The integration of high-penetration distributed photovoltaic (PV) systems has introduced challenges to distribution grids, including reverse power flow, distribution transformer overload, and voltage uplift. To address the challenges posed by high-penetration distributed PV grid connection and large-scale electric vehicle (EV) integration, this paper proposes a vehicle-grid cooperative interaction strategy based on a cooperative game model. A cooperative alliance between vehicles and the grid is established, with a collaborative interaction game model designed to maximize alliance profits. The profit distribution weights for both parties are determined based on their respective contributions to cooperative gains. Shapley values are applied to allocate interaction profits, yielding time-of-use electricity prices at different moments to guide EVs charging and discharging, thereby achieving coordinated operation between EVs and the grid. Simulation results demonstrate that after implementing the optimized cooperative interaction strategy: surplus PV power generation decreased by 939.10 kW, node voltage fluctuation rate reduced by 5.90%, mitigated the impact of PV fluctuations on the grid, enhanced PV power absorption capacity, and alliance cooperative profits increased by 606.91 yuan compared to non-cooperative grid scenarios.
Customer churn remains a critical concern in the mobile telecommunications industry, where retaining subscribers is essential for long-term profitability. This study presents an explainable machine learning framework to detect and interpret network-driven churn by integrating demographic data with key network performance indicators, including signal quality metrics e.g., reference signla received power, signal to interference plus noise ratio, download/upload throughput, and latency-related measurements derived from the radio access network. Using the CatBoost classifier on a proprietary dataset of 115 264 subscribers constructed with a balanced 50/50 class distribution to prevent majority-class bias, the model achieves an F1 score of 81.0% and an area under the receiver operating characteristic curve of 0.884, demonstrating strong discriminative performance. A central contribution is the introduction of a systematic SHapley Additive exPlanations (SHAP)-based attribution method that distinguishes network-driven churn from other churn types. For each churn prediction, positive SHAP values are extracted and categorized as network-related or non-network-related; instances where network features contribute the dominant share of the total positive SHAP attribution are labeled as network-driven churn. The framework is validated on both the proprietary Turkcell dataset and three publicly available telecom churn benchmarks, demonstrating generalizability. The proposed approach bridges predictive modeling and realworld network optimization, offering telecom operators a scalable and interpretable tool for targeted customer retention.
A compact coplanar waveguide (CPW) fed quad-notched ultra-wideband (UWB) antenna is proposed in this article. The proposed antenna consists of a set of regular geometric slots, including a rectangle, semicircular, and pentagonal stub, as well as a microstrip feedline with a total length of 35.4 mm and a width of 28.82 mm. To generate one notched frequency band at (3.78–4.15) GHz (9.34%), the antenna uses one inverted U-shaped slot at the pentagonal stub of the antenna. A combination of L-shaped slots are etched out in the upper surface of the antenna to obtain the dual WLAN frequency band notches at (5.15–5.31) GHz (3.05%) and (5.6-5.9) GHz with (5.21%). Rectangular Split Ring Resonator (SRR) is then embedded in the antenna’s lower surface to realize a notch band at (7.25–7.80) GHz (7.3%). Except in quad frequency stop bands at (3.78–4.15) GHz, (5.15–5.31) GHz, (5.6– 5.9) GHz, and (7.25–7.80) GHz, the proposed UWB microstrip antenna has an operational impedance bandwidth ranging from 3 to 11.2 GHz. A systematic development of the UWB antenna on FR4 substrate is reported, beginning with a pentagonal stub fed by CPW for wide impedance bandwidth, improved through asymmetrical geometry over 3.1–10.6 GHz, and gradually adding a U-shaped slot for C-band notch (3.78–4.15 GHz), dual L-shaped slots for WLAN notches (5.15–5.31 GHz and 5.6–5.9 GHz), and ground-plane SRR for SHF notch (7.1–7.8 GHz), ending with an optimized quad-notched design reducing mutual coupling for interference-resistant applications. The proposed quad-notched UWB antenna has been effectively developed, prototyped, and verified. The simulation and measurement results are thoroughly examined and evaluated. The fabricated antenna is justified for use in wireless applications such as military radar systems, medical imaging, and consumer electronics.
