Abstract A hybrid energy storage system (HESS) plays a crucial role in stabilizing DC microgrids against power fluctuations from renewable sources and loads. To mitigate severe bus voltage deviations under complex disturbances, this paper proposes a composite control method integrating a super-twisting sliding mode controller (STSMC) with an extended state observer (ESO). Combined with virtual impedance droop control, the proposed method enables coordinated power sharing in the HESS, where the ESO estimates disturbances and the STSMC improves transient performance. A decentralized secondary voltage recovery (SVR) strategy with decoupled battery (BAT) and supercapacitor (SC) control is further introduced to restore the bus voltage and rapidly recover the SC state of energy (SOE). Under the proposed scheme, the BAT handles low-frequency power components, while the SC compensates for high-frequency fluctuations. Simulation and controller hardware-in-the-loop (CHIL) tests validate the effectiveness of the proposed method in improving the dynamic response and stability of the HESS.
Abstract Public-building central air conditioning (CAC) systems provide substantial load flexibility because of their significant thermal inertia. However, existing optimization studies of park energy systems commonly represent them as conventional flexible loads, with insufficient consideration of zonal differences and occupant thermal comfort. This paper proposes a coordinated optimization method for CAC systems participating in park-level demand response. First, an aggregated demand response model is developed by considering building thermal inertia, indoor-temperature dynamics, load-ramping limits, post-response rebound, and terminal-temperature recovery. Second, a regional temperature control considering zonal demand (RTCD) strategy is proposed by evaluating occupant density and temperature demand in different building zones. The CAC system, residential demand response, multiple energy-storage systems, and electricity–gas–heating–cooling conversion equipment are then incorporated into a unified day-ahead scheduling model. The RTCD strategy is compared with the temperature-based demand response (Tem), Model Predictive Control (MPC), and central-air-conditioning non-demand-response (Non-DR) strategies. Relative to Non-DR, RTCD reduces operating costs by 2.46 % and 1.67 % on working and non-working days, respectively, and decreases peak grid-purchased power by approximately 5.7 % under both conditions. The renewable-energy utilization rate increases to 95.55 %, while the maximum reduction in grid-purchased power during the working-day peak period reaches 11.7 %. The average indoor-temperature deviation remains at 0.29 °C. These results demonstrate that RTCD improves operating economy, peak shaving, and multi-energy coordination while maintaining acceptable thermal comfort.
Abstract The selection of earthing system configuration and design play the critical role in operational integrity stability and reliability of electrical power systems. This article presents comparative assessment of different earthing grid configurations to obtain the most suitable grid shape with minimum values of safety parameters. Four configurations of earthing grid (Square, rectangular, T-shape and L shape) were used in this analysis. Considered earthing grids are buried in two-layer soil structure. Equal coverage area of substation was considered for all adopted configurations of earthing system. E-EG and U-EG designs were used in the analysis and, it is observed that U-EG design gives better results in terms of safety parameters. Performance of earthing grid with various configurations, grid burial depth, and different soil models are analyzed. IEEE standard 80–2,000 and variable spacing method are adopted to design the earthing system and all safety requirements are strictly followed as per the prescribed standard. Effects of varying resistivity in two layer soil model on earthing system resistance are also evaluated for E-EG. Rectangular U-EG design exhibited consistent results in terms of GPR, touch voltage and step voltage that improve system safety and performance. Vertical earthing rods were located at the periphery of earthing grids.
