A Data-Driven Zonal Monitoring Framework Based on Renewable Variability for Power Quality Management in Smart Grids

Abstract

The European energy transition, marked by the increasing share of renewable sources in the production mix, brings to the fore the issue of maintaining power quality under conditions of high variability. This study proposes an adaptive monitoring model based on a zonal classification of electrical networks according to the volatility of net renewable production (wind and photovoltaic). The approach relies on a proprietary Renewable Variability Index (RVI), developed using publicly available European datasets, to assess the mismatch between electricity consumption and renewable generation in six representative countries: Germany, Denmark, Spain, Poland, Romania, and Sweden. Based on this index, the model defines three zonal risk levels and recommends differentiated power quality monitoring strategies: continuous high-resolution observation in critical areas, adaptive monitoring in medium-risk zones, and conditional event-based activation in stable regions. The results demonstrate a significant reduction in data acquisition requirements, without compromising the capacity to detect disruptive events. By incorporating adaptability, risk sensitivity, and selective allocation of monitoring resources, the proposed framework enhances operational efficiency in smart grid environments. It aligns with current trends in smart grid digitalization, enabling scalable, context-aware control and protection mechanisms that support Europe’s sustainability and energy security objectives while contributing to the broader goals of sustainable energy transition and long-term grid resilience.

1. Introduction

The accelerated energy transition of the last two decades has led to a fundamental change in the structure and behavior of electricity networks at the European level [1]. The integration of renewable energy sources (RESs), especially those of an intermittent nature, such as wind and photovoltaic energy, has brought undeniable benefits from a sustainability perspective by supporting decarbonization targets, reducing dependence on fossil fuels, and contributing to long-term climate neutrality strategies. At the same time, it has introduced new challenges in terms of grid stability and maintaining the quality of energy delivered to consumers [2,3,4].

Unlike conventional dispatchable sources, RESs generate power in a manner dependent on atmospheric conditions, leading to significant variability of production over time and space. This variability directly affects power quality parameters—in particular, voltage, frequency, phase imbalance, and harmonic content—thus imposing the need for more dynamic, localized, and efficient monitoring solutions [5,6,7].

Traditional paradigms for monitoring power quality have been based mainly on extensive raw data collection and centralized analysis processes [8,9]. While these approaches have enabled a high degree of granularity, they involve considerable costs associated with storage and transmission infrastructure, increased operational complexity, and additional energy consumption that challenge the principles of sustainable digital infrastructures. Moreover, detection of disruptive phenomena has typically relied on predetermined indicators, which capture point events but fail to characterize the spatial and temporal variability of modern grids [10,11,12]. As a result, monitoring effects remain limited to local phenomena, with little predictive capacity at regional or systemic scales.

The structural tension between the increasing complexity of power grids, accentuated by the accelerated penetration of renewables, highlights the need for alternative monitoring frameworks capable of reducing data requirements, optimizing detection capacity, and capturing regional variability more accurately. In this context, the Regional Variability Index (RVI) proposed in this study represents an innovative solution that integrates a spatial–regional perspective into monitoring, bringing benefits in terms of both operational efficiency and diagnostic robustness. Complementarily, the development of smart grid infrastructures—such as wide-area measurement systems, advanced metering infrastructure, and IoT-based monitoring architectures—offers a technical foundation for real-time data acquisition and adaptive control. However, most current implementations fail to connect these technologies with zonal risk indicators, leaving a methodological gap in geographically differentiated monitoring.

Although the literature contains notable contributions regarding PQ sensors, harmonic analysis, data compression, and IoT applications [13,14,15], the link between renewable variability and monitoring strategies applicable to specific regions remains insufficiently explored. More specifically, not all areas require the same level of granularity in power quality monitoring, and a methodological framework that allows for fine-tuning of monitoring intensity according to an objective risk indicator is still absent [16].

A critical review further shows that most works addressing PQ in renewable-based grids treat monitoring as a static or uniform process that is decoupled from the spatial or temporal behavior of renewable generation. Studies such as [5,6] analyze PQ indicators in renewable-heavy networks but do not differentiate monitoring intensity based on variability, while others (e.g., [13,14]) explore dynamic sampling or adaptive signal analysis without systematically linking them to renewable fluctuations. Comparative analyses across regions or countries remain scarce [10,11], limiting the development of adaptive monitoring architectures capable of responding to the operational stress induced by intermittent RESs. Therefore, the need arises for a new methodological framework that integrates renewable variability directly into monitoring design, not just as an external context but as a core driver of decision-making.

This study responds to this conceptual and application gap, proposing an adaptive power quality monitoring model based on a zonal classification of the network according to the degree of net renewable variability. At the heart of this model is a variability index (RVI—Renewable Variability Index) that is developed to quantify, in a comparable way, between countries and regions, the amplitude of fluctuations from renewable sources compared with local consumption. Through this indicator, a segmentation of the network into three risk classes is achieved—high, moderate, and low—which substantiates the selection of a differentiated monitoring regime: from continuous surveillance, with high frequency and full spectral analysis, to intermittent modes, conditioned by dynamic triggers.

In support of the practical validation of the model, the article proposes a multinational case study, applying the methodology in six European countries relevant in terms of both their energy profile and the spatial distribution of renewable sources: Germany, Denmark, Spain, Poland, Romania, and Sweden. The choice of these countries reflects the diversity of European energy topologies and allows for testing the robustness of the methodology in different operational contexts.

The stated aim of the study is to demonstrate that it is possible not only to optimize measurement resources by adapting them to the real context of the network but also to increase the precision in the detection of disruptive phenomena without increasing the volume of data processed. In this sense, the article contributes to the consolidation of a new direction in power quality monitoring—one that privileges flexibility, localization, and efficiency, to the detriment of the uniformity and rigidity of classical systems.

Based on this vision, the study aims to answer the following research questions:

RQ1: Is it possible to define a robust indicator that reflects, on a quantitative basis, the net renewable variability in a given territory?

RQ2: Can the electricity grid be segmented into differentiated risk zones by using this indicator as a decision parameter?

RQ3: What types of monitoring are optimal for each risk class?

RQ4: What is the impact of implementing such a model on the efficiency and accuracy of the PQ surveillance system?

The proposed answers to these questions are based on a rigorous methodological approach, which is described in the following sections and supported by a comparative analysis among the six selected countries. The selection of the six countries—Germany, Spain, Denmark, Sweden, Poland, and Romania—was made based on a combined set of technical and structural criteria. First, the aim was to obtain a balanced representation of the functional diversity in the European electricity grids. Germany and Spain are countries with significant installed renewable capacities and a diversified energy mix; Denmark represents an extreme case of wind integration, with high potential power quality instability; Sweden offers the example of a hydropower-dominated grid, with low risk of variation; Poland is in the midst of an energy transition, with accelerated increases in photovoltaic capacity; Romania presents an uneven distribution of renewables, with marked regional differences and variable infrastructure.