In the present study, a novel sliding mode controller (SMC) is proposed for robust and reliable voltage regulation in modern power systems. The proposed SMC is designed for a nonlinear model of an automatic voltage regulator (AVR) system having limiters in line with IEEE recommendations. The mathematical derivation of the control input is described in detail, and the asymptotic stability is guaranteed by Lyapunov stability theorem. The number of the controller parameters is reduced from four to two, and the control input is simplified without sacrificing the performance of the controller. This feature simplifies the control signal and makes the controller easy to optimize. It has been revealed that the number of controller parameters in the proposed study is less than the number ranging from three to seven in other AVR controller studies in the literature. The SMC algorithm differs from design-oriented studies in the literature in that it simplifies the control signal without affecting the controller’s performance. In the optimization stage of the study, the Sine Cosine Optimization Algorithm is conducted and its performance is evaluated with respect to a different number of agents and iterations. The results showed that the proposed controller parameters need very few simulations to find their optimum values. It can be easily seen that ten iterations with five agents are sufficient for the optimization of the proposed controller. It was clearly seen that the total number of simulations with 50 simulations in the proposed study was significantly less than the number of simulations in other related studies in the literature. The resulting optimum controller parameters were obtained as ksw = 744.5885 and k3 = 0.0711. A 0.5% overshoot at the output of the generator is allowed in the objective function of the optimization. Simulation results show that the proposed controller is robust and reliable against parameter uncertainties and various types of output disturbances.
This paper presents a systematic and bibliometric review investigating the role of emerging technologies, specifically the Internet of Things (IoT) and unmanned aerial vehicles (UAVs), in earthquake disaster management. Following the PRISMA 2020 framework, a structured search was conducted in the Scopus database covering publications up to 2025. The retrieved studies were screened and analyzed through quantitative bibliometric techniques to identify research trends, collaboration networks, and keyword co-occurrence patterns. Complementary qualitative content analysis was applied to classify the contributions of IoT and UAV technologies across the three main phases of disaster management: preparedness, response, and recovery. The findings indicate that IoT-based systems are predominantly utilized for real-time seismic monitoring, early warning dissemination, and data communication, while UAVs are extensively employed for rapid damage assessment, communication restoration, and logistics support in post-earthquake scenarios. The integrated use of IoT and UAVs strengthens situational awareness, enhances coordination efficiency, and promotes data-driven decision-making. The study concludes by highlighting research gaps in interoperability, data fusion, and energy efficiency, providing future directions for developing adaptive and collaborative frameworks for earthquake resilience.
The distribution network faces significant operational challenges in maintaining acceptable voltage levels under dynamic load conditions. Rapid load fluctuations often degrade the performance of localized voltage control schemes, leading to voltage instability and increased power losses. To address this issue, this paper proposes a centralized, optimization-based Volt-VAR control strategy that integrates Particle Swarm Optimization (PSO) with Conservation Voltage Reduction (CVR) for enhanced voltage regulation and energy efficiency in distribution systems. The proposed approach determines optimal tap positions of Automatic Voltage Regulators and On-Load Tap Changers using a system-wide optimization framework implemented in OpenDSS. A voltage-dependent load model is employed to realistically capture the impact of voltage variations on active and reactive power consumption. Simulation studies conducted on the IEEE-123 bus distribution system demonstrate that the PSO-based centralized control improves voltage regulation. It also reduces active power losses compared with conventional localized control. Additionally, peak load reduction is observed as a secondary benefit of CVR implementation. The proposed PSO-based control reduces active power losses from 4.72% to 4.59% and improves the minimum bus voltage from 0.8714 p.u. to 0.9459 p.u., ensuring voltage profiles remain within acceptable limits while enabling effective load reduction under CVR operation. The proposed framework provides a scalable and practical solution for advanced distribution management systems to improve voltage quality and energy efficiency.
The service transformer stands as a pervasive and essential component within the energy infrastructure. Service transformers are critical elements in power distribution, and their longevity hinges on effective monitoring. Traditional fault detection methods, especially dissolved gas analysis (DGA), though widely used, often suffer from delays due to manual sampling and lab-based analysis. This article explores machine learning techniques (MLTs) as a modern alternative to enhance the interpretation of DGA data for early-stage fault detection in service transformers. The method relies on the use of multiple DGA datasheets to investigate the fault typing capabilities and suitability of different MLTs. It evaluates multiple algorithms—logistic regression, support vector machines (SVM), random forest, and gradient boosting—using actual DGA datasets. To validate the best class algorithms, this article also looks at performance accuracy and then evaluates the top-performing algorithm. The results demonstrate that random forest and gradient boosting outperform others, achieving up to 98% accuracy, and are especially useful for condition monitoring professionals dealing with insulating oil analysis, as compared to the accuracy achieved of 81.4% with SVM and 76% with artificial neural network (ANN) in the case of previously published work.
