A Review of Assessment Indicators and Methods for Rural Energy Systems

Abstract

This study presents a systematic bibliometric analysis and critical review of assessment indicators and multi-criteria decision-making methods for rural energy systems from 2010 to 2025. It examines the evolving definitions and regional variations in these indicators and methods. The research hotspots of rural energy systems have shifted from basic rural electrification to multi-dimensional assessment indicators and hybrid multi-criteria decision-making methods. The assessment indicators for rural energy systems demonstrate a marked imbalance, dominated by economic and technical dimensions. Specifically, economic evaluations for rural energy systems frequently utilize net present cost and levelized energy cost, shifting from static capital comparisons to comprehensive lifecycle assessments. Meanwhile, loss of power supply probability is identified as the primary inherent constraint among technical assessment indicators for rural energy systems. Geographically, assessment indicators for rural energy systems priorities exhibit significant divergence. Developing regions prioritize basic power supply and affordability, whereas developed regions focus on grid stability and market risk resilience. In addition, environmental evaluations for rural energy systems remain fixated on carbon emissions. Developed nations emphasize global climate benefits, while developing nations focus on localized dividends like indoor air quality improvement. Critically, despite an increasing focus on rural livelihoods, social indicators remain systematically marginalized in rural energy systems, leading to the neglect of local requirements and increasing technical risks. The field of rural energy system assessment is advancing toward multi-criteria decision-making indicators. Future methodologies must integrate robust, dynamic adaptive mechanisms that respond to evolving developmental priorities in order to effectively address inherent data scarcity and complex socio-economic uncertainties of rural energy systems.

1. Introduction

Rural energy systems are essential for achieving the United Nations’ Sustainable Development Goals, notably SDG 7. As the global energy transition accelerates, the focus of rural electrification has shifted from basic accessibility to sustainability, reliability, and high-quality service. It is necessary to introduce scientifically rigorous evaluation indicators. Previous research has tracked the evolution of these assessments from static costs to multi-dimensional indicators. Mandelli et al. [1] defined the critical evaluation boundaries for decentralized systems. Early assessment paradigms were dominated by techno-economic indicators such as levelized cost of energy (LCOE) and lifecycle cost, as firmly established by Chaurey & Kandpal [2]. Subsequent studies integrated technical reliability and stability indicators [3,4] to constrain these economic evaluations [5]. In these contexts, indicators were often treated merely as objective functions for meta-heuristic algorithms [6,7] rather than subjects of critical inquiry. In recent literature, the scope of indicators has expanded significantly. Odoi-Yorke et al. [8] highlighted a crucial thematic shift towards identifying socio-economic impacts. However, a systematic critique of how these diverse indicators interact within a unified framework is still lacking.

A critical examination of the existing body of reviews reveals persistent gaps. First, the landscape of indicator systems remains fragmented, heavily prioritizing techno-economic indicators while relegating environmental and social indicators to secondary roles without scrutinizing their mathematical evolution. Second, existing surveys frequently generalize findings across global contexts, thereby failing to capture significant regional heterogeneity and distinct local cost drivers. Third, methodological disconnects remain within multi-criteria decision-making (MCDM) methods, particularly the under-explored transition from subjective weighting to data-driven, adaptive methods. Finally, many foundational assessments rely on static cost assumptions that fail to reflect dynamic economic viability and actual market fluctuations.

To address these gaps, a systematic bibliometric analysis and critical review of evaluation indicators for rural energy systems from 2010 to 2025 were conducted using CiteSpace. Unlike previous reviews that focus primarily on generation systems, this study centered on the assessment indicators themselves. It systematically categorized indicators into economic, technical, environmental, and social dimensions. The mathematical models and regional applications were analyzed subsequently. Additionally, it critically examined the advancement of hybrid weighting methods in MCDM, highlighting the transition from subjective judgments to data-driven, adaptive decision-making. The research framework is illustrated in Figure 1.

2. Bibliometric Analysis

Bibliometrics analyzes knowledge networks within published literature. It is widely applied to examine the structure and progression of research fields, including rural energy systems assessment and sustainable energy planning. This section introduces the bibliometric software tools (CiteSpace 6.3.R1) and literature acquisition methods. Subsequently, multidimensional analysis results are presented to identify academic trends and core issues within the rural energy systems domain.

2.1. Methods

Currently, several bibliometric visualization tools are available, such as Pajek (1996), CiteSpace (1999), SciTool (2009), VOSviewer (2010), and CitNetExplorer (2017). Among these, CiteSpace stands out, particularly in fields related to ecology, engineering, computer science, and environmental science. It is particularly effective for “rural energy systems assessment and optimization modeling” by enabling the analysis of knowledge structures and emerging trends. CiteSpace (6.3.R1), developed by Dr. Chen at Drexel University, Philadelphia, PA, USA, was employed to analyze the literature data. The analysis covers four dimensions: publication trends, collaboration networks, keyword hotspots, and citation relationships to reveal the evolutionary trajectory and knowledge landscape of rural energy systems assessment.

Data were retrieved from the Web of Science Core Collection, comprising the Science Citation Index Expanded (SCI-E) and Social Sciences Citation Index (SSCI). Initial search terms included “rural”, “energy”, and “assessment”. Subsequently, “village” and word variations were added to prevent omissions, such as “evaluate” and “evaluation”. The final search query employed was: TS = (“energy”) and TS = (“rural” or “village”) and TS = (assessm* or evaluat*).

Both peer-reviewed journal articles and prestigious conference proceedings (e.g., ASCE and Elsevier) were included to broaden the analysis scope. Initially, 4908 entries were retrieved. The PRISMA framework was used to screen and select the literature, with the workflow illustrated in Figure 2. Ultimately, 315 research articles were selected as of 1 August 2025.

Given the rapid growth of relevant literature since 2010, the study period was established as 2010–2025. Subsequently, the downloaded data records were imported into CiteSpace software (CiteSpace 6.3.R1), with the time slice set to one year. Node types were defined as institutions, authors, keywords, cited authors, and cited journals to enable network visualization.

2.2. Publication Overview

Figure 3 illustrates the number of publications from 2010 to 2025. The field exhibits a pronounced upward trend, divided into two distinct phases. The first phase (2010–2015) represents a nascent stage characterized by slow growth. The second phase (2016–2025) witnessed rapid development, with over 80 research papers published annually. Notably, publications after 2016 account for 86% of the total. The growth rate indicates increasing attention to the field. Driven by technological advancements, the publication volume is projected to sustain the rapid upward trajectory.

2.3. Co-Author Analysis

Collaborative relationships among countries, institutions, and authors were examined at both macro and micro levels.

2.3.1. Network of Co-Authors’ Institutions

The analysis reveals that the top 9 institutions by publication volume have collectively published 208 articles. The Chinese Academy of Sciences ranks first with 34 publications, followed by the Indian Institute of Technology System with 32 publications. Other top institutions include the Universitat Politècnica de Catalunya (Spain, 26), Egyptian Knowledge Bank (EKB, Egypt, 23), Polytechnic University of Milan (Italy, 22), University of Tehran (Iran, 21), Tsinghua University (China, 18), North China Electric Power University (China, 17), Lawrence Berkeley National Laboratory (USA, 15), and China University of Geosciences (China, 14). Figure 4 depicts 415 nodes and 447 links, reflecting widespread cooperation, particularly within China. To quantitatively evaluate this structure, overall network metrics reveal a remarkably low density of 0.0052, alongside a modularity of 0.9756 and a mean silhouette score of 1.0. These metrics indicate that while localized clusters have a high degree of cohesion, the global network remains highly fragmented. Furthermore, the centrality distribution is heavily skewed. While most institutions possess scores near zero, the Chinese Academy of Sciences acts as a primary structural bridge with a centrality value of 0.13, indicating its influential role among research organizations. However, institutions from other countries show fewer interconnections, highlighting a need for further international cooperation.

2.3.2. Co-Authorship Network

In the author collaboration map, there are 490 authors and 295 connections. As shown in Figure 5, the scholar with the highest number of publications is Tomonobu Senjyu (5 publications), followed by Emanuela Colombo (4 publications). Other notable authors include Syed Ali Abbas Kazmi, Laia Ferrer-Marti, Abdullah Altamimi, L. Ferrer-Marti, Shaorong Wang, Ramchandra Bhandari, and Mohammad Amin Vaziri Rad. Similar to the institutional network, researchers in China demonstrate strong collaborative ties, centered around core authors such as An and Hong. Conversely, some authors, such as Erik O. Ahlgren, have published numerous papers but engaged in less external collaboration.

2.4. Co-Word Analysis

Keyword network mapping was employed to visually represent the co-occurrence relationships and association strengths of phrases extracted from abstracts. Additionally, a timeline view is constructed to observe the developmental trends of topics. Connecting lines represent associative relationships between keywords, while the size of each node and label reflects the keyword’s frequency.