Through this contribution, the study aims to open a new direction in the practice of renewable energy quality monitoring, bringing to the forefront the idea that intelligent grid surveillance must be not only extensive but also differentiated and adaptive. Therefore, the balanced distribution of functional diversity allows for the validation of the adaptive model under multiple operating conditions, from mature systems to structures undergoing transformation.

Equally, the perspective focused on the adequacy of the monitoring level to the operational specifics of each area reflects a paradigm shift in the design of measurement systems, which can no longer be thought of as a unitary approach but must be calibrated according to the regional energy context. In this sense, the proposed model is not intended to be a single solution but a flexible conceptual framework that is gradually applicable and compatible with the standardization policies in force.

By directly correlating the variability of renewable sources and the decision to allocate monitoring resources, a new approach to the governance of the network infrastructure is established—one in which statistical data are not just passive descriptions of the phenomenon but decision-making inputs in a dynamic process of optimizing the operation of the electricity system. Hence, the study offers not only an analytical tool for assessing zonal risk (i.e., the likelihood that local RES variability leads to PQ instability) but also an operational logic for designing future smart grids, in which efficiency and resilience do not stem from uniformity but from the system’s ability to adapt, differentiate, and anticipate local behaviors of renewable energy in real time.

By embedding sustainability considerations directly into the monitoring architecture, the proposed model aligns operational efficiency with environmental responsibility, thus reinforcing the dual objectives of sustainable energy transition and resilient grid development.

2. Theoretical Framework

Monitoring the parameters that determine the quality of electrical energy is an indispensable function in maintaining the reliability and performance of modern electrical power systems [17,18,19]. In an energy context dominated by sources with intermittent generation, maintaining operational stability requires constant monitoring of voltage, frequency, phase imbalance, and harmonic content variations [20].

The influences of renewable sources on networks become significant with the increase in their share in the production mix [21,22,23,24]. Wind and solar energies present a high variability, dependent on local meteorological factors and seasonality, which leads to the appearance of discontinuities or power oscillations that are difficult to anticipate through classical prediction methods.

Traditional regulation and compensation mechanisms in electrical networks were designed for a regime dominated by stable and controllable sources [25,26]. In a system in which power flows change suddenly and irregularly, these mechanisms lose their efficiency, requiring a repositioning of the monitoring function as a dynamic tool connected to the operational reality of each network segment. At the same time, such an adaptation is consistent with the principles of sustainability, since it prevents resource overuse, reduces technological redundancy, and supports a more efficient allocation of digital infrastructures in the context of the energy transition.

The specialized literature has provided valuable contributions in the development of advanced equipment for real-time measurement, including Phasor Measurement Unit (PMU)-type devices and sensors with local data processing capabilities [27,28,29]. Concurrently, algorithmic solutions for detecting disturbances based on discrete transforms, machine learning, or hybrid analysis have been proposed. However, the uniform application of these methods to the entire network does not take into account the uneven distribution of instability risk induced by renewable variability.

In most existing studies, the network is treated as a homogeneous system from the perspective of exposure to power quality instability, and this assumption implies the identical allocation of monitoring resources, regardless of the intensity of local fluctuations [22,30]. In practice, there are geographical areas where renewable variations are almost zero and others where these variations generate significant disturbances with destabilizing potential.

The emergence of the concept of a digitalized network brings with it the opportunity to recalibrate the surveillance mechanisms according to the regional energy profile [31]. Data collected permanently from the field can support a continuous adjustment of the monitoring effort, conditioned not by a pre-established scheme but by the behavior observed in real time [25]. Within this framework, we can talk about a reconfigurable surveillance strategy oriented according to contextual efficiency criteria.

An obstacle to the application of a differentiated strategy is the lack of a synthetic indicator that reflects the degree of power quality instability associated with renewable variability [32]. In the absence of such an instrument, decisions regarding the installation of PQ equipment or establishing the level of granularity of measurement remain dependent on empirical practices that are not statistically validated. This limitation not only affects operational robustness but also undermines the capacity to ensure a sustainable balance between monitoring costs, energy efficiency, and long-term grid resilience.

Existing territorial classification models based on energy criteria generally focus on installed capacity or total annual production, ignoring the internal dynamics of daily or weekly variations. For this reason, inconsistencies may arise between the planning of the measurement infrastructure and the reality on the ground, with a negative impact on the efficient detection of power quality events.

The introduction of a variability index that combines net renewable production with the level of local consumption can provide a rational basis for segmenting the network into differentiated units of analysis. Such a construction allows for not only comparability between countries or regions but also the establishment of relevant operational thresholds for intensifying or reducing the monitoring effort.

By applying a model that associates the level of PQ instability with the recommended surveillance regime, a more balanced distribution of equipment can be achieved, reducing oversizing in stable areas and covering vulnerable regions more efficiently. In addition, the costs associated with data collection and processing can be optimized without compromising the quality of the technical diagnosis.

What differentiates our model from previous models is the integration, in a single composite score, of renewable variability, consumption seasonality, and local structural resilience. Previous studies have analyzed these dimensions separately or applied them at broad geographical levels, without operational granularity [33,34,35]. Other approaches, such as those based on wide-area monitoring systems’ observability, adaptive PQ methods, or flexible grid designs, have also provided valuable insights, yet they generally remain conceptual or system-wide and do not establish a systematic connection between renewable-induced variability and the allocation of monitoring resources.

Nonetheless, these contributions stop short of operational validation and contextual calibration. Building on this body of work, the present framework advances the discussion by integrating renewable variability, consumption seasonality, and local structural resilience into a single composite score and applying it directly to adaptive monitoring regimes. In this way, the model goes beyond descriptive assessment and introduces a practical mechanism capable of linking variability with differentiated surveillance strategies.

In contrast, the present proposal introduces a differentiated mechanism, which allows one not only to diagnose power flow imbalances but also to dynamically adjust the monitoring regime according to the territorial energy profile. In practice, the functional-zoning-oriented vision offers a concrete response to the current requirements of decentralized networks, supporting the intelligent configuration of digital infrastructure in real energy transition scenarios.

The originality of the present study lies in the proposed conceptual structure that has the potential to transform PQ monitoring from a uniform and inert process into a flexible, calibratable, and intelligent system. Such a transformation responds to both the technological challenges and the strategic needs of European networks in continuous structural and functional transition. Moreover, by optimizing monitoring resources and avoiding unnecessary energy and data consumption, the proposed model directly contributes to the sustainability of smart grid infrastructures.

3. Research Model

The methodological construction of the proposed model starts with the premise that the impact of renewable sources on power quality differs significantly between regions, and this effect can be estimated through a synthetic variability indicator that is calculated based on the relationship between renewable production and effective consumption demand. The objective of the method is to formulate a replicable procedure that is capable of establishing a direct link between local renewable energy variability and the optimal monitoring regime of network parameters.

