An Evaluation of Potential Strategies in Renewable Energy Systems and Their Importance for South Africa—A Review

Abstract

The ageing of coal-fired power stations in South Africa has led to regular power outages. Therefore, the country will need to urgently increase its electricity capacity to meet further energy demand from growing urbanization and population growth. This challenge has increased interest in alternative energy sources, such as renewable energy (RE). South Africa is gradually implementing appropriate renewable energy practices, reducing carbon emissions, cleansing the air, and assuring a more sustainable future. This paper summarizes the availability, current state, and future potential of renewable energy choices in South Africa. This paper also evaluates specific policy measures and government actions aimed at eliminating barriers and increasing renewable energy deployment in the future. It also considers the South African network’s specialized techno-economic analysis. The findings show that South Africa is still heavily reliant on coal, with 78% of the current installed capacity coming from coal power in 2022, compared to 9.3% for solar and wind energy.

1. Introduction

The growing energy demand in South Africa and the ageing of coal-fired power stations have led to frequent power outages. Therefore, the country will need to urgently grow its electricity capacity to meet further energy demand from increasing urbanisation and population growth [1]. This challenge has increased interest in alternative energy sources, such as RE. The advantages of RE include energy savings, environmental benefits, improved investment options, and improved power security, among others [2,3]. Wind and solar power generation are two of the most promising modern RE technologies [4]. However, integrating these resources into the power system is challenging since they are intermittent and have unknown maximum generation limits [5,6,7]. A solar or wind power system is usually used in remote areas without a mains electricity supply [8]. Daily and seasonal fluctuations limit standalone power systems based on renewable energy sources (RESs), which results in difficulties in regulating the output power to the load [9,10]. Solar and wind power systems cannot be used solely as a power source for electrical installations that require constant, guaranteed energy throughout the day, week, month, or year. Although many issues are associated with RESs, they can be overcome in various ways. A viable alternative to this is using hybrid energy systems in place of individual sources [11].

The RES estimation is essential in system design and implementation in the modern energy sector [12]. This requires digitalizing the entire system, from resource estimation to the energy needed on the demand side [13]. Smart grid technologies, using the internet of things, have come up with a practical framework model called the internet of energy to enhance the overall performance of the energy system [14,15,16]. Several research studies have assessed potential schemes of RES and their application in diverse environments. In addition, RESs provide various advantages in the energy sector, such as the production of green hydrogen [17]. In [18], a feasibility strategy is presented to evaluate blue and green hydrogen exploitation in Europe by 2030. The study looks at the incapacity of RESs in producing green hydrogen and doubts arising in the European energy market regarding short-term blue hydrogen implementation. The existing European infrastructure, regulations, and technologies inhibit the spread of hydrogen. This is why the potential evaluation of RES is necessary to judge the economic growth of nations [19]. In India, solar energy is considered a promising RE that can effectively meet the increasing energy demand in agriculture [20]. Li et al. [21] assess the intersection between RE development and financial inclusion in China. Through an empirical assessment, the interconnections between system evidence, policymakers, and environmentalists are established for the potential long-term economic benefits of RESs’ implementation. The modelling of resources provides an opportunity to handle the system uncertainty and improve the overall reliability and resilience of the power network. This scheme further guarantees the energy dilemma in which the power sector becomes secure, equitable, and environmentally sustainable for better management techniques [22].

The management of the electrical power grid poses several challenges, including the economic dispatch problem (EDP) [23]. This problem is concerned with continuously balancing power generation and load while minimizing the operational costs of all generation units in the grid [24]. Ref. [25] defines the EDP as the determination of generation levels such that the total cost of generation becomes the minimum for a defined level of load demand. Refs. [26,27] emphasize the importance of finding the optimal solution for the EDP to achieve sustainable power systems with a focus on low dispatch costs, meeting demand–supply balance, and maintaining constraints such as power generation limits, ramp-rate limits, prohibited operating zone and valve-point loading [28,29]. Therefore, this study aims to evaluate the effectiveness of optimal intelligent control in minimizing the operation costs of a hybrid system. The information and data utilized in this research are sourced from various reliable and relevant published research works. In addition, the contributions made by these sources to this paper can be summarized as follows:

• Assess the current energy scenario in South Africa. This indicates the existing energy landscape in South Africa, highlighting the challenges and opportunities associated with RE integration in modern power grid development.
• Explore the implementation of RE with real data in different locations. The study incorporates data from various places where RE systems have been implemented. Thus, the assessment of the empirical information adds credibility and practicality to the analysis.
• Provide a techno-economic analysis tailored to the South African power network. Therefore, the study includes a comprehensive techno-economic investigation that explicitly considers the characteristics and requirements of the South African power grid. Furthermore, the analysis examines the classification of economically viable solutions for RE integration and their application within the innovation power grid environment.

2. Modelling and Methodology of Renewable Energy Integration

A strategy based on an optimization algorithm aims to reduce the cost of the most effective solution. This classification includes two categories of methods: real-time optimization approaches and optimization algorithm strategies [30,31]. Optimisation algorithms are formulated based on previous understandings of future operational conditions [32]. Among the global optimization methodologies, particle swarm optimization (PSO) consists of mathematical models, dynamic software programs, evolutionary algorithms, and optimal controllers [33]. Real-time optimization algorithms aim to minimize problems by using power supply choices at each end [33]. This method must be mathematically modelled in real-time concerning processing and memory capacity [34].

2.1. Stochastic Methods

Ref. [35] describes the optimization and control of strategies utilizing improved combinatorial optimization by applying the genetic algorithm (GA) model. The study examined the possibility of iHOGA for developing multi-objective combinatorial systems for new smart cities in Tamil Nadu for the first time while simultaneously reducing three competing objective functions: cost, unmet demand, and CO₂ emissions. The best Pareto set of combinations of methods for reviewing the three contributions was utilized by optimization using the multi-objective evolutionary algorithm and GA. The findings suggested that the most efficient hybrid varieties can handle the load in upcoming smart cities. The evaluated power demand was likely to be met by combining RE systems at all the chosen sites; no single DG system was used.

Suresh et al. [36] developed an off-grid hybrid renewable energy system (HRES) for electrification in remote rural areas using an HRES. The objective of the study was to reduce the Total System Net Preset Cost, Cost of Energy (COE), unmet load, and CO₂ emissions using the GA and HOMER Pro Software. Comparisons were made between the two methods and four combinations of hybrid RE systems: solar photovoltaic (PV) system—wind turbine generator—biogas generator—biomass generator—fuel cell—battery, solar PV system—wind turbine generator—biogas generator—biomass generator—fuel cell without battery, solar PV system—wind turbine generator—biogas generator—biomass generator—battery without fuel cell, and solar PV system—wind turbine generator—biogas generator—biomass generator without storage. A total net present cost and energy cost evaluation was conducted for these four combinations. The findings indicated that, when comparing all four combinations of HRES using HOMER and GA, the GA-based optimization was more cost-effective than HOMER, with the lowest COE of 0.163 $/kWh and 0% unmet load. In addition, GA-based systems have higher PV penetration and lower CO₂ emissions than HOMER-based systems. In ref. [37] an optimal design of an integrated approach involving a Wind-PV-Diesel-battery (BT) system for an isolated island with CO₂ emission evaluation was performed using a GA. The overall annual cost, which includes capital prices, fuel expenses, replacement costs, and operation and maintenance costs, was reduced using a GA strategy. It was discovered that the system could lower its overall annual cost and lower the CO₂ emissions produced by DGs.