2.4.1. Network of Co-Occurring Keywords

Figure 6 visually represents 410 nodes and 2996 links in the keyword network, with a density of 0.0357. Furthermore, the overall network modularity of 0.9756 and weighted mean silhouette score of 1.0 confirm that the keyword clustering is both highly reliable and structurally coherent. High-frequency keywords primarily include “rural electrification” (44), “renewable energy” (41), “energy” (38), “optimization” (29), “design” (27), “systems” (28), “life cycle assessment” (24), and “performance” (19). Beyond mere frequency, “optimization” exhibits a notably high betweenness centrality, acting as a pivotal bridge linking fundamental energy models with economic constraints. This demonstrates that researchers in this field focus heavily on optimization methods and rural energy systems performance analysis.

2.4.2. Network of Co-Occurring Hotspots

Figure 7 depicts the network from a timeline perspective (modularity is 0.4169, and mean silhouette is 0.7328). During the period of 2010–2015, research hotspots included “rural area”, “renewable energy”, and “rural electrification”, with early structural bursts predominantly observed for foundational keywords like “developing countries”, “sensitivity analysis”, and “sustainable development”. Between 2015 and 2020, the focus shifted towards specific techno-economic parameters, characterized by strong citation bursts in “net present cost (NPC)”, “diesel generators”, “techno-economic analysis”, “energy transition”, and “hybrid energy system”. After 2020, hotspots became numerous but lower in frequency, with recent emergent bursts concentrated around “energy cost” and “sustainable development goals”, indicating a trend towards diversification and refinement in rural energy systems assessment research. The distribution of major clusters along the time axis suggests that renewable energy and rural electrification are enduring topics. In contrast, the topic of sustainable development has gained prominence in recent years, suggesting a research shift from qualitative analysis to quantitative assessment.

2.5. Co-Citation Analysis

Co-citation analysis reveals the developmental trajectories of knowledge connections within relevant research fields. Citation frequency serves as an indicator for measuring the influence of journals, researchers, and publications. The most influential publications and prominent journals were identified.

2.5.1. Document Co-Citation Network

The co-citation network identified 834 document nodes and 2728 links, as illustrated in Figure 8. Node size denotes citation frequency, with larger nodes indicating greater relevance to the topic of rural energy system assessment. It is evident that certain researchers in the co-authorship network, such as Odou O.D.T. and Li J.Z., exert significant influence. The most frequently cited article was published in Renewable Energy in 2020, with 50 citations within the database.

2.5.2. Author Co-Citation Network

From the author’s co-citation analysis (Figure 9), publications with anonymous authors (typically representing organizations or reports) have the highest citation frequency (1149 times). Hong ranks second with 133 citations. Yan ranks third (132 citations), followed by ASHRAE (125), Doca (101), Haldi (27), Wang (87), Gunay (81), Dong (67), and Kim (66).

2.5.3. Journal Co-Citation Network

Figure 10 distinctly shows the co-cited journals in rural energy systems assessment. Renewable and Sustainable Energy Reviews has the highest number of citations (180), followed by Renewable Energy (160) and Energy (128). Energy Policy ranks fourth with 110 citations. Other notable journals include Applied Energy, Journal of Cleaner Production, Energies, Energy for Sustainable Development, Energy Conversion and Management, Sustainability, and Solar Energy. These represent the most widely read and cited periodicals in the subject area.

2.6. Distribution Characteristics of Evaluation Dimensions and Indicators

Based on the bibliometric statistics, the distribution of evaluation dimensions (Figure 11) delineates the macroscopic framework of current rural energy assessments. The framework is primarily composed of four dimensions: technical, economic, environmental, and social. Among these four dimensions, the economic dimension constitutes the largest proportion of the research focus, while the social dimension accounts for the smallest.

Figure 12 further reveals the microscopic priorities through the selection frequency of specific indicators. Statistical analysis identifies a “hard-indicator dominant” pattern. Energy consumption, energy efficiency, and greenhouse gas (GHG) emissions emerge as the three most frequently selected indicators, statistically corroborating the field’s primary focus on output efficiency and environmental burdens. Similarly, the widespread adoption of energy cost and per capita income reflects the entrenched integration of socio-economic factors. However, a discrepancy is observed between objective and subjective indicators. While physical and monetary quantities (e.g., renewable share, energy security) are highly represented, subjective indicators such as user satisfaction and policy support intensity show disproportionately low selection rates. This highlights a “heavy on physics, light on perception” bias in existing indicators.

These distributional characteristics provide the data foundation for the following in-depth critical review. The subsequent sections will systematically deconstruct these dimensions, traversing from the dominant techno-economic and environmental indicators to the marginalized social indicators, exploring their mathematical evolution and methodological integration.

3. Rural Energy Systems Evaluation Indicators

The evaluation of rural energy systems has fundamentally evolved from isolated, single-domain assessments into integrated indicators. Historically, system feasibility was determined by standalone indicators within distinct technical or economic boundaries. However, current methodologies emphasize an integrated approach that involves coupling economic viability indicators (e.g., LCOE, NPC) with technical reliability constraints like loss of power supply probability (LPSP). Furthermore, they are integrated with broader environmental indicators, including lifecycle carbon emissions (CE), and social welfare indicators such as human development index (HDI). This section details core indicators within each domain and subsequently explores how MCDM synthesizes these competing indicators.

3.1. Economic Indicators

The economic assessment of rural energy systems has transformed from simple cost accounting to a more comprehensive evaluation of long-term viability. Economic evaluation is fundamental to rural energy systems planning, directly influencing technology selection, system configuration, and policy formulation. Economic evaluation indicators have shifted from static cost comparisons to comprehensive indicators integrating life-cycle analysis and dynamic optimization. Advanced software platforms such as hybrid optimization model for electric renewables (HOMER) have enabled multi-scenario life-cycle optimization for rural energy systems [9,10]. Moreover, the evaluation scope has expanded to incorporate policy incentives and financing mechanisms. Emphasis is placed on financing capacity, local market activation, and long-term growth [11]. This section reviews the advancement of key economic indicators, their regional applications, and the impact of policy and financing mechanisms on the economic indicators.

3.1.1. Evolutionary Trajectory of Economic Indicators

The core economic evaluation indicators and their measurable equations are shown in

Table 1. Core economic evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Total annual cost ($TAC$)	$TAC=C_{\mathrm{acap}}+C_{\mathrm{amain}}$ $CRF={\displaystyle \frac{[r{(1+r)}^d]}{[{(1+r)}^d-1]}}$	[17,18]
Annualized cost of system ($ACS$)	$ACS=C_{\mathrm{acap}}+C_{\mathrm{arep}}+C_{\mathrm{amain}}$	[19]
Cost of energy ($COE$)	$TAEP={\displaystyle \sum _{i=1}^{365}{E}_{\mathrm{g}}(i)}$ $COE={\displaystyle \frac{{\displaystyle \sum _{i=1}^{n}AC{S}_i}}{{\displaystyle \sum _{i=1}^{n}TAE{P}_i}}}$	[20]
Life cycle cost ($LCC$)	$LCC=C+O{M}_{\mathrm{npv}}+R_{\mathrm{npv}}-S_{\mathrm{npv}}$	[21,22]
Net present value ($NPV$)	$\begin{array}{l}NPV={\displaystyle \sum NP{V}_{\mathrm{income}}}+{\displaystyle \sum NP{V}_{\mathrm{end}}\hspace{1em}}-\\ \hfill C_{\mathrm{invest}}-{\displaystyle \sum NP{V}_{\mathrm{OM}}}-\hfill \\ \hfill {\displaystyle \sum NP{V}_{\mathrm{r}}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}}\hfill \end{array}$	[5,22]
Levelized cost of energy ($LCOE$)	$LCOE={\displaystyle \frac{TAC}{E_{\mathrm{tot}}}}$ $LCO{E}_{\left({\displaystyle \frac{\$}{kWh}}\right)}={\displaystyle \frac{ICC·CRF}{{\displaystyle \sum _{i=1}^{365}\hspace{1em}}{E}_{\mathrm{g}(i)}}}$	[5,23]

. Rather than merely accounting for basic expenditures, the objective of economic assessment has fundamentally shifted toward evaluating long-term financial viability and socio-environmental cost-effectiveness. Early economic assessments in rural energy systems primarily focused on fundamental cost accounting and financial viability. To evaluate micro-level profitability based on discounted cash flows, core indicators are widely utilized, such as the total annual cost, net present value, internal rate of return (IRR), and payback period [5,12]. These conventional indicators were predominantly employed. Comparisons between kerosene lamps and solar home systems were conducted [13]. However, with the increasing integration of distributed renewable energy, these static indicators proved insufficient to capture long-term operational complexities and high initial capital costs. Therefore, the evaluative scope has expanded to a comprehensive life-cycle perspective. Indicators, such as life cycle cost and levelized cost of energy, have emerged as the definitive benchmarks [14]. These indicators allow for robust techno-economic comparisons across different technologies over their entire lifespans. In the contemporary “Carbon Neutrality” era, the economic evaluation indicators of rural energy systems have transcended conventional cost accounting by internalizing environmental externalities (e.g., carbon trading value [15]) and socio-economic co-benefits (e.g., energy poverty alleviation and systemic resilience) [16]. The trajectory reflects a strategic shift in rural energy systems development from resource-driven growth toward a synergistic equilibrium of decarbonization, economic efficiency, and social equity.

