Techno-Economic Optimization of an Isolated Solar Microgrid: A Case Study in a Brazilian Amazon Community

Abstract

Many communities in the Brazilian Amazon region remain without reliable access to electricity due to geographical barriers and the high cost of connecting to the national grid. This study aims to evaluate the techno-economic feasibility of implementing battery storage systems in an existing isolated solar–diesel microgrid located in Tunui-Cachoeira, in the district of São Gabriel da Cachoeira (AM). The analysis uses an energy balance methodology, implemented through the HOMER Pro simulation platform, to assess three scenarios: (i) without batteries, (ii) with lithium-ion batteries, and (iii) with lead–acid batteries. Technical and economic indicators such as net present cost (NPC), levelized cost of energy (LCOE), diesel consumption, and renewable fraction were compared. The results indicate that incorporating lead–acid batteries yields the lowest LCOE (1.99 R$/kWh) and the highest renewable fraction (96.8%). This demonstrates that adding energy storage systems significantly enhances the performance and cost-effectiveness of microgrids, offering a viable path to electrify remote and hard-to-reach communities in the Amazon.

1. Introduction

The growth in energy demand as a consequence of population increase and limited reserves of conventional energy, which are not only depletable but also the main source of greenhouse gas generation worldwide, are contributing factors to the various current energy problems [1,2]. For the current development of society, electrical energy is an essential resource and an important indicator of a country’s economic activity. Its consumption is directly related to the quality of life of the population, as it reflects the use of essential goods and services for society [3].

As a result of limited reserves of conventional energy, there is currently a global push for diversifying the electricity matrix with the aim of achieving reliability and availability in the electrical system while adhering to sustainable development guidelines [2]. Therefore, for developing countries, the use of renewable energy sources represents an important tool for their economic growth [2]. Brazil, despite having a significant contribution from renewable energies, does not have a great diversification of its electricity matrix [2]. This is due to the low cost of electricity generated by hydroelectric power plants, which account for 64.9% of the electricity generated in the country [2,3].

Within this context, it is important to examine how electrical energy is distributed across the Brazilian territory. In Brazil, the National Interconnected System (SIN) is responsible for supplying most of the country’s electrical energy. The system connects all states except Roraima. This interconnection allows consumers connected to SIN to be served with greater reliability, as it minimizes all risks of interruption while also enabling the regional complementarity of renewable sources [4]. However, in the states belonging to the northern region, SIN coverage [4] is still limited to capitals and metropolitan areas. In municipalities where connection to SIN is challenging due to technical and economic reasons, electrical energy is supplied through isolated systems (SISOL) [4]. Figure 1 contains information about the location of SISOL and SIN in Brazil.

Most SISOL energy production is located in the Amazon region, with the exception of Fernando de Noronha [4]. All SISOL energy systems together correspond to 1160 MW and are responsible for supplying electrical energy to 3.3 million inhabitants, the majority of whom are located in urban areas and their surroundings [4]. According to the National Electric System Operator (ONS) of SIN, there were 212 SISOL energy systems in Brazil as of 2021 [5].

According to the Energy Research Company (EPE), the primary fuel used for electrical energy generation in SISOL is diesel oil, accounting for 97% of its generation capacity [4]. As per [5], diesel generators (DGs) are commonly chosen for their relatively low initial costs and rapid implementation; however, they have several disadvantages, including (i) ongoing fuel costs [6]; (ii) greenhouse gas emissions [7]; (iii) environmental risk of fuel spills during transport to remote locations; (iv) noise when used without proper insulation or at a large scale; and (v) fuel loss due to theft or leakage [8].

In addition to the inhabitants supplied by SIN and SISOL, there are communities located in remote areas far from municipal centers, facing various difficulties and high costs in accessing urban distribution lines [4]. For these communities, it is more convenient to use the natural resources present in the region as a source for local electricity generation. In such cases, supply must be carried out through small-scale systems with decentralized generation, such as photovoltaic systems [4,5].

However, despite Brazil’s high national electrification rate of approximately 99.8% (Brazilian Institute of Geography and Statistics—IBGE), this percentage drops significantly in the Amazon region. It is estimated that approximately 990 thousand people without access to electricity in Brazil are located in the Amazon, with 19% living on indigenous lands, 22% in conservation units, and 10% in rural settlements [4].

Providing quality electricity to the Amazon is a challenge for the electricity sector. Distributing electricity to these populations living far from the national grid is difficult and costly for the federal government, private entities, and individual households, both economically and environmentally [5]. The characteristics of the region’s localities, which include long distances, logistical difficulties, and high installation costs, are obstacles to ensuring formal access to the public electricity service.

For a long time, resorting to diesel generators has been the solution implemented by residents of rural areas of the country. For example, the Ribei-Rinha community of Santa Helena do Ingles, located in the Sustainable Development Reserve (RDS) of the Rio Negro, relied on diesel generators. However, due to the difficulties they faced, they could only generate electricity for about four hours a day, which considerably limited the development of economic activities and the quality of life of the inhabitants [9].

The Amazon region, as an alternative to the use of diesel generators, has local renewable energy sources, such as wind, solar, hydroelectric, and biomass energy, which are alternatives to increase electrification with microgrids and reduce the use of fossil fuels [10]. Decentralized electricity generation, transmission, and distribution are proving to be the most important solution thanks to the availability of renewable energy sources and the deployment of microgrids [10].

The use of DGs has been the solution implemented by rural residents in the country. This fuel, besides being environmentally unfriendly, is costly to acquire and transport, making this alternative expensive. Furthermore, the lack of maintenance of generators in these areas increases fuel consumption and decreases generator efficiency [11]. For example, in Tunui Cachoeira, diesel fuel cost BRL 45 in 2016, due to the cost of transportation that is carried out for 10 days by raft from Manaus [11]. According to PETROBRAS, for 2020 the percentage of mandatory biodiesel in the diesel fuel that is distributed in Brazilian territory is 12% [12] and its reference value according to ANP is BRL 6.80 per liter in 2023 [13]. If biodiesel is produced close to the place of use, it can become an alternative to replace diesel and thus reduce the cost of producing electrical energy.

The Amazon region, as an alternative to DG, has access to local renewable energy sources such as wind, solar, hydroelectric, and biomass energy. These sources offer a way to increase electrification with microgrids and reduce the share of fossil fuels in the electricity generation mix [10]. Decentralized generation, transmission, and distribution of electricity appear to be the most important solution due to the availability of renewable energy sources and the deployment of microgrids [10].