Bhongade and Agarwal [38] optimized the Combined Economic and Emission Dispatch problem by comparing two optimization methods, Artificial Bee Colony and GA. They aimed to minimize the total generation cost subject to generator and power constraints. The findings indicated that the Artificial Bee Colony algorithm was superior in both cases, with or without losses, compared to the GA. Zhu et al., the authors of [39], developed a hybrid intelligent algorithm-based interval forecasting approach for wind power generation established on an extreme learning machine and PSO. Hybrid intelligent algorithms aimed to gain optimal PIs without prior knowledge, statistical inference, or distribution assumptions of forecasting errors required by most traditional approaches. The extreme learning machine applied in the proposed method is a learning algorithm designed for training single-hidden layer feedforward neural networks with high learning speeds and generalization capability. As compared with the benchmarks applied, experimental results demonstrated that the developed approach was highly efficient and reliable.

The authors of [40] proposed a multi-objective PSO technique for economic dispatch in a power system considering land pollution. The aim was to use an evolutionary search algorithm called the multi-objective PSO algorithm to solve the EDP in power systems while considering environmental pollution. It was found that when two objectives were optimized simultaneously, related costs were highly reduced. In Ref. [41], the most efficient design for an off-grid and on-grid hybrid PV-wind generating system with BT storage was investigated. They aimed to meet annual load demand while considering costs associated with energy supply, load losses, and pollutant emissions. The objective function aimed to minimise the components’ net present, emission, and loss-of-load costs. The decision variables, including the capacity of the renewable sources and the BT bank, were optimally determined using the spotted hyena optimization algorithm according to the solar irradiation and wind speed patterns, and they were proportional to the objective function, component constraints, and load interruption probability constraint. The findings demonstrated that, while total net present cost decreased, MAPnet increased, improving system reliability.

A hybrid metaheuristic optimization technique was developed by Ellahi and Abbas [42], who presented a hybrid metaheuristic optimization algorithm designed to solve the EDP encountered in different combinations of power plants. The study aimed to accelerate cost reduction and convergence with the Bat Algorithm (BA) by combining PSO and PSO-BA. The critical elements of PSO and BA were used to construct the algorithm. Additionally, it added a brand-new parameter, ‘α’, multiplied by the PSO velocity equation and dependent on BA with PSO. All RES-based energy systems have been used to test the designed algorithm (without restrictions, with time-varying demands, and multi-area economic dispatch). The results indicated a reasonable reduction in cost, faster computation, and faster convergence. A model predictive control (MPC) approach for grid-connected solar-wind with pumped hydro storage was proposed by Siti et al. [43]. The purpose was to demonstrate the benefit of the closed-loop optimal control model for integrating RESs into the grid to sell power into the grid during peak periods using time-of-use tariffs. The open-loop system was more vulnerable to external disturbances and could not be corrected without feedback. Therefore, the benefit of using MPC in open-loop control was that it could respond to external disturbances or inaccurate supply sources. In addition, the MPC could react to output disruptions and re-optimize the system during its control horizon [44,45].

2.2. Deterministic Methods

In [24], the authors proposed distributed learning control for economic power dispatch. A reinforcement learning strategy and consensus strategies were used to address frequency regulation and EDPs in smart grids while preserving the privacy of units. The results indicated that the proposed control approach could handle scenarios in which finding an accurate cost–function model was complex or the cost–function was non-convex. Moreover, each unit’s cost–function or control policy was private and was not shared with neighbours.

Ref. [46] investigated a multi-time-scale robust economic dispatching method for the power system with clean energy. The study attempted to build robust uncertainty models of wind, solar, and load projections for diverse time scale scheduling, converting deterministic constraints into strong constraints considering uncertainty. As the timeline shrank, more conservative scheduling was introduced to increase the degree of scheduling conservatism. This tactic successfully lessened the effects of unpredictable wind, sun, and load forecasts. A significant balance of security, economic, and environmental benefits was also achieved by alleviating adjustment pressure and lowering wind and load curtailments.

The authors of [47] developed a machine-learning strategy to boost power use in Tanzania’s rural areas. This study aimed to monitor shared load from the demand side in an off-grid solar-powered microgrid. The technique’s input was the reported electricity usage statistics of all the homes in a residential community throughout a specified time. The measurements had a time scale of 15 min, and the measurement length was a week. The output was the result of a power usage type assessment. The proposed approach is divided into two sections. The first stage was the detection selection process, intended to restrict the data extraction targets and save calculation time. The second component was the home energy use assessment, which was used to analyse the interrelated aspects of household power usage from the standpoint of usage behaviours and time. The results showed that, in addition to encouraging users to use more electricity, the proposed strategy extends off-grid power supply time, enhances system dependability, and, ultimately, improves residents’ experiences.

Zhou et al. [48] developed a data-driven machine learning model using supervised machine learning with excellent computational efficiency to estimate on-site renewable power production, which was then used for uncertainty and sensitivity assessments. The goal was to explore the on-site power production of an ideal mixed renewable network utilising deterministic characteristics under the uncertainty of high-level parameters. The results showed that, as opposed to deterministic situation factors, instances with scene uncertainty could improve the peak demand and overall amount of on-site renewable energy output. Patel et al. [49] designed a machine-learning algorithm for estimating power in a modest off-grid PV installation. The study aimed to create a revolutionary scaled-down model of an internet of things-enabled datalogger for PV panels in isolated areas where human interaction was not allowed owing to hostile climates or other factors. The recorded data were stored and visualised using an internet of things platform. The datalogger’s acquired data were used as the basis for training machine learning techniques. A linear regression approach was used to estimate electricity generation. The results demonstrated that the offered strategies for predicting power generation performed better regarding the derived features.

Yin and Lei [50] proposed a deep reinforcement learning (DRL) technique and a deep deterministic policy gradient (DDPG) algorithm for the integrated offshore wind and PV power system’s combined functions. The goal was to enhance energy production performance while dampening oscillations. Furthermore, the authors have highlighted DRL’s advantages in dealing with recurrent complicated decision-making situations, such that the cumulative reward can be maximised when engaging in an unpredictable environment. It is also flexible and model-free and requires no prior understanding of the environment to learn the generalised optimal control strategy from historical data. The DDPG method, on the other hand, is a model-free strategy appropriate for sequential unpredictable optimum control challenges. As a result, the results demonstrated that an OWT can respond quickly to unexpected changes in input wind conditions to maximise the total power supply. Significant fluctuations in total power production were also effectively reduced by adjusting generator torque, indicating that similar functioning of offshore wind and PV electricity is possible.