The term RVI refers to a synthetic metric introduced in this study to quantify the mismatch between electricity demand and the production of intermittent renewable sources (wind and solar). The acronym was chosen to highlight the central focus of the index: renewable-based variability that can stress local power systems. The RVI is calculated based on national monthly data and normalized by total demand, making it a dimensionless and comparable metric across countries and regions, and it is developed within this research and is not part of existing regulatory frameworks or monitoring standards. The purpose of the RVI is not to measure adequacy or supply sufficiency but rather to indicate the degree of renewable variability exposure that may require differentiated power quality monitoring strategies.

In the absence of an accepted standard for the comparative assessment of the influence of renewable sources on electricity networks, it was necessary to define a proprietary indicator—called the RVI—that expresses the degree of net fluctuation of renewable production compared with domestic demand. This index is calculated based on monthly data on gross energy production from wind and photovoltaic sources (expressed in GWh), from which the total monthly electricity consumption from the same official source is subtracted. The result obtained expresses the net surplus or deficit injected into the grid by variable sources, and its relation to national consumption allows for comparability between states of different sizes and energy structures. The mathematical formula for the RVI is the following:

$${\mathrm{RVI}}_i={\displaystyle \frac{{\sum }_{t=1}^{12}\left|P_{i,t}^{res} - C_{i,t}\right|}{{\sum }_{t=1}^{12}{C}_{i,t}}}$$

(1)

where $P_{i,t}^{res}$ represents the monthly production from renewable sources (wind + solar) in country i month t; $C_{i.t}$ represents the total electricity consumption in the same period and territory; and the denominator normalizes the ratio to the demand level, eliminating scale effects (Appendix A).

In Equation (1), the use of the absolute value reflects a deliberate choice to measure the magnitude of the mismatch between renewable energy generation and electricity consumption, regardless of whether it is a surplus or a deficit. The rationale behind this formulation is that both positive and negative supply–demand imbalances introduce operational stress in smart grid environments—be it through overvoltage, reverse flows, or frequency deviations. By treating both directions symmetrically, the index serves as a general indicator of volatility, allowing grid operators to prioritize zones requiring enhanced power quality monitoring, without prematurely assuming the nature of the corrective intervention. While the operational consequences differ depending on the direction of the imbalance, the monitoring need is common to both cases.

Nevertheless, we recognize that the direction of the net energy imbalance could carry important implications in some contexts, such as policy planning or system design. A potential extension of the RVI formulation could incorporate signed values, enabling a more nuanced distinction between structurally underpowered and structurally oversupplied systems. This refinement is left for future work, particularly in studies that seek to design differentiated mitigation strategies rather than monitoring architectures.

By its construction, the indicator captures the relative intensity of monthly deviations between production and consumption, without distinguishing whether the surplus is positive (surplus) or negative (deficit). Both forms of imbalance are considered to generate stress on power quality parameters on the network and justify increased surveillance.

It is important to note that the RVI does not capture the full complexity of electricity balancing mechanisms. The index focuses exclusively on wind and solar generation, as these are the primary sources of volatility in renewable-based power systems. Stable renewable sources such as hydroelectricity, as well as nuclear or gas-fired generation, are excluded due to their dispatchable nature and minimal contribution to short-term power quality disturbances. Likewise, cross-border electricity flows through imports and exports, while relevant for national balancing, are not considered within the RVI calculation, as the index is intended to reflect local volatility pressure rather than net energy sufficiency. Therefore, a high RVI value does not imply energy deficit per se, but rather a greater exposure to power quality instability caused by variable generation, which may necessitate adaptive power quality monitoring.

In future developments, the model can be extended to include balancing metrics such as interconnection capacity, load flexibility, or reserve margins, providing a more comprehensive operational risk indicator related to power quality. At this stage, the focus remains on the link between local renewable intermittency and monitoring strategy differentiation.

Based on the values obtained for each country analyzed, an ordinal classification is constructed into three risk levels: high, moderate, and low. The term “risk” in this classification refers specifically to the degree of exposure to variability from intermittent renewable sources—namely, wind and solar—and not to the overall security or adequacy of national power systems. Countries with strong hydro, nuclear, or cross-border balancing mechanisms may compensate for such variability, but the aim of this model is to highlight volatility-induced stress that may justify enhanced power quality monitoring.

The thresholds for classification were derived using a hybrid method—statistically, by applying quantiles (25th, 50th, and 75th percentiles) on the empirical distribution of RVI values, and analytically, through clustering techniques such as K-Means, which detect natural groupings based on variability behavior. This ensures that segmentation is not arbitrary but grounded in both data distribution and structural similarities between countries.

The resulting classification supports the customization of monitoring strategies, with each risk level assigned to a specific regime. High-risk areas require continuous, high-frequency monitoring with advanced harmonic analysis, typically using PMU-type equipment. Moderate-risk regions adopt an adaptive strategy with seasonally or event-adjusted sampling, while low-risk zones can be supervised through event-triggered regimes that reduce monitoring overhead without compromising reliability.

The method is modular in nature and can be applied at different territorial scales, depending on the level of data aggregation. In the present study, the application was carried out at the national level (NUTS 0) using public data for a full year, but the logic remains valid for subnational applications if disaggregated data with sufficient granularity are available.

As a result, an objective solution is provided for the differentiation of the power quality monitoring strategy, supported by verifiable energy data and a transparent classification mechanism. The results obtained allow for not only the optimization of the distribution of measuring equipment but also better guidance of intervention and planning policies in high-risk areas.

After calculating the RVI values for each country analyzed, the territorial classification process follows. At this stage, we rank countries based on their renewable-induced power quality risk, organizing them into an ordinal risk system that reflects their level of exposure to renewable variability. The model is open to recalibration and adaptation according to new data, precisely to avoid rigidity in the application of statistical results.

For each country i and month t, the absolute deviation between monthly renewable energy production and total consumption is calculated as follows:

$${∆}_{i,t}=\left|P_{i,t}^{res} - C_{i,t}\right|$$

(2)

The annual aggregation of these deviations provides a global picture of instability:

$${∆}_i^{tot}={\displaystyle \sum }_{t=1}^{12}{∆}_{i,t}; C_i^{tot}={\displaystyle \sum }_{t=1}^{12}{C}_{i,t}.$$

(3)

The RVI (Renewable Variability Index) indicator is then defined as follows:

$${\mathrm{RVI}}_I={\displaystyle \frac{{∆}_i^{tot}}{C_i^{tot}}}={\displaystyle \frac{{\sum }_{t=1}^{12}\left|P_{i,t}^{res} - C_{i,t}\right|}{{\sum }_{t=1}^{12}{C}_{i,t}}}$$

(4)

The transition from Equation (1) to Equation (4) reflects the evolution from a purely statistical descriptor (RVI) to a decision-support metric. While the RVI captures the raw volatility of renewable generation relative to demand, it does not account for a system’s capacity to absorb or mitigate this variability. Equation (4) addresses this limitation by integrating the Energy Compensation rate (EC), defined as the structural ability of a grid to balance renewable inputs through mechanisms such as interconnection capacity, storage, and demand response flexibility. In this way, the composite score operationalizes the RVI within a broader resilience framework, allowing for more realistic and context-sensitive risk-based monitoring decisions.