Thus, the transition toward renewable energy microgrids emerges as a strategic pathway to promote energy inclusion in the Amazon, while aligning with sustainability goals and overcoming logistical barriers. In light of this context, the present study is organized as follows: Section 2 presents a review of the scientific literature related to microgrids, photovoltaic solar energy, and energy storage systems, with a focus on applications relevant to remote and isolated areas. Section 3 details the methodological approach adopted for system sizing, including the case study, modeling tools, and input parameters. Section 4 discusses the simulation results and sensitivity analyses of different scenarios involving battery technologies. Finally, Section 5 summarizes the main findings and offers recommendations for future research and applications.

2. Distributed Energy Resources (DERs)

The application of renewable energy technologies in isolated microgrids has been widely explored in the literature, particularly for rural and remote regions with limited or no access to the conventional electrical grid. Several studies have proposed different configurations combining photovoltaic generation, diesel backup, and energy storage systems to address technical, economic, and environmental challenges. This section reviews recent works related to microgrid planning, component selection, and energy forecasting techniques relevant to off-grid electrification. Rather than presenting broad classifications of photovoltaic or storage technologies, the focus here is on approaches and systems directly comparable to the present study. Following this review, the unique contributions of this work are discussed in detail.

In the context of these studies, a microgrid is typically understood as the interconnection of low-voltage generation units, energy storage systems, and loads within a local area. These systems can operate either in connection with the traditional electrical grid or in complete isolation [14,15,16]. Microgrids connected to the grid can participate in the energy market as sellers or buyers, enabling them to exchange energy with the energy distribution company [15]. On the other hand, an isolated microgrid operates independently, disconnected from the grid [15]. According to [14,17], various countries around the world conduct research on various topics related to microgrids, owing to the variety of benefits that can be obtained from their application, including the following: (i) Increased reliability of energy supply. (ii) Reduction in the environmental impact of electrical energy supply. (iii) Reduced investment in plant, equipment, and cost. (iv) Increased stable energy efficiency. (v) Ensuring energy supply diversity. (vi) Supplying energy to remote sites. (vii) Capacity provided by energy storage.

Microgrids can be composed of renewable generation units, conventional power generation systems, such as internal combustion (IC) engines, electrical energy storage systems, thermal energy storage systems, electrical loads, and thermal loads [14,15]. Figure 2 shows the structure of a microgrid.

Microgrids can be connected to the grid through a point of common coupling (PCC). The isolation device isolates the microgrid from the electrical energy grid. Distributed Generation (DG) units, both renewable and conventional, are responsible for generating electrical energy in the system. The energy storage system balances and controls the flow of energy. The control system, which may consist of a central controller or distributed controller, is responsible for maintaining the microgrid safely in different operating modes [10].

Microgrids are designed to supply energy on a small scale, such as small communities, university campuses, commercial areas, and more [14,16]. They often face issues with adequate electrical power supply due to insufficient energy generation. This is primarily due to the irregularity of loads and renewable energy sources. Therefore, an Energy Management System (EMS) is required to address these issues [15].

Microgrids that use renewable energy sources for electricity generation are an alternative to ensuring energy supply in remote communities. Below are descriptions of some microgrid projects:

In [18] clairand and the other authors aim to study the long-term power generation planning of microgrids in remote communities in the Amazon in Ecuador, in order to evaluate new alternatives for electrifying the region. The work carried out by Ustun in [10] presents a feasibility study of microgrids based on renewable energy for electrification in the Amazon. Solar and hydropower are used as the primary sources of generation.

In [14] Melo and the other authors perform the sizing and simulation of a microgrid model using renewable energy sources such as solar and wind energy, along with an energy storage system using batteries.

Reference [19] introduces a new hybrid prediction method for short-term solar irradiance forecasting. The proposed method employs a set of image processing metrics to extract features from all-sky images, which are then used as input for machine learning models. This approach is noted for its greater interpretability compared to deep learning models, demonstrating competitive results and offering promising directions for future research in photovoltaic power prediction.

Reference [20] presents the influence of a set of all-sky image processing features on the prediction accuracy of the HPM Artificial Neural Network. Using correlation-based feature selection, three predictive models with different sets of input features were evaluated. This study provides new insights into the optimization of solar irradiance prediction using HPM, contributing to advances in photovoltaic power forecasting.

Reference [21] presents an original and comprehensive dataset with nine attributes extracted from all-sky images developed using image processing techniques. This dataset and the analysis of its attributes offer new avenues for research in solar irradiance forecasting.

The study carried out by Kalamaras in [22] designed a renewable energy-based system to meet the electrical and thermal demands of a remote home on a Greek island without a grid connection. The research conducted by Dawood in [23] evaluates the technical and economic feasibility of renewable energy-based systems using hydrogen as an energy storage for an off-grid microgrid.

Holguín’s work in [24] aims to design a hybrid photovoltaic system considering energy data from the distribution center in Corona Colcerámica using HOMER PRO software (Version 3.16.0). The study in [25] presents the development of the dimensioning of a modular and expandable system for electricity generation with off-grid energy storage to serve single-family riverside homes (two to eight people) in isolated communities in the Amazon, with the help of the Proknow-C systematic methodology, which demonstrates a systematic approach for rigorous and structured bibliographic reviews. In addition, the dimensioning of the Amazon Kit is carried out from data mapping to the estimation of consumption per person in homes, followed by the analytical calculation of the off-grid photovoltaic solar system, considering the base of the CRESESB portal. SAM (and HOMER PRO^® software) are used to simulate and validate the systems.

Finally, Witt’s thesis [26] investigates the feasibility of a hybrid hydrogen fuel cell and lithium-ion battery system for energy storage, while Ramos in [27] conducts the optimization of a hybrid system composed of photovoltaic panels, a diesel generator, and batteries for off-grid applications.

Photovoltaic solar energy continues to stand out as a promising alternative for clean and distributed electricity generation, particularly in regions where expanding transmission lines is economically or geographically unfeasible [2]. Among its benefits are reduced electricity expenses, local job creation, renewable energy expansion, CO₂ emission reductions, and decreased losses in transmission and distribution. According to [2], Brazil had installed capacities of 2.99 GW (centralized) and 4.25 GW (distributed) by December 2020. This diversification has improved the reliability and resilience of the national electrical system.

The trend of decreasing investment costs reflects the evolution of CAPEX values for photovoltaic generation in BRL/kW. Similarly, the evolution of O&M costs, which initially showed high dispersion, has stabilized over time as the technology matured and operations became standardized [28].

Photovoltaic energy, by nature, is generated in direct current (DC), requiring converters to produce alternating current (AC) at standardized frequencies. According to [29], it is often more efficient to install small-scale PV systems close to consumption points.