Abstract Every year, the load growth in distribution networks causes various challenges such as voltage drop. Since some of these challenges are due to the lack of reactive power sources, these problems can be solved by installing reactive power resources such as capacitors. On the other hand, considering that the installation of capacitors in the network based on one or more hours of the study period (due to the impracticality of carrying out load flow studies in all the hours of the multi-year study period) caused to over investment in capacitors and becomes uneconomical, a solution should be provided in order to optimally allocate these resources based on all the hours of the multi-year study period. Therefore, in order to solve the problem raised and in order to rationally allocate reactive power compensation devices in distribution networks, a method based on fuzzy logic is developed in this paper to take into account the amount of network load in all hours of the study period and at the same time the execution of simulations be very short in order to solve such problems. Hence, in this paper, as the first contribution, the considered problem has been solved using the method based on standard fuzzy logic, taking into account all the hours of the study period, and then as the second contribution, the optimization of the parameters of the fuzzy membership functions has been discussed, and the proposed method has been upgraded. The simulation results showed that the developed method has sufficient efficiency and it can be used to optimally allocate capacitors in the network by considering all the hours of the study period. Also, the developed method was able to improve the objective functions to a greater extent by optimally setting the parameters of the fuzzy membership functions. Optimizing the parameters of the fuzzy membership functions made the objective function of the total costs to be improved by 0.8, 1.9 and 2.5 percent by installing one, two and three capacitors, respectively. Also, this optimization caused the objective function of voltage profile to improve by 0.4, 0.7, and 1.2 percent, respectively, by installing one, two, and three capacitors.
Abstract In this study a cost effective and simplified topology of modified thyristor-controlled LC compensator-based fault current limiter (MTCLC-FCL) has been proposed to enhance the system stability and improve the power quality in microgrids. In microgrid environment with high usage of power electronic converters, issues such as harmonic distortion, voltage variations, and dynamic instability frequently arise. The proposed MTCLC-FCL is designed to mitigate power oscillations and reduce voltage sags caused by the intermittency of solar and wind generation as well as to suppress the fault currents under severe fault conditions. The performance of the MTCLC-FCL, integrated with microgrid is evaluated through MATLAB/Simulink studies. The simulation outcomes demonstrate that the overall dynamic response improves and voltage sag along with total harmonic distortion (THD) substantially reduces, thereby the overall power quality of the microgrid enhances. The comparative assessment on the basis of THD, transient stability indices, and fault current reduction levels, indicates that the proposed MTCLC-FCL topology gives better system performance compared to traditional superconducting FCLs (SFCL) and thyristor-controlled series compensation (TCSC) devices.
Abstract While doubly fed induction generators (DFIGs) are critical to modern high-power wind energy conversion systems (WECSs), their nonlinear dynamics and high sensitivity to parameter drift present formidable control challenges. This paper proposes a high-performance, sensorless adaptive control framework for the rotor-side converter (RSC) utilizing a three-level neutral-point-clamped (3L-NPC) topology. The core innovation seamlessly integrates finite-set model predictive control (FS-MPC) with a parallel moving horizon estimation (P-MHE) scheme, driven by a compact, complex space-vector DFIG model mapped in the stationary frame. This novel formulation mathematically reduces the conventional fourth-order α – β model into a second-order representation. Crucially, it eliminates explicit dependence on the synchronous speed ( ω s ), bypassing complex linear parameter-varying (LPV) structures to ensure high computational efficiency for real-time execution. The decoupled P-MHE architecture simultaneously delivers rapid state estimation and high-fidelity tracking of both rotor speed ( ω m ) and critical ohmic resistances ( R s , R r ). These continuous parameter updates yield a self-correcting FS-MPC law that actively maintains optimal torque and current regulation despite severe parametric drift and measurement noise. Extensive MATLAB/Simulink validations confirm the strategy’s exceptional transient response, steady-state accuracy, and robust sensorless operation across a diverse spectrum of dynamic wind speeds and load conditions.
Abstract The voltage stability limit (VSL) is crucial for power system operators to ensure grid security by defining the maximum loading before risking voltage collapse, thus preventing cascading events and maintaining reliable power supply. Higher proportion of renewable integration necessitates the accurate consideration of forecast uncertainties when determining VSL distribution. However, probabilistic VSL evaluation has received limited attention in the literature. This article puts forth a novel framework for efficient and accurate probabilistic VSL evaluation in large, contemporary power systems. Addressing the limitations of existing methods, the proposed framework incorporates several key advancements. Firstly, Transversality Enforced Newton-Raphson (TENR) is employed as an efficient alternative to Continuation Power Flow (CPF) for VSL determination, significantly reducing computational burden. Secondly, the framework utilizes R-vine copula, a flexible and data-driven model, to accurately capture complex dependence structure among wind power forecasts, surpassing the limitations of standard Gaussian and restricted vine copula models. Finally, a novel scenario generation approach integrating Maximum Projection Design (MPD) based R-vine sampling is proposed to enhance the accuracy-efficiency balance of probabilistic VSL evaluation. The proposed framework is rigorously validated on modified 39-bus, 118-bus, and 2383-bus test systems against four state-of-the-art benchmark methods. Numerical results substantiate the superior VSL distribution estimation accuracy of the proposed framework, alongside an approximate 25-fold computational speedup compared to existing benchmarks for large-scale networks, fulfilling a critical industry need for a scalable and accurate VSL evaluation tool for step-ahead and day-ahead security studies with N -1 contingency compliance.
