalable means of future energy needs. Scenario 3 demonstrates the ability to transition short-term and long-term energy shortages effectively, yielding a robust and stable hybrid energy system. Its versatility and emphasis on long-term energy security enable it to be an appropriate design for remote or resource-poor locations with highly variable renewable energy conditions.

Scenario 4: PV–wind–battery–fuel cells–EVB

PV panels and WTs are paired with three types of storage systems: BESS, hydrogen-based FCs, and EVBs. The hybrid setup makes use of the advantage of each storage system to generate a powerful and highly versatile hybrid energy system capable of meeting short-term, long-term, and dynamic energy needs. The BESS manages short-term variations in energy by reacting promptly to generation from renewable energy and changes in loads. Hydrogen FCs provide long-term storage using excess renewable energy to produce hydrogen through electrolysis and the reverse process to create electricity when needed. EVBs also provide an additional level of flexibility in load management and energy storage. The EVBs can sell back to the grid during peak load and purchase during peak renewable generation, allowing for the best use of energy and grid stability. The system achieves maximum reliability by leveraging the complementary advantages of the three storage systems. The use of EVBs includes dynamic participation by the EV owners, while the hydrogen FCs offer guarantees of availability of energy during prolonged low phases of renewable energy generation. It attains minimized power loss, reduced fossil energy dependence, and maximum utilization of accessible renewable energy. Scenario 4 performs best for complex energy systems in constraint-rich or high-variance environments. It has the highest scales of scalability, renewable utilization, and reliability among all the scenarios. Though the initial investment costs will be higher due to the advanced infrastructural requirements, the long-term benefits of cost-effectiveness, energy independence, and environmental sustainability make this system an ideal design for huge hybrid energy systems.

The flowchart in Fig. 4 outlines how energy management is optimized in each scenario. It illustrates the decision-making for energy dispatch, storage management, and backup power engagement against the variation in energy supply and demand. The flowchart highlights how different combinations of renewable sources, storage, and backup systems are managed to offer optimal system performance in various operating conditions.

Fig. 4

Energy management process across different scenarios for optimizing dispatch, storage, and backup systems.

Results and discussion

The proposed MAS architecture is envisioned to simplify energy management in hybrid microgrids for addressing the unique challenges of remote Egyptian regions. These regions often experience inadequate infrastructure, unstable grid connection, and fluctuating energy demand, which demand a smart and adaptive solution to integrating renewable energy sources and hybrid storage technologies. The MAS framework enhances the efficiency of energy dispatch by using real-time decision-making and advanced optimization techniques, which substitutes traditional forecasting-based approaches. To compare the performance and efficiency of the proposed system, four scenarios were assumed, each with a different integration mix of solar PV, WTs, BESS, EVBs, and hydrogen-based FCs. These installations were put through real-time operational constraints to determine their feasibility in providing cost-effective, sustainable, and reliable energy solutions for off-grid settlements.

The results are focused on some of the key performance indicators that determine the effectiveness of the proposed MAS architecture in coordinating hybrid energy systems. LPSP is considered as a key system reliability indicator expressing the probability of power shortage in remote areas where secure energy supply is essential. O&M Costs are examined to establish the economic feasibility of the system so that the proposed energy management strategy remains cost-effective and sustainable throughout its long-term implementation. Excess energy use is also examined to see how efficient the system is in avoiding wastage of energy compared to optimization of excess renewable power storing and dispatch. Finally, the consumption of renewable energy is considered to see how much the MAS maximizes the contribution from wind and solar power and restricts the consumption of fossil fuels, thereby enhancing the environmental sustainability. These performance indicators provide a general assessment of the ability of the MAS system to improve the reliability of energy, reduce costs, and facilitate efficient energy management in remote Egyptian villages.

A key aspect of the MAS framework is its pairing with the GTO, which is an advanced metaheuristic algorithm with the ability to adapt energy dispatch strategies dynamically to real-time changes in energy production and consumption. This enables optimal resource allocation while maintaining system stability even under conditions of high energy variability. The performance of this optimization method was tested under the different energy storage scenarios, providing an insight into the best performing configuration for economically viable and large-scale microgrid deployment at remote areas. The following sections provide comparative performance comparison of each one of the scenarios, showing the impact of the MAS on energy cost minimization, power reliability, and sustainability. The findings confirm the potential of this smart energy management system to transform the performance of remote microgrids, offering a way towards affordable, reliable, and renewable-powered energy prospects for Egyptian off-grid communities.