In isolated systems, energy storage is a key factor in improving operational reliability and maximizing the use of renewable sources. Storing electrical energy typically involves converting it into other forms. Various storage technologies exist, each with specific characteristics and applications [30].

Battery technologies such as lithium-ion, lead–acid, sodium–sulfur, and flow batteries are most commonly applied. Their selection depends on technical and economic evaluations, which vary by use case [30]. Over the years, significant cost reductions and improvements in response time, efficiency, and flexibility have driven the expansion of battery projects [30].

Lithium-ion batteries are particularly relevant due to their high energy density, efficiency, and expected cost reductions driven by their widespread use in electric vehicles and electronics [30]. Their potential for second-life applications in stationary systems further enhances their viability. Lead–acid batteries, though older, remain cost-effective and widely used despite their lower energy density and limited tolerance to deep discharges [31].

Battery storage is particularly advantageous in isolated systems, which often rely solely on diesel generation. Adding batteries allows for greater renewable integration, fuel cost reduction, and lower emissions [30]. According to EPE [32], in some cases, the combination of photovoltaic systems and battery storage is already economically advantageous. While regulatory and commercial barriers remain, the high generation costs in isolated regions create opportunities for deploying storage technologies that are still less competitive in the national grid.

Recent studies also highlight the importance of smart battery management through State-of-Health (SoH) estimation techniques. For example in [33,34,35,36,37] presents the use of the ProKnow-C methodology combined with machine learning and transfer learning to improve SoH prediction. These technologies have applications in smart grids, IoT systems, and wireless sensor networks, and are essential for extending battery life, improving efficiency, and supporting sustainable energy practices.

While various studies have addressed the application of renewable energy in off-grid microgrids, this work offers several distinctive contributions to the literature, particularly in the context of remote communities in the Brazilian Amazon.

First, this study is based on real data from a remote military facility in Tunui-Cachoeira, provided by ITAIPU Binacional. This allows the modeling and simulation to reflect an actual load profile, which improves the technical reliability and realism of the proposed system designs—an aspect not commonly found in similar studies [10,18].

Second, the work performs a comparative techno-economic analysis of three storage scenarios: (i) no battery, (ii) lithium-ion batteries, and (iii) lead–acid batteries. This structured comparison highlights not only the technical feasibility but also the cost-effectiveness of different storage technologies for isolated systems.

Third, the research is grounded in the specific challenges of the Amazon region, including diesel transport logistics, long distances, and limited infrastructure. These regional constraints directly influence system design and feasibility, making this study uniquely tailored to the reality of Amazonian microgrids.

Finally, the integration of energy balance methodology with HOMER Pro as a decision-support tool reinforces the reproducibility of the simulation and optimization approach, providing a structured framework that can be replicated or adapted for similar isolated environments [22,24]. To highlight the distinctiveness of this study,

Table 1. Comparative overview of related studies.

Ref.	Region	Storage Technology	Real Data?	Method	Contribution
[18]	Ecuador (Amazon)	Li-ion	No	Simulation	Long-term microgrid planning
[10]	Amazon (general)	PV + hydro	No	Feasibility study	Source diversification
[22]	Greece (island)	Thermal + electrical	No	System design	Dual-purpose microgrid
[23]	Off-grid (general)	Hydrogen	No	Techno-economic	Hydrogen storage application
[24]	Colombia	PV + batteries	No	HOMER	Industrial application
Paper	Brazil (Amazon)	None, Li-ion, lead–acid	Yes	HOMER + energy balance	Field data + regional focus

presents a comparison with related works from the last five years.

3. Materials and Methods

This section presents the methodology adopted for the sizing of an isolated microgrid located in the Brazilian Amazon, using solar photovoltaic energy as the primary generation source. The overall procedure applied in this study is summarized in Figure 3, which outlines the steps involved in system modeling, data collection, simulation, and performance evaluation.

To support the computational simulation in HOMER Pro and reinforce the methodological transparency of the study, the mathematical models used to represent the key system components are presented below. These equations are consistent with the modeling approach adopted by HOMER and are derived from its documentation and established literature [25,26,27].

The power output of the photovoltaic array, for instance, is modeled according to the irradiance and temperature conditions at each time step, as follows in Equation (1), where $P_{PV}(t)$ is the actual power output at time t (kW); $Y_{PV}$ is the rated capacity under standard test conditions (kW); $f_{PV}$ is the derating factor; $G_t$ is the solar irradiance on the array surface (W/m²); ${ G}_{STC}$ is the irradiance under STC, set at 1000 W/m²; ${ \alpha }_p$ is the temperature coefficient (%/°C); $T_C$ is the cell temperature (°C); and ${ T}_{STC}$ is the standard cell temperature (25 °C) [25,26,27].

$$P_{PV}(t)=Y_{PV} . f_{PV} . {\displaystyle \frac{{ G}_t}{{ G}_{STC}}} . [1+{ \alpha }_p\left(T_C-{ T}_{STC}\right)]$$

(1)

In turn, the fuel consumption of the diesel generator is represented by a linear model that accounts for both rated capacity and instantaneous generation, as expressed in Equation (2), where $F\left(t\right)$ is the fuel consumption at time ttt (L/h); $F_0$ is the fuel curve slope (L/kWh); $F_1$ is the intercept coefficient (L/h); $P_{gen}\left(t\right)$ is the output power at time ttt (kW); and $P_{rated}$ is the rated generator capacity (kW) [25,26,27].

$$F\left(t\right)=F_{0 .}{ P}_{gen}\left(t\right).+F_1 . P_{rated}$$

(2)

The state of charge (SOC) of the battery bank is another essential element of the system modeling. It is updated dynamically according to the energy flows during charging and discharging, as shown in Equation (3), where $SOC\left(t\right)$ is the state of charge at time t; $E_{ch}(t)$ and $E_{dis}(t)$ are the energy amounts charged and discharged (kWh), respectively; ${\mathrm{ƞ}}_{ch}$ and ${\mathrm{ƞ}}_{dis}$ represent the charge and discharge efficiencies; and $C_{bat}$ is the total capacity of the battery bank (kWh) [25,26,27].

$$SOC\left(t\right)=SOC\left(t-1\right)+{\displaystyle \frac{{\mathrm{ƞ}}_{ch.}{E}_{ch}(t)-\frac{E_{dis}(t)}{{\mathrm{ƞ}}_{dis}}}{C_{bat}}}$$

(3)

Finally, the power converter is modeled based on its efficiency in converting energy between alternating current (AC) and direct current (DC), as shown in Equation (4), where ${\mathrm{ƞ}}_{conv}$ is the conversion efficiency (typically ranging from 90% to 98%); $P_{in}$ is the input power (kW); and $P_{out}$ is the output power delivered to the system (kW) [25,26,27].