Abstract The mass deployment of solar photovoltaic generation in distribution networks induces rapid voltage variations, reverse power flow, and frequent on-load tap changes, thereby undermining voltage quality and the longevity of on-load tap changing units. Classical on-load tap-changing controllers can only act on instantaneous voltage readings and, as a result, exhibit a reactive response to disturbances caused by solar photovoltaic generation, inevitably leading to unnecessary tap changes and tap-hunting behaviour. The current work presents a predictive control algorithm based on a family of rules that follows the principles of model predictive control logic and tap-budget control, similar to the token-bucket algorithm used in computer networks. This on-load tap changing-predictive control algorithm model includes: (i) a low-price, field-tested on-load tap change retrofit controller with phase-sequence detection and hysteresis-based intelligent tap logic, and (ii) short-term solar photovoltaic generation voltage prediction based on different deep learning techniques. The tap actions are also activated when the measured and predicted deviations exceed their respective limits, making predictive voltage regulation possible and reducing mechanical wear and tear. When testing, validation is performed using a 1-year dataset for an 11 kV/433 V campus feeder with a 2 MW peak load and optional 0.5, 1, and 1.5 MW distributed solar photovoltaic plants. Findings indicate improved voltage regulation performance, with approximately 57 % fewer tap operations, while maintaining stable voltage regulation under reverse power flow conditions and phase reversal conditions. The suggested methodology provides a scalable, cost-effective solution for modernizing on-load tap-changing systems in distribution networks in developing regions.
Anomaly detection and disaster response based on real-time power grid data are essential for ensuring the stability and robustness of power grids. This paper proposes an integrated framework combining multi-scale 1D-CNN for local feature extraction, LSTM for temporal dependency modeling, and dual-branch temporal-channel attention for adaptive feature weighting, coupled with a D3QN-based disaster response module for low-latency sequential decision-making. Experimental validation on a real-time power monitoring dataset demonstrates an anomaly detection accuracy of 95.2 %, an F1 score of 93.8 %, a disaster recovery rate of 94.6 %, and an overall system response latency of 112.6 ms, outperforming CNN-BiLSTM-AdaBoost and four other baselines. Future work will focus on reducing computational overhead and improving model robustness under large-scale deployment conditions.
Abstract Data centers are recognized as major energy consumers in the modern digital economy. The low-carbon transformation of these facilities is considered an urgent task. In this paper, a capacity planning method for multiple data centers with shared hydrogen energy storage is proposed. The data center loads are classified into interactive loads and deferrable loads. A temporal shifting model is established for deferrable loads. The spatio-temporal complementarity among multiple data centers is utilized. A shared hydrogen energy storage framework is developed. The framework consists of electrolytic cells, hydrogen tanks, fuel cells, and thermal storage systems. The decoupled operation of hydrogen production, storage, and utilization is exploited. A combined heat and power model is established for thermal energy utilization. A cooperative game theory approach is adopted for capacity optimization. An improved Shapley value method is proposed for fair cost allocation. Case studies are conducted to verify the proposed method. The investment cost is reduced by 30.2 % under the shared storage mode. The carbon emission cost is reduced by 41.9 %. The utilization efficiency of energy storage resources is significantly improved.
