Research on Low Voltage Series Arc Fault Prediction Method Based on Multidimensional Time-Frequency Domain Characteristics

The load types in low-voltage distribution systems are diverse.Some loads have current signals that are similar to series fault arcs,making it difficult to effectively detect fault arcs during their occurrence and sustained combustion,which can easily lead to serious electrical fire accidents.To address this issue,this paper establishes a fault arc prototype experimental platform,selects multiple commonly used loads for fault arc experiments,and collects data in both normal and fault states.By analyzing waveform characteristics and selecting fault discrimination feature indicators,corresponding feature values are extracted for qualitative analysis to explore changes in timefrequency characteristics of current before and after faults.Multiple features are then selected to form a multidimensional feature vector space to effectively reduce arc misjudgments and construct a fault discrimination feature database.Based on this,a fault arc hazard prediction model is built using random forests.The model’s multiple hyperparameters are simultaneously optimized through grid search,aiming tominimize node information entropy and complete model training,thereby enhancing model robustness and generalization ability.Through experimental verification,the proposed method accurately predicts and classifies fault arcs of different load types,with an average accuracy at least 1%higher than that of the commonly used fault predictionmethods compared in the paper.

Study on the Detection Methods of DC Series Arc Fault

It is difficult to detect and extinguish direct current(DC)arc in power electronics systems,and the arc could easily lead to a fire and cause great damage to surrounding equipment.A DC arc generation simulation unit is established,in which DC series arcs are generated by dragging the moving electrode away from the fixed one with the help of the stepper motor.In addition,a ferrite rod antenna is used to receive the electromagnetic radiation signals induced by the arcs.Based on experiments using the unit,the general characteristics of DC arc,including the pulse characteristics of arc current and source output in corresponding time window,and the frequency-domain characteristics of arc current,are studied.With discussion on three detection methods,it is concluded that the variation of current and voltage of arc,the spectrum of the arc current during the discontinuous intervals and the radiating electromagnetic signal are all features that can be adopted for detecting DC series arc.Therefore,a synthetic judgment method is suggested for further study.

Arc Grounding Fault Identification Using Integrated Characteristics in the Power Grid

Arc grounding faults occur frequently in the power grid with small resistance grounding neutral points.The existing arc fault identification technology only uses the fault line signal characteristics to set the identification index,which leads to detection failure when the arc zero-off characteristic is short.To solve this problem,this paper presents an arc fault identification method by utilizing integrated signal characteristics of both the fault line and sound lines.Firstly,the waveform characteristics of the fault line and sound lines under an arc grounding fault are studied.After that,the convex hull,gradient product,and correlation coefficient index are used as the basic characteristic parameters to establish fault identification criteria.Then,the logistic regression algorithm is employed to deal with the reference samples,establish the machine discrimination model,and realize the discrimination of fault types.Finally,simulation test results and experimental results verify the accuracy of the proposed method.The comparison analysis shows that the proposed method has higher recognition accuracy,especially when the arc dissipation power is smaller than 2×10^(3) W,the zero-off period is not obvious.In conclusion,the proposed method expands the arc fault identification theory.

Spectrum-Sensing Method for Arc Fault Detection in Direct Current System with Lithium Batteries

We mainly study the detection of arc faults in the direct current (DC) system of lithium battery energystorage power station. Lithium battery DC systems are widely used, but traditional DC protection devices areunable to achieve adequate protection of equipment and circuits. We build an experimental platform based onan energy storage power station with lithium batteries. Then, the data collection of normal current and arc-faultcurrent is completed under multiple conditions, and the waveforms of obvious and weak signals as the arc occursare presented. We analyze the principles and application range of several common spectrum-sensing methods andstudy the feasibility of applying them to the arc detection field. Finally, the covariance absolute value detectionalgorithm is selected, and the average value of the current is filtered out to make the algorithm adapt to the arcdetection field. The result shows that the detection probability in 500 sets of experimental data has reached 98%.

Intermittent Arc Fault Detection Based on Machine Learning in Resonant Grounding Distribution Systems

In resonant grounding systems,most single-phaseto-ground faults evolve from IAFs(Intermittent Arc Faults).Earlier detection of IAFs can facilitate fault avoidance.This work proposes a novel method based on machine learning for detecting IAFs in three steps.First,the feature of zero-sequence current is automatically extracted and selected by a newlydesigned FINET(“For IAFs,Neuron Elaboration Net”),instead of traditional feature selection based on time-frequency decomposition.Moreover,data of the zero-sequence current divided by different time windows are successively input into the trained FINET.A proposed PSF(principal-subordinate factor)analyses the results obtained from FINET to improve anti-interference in the mentioned IAF detection algorithm.Experiments using PSCAD/EMTDC software simulation data show the proposed method is feasible and highly adaptable.In addition,the detection result of on-site recorded data demonstrates the effectiveness of the proposed method in practical resonant grounding systems.