$$P_{out}={\mathrm{ƞ}}_{conv.}{ P}_{in}$$

(4)

To ensure consistency throughout the simulation horizon, the overall energy balance of the system is governed by Equation (5), which guarantees that energy supplied by the generation sources and storage matches the load demand, discounting system losses, as shown in Equation (5). This condition is fundamental to accurately represent the dynamic interactions among generation units, storage components, and load requirements. $E_{load}$ is the total energy demanded by the load (kWh); $E_{PV}$ is the energy supplied by the photovoltaic system (kWh); $E_{diesel}$ is the energy generated by the diesel generator (kWh); $E_{battery}$ is the net energy provided by the battery bank (kWh), considering charging and discharging cycles; and $E_{losses}$ represents the energy lost in the system due to inefficiencies, including conversion and transmission losses (kWh) [25,26,27].

$$E_{load}=E_{PV}+E_{diesel}+E_{battery}+E_{losses}$$

(5)

These mathematical formulations, implemented within the HOMER Pro^® simulation environment, form the basis for energy flow management and performance evaluation of the hybrid microgrid. Once the system components and their dynamic interactions are represented through Equations (1)–(5), the next step is the application of this modeling framework to a real-world context.

Accordingly, the present study focuses on a microgrid system located near the Army’s Special Border Platoon (PEF) in Tunuí-Cachoeira, in the municipality of São Gabriel da Cachoeira (Amazonas state). This site was chosen due to its representativeness as a remote and hard-to-reach community—characteristic of the Amazon region—where isolated microgrids can offer concrete benefits in terms of energy access and sustainability. Moreover, the project was supported by ITAIPU Binacional, which provided actual on-site electrical consumption data. This enabled the simulation to be carried out using a real and accurate load profile, which is essential for proper system sizing. It also addresses a major gap in the literature, as reliable databases for such communities are often not publicly available. It is worth noting that the system must operate 24 h a day to supply electrical energy to approximately 260 inhabitants—60 of whom are members of the Army platoon and 200 from the Baniwa indigenous community. Figure 4 illustrates the geographic location of the system installation.

Figure 4. The location of the system installation.

Figure 4. The location of the system installation.

The community and the platoon are powered by a system composed of a 178 kW photovoltaic generator, a 100 kW DG that produces energy in alternating current (AC), and a sodium battery bank that stores energy in direct current (DC). The photovoltaic panels are connected to inverters to convert DC into AC to supply the load, and the excess electrical energy is converted back into DC to be stored in the battery bank.

Due to problems in the sodium bank storage system, the system will be upgraded by replacing the battery bank with a new one while retaining the current photovoltaic generator and diesel DG.

Considering all the information described above, the main objective of this work is to size an isolated microgrid for a community located in the Brazilian Amazon using solar energy, which is abundantly available in the region, as a generation source. The Brazilian Amazon, which has a typically equatorial climate, is located less favorably in relation to the equator compared to the northeastern, central-western, and southeastern regions. As a result, its capture of solar radiation may be somewhat compromised due to some climatic and geographical characteristics [38]. However, despite these factors, its efficiency in generating electrical energy is not significantly affected, as its average solar radiation reaches levels of around 5.5 kWh/m², according to the Brazilian Solar Energy Atlas [39]. Additionally, it presents one of the higher rates of development for photovoltaic potential.

The secondary objectives are to size an isolated microgrid using lithium-ion batteries as a storage system, size an isolated microgrid using lead–acid batteries as a storage system, and conduct a technical and economic feasibility analysis of the different battery technologies selected to potentially replace the currently installed system. In the subsequent literature review, the reasons for selecting these two battery technologies will be justified.

The sizing of the system will be carried out considering three scenarios. Scenario 1 will consist of photovoltaic panels and a DG, assuming that the storage system cannot be replaced and must be removed from the installation. For the two following scenarios, sodium batteries will be replaced with lithium-ion and lead–acid batteries, respectively. Therefore, the proposed Scenario 2 will consist of photovoltaic panels, a lithium-ion battery bank as a storage system, and a DG as a backup system. Finally, Scenario 3 consists of photovoltaic panels, a lead–acid battery bank as a storage system, and a DG as a backup system. The diagram in Figure 5 shows the components of the system for the proposed scenarios.

To simulate and evaluate the technical and economic performance of each proposed configuration, the HOMER Pro microgrid software was adopted. This tool is widely used for the optimization of energy systems, both grid-connected and off-grid. It simplifies the evaluation of different configurations by performing system simulations, optimization routines, and sensitivity analyses across a range of technical and economic variables [37]. The software simulates all viable combinations of the equipment defined by the user, ranking and filtering them based on predefined criteria such as cost, performance, and reliability.

The optimization process identifies the most cost-effective solutions from a large number of possible configurations. Additionally, the sensitivity analysis module models the impact of variables beyond the user’s control—such as wind speed, fuel price, or demand variability—allowing designers to assess how these uncertainties affect the optimal system [40]. Because of these features, HOMER Pro is considered one of the most universal and practical tools for long-term microgrid planning. The sequence of steps followed in this study during system sizing is illustrated in Figure 6.

To carry out the simulations, several input variables were defined. These include optimization parameters, availability of energy resources, energy demand profiles, and component cost data. The general assumptions used in this study were as follows: the nominal discount rate is 13.75% and the expected inflation rate is 7.71%, as reported in [41]. The system was designed to supply 100% of the load, with a project lifespan of 25 years. The base price of diesel fuel in the State of Amazonas was considered to be 5.4 BRL/L, in accordance with [12]. Furthermore, a 10% logistical surcharge was added to the diesel cost, based on the methodology proposed by EPE [35]. For simplification, annual degradation of the photovoltaic modules was not considered in the simulation.

Among these input parameters, the availability of energy resources plays a fundamental role, as it directly influences the system’s energy production and operational performance [14]. In this case, solar energy is the primary generation source. Therefore, a site-specific analysis of the solar generation potential was conducted. By entering the geographical coordinates of the installation into HOMER Pro, the software calculated the solar resource profile and local temperature trends using data from the NASA Prediction of Worldwide Energy Resource (POWER) database. The results are illustrated in Figure 7a,b.

Following the solar resource assessment, the system’s electrical demand profile was defined based on field measurements collected at the installation site. Data was gathered between 28 June and 13 July 2022, and used to generate an annual load profile file in CSV format, which was then imported into HOMER Pro. The simulation software identified a daily variability of 14.13% and a time-step variability of 13.87%. The resulting daily and monthly consumption patterns are presented in Figure 7c. Based on these values, the average daily energy demand is 222.63 kWh, with a peak load of 16.83 kW. The month of June was identified as the period of highest consumption.