Scenario 1

Scenario 1 integrates WTs, BESS, and solar photovoltaic panels into the MAS structure, which is optimized by the GTO. The MAS-GTO ensures real-time energy management, with dynamic balancing between demand and supply to achieve optimum power distribution in Egyptian off-grid sites. The configuration is aimed at optimizing the utilization of renewable energy and minimizing the reliance on fossil fuel, hence being most appropriate for off-grid communities with limited infrastructure and unreliable grid connection. Figure 5 shows the power distribution over a 24-h period, indicating the role of solar PV, WTs, and BESS in fulfilling the variable energy demand. The load profile is highly variable, with demand ranging from 9640 to 24,070 W. The peak demand occurs at 9:00 PM, followed by a gradual decrease during early morning hours. Wind power is available throughout the day but fluctuates between 0 and 10,856 W. Wind power is maximum at 2:00 AM, which is consistent with demand and reduces battery storage reliance. Solar PV generation commences at 6:00 AM and increases gradually to 10:00 AM to 2:00 PM. It is the cheapest and most energy-efficient time as there is spare power to recharge the BESS. The BESS dynamically follows energy availability by recharging during the middle of the day when solar generation is highest and discharging during nighttime and evening hours to compensate for low renewable generation. Most of the charging is performed during midday hours, and most of the discharge is performed between 6:00 PM and 12:00 AM. With these optimizations, the system’s reliance on BESS alone becomes an issue during nighttime hours. At 6:00 PM, the system is relying on stored energy, and between 11:00 PM and 12:00 AM, there is greater risk of depletion, which is a potential energy reliability issue. Figure 6 plots the LPSP for 24 h, a critical measurement of system dependability. LPSP is near zero during the day, indicating system reliability is good when solar PV and wind power are providing. After sunset, LPSP starts to increase with intensified battery discharge during the peak between 11:00 PM and 12:00 AM. The peak LPSP spike happens late in the night. Despite that, it is still within the acceptable limit of 0 to 0.05. This indicates that despite the system being reliable within given limits, the lack of other backup storage or demand-side management measures could still result in possible power shortages.

Fig. 5

Energy generation, storage, and demand profile for scenario 1 optimized using GTO.

Fig. 6

System reliability assessment for scenario 1.

The economic operation of the system is to be blamed for determining its viability. The sustainable operating cost at the end of the day is $10,785.60, the combined cost of maintaining energy storage and generation. Scenario 1 is more energy-independent and less expensive but counts on BESS to maintain reliability at night and, therefore, increases the operation cost. If wind energy generation drops sharply at night, the risk of power supply failure rises, and strategic increase of storage or integration of hybrid energy storage is needed to enhance overall resilience. Scenario 1 provides a highly viable and sustainable option for Egyptian off-grid communities with cheap and reliable daytime energy supply. Ideal use of wind energy, solar PV, and battery storage in the system maintains LPSP levels at their lowest, cutting shortages during the day considerably. Higher consumption of stored power during the night and the corresponding rise in LPSP levels requires tactical addition of storage units and standby supplies. By integrating leading-edge energy storage solutions like hydrogen FCs or EVBs and enforcing best-practice demand-side management techniques, Scenario 1 can be upgraded to a more scalable, robust, and completely self-sustaining microgrid solution. This will provide round-the-clock, dramatically enhanced energy security as well as economic viability to remote areas in Egypt. The complementarity between solar PV, wind, and BESS leads to low LPSP values and energy dispatch optimization.

Scenario 2

Scenario 2 integrates solar photovoltaic panels, WTs, a BESS, and EVBs in the MAS optimized by the GTO to maximize energy management in Egypt’s rural areas. The system enhances energy flexibility and reliability by leveraging EVBs as an additional energy storage component. The integration of EVBs provides dynamic distribution of stored energy to supplement the BESS and avoid the risk of power shortages, particularly during nighttime. Optimization of real-time dispatch of energy through MAS provides efficient use of renewable sources of energy, equilibrium of demand and supply, and reduced reliance on conventional backup sources.

