Electricity

Reliable discrimination between power swings and short-circuit faults is essential for secure transmission line protection, since misclassification may lead to unnecessary tripping or delayed fault clearing. Conventional power swing blocking techniques, particularly those based on impedance trajectory analysis, often require system-dependent settings and may exhibit reduced reliability under dynamic operating conditions with increasing renewable generation penetration. This paper proposes a setting-independent method for power swing and fault discrimination based on the Second Central Moment (SCM) of normalized instantaneous voltage and current signals. In this context, setting-independent means that the method does not require line-specific protection settings or case-by-case threshold tuning, although nominal voltage, nominal current, system frequency, and sampling frequency are required for signal normalization and sliding-window implementation. The SCM provides a statistical measure of signal dispersion that enables classification into three operating states: steady-state operation, power swing conditions, and fault events. Common SCM decision boundaries are applied without adjustment across the evaluated transmission lines, operating conditions, fault characteristics, power-swing frequencies, and levels of inverter-based resource penetration. The proposed method is validated through time-domain simulations using the Kundur two-area benchmark system and the IEEE 14-bus network under a wide range of disturbance scenarios, including oscillatory conditions, symmetrical and asymmetrical faults, renewable integration, and swing–fault sequences. For benchmarking purposes, the SCM-based algorithm is compared with a commercial Swing Center Voltage (SCV)-based blocking scheme widely implemented in digital relays. The results show that the proposed method achieves reliable swing–fault discrimination with low computational complexity while providing earlier blocking activation under slow oscillatory conditions and inherent fault discrimination capability. These characteristics support its practical application in real-time transmission line protection. Full article

Short-term electricity demand forecasting is essential for power grid stability and generation scheduling. Although forecasting models are commonly evaluated using one-step-ahead predictions, high accuracy at a short horizon does not necessarily imply robust performance when forecasts are recursively extended over longer horizons. This study systematically investigates this issue by evaluating ten deep-learning models: recurrent architectures (LSTM, GRU, and bidirectional LSTM), four ResNet variants with different optimizer–activation configurations, two ResNet-based hybrid models, and the PatchTST Transformer. To ensure a fair assessment, all models are trained and evaluated using identical datasets, experimental settings, and training budgets. Their performance is assessed for one-step-ahead forecasting and recursive multi-horizon forecasting at 24, 48, and 168 h, with each experiment repeated over six independent random seeds. The results reveal a substantial reversal in model ranking as the forecasting horizon increases. For the next single step, the recurrent models are the most accurate, with the lowest mean average percentage error (MAPE), while the ResNet models trained with the SGD optimizer are the weakest. As the horizon extends to 168 h, the ranking reverses, and the two SGD-trained ResNets with ReLU and tanh activation functions rise to first and second rank, respectively, while the three most accurate one-step models fall to 6th, 4th and 9th rank with respect to MAPE. The more complex PatchTST transformer never leads. Therefore, demand forecasting models should be evaluated at the operational forecasting horizon for which they are intended to be deployed rather than selected solely based on one-step accuracy. Full article

Synthetic load profiles are essential for distribution network planning, protection sizing, and demand-side management studies. However, most generation methods are validated on a single load typology, and their transferability remains unexamined. Two broad paradigms dominate the generation literature: data-driven approaches that learn the demand distribution directly from historical records (generative adversarial networks, diffusion models, Markov-chain generators) and bottom-up, physically motivated approaches that reconstruct demand from the superposition of discrete appliance ON/OFF events. Same-data, same-metric comparisons across these two families for structurally distinct load typologies remain absent from the literature, and this gap is the one this paper addresses. This paper presents a systematic comparison of three stochastic models (a first-order autoregressive (AR(1)) profile, a physically constrained ON/OFF event model, and a nonlinear-least-squares (NLS) calibrated variant) applied to two fundamentally different load typologies measured with a Class A power-quality recorder at 10-min resolution: a single-family residential dwelling (13 days, 1860 samples) and an institutional commercial building (9 days, 1333 samples). Evaluation spans six distributional statistics (mean, standard deviation, and the percentiles $p_{50}$, $p_{90}$, $p_{95}$ and $p_{99}$), the Kolmogorov–Smirnov (KS) statistic, root mean square error (RMSE), and hourly variance profiles. In this two-site study, load typology, not model sophistication, emerges as the dominant factor shaping fit quality. The simple AR(1) reproduces all percentiles within 7% for the near-Gaussian commercial load (skewness 3.47). By contrast, no model reproduces the centre, the dispersion and the upper tail of the highly skewed residential load (skewness 6.56) simultaneously to within 10%. On the residential tail the AR(1) ensemble is the least biased ($p_{99}$: −5.8%) but by far the most dispersed across realizations (CV = 15.3%), whereas the event-based models are more stable but biased, so model choice on skewed loads is a bias–stability trade-off rather than an accuracy ranking. The NLS-calibrated model attains near-exact residential $p_{95}$ reproduction (−0.3%) and commercial upper-tail errors below 2%, at a computational cost roughly three orders of magnitude above the AR(1). The dynamic characterization of demand obtained from these models, including the magnitude and frequency of the detected events, provides elements that may be of interest for the sizing and operation of photovoltaic systems in the context considered. Based on these two cases, a preliminary recommendation matrix mapping models to engineering applications is proposed, motivating typology-specific model selection rather than universal approaches; broader validation on additional sites is identified as future work. Full article