Unlike conventional reflection-only reconfigurable intelligent surfaces (RISs), simultaneously transmitting and reflecting RISs (STAR-RISs) constitute a paradigm shift in intelligent surface design by facilitating electromagnetic wave manipulation in both transmission and reflection domains, thereby achieving full-space coverage and mitigating the inherent directional limitations of traditional RIS implementations. Orthogonal time frequency space (OTFS) modulation has emerged as a highly promising waveform candidate for enabling reliable and efficient communication in high-mobility environments, owing to its resilience against severe Doppler effects and time-frequency channel variations. In this paper, a new STAR-RIS aided OTFS system (STAR-RIS-OTFS), which is a promising technique for high-mobility future wireless communication systems, is proposed. The combination of these two methodologies yields enhanced bit error rate performance over conventional OTFS systems. The performance of the proposed STAR-RIS-OTFS scheme is evaluated under M-ary quadrature amplitude modulation over both Rayleigh and Nakagami-m fading channels. In addition, a detailed computational complexity analysis is carried out for the maximum likelihood detector and the enhanced low-complexity detector, considering both the proposed STAR-RIS-OTFS architecture and the benchmark system. Moreover, the effects of imperfect channel state information (CSI) on the proposed STAR-RIS-OTFS system are investigated, and comparative performance results are provided under perfect and imperfect CSI assumptions. The proposed STAR-RIS-OTFS system is evaluated through Monte Carlo simulations over high-mobility Rayleigh channels for various system parameters. It is shown that the proposed STAR-RIS-OTFS scheme outperforms the traditional OTFS technique.
Frequency control has emerged as a significant challenge in microgrid systems due to the increasing penetration of renewable energy sources. To address this issue, a virtual inertia (VI) based control scheme is proposed. A cascaded PID (C-PID) controller which consists of a first stage PD controller with derivative filter followed by a second stage one plus a PI controller is used in the VI control scheme to improve the frequency regulation. The recently proposed Bald Eagle Search (BES) algorithm is employed to optimize the controller parameters. Initially, the PI controller is considered, and the results of BES are compared with the Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). It is demonstrated that the objective function value with BES is decreased by 25.15% and 26.1% compared to GA and PSO, respectively. In the next stage, PID and C-PID controllers are considered, and controller parameters are tuned by the BES algorithm. The efficacy of C-PID is evaluated with PID and PI for different cases like constant and variable load, solar and wind power. It is shown that the percentage improvements in integral of squared error (ISE), integral of time-weighted absolute error (ITAE),integral of time weighted-squared error( ITSE), integral of absolute error(IAE), and mean squared error( MSE) are 97.75%, 89.35%, 98.12%, 88.55%, and 98.27%, respectively, compared to the BESoptimized PI controller, and 81.42%, 65.84%, 84.89%, 62.06%, and 85.71%, respectively, compared to the BES-optimized PID controller under simultaneous variations in solar power, wind power, and load demand. Sensitivity analysis, stability analysis, effect of nonlinearities and comparison with recent control schemes of proposed approach is also performed. Finally, real-time validation through Open Platform for Advanced Laboratory-RealTime(OPAL-RT) is done.
This paper presents a novel adaptive sliding mode control approach based on a model-free technique for nonlinear active suspension systems in cars. The proposed method comprises four key components: First, an intelligent proportional-integral control based on model-free principles is employed to introduce desired dynamics. Second, an extended state observer is utilized to estimate the system’s nonlinearities and unknown dynamics. Third, an integral sliding surface is integrated to eliminate the reaching phase, along with a noise-sensitive switching mechanism. Last, adaptive gain dynamics are incorporated to improve accuracy. This controller offers several advantages, including its simple structure and ease of adjustment. The study also includes analyses of theoretical system stability, convergence speed, and control accuracy. Finally, the proposed approach is validated using MATLAB simulations and compared with a recent paper on a two-degree-of-freedom nonlinear quarter-car active suspension system.