3.1.2. Critical Assessment of Key Economic Indicators

In the economic assessment of rural energy systems, LCOE, NPC, and IRR are three fundamental economic indicators. In the context of rural energy systems, LCOE serves as the benchmark for unit generation efficiency, NPC evaluates the total lifecycle financial liability, and IRR determines investment attractiveness.

(1) Levelized cost of energy

LCOE formulations across the reviewed literature exhibit structural variations and diverse underlying financial assumptions [5,22]. In the absence of strict harmonization, the subsequent comparisons function as macro-level indicators of relative cost trends rather than absolute micro-financial equivalencies [1]. The LCOE of rural energy systems is primarily governed by technological configuration and reliability standards. As quantified in Figure 13, pronounced disparities in the LCOE distribution of rural energy systems are evident across systems. Superior LCOE resilience in rural energy systems is exhibited by micro-hydro and biomass configurations, supported by mature systems and reliable local resources. Therefore, the lowest LCOE benchmarks for rural energy systems are ascribed to micro-hydropower solutions in regions defined by permissible topography [24]. Intermediate economic performance is observed in wind power, solar photovoltaic (PV), and grid extension scenarios in rural energy systems. A compact cost distribution is demonstrated by wind power, reflecting the economic consistency of rural energy projects deployed in high-potential sites. In contrast, moderate dispersion is noted in solar PV, attributable to spatial heterogeneity in irradiation profiles. Similarly, grid extension in rural energy systems generally occupies this median cost range, serving as a centralized benchmark against which decentralized renewable options are evaluated in such systems. Instead, the highest LCOE baseline and most pronounced variability are displayed by diesel generation in rural energy systems. This extreme volatility quantifies the systemic risk of fossil fuel dependence, where costs are dictated by fragile supply chains and global price fluctuations rather than local resource availability. The LCOE of rural energy systems is significantly optimized through multi-energy integration to mitigate these limitations. A reduction of 30% to 78% in the LCOE of rural energy systems is typically achieved by hybrid PV-diesel-battery architectures compared to single-source models [3]. Furthermore, the LCOE of rural energy systems is highly sensitive to the stringency of stability requirements. A decrease of approximately 20% to 27% in the LCOE indicator for rural energy systems is observed when reliability targets are modified from 100% to 80% [16].

Significant heterogeneity in the LCOE of rural energy systems across national boundaries is evidenced by the cross-country comparisons in Figure 14. Beyond systems parameters, the economic viability of rural energy systems is fundamentally reshaped by regional industrial maturity and strategic valuation indicators [25]. First, LCOE baselines are heavily dependent on local infrastructural capabilities in rural energy systems. In resource-rich regions like India, hybrid systems achieve competitively low LCOE (0.111 USD/kWh), supported by economies of scale and favorable industrial policies [26]. In contrast, rural energy systems in developing regions are often constrained by logistical deficits. In sub-Saharan Africa, where high costs create critical affordability barriers, LCOE-centric indicators are prioritized in 94% of studies about rural energy systems [25]. Similarly, in remote territories such as the Philippines [27] and the Amazon [28], extreme cost outliers are driven by fragile supply chains and grid extension barriers. Therefore, solutions in these contexts focus on utilizing indigenous feedstocks (micro-hydro/biomass) to insulate projects from external supply risks. Second, economic evaluations for rural energy systems are increasingly redefined by policy imperatives beyond strict cost minimization. In developed contexts, higher LCOE thresholds are often tolerated to achieve broader socio-economic goals in rural energy systems. For instance, strong production inducement effects are prioritized in Japan’s rural energy systems despite resulting in higher LCOE values [29]. Similarly, in the EU rural energy systems, the deployment of local biomass is justified by the strategic imperative of regional self-sufficiency, frequently overriding strict LCOE minimization targets [30]. Furthermore, financial mechanisms also play a regulatory role. For example, a reduction of 19% in LCOE is observed under a carbon price of USD 10/t CO₂, reflecting the internalization of environmental values [2].

The evaluation of LCOE for rural energy systems has evolved from static, deterministic assumptions to dynamic, spatiotemporal indicators. This evolution has been driven by the need to capture the spatial and financial complexities. In advanced models, the LCOE of rural energy systems is transformed into a continuous distribution dependent on geography and infrastructure via dynamic cash flow analysis [31]. However, exclusive reliance on the LCOE of rural energy systems for system selection is increasingly being viewed as insufficient due to associated risks of excessive emissions or compromised reliability [32]. Furthermore, LCOE is inapplicable for non-generating systems (e.g., peak-shaving storage), where it yields misleadingly high values and must be replaced by annualized capacity cost indicators [14]. Consequently, the optimization of LCOE for rural energy systems is prioritized within multi-objective indicators. These indicators are designed to balance economic targets against technical [33], environmental, and social indicators [34]. The LCOE of rural energy systems is frequently minimized under strict constraints using artificial intelligence approaches, including genetic algorithms and particle swarm optimization [6]. Through this evolution, the LCOE of rural energy systems is evaluated simultaneously with load loss probability, GHG emissions, and social indicators like the HDI and job creation (JC) [35]. This is achieved through a multi-objective optimization that balances economic, technical, and social dimensions [35].

(2) Net present cost

While the LCOE indicator quantifies the cost per unit of energy generation, the NPC of rural energy systems aggregates the total financial burden over the project lifecycle, serving as the primary objective function for systems sizing. Therefore, the NPC of rural energy systems is fundamentally dictated by regional resource endowments. Minimized NPC values for rural energy systems are observed in standalone PV-storage architectures within solar-abundant regions (e.g., Iran [36], Iraq [37]). Similarly, superior NPC performance in rural energy systems is achieved by hybrid wind-solar configurations in locations characterized by complementary resource profiles [3]. In developing nations such as Myanmar [9], Yemen [38], and Nigeria [39], LCOE and NPC reduction is emphasized by optimized configurations to mitigate high investment and transport costs [40]. Furthermore, cost drivers of NPC in rural energy systems are observed to diverge significantly based on grid connectivity. For off-grid applications, the initial capital component of NPC for rural energy systems is increased by battery integration, yet the total lifecycle NPC is reduced through fuel displacement [41]. Comparative analyses in South Africa demonstrate that off-grid systems are more cost-effective than grid expansion over the project lifecycle [42]. Conversely, for grid-connected configurations, the net NPC burden is primarily mitigated by revenue from electricity sales. Driven by these resource-based advantages and operational mechanisms, the economic superiority of renewable hybrid configurations in rural energy systems is consistently established by NPC evaluations. NPC analyses of rural energy systems identify significantly lower lifecycle costs for hybrid models compared to standalone diesel generator (DG) baselines [9].

The NPC of rural energy systems is highly sensitive to dynamic external factors, particularly fuel market volatility and policy incentives. Rising diesel prices further amplify the NPC efficiency gap between hybrid rural energy systems and fossil-fuel counterparts [43]. The NPC of rural energy systems is simultaneously influenced by carbon trading revenues, which offset project costs [44]. Regarding assessment methodology, the reliability of NPC calculations for rural energy systems is strictly dependent on the sophistication of the modeling framework used. Intrinsic NPC algorithms within platforms like HOMER are widely adopted for rural energy systems [7], though dynamic operational characteristics are often neglected by such simplified tools [45]. Furthermore, NPC’s rigid assumptions regarding static lifespans and discount rates make it inapplicable for evaluating the flexibility of dynamically scaled phased projects. Accordingly, the evaluation of NPC for rural energy systems has necessitated a transition toward refined cost decomposition and dynamic robustness analysis [46]. The robustness of NPC in rural energy systems is now examined under parameter perturbations [47], with risk-adjusted discount rates incorporated to quantify market impacts [48]. Ultimately, the minimization of NPC and LCOE for rural energy systems primarily ensures affordability for end-users rather than guaranteeing investment returns. This distinction highlights the necessity of incorporating profitability indicators alongside NPC, such as the IRR.