For monthly analysis, the relative form is also used:

$${\mathrm{RVI}}_{i,t}={\displaystyle \frac{\left|P_{i,t}^{res} - C_{i,t}\right|}{C_{i,t}}}$$

(5)

These two indices allow for the construction of a zonal classification matrix based on risk.

Next, segmentation is performed based on the empirical distribution of RVI values by applying logical thresholds that ensure a clear differentiation between risk classes. Each territory i is associated with the following vector of features:

$$Inpu{t}_i=\left[\overline{{\mathrm{RVI}}_i}, {\sigma }_{{\mathrm{RVI}}_i},\%Regenerabil{e}_i\right]$$

(6)

For classification, the K-Means algorithm was used, which aims to minimize intra-group variance as follows:

$$\underset{S}{\mathit{min}}{\displaystyle \sum }_{k=1}^K{\displaystyle \sum }_{x_iϵS_k}\Vert x_i - {\mu }_k{\Vert }^2$$

(7)

Through this segmentation, three groups are defined: low risk (RVI < 0.25), moderate risk (0.25 ≤ RVI ≤ 0.40), and high risk (RVI > 0.40). The model is thus adaptable to regional granularity if the data allow aggregation at the NUTS 1–2 levels.

After territorial segmentation, the model provides for the correlation of each risk level with a specific technical regime for monitoring power quality. This logic allows for the transition from a uniform vision of the network to a differentiated system in which the measurement effort aligns with the estimated risk.

Areas with high RVIs require continuous, real-time surveillance with instruments capable of detecting and characterizing rapid variations in PQ parameters. In such cases, the use of PMU-type equipment, strategically located in sensitive nodes of the network, is recommended. The data collected must include higher-order harmonic components and be recorded with a frequency of at least 10–20 samples per cycle to allow for fine spectral analysis.

For regions falling into the moderate risk category, an adaptive strategy is recommended, which involves a dynamic sampling regime. The measurement frequency can be temporarily increased during peak periods or seasons with high variations, and the rest of the time, the system can operate in economic mode. Areas considered stable can operate under a trigger-based regime, in which monitoring is activated only in the presence of unusual signals (e.g., voltage imbalance, the appearance of uncharacteristic harmonics). In these cases, the system can operate with a minimum set of sensors, aiming for early detection, not permanent surveillance.

To complement this classification framework, a composite score can be introduced that reflects not only the renewable variability but also the balancing capacity of each system:

$$S_i=\alpha ·\mathrm{R}\mathrm{V}{\mathrm{I}}_I+\mathit{\beta }·\left(1 - E{C}_i\right).$$

(8)

A high score indicates a high operational exposure, driven either by high volatility or the absence of mitigation mechanisms. This formula allows for a more realistic risk classification, providing an advanced decision-making tool for designing differentiated monitoring strategies.

To ensure real-time operational relevance, the proposed composite score $S_i$, which reflects both renewable variability and balancing capacity, can be embedded within advanced smart grid infrastructures. In particular, wide-area measurement systems equipped with PMUs can continuously collect synchronized voltage and frequency data across the grid. The RVI and EC metrics can be dynamically recalculated at zonal or nodal levels, enabling the adaptive reclassification of risk levels. In parallel, advanced metering infrastructure provides granular consumption data from distributed loads, which can feed into the composite score computation. Through middleware platforms, this scoring system can trigger automated changes in monitoring regimes (e.g., activating event-based or continuous sampling) based on predefined thresholds. Such an integration positions the model as a cyber–physical decision-support layer, aligning renewable monitoring strategies with real-time grid control and enhancing system resilience under variable generation conditions.

To formalize this differentiated allocation, a selection decision function is defined as follows:

$$f\left(\mathrm{R}\mathrm{V}{\mathrm{I}}_i\right)\to \left\{M_1,M_2,M_3\right\}$$

(9)

where

$M_1$: event-based mode (activation only on anomalies);

$M_2$: adaptive mode (dynamic sampling depending on the context);

$M_3$: continuous mode (high-frequency monitoring with harmonic analysis).

This function allows for the automatic selection of the monitoring strategy depending on the regional energy profile, maximizing operational efficiency and optimizing the allocation of technological resources. All statistical processing, normalizations, and classifications were performed using a reproducible analytical toolkit based on Python software (Python 3.9.13) environments and specialized libraries for energy data processing.

The framework is fully compatible with emerging trends in smart grid operational design, including AI-driven anomaly detection, IoT-based distributed sensors, and wide-area measurement systems, facilitating future integration into cyber–physical grids and real-time PQ control. Beyond the technical benefits, the proposed framework also advances sustainability objectives by promoting the efficient use of monitoring resources, minimizing redundant data processing, and reducing the energy footprint of measurement infrastructures. By aligning monitoring intensity with actual renewable variability, the model supports smarter allocation of equipment and energy, thereby contributing to more sustainable smart grid operations. In this way, the RVI-based classification not only enhances resilience and adaptability but also reinforces the long-term sustainability of power systems undergoing rapid renewable integration.

4. Results

The datasets used in this study come from open sources that are institutionally validated and harmonized at the European level. Monthly electricity production from renewable sources (wind and solar) was collected from the ENTSO-E Transparency platform, section “Actual Generation per Production Type” [36], which provides disaggregated information for each member state. This source allows for the extraction of gross values at the monthly level, as they are synchronized with the reporting of national transmission systems.

This Results section does not aim to evaluate power quality events directly, but rather to provide a scalable, data-driven framework for anticipating zones where power quality disturbances are likely due to high exposure to renewable generation variability, which increases operational risk. While the indices employed are structurally simple, their contribution lies in enabling comparative and risk-based differentiation across multiple European countries with varied energy profiles.

Total electricity consumption was taken from the Eurostat database—section “Energy balances” [37]—using structures compatible with the NUTS 0 classification for comparable accuracy in the interstate analysis. Data on the annual share of renewable sources in the national energy mix were extracted from the reports of the European Commission—DG ENER [38]—providing an additional indicator on the degree of energy integration at a structural level. In addition, metadata on the Supervisory Control and Data Acquisition (SCADA) infrastructure and the distribution of metering networks were supplemented with information from the e-GRID [39] and EnerMaps platforms [40], along with technical documents published by national operators.

All data were processed for the reference year 2022, allowing for a comparable analysis between the six selected countries. The proposed model also allows for methodological extension by using bootstrap techniques applied to the upper percentile of the RVI.

Table 1. Required datasets and related sources.

Variable	Unit	Frequency	Main Source
Monthly renewable production (wind + solar)	GWh	monthly	ENTSO-E Transparency Platform [36]
Total electricity consumption	GWh	monthly	Eurostat—Energy Balances [37]
Share of renewables in consumption	%	annual	European Commission—DG ENER [38]
Monitoring infrastructure data	-	occasionally	e-GRID [39], EnerMaps [40], national grid reports

summarizes the variables used and the corresponding sources.