Figure 7 depicts the 24-h power sharing, illustrating that solar photovoltaic system and WTs are the majority sources of energy while the buffering is provided by the BESS and EVBs to ensure smooth availability of power. The BESS charges in an appropriate manner during the period when the maximum solar photovoltaic generation occurs by absorbing surplus energy to prevent curtailment and storing the energy for future utilization. This enables surplus renewable energy to be efficiently stored rather than wasted. System discharge takes place when demand exceeds supply, precisely after 6:00 PM when solar photovoltaic production declines. At night, as solar production drops, the BESS becomes the primary stabilizing component, seamlessly compensating for the deficit in the availability of renewable energy. Its peak release is between 8:00 PM and 12:00 AM, when both wind power deviation and energy demand require additional backup. Without the BESS, the system would have significant power shortages during these late hours, increasing the reliance on alternative solutions.

Fig. 7

Energy generation, storage, and demand profile for scenario 2 optimized using GTO.

EV Batteries provide a dynamic layer of energy storage, charging and discharging adaptively based on real-time variability in supply and demand. Unlike the fixed BESS, EV Batteries charge and discharge in a flexible manner in accordance with user patterns, maintaining a 70–80% state of charge to give confidence to vehicle owners that there is stored energy when needed. When renewable generation exceeds demand, EV Batteries absorb excess energy to prevent curtailment. At night, they release supplemental power, decreasing stress on the BESS and minimizing the risk of complete depletion. In scenario 2, EV Batteries contribute actively to load balancing during the peak hours of demand, with their discharge dynamically regulated to balance the system. Interestingly, EV Batteries provide 44.44 W at 8:00 PM, 3093.58 W at 9:00 PM. The concurrent operation of BESS and EV Batteries enhances energy reliability through maintaining stability, minimizing interruptions, and making power available round the clock. This integration has greater flexibility in terms of power allocation than scenario 1 and yields a more stable supply of energy during night.

Figure 8 illustrates the 24-h LPSP, one of the key system reliability metrics. LPSP is near zero for a large part of the day because of successful operation under the availability of solar and wind power. Power deficiency probability increases after 8:00 PM with the highest values between 10:00 PM and 12:00 AM. Despite this rise, LPSP remains within acceptable ranges (0–0.05), indicating overall system reliability. Scenario 2’s economic performance is evaluated by its cumulative operating cost, which totals $10,709.10 at the end of the 24-h period. The inclusion of EV Batteries in conjunction with BESS leads to more efficient use of energy, distributing stored energy effectively and reducing excessive high-cost discharges. Economic efficiency of scenario 2 is quite evident in the evening time period since EV Batteries reduce the dependency on BESS; hence, it results in stabilized and optimized energy dispatch plan. The cost, though equal in total value with that of scenario 1, has the addition of EV Batteries providing system flexibility and enhancing its economic efficiency in fulfilling energy requirements at late hours.

Fig. 8

System reliability assessment for scenario 2.

Scenario 3

Scenario 3 integrates PV panels, WTs, a BESS, and FCs into the MAS with the GTO optimization to enhance improved energy management in remote Egyptian areas. The configuration focuses on improving energy reliability, particularly during nighttime, by incorporating FCs as an additional source of power. FC integration significantly enhances the stability of the system by including a backup power source during maximum demand to ensure minimal overall reliance on BESS alone. Maximum utilization of the renewables is accomplished by MAS by optimal real-time dispatch of power with dynamic supply and demand control.

Figure 9 shows the power contribution pattern across the 24-h period by PV panels, WTs, BESS, and FCs. Load demand is different throughout the day with the highest demand at 9:00 PM and a maximum of 24,070 W. BESS adapts dynamically with energy supply by charging during high solar hours when there is surplus solar energy and discharging to meet the demand in the evening and night. The BESS is discharging up to 7838.65 W at 1:00 AM, utilizing the stored energy to meet demand. BESS discharge occurs throughout the day, with maximum discharges of 6930.6 W at 8:00 AM, 5803.4 W at 10:00 AM, and 4841.54 W at 1:00 PM, indicating its use in smoothing out intermittent renewables. Interestingly, from 8:00 to 10:00 PM, BESS supplies critical backup power, charging other sources to stabilize the system performance. The addition of FCs provides a critical buffer of power, particularly from 8:00 to 10:00 PM, when they supply 8000 W to stabilize the power supply and prevent blackouts. Having FCs guarantees a more stable energy system, less dependent on stored battery energy alone. FCs supplement the 8000 W gaps from 8:00 to 10:00 PM and serve as a valuable energy cushion to balance supply and prevent shortages. The contribution is critical with wind generation fluctuation and without the contribution of solar, supporting power throughout the clock. FCs also continue to support the system in 9:00 PM and 10:00 PM with 8000 W, significantly boosting nighttime energy reliability. This supplementary source of power ensures the system to have an uninterruptible power supply with energy compensation for variability in the solar and wind generation. The combined impact of FCs greatly enhances system ruggedness, making it to reduce the dependence on stored battery reserves and diminishing potential supply deficits. This is an effective hybrid system that performs effectively in balancing energy supply accordingly, utilizing maximum renewable energy and reserved reserve energy.

