AC series arc is dangerous and can cause serious electric fire hazards and property damage. This article proposed a convolutional neural network -based arc detection model named ArcNet. The database of this research is collected from eight different types of loads according to IEC62606 standard. The two most common types of arcs, including arcs from a loose connection of cables and those caused by the failure of the insulation, are generated in testing and included in the database. Using the database of raw current, experimental results indicate ArcNet can achieve a maximum of 99.47% arc detection accuracy at 10 kHz sampling rate. The model is also implemented in Raspberry Pi 3B for classification accuracy. A tradeoff study between the arc detection accuracy and model runtime has been conducted. The proposed ArcNet obtained an average runtime of 31 ms/sample of 1 cycle at 10 kHz sampling rate, which proves the feasibility of practical hardware deployment for real-time processing.

Due to detection difficulties, dc arc faults are one of the most dangerous risks in dc power systems. Most traditional studies are based on arc current, which may change during normal operation and cause unwanted trips. Another problem with the traditional methods is that the detection threshold may need to be adjusted for different photovoltaic (PV) systems; otherwise, it will cause malfunctions. To solve the aforementioned problems, a series arc fault detection method based on steady patterns of the frequency domain is proposed. The proposed method utilizes the electromagnetic radiation (EMR) emitted by an arc as a testing basis, avoiding the occurrence of unwanted trips. Patterns such as the structural similarity index (SSIM) and 6-dB bandwidth bins (6-dB BWBs) are calculated to extract the similarity of the steady-burning arc spectra. The experimental verification shows that the proposed steady-pattern-based method can accurately identify arc faults in different dc power systems, discriminate arc faults from normal operations, effectively avoid the occurrence of malfunctions, and can be used as a supplementary technique to traditional methods.

Although photovoltaic (PV) systems play an essential role in distributed generation systems, they also suffer from serious safety concerns due to DC series arc faults. This paper proposes a lightweight convolutional neural network-based method for detecting DC series arc fault in PV systems to solve this issue. An experimental platform according to UL1699B is built, and current data ranging from 3 A to 25 A is collected. Moreover, test conditions, including PV inverter startup and irradiance mutation, are also considered to evaluate the robustness of the proposed method. Before fault detection, the current data is preprocessed with power spectrum estimation. The lightweight convolutional neural network has a lower computational burden for its fewer parameters, which can be ready for embedded microprocessor-based edge applications. Compared to similar lightweight convolutional network models such as Efficientnet-B0, B2, and B3, the Efficientnet-B1 model shows the highest accuracy of 96.16% for arc fault detection. Furthermore, an attention mechanism is combined with the Efficientnet-B1 to make the algorithm more focused on arc features, which can help the algorithm reduce unnecessary computation. The test results show that the detection accuracy of the proposed method can be up to 98.81% under all test conditions, which is higher than that of general networks.

Arc faults in DC microgrids or photovoltaic (PV) systems are difficult to detect and isolate. Especially in location research, there are few related results, and arc fault detection is immature. This paper proposes a planar location method that requires only two detection points. Because the height of many DC systems is much smaller than their horizontal dimension, an antenna array and a fault source form a horizontal triangle to locate DC arc faults. Then, signal pulses are distinguished and extracted using a cross-correlation technique. A neural network (NN) and received signal strength indicators (RSSIs) are utilized to estimate the arc distance. However, the data from the dual detection points are insufficient for meeting the demands of NN training and may lead to a large location error. Therefore, a data-augmented neural network (DANN) technique was introduced into the algorithm to enhance the accuracy and robustness of distance estimation. The experimental verification results show that the method can effectively locate faults under a variety of arcing conditions. Moreover, the proposed method has intrinsic immunity to normal current fluctuations (NCFs). It can be used in PV plants or smart buildings with solar cells to avoid nuisance trips and improve the effectiveness of fault protection.

Photovoltaic (PV) systems are increasingly used for renewable energy generation but remain vulnerable to series arc faults, which can cause serious safety risks, fire hazards, and system failures. Detecting these faults in DC circuits is challenging due to their subtle electrical signatures and the presence of noise from system and environmental sources. Traditional detection methods often fall short in terms of accuracy, prompting growing interest in artificial intelligence (AI)-based solutions. This review article provides a comprehensive analysis of AI-based techniques for series arc fault detection in PV systems, covering key aspects such as data preprocessing, feature extraction, model optimization, and hardware implementation. It presents a structured comparison of existing methods, including their strengths and limitations, through descriptive discussion and summary tables. The review also includes a simplified flowchart to illustrate the typical AI-based DC arc fault detection process. Key challenges are discussed, along with future directions such as hybrid models, transfer learning, and implementation in resource-constrained edge devices. This work aims to support continued research and development by helping researchers and engineers better understand the strengths and limitations of current approaches and identify practical ways to improve arc fault detection for enhanced PV system safety and reliability.

Residual current device is an important switch apparatus to ensure the safe and reliable operation of power distribution system. This paper proposes a novel fault leakage current detection method to eliminate the protection deadzone that is typically associated with classical leakage current detection method. The proposed method takes leakage current variation in addition of instantaneous RMS value of leakage current as the indicator of leakage fault. Three mathematical models of leakage current variation calculation are illustrated in details and multi-level leakage fault protection strategy is proposed. Simulation and experimental results are presented to validate the effectiveness of proposed leakage current detection method, and comparative evaluation of classical and proposed leakage current detection methods is conducted to verify the superiority of proposed method.