Conventional model predictive control-based virtual synchronous generator control technology usually has obvious active power oscillation with static error and large current control errors. To solve these demerits, an improved active power oscillation suppression method for virtual synchronous generator with reduced current control errors is proposed in this paper based on damping power adjustment and model predictive control. The principles of the conventional virtual synchronous generator control technology are reviewed firstly, and its drawbacks are analyzed. Then, an improved active power oscillation suppression method is designed by introducing a PI controller to dynamically adjust the damping power. Theoretical analysis is carried out to validate the effectiveness of the proposed method. Next, for the current control of the virtual synchronous generator, the conventional model predictive control method is explained, and a simple double-vector model predictive current control method is further proposed to reduce the current control errors. A theoretical analysis is also carried out, which verifies the effectiveness of the proposed double-vector current control method. Finally, comparative experimental researches are carried out, which verify the validity of the proposed control methods in this paper.
Abstract This manuscript proposes a single-phase hybrid fault-tolerant seven-level inverter for grid-connected applications. In conventional multilevel inverters, a fault in any switching device can adversely affect the output voltage levels, making the inverter unsuitable for reliable grid integration. Therefore, incorporating fault-tolerant capability is essential to ensure uninterrupted operation. In the proposed approach, fault tolerance is achieved in a cascaded H-bridge inverter by integrating a reduced-switch inverter. The developed topology comprises a reduced-switch seven-level (RSSL) inverter in combination with two cascaded H-bridge modules. A level-shifted pulse-width modulation (LSPWM) technique is employed to generate the required gating signals for the proposed inverter. Furthermore, the capacitors in the RSSL inverter maintain inherent self-voltage balancing through the adopted control strategy. The performance, suitability, and effectiveness of the proposed fault-tolerant inverter for grid-connected applications are validated through both MATLAB/Simulink simulations and experimental hardware implementation results.
Transformer oil temperature is an important online-monitoring indicator reflecting the thermal state, insulation stress, and operational safety margin of power transformers. Accurate short-term prediction of oil temperature can provide actionable support for condition monitoring, early warning, and operation and maintenance decision-making under fluctuating load and environmental conditions. However, oil-temperature series usually exhibit strong nonstationarity, multiscale coupling, and thermal-inertia-induced lag, which make direct end-to-end prediction on the raw sequence prone to unstable generalization during operating-condition transitions. To address this issue, this study develops a decomposition-enhanced hybrid temporal learning framework for short-term transformer oil-temperature prediction. The raw oil-temperature sequence is first decomposed by VMD and CEEMDAN into time-aligned multichannel components to reduce scale aliasing and improve the separability of operational fluctuations and disturbance-related patterns. These components are then modeled by a hybrid Transformer-BiGRU architecture to jointly capture cross-time-step dependency and local temporal evolution. Experiments on the ETTh2 transformer temperature dataset show that the proposed method outperforms representative baseline models, achieving RMSE = 0.0219, MAE = 0.0160, R2 = 0.9977, MAPE = 4.87 %, sMAPE = 4.73 %, and WQE = 4.835. Compared with the BiGRU benchmark, the proposed model reduces RMSE, MAE, MAPE, sMAPE, and WQE by 24.33 %, 26.34 %, 19.90 %, 19.97 %, and 20.07 %, respectively. Compared with the strongest conventional baseline, Ridge, it further reduces RMSE, MAE, MAPE, sMAPE, and WQE by 5.19 %, 6.43 %, 3.37 %, 3.47 %, and 3.65 %, respectively. In addition to improved numerical accuracy, the model demonstrates better tracking of peak-valley variations and a more concentrated residual distribution, indicating stronger robustness under nonstationary operating conditions. These results suggest that the proposed framework is a promising data-driven tool for transformer condition monitoring and early warning, and may support more reliable thermal-state assessment in practical power-system operation.