Ref. [51] investigated deep learning (DL) approaches for predicting power and renewable energy demand in smart microgrids. This research aimed to identify, categorise, and evaluate the DL techniques used for demand prediction or power prediction in solar and wind energy. Their work also discussed the limitations and future directions of DL approaches for predicting wind, solar, and electricity demand. The findings suggested that DL-based approaches might improve the efficiency of intelligent microgrids by providing prospective strategic solutions for accurate energy production predictions from RESs and load demand prediction. However, DL techniques can only attain more excellent performance when large amounts of high-quality data are accessible [52].

2.3. Optimal Control Scheme

The power flow control (PFC) algorithm and fuzzy logic controller (FLC) were used by the authors of [52] to build a novel method for optimizing an off-grid PV-wind system with energy storage and load management. The comparison was between a traditional and hybrid system that uses an FLC and a PFC algorithm to manage the load. A sizing technique was created based on solar and wind energy concentrations and a series of overcast days without wind. It helped identify the critical months of the year. The rated power and usage pattern of household equipment should be considered while estimating household power consumption [53]. According to the results, the power flow management algorithm (PFC algorithm) lowered BT investment costs by 33%, increased power supply dependability by 1.35%, and decreased consumption by 1.35% while preserving user comfort. Aside from that, the FLC application has led to a 50% reduction in BT use, a 4.9% drop in energy usage, and no noticeable changes in user comfort.

Al-sakkaf et al. [54] demonstrated an autonomous DC MG that implemented a simple FLC to control power flow to provide a residential dwelling load. The research aimed to design, simulate, and operate a DC MG controlled using an efficient and straightforward fuzzy logic energy management system that used real-world data. An Artificial Bee Colony was used to address the optimal control problem based on the mathematical model, which led to a 10.89% increase in control and energy savings. It has also been determined that the proposed fuzzy-based energy control system for MGs was a dispatch problem with economic implications. The findings of the simulation indicated an improvement in control and an energy reduction of 10.79% when compared with the fuzzy logic energy management system, which had not been optimized. This was 3.17% higher than the fuzzy logic energy management system tuned in the first stage. In addition, compared to traditional economic dispatch strategies, generation costs were decreased by 11.19%.

In [55], an energy management strategy for the DC MG based on a fuzzy logic-based controller was examined. This research aimed to design an energy management system for DCMG current mode control and a controller for a DC-DC converter. The charging and discharging of the combined energy storage system were controlled by current mode control using a fuzzy logic-based system as a voltage controller. Renewable energy resources (RERs) were linked to the DC bus in the DC-DC converter controller using DC-DC boost converters. These boost converters were implemented using maximum power point tracking (MPPT). In MPPT, the match between the RER and DC bus was optimized for maximum power extraction from RER. According to the results, the tested EMS performed well under load, solar irradiation, and wind speeds. There were five scenarios where the employed PI and fuzzy logic-based system controllers kept the DC voltage at the required level. This was with fluctuations of no more than 4% and 3.3%. The controllers handled the RER intermittently by charging and discharging the hybrid energy storage system. Due to the control approach, the SC responded faster to rapid load changes than the BT, while the BT provided long-term power. In times of need, the DG performed well.

A Hopfield neural network approach was developed by Ganesan et al. [56] for the optimal design of hybrid energy systems with diverse RESs. The purpose was to optimize the design parameters of a hybrid distributed generation power system with alternative energy sources (solar and wind power) regarding power balance and design constraints [57]. Therefore, it was necessary to modify the power system’s design characteristics to reduce costs, increase stability, and reduce pollution levels. As a result of meteorological circumstances being non-deterministic, the results showed that incorporating the fuzzy environment into the equation gave more realistic outcomes. In addition, this method performed an excellent job of optimizing the objective function. Ref. [58] created a more efficient power management controller for a grid-connected hybrid system consisting of energy storage and PV generation. Elman neural network was used to design a hybrid power management strategy. The first contribution of this research was the design of an intelligent supervisory controller based on Elman neural networks, which eliminates the requirement for a rule-based structure or earlier mathematical modelling. Second, the authors developed the power management strategy utilising state flow to collect the training and test datasets for designing the Elman neural network controller. Results showed the suggested technique was reliable for appropriate forecasting model-based power management control.

A power management system based on artificial neural networks (ANNs) was presented by Sahoo et al. [33] for controlling the power in AC-DC mixed-distribution channels. In autonomous renewable microgrids, they aimed to develop a HESS voltage control technique based on artificial neural networks to lower Li-ion BT load when switching devices were replaced. The ANN acquired data on the power system, distributed generation energy, and charge level state of charge to select the best operating mode. The outcomes indicated that the energy management technique could control the ESS energy flow. Additionally, it managed their charging and discharging power throughout operation based on their state of charge condition and the load demands of the system. Ref. [59] discussed how ANNs have been used to model different solar energy technologies. The review covered several ANN kinds, activation function types utilized in various ANNs, and common statistical performance evaluation standards for assessing ANN performance. ANNs have also been used in multiple solar-powered appliances, including solar collectors, solar-assisted heat pumps, solar air and water heaters, PV/thermal (PV/T) systems, solar stills, solar cookers, and solar dryers. Despite the simplifying presumptions in the ANN-based mode, the results showed that utilizing ANNs eliminates solving complex mathematical models. In addition, fewer trials were needed to compare input/output relationships to experimental investigations.

An effective PV-Wind scheme was reported by Pancholi and Chahar [60]. In the study, the PV system’s MPPT utilized the Nonlinear Autoregressive Moving Average (Narma-L2) Neural Network approach to boost productivity. The wind power system also included permanent magnet synchronous machine-based generators to reduce output variability and increase efficiency. The system used an FLC for parameter control to verify the dynamic voltage restorer layout. The dynamic voltage restorer was designed to handle the system’s voltage instability and surges. The simulation output was subjected to a MATLAB examination, and the results were verified. An intelligent combined generation system with a PV and WT was introduced in [4]. To show the efficiency and supremacy of the blended fuzzy neural controller, they compared it to three MPPT controllers: perturb-and-observe, incremental conductance, and FLC. Utilizing Matlab/Simulink, a mixed system simulation was created. According to the results, the proposed method demonstrated high dynamic performance in monitoring the blended power system’s maximum power output swiftly and precisely. Moreover, the proposed approach tracked the maximum power point faster and more precisely than perturb-and-observe and FLC. The radial basis function network-sliding mode in the WT also demonstrated improved transient performance, increased efficiency, and higher strength.

3. Distributed Energy Resources

Several forms of distributed energy resources exist, from conventional to renewable resources [61,62]. Thus, this study evaluates the most relevant distributed energy resources based on RE for a suitable modernization of the power sector in South Africa. Energy storage is also considered a potential distributed generation that serves to improve the operability of RE due to its variable nature.

3.1. Solar Power

Solar PV systems transform sunlight into electricity using the PV effect. As a result of the sun, semiconducting materials produce voltage and current. In practical applications, solar cells are discrete objects with electrical properties that change in response to light [63]. In a solar PV system, the PV panel absorbs sunlight and converts incident photons into electricity. The energy generated by the panel is transferred from direct current (DC) to alternating current by an inverter [64]. Solar energy is more affordable than other energy production methods. They are also plentiful and valuable for many things. In addition, solar power systems require relatively low maintenance [65]. However, their fundamental drawback is that they depend on weather intermittency, necessitating a BT storage system that would raise the cost of the innovation altogether [66,67]. It began as a small-scale application but has grown into a mainstream electricity source.