(3) Internal rate of return

In addition to cost-minimization indicators (LCOE/NPC), the IRR of rural energy systems shifts the focus to capital efficiency, serving as a decisive “go/no-go” gate for private investors in rural energy systems. The application of IRR in rural electrification has progressed from fundamental financial evaluation to the comprehensive analysis of policy correlations. IRR evaluations are utilized to quantify the critical impact of subsidies on the viability of rural energy projects. For instance, capital subsidy rates of 35–60% for rural energy systems in Ghana were shown to elevate the IRR from unfeasible levels to a viable range of 15–25% [8], thereby identifying effective intervention thresholds for policymakers. Similarly, in policy-driven contexts like Brazil, 80% of investments under the “Light for All” program are covered by government subsidies to ensure project feasibility [49].

Furthermore, the IRR of rural energy systems is intrinsically linked to revenue models and regional risk profiles. Analytical scopes have expanded to demonstrate that rural energy systems enabling income-generating activities substantially outperform basic lighting models in terms of IRR [50,51]. Concurrently, national benchmarks are integrated to align evaluations of rural energy systems with local investment realities. Evaluation benchmarks for rural energy systems vary significantly by region due to differing risk profiles. For instance, a 20% IRR was deemed economically viable against a 15% minimum standard in Venezuela [52]. In contrast, due to higher market volatility, investors in contexts like Tanzania typically demand IRR thresholds exceeding 20% to justify the investment risk [51]. Despite these benchmarks, persistent obstacles for rural energy systems, notably high upfront costs and inadequate maintenance funding, often suppress IRR levels [53,54]. Therefore, the IRR of rural energy systems is stabilized through specialized financing instruments. Initial investment barriers are mitigated through soft loans and microcredit [55,56]. At the user level, a distinct preference for short-term credit options is leveraged to secure cash flows and sustain the project IRR [57,58]. In parallel, the evaluation methodology for rural energy systems has shifted toward rigorous refinement. Robust indicators, such as the Adjusted IRR, have been proposed as a means of accounting for the reinvestment of mid-term cash flows. This approach addresses the shortcomings of traditional calculations for long-term rural projects [59]. Ultimately, the strategic utility of IRR extends to the mobilization of private capital. Innovative mechanisms, notably blended finance, are increasingly leveraged to elevate the IRR of rural energy systems beyond commercial thresholds. This alignment of public subsidies with private return expectations transforms rural energy systems into bankable assets, ensuring long-term market scalability.

However, presenting economic indicators as static deterministic values significantly limits their practical applicability in complex rural environments. To address this, contemporary evaluations increasingly incorporate sensitivity analyses to assess system robustness against critical parameter perturbations. Key variables such as discount rates, fuel prices, and system lifetimes drastically alter the economic hierarchy of rural energy system. For instance, high discount rates disproportionately penalize capital-intensive renewable systems by inflating the annualized cost, while escalating diesel prices rapidly improve the relative net present cost competitiveness of hybrid architectures [43]. Similarly, variations in the projected system lifetime fundamentally skew the capital recovery factor and total life cycle costs. Therefore, modern economic assessment methodologies advocate for dynamic sensitivity modeling, transitioning from single-point static estimates to probabilistic economic evaluations that account for long-term financial uncertainties and market volatility [48].

Beyond these financial and temporal uncertainties, the practical application of these economic models in rural contexts is fundamentally constrained by resource availability and logistics. The geographical distribution of natural resources (e.g., solar irradiation, wind velocity, and local biomass feedstocks) directly dictates the system’s capacity factor and annual energy production, which critically shapes the denominator of the LCOE [53]. Concurrently, logistical challenges, such as underdeveloped transportation networks and long distances to urban centers, significantly inflate both the initial capital costs for equipment delivery and the lifecycle operation and maintenance costs due to fragile fuel supply chains [40]. Therefore, traditional economic evaluations are insufficient for rural energy systems unless they dynamically incorporate these spatial and logistical premiums.

3.2. Technical Indicators

3.2.1. Key Technical Indicators

Technical indicators serve as the quantitative basis for assessing the physical performance of rural energy systems, acting as constraints to balance reliability against economic viability. The core technical evaluation indicators and their measurable equations are shown in Table 2. These indicators are generally categorized into three subsets: reliability indices, performance indicators, and operational constraints. As the primary subset, reliability indices play a critical role in the capacity allocation of rural energy systems [6]. Probabilistic indicators, such as the LPSP and loss of load probability, fundamentally quantify the percentage of time or energy with unmet load demand. Because they measure this same ratio, current literature reveals that applying them simultaneously is mathematically redundant. LPSP is a foundational constraint indicator. In rural energy systems design, an LPSP of 5% is frequently selected as the technical constraint to optimize economic viability without compromising essential reliability. Conversely, stringent constraints (e.g., 0%) significantly escalate the costs of rural energy systems due to the requirement for extensive backup capacity [12]. Similar probabilistic constraints include the probability of deficiency in power supply, the ratio of power supply faults to total load, and the loss of load risk, which is defined as the probability of daily supply failure from renewable deficits. To evaluate absolute shortfalls, indicators such as the expected energy not supplied, loss of energy expected, and unmet load are employed to measure the anticipated unsupplied energy and total power shortage. For temporal evaluation, the loss of load expected measures the expected annual hours of capacity deficit, broadening the evaluative scope beyond fundamental indicators like LPSP. Similarly, loss of load expected and loss of energy expected [57] overlap substantially, merely changing the scale from time to volume. Explicit thresholds are often applied to the loss of load expected in rural energy systems, such as requiring values under 2% to ensure operational stability. Meanwhile, the expected energy not supplied is utilized to derive reliability indices for optimizing rural energy systems in remote terrains, such as the mountainous regions of India. At a macro level, the reliability of rural energy systems is further characterized by the equivalent loss factor. This is defined as the ratio of effective forced outage hours to total hours. Values below 0.01 are typically deemed acceptable for rural energy applications [60].

Beyond reliability, additional technical indicators acting as performance indicators and operational constraints serve as key boundaries in system modeling. To eliminate computational redundancy and a lack of standardized prioritization, practical system evaluation must prioritize these indicators hierarchically rather than deploying the entire suite indiscriminately. Regarding performance indicators, capacity factors reveal significant performance disparities across generation systems employed in rural energy systems. For instance, small hydropower in Uttarakhand, India, achieves markedly superior capacity factors (55–90%), illustrating the performance advantages of site-matched systems [53]. Among intermittent renewable options, wind power typically demonstrates higher capacity factors (10–30%) than solar PV (~20%). Nevertheless, PV remains favored in many rural contexts due to its reduced capital costs and installation simplicity [4]. However, the inherently low capacity factor and intermittency of solar PV necessitate robust energy storage solutions. Therefore, the battery state of charge, representing the proportion of available energy in the storage system, emerges as a core dynamic constraint in these storage-integrated rural energy systems. Managing this operational parameter requires advanced modeling approaches (e.g., the Kinetic Battery Model [48]) to handle complex discharge cycles, prevent over-discharge, and ensure component longevity. It is crucial to recognize when specific technical models are inapplicable and must be excluded. For instance, off-grid reliability indices (e.g., LPSP) and performance indicators (e.g., level of autonomy) are systematically excluded when evaluating grid-connected microgrids. Similarly, operational constraints like the battery state of charge are inherently inapplicable to systems lacking electrochemical storage [50].

Initially focused on singular reliability indicators, assessment protocols have expanded from a predominant reliance on the LPSP to a complex indicator matrix addressing overall system resilience. Rather than isolating indicators, parameters such as the aforementioned equivalent loss factor [61], expected energy not supplied, and total energy loss are now employed concurrently to rigorously evaluate system robustness [5,61]. Hierarchically, probabilistic constraints like LPSP act as primary thresholds for initial capacity sizing [6], while operational diagnostic indicators like level of autonomy [19] and battery state of charge [5] serve as secondary indicators to assess dispatch efficiency and storage health. Concurrently, indicators of operational independence, specifically the level of autonomy, defined as the fraction of operational hours without loss of load, and the battery state of charge, have been integrated to assess the standalone capabilities of off-grid systems [38,62]. Furthermore, with the widespread adoption of hybrid rural energy systems, the evaluative scope has broadened from individual technology characterization to the synergistic performance of integrated systems. The technical potential is typically quantified by integrating geographic endowments with the previously discussed capacity factors and conversion efficiencies [63], based on the standardized indicators like the irradiance-area-efficiency model for solar applications [64]. Methodologically, reliance on empirical models and static assumptions [15] has been superseded by high-resolution time-series simulation and stochastic optimization. Advanced software tools (e.g., HOMER) and AI-driven algorithms, such as Genetic Algorithms [39], are now extensively utilized to compute these technical indicators dynamically. Consequently, these tools enable a more precise optimization of the trade-off between system reliability and economic constraints [4,65].