Fig. 9

Energy generation, storage, and demand profile for scenario 3 optimized using GTO.

Figure 10 is 24-h LPSP, which tests the system reliability. LPSP is nearly zero throughout the day, confirming that scenario 3 is working fine with solar and wind power. LPSP is gradually rising after 6:00 PM with the highest during 7:00–9:00 PM. FC integration ensures greater system dependability through the availability of firm energy, reduction in variability, and minimization of the likelihood of shortages. Compared to the other two scenarios, scenario 3 is a more secure and reliable energy system. Economic efficiency of scenario 3 is verified through its total cost of operation, which is $10,784.79 within 24 h. Although additional cost of FC integration, overall cost is equal to scenario 2 but system reliability is enhanced and LPSP is reduced. Cost effectiveness of scenario 3 is adequately demonstrated through nighttime periods, where FCs prevent batteries from over-discharging, reducing energy shortage and increasing dispatch efficiency.

Fig. 10

System reliability assessment for scenario 3.

Scenario 4

Scenario 4 integrates PV panels, WTs, BESS, FCs, and EVBs into the optimized MAS via the GTO. The configuration offers enhanced flexibility and reliability in energy management through the application of different storage systems to enhance energy dispatch, particularly during peak load. The integration of EVBs alongside FCs and the BESS offers enhanced system robustness to support load balancing and reduce reliance on any specific energy source. The MAS manages the power distribution adaptively in a manner so as to best utilize renewable resources with minimum economic losses and with minimized shortages in power. Figure 11 illustrates power dispatch over 24 h showing how solar PV arrays, WTs, BESS, FCs, and EVBs supplement the dynamic demand requirement. Demand levels vary from one region of the day to another with utilization of continuous availability of wind power and utilization of solar PV generations when there is sunlight. Solar power follows a typical daily cycle, beginning early in the morning and reaching peak production at midday, which enables efficient charging of storage devices.

Fig. 11

Energy generation, storage, and demand profile for scenario 4 optimized using GTO.

Peak solar time offers efficient charging of EVBs and BESS to store energy for later evening peak demand. The BESS responds variably to available energy, charging in the time of peak sunshine and discharging when demand surpasses real-time output in the evening and at night. Maximum charging occurs from morning to mid-day to cover sufficient stored power to be used afterward. BESS discharges significantly in the evening, with high discharges in times of maximum demand, supplementing the system’s energy stability. At 1:00 AM, BESS provides peak discharge of 8478.65 W, ensuring sufficient energy supply in the early morning time. BESS operates dynamically day and night, charging in the time of surplus energy and discharging when demand is more than supply. BESS discharges heavily at 1:00 AM with 8478.65 W, 6:00 PM with 10,054.45 W, and 9:00 PM with 10,800 W, ensuring stable energy distribution. Conversely, BESS charging occurs when renewable energy is most available, e.g., 8:00 AM at 6290.6 W and 10:00 AM at 5163.3 W, proving its role in regulating variable energy supplies and maintaining system stability. Most intense BESS charging occurs from 8:00 AM to 2:00 PM, matching peak solar PV power output, and deep discharge from 6:00 PM and later, filling in demand when solar PV is not generating power.

The EVBs operate dynamically, discharging primarily at peak demand periods in aid of energy supply. Discharges occur significantly at 9:00 PM with 1143.66 W, offering a stable energy supply during high load periods. EVs do not participate in night-time energy storage as BESS but act as an extra energy source during vital demand peaks. FCs integration takes precedence in late evening hours to supply stored energy between the hours of 7:00 PM and 11:00 PM through a maximum capacity of 6000 W. The additional generation becomes critical in maintaining energy reliability and preventing gaps in supply during peak evening times. Compared to previous states, scenario 4 indicates improved supply reliability due to the integration of EVBs and FCs to counterbalance overreliance on BESS alone.