Multilevel inverters are essential for maintaining the functionality of safety-critical applications. This article describes the Triac-Assisted fault-tolerant (FT) 5-L inverter (TFT5LI), which is developed to withstand the single and simultaneous multiple switches' open circuit (OC) faults. It is designed by connecting a redundant switch unit to the main circuit. The main switch fault can be avoided by detecting and diagnosing the fault in a shorter period of time to continue the healthy operation. The suggested TFT5LI maintains the same output power as the healthy mode in all fault scenarios with the aid of a redundant unit. The results of simulated and experimental investigations under different faulty scenarios, that confirms the proposed TFT5LI's ability to withstand the faults for 1,000 W output power, have been presented. The efficiency of the proposed converter was found to be 97.9 % under healthy conditions, with a current THD as 0.58 %, and the level to switch ratio is to be calculated as 0.45. This is a dependable option for emergency loads because of its improved qualities, which include a reduced device count, increased efficiency, a level-to-switch ratio, and a lower overall blocking voltage. To support the findings, comparative analysis along with component count analysis is presented in detail to show the effectiveness of the proposed topology over other modern topologies. A cost analysis of the proposed TFT5LI has been carried out to highlight its economic benefits.
Industrial and commercial user-side photovoltaic-energy storage systems face challenges in coordinating multiple objectives such as economic and technical performance during the planning and configuration phase. Traditional methods often fail to accurately account for the impact of demand charges on system benefits. This paper proposes a bi-level multi-objective configuration and operation optimization method for photovoltaic-energy storage systems that considers the demand increment effect. A bi-level optimization framework is constructed, combining an upper-level multi-objective genetic algorithm with a lower-level mixed-integer linear programming model. The upper level optimizes the photovoltaic and storage capacity, while the lower level solves for the optimal operation strategy and quantifies the demand increment effect. An accelerated evaluation strategy based on "typical months" is designed to align with the monthly settlement characteristics of demand charges. Simulation verification using data from a typical industrial and commercial park shows that the method yields a set of Pareto optimal solutions revealing the trade-off relationship among objectives such as return on investment and photovoltaic self-consumption rate. The proposed method achieves coordinated optimization of planning and operation for photovoltaic-energy storage systems, providing decision-making support for users to develop configuration schemes that balance both economic and technical objectives under multiple goals.
For the neutral-point-connected open-winding induction motor based on auxiliary bridge arms topology (AB-NPC-OEWIM), the traditional single vector model predictive control (SVMPC) suffers from issues such as large current ripples and significant zero-sequence current (ZSC). To address these problems, a simplified hybrid vector model predictive control (HVMPC) method is proposed in this paper. The studied AB-NPC-OEWIM is supplied by double three-phase four-leg inverters, where the DC bus voltage ratio between the high-voltage inverter (INV1) and the low-voltage inverter (INV2) is set to 2:1. For the proposed HVMPC, the three-dimensional (3D) voltage vectors (VVs) of the three-phase four-leg inverter is first projected onto the alpha beta coordinate by the 3D voltage vector (VV) to two-dimensional (2D) VV conversion method. Then, the optimal VV of the INV1 is selected firstly. Next, six virtual VVs with the duty cycle calculated by the cost function values are defined for the INV2 and the optimal virtual VV is selected. Finally, by evaluating zero-sequence current using the redundant vectors, the optimal VVs for INV1 and INV2 are finally determined. Additionally, to show the effectiveness of the proposed virtual VV based MPC method for INV2, a new visualization analysis method is proposed, which depicts the control error vividly using a 3D diagram, indicating that the proposed MPC can reduce the control error obviously, resulting in reduced current ripples. Finally, experimental studies are carried out, which verify the effectiveness of the proposed HVMPC method for AB-NPC-OEWIM.
The systems in power plants exhibit intricate thermal and mechanical behaviours, making early fault detection and accurate forecasting crucial for maintaining safety, efficiency, and operational reliability. Existing methods face issues such as inconsistency among multi-source data, poor representation of thermal dynamics, and lack of ability to model long-term dependencies, which result in the delay or inaccuracy in fault detection. In response to these challenges, an AI-oriented multi-source data fusion framework is proposed in this study that combines advanced deep learning techniques for better thermal control, fault detection, and forecasting of future system behaviour in the power plant area. In this study, the dataset was obtained from the Kaggle repository of Power Plant Data: Steam Turbine and Boiler Metrics which contains normal as well as faulty operational conditions. Once the data was collected, extensive preprocessing was performed consisting of missing value handling, Minmax normalization, and noise reduction to keep the data high-quality. The extraction of features was based on analysis, domain knowledge, and time-series representations, which reinforced the thermal characterizations. A combined deep learning model of CNNs for spatial feature extraction and LSTMs for temporal pattern learning was developed for the purpose of fault detection and remaining useful life (RUL) prediction. The new method not only traditional ways but also gave very high measurement values such as 0.9988 for accuracy, 0.9968 for precision, and 0.9955 for recall and 0.9961 for F1-score. Besides, the remaining useful life prediction had so small errors the MAE, RMSE, and MAPE all indicated convergence. Thus, the results support the hybrid model's power and reliability in prediction. This research presents a highly effective and adaptable approach for the continuous monitoring, proactive maintenance, and enhanced decision-making of modern power plant operations.