Furthermore, the practical optimization of rural energy systems demands a rigorous quantification of the trade-offs between technical reliability and economic costs. The marginal cost of reliability exhibits a non-linear increase as systems approach zero-outage targets [12]. For instance, modifying reliability targets from an absolute 100% (i.e., 0% LPSP) down to 80% can yield a substantial decrease of approximately 20% to 27% in the LCOE [16]. This severe cost escalation for absolute reliability is primarily driven by the necessity for exponentially larger battery banks and redundant generation capacities to cover rare, extreme weather events. Therefore, rather than pursuing absolute technical reliability, practical optimization of rural energy systems must identify the economic inflection point. At this point, the marginal cost of further reducing the LPSP must be carefully weighed against the local community’s affordability and the regional value of lost load, ensuring that strict technical constraints do not inadvertently paralyze the project’s financial viability.

Table 2. Core technical evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Loss of power supply probability ($LPSP$)	$LPSP={\displaystyle \frac{{\displaystyle \sum _{t=1}^T }DE(t)}{{\displaystyle \sum _{t=1}^T }{P}_{\mathrm{load}}(t)\Delta t}}$	[20,23]
Loss of load probability ($LOLP$)	$LOLP={\displaystyle \frac{{\displaystyle \sum _{t=1}^{8760} }ES(t)}{{\displaystyle \sum _{t=1}^{8760}\hspace{1em}}LD(t)}}$	[34,66]
Expected energy not supplied ($EENS$)	$EENS={\displaystyle \sum _{t=1}^{8760} }{E}_{\mathrm{u}}(t)$ $EIR=1-{\displaystyle \frac{EENS}{E_0}}$	[67,68]
Deficiency of power supply probability ($DPSP$)	$DPS={\displaystyle \sum _{t=1}^T }LD(t)-E_{\mathrm{sist}}(t)$ $DPSP={\displaystyle \frac{{\displaystyle \sum _{t=1}^T }DPS(t)}{{\displaystyle \sum _{t=1}^T }LD(t)}}$	[21,69]
Loss of load expected ($LOLE$)	$LOLE={\displaystyle \sum _{h=1}^H }{\displaystyle \sum _{i\in S} }{P}_i\times T_i$	[20,69]
Loss of energy expected ($LOEE$)	$LOEE={\displaystyle \sum _{h=1}^H }{\displaystyle \sum _{i\in S} }{P}_i\times LO{E}_i$	[57,64]
Unmet load ($UL$)	$UL={\displaystyle \frac{{\displaystyle \sum _{t=1}^n }{P}_{\mathrm{failure}}}{{\displaystyle \sum _{t=1}^n }{P}_{\mathrm{total}}}}$	[36,70]
Loss of load risk ($LOLR$)	$LOLR=1-p \mathrm{or} LOLR=q$	[23]
Equivalent loss factor ($ELF$)	$ELF={\displaystyle \frac{1}{h}}{\displaystyle \sum _{n=1}^H }{\displaystyle \frac{E(Q(h))}{D(h)}}$	[40]
Level of autonomy ($LA$)	$LA=1-{\displaystyle \frac{H_{\mathrm{LOL}}}{H_{\mathrm{tot}}}}$	[19]
State of charge ($SOC$)	$SOC(t+1)=SOC(t)\sigma +I_{\mathrm{bat}}(t)\Delta t\eta (I_{\mathrm{bat}}(t))$	[5,71]

Table 2. Core technical evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Loss of power supply probability ($LPSP$)	$LPSP={\displaystyle \frac{{\displaystyle \sum _{t=1}^T }DE(t)}{{\displaystyle \sum _{t=1}^T }{P}_{\mathrm{load}}(t)\Delta t}}$	[20,23]
Loss of load probability ($LOLP$)	$LOLP={\displaystyle \frac{{\displaystyle \sum _{t=1}^{8760} }ES(t)}{{\displaystyle \sum _{t=1}^{8760}\hspace{1em}}LD(t)}}$	[34,66]
Expected energy not supplied ($EENS$)	$EENS={\displaystyle \sum _{t=1}^{8760} }{E}_{\mathrm{u}}(t)$ $EIR=1-{\displaystyle \frac{EENS}{E_0}}$	[67,68]
Deficiency of power supply probability ($DPSP$)	$DPS={\displaystyle \sum _{t=1}^T }LD(t)-E_{\mathrm{sist}}(t)$ $DPSP={\displaystyle \frac{{\displaystyle \sum _{t=1}^T }DPS(t)}{{\displaystyle \sum _{t=1}^T }LD(t)}}$	[21,69]
Loss of load expected ($LOLE$)	$LOLE={\displaystyle \sum _{h=1}^H }{\displaystyle \sum _{i\in S} }{P}_i\times T_i$	[20,69]
Loss of energy expected ($LOEE$)	$LOEE={\displaystyle \sum _{h=1}^H }{\displaystyle \sum _{i\in S} }{P}_i\times LO{E}_i$	[57,64]
Unmet load ($UL$)	$UL={\displaystyle \frac{{\displaystyle \sum _{t=1}^n }{P}_{\mathrm{failure}}}{{\displaystyle \sum _{t=1}^n }{P}_{\mathrm{total}}}}$	[36,70]
Loss of load risk ($LOLR$)	$LOLR=1-p \mathrm{or} LOLR=q$	[23]
Equivalent loss factor ($ELF$)	$ELF={\displaystyle \frac{1}{h}}{\displaystyle \sum _{n=1}^H }{\displaystyle \frac{E(Q(h))}{D(h)}}$	[40]
Level of autonomy ($LA$)	$LA=1-{\displaystyle \frac{H_{\mathrm{LOL}}}{H_{\mathrm{tot}}}}$	[19]
State of charge ($SOC$)	$SOC(t+1)=SOC(t)\sigma +I_{\mathrm{bat}}(t)\Delta t\eta (I_{\mathrm{bat}}(t))$	[5,71]

3.2.2. Selection of Technical Indicators

The selection of indicators and thresholds for rural energy systems is heavily dictated by regional resource endowments and developmental stages. In Asia, emphasis is placed on LPSP and load satisfaction to ensure basic access for rural energy systems under strict budget constraints, as evidenced by studies in India and Bangladesh [13,53]. Conversely, Latin American research on rural energy systems prioritizes system adaptability and local resource integration, exemplified by biomass efficiency assessments in Brazil [28,54]. In Africa, the resilience and long-term performance stability of rural energy systems under extreme environmental conditions remain the focal points. Furthermore, structural distinctions exist within rural energy systems by connection type. Off-grid systems prioritize autonomous operation, whereas grid-connected rural systems focus on stability and power regulation [62].

Despite the proliferation of indicators for rural energy systems, a unified global benchmark remains unestablished, leading to fragmentation in definitions [5,53]. More critically, substantial discrepancies persist between theoretical optimization and practical performance in rural energy systems, often stemming from inadequate operation and maintenance. Evidence of this gap includes premature battery failures in rural energy systems on Pangon Island, Philippines [27], and high failure rates in Brazil’s Prodeem project [54]. Current planning methodologies for rural energy systems often prioritize static design simulations while neglecting dynamic challenges such as equipment aging and supply chain resilience [39]. To bridge this gap, future research must pivot from static design indicators to dynamic, lifecycle-focused indicators, such as real-time degradation rates and predictive maintenance parameters. Meanwhile, a paradigm shift toward “technical resilience” is proposed for rural energy systems, utilizing IoT and AI to create a closed-loop cycle of design, deployment, and re-optimization [65,72].

3.3. Environmental Indicators

3.3.1. Evolution of Environmental Indicators

The environmental indicators for rural energy systems have undergone a paradigm shift, transitioning from a singular indicator of CE [73] to a comprehensive suite of multi-impact, lifecycle-oriented indicators. However, a critical methodological challenge in current assessments is the lack of a standardized system boundary, with studies inconsistently applying either purely operational or comprehensive “cradle-to-grave” scopes [74,75]. This ambiguity significantly affects emission results and severely limits comparability.