Accurate forecasting of electricity production has become increasingly critical for energy systems, particularly in countries where renewable energy penetration and climate variability are rapidly increasing. In this context, this study investigates the interaction between meteorological indicators and electricity production in Türkiye using machine learning–based time series forecasting methods. The analysis is conducted using daily electricity production data obtained from the Turkish Energy Market Operator (EPİAŞ) and corresponding meteorological variables—namely apparent temperature, wind speed, and humidity—sourced from the Open-Meteo platform, covering the period from late 2015 to 2024. After aggregating hourly observations into daily series and applying appropriate preprocessing steps, several forecasting models are implemented and compared, including recurrent neural networks (RNN), long short-term memory (LSTM), gated recurrent unit (GRU), and the Prophet time series model. Model performance is evaluated using mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE), and the coefficient of determination (R2). The results indicate that gated RNN architectures significantly outperform classical approaches in modeling electricity production dynamics. Among all models, the GRU model achieves the best performance, yielding the lowest error values (MAE = 1241.4, RMSE = 1951.5, and MAPE = 3.5%) and the highest explanatory power (R2 = 76.9%). The LSTM model provides comparable but slightly weaker results, while the classical RNN and Prophet models exhibit substantially lower predictive accuracy, particularly in capturing nonlinear patterns and sudden production fluctuations. Overall, the findings demonstrate that GRU-based models offer a robust and reliable framework for electricity production forecasting in meteorologically sensitive energy systems and provide valuable insights for energy planning, grid operation, and data-driven energy policy development.
This study simulates and develops a dynamic barrier defense system utilizing a drone swarm against airborne threats. The dynamic barrier system begins with a formation of maximum coverage width to account for potential route changes based on the target distance; as the threat approaches, it transitions to an optimally minimized configuration. Unlike traditional static defense lines, in this study, the barrier formation point and size are calculated based on swarm size, the positions of both threat and defense elements, and their respective velocity vectors. To validate the study through simulation, the reaction process of a swarm consisting of 100 drones in a 10 × 10 grid topology, the dynamic interception point, the maximum formation gap based on kinematic constraints, and the asynchronous arrival condition (shielding duration) arising from the swarm's transit geometry were analyzed in detail. Formation selection and variable barrier size were optimized according to these parameters. Simulation results demonstrate that the proposed dynamic model enhances operational flexibility compared to static systems.
The present system of photovoltaic (PV)-battery-supported dynamic voltage restorer (DVR) has been developed for power quality improvement in the distribution system. The proposed system has the ability to mitigate voltage power quality issues like voltage sag, voltage swell, and voltage harmonics. For the proposed system, a type-3 second-order generalized integrator (SOGI) based phase-locked loop controller is utilized to improve the performance of the battery-supported DVR system. The proposed type-3 SOGI for the PV-battery-based DVR system has the ability for immediate dynamic behavior regulation with phase lead mitigation. Along with the type-3 SOGI, a modified synchronous reference frame controller is applied for the easy computation of the system controller. The proposed controller also utilizes an adaptive filter based on a double band structure to eliminate the harmonics and disturbances in the source voltages. This adaptive filter is also applied for the regulation of the direct current (DC) link voltages supported by the PV and battery. The integration of the PV and battery into the DVR system improves energy conservation and utilization of renewable energy sources. The system has been simulated under various dynamic grid conditions, and the conditions for voltage disturbances are analyzed in the MATrix (MATLAB)/Simulink platform. After analyzing the results, it has been concluded that the proposed controller is capable enough to mitigate voltage quality issues under various environmental conditions.
Hamming Quasi-Cyclic (HQC) is a code-based key encapsulation mechanism that stands out in the National Institute of Standards and Technology Post-Quantum Cryptography standardization process. Almost all of the execution times of the HQC algorithm stem from high-order polynomial multiplication operations. These high-order polynomials constitute the execution times of the HQC algorithm. In this study, different hierarchical combinations of Toom-Cook (TC) and Karatsuba algorithm (KA) were analyzed to optimize the performance of HQC. Tests revealed that the highest throughput on the Raspberry Pi 4 Model B (ARM Cortex-A72) platform was achieved with the TC4 + TC3 + KA2 combination. By optimizing the hardware's pipeline and cache structure, improvements in processing times of 50.5%, 52.8%, and 41.6% were achieved in the HQC-128, 192, and 256 variants, respectively, compared to the standard versions; this approximately doubled the performance. Cite this article as: C. İnce, “Performance analysis and acceleration of Hamming Quasi-Cyclic algorithm in raspberry pi 4 environment,” Electrica, 26, 0091, 2026. doi: 10.5152/electrica.2026.26091.