Improved Control Strategy for Carbonized Cable Preparation in Low-voltage Arc Fault Test

Outdated testing methods hinder the success rate of carbonized cable preparation in low-voltage arc fault tests,leading to incomplete tests and high failure rates.To address this issue,we finely categorized the preparation results of carbonized cable specimens by analyzing the experimental phenomena during the carbonization process and assessing the impact of high-voltage energization time on the outcomes,presenting a process control strategy aimed at optimizing the preparation results of carbonized cable specimens.This method utilizes three periodic moving algorithms(root-mean-square,average,and shoulder percentage)to classify the cable specimens into four preparation categories:open-circuit carbonization,under-carbonization,short-circuit carbonization,and successful carbonization.The high-voltage energization time during carbonization or secondary carbonization was adjusted to optimize the preparation of the carbonized cables by considering different discrimination outcomes.Finally,the proposed method was tested on a purpose-built carbonized cable experimental platform,which confirmed its effectiveness in differentiating the preparation outcomes of the carbonized cable specimens and improving the success rate of the carbonized cable preparation.The proposed method has significant potential for application in low-voltage arc fault test systems.

A fast arc fault detection method for AC solid state power controllers in MEA

Arc fault detection is desperately required in Solid State Power Controllers（SSPC） in addition to their fundamental functions because arcs will provoke growing harm and threat to aircraft safety. Experimental study has been done to obtain the faulted current data. In order to improve the detection speed and accuracy, two fast arc fault detection methods have been proposed in this paper with the analysis of only half cycle data. Both Fast Fourier Transform（FFT） and Wavelet Packets Decomposition（WPD） have been adopted to distinguish arc fault currents from normal operation currents. Analysis results show that Alternating Current（AC） arcs can be effectively and accurately detected with the proposed half cycle data based methods. Moreover,experimental verification results have also been provided.

Novel detection method for DC series arc faults by using morphological filtering

It has been known that DC arc faults pose a hazard in power electronics systems. Due to the fault current is lower than normal current load, series arc is more difficult to be discovered than parallel arc. The traditional methods of DC series arc faults detection have some disadvantages, such as large computation, long delay and easier influence by interferences. In this paper, a novel method with multi-scale morphological filtering was utilized for the fault detection of DC series arc. Compared to the tradition method, the proposed method is more simple and efficient. It was verified that it was convenient for field application of on-line monitoring and diagnosis with a good ability to prevent misjudgments from the environment interference.

Analysis of Fault Arc in High-Speed Switch Applied in Hybrid Circuit Breaker

The behavior of fault arc in a high-speed switch （HSS） has been studied theoretically and experimentally. A simplified HSS setup is designed to support this work. A two-dimensional arc model is developed to analyze the characteristics of fault arc based on magnetic-hydrodynamic （MHD） theory. The advantage of such a model is that the thermal transfer coefficient can be determined by depending on the numerical method alone. The influence of net emission coefficients （NEC） radiation model and P1 model on fault arc is analyzed in detail. Results show that NEC model predicts more radiation energy and less pressure rise without the re-absorption effect considered. As a consequence, P1 model is more suitable to calculate the pressure rise caused by fault arc. Finally, the pressure rise during longer arcing time for different arc currents is predicted.

Logistic Regression Based Arc Fault Detection in Photovoltaic Systems Under Different Conditions

This paper investigates direct current(DC) arc fault detection in photovoltaic system. In order to avoid the risk of fire ignition caused by the arc fault in the photovoltaic power supply, it is urgent to detect the DC arc fault in the photovoltaic system. Once an arc fault is detected, the power supply should be cut off immediately. A lot of field experiments are carried out to obtain the data of arc fault current of the photovoltaic system under different current conditions. Cable length, arc gap, and the effects of different sensors are tested.These three conditions are the most significant features of this paper. Four characteristic variables from both the time domain and the frequency domain are extracted to identify the arc fault. Then the logistic regression method in the field of artificial intelligence and machine learning is originally used to analyze the experimental results of arc fault in the photovoltaic system. The function between the probability of the arc fault and the change of the characteristic variables is obtained. After validating 80 groups of experimental data under different conditions,the accuracy rate of the arc fault detection by this algorithm is proved to reach 100%.

Simulation of Fault Arc Based on Different Radiation Models in a Closed Tank

This paper focuses on the simulation of a fault arc in a closed tank based on the magneto-hydrodynamic （MHD） method, in which a comparative study of three radiation models, including net emission coefficients （NEC）, semi-empirical model based on NEC as well as the P1 model, is developed. The pressure rise calculated by the three radiation models are compared to the measured results. Particularly when the senti-empirical model is used, the effect of different boundary temperatures of the re-absorption layer in the semi-empirical model on pressure rise is concentrated on. The results show that the re-absorption effect in the low-temperature region affects radiation transfer of fault arcs evidently, and thus the internal pressure rise. Compared with the NEC model, P1 and the semi-empirical model with 0.7 〈 α 〈 0.83 are more suitable to calculate the pressure rise of the fault arc, where is an adjusted parameter involving the boundary temperature of the re-absorption region in the semi-empirical model.