Photovoltaic fault detection and maximum power point tracking are commonly treated as separate functions, although both depend on irradiance, temperature, electrical state and data quality. This study develops a physics-guided framework that links measured fault detection to diagnosis-dependent supervisory MPPT while keeping measured diagnostic evidence separate from control-software evidence. The Lahore dataset contains 217,196 timestamped records and 19,452 unique physical fault-event groups from one grid-tied inverter with two monitored MPPT channels. Fault events are separated chronologically at the physical-event level, and calibration and threshold selection use Training + Validation data only. The HGB detector using the complete operational representation achieved 99.46% Test accuracy and macro-F1, with MCC 0.9892 and AUROC 0.9998. Permutation importance showed strong dependence on sequence-availability descriptors; a residual-free and quality-free design using only logged electrical, environmental and grid channels achieved 88.79% Test accuracy. In a separate static P–V software benchmark, the diagnosis-dependent supervisory P&O controller achieved 99.96% mean tracking efficiency and 0.193 V mean simulated voltage-reference oscillation on a restricted 240-event subset. The control experiment contains no converter dynamics or physical time base and the measured diagnosis evidence is limited to one PV installation. External plant validation and converter-level testing are therefore required before broader operational transfer. Full article

The frequent occurrence of extreme high-temperature events has significantly affected the operating characteristics and distributed photovoltaic (PV) hosting capacity of distribution networks. However, existing hosting capacity assessment methods rarely consider the accumulated heat effect caused by sustained high temperatures. To address this issue, this paper proposes a coordinated planning method for enhancing distributed PV hosting capacity under extreme high-temperature scenarios. First, an accumulated heat load model is developed to characterize the temporal cumulative influence of sustained high temperatures on temperature-sensitive loads. Meanwhile, the uncertainties associated with PV output fluctuations and load demand variations are considered to represent the stochastic characteristics of source-side generation and load-side consumption. Subsequently, a multi-objective source–network–load coordinated planning model is established to maximize distributed PV hosting capacity while minimizing the hosting capacity enhancement cost. A multi-objective artificial bee colony (MO-ABC) algorithm incorporating Sobol sequence-based quasi-Monte Carlo sampling (Sobol-MC) and a constraint domination-based constraint handling strategy are further developed to solve the proposed model efficiently. Simulation results on the modified IEEE 33-bus distribution system show that the proposed method increases distributed PV hosting capacity by 69.52% under extreme high-temperature scenarios through coordinated optimization of PV inverter reactive power control, VAR compensation, and Incentive-based Demand Response (IDR). Full article

In high-speed drive applications, including drone propulsion systems and high-speed spindle drives, switching frequency is often limited by thermal constraints or cost considerations when the adoption of wide-bandgap power devices is not economically justified. Under these conditions, natural-sampled PWM offers significant advantages over regular-sampled PWM, particularly at low switching-to-fundamental frequency ratios, by improving output waveform quality and reducing control-loop delay. This paper proposes an adaptive modulation strategy for two-level three-phase voltage-source inverters, enabling a seamless transition from space-vector PWM (SVPWM) to generalized discontinuous PWM (GDPWM). The proposed approach preserves the number of switching events by synchronizing the discontinuities of the modulation signals with the corresponding carrier peaks, thereby ensuring a consistent switching pattern while exploiting the benefits of discontinuous modulation. Full article