Figure 12 displays LPSP for 24 h, a key measure of system reliability. LPSP is close to zero for most of the day, verifying that the system effectively leverages solar and wind resources to satisfy energy needs. Comparison to scenario 3, where nighttime reliability was also slightly lower, scenario 4 diminishes power shortages by a great margin, resulting in a balanced supply over the 24-h period. Overall effect of EVBs and FCs minimizes LPSP the most since it is certain that this combination has the best energy dispatch among all the configurations that have been tested. Scenario 4 economic performance is also determined by cumulative operating cost, amounting to $10,688.06 at the end of the 24-h period. Compared to the previous scenarios, it is less cost but more system flexibility and reliability. Scenario 4 cost-effectiveness is best achieved during peak times when FCs and EVBs are operated in conjunction with each other to ease over-dependence on BESS, thereby lowering operating costs. This balance between cost, energy availability, and system reliability makes scenario 4 a superior energy management solution.

Fig. 12

System reliability assessment for scenario 4.

Backup system performance analysis

Backup power generation and storage equipment is crucial in microgrids powered by renewables to provide power stability, especially where primary sources of energy are not stable. Incorporating backup equipment such as BESS, EVBs, and FCs maintains the system stable and provides a continuous supply of power during low solar and wind production. This section discusses the performance of the backup sources for all four cases and illustrates how well they performed in handling energy, load balancing, and system reliability as a whole. Each case uses different backup system configurations that influence energy availability, system reliability, and cost-saving. This is a point-to-point comparison of the performance of BESS, EVBs, and FCs for each case.

The BESS is essential in all scenarios by storing excess renewable energy and releasing it when supply is insufficient. In case 1, BESS is the primary backup system charging during the day and discharging heavily at night. Night discharge is the highest, resulting in possible energy shortages from insufficient capacity. In scenario 2, adding EVBs reduces BESS reliance, energy distribution more even; yet, BESS still discharges heavily from 6:00 PM to 12:00 AM. Scenario 3 introduces FCs, which reduce evening BESS discharge, particularly between 8:00 and 10:00 PM, when FCs supply 8000 W. Scenario 4 achieves maximum FC and EVB combination for optimal BESS load and system stability. EVBs are used in scenario 2 and scenario 4 as a second backup system for energy stability. There are no EVBs in scenario 1, resulting in the over-reliance of BESS for energy deficiency correction. Scenario 4 uses peak EVB integration, with peak discharges at 9:00 PM, and greatly enhances load balance and lessens reliance on BESS.

FCs are introduced in scenarios 3 and 4, which significantly increase evening reliability. In scenarios 1 and 2, with no FCs, LPSPs are larger since they rely so heavily on battery storage. Scenario 3 adds FCs, generating up to 8000 W during the 8:00–10:00 PM window, effectively overcoming energy shortages. In scenario 4, FCs retain the main role, particularly at 9:00 PM, serving as an extra contribution to total system reliability in filling in from renewable energy and contributing stable supply at peak times. Scenario 4 presents the best balanced backup system by employing a combination of BESS, EVBs, and FCs in minimizing dependence on any individual source. EVBs enhance the system flexibility to mitigate the pressure on BESS, particularly nighttime. FCs significantly improve night-time reliability, decreasing LPSP values by mitigating energy deficits.

Comparing the overall operation expense in the different scenarios reveals gigantic differences based on the composition of backup sources utilized. Scenario 1, using BESS alone, has the highest operation expense of $10,785.60 because of extensive battery cycling and degradation. Scenario 2 incorporates EVBs into BESS, which doesn’t have much effect on expenses and maintains the similar cost of $10,709.10. Scenario 3 includes FCs, reducing reliance on BESS and realizing a modest cost saving to $10,784.79. Scenario 4 is the combination of the best BESS, EVBs, and FCs to meet the lowest cost of operation of $10,688.06. Scenario 4’s advantage is having the capacity to distribute energy usage evenly over multiple storage sources, avoiding maximum use on any one component. Compared to scenario 1, scenario 4 achieves a cost saving of $97.54, reflecting more efficient energy use and reduced degradation expenses. The introduction of FCs in scenario 3 stabilizes the cost, yet their introduction in scenario 4 coupled with EVBs achieves the most balanced and cost-effective solution. This comparison reflects that a combined strategy using multiple backup sources not only boosts system reliability but also optimizes long-term financial sustainability.