Concurrently, the indicator scope applied to rural energy systems has broadened beyond GHGs, involving SO₂, NO₂, and particulate matter [50,76]. Contemporary environmental indicators further integrate resource utilization intensity [53] and ecological dimensions relevant to rural contexts, such as land use and water consumption [52,60]. Moreover, environmental impacts are jointly evaluated with economic parameters, such as abatement costs and carbon externalities, which are explicitly quantified to support holistic decision-making in rural planning [31,77]. The core environmental evaluation indicators and their measurable equations are shown in

Table 3. Core environmental evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Carbon emission (CE)	$E_{\mathrm{tot}}={\displaystyle \sum _{t\in T}\hspace{1em}}{\displaystyle \sum _{n\in N}\hspace{1em}}{E}_n{G}_{nt}$	[5,32]
Embodied energy ($EE$)	$E{E}_{\mathrm{pv}}=3379\times A_{\mathrm{pv}}$ $E{E}_{\mathrm{wt}}=28.342\times A_{\mathrm{wt}}^2+2361.3\times A_{\mathrm{wt}}$ $E{E}_{\mathrm{bat}}=60\times C_{\mathrm{n}}$	[78]
Carbon footprint of energy ($CFOE$)	$\begin{array}{ll}CFOE=& {\displaystyle \frac{Y_{\mathrm{PV}}\left(P_0)+Y_{\mathrm{PVstr}}\right(P_0)+Y_{\mathrm{B}}\left(K_{\mathrm{B}}\right)}{d·E_{\mathrm{L}}}}\\ & +{\displaystyle \frac{Y_{\mathrm{inv}}\left(P_0)+Y_{\mathrm{binv}}\right(P_{\mathrm{binv}})+Y_{\mathrm{W}}\left(P_0)+Y_{\mathrm{G}}\right(P_0,K_{\mathrm{B}})}{d·E_{\mathrm{L}}}}\end{array}$	[79]
Life cycle assessment ($LCA$)	$\begin{array}{cc}\hfill LCA& =\left(E{M}_{\mathrm{w}}·{\displaystyle \sum _{t=1}^{24}{P}_{\mathrm{w}}^t·N_{\mathrm{w}}·\Delta t}+E{M}_{\mathrm{pv}}·{\displaystyle \sum _{t=1}^{24}{P}_{\mathrm{pv}}^t·N_{\mathrm{pv}}·\Delta t}\right.\hfill \\ & \hspace{1em}+\left.E{M}_{\mathrm{batt}}·S_{\mathrm{batt}}·N_{\mathrm{b}}+E{M}_{\mathrm{d}}·{\displaystyle \sum _{t=1}^{24}{E}_{\mathrm{d}}^t}\right)·365·d\hfill \end{array}$	[32,80]

.

Figure 15 illustrates the transition from single environmental indicators to comprehensive assessments for the rural energy systems, encompassing a shift from emissions-oriented to holistic impact assessments. However, despite the evolutionary progress, critical challenges persist within the current literature on rural energy systems. These challenges are primarily characterized by regional heterogeneity and inconsistencies in quantitative assessments. Indicator selection remains biased by socioeconomic development levels, as local economic priorities often dictate environmental targets and technological contexts. Developed economies, such as those in Europe, emphasize global GHG emissions and overall ecological efficiency of rural energy systems [81,82]. Conversely, developing nations (e.g., India, China) prioritize localized impacts of rural energy systems. These studies focus on indoor air quality [83], forest conservation, and fuel substitution [84]. Similarly, technological biases for the rural energy systems are evident in rural contexts. Biomass assessments often focus on resource cycling [30,85], while PV and wind studies for rural applications prioritize lifecycle carbon footprints. In addition to these disparities, a significant imbalance between quantitative and qualitative assessments exists in the evaluation of rural energy systems. Non-climate benefits, such as health improvements, are often only described descriptively rather than numerically [30,86]. This hinders their direct comparison against quantifiable economic costs during system optimization. Furthermore, methodological support for aligning environmental indicators with reliability and affordability in rural energy systems remains insufficient [49,87]. Therefore, a unified, adaptive indicator system for rural energy systems should be developed in future research.

3.3.2. Critical Analysis of Environmental Indicators

(1) Carbon emissions

CE, which evaluates the total carbon dioxide emissions over a specific period of time, has been established as the core environmental impact indicator in rural energy system sustainability assessments. The CE evaluation methodology has transitioned from operational emissions accounting to comprehensive life cycle assessment. Prior research prioritized direct GHG reduction benefits. For instance, solar home systems replace 5.0–21.3 L of kerosene monthly [2]. However, the distinct emission characteristics of renewable energy systems require a broader scope. Unlike fossil-fuel systems, where 95% of emissions originate during operation [82], renewable energy systems are “manufacturing-intensive” regarding emissions. While PV systems exhibit zero operational emissions, they retain a substantial manufacturing global warming potential of approximately 80 kg CO₂eq per panel [2]. And thus, excluding embodied emissions would artificially inflate the environmental benefits of renewables. This structural difference necessitates the minimization of lifecycle CE (rather than merely operational CE) as the optimization target. Therefore, minimizing lifecycle CE is frequently adopted as a primary optimization target for rural energy systems. Lifecycle CE is simulated via tools including HOMER [65], LEAP [39], and RETScreen [88]. Particularly in hybrid renewable energy systems, CE is evaluated alongside renewable energy share [48], fuel consumption, and system costs [62].

Comparative life-cycle analyses of rural energy systems reveal significant performance disparities across technical pathways, as visualized in Figure 16. While pure diesel systems generate tens to hundreds of tons of carbon annually [38], renewable integration substantially mitigates these emissions [26]. Integrated hybrid systems can achieve reductions approaching 100% [81], attributed to multi-source complementarity and biomass carbon cycling. Solar and wind hybrid systems typically achieve reduction rates of 60–90% (e.g., 4.82 tons reduction over 20 years for Bangladeshi solar homes [3]). However, intermittent renewable energy systems, such as solar and wind, require storage to mitigate fluctuations [62]. Emission reduction performance is strictly governed by system boundaries and connection types. For standalone systems, the baseline is typically diesel generation. Thus, renewable integration achieves linear reductions proportional to penetration. In addition, grid-connected potential is constrained by the local grid’s marginal emission intensity. If the regional grid is already low-carbon (e.g., hydro-dominated), the marginal benefit of adding distributed PV is diminished (40–70% reduction) [29]. Therefore, grid-connected evaluations require dynamic correction based on the real-time carbon emission intensity of the power grid, whereas off-grid assessments focus on the depth of fossil substitution.

Nowadays, the economic viability of emission reduction is increasingly reinforced by policy instruments. Economic externalities are integrated into assessments through social carbon costs and carbon taxes. Policy indicators leverage carbon pricing (e.g., $20/ton [89]) to internalize these costs and incentivize the optimization of rural energy systems. With certified emission reduction prices at €40/ton, carbon revenues account for 15% of total investment [77]. Ultimately, under these market mechanisms, integrated hybrid systems can provide the optimal balance between emission efficiency and financial feasibility [31].

(2) Renewable fraction

The renewable fraction serves as a vital indicator for evaluating the extent of fossil fuel substitution and decarbonization. Figure 17 illustrates global renewable fraction trends, identifying hybrid renewable energy systems as the dominant pathway for deep decarbonization. Hybrid renewable energy systems consistently outperform single-source systems through spatiotemporal complementarity. For instance, PV-wind hybrid systems in Malaysia and India achieved renewable fractions of 64–99% [4], which enables substantial reductions in CO₂ emissions, ranging from 70% to 99%, and a significant decrease in diesel consumption by over 80% [38]. Generally, resource availability and logistics constrain the renewable fractions of single-source configurations. Solar PV in sub-Saharan Africa reached renewable fractions of 38–54% [9,51]. Conversely, biomass utilization drops to 15–25% when collection radii exceed 5 km [54].

However, while an increasing renewable fraction yields environmental benefits, the marginal abatement cost escalates exponentially as the system approaches 100% renewable generation. High-fraction scenarios (e.g., in Saudi Arabia [3]) realized maximum CO₂ reductions but at levelized costs 3–5 times those of diesel systems due to massive seasonal storage needs [10]. Conversely, low-to-medium fraction systems (30–60%) achieve substantial fossil displacement at minimal cost [84]. This non-linearity indicates that simply maximizing the renewable fraction is not always economically optimal. Therefore, robust system design requires identifying the “economic inflection point”. This is achieved by balancing renewable fraction targets [12] with levelized costs within a comprehensive techno-economic perspective [14].

3.4. Social Indicators

The core social evaluation indicators and their measurable equations are shown in Table 4. Social performance evaluation of rural energy systems has evolved from isolated indicators into integrated techno-socio-economic indicators. This integration involves the coupling of social dimensions, such as macro-economic indices and social acceptance, directly with technical sizing and system costs.

At the macro level, rural energy systems are primarily evaluated in social performance based on their contribution to regional development. The HDI is utilized as the core social assessment indicator at this level, serving as a comprehensive measure of life expectancy, education, and per capita income. This indicator serves as the cornerstone linking energy access to human well-being. HDI is quantified as a logarithmic function of annual per capita electricity consumption [7] and usually integrated into rural energy systems optimization models [60]. The logarithmic growth trajectory of the function implies that the initial transition in rural energy systems from “no energy” to “basic access” yields the highest marginal utility for the poorest populations [2]. This theoretical correlation is corroborated by regional data disparities within rural energy systems. Rural areas with higher HDI (0.780–0.919) have achieved near-universal electricity service rates (99.26–100%). Conversely, rural regions with lower HDI (0.467–0.672) exhibit significantly lagging electrification rates with a wide variation (16.85–77.50%) [28]. Social performance assessment further decomposes HDI into specific welfare indicators to capture the tangible benefits underlying these macro-statistics. In the educational domain, rural energy systems extend nighttime study duration, a benefit exemplified by a 40.7% increase in Brazil’s “Light for All” program [49]. Simultaneously, regarding health, these systems mitigate indoor air pollution following the displacement of kerosene [2,84]. Regarding living standards, rural energy systems act as catalysts for productive uses like micro-enterprise growth. Such activities in rural energy systems lead to measurable economic gains, such as a 35.6% increase in family income in specific case studies [26]. The multi-tier indicators advance standardization by integrating attributes like affordability, health, and safety into a consolidated assessment system of rural energy systems [42].