Current electrical systems are undergoing a profound transformation due to the massive integration of renewable energy sources and power electronics, exceeding the capabilities of traditional simulation tools. Digital Twins (DTs) have emerged as a strategic solution by enabling a bidirectional and real-time connection between physical assets and their virtual replicas. This paper presents a conceptual design proposal for a DT laboratory for electrical networks, developed from the analysis of twelve representative case studies of internationally recognized DT laboratories. The analyzed laboratories were evaluated using a technological integration scale to identify common technological trends and experimental capabilities. The results indicate that most of the selected laboratories operate at Hardware-in-the-Loop (HIL) and Power Hardware-in-the-Loop (PHIL) maturity levels, with a predominant focus on smart grids and distribution systems. Based on these findings, a conceptual laboratory architecture organized into four functional stages is proposed as a scalable cyber–physical environment for technology validation and advanced training in the electrical sector. Full article

This research presents a comparative performance evaluation of two advanced maximum power point tracking (MPPT) methodologies, namely sliding mode control (SMC) and the Kalman filter (KF), specifically applied to a standalone photovoltaic water pumping system (PVWPS). To achieve economic viability, the system is designed for storage-less operation, driving a three-phase induction motor (IM) via a high-dynamic direct torque control (DTC) scheme and a three-level inverter. The core technical contribution addresses the critical challenge of maximizing energy yield under partial shading conditions (PSCs). PSCs result in a complex, non-convex power–voltage (P−V) characteristic, containing multiple peaks, where conventional MPPT algorithms fail to consistently locate the global maximum power point (GMPP). To overcome this deficiency, we implemented the SMC-based MPPT algorithm to exploit its inherent robustness and rapid dynamic response, and compared it with the Kalman filter MPPT, which relies on stochastic state estimation to achieve accurate tracking and effective disturbance rejection. MATLAB/Simulink analysis compares the proposed techniques with the perturb and observe (P&O) MPPT method. The comparison considers tracking efficiency, convergence speed, and steady-state ripple under various shading conditions to identify the most effective control strategy for improving PVWPS performances. The reported performance evaluations are based on numerical simulations conducted within the MATLAB/Simulink environment, using a validated system model. Full article

This paper presents an approach for optimal operation of self-healing networked microgrids (NMGs) under both normal operation and emergency conditions using the pufferfish optimization algorithm (POA). The proposed methodology is based on an energy management system (EMS) with two levels and independent functions. The lower-level is designed for normal operation, where the local controller of each microgrid (MG) performs the optimal dispatch of power from the dispatchable sources. During an emergency case in any MG, the higher-level EMS is activated, and the global controller is brought into operation. Physically, the NMGs are connected by tie-lines, while cyber links are established to exchange information and control signals for coordinated operation. Each MG operates to supply its local demand during normal operation conditions, resulting in no electrical power exchange between MGs. İn case of generation deficiency or a fault leading to generation outage, electrical power can be exchanged through the existing interconnections, enabling the affected microgrid to receive support from neighboring MGs. The main objective of POA is to minimize the total operating cost, in which the economic impact of network power losses is incorporated into the single objective function. Simulation studies were conducted using MATLAB and DIgSILENT software over one day. The performance of POA was compared with Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and Grey Wolf Optimizer (GWO) under the same computational settings. Statistical and convergence analyses show that POA achieves the lowest mean operating cost across all studied cases, with low run-to-run variability and favorable convergence behavior. The results demonstrate the effectiveness of the proposed approach in improving the economic operation of NMGs under both normal and emergency conditions. Full article

Lithium-ion battery (LIB)-powered locomotives have emerged as a promising alternative to conventional rail electrification by reducing dependence on overhead catenary systems, lowering infrastructure costs, and improving operational flexibility. However, train failures involving power loss or mechanical faults present significant challenges for battery-powered rail systems, particularly on single-track corridors where alternate route options are limited. Consequently, the implementation of a reliable limp-home strategy is more complex than in road electric vehicles (EVs), requiring consideration of battery availability, rescue logistics, and track accessibility. This study investigates limp-home operation and rescue planning for a battery-powered historic trolley operating on a 20 km heritage route in North Carolina. The trolley is powered by a dedicated LIB trailer and supported by battery-charging (BC) infrastructure based on inductive power transfer (IPT) technology. A comprehensive framework is developed that integrates time–space analysis, cellular automata (CA)-based failure-risk modeling, and battery-energy assessment to evaluate train-failure scenarios and recovery strategies. The results identify critical failure regions along the route and demonstrate the benefits of strategically deploying additional LIB rescue trailers and wireless power transfer (WPT) infrastructure. A revised rescue strategy incorporating two additional LIB trailers, together with static and dynamic WPT systems, substantially reduces recovery time and improves the likelihood of maintaining scheduled excursions. Emergency energy analysis further shows that WPT-assisted operation can significantly extend limp-home capability under low state-of-charge conditions. Based on the integrated analysis, a tiered limp-home decision framework is developed to support operator decision-making during train failures. The proposed methodology provides a practical approach for enhancing the resilience, recoverability, and operational reliability of battery-powered heritage rail systems and can serve as a foundation for future battery-electric rail applications. Full article