Financial viability and cost analysis

The economic viability of microgrid systems from renewable sources is a key factor for their long-term sustainability and applicability. In this section, the economic efficiency of each scenario is evaluated on the basis of cumulative operating expenditures, cost sharing between storage units, and the impact of energy management policies on overall expenditure. By comparing different configurations of FCs, EVBs, and BESS, this research presents the most cost-effective energy management policies. Scenario 1, with its operation cost sum as a main index of economic effectiveness, is most expensive to run since it constitutes excessive battery cycling and premature ageing at a cost of $10,785.60. Scenario 2 with EVBs reduces BESS usage and is less expensive at $10,709.10, a reduction of 0.71% compared to scenario 1. Scenario 3 incorporates FCs to minimize cost by preventing battery degradation at a price of $10,784.79, a reduction of 0.0075% compared to scenario 1. Scenario 4, optimally combining BESS, EVBs, and FCs, carries the lowest cost of $10,688.06, a 0.91% reduction from scenario 1, reflecting the cost advantage of an diversified energy storage plan.

Cost analysis of different energy storage units shows higher economic efficiency. The BESS is the largest cost factor in all the cases since it experiences ongoing charging and discharging processes, thereby involving more maintenance and replacement expenses. The use of EVBs in scenarios 2 and 4 reduces the load on BESS by providing energy supply during peak demand hours, which reduces overall cycling rates for BESS and reducing long-term operation costs. FCs, taken under scenarios 3 and 4, complement evening energy support, reducing dependence significantly on battery storage. Although FCs have a higher initial investment, their ability to enhance energy stability and reduce LPSP values makes them worth incorporating. Energy management plays a crucial role in cost minimization. Methods such as peak load shifting, diversified use for storage, LCPS, and LPSP reduction are cost-saving. Methods such as EVBs and FCs combined offer enhanced load balancing and reduce pressure of peak demand on BESS. Multiple sources of energy storage, as in the case of scenario 4, reduce pressure on each one and therefore reduce the cost of maintenance. LPSP values for all scenarios remain within acceptable limits. Scenario 4 is the most cost-effective option with the lowest total operating cost and high energy reliability. The use of EVBs and FCs reduces the economic burden on BESS tremendously, especially enhancing cost-effectiveness overall. Scenario 1, however, with its high cost of operation and high rate of degradation, depending on BESS alone, is least economically viable. The addition of EVBs in scenario 2 adds some cost-effectiveness, while the addition of FCs in scenario 3 stabilizes costs. But scenario 4 maximally utilizes all the backup systems, with the best cost-performance ratio. The cost calculation highlights the importance of a diversified energy storage system for economic feasibility. Scenario 4 indicates the lowest total operating cost, which verifies the benefits of employing multiple backup sources. By reducing dependence on any single component and judiciously balancing energy storage, renewable-based microgrid systems can achieve higher cost-effectiveness and long-term.

Comparative performance analysis of scenarios

Apart from examining economic feasibility, a relative performance comparison among all four scenarios was made in order to assess their reliability, flexibility, utilization of renewables, and overall efficiency of the system. Although scenario 4 is found to achieve the lowest overall operating cost at $10,688.06 (0.91% lower than scenario 1), its excellence is also reflected through various important technical merits. Scenario 4 presents the most balanced and diverse energy storage configuration with BESS, vehicle-to-grid enabled EVBs, and HESS. This combination allows for the coordination of short-term, medium-term, and long-term energy storage efficiently, which grants the system a more versatile capability to address fluctuations on either the demand or supply side. Scenario 1 is fully reliant on BESS, which lags during long stretches of peak demand or low renewable supplies. Scenarios 2 and 3 enhance performance by incorporating EVBs or HESS respectively, but scenario 4 relies solely on all three storage technologies combined in optimized and integrated state.

The combination of EVBs and HESS in scenario 4 enhances the amount of energy on hand during peak hours, particularly during the evening and night, and offers a more stable electricity supply. Through real-time coordination and optimization by agents, the system can efficiently allocate the energy load to the storages on hand without surpassing the capacity of any individual storage unit. This approach makes the system more robust and reduces the chances of performance loss under dynamic loads. Scenario 4 also shows the highest efficiency in the use of renewable energy. Its storage form allows it to easily take up excess energy generated during peak sun and wind hours and redistribute it with little energy wastage. Condition of storage capacity limitation is less efficient in consuming and utilizing this excess power, thereby leading to inefficiencies such as power curtailment or unutilized generation capacity.