Several countries have relied on solar energy since the origin of solar panels in the 1950s [65]. In the past, the United States was the world’s top solar electricity generator, followed by Japan and Germany, but today, China remains the most significant generator [68]. PV cells are primarily powered by solar energy. Weather conditions are essential in this power source, although it is relatively cheap and clean. As a result, solar power is unreliable due to weather fluctuations [56]. A PV system, however, offers better reliability, ease of use, lack of noise, no pollution, a long life cycle, and a cheaper cost per unit of energy produced [69].

Kannan and Vakeesan examined the future of solar power [70]. Their overview included the fundamentals of PV technology and the factors driving its advancement, possible uses, and challenges. For the solar industry to progress, they also highlighted some of the excellent research carried out in solar power generation, PV/thermal collectors, solar heaters, design improvement sizing, and materials for effective light absorption. Ref. [71] examined PV and solar thermal potential in domestic buildings to determine how they affected almost-zero-energy buildings. In this study, steady-state energy calculations were applied, which offered relatively less correct information than dynamic analysis but were readily implementable for various building topologies. Based on the results, a solar combi system combined with a modest PV system could supply enough energy to make a building a net-zero-energy building. Ref. [72] presented an overview of solar energy technologies and discussed how they impact human health, safety, and the environment. In the study, solar energy was highlighted as having environmental benefits. However, greenhouse gas emissions from solar systems, especially during production, are a disadvantage and must be handled carefully, just as e-garbage. The study concluded that innovative production technologies could improve PV modules efficiency, lengthen their lifetime, reduce silicon mass per module, use less electricity during production, and recycle properly to minimize environmental impact.

In [65], Rabaia et al. investigated and addressed the environmental implications of solar power systems, using the most recent and effective life cycle assessments and ecological impact analyses available on the market and developing solar PV and concentrated solar power systems. The main stages were examined during the discussion, from early design to manufacture, product use, building or instalment, through their lifespan, and finally during their decommissioning. As a result, it was determined that several well-known systems have had their environmental effects well researched. In addition to technical and ecological considerations, developing wafer-based and thin-film technologies, including Gallium Arsenide, Copper Zinc Tin Sulfide, Organic PV, Dye-Sensitized Solar Cell, and Colloidal Quantum Dot PVs, need more attention. Shah and Ali [73] reviewed studies discussing hybrid nanofluid-based solar energy systems.

Furthermore, mono-nanofluid solar energy systems have been analysed for their performance. A discussion has been conducted regarding the significance, fabrication methods, characteristics, and implications of hybrid nanofluid systems on solar system performance parameters. Results showed that the hybrid nanofluid solar system showed a performance expansion of over 200 percent due to its optimal characteristics.

Ref. [74] reviewed solar PV system cooling technologies. Its purpose was to provide an analysis of various technologies for lowering the surface temperature of PV modules. Different cooling technologies were reviewed: a floating tracking concentrating cooling system; a hybrid solar PV/thermal system cooled with water spraying; a hybrid solar PV/thermoelectric system cooled with a heat sink; and a hybrid solar PV/thermal cooled with forced water circulation. More technologies included improving the performance of solar panels through phase-change materials, solar panels with water immersion cooling technique, solar PV panels cooled by transparent coating (photonic crystal cooling), and hybrid solar PV/thermal systems utilizing thermoelectric cooling. It was found that any adequate cooling technology for PV panels should maintain a low, stable operating surface temperature and be simple and reliable. In addition, it should allow the extraction of thermal heat to increase conversion efficiency if possible [75].

3.2. Wind Power

Wind power is one of the eco-friendly RESs in the fundamental representation of the wind configuration system, with a source from converting wind speed into mechanical power. Thus, it is a process where the kinetic energy of the moving air is converted into the mechanical power of the turbine shaft, which will generate electrical energy [76]. A wind power conversion system is a networked system comprising connected parts. This system converts the kinetic energy of the wind into mechanical energy and then into electrical energy by using generators [63]. The WT is the main component of a wind power conversion [77]. A WT is a rotor with two or three attached propellers that harness wind energy. In windy situations, low-pressure air collects on one side of the blade and is drawn toward it, rotating the rotor. Additionally, as WT generators produce DC, an inverter is needed for alternating current applications [63]. Despite the initial cost, wind power is one of the cleanest and most economical sources because it is readily available and generates no pollutants [56]. However, the weather significantly impacts how much wind energy is produced. As a result, the WT’s position is significant. It is impractical to rely solely on wind energy because its dependability is unstable due to changing weather conditions [56,78]. Additionally, turbulence, a random fluctuation added to the average wind speed, is a common cause of the change in wind speed [79,80].

Furthermore, the wake impact of a wind farm exposes downstream WTs to increased turbulence [81]. As the wind patterns of the WTs in a wind farm alter over time, some swings are stabilized, and the output energies of the turbines match one another [82]. BESS is suggested to be utilized with other theories, such as a dynamic thermal rating system [77,78,79,80], to integrate wind energy systems and manage transmission line congestion and wind energy fluctuations. Furthermore, the use of wind energy is highly dependent on how well it protects the environment and requires careful study [83]. By carefully locating and developing wind farms and designing WTs, noise levels from wind farms can be significantly lowered [84]. The number of bird deaths brought on by wind energy is relatively low. Evidence suggests it is less extensive than other energy sectors or structures like power lines [85,86,87].

Novacheck and Johnson [88] presented a revised version of mean-variance portfolio optimization to examine the feasibility of lowering wind variation using various wind sources. The study aimed to understand better how different wind affects power systems by reviewing optimization strategies for exploring multiple winds. The power systems model’s findings illustrated how the complexities of the existing power system could unexpectedly affect the outcomes that can be anticipated when employing mean-variance portfolio optimization to vary wind power portfolios. Additionally, the findings demonstrated that wind diversity lowers wind restriction and transmission congestion, increasing diversity’s value beyond that expected by mean-variance portfolio optimization.

Verdejo et al. [89] developed a method for characterizing the stochastic and self-sustaining temporal pattern of a wind farm’s produced power and the electricity demand. The primary goal of this study was to stochastically simulate, in continuous time, the performance of wind farm power production and the power demand due to domestic consumption using a model that considers stochastic and self-sustaining dynamics over time. Two one-dimensional and two-dimensional short and continuous time representations have been proposed for modelling wind generation. The potential interaction between windmills in the multidimensional scenario, including linked Brownian motion, was also considered. The results showed that the one-dimensional method better described the output.

An extensive evaluation of the most recent techniques and advancements in wind power forecasting and prediction was provided by Foley et al. [90]. First, upscaling and downscaling methods, ensemble forecasting, and numerical wind prediction approaches from global to local sizes were examined. According to [91], every electricity market player can lower their financial and technical risk by making accurate wind power predictions. It was further noted that integrating wind power will be more accessible when innovative grid technology and intelligent load management techniques are implemented [92]. Furthermore, it was shown that system operators would require less operational effort to integrate wind power if their generation portfolio included energy storage facilities, offshore wind, wave, and tidal energy.