Beyond direct household welfare, rural energy systems are evaluated based on their broader local socio-economic dividends, such as user income generation and employment impacts. A key indicator is JC, which quantifies labor market benefits, spanning component production, transportation, installation, and long-term operation and maintenance, using technology-specific coefficients [65]. For instance, biomass energy systems generate approximately 0.21 jobs/GWh/year, whereas natural gas systems contribute 0.11 jobs/GWh/year [6]. Additionally, social performance assessment of rural energy systems increasingly considers livelihood resilience against economic shocks. To this end, portfolio risk indicators are employed to minimize the instability of fuel prices. This approach serves to safeguard vulnerable populations from living cost fluctuations and maintain socio-political stability [22,90].

Beyond objective macro-indicators, the long-term sustainability of rural energy systems is heavily dependent on subjective factors like social acceptance and equity criteria. Social acceptance of rural energy systems is recognized as a pivotal determinant of project success [5] and is structurally subdivided into socio-political, community, and market acceptance [74]. Rural energy projects face varying degrees of acceptance, often determined by the trade-off between the community’s technological understanding and negative externalities like land-use conflicts, visual impacts, acoustic noise, and electromagnetic interference [38,56]. Closely related to acceptance, the governance of rural energy systems relies on deep community involvement in project design. High participation rates serve as a proxy for local ownership, which is essential for ensuring operational continuity and capacity building [91,92]. Moreover, modern rural energy planning prioritizes distributional equity [93]. Evaluation indicators now extend beyond simple access to encompass energy poverty alleviation and gender equity for vulnerable groups [28,52].

The practical application of social indicators within rural energy system optimization models remains constrained. While the case studies (e.g., hybrid systems in Iran) simultaneously optimize NPC and social objectives, mainstream simulation software lacks integrated social assessment modules [88]. Due to the persistent difficulties in quantifying subjective human perceptions, most models are forced to rely solely on easily measurable macro-indicators such as HDI and JC [6]. Therefore, subjective indicators of rural energy systems like social acceptance are often assigned significantly lower weights (0.1–0.2) in MCDM. To bridge the quantification gap of subjective variables like social acceptance, recent methodologies utilize Fuzzy Set Theory [5]. Primary data from surveys and questionnaires [57,74] are converted from linguistic variables into fuzzy numbers to mathematically formalize subjective feedback. These fuzzy numbers are subsequently embedded into hybrid weighting algorithms, such as Fuzzy-AHP or Fuzzy-TOPSIS. This translation allows social acceptance to be evaluated as a computable parameter alongside deterministic techno-economic indicators [33]. Future research must bridge the gap by advancing quantification methodologies. Other promising approaches include incorporating the social cost of carbon, a local social-economic estimate of the societal damages caused by carbon dioxide emissions, and health-related penalty functions [12], or adopting mixed quantitative–qualitative methods to capture complex social impacts [49].

Table 4. Core social evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Human development index ($HDI$)	$HDI=0.0978\times \ln (E_{\mathrm{load}, \mathrm{pc}})-0.0319$	[35,94]
Job creation ($JC$)	$\begin{array}{cc}JC\hfill & =J{C}_{\mathrm{pv}}\cdot P_{\mathrm{pv}}+J{C}_{\mathrm{w}}\cdot P_{\mathrm{w}}+J{C}_{\mathrm{d}}\cdot E_{\mathrm{d}}\hfill \\ & +J{C}_{\mathrm{BAT}}\cdot E_{\mathrm{BAT}}+J{C}_{\mathrm{h}}\cdot P_{\mathrm{h}}\hfill \end{array}$	[35,95]
Portfolio risk ($PR$)	$PR={\displaystyle \sum _{t\in T}\left[{\displaystyle \sum _{j\in F}{\alpha }_{jt}{\displaystyle \sum _{n\in N_j}{G}_{nt}}}\right]}$	[22,90]
Social acceptance (SA)	Public participation	[5]
Social cost of carbon ($SCC$)	$SCC=\left\{\begin{array}{ll}{\displaystyle \sum _{t=1}^d\frac{O_{\mathrm{DG}}\cdot C_E}{{(1+r)}^i}}\hfill & \mathrm{if} \mathrm{DG}\in M\hfill \\ 0\hfill & \mathrm{otherwise}\hfill \end{array}\right.$	[96]

Table 4. Core social evaluation indicators for rural energy systems.

Indicator	Equation	Reference
Human development index ($HDI$)	$HDI=0.0978\times \ln (E_{\mathrm{load}, \mathrm{pc}})-0.0319$	[35,94]
Job creation ($JC$)	$\begin{array}{cc}JC\hfill & =J{C}_{\mathrm{pv}}\cdot P_{\mathrm{pv}}+J{C}_{\mathrm{w}}\cdot P_{\mathrm{w}}+J{C}_{\mathrm{d}}\cdot E_{\mathrm{d}}\hfill \\ & +J{C}_{\mathrm{BAT}}\cdot E_{\mathrm{BAT}}+J{C}_{\mathrm{h}}\cdot P_{\mathrm{h}}\hfill \end{array}$	[35,95]
Portfolio risk ($PR$)	$PR={\displaystyle \sum _{t\in T}\left[{\displaystyle \sum _{j\in F}{\alpha }_{jt}{\displaystyle \sum _{n\in N_j}{G}_{nt}}}\right]}$	[22,90]
Social acceptance (SA)	Public participation	[5]
Social cost of carbon ($SCC$)	$SCC=\left\{\begin{array}{ll}{\displaystyle \sum _{t=1}^d\frac{O_{\mathrm{DG}}\cdot C_E}{{(1+r)}^i}}\hfill & \mathrm{if} \mathrm{DG}\in M\hfill \\ 0\hfill & \mathrm{otherwise}\hfill \end{array}\right.$	[96]

3.5. Regional Heterogeneity in Evaluation Indicator Priorities

While the assessment dimensions of rural energy systems are well-established, the prioritization of these indicators is highly context-dependent and fundamentally shaped by regional resources, logistics, and developmental stages, as shown in

Table 5. Regional heterogeneity in rural energy system evaluation priorities.

Region	Primary Developmental Goal	Prioritized Economic Indicator	Prioritized Technical Indicator	Prioritized Environmental and Social Indicators
Developing regions (cost-driven) (e.g., Sub-Saharan Africa, South Asia)	Basic power supply and affordability	Strict LCOE and NPC minimization	LPSP and load satisfaction	Localized impacts (e.g., indoor air quality, forest conservation)
Developing regions (policy and resource-driven) (e.g., Latin America, Amazon Region)	Adaptability, resource integration, and poverty alleviation	Viable IRR (subsidy-reliant) and supply chain insulation	Biomass efficiency and environmental adaptability	Energy poverty alleviation, extended education, and distributional equity
Developed regions (resilience and climate-driven) (e.g., Europe, Japan)	Grid stability, resilience, and self-sufficiency	Higher LCOE tolerance for broader socio-economic benefits	Grid interaction and system robustness	Global climate benefits, GHG reduction, and ecological efficiency

. Indicator prioritization exhibits distinct regional trajectories. In cost-driven developing regions such as Sub-Saharan Africa, strict budget constraints compel planners to prioritize affordability and basic access [25,40]. This leads to a strong emphasis on minimizing LCOE and NPC alongside basic technical indicators like loss of power supply probability. Therefore, environmental assessments in these areas focus on localized benefits, including improved indoor air quality and forest conservation [83,84].

Policy and resource-driven developing regions such as Latin America emphasize system adaptability, indigenous resource integration, and social equity to overcome geographic isolation and supply chain vulnerabilities [28,52]. Economic indicators in these areas often rely heavily on subsidies to ensure viability [49]. In stark contrast, developed regions such as Europe and Japan frequently tolerate higher economic costs to achieve overarching systemic resilience and grid stability [29,30]. Their environmental priorities are strictly aligned with global climate targets, including total greenhouse gas emission reductions and comprehensive ecological efficiency [81,82]. Ultimately, this regional heterogeneity demonstrates that a uniform assessment framework is fundamentally inadequate for rural energy planning. It reinforces the necessity of adopting dynamic MCDM methods to tailor the evaluation process to localized developmental goals and constraints.

