Faults and fast transients caused by lightning or the malfunction of power equipment, such as substation gas-insulated switchgears (GISs) and transmission tower insulators, can result in severe voltage sags, equipment damage, and wide-area outages in extra-high-voltage (EHV) power networks. To capture the arc physics governing these events, this paper presents a dynamic arc modeling framework in EMTP/ATP that integrates Mayr’s arc model with controlled switches for event reconstruction and lightning flashover studies. The framework simulates arc initiation, conduction, and extinction under realistic network conditions. The proposed approach is validated through three case studies: (1) a 345-kV GIS fault reconstruction based on an actual event, (2) reconstruction of a near-zero-impedance short circuit that occurred on a 345-kV transmission corridor and caused GIS failure in an EHV substation, and (3) a simulation study of lightning-induced flashover on a double-circuit transmission line, comparing air gaps without and with an external gapped line arrester (EGLA). Results obtained from the first two cases closely match recorded waveforms and reveal sub-cycle transients beyond the resolution of field recorders. In the third case, the results show that the EGLA protection mechanism achieves shorter low-voltage durations, lower insulator stress, and reduced arc currents. These findings demonstrate that the proposed arc model is a practical tool for fault event reconstruction, insulation coordination, and protection planning in high-voltage transmission systems.

Offshore wind farms (OWFs) significantly produce capacitive reactive power through their submarine connection to the main grid. Grid-connected resources must regulate their reactive power output to maintain reliable system operation. Excessive capacitive reactive power can be economically mitigated with the optimal planning of shunt reactors. This paper proposes an optimal shunt reactor planning method using the Equilibrium Optimizer (EO) algorithm. The proposed method is assessed in an actual OWF located on the west coast of Taiwan. The test consists of four cases with different shunt reactor configurations each. The results indicate that the ideal allocation of shunt reactors for the studied system consists of a smaller offshore shunt reactor capacity and a larger onshore shunt reactor capacity, each split into several individual shunt reactors. However, limited space availability and weight capacity on the OWF and the complexity of controlling too many shunt reactors may warrant less ideal allocation of shunt reactors. Furthermore, the solutions from the proposed EO-based method are superior to three other metaheuristic algorithms and one traditional solver.

In distribution systems, finding the exact location of a fault can be challenging due to uneven load distribution and different grounding methods. This paper introduces a hybrid data-driven method that uses data obtained from fault-induced transients to pinpoint the single-line-to-ground (SLG) fault occurring in the system. The process begins by analyzing voltage waveforms measured at the substation using the Clarke transform to detect whether a ground fault has occurred, based on the presence of zero-mode voltage. Then, the wavelet transform is applied to current waveforms to determine the fault type. To further analyze the fault, high-frequency spectral features of the zero-mode voltage are extracted using a high-pass filter and the fast Fourier transform (FFT). These features are then fed into a hybrid algorithm that combines clustering techniques with five different machine learning models, topped with an ensemble model for final classification and regression. This approach helps accurately identify both the area and the exact location of the SLG fault. The method is tested on two benchmark distribution systems and shows better performance than any single machine learning model alone. It also provides more accurate results than the traditional impedance-based method for locating SLG faults.

Photovoltaics (PV) integration in a distribution system (DS) is often limited by overvoltage or line overload concerns. With optimal placement, it is possible to maximize the PV capacity of integration while maintaining safe system operation. This paper compares several metaheuristic algorithms for the simultaneous optimal siting and sizing of several PV units for a DS. Compared to exact methods, metaheuristic algorithms are easily implemented and facilitate integration with off-the-shelf power system simulation tools. The algorithms are verified on two test systems. The test shows that metaheuristics are suitable for optimal PV planning. Additionally, proper placement of PV can improve system operations by reducing line loading and power losses.

This paper proposes a data-driven incipient fault detection framework based on voltage waveform data collected from GPS-enabled power quality monitors deployed across Taiwan’s southern science parks. The framework integrates a multi-stage process, including waveform-based incipient fault detection, fault identification, feature extraction, and incipient fault classification. By leveraging time-synchronized waveform events and historical fault records, the system identifies incipient fault events and establishes the correlation between incipient and permanent fault events. Based on the reliable classification performance, case studies are conducted to investigate the waveform characteristics of incipient faults across different equipment types. These findings contribute to a deeper understanding of incipient fault behaviors and offer insights for the implementation of incipient fault detection and classification framework.

Excessive capacitive reactive power in offshore wind farms, mainly due to submarine cables, can cause voltage instability and the violation of grid codes. This paper investigates optimal shunt reactor sizing to minimize the reactor-induced increase in system losses while satisfying regulatory requirements. Traditional methods, including a pared-down approach and sequential quadratic programming, are compared against metaheuristic algorithms such as particle swarm optimization (PSO), genetic algorithm (GA), and grey wolf optimizer (GWO). A two-stage case study evaluates reactive power imbalance and optimizes reactor sizing for a 495 MW offshore wind farm located off Taiwan’s west coast. Results show that metaheuristic algorithms, especially GWO, achieve better loss reduction and adaptive sizing than traditional methods. Thus, highlighting the necessity of onshore reactive compensation for grid compliance.

