International Transactions on Electrical Energy Systems最新文献

Power quality disturbances (PQDs), including voltage sags, swells, harmonics, transients, flickers, and interruptions, affect the reliability and efficiency of modern power distribution systems. This study introduces a novel heuristic model that integrates a long short-term memory (LSTM) deep neural network with the sea lion optimization (SLO) algorithm for precise classification and elimination of PQDs. The system comprises a voltage-controlled distribution static compensator (VCM-DSTATCOM), including an LSTM-SLO optimized PI controller to enhance reactive power compensation and voltage regulation performance. The proposed LSTM-SLO classifier is executed using the SLO algorithm, which enhances hyperparameter optimization, increases accuracy, and reduces computing time. All simulations and coding were conducted in MATLAB/Simulink 2019b on a 400 V, 50 Hz distribution network and PyCharm 2022. The classifier achieved a test accuracy of 99.10% with a convergence rate of 0.97%. The proposed VCM-DSTATCOM, utilizing optimal gains of PI controllers, effectively eliminated PQDs and reduced total harmonic distortion (THD) to 0.52%, compared to 15.17% with a conventional PI controller; furthermore, voltage stabilization was achieved with instrument response times under 20 ms. This study offers a practical method for addressing real-time PQ problems, with potential applications in smart grid power distribution. Future endeavors can focus on a customized hardware solution that can be integrated with the IoT environment for enhanced monitoring and control tasks.

This research investigates the application of the parrot optimizer (PO) algorithm and the artemisinin optimization (AO) algorithm to determine the optimal placement of solar-powered distributed generators (SP-DGs) within the standard IEEE 33-node, 69-node, and the Vietnamese Hoang Dieu 26-node distribution power networks (DPNs). Both algorithms were implemented to minimize active power loss under peak load conditions for a one-hour period. The optimal placement problem was analyzed for two scenarios of renewable-based distributed generation (RBDG) operation: Case 1 (pure active power generation) and Case 2 (active and reactive power generation). The performance of the two algorithms was compared against previous studies to identify a more effective approach. Furthermore, the AO algorithm was applied to minimize the total one-year energy loss (OYEL) in the Hoang Dieu 26-node DPN, utilizing real solar radiation data for Hoang Dieu Street in Ho Chi Minh City, Vietnam, obtained from a solar global atlas. The findings indicate that AO demonstrates superior effectiveness compared to PO and certain existing algorithms for the two standard DPNs. For the Hoang Dieu 26-node DPN, AO achieved significant reductions in OYEL, suggesting its suitability for optimally placing and operating RBDGs to yield substantial benefits for this specific network.

This paper presents a three-step identification-based dynamic equivalencing approach for multimachine electric power systems. Viewed from the bus of a local machine, a multimachine power system is modeled as a lossy two-machine equivalent (LTME) system. The LTME system comprises the local machine, an equivalent transmission line, and an equivalent synchronous machine connected in series. The LTME system parameters are estimated using the simulated time responses of the multimachine system following a disturbance. The required parameters are categorized into two types: electrical parameters (EPs) and mechanical parameters (MPs). The first two steps of the proposed method are formulated as linear recursive least-squares (RLS) estimation problems. An RLS algorithm with a memorizing factor is then employed in these steps to estimate the LTME system parameters. In the third step, a critical clearing time (CCT)–based method is used to determine appropriate values for the memorizing factors applied in the first two steps. Transient stability studies of power systems can be conducted using the resulting LTME system. The proposed approach is validated on the New England 10-generator, 39-bus test system using the classical machine model, and the simulation results are discussed in detail.

Accurate determination of critical clearing time (CCT) remains a fundamental challenge in transient stability analysis of power systems, primarily due to the highly nonlinear and complex dynamic behavior of multimachine networks. Existing methods based on the loss of synchronism (LOS) condition often require intensive numerical iterations and may suffer from convergence issues, particularly near the critical stability boundary. To address these limitations, this paper proposes a novel CCT estimation approach based on the critical trajectory method with a modified loss of synchronism (MLOS) condition. Unlike conventional LOS methods, the proposed MLOS approach linearizes the critical condition by assuming that the directional variation of the eigenvector associated with the zero eigenvalue of the Jacobian matrix is negligible, allowing it to be approximated as an identity matrix. This modification simplifies the critical trajectory formulation, significantly reduces the computational burden, and improves the numerical stability of the CCT determination process without sacrificing accuracy. Simulation results conducted on multiple benchmark test systems demonstrate that the proposed MLOS method achieves higher computational efficiency and comparable, if not superior, accuracy relative to traditional LOS-based approaches. These results highlight the effectiveness and robustness of the MLOS method, making it a promising tool for accurate and efficient transient stability assessment in modern power systems.

Ferroresonance poses a major threat to the quality and reliability of power distribution systems due to its inherent characteristics of sustained overvoltages and currents. This paper aims to enhance the understanding and reduce the ferroresonance threat by investigating the susceptibility of different transformer configurations using MATLAB/Simulink simulations. To achieve this, four 200 kVA transformers with different vector groups (D11-Yn, Yg-Yg, Yn-Yn, and Y-D11) and core types (3-limb and 5-limb) were systematically exposed to controlled ferroresonance conditions. The influence of varying the length of the 11 kV cable connected to the transformers was also examined. Unlike previous studies, which primarily relied on waveform analysis, our approach integrates total harmonic distortion of voltage (THDv), total harmonic distortion of current (THDi), peak overvoltage, peak current, and energy content analysis of the ferroresonance oscillations. This methodology facilitates a more rigorous and comparative evaluation of transformer susceptibility, equipping utilities and manufacturers with practical tools to assess and mitigate ferroresonance risks in real-world applications. The findings indicate that the Y-D11 configurations exhibited lower susceptibility to ferroresonance than the others. It was also observed that ferroresonance effects are most pronounced within a cable length range of 1.5 km–2 km, beyond which the distributed capacitance helps to moderate the severity. A key contribution of this research is the development of a multimetric ferroresonance susceptibility framework. This framework advances beyond traditional qualitative assessments by providing a data-driven methodology for evaluating transformer vulnerability.