3.6. Multi-Criteria Decision-Making Methods

As highlighted by the regional disparities and multidimensional conflicts discussed above, the configuration of rural energy systems is defined by complex trade-offs. Economic objectives often conflict with environmental targets, while technical requirements may diverge from social acceptance. Before integrating these diverse dimensions, it is crucial to recognize the inherent implementation barriers associated with each evaluation category. Figure 18 illustrates the relationships and proportional distribution of the main obstacles facing economic, technical, environmental, and social indicators. While economic indicators are predominantly constrained by market volatility and dynamic supply chain risks [25,43], technical assessments are primarily hindered by high-resolution data scarcity in remote areas [6]. Conversely, environmental and social indicators suffer heavily from quantification difficulties and regional biases. Specifically, the subjective nature of social acceptance and the localized health impacts of environmental emissions lack standardized monetization methodologies [74,88]. This systemic imbalance, where easily measurable techno-economic indicators dominate subjective socio-environmental factors, necessitates the deployment of robust weighting methodologies to resolve these conflicting obstacles. Accordingly, hybrid weighting approaches mark a shift from isolated techno-economic optimization toward integrated techno-economic-socio-environmental trade-offs. As illustrated in Figure 19, the progression is categorized into three distinct evolutionary stages.

To systematically clarify how different evaluation dimensions are integrated into the decision-making process, the mapping relationships between specific evaluation indicators and their corresponding MCDM methods across evolutionary stages are summarized in

Table 6. Mapping relationships between evaluation indicators and multi-criteria decision-making methods.

MCDM Evolutionary Stage	Representative Methods	Dominant Evaluation Indicators Coupled	Main Characteristics and Limitations
Stage 1: Tool-driven implicit hybridization	HOMER, RETScreen	Economic: LCOE, NPC Technical: LPSP, Capacity Factor	Economic-centric; treats technicals as constraints; marginalizes social dynamics.
Stage 2: Explicit weight combinations	AHP, Entropy weighting, CRITIC, Linear averaging	Social: HDI, Job Creation Environmental: CE Economic and Technical: Integrated trade-offs	Assigns explicit weights; balances objective data with subjective expert judgments.
Stage 3: Embedded decision governance	Multi-objective evolutionary algorithms, Pareto frontier, Fuzzy logic	Integrated Techno-Economic-Socio-Environmental	Handles high uncertainty and subjective rural social dynamics (e.g., community acceptance).

.

The initial phase of the MCDM for the assessment of rural energy systems is characterized by tool-driven implicit hybridization, relying on platforms like HOMER. While offering rapid screening for rural electrification, it is important to note that the model results in a “black-box” process, meaning that it prioritizes purely economic efficiency. Besides, complex rural social dynamics are marginalized, such as extreme price sensitivity and local community acceptance [9,26]. The subsequent phase in the assessment of rural energy systems shifts to explicit mathematical integration. Early mathematical methodologies primarily relied on simple linear averaging, which assumes equal weight importance and fails to address the inherent conflicts of rural energy systems in low-income areas (e.g., balancing high technical reliability against rural affordability constraints). Consequently, assessments of rural energy systems have progressed to advanced couplings like the analytic hierarchy process combined with entropy weighting [97]. This paradigm balances subjective expert judgment regarding rural development priorities with objective data dispersion, accommodating the scarcity and high uncertainty of meteorological and load data in remote areas [6]. To further mitigate the correlations between heavily intertwined evaluation criteria (e.g., fuel consumption and carbon emissions, or NPC and LCOE), techniques like CRITIC have been adopted, though with higher computational complexity. Beyond mathematics, participatory weighting has emerged to address decentralized governance challenges. This approach involves the direct involvement of local villagers and stakeholders to bridge the gap between algorithmic design and rural social acceptability [98]. The implementation of hard constraints serves to guarantee fundamental performance standards specific to rural off-grid contexts (e.g., maintaining 95% system availability or a 0.95 energy index ratio) [4]. The current frontier focuses on deep integration of MCDM for rural energy systems, embedding algorithms directly into the decision lifecycle. A standard paradigm couples front-end multi-objective optimization with back-end hybrid weighting. Examples of such optimization include the generation of a Pareto frontier to balance high battery storage costs against rural blackout risks [99,100]. This approach effectively bridges technical feasibility with the localized decision preferences of village committees and off-grid investors. Additionally, fuzzy logic is increasingly used to handle the high social uncertainty inherent in remote rural projects, such as unpredictable agricultural load profiles, fluctuating household incomes, and subjective community acceptance [5].

Despite the theoretical advancements in MCDM methods, their practical application in rural energy systems faces profound challenges, particularly in developing countries. First, the implementation of advanced data-driven models is heavily constrained by severe data scarcity. Remote villages often lack historical meteorological records and high-resolution load profiles, forcing a reliance on deterministic assumptions that amplify system uncertainty [6]. Second, extreme economic constraints create an “affordability override.” In impoverished regions, high initial capital costs act as strict survival thresholds rather than negotiable weights. This extreme price sensitivity [9,26] often paralyzes complex algorithms that attempt to mathematically trade off basic economic viability for higher environmental or technical performance. Finally, significant stakeholder capability barriers exist. Modern MCDM methods increasingly emphasize participatory governance and the direct involvement of local villagers [98]. However, local communities in developing contexts frequently lack the technical literacy required to interpret complex algorithmic trade-offs or Pareto frontiers. This disparity hinders genuine community acceptance and risks reverting to top-down decision-making that marginalizes local realities.

Looking forward (Figure 20), research on the multi-criteria evaluation of rural energy systems must transition from static to dynamic weighting frameworks. This requires adaptive mechanisms where indicator weights evolve alongside the progressive phases of rural electrification (e.g., shifting priority from basic electricity access in the early stages to income generation and system profitability in later stages), utilizing Bayesian updates as data accumulates. Therefore, robust decision-making methods (e.g., interval analytic hierarchy process) are critically needed to handle the severe data scarcity and unmetered load uncertainty typical of off-grid areas. Additionally, visual decision support tools must be developed to facilitate consensus among diverse rural stakeholders, such as local villagers, government agencies, and private operators [42,98]. Furthermore, the integration of advanced machine learning algorithms, such as reinforcement learning, is expected to elevate the MCDM from a purely computational technique to a dynamic, community-centric governance tool capable of autonomous adaptation to real-time rural microgrid conditions and demographic shifts.

4. Conclusions and Outlook

A systematic bibliometric analysis and critical review of the assessment indicators and methods for rural energy systems were conducted. Specifically, the progression of definitions, regional variations, and coupling optimization characteristics of economic, technological, environmental, and social indicators and the MCDM methods were examined across the literature from 2010 to 2025. The following key conclusions are drawn.

• The assessment of rural energy systems has experienced rapid growth, shifting its core focus from basic technical feasibility toward comprehensive techno-economic evaluations. While collaborative networks in China are well-established, international cooperation remains highly fragmented, highlighting a critical need for cross-border integration to address global rural electrification challenges.
• Economic and technical dimensions continuously dominate current assessment indicators for rural energy systems. Economically, contemporary assessments synthesize the levelized cost of energy, net present cost, and internal rate of return to balance unit generation efficiency, total financial liability, and investment viability. This integration is accompanied by an increasing internalization of socio-environmental externalities. Technically, evaluations have evolved from singular reliability constraints to comprehensive matrices assessing overall system robustness. Furthermore, evaluation priorities exhibit significant regional divergence: developing regions prioritize fundamental affordability and basic energy access, whereas developed regions focus on rural grid stability and market risk resilience.
• Environmental and social evaluations of rural energy systems have expanded to life-cycle-oriented frameworks. Modern environmental assessments prioritize life-cycle carbon footprints over merely direct operational emissions, seeking an optimal balance between renewable fractions and marginal abatement costs. Meanwhile, social evaluations for rural energy systems have been increasingly integrating macro-economic indicators, such as the human development index and job creation, directly into techno-economic optimization. However, a significant quantitative–qualitative imbalance of current indicator applications for rural energy systems persists, severely marginalizing subjective factors like social acceptance. Overcoming methodological disparities of indicator selection and integration requires developing unified assessment indicators for rural energy systems to ensure a balance among decarbonization, economic affordability, and holistic human well-being.
• To navigate the inherent conflicts among these diverse indicators, multi-criteria decision-making methods have transitioned from early tool-driven models to advanced hybrid mathematical approaches. These modern methodologies effectively balance socio-environmental trade-offs and address inherent off-grid uncertainties in rural energy systems. The decision-making paradigms of rural energy systems must evolve into dynamic, community-centric governance tools in the future. It is necessary to introduce an adaptive weighting mechanism that comprehensively reflects the progressive phases of rural energy systems electrification. This mechanism should be supported by advanced machine learning and visual decision-support platforms to foster consensus among diverse stakeholders and ensure the long-term viability of rural energy systems.

Finally, it should be noted that the scope of this study is inherently limited by the selected search keywords and the exclusive reliance on the Web of Science database. Therefore, relevant research falling outside these specific search boundaries may have been omitted, highlighting the need for future studies to employ broader search strategies and integrate multiple databases.