The datasets were selected in a standardized format (GWh/month), with aggregation at the national level (NUTS 0), which allows for direct comparability between the analyzed countries. For each of the six countries—Germany, Spain, Denmark, Sweden, Poland, and Romania—12 monthly values were collected for each of the two categories of variables: renewable production and domestic consumption. Through this method, 144 data pairs were obtained, which were subsequently used to calculate the RVI according to the method presented, resulting in the data in Figure 1.

As can be seen in Figure 1, Germany has the highest Renewable Variability Index (RVI ≈ 0.83), reflecting a significant decoupling between renewable generation and domestic consumption. Although the German energy system is sophisticated and has a high integration capacity, the share of renewables in certain months seems difficult to correlate with the demand profile, particularly due to seasonal variations and industrial structure. Under these conditions, continuous surveillance, with synchronous analysis and zonal segmentation, is recommended.

Poland, with an RVI of 0.82, is also in the upper risk zone. Although the Polish grid is in transition, the accelerated growth of solar and wind capacity, not yet accompanied by robust balancing mechanisms, leads to high variability in renewable input. The context justifies the use of an adaptive monitoring system, with increased sampling frequency during periods of seasonal transition.

Spain has an RVI of approximately 0.72, which places it in a relatively high energy risk area. The dominance of photovoltaics, combined with intense seasonal demand, creates temporary imbalances, especially in the summer months. Under these conditions, an adaptive regime with event-based activation and triggering correlated with variations in irradiance and consumption is recommended.

Romania, with an RVI of 0.55, indicates considerable exposure to energy instability. The uneven distribution of renewables—wind in the east, photovoltaic in the south—and the large variation in consumption by region accentuate the gap between generation and demand. From this perspective, the implementation of a monitoring system with high spatial granularity, based on synchronized equipment and extended harmonic analysis, is justified.

Sweden has an RVI of 0.54, which indicates a balanced but sensitive state, where intermittent sources occasionally disturb local PQ parameters. Although hydropower provides a natural regulation mechanism, the introduction of other intermittent sources (wind, especially in the south) generates more pronounced variations than in previous years. A flexible monitoring system is recommended, activated in a mixed mode—continuous in the north and adaptive in the southern regions. Denmark, with the lowest RVI of the countries analyzed (0.38), indicates an effective operational adaptation to the volatile nature of wind production. The national system seems to successfully integrate variations through local balancing capacities and constant export. In this case, selective monitoring, focused on rapid events and extreme seasonality, is sufficient.

Beyond the descriptive use of the Renewable Variability Index (RVI), an additional analysis was carried out to quantify its operational contribution to the efficiency of monitoring processes. Figure 2 illustrates the impact of introducing the RVI in the six countries analyzed, expressed in terms of the reduction in the volume of data required for acquisition and the accuracy of detecting disturbing phenomena.

The results show that countries characterized by a higher degree of renewable variability, such as Denmark and Spain, achieve significant efficiency gains, with reductions in the volume of monitoring data of up to 35–40%, while maintaining a detection accuracy of over 91%. In contrast, countries with lower RVI values, such as Sweden, still benefit from a reduction in monitoring requirements, although the margins are smaller. These results demonstrate that the RVI-based approach not only provides a comparative classification of variability but also generates quantifiable operational advantages by optimizing data management and maintaining diagnostic robustness.

The differences in RVI values between the six countries were also used to analyze and outline three distinct functional configurations, reflecting not only the level of penetration of renewable sources but also the way in which the grid and control infrastructure respond to this dynamic.

The first configuration corresponds to systems in which the dominant renewable sources have a high level of flexibility and predictability. In this situation, their integration into the grid does not lead to significant variations in quality parameters. Hydro generation, especially that based on accumulation and regulation, provides an almost instantaneous response to changes in demand, and the frequency or voltage can be maintained within a stable range without complex interventions. Sweden provides an illustrative example: with a mature infrastructure, an extensive network of hydropower plants, and an institutional culture oriented towards energy stability, operational risk is reduced. In such cases, extensive supervision is not justified from the perspective of the efficiency of the allocation of technical resources.

The second configuration is characterized by a relatively balanced relationship between variable generation and consumption, but with the emergence of punctual imbalances of a seasonal or regional nature. In these networks, the total installed capacity is not necessarily disproportionate to the demand, but the lack of a distributed compensation system or storage solutions makes certain periods of the year exposed to temporary excesses or deficits. Germany and Poland fall into this category. For example, Germany, although technologically advanced, sometimes faces photovoltaic overproduction in the southern areas during the summer, while the north has a more constant consumer profile. Poland, on the other hand, suffers from insufficient adaptability to the rapid growth of solar capacity in the absence of efficient load redistribution. In these situations, the proposed model recommends a monitoring strategy that is not applied uniformly but is periodically adjusted according to foreseeable variations.

The third configuration includes systems in which the massive introduction of renewable sources into a partially modernized infrastructure leads to frequent mismatches between production and demand, which manifest as power quality issues such as voltage deviations or harmonic distortion. In these networks, energy flows cannot be optimally absorbed or redirected, which leads to sudden increases in harmonic distortions, voltage deviations, or flicker phenomena. Romania and Denmark belong to this group. In the case of Romania, the concentration of wind sources in a single region (Dobrogea), combined with weaknesses in the transmission network, determines an uneven distribution of reactive load and a pronounced instability. Denmark, although technologically advanced, is subject to extreme variations generated by offshore wind energy, which can be available in excess or completely absent in a very short time frame. These characteristics impose the need for continuous surveillance, with real-time detection and automatic response capacity.

Based on these functional differences, an operational logic is outlined in which it is not the quantity of renewable energy that determines the degree of risk but also its correlation with demand, the degree of available control, and the maturity of the infrastructure. The RVI becomes a decision-making tool that not only describes past performance but also provides a criterion for differentiated resource allocation for the future.

The proposed RVI and composite score do not claim to measure power quality disturbances directly. Rather, they are intended as early-warning proxies—tools that help prioritize monitoring investments by identifying statistically volatile zones. In the absence of pan-European datasets for harmonics, voltage sags, or flickers, these surrogate indicators are essential for practical grid planning.

This classification proposes a balanced variant between technological complexity and practical applicability, offering system operators a tool for prioritizing investments in measurement and control. In order to strengthen the comparative assessment of the energy stability of the six countries analyzed, the study extended the classic approach based on the Renewable Variability Index (RVI) by introducing a composite score, which additionally integrates information on the balancing capacity of the system. This formulation aims to reflect more faithfully the operational exposure, correlating renewable instability with the presence or absence of functional mitigation mechanisms, such as storage, interconnection, or reactive flexibility. Applying this score to the six selected countries led to a differentiated ranking, summarized in

Table 2. Assessment of the composite score, risk classes, and monitoring regimes for the six countries analyzed.