This study presents the Turkish Offensive Language Identification Dataset (TOLID), a large-scale, high-quality dataset for the automatic detection of offensive language in Turkish social media posts and evaluates the performance of several transformer-based models. Although most previous studies have been constrained by small sample sizes, imbalanced class distributions, or narrow topical focus, TOLID includes a wide range of offensive expressions without imposing restrictions on topics, individuals, or groups. The dataset was annotated by three independent experts using a hierarchical and fine-grained scheme that addresses not only general offensive language but also specific subcategories such as sexist, racist, political, and religious insults. To evaluate the dataset, several transformer-based models adapted for Turkish, including BERTurk (Bidirectional Encoder Representations from Transformers for Turkish), ConvBERTurk (Convolutional Bidirectional Encoder Representations from Transformers for Turkish), and ELECTRA-Turkish (Efficiently Learning an Encoder that Classifies Token Replacements Accurately for Turkish), were trained and tested. Among these, ConvBERTurk achieved the highest scores, reaching 82.76% macro F1 in offensive vs. non-offensive classification and 78.14% in targeted vs. non-targeted offensive classification, outperforming previous research. These results demonstrate that combining a balanced, multi-annotated dataset with advanced deep learning architectures can effectively address the linguistic richness, contextual complexity, and informal nature of Turkish social media text. Additionally, a web-based application was developed to provide a practical interface for analyzing text and visualizing model outputs, extending the study's impact beyond academic research to real-world applications. Overall, this study addresses key limitations of prior research and makes a significant contribution to Turkish natural language processing by providing a meticulously constructed dataset and extensive benchmarks with state-of-the-art models. TOLID establishes a robust foundation for future work on offensive language subtypes, automatic moderation, and toxicity analysis in Turkish social media. Cite this article as: M. S. Kurt and E. Yücel, “TOLID: Turkish offensive language identification dataset and transformer-based benchmarks,” Electrica, 26, 0404, 2026. doi: 10.5152/electrica.2026.25404.
This study shows a new design for an approximation unsigned multiplier that keeps good computational accuracy while cutting down on size and power use significantly. The suggested design has three main parts: the lower essential portion (LEP), the middle essential portion (MEP), and the accurate region. In the LEP, a preset compensation term takes the role of partial products (PPs). This makes the hardware more efficient without losing accuracy. Two new approximate compressors are used in the approximate region to lower the PPs even more. To reduce errors made by these compressors, a special and effective error-correcting module is built in. The testing on the 8-bit multiplier shows that the suggested designs can improve power by up to 43.71% compared to the exact multiplier. The proposed architecture achieves up to 23.41% more power savings and up to 31.9% better power-delay product than other approximate architectures. Moreover, the proposed designs' practical usefulness is substantiated by assessments in image processing and neural network inference tasks. Cite this article as: D., P. Upendranath and V. Krishnanaik, “Design of power efficient approximate multiplier using enhanced adder compressors,” Electrica, 2026, 26, 0348, doi: 10.5152/electrica.2026.25348.
This study presents the design of a compact analog I-Q type vector modulator (VM) for phased array systems in the S-band. The design and implementation of the subcomponents of the produced VM, which are a stripline-based symmetric coupled directional coupler, two 0°/180° bi-phase modulators, two analog variable attenuators, and a Wilkinson combiner, are realized in detail. Bi-phase modulators, analog variable attenuators, and a Wilkinson combiner were combined on a single board using RO4003 material. Simulations were performed in the Computer Simulation Technology (CST) and AWR Applied Wave Research (AWR) Design Environments. The measurement results of the produced VM show a dynamic range of around 30 dB, a 360° phase shift range, a phase resolution of 0.5°, 12 dB insertion loss at 3 GHz, and a return loss higher than 10 dB between 2 and 4 GHz. Moreover, in this study, a compact analog I-Q type VM has been designed and fabricated with lower insertion loss than available literature, considering similar operating bands. Cite this article as: C. Çindaş and S. Şimşek, “Design of a compact analog I/Q type vector modulator in S-band for phased array antennas,” Electrica, 2026, 26, 0293, doi: 10.5152/electrica.2026.25293.