Yang et al. [93] reported estimates of wind power supply in dense urban areas. The study aimed to create evaluation methods based on computational fluid dynamics that took into account the specifics of the surrounding urban terrain and boundary conditions of microenvironments to identify possible locations for WT mounting and estimate wind output. The predictions, which included those of wind speed and direction as well as fluctuation intensity, were compared with real measurements obtained using ultrasound anemometers and thermal flow speed meters. These measurements were obtained at 10 controlled locations at five levels in an intended building. This was performed to validate the computational model and better understand how wind interacts with buildings in complex terrain.

Pinson and Girard [94] offered event-based verification-based diagnostic methods in addition to multivariate verification tools. The study aimed to describe a methodology for assessing the quality of short-term wind power supply scenarios using multivariate validation methods already in use and a diagnostic approach. According to the findings, their use in evaluating various short-term wind power generating scenarios shows them to be effective discrimination tools. A survey of cutting-edge techniques and recent wind power uncertainty forecasting advancements was presented [13]. There was a brief introduction to three distinct representations of wind power uncertainty. Then, various predicting techniques were reviewed. The findings showed that evaluating uncertainty forecasting was more challenging than assessing spot forecasting, which needed specific properties to be defined and an evaluation framework to be established.

3.3. Hybrid System

A combined power system is a power generation system comprising more than one distinct power technique [95]. This power system uses various power streams to complement one another and improve the reliability of the power supply [96]. Occasionally, such an energy system might consist of more than one independent power source, each driven by a different technique. This includes solar PV, wind power, geothermal, small hydropower, and DGs. However, when we refer to a combined renewable power system, we mean that all contributing power generators must be powered exclusively by renewable sources [62]. They may even have power storage facilities. In addition to producing power, this system protects the environment from the dangers associated with conventional power generation methods that burn fossil fuels [63].

Many studies have been conducted on combinatorial PV-wind-BT systems and BT capacity improvements. The main focus of several studies [97,98,99] is the synchronization of load patterns, wind power output, and BT energy. Furthermore, the inherent intermittency of renewable energy resources makes it challenging to regulate and control power flow in distribution networks [100,101]. Ref. [102] claims to have devised a way to overcome these challenges. Microgrids (MGs) are local distribution systems with the capacity for generation, storage, and load, according to [55]. The MG in [55] comprised distributed generators, diverse loads, power storage innovations, and a control system.

The novel generation of produced electricity focuses mainly on RESs, such as solar and wind. Although it has several advantages in lowering greenhouse gas emissions, its drawbacks include irregularity and unpredictability. These restrictions must be overcome for the system to perform like conventional generation units, such as power dispatching [103]. Given the significant unreliability of solar and wind power, it is highly desired to incorporate storage batteries into the distributed generation system [104]. As a result, the energy storage system can eliminate the variations and provide a constant power source over time [105]. The storage batteries can thus act as a regulator to regulate the variability in source and demand [106]. When it comes to dependability, cost-effectiveness, and pollutant emissions, each power source differs from the others [106]. Therefore, the creation of a combinatorial distributed generation network appears to be a desirable solution for simultaneously satisfying all three criteria [56]. Innovative grid technologies provide opportunities to dynamically handle the variables of renewable energy resources [107].

Hybrid power suppliers play an essential role in the stability and reliability of the network, especially in developing countries where the relationship between suppliers and consumers is critical [108]. In addition, the three main integration problems associated with the structural balancing of the load with wind and solar electricity supply are low production, poor dispatchable plant utilization, and overproduced supply, according to Ueckerdt et al. [109]. Although it was acknowledged that these challenges were not exclusive to varying RE, the study addressed the challenges of integrating varying RE into present and future power networks. Based on the results, varying RE has to be planned from a system perspective so that wind and solar supply match consumption. However, production costs should not be the only factor influencing wind and solar energy utilization.

Weschenfelder et al. [110] studied the compatibility of grid-connected solar and wind generation systems. The review provided a foundation for understanding the utilization of solar PV–wind hybrid systems, which primarily concentrated on sizing, modelling, and control. The study aimed to undertake a comprehensive review of innovative techniques to determine the compatibility of grid-connected solar and wind power systems, an essential element of extensive grid integration. The results showed that the grid might be more secure by combining different amounts of wind and solar energy in various regions. This would smooth out the total amount of power produced by these sources.

Adaramola et al. [111] performed an economic analysis of the practicality of employing a mixed power system consisting of solar, wind, and DGs for usage in rural parts of southern Ghana. This was performed using the levelized energy cost and net current cost of the system. This study investigated the economics of a mixed power system that included solar, wind, and conventional DGs. It was used in rural southern Ghana. The analysis used a hybrid optimization model for electric renewables (HOMER) software, and viability was evaluated based on the net current cost and power price. The results indicated that the hybrid power plant produced 791.1 MWh of total electricity per year, of which 126.9 MWh (16.04%) came from solar PV, 248.2 MWh (31.38%) from WTs, and 416.0 MWh (52.58%) from the two DGs.

3.4. Energy Storage System

Energy storage captures and preserves power for future usage [63]. In order to accomplish this, energy must be changed from one form, which is typically challenging to store, to another, more convenient form [63]. This objective is accomplished using an energy storage device. An energy storage system is a critical component of a comprehensive energy system [112]. As a result, it contributes to stable operation and proper load transfer [113]. In addition, power grid security is enhanced, RE is increased, energy utilization efficiency is improved, and sustainable energy development can be realized [114,115]. Furthermore, in order to introduce RESs into the power system, energy storage systems (ESSs) are crucial [116]. In addition to providing fast active power compensation, it is a critical component of power systems [117]. ESSs better mitigate the effects of non-dispatchable variable generation [37,38]. Energy storage systems, particularly BT storage, are commonly used to cope with wind power intermittency, prediction errors, and participation in power markets [118,119,120,121,122].

The deployment of Battery Energy Storage (BES) systems is expected to increase significantly in the coming years due to the abundance of resources, fewer geographical restrictions, and the speed with which they start and ramp up [123]. BES can reduce renewable power curtailment at the grid scale [124]. Furthermore, BES can balance load generation, undertake energy arbitration, reduce peak demand, and defer transmission and distribution [125,126]. However, their uptake is limited by high capital costs and the non-recognition or valuation of the services provided [127]. The limitation of the services BES systems can use results in lower revenue and more extended investment payback periods [127]. In contrast, using appropriate control mechanisms, BESs can provide multiple services simultaneously [127]. Therefore, BES must participate in multi-value services to boost its revenue and ensure economic viability [127].

Battery-based energy storage also provides a variety of advantages, including improved power system reliability [104]. It is proposed to integrate BESSs with other theologies, such as dynamic thermal rating systems, to manage wind power intermittency to alleviate congestion on transmission lines [128,129]. As a result of the high power and energy density of Battery Energy Storage, it is ideal for compensating for fluctuations and forecasting errors [130,131,132]. In contrast, a lack of BESS capacity may compromise reliability and safety [133]. Ref. [134] compared long-duration flywheel BT storage, lithium-ion BT storage systems, and lead–acid BT storage systems using the Monte Carlo approach. The results indicated that long-duration flywheels were expected to yield a lower net energy rate of storage and a lower net energy charge than lithium-ion batteries.