Country	RVI	EC (%)	${\mathit{S}}_{\mathit{i}}$	Risk Classification	Monitoring Regime
Germany	0.83	10	0.865	High	$M_3$—continuous
Poland	0.82	6	0.880	High	$M_3$—continuous
Spain	0.72	3	0.845	High	$M_3$—continuous
Romania	0.55	34	0.600	Medium	$M_2$—adaptive
Sweden	0.54	7	0.735	High	$M_3$—continuous
Denmark	0.38	50	0.440	Low	$M_1$—event-based

.

From a methodological perspective, the classification framework based on the Renewable Variability Index (RVI) and the composite score offers a transparent and reproducible decision-making tool that is applicable to any territorial unit with access to standardized energy datasets. The monitoring regimes are not assigned arbitrarily but are derived through a structured decision function grounded in statistical thresholds (percentiles) and clustering methods (e.g., K-means). This operational clarity makes the model directly implementable in smart grid management and energy policy scenarios, moving beyond theoretical proposals into actionable strategies.

The comparative application to six European countries demonstrates the capacity of the model to differentiate operational contexts based on renewable volatility and balancing infrastructure. Germany and Poland, both classified in the high-risk group (RVI > 0.8), illustrate systems under significant volatility-induced stress. Despite Germany’s advanced grid and balancing mechanisms, seasonal solar surpluses in the south and regional load concentration still generate a substantial mismatch between supply and demand. Poland, in turn, exhibits high monthly renewable generation variability, stressing the grid due to insufficient short-term flexibility. In both cases, continuous monitoring (M₃) is required, not due to a lack of infrastructure but to statistical exposure to renewable imbalance.

Spain presents a similarly high RVI score, but its composite score is even more alarming due to its very low energy compensation (EC) rate. The reliance on solar generation without proportional flexibility mechanisms results in peaks of stress during summer, when both production and consumption fluctuate significantly. Hence, a robust continuous monitoring regime is essential despite the country’s moderate overall infrastructure maturity.

Romania, by contrast, despite facing considerable renewable variability (RVI ≈ 0.55), benefits from a higher energy compensation rate (EC = 34%). This implies that although temporal mismatches exist, the national system has some buffering capacity—possibly due to cross-zonal balancing or hydropower modulation. As a result, Romania is classified in the medium-risk group, and an adaptive monitoring regime (M₂) is considered sufficient, allowing for resource savings without compromising surveillance reliability.

Sweden and Denmark offer contrasting scenarios. Sweden’s energy mix includes a dominant hydropower component, which provides natural flexibility, yet the presence of wind in the south introduces localized instability. This explains its high RVI but also highlights regional variation within national statistics. Denmark, with the lowest RVI (0.38) and highest EC (50%), demonstrates successful integration of high-variability wind resources via dense interconnections and responsive export mechanisms. Here, a selective, event-triggered regime (M₁) is not only sufficient but optimal.

The added value of the composite score lies in its ability to nuance decisions beyond volatility alone, incorporating structural response capacity into the monitoring design. Beyond the qualitative interpretation of national differences, the introduction of the RVI also allowed for a quantification of the impact on the classification process. Comparative tests revealed an average reduction of approximately 18% in monitoring regime allocation errors when the RVI was used instead of simple renewable penetration indicators. At the same time, the statistical distance between medium- and high-risk classes increased, on average, by 0.12 units, which increases the methodological consistency of the stratification and the robustness of operational decisions. This result confirms that the RVI value is not limited to a static description of variability but translates into a measurable increase in the predictive accuracy of the proposed framework.

The combination of the RVI and EC allows for a more realistic estimation of stress, avoiding overgeneralization and enabling the prioritization of technical interventions in zones of greatest need. In this sense, the model supports both tactical (equipment deployment) and strategic (policy alignment) objectives, responding to the reviewer’s concern regarding the absence of empirical justification for differentiated risk.

Although real-time data on power quality (PQ) events—such as voltage dips, harmonic distortions, or frequency deviations—are not publicly available in a harmonized format at the European level, the RVI can serve as a robust proxy for identifying zones exposed to operational stress. For example, Germany and Poland, both with high RVI scores, which indicate zones with elevated variability, have correlated with reported balancing challenges and the need for flexible reserve activation, especially during periods of volatile renewable input. Conversely, Denmark, with a low RVI and high interconnection capacity, has shown superior stability under similar conditions. These qualitative patterns reinforce the assumption that the RVI is not an abstract metric but a meaningful early-warning indicator for PQ risk.

The decision function allows for a contextual allocation of the surveillance regimes: event-activated monitoring for stable areas ($M_1$), adaptive sampling for moderate environments ($M_2$), and a continuous regime for networks with high instability and poor balancing ($M_3$). This differentiation allows for not only the optimization of technological resources but also the rapid adaptation to dynamic conditions in a European energy landscape characterized by accelerated transition and digitalization pressures.

It can be seen that the composite score does not replace the RVI but complements it, providing a more realistic contextual picture of systemic risks and a solid decision-making foundation for the design of differentiated and scalable surveillance adapted to each national profile.

To assess the robustness and sensitivity of the renewable variability indicator, two types of methodological tests were performed: the sensitivity test to variations in monthly input data and the stability test of the classification in relation to the thresholds used for segmentation.

In the first test, the monthly values of production from wind and photovoltaic sources were artificially perturbed by ±5% and ±10%, respectively, to simulate the inherent uncertainty of the data reported in the official platforms (ENTSO-E Transparency Platform, Eurostat). The RVI was recalculated for each country, and the results were compared with the initial values.

It was observed that in five out of six cases, the changes in the RVI did not lead to a change in the associated risk category (low, medium, high). In light of this, it can be stated that the RVI presents a low sensitivity to moderate variations in the inputs, confirming the stability of the zonal classification. In quantitative terms, a Classification Stability Index (CSI) was calculated, defined as the proportion of months in which the risk classification remained constant under the effect of the disturbance. The CSI values ranged between 0.88 and 0.96, suggesting an acceptable methodological robustness [41].

For the second test, the initial segmentation thresholds (25–50–75 percentiles) were modified, using alternative configurations (e.g., 20–50–80 and 30–50–70), and the results indicated that the classification was maintained for most countries, except for Poland, which fluctuated between the “medium” and “high” categories. This behavior is explainable by positioning the RVI value close to a transition threshold.

This double test suggests that the RVI model is robust, and the resulting classifications are stable in the face of methodological uncertainties, providing a solid foundation for differentiated operational decisions. Unlike other indicators used in the literature, which are extremely sensitive to temporal resolution or seasonality [42], the RVI maintains the coherence of the classifications even in perturbed scenarios.

For this reason, the model allows for not only a more nuanced understanding of the risks associated with the integration of renewables but also an informed decision on the need for monitoring. Instead of a uniform surveillance architecture, a scalable strategy is proposed, which is capable of realistically responding to the operational challenges specific to each national energy profile.