In the context of the 2050 net-zero emissions goal, accurately forecasting solar photovoltaic (PV) generation is crucial for maintaining a reliable grid operation. This paper explores clustering techniques in PV output forecasting using historical data from a PV power plant. Four clustering methods are employed, and clustering validation indicators assess tightness, dispersion, and similarity between clusters. By combining the most suitable method with the optimal clustering number of PV output patterns, Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (Bi-LSTM) prediction models are applied to forecast PV output. This is followed by a comparison between the predicted and actual PV generation.

Solar PV generation is widespread in the distribution system nowadays, and high-penetration PV brings challenges to reliable grid operation. Taiwan’s government has set a policy goal of achieving 20% of total electricity generation from renewable sources by 2025. However, the intermittent and unpredictable nature of renewable energy leads to voltage deviation and overvoltage in distribution system buses. Nevertheless, increasing PV capacity in distribution feeders is crucial to achieving the 2025 target. This study employs a stochastic analysis method to simulate PV hosting capacity accommodated in a distribution feeder. The study incorporates dynamic reactive power compensation devices for distribution feeders to assess the maximum hosting capacity of PV, in addition to smart inverters and energy storage systems (ESSs). Simulations are performed using OpenDSS with Matlab for power flow calculations and result analysis. The results show that the proposed method is effective for assessing PV hosting capacity in a distribution feeder.

Designing a large offshore wind farm necessitates thorough research to optimize economic returns. A significant consideration in this planning process is the electric power loss, which greatly impacts operational costs. Typically, wind power generation estimations are based on actual wind measurements, wind speed distributions, or equivalent capacity factors. This paper presents a comparative analysis of three methods for determining the annual energy production of an offshore wind farm. The study employs Monte Carlo simulation to facilitate a probabilistic load flow analysis. Furthermore, the equivalent capacity factor is obtained from actual wind speed measurements. Upon analysis, the test results highlight that the method utilizing monthly data as input shows superior outcomes in annual energy production and annual energy losses compared to the methods based on yearly data and capacity factors.

When the inverter-based microgrid (MG) operates in islanded mode, effective frequency regulation and equitable sharing of inverter power output among distributed generators (DGs) require the adoption of suitable control strategies. To simultaneously achieve these objectives in inverter control, this article proposes two methods: a novel adaptive inertia-based virtual synchronous generator (VSG) for frequency regulation, and an accurate power sharing mechanism based on a new feedforward line impedance approach for multiple parallel inverter-based DGs acting as voltage sources with stability analysis. By implementing the proposed methods, the VSG-based DG inverters can be appropriately controlled to support MG frequency and ensure even power sharing among inverters. Simulations using PSIM and hardware tests conducted with TMSF28335 DSP for a three-inverter MG are performed to demonstrate the efficacy of the proposed methods.

The widespread power electronics-based renewable energy resources and modern nonlinear loads in power systems present challenges in accurately assessing power quality due to their produced time-varying voltage and current waveforms. While several time-frequency analysis methods have been proposed to overcome the limitations of traditional frequency domain analysis, parameter adjustments are often needed for different types of signals. This paper focuses on applying Bayesian optimization to select the suitable parameters for the short-time Fourier-based multi-synchrosqueezing transform. The results obtained from simulations and actual measurements provide an appropriate range of parameter settings that reduce assessed time and validate the performance of the algorithm. Moreover, two feature importance indices emphasize the relevance and significance of the parameters.

This paper presents an efficient mixed-integer linear programming approach to unit commitment (UC) in isolated power systems while maintaining frequency stability under loss of generation contingencies. The dynamic frequency constraints considered are formulated using the center of inertia (COI) concept to determine the dominant parameters affecting system frequency response, specifically addressing the rate of change of frequency and the frequency nadir. Another significant contribution of this work is to propose a simple and computationally efficient method to obtain the lower bounds for the frequency constraints through the random sampling of generation unit combinations. Additionally, this paper introduces a load frequency sensitivity index-based method for determining the procured volume of dynamic regulation reserve from the ancillary service market. By incorporating these methods into the UC framework, the system can better adapt to system frequency dynamics. The proposed approach is validated using the benchmark system of Taiwan Power Company. Simulation results demonstrate the efficacy of the proposed methods in enhancing frequency stability during both low and high load demand days. Comparative analyses also highlight the benefits of Taipower’s internally provided dynamic regulation reserves. These reserves are enforced through frequency constraints in the proposed UC model, resulting in increased online units and enhanced frequency nadir compared to the conventional UC method. External procurement of dynamic regulation reserves is more cost-effective and recommended during low load demand days.