Stand-alone electric propulsion systems require high fault tolerance, and multiphase induction motor drives offer significant advantages over their three-phase counterparts in this regard. These drives can maintain functionality under fault conditions, ensuring reliable post-failure operation. This paper presents a predictive torque control technique applied to a five-phase induction motor. The proposed approach introduces a notable feature: the use of two distinct cost functions for torque and flux. This dual-function strategy simplifies the control process and enables a rapid dynamic response. The proposed technique's performance is evaluated against classical field-oriented control and direct torque control under both normal and open-phase fault conditions. Simulation results demonstrate the method's effectiveness in accurately tracking flux and speed references. Moreover, the proposed strategy significantly reduces flux and electromagnetic torque ripples, both during normal operation and post-fault. Cite this article as: F. Adil, Y. Meslem and M. Gouichiche, “An effective predictive torque control strategy for five-phase induction motors under normal and fault conditions,” Electrica, 2026, 26, 0172, doi:10.5152/electrica.2026.25172
This study presents a comprehensive analysis of the prevailing themes and technological advancements within the Internet of Medical Things (IoMT) literature over the past decade (2016–2025). Given its potential to fundamentally transform healthcare delivery, the IoMT represents a rapidly evolving and critical research domain. This analysis aims to identify future research trajectories and pinpoint pivotal challenges within biomedical engineering. Approximately 2000 articles were selected from the Web of Science database using the keyword "Internet of Medical Things." The derived text data was analyzed through the implementation of topic modeling, leveraging the Bidirectional Encoder Representations from Transformers (BERT) Topic method within a Python environment on Google Colab. The findings are presented as document and topic maps, visualized using topic probability distributions and dimensionality reduction techniques such as Uniform Manifold Approximation and Projection (UMAP)/t-Distributed Stochastic Neighbor Embedding (t-SNE). The BERT Topic analysis reveals that IoMT research is predominantly focused on enhancing security and privacy through edge computing and blockchain-based approaches. Furthermore, federated learning and deep learning algorithms demonstrate significant potential for specific medical applications, such as heart and mental disease prediction and COVID-19 management. The results highlight that data security, privacy, and reliability remain the three core themes of the IoMT ecosystem. Ultimately, this study demonstrates that cybersecurity and privacy vulnerabilities constitute the primary barriers to IoMT adoption, suggesting that future research must prioritize the integration of these protective mechanisms within real-time, resource-constrained devices. Cite this article as: H. Kuduz, G. Tonguç and K. K. Çevik, "A transformer-based semantic approach to internet of medical things applications: integrating bidirectional encoder representations from transformers topic modeling with dimensionality reduction," Electrica, 26, 0060 2026. doi: 10.5152/electrica.2026.26060.
This paper presents a design of flexible coplanar waveguide–fed slot-loaded antenna for wideband vehicle-to-everything communications in curved automotive environments. The proposed design incorporates strategically placed slots in both patch and ground planes to enhance impedance matching and bandwidth characteristics. Experimental results demonstrate an operational bandwidth from 2.44 to 7.26 GHz (99.38% fractional bandwidth) with stable radiation patterns, achieving gains between 3.45–4.6 dBi and maintaining 90% total efficiency across the band. The antenna's conformal capabilities are validated through fabrication and measurement, showing strong agreement with simulation results. Practical implementation is demonstrated via a careful integration on a vehicle wing mirror, with far-field performance analyzed using Ansys Savant tool. The compact 32 × 32 × 0.1 mm³ design exhibits excellent mechanical flexibility while meeting the demanding requirements of modern vehicular communication systems. Cite this article as: S. Virothu, R. Boddu, J., S. Varma and J. S. Roy, "A flexible coplanar waveguide-fed slot–loaded antenna for wideband vehicle–to–everything communications on curved automotive environments," Electrica, 26, 0266, 2026. doi: 10.5152/electrica.2026.25266.