With the system demand and solar potential defined, the next step was to establish the technical and economic parameters for sizing the system components. For the current configuration, the photovoltaic generator must provide a capacity of 187 kW, while the backup diesel generator (DG) offers a nominal capacity of 100 kW. The economic parameters—including capital expenditure (CAPEX), replacement cost, and operation and maintenance (O&M) costs—were based on the methodology proposed by EPE [33], which defines both fixed and variable components for these values. Some CAPEX and O&M costs are common to any system configuration [35]. According to [33,38], the total CAPEX includes not only power generation units but also auxiliary systems, project development, land acquisition, civil works, substation construction, and electronic assembly. Fixed O&M covers expenses such as staff salaries and preventive maintenance [36,41].

Based on EPE’s cost breakdown, fixed CAPEX accounts for 65% of the total investment, fixed O&M for 5%, and the DG itself for the remaining 35%. Diesel generators are assumed to have a replacement cost equivalent to 60% of their initial value. For this study, the GE 140 FSX model from the MOSA brand was selected [35] and its technical and cost parameters are summarized in

Table 2. DG technical characteristics, fuel consumption, and economic parameters.

Technical Characteristic
Rated power (kVA/kW)	125/100
Useful life (h)	50,000
Minimum load (%)	20
Fuel consumption
100% Prime Rated Power (PRP)	26.30
75% PRP	20.10
75% PRP	11.50
Economic parameters
Total CAPEX (BRL)	415,994.28
Fixed CAPEX (BRL)	270,396.28
Generator cost (BRL)	145,598.00
Replacement cost (BRL)	87,358.80
Fixed O&M cost (BRL)	13,519.81
Variable O&M cost (BRL/op. h)	3.24
Diesel cost (BRL/L) + 10%	6

.

The photovoltaic subsystem includes modules, support structures, charge controllers, and freight to the site. According to [32], O&M expenses for PV systems are approximately 1% of their CAPEX. Current installation costs range between BRL 4000 and BRL 4500 per kW [32,42], and, for this study, the upper value of BRL 4500/kW was adopted. The panel selected was the RS6C-P model from RESUN, with technical specifications shown in

Table 3. Technical and economic parameters for sizing the photovoltaic generator.

Characteristic [31]
Rated power (W)	26 000
Maximum power (W)	19 200
Maximum power voltage (V)	30.30
Full power current (A)	8.58
Open circuit voltage (V)	37.70
Short circuit current (A)	8.95
Rated efficiency (%)	16.10
Temperature coefficient (%/°C)	−0.39
Normal cell operating temperature (°C)	4500
Useful life (years)	2500
Annual degradation (%)	0.50
Economics parameters
Cost per photovoltaic panel (BRL)	650.07
CAPEX (BRL/kW)	4500
Replacement cost (BRL/kW)	4500
O&M cos (BRL/kW)	45

[43].

The battery system CAPEX includes purchase, transportation, and installation. However, to focus the economic sensitivity analysis, only the purchase cost was used as the sizing input, assuming equivalent logistics for both battery technologies. The lead–acid battery selected was the Moura Solar 12MS234 (Pernambuco City, Brazil) (220 Ah), a model specifically designed for small and medium photovoltaic systems. These batteries tolerate high temperatures, operate in 12 V to 48 V systems, and require low maintenance [44]. The lithium-ion battery used in this study is a LiFePO4 model from Energy Source, known for durability and high efficiency under variable thermal conditions. Its depth of discharge reaches up to 90%, and both battery types are characterized in

Table 4. Technical and economic sizing parameters for Scenario 1 and Scenario 2.

Indicators	Lead–Acid Battery	Lithium-Ion Battery
Rated capacity (Ah)	220
Rated capacity (kWh)	2.64	10
Voltage (V)	12	48
Maximum number of cycles	1500	4000
Discharge depth (%)	50	90
CAPEX (BRL)	1580.07	22,489
Replacement cost (BRL)	1580.07	22,489
O&M cost (BRL)	15.80	224.49

.

The controller component in the HOMER Pro software allows specifying the system’s operation during simulation [42]. Each controller type uses a unique control algorithm or dispatch strategy. According to [43], a dispatch strategy can be defined as follows: a set of rules used to control the operation of the generator and energy storage system whenever there is not enough renewable energy to meet the load. Considering the components of the proposed system, the dispatch strategies to be analyzed are as follows [28,40,44,45,46,47,48,49]:

• Cycle Charging (CC): Under the cycle charging strategy, each time a generator is required, it operates at full capacity, and the surplus energy charges the battery bank. Cycle charging tends to be optimal in systems with little or no renewable energy.
• Load Following Strategy (LF): Under the load following strategy, when a generator is needed, it produces only enough energy to meet the demand. Load following tends to be optimal in systems with a lot of renewable energy that sometimes exceeds the load.
• Combined Dispatch (CD): The combined dispatch strategy can improve performance compared to cycle charging and load following strategies by making more efficient use of the generator [43].
• Predictive Dispatch (PS): Under the predictive dispatch strategy, the dispatch algorithm knows the upcoming electrical and thermal demand, as well as the availability of solar and wind resources. It often produces results with lower system operating costs compared to other dispatch strategies in HOMER Pro software.
• Evaluation of costs and energy production: Once all the input variables are defined, the simulation and optimization of the system are initiated. If the load demand is not met, components must be resized; otherwise, the costs and energy production of the proposed scenarios need to be evaluated. The costs to be analyzed include the total net present cost (NPC), the levelized cost of energy (LCOE), and the initial investment cost. Other system variables that will also be analyzed include the renewable fraction, diesel consumption, and excess electricity.