Regular monitoring of outdoor insulators is crucial to ensure the reliable functioning of the power grid. With recent progress in computer vision technologies, traditional manual and expensive visual inspections can now be replaced by automated analysis using images captured by unmanned aerial vehicles (UAVs). In such applications, a practitioner might opt to choose a state-of-the-art object detection and classification deep learning architecture, including You Look Only Once (YOLO). The variety of existing YOLO architectures per se makes selecting the best application-dependent YOLO model challenging. However, selecting the best architecture solely based on performance without considering the model complexity limits its deployment on resource-limited embedded devices. Consequently, we conduct a rigorous, systematic model selection based on the performance–complexity trade-off across 13 YOLO architectures to determine the most effective model for detecting common mechanical faults in insulators using images captured by UAVs. A dataset comprising 15,000 images of insulators, categorized into normal condition, bird-pecking damage, cracks, and missing caps, has been compiled for training the models. Specifically, all considered YOLO architectures are compared using model complexity and the [email protected]:0.95. During the model selection stage, YOLOv8l proved to be the best model in terms of [email protected]:0.95, while YOLOv5n was the model of choice in terms of complexity at the expense of a slight reduction in performance. Alongside YOLOv8l and YOLOv5n, an “optimal” model (OP-YOLO) was selected using a multicriteria decision-making approach, balancing detection accuracy and computational efficiency. In particular, in terms of test-set performance, YOLOv8l, YOLOv5n, and OP-YOLO achieved 0.919, 0.901, and 0.896 [email protected]:0.95, respectively. Although YOLOv8l reported a higher [email protected]:0.95, YOLOv5n requires ∼20.9 times less memory and ∼40.2 times less floating-point operations per second (FLOPs). Also, YOLOv5n outperforms the OP-YOLO model, still requiring ∼12 times less memory and ∼19 times less FLOPs.

In grid-connected smart distribution systems, the prediction and mitigation of flicker caused by harmonics and interharmonics present a significant challenge for stable grid operation, particularly in the presence of distributed renewable energy sources (DRESs). The intermittent nature of DRES introduces dominant low-frequency components that exacerbate flicker issues in smart grid environments. To address these challenges, this research proposes a deep convolutional neural network (DCNN) model, employing a mean squared error loss function, designed to outperform conventional active power filters and static Var compensators (SVCs) in flicker mitigation. The training dataset for the proposed DCNN was obtained from real-time measurements at the Muppandal Wind Farm in Tamil Nadu, India. Numerical evaluations based on flicker sensation, prediction accuracy, perceptibility, and error rates demonstrate the superior performance of the proposed method compared to existing techniques. The results confirm that the proposed DCNN model is a viable solution for real-time flicker prediction and mitigation in smart grid applications, especially those integrating intermittent renewable energy sources.

This paper addresses a novel single-ended Zeta (SEZE) converter topology for bidirectional power flow in grids, enabling both vehicle-to-grid (V2G) and grid-to-vehicle (G2V) operations. The primary challenge tackled is the reduction of total harmonic distortion (THD) and improvement of power quality in grid-connected charging systems. The SEZE topology distinguishes itself with a modular structure, fewer switching devices, and enhanced energy transfer efficiency, directly addressing conventional limitations related to size, complexity, and harmonic performance. The SEZE framework uses the fuzzy logic controller (FLC) for operating the converter in V2G and G2V modes. During the V2G operation, the FLC controller provides the lowest current THD of 3.51% whereas the PID and PI controllers provide a THD of 4.83% and 10.83%. Moreover, in the G2V mode, the FLC controller provides the lowest current THD of 3.02% whereas the PID and PI controllers provide a THD of 5.02% and 6.64%. From the results, it is identified that the FLC works better than the conventional PID and PI controllers. The proposed converter also shows an efficiency of 97.1% and grid compliance than existing charging architectures. The SEZE converter, with fuzzy control, greatly improves power quality and makes sure that power can be transferred efficiently from V2G and G2V.

To leverage the complementary advantages of hydrogen, electricity, and heat, this paper proposes a coordinated optimization scheduling method for an integrated electricity–hydrogen–heat energy system, considering the dynamic operating temperature of proton exchange membrane electrolyzers and the thermal storage capacity of pipeline networks. First, a refined PEM operational model is established by analyzing its hydrogen production efficiency and temperature characteristics under fluctuating power scenarios. On this basis, a comprehensive hydrogen lifecycle model is established, encompassing production, storage, and utilization. On the thermal supply side, a quantitative pipeline network thermal storage model is constructed using a fictitious node method to further explore the system’s flexibility potential. Finally, to achieve optimal economic performance and maximize the utilization of wind and solar energy, an integrated optimization scheduling model is formulated, considering the operational constraints of all devices. Case study results demonstrate that PEM electrolyzers can convert excess electricity into stored heat at the cost of reduced hydrogen production efficiency, effectively facilitating energy flow coordination. Moreover, the thermal storage capability of the pipeline network enhances the system’s overall heat regulation capacity, maintaining PEM hydrogen production efficiency and promoting the local consumption of renewable energy.