This paper presents a novel master–slave stochastic optimization framework for the optimal siting and sizing of fixed-step capacitor banks in medium-voltage distribution networks, explicitly addressing the inherent variability of load demand that is typically neglected in conventional deterministic approaches. The proposed methodology integrates a scenario-based stochastic optimization model with a Chu and Beasley genetic algorithm (CBGA) as the master stage, which handles discrete placement decisions, and a successive-approximation power flow method (SAPF) as the slave stage, which evaluates the technical and economic performance of each candidate solution under multiple load scenarios. To capture demand uncertainties, 365 daily load realizations are generated using independent Gaussian noise with a relative standard deviation of 10% applied to each load point. These are subsequently reduced to ten representative scenarios via k-means clustering, reducing the number of power-flow evaluations per candidate solution from 365 to 10 (a 36.5-fold reduction); the reduced scenarios exhibit a low mean absolute error (MAE: <2%) with respect to the original mean, indicating faithful representation of the average load behavior, while the silhouette score is modest (approximately 0.25), consistent with the unimodal nature of the generated data and implying that the clusters are not well separated. Extensive simulations on a 33-bus test feeder considering three energy-cost-escalation scenarios (0%, 10%, and 20%) over a 20-year planning horizon demonstrate that both the deterministic and stochastic approaches reduce the total net present cost by 16.52% to 17.34% compared to the uncompensated network; the stochastic approach consistently delivers solutions that are either superior or comparable to deterministic planning (yielding up to approximately 0.16% additional cost reduction) while offering enhanced robustness against load variability. The stochastic framework offers distinct advantages, including robust solutions across a wide range of operating conditions, an inherent ability to adjust investment levels in response to probabilistic load distributions, and the ability to quantify uncertainty in decision making, with the most significant benefits observed when energy costs are low and load variability is high. The convergence of both approaches at a 20% escalation level further validates the reliability of high-resolution deterministic modeling when economic factors strongly dominate the optimization objective. This study underscores the importance of probabilistic modeling for modern distribution network planning, providing a practical and computationally efficient decision-support tool for utility planners to enhance grid resilience and operational efficiency in the context of increasing demand variability and renewable energy integration. Full article

The availability of reliable digital models of low-voltage (LV) networks is a prerequisite for operation, planning, hosting-capacity, and asset-management studies. In practice, however, utility geodatabases often contain geometric discontinuities, implicit relationships among assets, and incomplete connectivity, which hinders their direct use as topological models for electrical studies. This work proposes a reproducible framework for topological-quality and fitness-for-use screening, applied to a real LV network derived from a utility geographic information system (GIS). The methodology converts geospatial layers into a graph representation, audits load-to-pole and endpoint-to-known-node distances, evaluates structural sensitivity to the endpoint connection tolerance, computes a load-level screening confidence score, characterizes transformer-level coverage under an explicit demand scenario, and evaluates structural sensitivity under synthetic perturbations of the line geometry. The case study comprises 38 urban blocks, 155 poles, 13 distribution transformers, 429 loads, 777 overhead LV segments, and 23 underground LV segments. The mean load-to-pole distance is 10.91 m and the mean endpoint-to-known-node distance is 4.39 m. The fixed 30 m load-to-pole assignment rule retains 416 loads (96.97%) throughout the tolerance sweep, while the number of connected components decreases from 245 to 82 and the number of virtual nodes from 817 to 439 as the endpoint tolerance increases from 0.5 to 5.0 m. The confidence score classifies 32.40% of loads as High, 59.44% as Medium, 5.13% as Low, and 3.03% as Review/Unassigned. Under the stated residential screening assumptions, four transformers exceed the adopted 80% loading threshold. Synthetic perturbations leave the load-to-pole control metrics unchanged by construction, because only line-segment geometry is perturbed, but the number of connected components varies from 78 to 236, with a maximum relative change of 187.80% with respect to the 5 m base graph. The contribution is therefore a transparent GIS quality-screening procedure, not an exact reconstruction or electrical validation of the physical LV topology. Full article