The NPC of a system is the present value of all the costs incurred by the system during its lifetime, minus the present value of all the revenues it receives during its lifetime [48]. Costs include CAPEX costs, replacement costs, O&M costs, fuel costs, emissions penalties, and the costs of purchasing energy from the grid [28,40]. Revenues include the residual value and revenues from grid sales [43]. HOMER calculates the total NPC by summing the total discounted cash flows for each year of the project’s life [44,45,46,47,48,49,50]. The total NPC is the primary economic result of HOMER, the value by which it ranks all system configurations in the optimization results, and the basis from which it calculates the total annualized cost and the LCOE, as showed in Equation (6) [43], where $CRF:$ capital recovery factor, $C_{ann,Total}:$ total annualized cost, and ${NPC}_{Total}:$ total net present cost.

$${NPC}_{Total}={\displaystyle \frac{C_{ann,Total}}{CRF}}$$

(6)

HOMER software defines LCOE as the average cost per kWh of useful electrical energy produced by the system [48]. To calculate the LCOE, HOMER divides the annualized cost of producing electricity, which is the total annualized cost minus the cost of meeting the thermal load, by the total electric load served, using Equation (7), where $C_{annual,total}:$ total annual system cost (BRL/year), $C_{boiler}:$ boiler marginal cost (BRL/kWh), $H_{served}:$ total thermal load served (kWh/year), and $E_{served}:$ total electric load served (kWh/year).

$$LCOE={\displaystyle \frac{C_{annual,total }-C_{boiler}{H}_{served}}{E_{served}}}$$

(7)

The initial investment cost is the sum of all costs incurred at the beginning of the project. In this work, this variable consists of the initial costs of photovoltaic panels, DG, batteries, converters, and other things [40]. The renewable fraction is the fraction of the energy delivered to the load that originated from renewable energy sources. The software calculates the renewable fraction using Equation (8) [40], where $E_{non-renowable}:$ non-renewable electrical production (kWh/year), $H_{no-renowable}:$ non-renewable thermal production (kWh/year), $E_{served}:$ total electric load served (kWh/year), and $H_{served}:$ total thermal load served (kWh/year).

$$f_{renowable}=1-{\displaystyle \frac{E_{non-renowable}+H_{non-renowable}}{E_{served}+H_{served}}}$$

(8)

The diesel consumption is the annual cost of feeding the generator. HOMER calculates this value by multiplying the fuel price by the amount of fuel used by the generator in a year [49].

The excess electricity is surplus electrical energy that must be discarded because it cannot be used to power a load or charge batteries. Excess electricity occurs when there is an energy surplus, and the batteries cannot absorb it all [40]. The excess electricity fraction, which is the ratio of total excess electricity to total electrical production, is calculated by HOMER software at the end of each simulation using Equation (9), where $E_{excess}:$ total excess electricity (kWh/year) and $E_{production}:$ total electrical energy production (kWh/year).

$$f_{excess}={\displaystyle \frac{E_{excess}}{E_{production}}}$$

(9)

4. Results and Discussion

This section presents the results obtained from the simulations performed using the HOMER Pro^® platform, aiming to evaluate the technical and economic viability of different microgrid configurations for the isolated community of Tunuí-Cachoeira, located in the Brazilian Amazon. The analyses are based on real load data collected in the field and consider three scenarios with variations in the energy storage strategy: (i) a system without batteries, relying solely on photovoltaic generation and a diesel generator; (ii) a hybrid system with lithium-ion battery storage; and (iii) a hybrid system with lead–acid battery storage.

The results are discussed in an integrated and comparative manner, considering performance indicators such as energy production, diesel consumption, renewable fraction, system losses, investment and operational costs, and sustainability metrics. The discussion also includes a sensitivity analysis to support technology selection for real-world deployment, given the logistical constraints and economic conditions of the Amazon region.

The assumptions adopted, the computational models used, and the definition of scenarios underpinning this comparative analysis are based on established methodologies. For the analytical modeling, the research adopts as a reference the established models presented in [47], which serve as the analytical basis for structuring the computational models used for simulation and optimization of the proposed systems. The simulations were carried out using the computational tool HOMER Pro^® (Hybrid Optimization of Multiple Energy Resources from NREL) [40], widely used in studies on microgrids and isolated hybrid systems [24,28,51,52].

The computational modeling considered the following elements: (i) real electric load profile, collected from 28 June to 13 July 2022, in the locality of Tunuí-Cachoeira, municipality of São Gabriel da Cachoeira—AM, providing an empirical database representative of the annual demand; (ii) solar resource data obtained from NASA’s Prediction of Worldwide Energy Resources (POWER) database, using geographic coordinates specific to the installation site; (iii) a simulation horizon of 25 years, with a real discount rate of 13.75% and an estimated inflation rate of 7.71%; (iv) a base diesel price of BRL 6.00/L (including a 10% logistical surcharge).

The modeling applied is based on an hourly energy balance, optimizing the joint operation of renewable sources, thermal generator, and storage system according to dispatch strategies defined in the software (load following, cycle charging, and combined dispatch).

For the configuration and analysis, three scenarios were defined to represent different strategies for energy storage and solar energy utilization, tailored to the reality of the application under study.

Table 5. Characterization of the three simulated scenarios in HOMER Pro.

Element	Scenario 1	Scenario 2	Scenario 3
Solar Generation (PV)	187 kW	187 kW	187 kW
Diesel Generator (DG)	100 kW	100 kW	100 kW
Energy Storage	—	22 LiFePO4 batteries	192 Lead–Acid batteries
Converters	—	27.3 kW	33.2 kW
Dispatch Strategy	Load Following (LF)	Combined Dispatch	Combined Dispatch
Load Data	Real	Real	Real
Project Horizon	25 years	25 years	25 years
Diesel Price	BRL 6.00/L	BRL 6.00/L	BRL 6.00/L

presents the comparative characterization of the systems analyzed. The simulation was conducted in HOMER Pro^® using real load data collected between 28 June and 13 July 2022, from the Tunuí-Cachoeira community (AM). The modeling includes 187 kW PV generation, a 100 kW diesel generator, a 25-year simulation horizon, a 13.75% discount rate, and a diesel price of BRL 6.00/L.

4.1. Technical and Economic Assessment of the Scenarios

This study aims to evaluate the technical and economic viability of alternative hybrid microgrid configurations based on photovoltaic solar energy for a remote community in the Brazilian Amazon. To this end, the results of the simulations for the three proposed scenarios (Figure 5) are presented, assessing in an integrated manner the technical, economic, and operational aspects of the isolated photovoltaic microgrid located in the community of Tunuí-Cachoeira (AM). The comparative analysis enables the identification of the configuration with the best overall performance, based on data obtained using the HOMER Pro^® software, considering the actual electric load profile and the assumptions previously described (Table 6).

In terms of technical configuration, all simulated scenarios adopt a photovoltaic capacity of 187 kW and a 100 kW diesel generator as a backup source. The main difference among them lies in the presence and type of energy storage. Scenario 1 does not use batteries, operating exclusively with instantaneous solar generation and diesel generator support. Scenario 2 incorporates 22 lithium-ion (LiFePO₄) batteries, characterized by high efficiency, 90% depth of discharge, and long lifespan. Scenario 3 employs 192 lead–acid batteries (Moura Solar 12MS234), with a depth of discharge of 50%, requiring more physical space and more frequent maintenance.