A further key benefit of scenario 4 is the support of long-term system sustainability. Through the sharing of charging and discharging duties among multiple storage systems, wear and stress on the individual units, particularly the BESS, are significantly reduced. This balanced operational load leans towards longer component life, less maintenance requirements, and improved overall system longevity. All these performance advantages capture the argument that, while all four scenarios are excellent, Scenario 4 is technologically most robust, flexible, and resilient design. Its adaptability, energy efficiency, and reduced operating stress make it well suited for off-grid and rural areas where energy stability, reliability, and sustainability are a priority.

Conclusions

The need for low-cost and reliable energy solutions for off-grid areas has fostered the development of complex energy management approaches. Hybrid energy storage systems, combining more than one renewable source and different storage technologies, are an attractive approach to building energy resilience and efficiency. Traditional energy management methods, however, are often founded on forecasting approaches that generate uncertainties and inefficiencies. This study developed a large-scale and adaptive MAS for hybrid energy storage optimization, including PV panels, WTs, BESSs, HESSs, and EVBs. The MAS employs real-time optimization with the GTO and exploits V2G technology to efficiently balance energy supply and demand. The system was tested in four different energy management cases in a small-scale microgrid with a peak load demand of 24,070 W to investigate the most cost-effective and reliable configuration. The main findings concluded in this study are summarized as follows:

• 1. Scenario 4 was the most cost-effective and lowest cost setup. It also possessed the lowest operating expense of $10,688.06, a reduction of 0.91% from the operating expense of scenario 1 at $10,785.60. The reduction percentage is small but significant in the small-scale microgrid setup where efficiency and energy conservation are the issues. Moreover, for larger applications, the cost savings would be even higher with the larger energy requirement and scope for optimization. Scenario 4 also enhanced energy resilience and stability significantly, offering more stable power supply with negligible fluctuation. By combining a set of energy storage solutions, it cut down operating expense without sacrificing a reliable and smooth supply of energy, thus making it the most suitable solution for remote energy management.

• 2. V2G technology integration with HESS was behind improving system performance. HESS balanced the energy supply by alleviating volatility in low renewable generation periods, and V2G technology enabled real-time energy trade between EVBs and the grid. This dual integration maximized the availability of energy, reduced unwanted cycling of batteries, and enhanced battery lifespan, hence reducing maintenance costs and improving system life.

• 3. Comparison of the other scenarios made the benefits of a diversified energy storage approach more pronounced. Scenario 2 (PV–WTs–BESSs–EVBs) reduced the cost to $10,709.10, down by 0.71%, but without the advantage of HESS’s system stability enhancement. Scenario 3 (PV–WTs–BESSs–HESSs) reduced the cost by a small margin to $10,784.79, but along with experiencing the advantages of reducing battery aging and system stability improvement. Scenario 1 (PV–WTs–BESSs) was the most costly to run and did not benefit from the resilience and flexibility brought in by EVBs and HESS and was hence least effective.

• 4. The MAS, optimized with GTO, effectively managed real-time energy dispatch. The system boasted an LPSP of below 0.05, ensuring a reliable and uninterrupted power supply. The low LPSP indicates the dynamic sensitivity of the system to fluctuations in renewable energy resources, preventing shortages and optimizing the passage of energy, thus also ensuring its reliability in areas of poor resources.

• 5. The MAS solution provides a cost-efficient and scalable means towards sustainable energy management in rural communities. The article demonstrates the technical and economic feasibility of integrating multi-energy storage options with real-time optimization. Through this approach not only is operational cost reduced but also enhanced grid reliability is guaranteed, and increased energy availability is promoted in under-served communities. By employing hybrid energy storage, V2G technology, and smart optimization techniques, the MAS architecture provides a healthy blueprint for prospective renewable energy setups in off-grid and resource-limited settings.

While the present study focused on a 24-h operational simulation to evaluate the short-term performance and economic efficiency of the proposed hybrid energy system, future work will aim to expand the analysis to a larger-scale system over an extended planning horizon. This will include a comprehensive financial assessment that accounts for capital investment, operational and maintenance costs, and overall lifecycle performance. In addition, more realistic modeling of EV availability will be considered by incorporating time-dependent connection patterns and user behavior to better reflect practical V2G operation. Incorporating these aspects will provide a more robust understanding of the system’s long-term economic viability, operational flexibility, and scalability in real-world applications.