The increasing penetration of variable renewable energy sources into electric power systems requires advanced optimization tools to address the complexity of hybrid hydrothermal scheduling while minimizing generation costs and carbon emissions. This study investigates the application of four deep learning architectures—Kolmogorov–Arnold networks (KANs), long short-term memory (LSTM), gated recurrent unit (GRU), and deep feedforward (DFF)—to solve the hydrothermal scheduling problem in hybrid power systems that incorporate wind power generation and pumped-storage hydropower (PSH) plants. The methods were evaluated on a 10-generator test system over a 24-h planning horizon in three objective-weighting scenarios, considering economic dispatch only, pure emission minimization only, and balanced objectives. All architectures successfully solved the integrated problem and satisfied the system constraints. This study reports the first application of the KAN to the hydrothermal scheduling problem, demonstrating its viability and interpretability potential for future applications in electric power systems. Full article

Power quality degradation in low-voltage (LV) distribution networks remains insufficiently documented in many rapidly urbanizing African cities despite its significant impact on electrical equipment, service reliability, and network operation. This study investigates the following research question: To what extent does the power quality of an urban low-voltage distribution network comply with international standards, and which network characteristics are most strongly associated with the observed disturbances? To address this question, an extensive field measurement campaign was conducted from October 2024 to February 2025 on five radial feeders supplied by the Babemba medium-voltage/low-voltage (MV/LV) substation in Lubumbashi, Democratic Republic of the Congo. Electrical parameters were monitored using a Class B Chauvin Arnoux Qualistar C.A. 8331 power quality analyzer and evaluated against internationally recognized power quality standards. The measurements revealed persistent power quality degradation characterized by chronic under-voltage, with prolonged voltage levels below 207 V, typical deviations ranging from −20% to −30%, and voltage dips reaching 70–80% of the nominal voltage during peak loading conditions. Power supply continuity was also severely affected, with a System Average Interruption Frequency Index (SAIFI) of 7.85 interruptions/year and a System Average Interruption Duration Index (SAIDI) of 491 min/year, while a medium voltage outage lasting approximately 48 h highlighted the limited resilience of the distribution system. Additional disturbances included phase voltage imbalance reaching 18%, neutral currents up to 327 A, and short-term flicker values (Pst) approaching 1.5, exceeding the recommended comfort threshold. Overall, the observed disturbances were associated with heterogeneous feeder loading conditions, network configuration, non-standard electrical connections, and documented physical deterioration of the infrastructure. This study provides a comprehensive field-based assessment of power quality and service continuity in an urban LV distribution network in the Democratic Republic of the Congo and offers a quantitative basis for prioritizing feeder reinforcement, phase balancing, infrastructure rehabilitation, and the establishment of continuous local power quality monitoring. Full article

Distribution networks face low supply reliability and limited fault-recovery capability. Flexible interconnection devices enable looped operation of AC feeders and facilitate optimal power-flow control, which helps address these operational challenges. This paper proposes a novel medium-voltage flexible interconnection device with a star-connected topology, together with a compact power-module structure and its main-circuit topology. Because distribution networks often contain high background harmonics, a control scheme is developed for the device, comprising a phase-locked loop, hierarchical capacitor-voltage control, and an inner current-control loop. A harmonic-current compensation control strategy is further designed to suppress current harmonics. A hardware-in-the-loop (HIL) testing platform is built to validate the proposed strategy. The results show that the proposed topology enables flexible interconnection of distribution networks, and the proposed control strategy operates stably under high background harmonics while effectively compensating current harmonics. Full article