3.5. Hybrid Energy Storage System

Batteries and supercapacitors (SCs) are among the several energy-storage options offering high power density and efficiency [135]. In contrast to batteries, SCs have several advantages, including fast charging, safety, and excellent cycling life. According to their energy storage mechanism, SCs can be classified into two categories. In one case, the electrical double-layer capacitor stores charge near the electrode interface via the adsorption of electrolytes. In contrast, in the other case, a pseudo capacitor is created by reversible Faradaic reactions to produce higher specific capacitance and energy density [136,137,138,139,140,141]. There have been vigorous developments in pseudo-capacitor electrode materials such as transition metal oxides, sulphides, hydroxides, and conducting polymers up until now [142,143]. Despite this, SCs with higher energy density cannot meet the growing demand.

Wang et al. [144] examined the power storage application in a combined energy system. The study aimed to explore how power storage technologies were used concerning RESs thoroughly. This was carried out by focusing on how well they can manage unpredictability and variations. The results showed that most power storage technologies could autonomously regulate frequency and voltage based on the innovative properties of power storage technologies and RE. In addition, they could level the power of RE within a predetermined generation gauge. Authors of [104] examined determining the optimal BT capacity in a combined wind–BT system at the design stage to participate in the unit commitment program and provide constant power at specified intervals. The study was conducted to determine the BT capacity needed at the design stage based on long-term wind speed data. The results indicated that the amount of BT capacity required to compensate for variations in wind farm output power and to minimize the cost of deviating from the grid could be determined.

Schaefer et al. [145] examined a technique and tool to determine the necessary storage capacities, power rates, and ramp rates for a specific MG case. This was calculated for both a single power storage system and a mixed power storage system design. The study aimed to estimate the above values for each power storage system in a hybrid power storage system. Low-pass filters with cut-off frequencies were used to create a set of sub-profiles based on time-scale categories associated with storage systems. Based on these sub-profiles, the requirements were estimated using a rudimentary storage model. The results indicated that the methodology allowed for an easy and quick analysis of the necessary storage flexibility for a given scenario.

Mohammadi et al. [146] introduced an energy control system based on PV and BT compatibility for socializing smart homes. Smart homes with different lifestyles and uncertainties were studied simultaneously to determine BESS behaviour and size. The results showed that the suggested method offered a means of figuring out the size of the BT storage system. It also provided a schedule for using home appliances in several intelligent houses. Additionally, the model was developed to incorporate smart homes, varied lifestyles, and different energy storage technologies, considering the consumer’s specific needs.

Jaszczur and Hassan [147] used high-resolution sequences to investigate a grid-combined PV system for a typical house with a rapid storage unit and an SC. The goal was to minimize high-current flows between the local power system and distributed generation by optimizing SC size as per demand and PV system size. The estimation of the research time interval allowed for the evaluation of the impact of electrical demand and temporal resolution on PV system power flows. This was carried out to determine the most appropriate time resolution to guarantee minimal calculation error. The results showed that, compared to a system without local power storage, adding small but rapid power storage improved self-consumption by approximately 83% and 114% for a bright and partly overcast day, respectively, with an average increase of 100%.

Ref. [148] compared the effectiveness of PV systems containing and without an SC. The aim was to improve the system’s efficiency, reduce current and energy losses, and protect the BT from peak currents that reduce its lifespan. Findings showed that combining the SC and the BT in the solar cabin led to many significant improvements. In addition, when solar power was available, the SC supplied most of the power to maintain the BT voltage. Additionally, the SCs could provide peak demands, reducing BT current fluctuations. Moreover, SCs significantly reduce BT power and BT current.

Using MATLAB/Simulink, the authors of [149] investigated a combinatorial PV-BT/SC system with active power management. The benefits and drawbacks of SC were examined and analysed in a combinatorial approach. The study also suggested a unique combinatorial power storage system topology to share and lower peak PV power. The findings showed that SCs reduced BT load and stabilized voltage on both sides. Longer BT life and lower system costs were the immediate results of this.

Abbassi et al. [150] proposed a statistical technique for estimating a blended energy storage system. This study outlined a dynamic programming approach for evaluating a combined energy storage system. It involved determining how much energy was stored in the BT and SC based on their initial charge point. The power of the blended energy storage system was split into two distinct shapes using frequency management: one for the BT and the other for the SC. Results showed that hybridizing the storage system improved performance indices more than batteries alone. Additionally, the allocation of SCs permitted a reduction in power consumption, the avoidance of fast charge/discharge cycles, and, consequently, a decrease in BT usage.

In order to address issues with power fluctuation, Mamen and Supatti [151] addressed the uses and overviews of combined energy storage systems. These systems integrate SCs’ storage devices with BT systems. The study aimed to evaluate the combined energy storage system combining BT storage and SCs. Additionally, initial investment, rated power, and storage system features were used to determine the most effective pairing of magnet energy storage with a BT and an SC with a BT. According to the findings, magnet energy storage and batteries were a suitable match for energy generation systems with high energy consumption, such as WTs. Furthermore, SC and batteries were appropriate for systems with low power penetration, such as solar PV generation.

Ref. [152] presented an efficient blended power storage system employing a distributed online fuzzy control system for managing reverse power in an isolated wind-diesel system. The study aimed to combine an SC and a BT storage system with an appropriate control strategy. This was performed to distribute the power differential between the system power supplies and user loads. Severe operational conditions, such as user load and a stepping wind speed, were simulated for the complete system. The findings showed that the proposed fuzzy type of DSC could enable the combination supercapacitor–Battery Storage System to appropriately balance actual power under user load variations and those caused by wind energy variants, effectively controlling system frequency/voltage dynamics and improving the wind–diesel combination system power quality. Ref. [153] suggested a combined storage solution to address the primary issues on the load and supply sides by storing energy in batteries and SCs. The objective of the study was to evaluate a novel topology combined with a BT/SC system and real power control in MATLAB/Simulink. The results showed that employing SCs will reduce BT degradation and boost effectiveness.

4. System Estimation

4.1. Potential of Renewable Energy in South Africa

Since 1994, South Africa’s electricity consumption has increased by 4% per year due to the government’s goal of guaranteeing universal power access and economic impact [154]. South Africa’s electricity reserves fell from 25% in 2001 to between 8% and 10% in 2007 [155]. Eskom has struggled to meet electricity consumption since 2007 [156]. This electricity shortage had a negative impact on the South African government’s target to expand the economy by 6% annually between 2010 and 2014 [155]. Eskom is having difficulty supplying the country’s electricity needs, particularly during peak hours, due to the ageing of power stations [157], marginalized energy reserves, and the high capital cost of new coal-fired power stations [158] using emerging technologies [159]. Pollutant gas and particle emissions from coal-fired power plants harm the environment and society in the immediate vicinity of the power plants and distant zones, including neighbouring countries [160]. Africa, an emerging continent, remains vulnerable to the conversion of fossil fuels designed to develop the countries to which it exports the raw oil [53,161]. Figure 1 shows the South African power grid in 2022. Given recent advancements in the manufacturing and cost-effective installation of renewable power generation systems, particularly PV and windmills, RE is an option for addressing South Africa’s energy problem [162]. In February 2023, the South African government implemented a solar panel tax incentive for individuals to motivate households to invest in the RE-producing capacity that augments electricity supplies [163]. Figure 2 presents the potential of renewable sources in South Africa between 2022 and 2023.