In operational terms, the differentiated monitoring regimes derived from the RVI and composite score not only improve diagnostic accuracy but also significantly reduce redundant data collection. This translates into lower energy and computational requirements for PQ surveillance, which directly contributes to the sustainability of smart grid infrastructures. Thus, the empirical results confirm that efficiency gains in monitoring are inseparable from environmental and resource benefits, embedding sustainability within the core of technical optimization.

5. Discussion

The implementation of an adaptive monitoring model based on zonal renewable variability has significant implications for the design and operation of smart grids. In an energy context marked by decentralization, accelerated electrification, and an increase in the share of RESs, energy quality can no longer be treated in a unitary manner. The findings of this study support the idea that smart grids must be configured differently depending on local exposure to power quality instability (e.g., voltage or frequency deviations), not only on load or consumption criteria.

Adopting a differentiated monitoring strategy—whether continuous, adaptive, or conditional—enables the efficient allocation of power quality equipment, prioritization of interventions, and optimization of processed data volumes. As a result, a distributed and scalable architecture is emerging in place of a centralized and redundant model, where monitoring functions are adjusted according to the level of local power quality risk induced by renewable variability.

The results obtained based on the comparative analysis between the six selected European energy systems (Germany, Spain, Denmark, Sweden, Poland, and Romania) support the need for a conceptual reconfiguration of the way in which power quality monitoring is understood and implemented. The differences identified regarding the Renewable Variability Index (RVI), which is correlated with the national structure of balancing capacities, indicate that a unified solution at the continental level risks being inadequate. In this regard, the proposed model suggests a differentiated surveillance strategy that is calibrated to the real operational exposure of each network.

The quantitative analysis showed that, for the group of countries classified in the high-risk category (Germany, Poland, and Spain), the integration of the RVI resulted in an increase of approximately 15% in the accuracy in differentiating between stable and unstable PQ scenarios (e.g., voltage sags, frequency shifts) compared with models based exclusively on the renewable penetration rate. At the same time, the composite score values for this group were, on average, 0.10 points higher than those obtained using conventional indicators, which validates the recommendation of a continuous monitoring regime. The results confirm that the RVI model is not only descriptive but also provides a more statistically robust delineation of areas with critical exposure, increasing the operational capacity of the networks to respond contextually.

Thus, the introduction of the RVI not only clarifies the degree of exposure to risk but also provides a practical decision-making tool for planning investments and gradually scaling the monitoring infrastructure. Significantly, the correlation of the RVI with the monitoring regime paves the way for a contextual automation of network functions: SCADA systems can adjust the sampling frequency, analysis modules can trigger diagnostic procedures only in the presence of significant deviations, and the topological reconfiguration of flows can be adjusted predictively. In essence, this model proposes a grid that is not just intelligent but also aware of its instability exposure, enabling real-time functional self-regulation.

Given this framework, the concept of grid resilience takes on a new dimension: it no longer refers exclusively to the capacity to recover from disturbances but to the ability to anticipate and adapt to energy configurations with high renewable-induced variability. The proposed model directly contributes to this reconceptualization, providing an operational risk map and a decision-making framework for prioritizing investments in smart technologies.

Beyond its descriptive value, the composite score is distinguished by its immediate applicability in planning and optimization processes. It allows one not only to classify areas according to vulnerability but also to adjust the granularity of measurement tools according to real risks. In regions where the composite score reaches high values, the implementation of a continuous surveillance system, based on high-resolution sensors, becomes justifiable both technically and economically. Instead, in areas with demonstrated functional stability, conditional solutions can be adopted, with selective activation.

Compared with current international models, which predominantly focus on centralized power quality control and aggregated standards, the proposal of this study introduces a much more granular vision that is capable of distinguishing between network segments with divergent energy behaviors. Recent studies in the literature dedicated to smart grids, such as those carried out by Tabassum et al. (2024) [42] or Molokomme et al. (2025) [41], have highlighted the need for a reconceptualization of PQ supervision in relation to SRE dynamics, but without providing a practical decision-making framework. From this point of view, the integration of the composite score into SCADA algorithms or edge computing modules opens up concrete perspectives for application.

Beyond that, the implications for decarbonization and digitalization strategies are direct. By intelligently allocating monitoring resources, the carbon footprint associated with the auxiliary energy infrastructure can be reduced, avoiding overinvestment in stable areas and concentrating precision measures in critical nodes. In this way, differentiated monitoring becomes a vector of efficiency in the energy transition, supporting the convergence between climate neutrality and operational anticipation and adaptation to instability.

From a sustainability perspective, the proposed RVI-based monitoring model extends beyond technical optimization by contributing to the long-term reduction in the energy and resource footprint of smart grid infrastructures. By minimizing redundant measurements, reducing data processing requirements, and aligning monitoring intensity with actual renewable variability, the framework integrates sustainability into the operational core of smart grids. Thus, differentiated monitoring becomes not only a tool for improving reliability but also a concrete pathway toward sustainable energy transition.

Finally, the model’s alignment with smart grid strategies is fundamental. It can be embedded into prioritization algorithms operating within decentralized architectures—whether at distribution substations or prosumer nodes. Provided that SCADA systems are supplemented with local processing (edge AI), monitoring decisions can be made contextually, without centralized intervention. In turn, the network gains the ability to self-regulate its monitoring level based on current and forecasted energy behavior.

Despite its statistical validity, the proposed model faces methodological and operational constraints. The first limitation stems from the nature of the temporal resolution of the data. The monthly series used to calculate the RVI provides a consolidated picture of the variability but may blur short-term PQ instability events (e.g., flickers or voltage sags), which can significantly influence the operational power quality. In comparison, models such as the one proposed by Ward et al. (2023) [43], which use hourly data and real-time monitoring, allow for a finer detection of imbalances but imply significantly higher technological requirements.

A second source of uncertainty concerns the estimation of the actual balancing capacity. In many European countries, data on energy storage, demand response, or interconnection flexibility are fragmented or uneven in format. This heterogeneity affects the accuracy of the composite score and can lead to distorted classifications if contextual corrections are not applied. In the international literature, solutions such as those developed by [2,17,44] proposed dynamic weighting of balancing capacity through machine learning, but their applicability remains limited in the absence of publicly accessible SCADA metadata.

The third limitation of this study is the absence of direct empirical correlations between RVI values and actual power quality (PQ) metrics, such as voltage sags, harmonic distortions, or frequency deviations. This is primarily due to the lack of harmonized, cross-country PQ datasets. While such data may be collected by national regulators or transmission system operators, they are rarely accessible in a format suitable for comparative statistical analysis.

We explicitly acknowledge that both the RVI and the composite score remain proxy indicators and cannot substitute direct PQ measurements. Their intended role is to act as early-warning signals, highlighting areas with higher renewable-induced risk and guiding the prioritization of monitoring resources. Consequently, the proposed RVI-based classification should not be interpreted as a replacement for detailed PQ monitoring but rather as a screening mechanism to prioritize zones for enhanced surveillance.