Optimal distribution network reconfiguration modifies feeder topology through switching actions to improve electrical performance and service continuity. However, integrated and reproducible approaches that explicitly combine Pareto-based loss–reliability trade-off analysis, AC electrical evaluation, and clearly defined feasibility criteria remain comparatively limited, particularly in studies extending beyond standard benchmark systems. This paper presents a Pareto-based computational framework for radial distribution network reconfiguration that combines an NSGA-II-type evolutionary search with AC power-flow evaluation in OpenDSS^®. The search retains non-dominated sorting, crowding-distance-based diversity preservation, and elitist environmental selection, while offspring are generated through a graph-aware close–open branch-exchange mutation adapted to radial DNR. Accordingly, the evolutionary component is used as a problem-oriented search strategy rather than as a new canonical NSGA-II variant. Active power losses and SAIDI are minimized simultaneously, while SAIFI is reported as a derived reliability indicator and ENS as a complementary metric. Connectivity, radiality, and power-flow convergence define feasibility, while voltage compliance and load supply are subsequently assessed to characterize the operational status of each configuration. The framework is evaluated on a synthetic 5-node system, the IEEE 33-node benchmark, and an anonymized planning-oriented model of a real 13.2 kV Colombian distribution network. For the IEEE 33-node system, the minimum-loss configuration reduces losses by 26.03%, whereas the compromise solution achieves a 25.88% loss reduction and a 21.52% improvement in SAIDI. Across 20 paired runs, both the graph-aware strategy and a canonical NSGA-II baseline with topology repair recovered the same five-point empirical non-dominated set, including the classical minimum-loss topology without explicit seeding. For the real system, the operational compromise reduces active power losses by 38.88%, improves SAIDI by 23.11%, and increases the minimum voltage from 0.826 to 0.915 p.u. These results show that minimum-loss operation does not necessarily provide the most balanced trade-off between electrical efficiency and expected service continuity, supporting the use of explicit Pareto analysis for planning-oriented distribution network reconfiguration. Full article

The rapid deployment of distributed photovoltaic (PV) systems has increased renewable energy generation while introducing operational challenges for distribution networks, particularly surplus PV export during periods of high solar production. This study proposes a measurement-driven, rule-based smart electric vehicle (EV) charging framework to enhance the local utilization of surplus PV generation using real operational measurements acquired from a Siemens PAC3200T power quality analyzer installed at the point of common coupling (PCC) of the Far’ata–Immatain distribution feeder in Palestine. The proposed framework coordinates EV charging based on representative surplus PV operating states and PCC operating conditions while considering user charging requirements and network operational limits. Four MATLAB-based simulation scenarios representing increasing EV penetration were evaluated. The results demonstrate progressive reductions in reverse active power export together with corresponding improvements in local PV utilization as EV penetration increases. Under the highest investigated charging scenario, up to 200 kW of the investigated surplus PV generation was locally utilized, and reverse active power export was eliminated under the investigated operating conditions. Overall, the proposed framework provides a practical and scalable approach for improving renewable energy utilization in data-limited distribution networks without requiring comprehensive feeder models or immediate network reinforcement. Full article

With the increasing conflict between economic and environmental objectives in power systems, the economic emission dispatch (EED) problem has become a highly constrained, nonlinear, and strongly non-convex multi-objective optimization problem due to valve-point effects and nonlinear constraints such as network losses. To address this challenge, this paper proposes an improved multi-objective Grey Wolf Optimizer (IMOGWO). The proposed method enhances search performance through four strategies: a hybrid initialization scheme combining circle chaotic mapping and Latin hypercube sampling to improve population diversity, a dream-inspired group perturbation mechanism to strengthen global exploration, a nonlinearly decreasing convergence factor to dynamically balance exploration and exploitation, and a hybrid update strategy incorporating Lévy flight to avoid local optima. Experimental results demonstrate that IMOGWO can effectively balance the trade-off between generation cost and pollutant emissions while exhibiting competitive performance in terms of convergence behavior, solution quality, and stability. Full article

False data injection attacks (FDIAs) pose a growing threat to distributed energy resource (DER) aggregators because a compromised aggregator can expose and affect a number of enrolled DERs. However, DER aggregators only have access to limited measurements from DERs and the systems. This makes existing FDIA detection methods ineffective in this setting due to two main limitations: (1) they are grounded in full observability of a microgrid or distribution system and therefore are incompatible with the limited observability of a DER aggregator; (2) they can only either detect anomalies or explain why a deviation occurs, but not both simultaneously. To address these limitations, we propose a physics-guided fusion-based cyberattack detection method specifically designed for DER aggregators. This approach integrates two complementary modules: a forecasting-assisted residual method for rapidly anomaly detection, and a PV-aware sensitivity-based method to diagnose and explain their underlying physical causes. A gradient boosting machine (GBM) is then leveraged to fuse these outputs, optimizing the precision–recall tradeoff. The proposed method is tested on one microgrid test system with different bus observability levels across static and gradual attack scenarios with multiple levels of attack sophistication. Across the scenarios, the proposed method achieves a 0.91–0.93 precision–recall area under the curve score (PR-AUC), demonstrating the method’s effectiveness in securing DER aggregators with partial system visibility. Full article