Beyond the differences in storage systems, there are also variations in the converter capacities required for each scenario. Due to the number and characteristics of the storage devices, the scenario with lithium-ion batteries requires converters rated at 27.3 kW, while the lead–acid battery scenario requires 33.2 kW converters. These specifications directly influence the sizing of auxiliary components and the complexity of the system’s installation.

Regarding energy performance, although all scenarios exhibit the same estimated annual photovoltaic energy production (approximately 215,800 kWh/year), the way this energy is utilized varies significantly. Scenario 1 shows a high level of energy waste, with 69.8% of the solar energy being discarded due to the absence of storage. In contrast, the scenarios with batteries exhibit a considerable reduction in surplus: 57.4% in Scenario 2 and 53.9% in Scenario 3, indicating more efficient utilization of the solar resource.

The operation of the diesel generator also differs substantially among scenarios. Scenario 1 requires approximately 21,796 L of diesel per year, reflecting intensive generator use throughout the day. In contrast, Scenario 2 reduces this consumption to 1488 L per year, while Scenario 3 records the lowest consumption of all, with just 1056 L annually. This reduction is directly attributed to the superior performance of the lead–acid battery storage system, which operates with a state of charge (SOC) above 90% for more than 90% of the time, ensuring greater stability and autonomy for the microgrid.

Table 6. The main economic results obtained for the three scenarios.

Indicator	Scenario 1	Scenario 2	Scenario 3
Net Present Cost—NPC (BRL)	3,570,000	2,330,000	2,140,000
Levelized Cost of Energy—LCOE (BRL/kWh)	3.31	2.16	1.99
Initial Investment (BRL)	1,260,000	1,780,000	1,590,000
Diesel Consumption (L/year)	21,796	1488	1056
Excess Energy (%)	69.8%	57.4%	53.9%

Table 6. The main economic results obtained for the three scenarios.

Indicator	Scenario 1	Scenario 2	Scenario 3
Net Present Cost—NPC (BRL)	3,570,000	2,330,000	2,140,000
Levelized Cost of Energy—LCOE (BRL/kWh)	3.31	2.16	1.99
Initial Investment (BRL)	1,260,000	1,780,000	1,590,000
Diesel Consumption (L/year)	21,796	1488	1056
Excess Energy (%)	69.8%	57.4%	53.9%

Analyzing the data presented in Table 2, from an economic standpoint, the results also reinforce the feasibility of the configuration using lead–acid batteries. Despite the need for a larger number of batteries and the requirement for periodic maintenance, Scenario 3 exhibited the lowest total life cycle cost (NPC) and the lowest levelized cost of energy (LCOE), highlighting its economic superiority in the long term. Scenario 2, which uses lithium-ion batteries, emerges as a technically viable and environmentally promising alternative, albeit with slightly higher costs. Meanwhile, Scenario 1, although requiring the lowest initial investment, proved economically unviable over time due to its high dependence on the diesel generator and associated operational costs.

In addition to the economic aspects, the sustainability of the scenarios was also significantly influenced by the presence of storage systems. The renewable fraction of the energy supplied to the load increased substantially with the inclusion of batteries. In Scenario 1, only 34.7% of the consumed energy comes from renewable sources, whereas in Scenarios 2 and 3 this percentage reaches 95.3% and 96.8%, respectively. This shift has a direct impact on reducing the emissions associated with diesel use, representing a major step forward toward energy sustainability. The use of batteries also allows for decoupling generation from consumption, ensuring a nighttime supply from solar energy stored during the day.

From a technical–operational perspective, the results also demonstrate the robustness of the configuration with lead–acid batteries. Although it requires more physical space and maintenance, the system exhibited excellent performance, with the lowest diesel consumption among the scenarios, generator operation limited to just 236 h per year, and the ability to supply the load for up to 12 continuous hours solely with stored energy. The combined dispatch strategy proved effective, alternating between cycle charging and load following based on operational conditions, thereby optimizing both cost and system efficiency. Scenario 2, with lithium-ion batteries, while technically feasible and promising, still requires acquisition cost reductions to become more economically competitive compared to the lead–acid battery technology—as will be further explored in the sensitivity analysis.

4.2. Sensitivity and Applicability Analysis

Given the real and applied nature of this study, decisions regarding which technologies to adopt cannot be based solely on isolated economic indicators. They must also consider factors such as technological maturity, local availability, maintenance complexity, and transportation and replacement logistics—especially critical aspects in hard-to-reach regions like the interior of the Amazon. In this context, it becomes essential to assess the thresholds and conditions that could, in the future, change the relative attractiveness among the analyzed scenarios, particularly in the case of lithium-ion batteries, which emerge as a promising alternative for future applications.

To this end, a sensitivity analysis was conducted by exploring three critical parameters through simulations in the HOMER Pro^® software (Figure 8). The first parameter analyzed was the price of diesel fuel, which varied from BRL 2.00/L to BRL 50.00/L, considering scenarios of steep increases associated with regional logistical challenges. The second parameter was the acquisition cost of lithium-ion batteries, with simulated variations from −50% to +50% relative to the current reference value. Finally, the third parameter evaluated was the influence of the lithium-ion battery lifespan, with variations of ±20% in the number of cycles.

It is important to note that the base diesel price adopted in the main simulations was BRL 6.00/L, already incorporating average logistical surcharges. However, it is well known that this value can be significantly higher in real-world situations. For instance, diesel must often be transported via river barges, with trips lasting up to 10 days, and is subject to shortages, seasonality, and additional costs related to storage and security.

The results of the analysis indicate that, with an estimated useful life of 40,000 kWh per battery, it would be necessary for the cost of lithium-ion batteries to decrease by approximately 30% for Scenario 2 (with this technology) to become more competitive than Scenario 3, which uses lead–acid batteries. In the case of a 20% reduction in lifespan (32,000 kWh), the price of lithium-ion batteries would need to fall by up to 40% to make replacement viable. Conversely, if the lifespan increased by 20% (48,000 kWh), the required price reduction would be smaller—around 20% below the current value.

Even under an extreme scenario in which diesel prices reach BRL 50.00/L, the system using lead–acid batteries remains the better choice if lithium-ion battery prices do not show significant reductions. These findings demonstrate that, under current conditions, the system with lead–acid batteries is more suitable for immediate implementation, primarily due to its balanced cost, robustness, and simplified maintenance requirements.

Figure 8. Sensitivity analysis for useful life of lithium-ion batteries equal to (a) 40,000 kWh, (b) 32,000 kWh, and (c) 48,000 kWh.