4.2. Installed and Potential Renewable Energy in South Africa

4.2.1. Biomass Energy

Biomass can be converted into electricity, heat, or liquid fuels [164]. In addition to being burned to generate electricity, it can be used as fuel for cooking and heating [165]. Moreover, it can be used indirectly by utilizing its biological activities to create ethanol, methanol, and gasoline for transportation and cooking [165].

Biomass fuel is mainly produced from wood. Plant and animal waste can be processed to create several types of biofuel [166]. Biodiesel-producing facilities are operational in South Africa [165]. South Africa has significant potential for biomass generation since it has 42 million hectares of natural wood and 1.35 million hectares of plantations [167]. Moreover, with a possibility of 313 MW in 2030, South Africa has installed 63 MW in 2022, see

Table 1. Installed electricity sources 2022 (MW).

Coal	Wind	Solar PV	Solar CSP	Biomass	Nuclear	Hydro-Electric
46,259	1705	1520	400	63	1800	1545

and

Table 2. Potential electricity sources 2030 (MW).

Coal	Wind	Solar PV	Solar CSP	Biomass	Nuclear	Hydro-Electric
51,509	8705	5880	1200	313	1800	2545

.

4.2.2. Hydropower

The low annual rainfall rate of 500 mm in South Africa limits hydroelectricity production [168]. Seasonal flows and recurrent droughts are challenges for South Africa’s hydropower industry [169]. The country’s highest hydro-potential lies in the Eastern Cape province [165]. Table 1 and Table 2 indicate that, presently, the country has 1545 MW of installed capacity with a potential of 2545 MW. Although they can generate up to 5091 MW of power, large-scale hydropower production systems can severely impact the environment [165]. The facilities occupy a large amount of land, and the volume of running water might harm river ecology [165]. Small-scale hydroelectric production systems can produce up to 69 MW of electricity without significant environmental impact.

4.2.3. Wind Energy

One of the most effective forms of RE is wind energy. Wind power generation is an alternative to fossil fuel generation since it does not release carbon emissions [165]. Over the past ten years, WT technology has advanced significantly, and numerous upcoming businesses have entered the market [170]. Wind farms have gained popularity, with more giant WTs, better availability, and increased efficiency [171]. Furthermore, wind energy may be created and harvested from South Africa’s large coastline and physical environment, including lowlands and the high veld escarpment [165]. South Africa has fair to reasonable wind resources globally [165]. Despite the recent start of the business, the economics of wind energy are now solid [171]. There is an expectation that wind energy costs will continue to decline. Moreover, the Diab wind atlas classified distinct wind energy-generating regions as high, moderate, or low [172]. The coastal parts of the Western and Eastern Cape areas were designated as having a high wind energy capacity because they had an average annual speed of more than 4 m/s at 10 m above ground level [173]. The Bushmanland, Drakensberg foothills, and Kwazulu-Natal areas were identified as having moderate wind energy possibilities. Finally, the Bushveld basin and Cape Middleveld were designated low-wind zones [173]. Figure 3 indicates the total potential for wind energy in South Africa.

4.2.4. Solar Energy

The two main types of solar energy used in South Africa are PV and concentrated solar power [173]. Solar energy is the process of turning solar radiation into electricity. This is accomplished by directly using traditional PV or concentrator PV systems to transform global horizontal insolation (GHI) into energy [162]. Concentrated solar power systems also transform energy indirectly. Linking the properties of the main system components to environmental data, particularly solar irradiance and temperature, is necessary to predict the energy output of a PV system [174]. The operational energy information of these systems is of high significance in the judgment calls by investors and developers who are in charge of designing power plants to generate electricity using PV systems [175]. High insolation incidence on the PV system results in the most significant outcomes for solar power systems [176]. PV technology becomes more efficient when combined with a heat collector [177]. By 2030, South Africa is expected to produce 7000 MW of solar energy [173], as seen in Table 2. According to predictions, South Africa’s demand may be satisfied with 3000 km² of solar-powered land [165]. On average, South Africa’s landscape experiences 2500 h of sunlight annually [165]. In South Africa, solar energy adoption influences RESs like wind, hydroelectricity, and biomass [165]. Figure 4 below indicates the total potential for solar energy in South Africa.

Pgap = Ppotential − Pinstalle

(1)

Equation (1) finds the power gap between installed electricity sources in 2022 and potential electricity sources in 2030.

4.3. Techno-Economic Analysis

It is essential to do a sizing analysis before installing a hybrid system. Solar irradiation, temperature, wind speed, wind direction, hub height, and biofuel calorific value impact how much energy a renewable power system can generate. Since RESs are not available everywhere, their electricity output varies. Since natural resource availability varies depending on latitude and longitude, it is essential to consider these factors while planning MG configurations. This includes creating optimal ratings and performing a techno-economic analysis of HRES. It is critical to do a techno-economic assessment of HRES before using it. HOMER Pro software can assess the enhancement of MGs, whether on or off the grid. Both power system costs and HRES sizes can be estimated through HOMER Pro [178]. Off-grid HRES in three different parts of Syria have been optimized by Merei et al. [179]. The usage of batteries in the HRES has been found to improve its efficiency, effectiveness, and environmental impact.

In order to fulfil a home’s load demand in the coastal region of Turkey under varying geographical and climatic conditions, Duman et al. [180] conducted a technical and economic analysis of an off-grid HRES-comprising solar system, WTs with fuel cells, and batteries. Off-grid HRES systems cost more. The BT is cheaper than the fuel cell. Javadi et al. [181] suggested using multiple residential MGs to swap electricity, maximise reward, and minimise COE. They claimed that their non-cooperative game theory-based control technique increased total payout by 169%. Wind-based RE systems require careful planning, primarily when intended for load increase uncertainty. Valinejad et al. [182] offer a wind energy conversion system growth planning model that maximizes investment and profit by subsidizing and considering load ranges. Balancing power generation and demand while cutting costs as much as possible with the MPC technique has been presented by Tavakoli et al. [183]. Plug-in electric vehicles can reduce peak load and enhance demand response in power systems.

A BT energy storage system helps the HRES maintain’s voltages if generated power exceeds power consumption, maximizing RE use. Series and parallel batteries of the same rating increase energy capacity and backup [138]. A power converter that transfers power between the HRES, AC, and DC buses costs 200 USD/kW to buy and 175 USD/kW to replace. The power converter, however, is low-maintenance and lasts 15 years. Peak demand for electricity delivery determines converter capacity [139]. GAs [140], hybrid genetic algorithms [141], graded particle swarm optimizations [142], meta-PSO [143], mixed-integer quadratic programming [144], graphical construction [145], and probabilistic approaches [146] have been studied to develop efficient methods for size and cost control and efficiency improvement. By utilizing simulation research, optimization, net-present-cost, COE, loss of power supply probability, and sensitivity analysis, HOMER Pro [147] is extensively used for energy planning and cost optimization.