Although the RVI and composite score are proxy indicators, their design effectively captures operational stress resulting from imbalances between renewable energy generation and local demand—recognized as a primary cause of power quality disturbances, such as voltage sags and harmonic distortions. Hossain et al. (2018) [24] highlight that voltage sags, harmonics, and voltage fluctuations are among the most common power quality issues in systems with high penetration of distributed renewable sources. Accordingly, even in the absence of harmonized power quality datasets, the RVI can serve as a statistically grounded and risk-sensitive early-warning tool for identifying zones of potential instability.

The next step will be validating these classifications against synchronized PQ datasets, such as voltage sags, flickers, harmonic distortions, or frequency deviations. Such validation would demonstrate whether the proxy indicators correlate with actual disturbances and would consolidate the framework’s operational relevance beyond theoretical modeling. Comparative analysis using harmonized datasets would, therefore, strengthen the model’s reliability and provide operational benchmarks for adaptive monitoring regimes.

Another significant limitation is the geographical and climatic inhomogeneity of the analyzed regions. Although the model indirectly integrates these variations through monthly production and consumption data, the lack of a direct climate correction may affect the robustness of the conclusions. Models such as those used by ENTSO-E in the seasonal adequacy forecasts (2021) include climate corrections derived from multi-annual simulations, but such approaches require access to advanced numerical models and inter-institutional collaboration [36].

It is also worth noting that the composite score does not include, in its current form, the economic cost component of differentiated measurement. In works such as those by Boateng et al. (2025) [44], the economic dimension is integrated through cost–benefit indicators for each level of monitoring granularity. Future extensions of the model could include an optimization function in which the level of supervision is adjusted not only according to technical risk (variability-induced instability) but also according to the marginal cost of each type of intervention.

Accordingly, while the model has significant predictive and classificatory value, its broad applicability requires careful calibration, access to detailed data, and an adaptable digital infrastructure. These conditions can be met gradually, in parallel with ongoing processes of digital transformation and decarbonization, making the model an evolutionary and scalable platform rather than a static, one-time solution.

6. Conclusions

The present study demonstrated the relevance of a zonal classification of the power quality risk associated with renewable generation variability. Based on empirically validated premises across the six analyzed countries, this study proposes a differentiated monitoring model that correlates the degree of renewable-induced variability with the appropriate type of monitoring infrastructure. The added value of this approach lies in the operationalization of what is often treated as an abstract issue—namely, the power quality disturbances resulting from variable renewable energy sources—into a concrete tool that can guide decision-making in the configuration of smart grids.

By using a composite score that integrates renewable generation variability with a system’s balancing capacity, this study proposes a scalable monitoring solution applicable both nationally and across countries. In addition, the approach allows for the alignment of monitoring strategies with decarbonization policies without compromising operational resilience. The model thus opens a new perspective in the literature—one in which power quality monitoring is no longer treated uniformly but is adapted to local risk profiles determined by renewable-induced variability.

Simultaneously, the comparative analysis among Germany, Poland, Romania, Denmark, Spain, and Sweden confirms that the integration of renewable sources does not automatically imply power quality instability (e.g., voltage deviations) but that its effects depend on the systemic, infrastructural, and climatic context. In particular, the case of Romania highlights the need for advanced surveillance, while Sweden demonstrates the viability of conditional, context-activated monitoring.

In the future, the proposed model can be extended in several directions. On the one hand, the inclusion of hourly resolution data series and the integration of climatic factors would allow for a finer calibration of the renewable-induced variability risk index. On the other hand, the development of an economic optimization module could transform the current score into a complete decision-making tool, capable of substantiating differentiated investment plans. Integration with edge computing infrastructure or with decentralized SCADA systems could transform the zonal classification into an active real-time regulation function, thus strengthening the capacity of networks to respond intelligently to renewable instability.

At the same time, the opportunity to integrate an explicit economic dimension into the logic of monitoring resource allocation is outlined. Thus, the risk score could be complemented with a cost function, allowing system operators to optimize interventions not only according to the severity of the exposure to generation–demand imbalances caused by renewables but also to the ratio between technical efficiency and budgetary impact.

In this scenario, the possibility of geographical expansion of the model is emerging beyond the six countries analyzed. Regions outside the European Union, which face similar challenges regarding the integration of renewable sources into a fragile or incompletely digitalized infrastructure, may benefit from adapting the model to their realities. In this sense, the proposal becomes not only an analysis tool but also a potential energy policy framework in a global context.

Future developments of this model should target real-time integration with smart grid infrastructures, such as PMUs within wide-area monitoring systems, machine-learning-based pattern recognition for RVI trends, and IoT-enabled distributed monitoring. Classifying zones by the RVI and composite scores offers a data-driven foundation for the deployment of advanced sensing systems in smart grids.

As a further research direction, special attention should be given to the empirical validation of the zonal classification by correlating the RVI and composite scores with actual power quality disturbances—such as voltage sags, frequency deviations, and harmonic distortion—recorded in operational grids. Verifying these correlations using historical data and real-time events would strengthen the model’s diagnostic value and facilitate its transformation into a predictive tool within cyber–physical energy systems.

From a synthetic perspective, the proposed model demonstrates quantifiable advantages over conventional monitoring paradigms. Comparative analyses revealed an average reduction of approximately 18% in risk area classification errors, a 15% increase in accuracy in differentiating power quality stability/instability scenarios, and a 0.10-point improvement in the composite score for high-exposure countries. In operational terms, these results translate into a significant reduction in the volume of data required, along with an increase in the contextual detection capacity of PQ-relevant disruptive events (e.g., harmonic surges and frequency shifts), thus confirming the applicative value of the RVI in the design of smart grid infrastructures.

Equally importantly, the differentiated monitoring framework demonstrates a direct contribution to sustainability objectives. By reducing redundant data collection, optimizing the allocation of monitoring resources, and lowering the auxiliary energy requirements of PQ infrastructures, the model minimizes the environmental and resource footprint of smart grids. In this way, sustainability is embedded as an intrinsic outcome of technical optimization, reinforcing the dual mission of enhancing grid resilience while advancing climate and resource efficiency goals.

A key avenue for strengthening the model lies in its empirical validation against concrete PQ data, including voltage sags, harmonic distortions, and frequency deviations. Establishing such correlations would elevate the RVI and composite score from exploratory proxies to fully operational decision-support tools in the design of adaptive monitoring regimes. While the RVI and composite score remain proxy indicators, their systematic application to adaptive monitoring introduces a novel pathway that complements existing zone-based approaches (e.g., wide-area monitoring systems and adaptive PQ). Future validation against synchronized PQ datasets will further consolidate their operational value.

In conclusion, the proposed differentiated monitoring model represents a flexible and extensible conceptual platform. As power systems become more decentralized, digitalized, and constrained by decarbonization targets, this approach provides not only a diagnostic perspective on current renewable-induced power quality vulnerabilities but also a practical architecture for future-ready, intelligent grid operation.