The offshore wind farm has become a growing focus in the renewable energy development and may cause significant harmonic distortion in power systems. Since each wind turbine can be regarded as a harmonic source, the accurate assessment of harmonic distortion produced by a wind farm is a crucial issue for the wind farm operation. When simulating harmonic sources, the model selections can substantially impact the calculated harmonic distortion results. Therefore, the accurate modeling of harmonic sources for the wind turbines becomes an indispensable issue. This paper aims to evaluate harmonic distortion in an offshore wind farm system under each wind turbine’s rated output. The study cases compare the differences in simulation results between the harmonic source based on the IEC 61000-3-6 and the DIgSILENT unbalanced phase correction (UPC) models. Then, the findings are reported to for harmonic assessment of an offshore wind farm.

Effectively sharing power among the distributed generators (DGs) in an inverter-based islanded microgrid often involves the use of virtual impedance. Achieving accurate power sharing while considering the effects of line impedance requires that the virtual impedance be relatively larger compared to the line impedance. However, the presence of line impedance still leads to significant power sharing mismatches, and both virtual and line impedances result in voltage drops between DGs and the point of common coupling (PCC). This paper proposes a practical method that utilizes virtual line impedance to mitigate the impact of line impedance on inverter output voltage and power sharing in an islanded microgrid. With this new method, PCC voltage is improved, and DGs' power output is accurately shared. Simulations were conducted using MATLAB/Simulink, and the results were subsequently validated using a commercial real-time simulator to demonstrate the effectiveness of the proposed method.

With the increasing demand for green energy, solar photovoltaic (PV) has become as a major source of electricity production. However, the high penetration of solar PV in the power grid poses challenges due to variations in electric power generation, making accurate PV output forecasting crucial for effective system operation. This paper proposes a hybrid method for day-ahead PV output forecasting based on historical data from an actual PV power plant. The study employs different weather clustering techniques and deep learning models. The forecast results of these models are then combined using an ensemble method. Results show that the proposed model improves accuracy across various weather conditions, and the method is suitable for day-ahead PV output forecasting.

This Task Force paper presents several important aspects on modelling single-phase power electronic converters (PECs) for power system distortion studies in the 2-150 kHz frequency range (also referred to as supraharmonics). The paper summarizes the state of the art and provides a systematic framework for the development of models of the unintentional emissions produced by modern PECs in this frequency range. Starting from a review of the basic concepts of the phenomena, model requirements, functionality and validation procedures are defined. The main modelling approaches proposed for power system distortion studies in the 2-150 kHz frequency range are discussed and compared. A detailed numerical case study of a fixed switching frequency converter is included to illustrate and compare the application and performance of different models.

Harmonics, inter-harmonics, and low-frequency (below the fundamental frequency) components produced by nonlinear loads and renewables may cause undesired effects on electric power equipment including overheating, resonance, and voltage flickers. This paper presents a hybrid method for the detection of frequency components of the measured waveform that is nonstationary (time-varying) in a holistic manner. In the proposed method, the multi-syncrosqueezing transform (MSST) is firstly applied to determine time-frequency ridges for the low-frequency, fundamental, harmonic, and interharmonic components. Next, the algorithm of density-based spatial clustering of applications with noise (DBSCAN) is adopted to precisely identify the prominent frequency components (i.e., intrinsic modes) using the MSST results. The time-domain waveforms corresponding to each detected frequency component and the original signal are then reconstructed via the inverse operation of MSST to assess the rms , and total harmonic and interharmonic distortion (THD and ITHD) trends of the recovered waveforms. The proposed method is validated by simulations and actual measurements to show its usefulness. Results are also compared with different time-frequency analysis methods to show the superiority of the proposed method.

When the inverter-based microgrid operates in grid-tied mode, the low voltage ride-through (LVRT) capability presents a challenge in its inverter controller design to support the grid voltage. This paper proposes a practical method to support the LVRT capability of an inverterbased grid-tied AC microgrid during grid fault. Simulations are performed using Matlab/Simulink and implemented in a real-time simulator with a hardware setup to validate the usefulness of the proposed method.

When the microgrid operates in grid-tied mode, the voltage ride-through (VRT) capability presents a challenge in its inverter controller design. This letter proposes an effective method to support the VRT capability of an inverter-based grid-tied microgrid during grid fault. In addition, the inverter control is implemented by a combined method to mitigate its output current harmonics under normal or faulty grid conditions, even if the grid voltage is distorted. Simulations are performed using Matlab/Simulink and implemented in a real-time simulator with a hardware setup to validate the usefulness of the proposed method.

Electrical power quality is a vital aspect when designing or assessing the operation of all modern power systems and forms an important part of the ongoing energy transition to more efficient and multi-vector systems. However, the ongoing proliferation and changing functionality of power electronic devices, coupled with new grid operating paradigms, such as renewable energy sources integration, microgrids, low-voltage dc distribution networks, and the large-scale integration of electric vehicles, present unique opportunities and challenges for grid operators the world over and require new assessment methods and fresh perspectives on the role of power quality.