The dynamic economic dispatch problems can be either single-objective or multi-objective depending on the concerns of the operator, and it is used in order to minimize the total generation cost and transmission losses. In this research, the day-ahead dynamic economic dispatch is investigated that involves diesel generators to find a solution with minimum transmission losses, a smaller number of iterations, and a smaller computational time. The aim of this study is to find the optimal schedule for the generation that minimizes the generation cost and transmission losses of generating units for a 24 h period. Grey wolf optimizer algorithm (GWO) is used in the study to deal with the dynamic economic dispatch problem, whose results are compared with results of many other powerful algorithms. The proposed approach is tested on two benchmark systems with 5 and 10 generating units over a 24 h scheduling horizon. The simulation results are achieved by grey wolf optimizer algorithm then compared with different algorithms and techniques in order to prove the high-quality effectiveness of grey wolf optimizer algorithm in handling the dynamic economic dispatch problem especially in minimizing transmission losses including all the supply-demands equality constraints, valve point effects, ramp rate constraints, upper and lower generation limits for each generating unit. The results show that the GWO attains competitive generation cost minimization while effectively reducing transmission losses, with rapid convergence and modest computational requirements. Comparisons with several state-of-the-art optimization methods further demonstrate the robustness and effectiveness of the proposed framework.
Soiling and degradation are two major challenges that reduce the performance and forecasting accuracy of solar photovoltaic (PV) systems. While degradation effects have been studied, the combined influence of soiling and degradation on power forecasting remains largely unexplored. Soiling is particularly difficult to model because it has no direct correlation with irradiance or temperature and varies with cleaning schedules. As a result, power output can differ under identical environmental conditions depending on the soiling ratio (SR), leading to increasing forecasting errors. This study presents a transformer-based forecasting framework that explicitly corrects for both soiling and degradation. The proposed method first removes soiling effects from historical data, enabling the model to learn clean generation patterns. Forecasted power is then adjusted in real time using the SR. To handle degradation, the model employs periodic retraining (PR), maintaining accuracy as system capacity decreases. Tested on a 504 kWp rooftop PV plant, the framework achieved a mean relative error of 3.57 %, RMSE of 30 kW, and MAPE of 9.8 % over one year. Long-term analysis shows RMSE reduction of 32 % and MAPE reduction of 69 % over 20 years, confirming its robustness for reliable PV forecasting.
Self-excited induction generators (SEIGs) are suitable for renewable energy applications due to their simple construction, low cost, and minimal maintenance requirements. SEIGs are superior to traditional synchronous generators in edge-of-grid pico-hydro systems because they offer inherent self-excitation and do not require external excitation. This research studies the power quality performance of SEIGs by analyzing voltage regulation, frequency variation, total harmonic distortion (THD), and power factor (PF) under varied rotor speeds and excitation capacitance. The behavior of SEIGs under abnormal operating conditions is also examined through extensive fault analysis, including open circuit faults and short circuit faults such as single line-to-ground (LG), line-to-line (LL), double line-to-ground (LLG), three-line (LLL), and three-line-to-ground (LLLG) faults. These source parameters variation and fault condition are performed in hardware setup and the results are executed. Experimental results reveal that terminal voltage and current THD increase with increasing speed and excitation capacitance. Depending on the load and capacitor configuration, the SEIG can tolerate both single- and double-conductor open circuit faults. The generator can withstand LG faults up to 40 % load with delta connected excitation capacitors, but other short-circuit faults result in stalling. These findings support SEIGs applicability for pico-hydro power production.