Cost-effective solutions require component selection and energy resource-based scaling. For decades, researchers have focused on multi-energy source integration to overcome techno-economic constraints using distributed RE systems [184]. The energy planning and cost optimization software HOMER Pro is utilized in [185]. It performs simulation, optimization, net present cost, loss of power-supply-probability, and sensitivity analysis [185]. Hybrid system sizing software includes RET Screen, iHOGA, INSEL, and HOMER. The RE Research Laboratory (RERL) has developed Hybrid2 software. This application can model hybrid solar/wind/fuel cell systems. RET Screen, established by the Canadian Ministry of Natural Resources in 1998, simulates energy systems from technical, financial, environmental, power efficiency, and other viewpoints. TRNSYS, a thermal network modelling tool from the University of Wisconsin in Madison, is widely used in the area. This programmed simulated a hybrid PV/thermic system, which outperformed the solar PV system technically and economically. Some commonly used software for techno-economic analysis is listed in Table 4.

There are four common approaches to sizing: (a) Analytical, (b) Iteration, (c) Probability, and (d) AI. The hybrid system is used as an analytical approach to numerically model the system’s feasibility in the South African hybrid solar-wind system analysis. This study seeks to improve the efficiency of the hybrid system water. Hybrid central generates 100,000 MWh per year at 0.97 EUR per kWh and utilizes 75,000 cubic metres of water. In the iterative method, the best system design ends this recursive procedure. This method was designed for a Brazilian community’s solar PV, wind power, and BT-based hybrid system. This research seeks a low-cost, reliable system. Probabilistic techniques factor wind speed isolation and system design changes into integrated system size. The data imply that one of the most fundamental sizing techniques may not be the best. A machine or artefact with artificial intelligence can think and reason in much the same way as a person can. AI algorithms such as GA, multi-objective self-adaptive differential evolution algorithm (MOSaDE), non-dominated sorting genetic algorithm (NSGA-II), mine blast algorithm (MBA), PSO, multi-objective line-up competition algorithm (MLUCA), Ant colony optimization (ACO), and preference inspired coevolutionary algorithm (PICEA) are used to estimate the optimal size of a hybrid system.

5. Discussion

South Africa has hot and temperate climates, making using solar and wind resources particularly appealing. Table 1 shows the installed electricity sources in 2022. It shows that South Africa is still heavily reliant on coal, with a capacity of 46,259 MW, followed by nuclear power at 1800 MW, wind at 1705 MW, hydro-electric at 1545 MW, solar PV at 1520 MW, solar CSP at 400 MW, and biomass at 63 MW. Table 2 also depicts potential electricity sources in 2030, indicating solar, wind, hydroelectric, and biomass growth. Furthermore,

Table 3. Gap in electricity sources between 2022 and 2030 (MW).

Coal	Wind	Solar PV	Solar CSP	Biomass	Nuclear	Hydro-Electric
5250	7000	4360	800	250	0	1000

illustrates how much energy is still required to close the deficit regarding electrical sources between 2022 and 2030.

Figure 3 also depicts windy places that are excellent for wind energy exploration. Furthermore, Figure 4 shows areas in South Africa that are hot and suited to solar energy exploration. Moreover, techno-economic analysis necessitates employing various software to evaluate the enhancement of MGs, whether on or off the grid.

Table 4. Some of the common software used for techno-economic analysis and sizing optimization.

Software Name	Advantages	Disadvantages
HOMER	Efficiency results graph. Very simple to understand	Employs linear equations of the first degree. Inability to recognize time series data.
iHOGA	Uses either single or multiple objective optimizers. With a short simulation step time, the computer will not need to restart as often.	Lack of Probability and Sensitivity Analysis. Everyday workload constraints (10 kWh).
Hybrid2	Numerous load electrical options. Options for Detailed Dispatch	Long simulation times. Although the project was created without significant problems, there were issues with the simulation.
RET Screen	Top meteorological database. Application based on Excel.	Reduced need for information entry. Failure to accept time series data.
TRNSYS	Simulator adaptability. The visuals are quite precise.	Not all power plants can be modelled, including hydroelectric plants. Optimisation choices are unavailable.

lists some of the most commonly used tools for techno-economic analysis and sizing optimization, as well as their advantages and disadvantages. As a result, the results demonstrate that HOMER’s outputs are easy to interpret but cannot recognize time series data. IHOGA is deficient in probability and sensitivity analysis. Hybrid 2, on the other hand, includes extensive load options for electricity and detailed dispatch options but has long simulation periods. RET Screen has the top meteorological database, while TRNSYS cannot model all power plants.

As the country grapples with frequent power outages caused by Eskom’s inability to meet energy demands and the challenges of reducing the amount of coal supplied to the South African power grid, as shown in Figure 1, using RE becomes critical in addressing electricity security and environmental concerns. Moreover, RE helps to mitigate the use of fossil fuels, which are harmful to the environment. According to Table 3, there is a substantial gap that South Africa needs to fill between 2022 and 2030 to achieve its energy target by 2030. To bridge this gap, bid windows 5, 6, and 7 are anticipated to procure 6800 MW of solar PV and wind power. These additional power capacities will be connected to the grid in late 2023 [186].

Several studies have been conducted to assess the potential of RE in South Africa. While most of these studies have primarily focused on the country’s RE policy, they have also evaluated the energy density of various RESs. These studies have identified and quantitatively assessed different RESs available in South Africa. Figure 2 shows that wind is the most widely installed renewable energy source in South Africa as of 2022, followed by solar PV, hydroelectric, and biomass. Furthermore, Figure 2 depicts the potential renewable energy in South Africa for 2030, indicating that wind is the most available renewable energy source, followed by solar PV, hydroelectric, solar CSP, and biomass. In recent years, there has been a significant surge in interest surrounding RESs. This can be attributed to the growing demand for affordable energy solutions and mounting concerns about greenhouse gas emissions and their impact on climate change [187]. Furthermore, the identification of operational load characteristics in MGs is crucial for proper system sizing. Table 4 summarises the different software tools used for techno-economic analysis and sizing optimization. These software tools aid in analysing the economic feasibility and optimizing the sizing of RE systems in MGs.

6. Conclusions

This study presented a range of information supplied by the various RE sources in SA and used maps to assess the possibilities of exploitable locations for generating electrical energy. Many prospective renewable energy sources have been discovered, and reference sites for SA have been analysed. The potential of the various RE sources available varies from one place to another depending on critical aspects such as geographical structure, resource types, and features. Furthermore, based on the data, South Africa is still heavily reliant on coal and is gradually switching to renewable energy. Moreover, the study presented a comprehensive review of previous works on the operative control of hybrid systems. Multiple strategies from diverse research domains have been thoroughly reviewed and classified based on their specific focus, contribution, and the type of control strategy employed to achieve optimal solutions. As a result, the knowledge acquired in this paper may be helpful for developing and modelling renewable energy generation systems in various places and in future research.