Figure 8. Sensitivity analysis for useful life of lithium-ion batteries equal to (a) 40,000 kWh, (b) 32,000 kWh, and (c) 48,000 kWh.

4.3. Final Considerations and Future Perspectives

The integrated analysis of the three proposed scenarios for the isolated solar microgrid of Tunuí-Cachoeira allowed the assessment of not only economic and energy aspects but also the practical constraints that directly impact deployment in remote regions of the Brazilian Amazon.

It is noteworthy that the scenarios are modeled based on real conditions, with load data measured in the field and within a challenging logistical context. The scenario with lead–acid batteries emerged as the most balanced and viable alternative for immediate implementation. Although it requires more physical space and maintenance, its cost–benefit ratio, technological robustness, operational simplicity, and wide commercial availability outweigh these limitations.

The proposed solution is capable of operating with a high renewable energy penetration (96.8%), reducing the annual diesel consumption to only 1056 L, while still ensuring continuous power supply for up to 12 h using storage alone. As a result, there is a concrete expectation of reduced operational costs, increased local energy independence, and improved quality of life for the served community.

On the other hand, Scenario 2 with lithium-ion batteries stands out as a technically competitive and environmentally superior alternative in the long term. The sensitivity analysis results indicate that reductions in acquisition costs and/or increases in battery lifespan would make this technology more attractive. Moreover, the lower number of units required, higher energy density, and reduced maintenance needs make this solution promising for future system upgrades.

Scenario 1, without storage, despite presenting the lowest initial investment, proved economically unsustainable over the project’s lifetime, with a high levelized cost of energy (LCOE of BRL 3.31/kWh), critical diesel dependency, and a low renewable fraction (34.7%).

Given that this is a project with real implementation prospects, the technological choice made at this stage considers not only the quantitative simulation results but also qualitative factors related to operation, technical support, and regional logistical availability. The solution based on lead–acid batteries aligns with these criteria, representing a conscious choice for the present without disregarding the technological trends that will shape the future of the decentralized energy sector.

The selection of this technology considers not only its current techno-economic superiority but also practical criteria that are highly relevant to the Amazonian context: wide commercial availability and a consolidated maintenance chain, low complexity for installation and operation, and reduced risk of obsolescence or warranty loss due to severe environmental conditions. However, the results from the scenario with lithium-ion batteries reveal a significant strategic potential for future system updates. This transition could become viable if the technology becomes more affordable due to the expansion of electric mobility, the implementation of reverse logistics or battery repurposing (second life), and regulatory improvements that promote clean and long-lasting solutions.

In this regard, the long-term strategy may include a planned transition to lithium-ion batteries or even to emerging technologies such as flow batteries, hybrid arrangements with supercapacitors, or hydrogen-based systems, depending on the actual performance of the installed system and the evolution of the market.

5. Conclusions

Supplying electrical energy to remote regions of the country through the National Interconnected System is very complicated. This is due to a series of problems, including their geographical location. The difficult access to these localities makes it nearly impossible to supply electricity in a conventional manner; therefore, other options must be explored to solve this problem. One of these solutions is the implementation of decentralized microgrids, which allow small power generators to be installed close to the point of consumption. Frequently, due to its ease of installation and a range of benefits, the technology most commonly used for isolated systems is diesel generators. However, on the downside, the price of the fuel used is high and continues to increase year after year, making it necessary to explore other alternatives.

One solution to this problem is the incorporation of renewable energy generators, such as photovoltaic systems. In this work, the sizing of a hybrid system for an isolated community located in the Brazilian Amazon was carried out using the HOMER Pro computational tool. Three possible scenarios for supplying electricity to the locality were analyzed. Different parameters obtained from the results of each energy storage technology were analyzed to define and select the most suitable technology for the system.

Considering the proposed assumptions and the technical and economic sizing parameters, the simulation and optimization results showed that, when diesel is priced at BRL 6.00/L, the most suitable technology for the system is lead–acid batteries. However, in the remote regions of the Brazilian Amazon, the price of diesel fuel experiences a significant increase due to the difficulties in transporting fuel to these locations. Therefore, a sensitivity analysis was conducted to analyze the influence of parameters on the selection of the scenario to be used by the system. The variables analyzed included the price of diesel fuel, the price of lithium-ion batteries, and the lifespan of lithium-ion batteries.

This study stands out by using real field consumption data from an isolated Amazonian community, which enhances the accuracy and relevance of the simulations. This study presented a techno-economic analysis of three isolated solar microgrid configurations to serve the community of Tunuí-Cachoeira, located in a remote region of the Brazilian Amazon. The investigation used real electric demand data and was conducted based on the HOMER Pro^® simulation and optimization tool, considering scenarios with and without storage systems.

The main results demonstrated the following: (i) The scenario with lead–acid batteries presented the best techno-economic performance under current conditions, with the lowest net present cost (BRL 2.14 million), the lowest LCOE (BRL 1.99/kWh), and the highest renewable fraction (96.8%), in addition to high efficiency in solar energy use. (ii) The scenario with lithium-ion batteries proved to be technically viable and environmentally promising, especially for future applications, depending on the reduction in acquisition costs and/or increase in lifespan. (iii) Scenario 1, without batteries, proved inadequate for the local reality, with critical dependence on diesel, high operational costs, and low energy efficiency.

The choice of Scenario 3 with lead–acid batteries for immediate implementation was supported not only by economic factors but mainly by its compatibility with the logistics, maintenance, and robustness required for applications in remote areas of the Amazon. From the sensitivity analysis, it was observed that, for high diesel fuel costs, lithium-ion batteries need to reduce their current cost by approximately 30–40% across all three battery lifespan options. Therefore, it was concluded that, for high diesel fuel costs, the technology that offers the best technical and economic parameters for the system is lead–acid batteries.

In addition to providing an effective energy solution for the local community, based on real data and considering practical constraints and field deployment challenges, the analytical framework adopted can be replicated for other isolated and/or sustainable rural electrification projects.

Suggestions for future work include the following: (i) Analyze the influence of increasing annual load on microgrid sizing. (ii) Develop projects in microgrids that use other renewable generation sources, such as wind and biomass, among others. (iii) Conduct a technical and economic feasibility analysis of microgrids that use other energy storage technologies, such as hydrogen-based energy storage. (iv) Analyze the possibility of connecting isolated microgrids to the grid in the future. (v) Size microgrids that have both electrical and thermal loads to be supplied.

Finally, this study represents a step toward the universalization of access to quality energy based on renewable sources, in line with the Sustainable Development Goals (SDGs 7, 13, and 15), and can serve as a technical reference for public policies, funding agencies, and implementers of energy transition projects in isolated contexts.