Comparison between open phase fault of arc suppression coil and single phase to earth fault in coal mine distribution network

When, in a coal mine distribution network whose neutral point is grounded by an arc suppression coil (ASC), a fault occurs in the ASC, compensation cannot be properly realized. Furthermore, it can damage the safe and reliable run of the network. We first introduce a three-phase five-column arc suppression coil (TPFCASC) and discuss its autotracking compensation theory. Then we compare the single phase to ground fault of the coal mine distribution network with an open phase fault at the TPFCASC using the Thévenin theory, the symmetrical-component method and the complex sequence network respectively. The results show that, in both types of faults, zero-sequence voltage of the network will appear and the maximum magnitude of this zero-sequence voltage is different in both faults. Based on this situation, a protection for the open phase fault at the TPFCASC should be estab-lished.

电动汽车驱动电机与负载模拟系统建模及电弧故障仿真研究

电弧故障是引发电动汽车电气火灾的重要原因之一。电动汽车行驶工况复杂,电机及其驱动系统电压高、电流大,且其电弧故障随机性强、隐蔽性高导致真车故障实验难以开展,因此提出一种借助小功率电机与负载系统模拟故障的方法,以便快速开展大量实验,研究电弧故障特性。首先,在电动汽车负载转矩计算和等效缩放的基础上搭建了模拟实验平台,采集三相永磁同步电机线路串联电弧故障电流。其次,运用MATLAB软件构建空间矢量脉宽调制控制的电动汽车驱动电机与负载模拟系统,引入Cassie电弧故障模型并进行改进,对电动汽车三相永磁同步电机线路串联电弧故障展开仿真分析。最后,采用基于平肩宽度占比和小波包分解能量占比的特征提取方法,将仿真数据与实测数据进行比较并定量评价。结果表明,所提出的高斯电弧故障复合模型的相对平均误差最小,仅为7.6%。所构建的仿真系统可有效模拟实际线路的电弧故障,对电动汽车电气火灾的防控具有重要意义。

基于多卷积和结构搜索的电弧故障检测模型

串联型电弧故障主要由电路中电气接触点接触不良引起,是引发电动汽车电气火灾的主要原因之一,直接威胁着车内人员的生命安全。为对其进行研究,论文搭建了电动汽车直流串联型电弧故障实验平台,获取了系统处于不同工作状态下的电源端电压信号,分析了电弧故障对电源端电压的影响。在构建检测模型时,论文使用了卷积神经网络,引入轻量型的卷积操作并考虑了其在实际应用中的局限性。将常规卷积和轻量化卷积操作结合,构建了电弧故障检测的初步模型。接着以网络的规模和准确率为评估指标,通过具有精英保留策略的遗传算法对模型的外部结构和内部参数进行搜索。最终建立了适合电动汽车的电弧故障检测(arc fault detection,AFD)的检测模型AFDNet。该模型的检测准确率达到93.73%,在嵌入式设备Jetson Nano(JN)中的运行时间为10.82 ms。模型建立后,论文在网络的规模、准确性及实时性方面,将搜索算法的搜索结果与其他的网络结构进行比较,验证了搜索算法所得结果的合理性。并通过与其他检测方法对比,证明了电动汽车电弧故障检测模型AFDNet性能的优越。

基于改进EfficientNetV2算法的三相串联故障电弧检测

串联电弧故障在电气火灾中的成因中占据着重要的比例,在具有变频器的三相电机线路中,由于多种电弧电流复杂行为以及多种负载的存在,很难准确识别变频器后部线路中发生的串联电弧故障。为了解决这一问题,提出了一种改进的EfficientNetV2算法。搭建低电压三相电弧故障数据采集平台,采集了所需要的正常状态和故障状态的电流信号。为了充分利用机器视觉的优势,采用马尔可夫变迁场(MTF)将采集到的时域电流信号编码为图像。将MTF图像送入模型中进行训练和测试,该模型具有轻量级高效的通道注意力和双池化空间注意力,更加专注于电弧特征,提高网络性能。实验结果表明,该方法的准确率可达98.99%。

基于1DCNN和D-S多信息融合的光伏系统直流母线串联电弧故障检测

直流母线是光伏系统输出能源的主干道,由于长期曝晒、风化等作用,电缆、连接器等组件劣化,光伏系统直流母线中发生电弧的可能性急剧上升,极易引发火灾、触电等事故。在光伏系统中,串联电弧故障将使回路电流下降,传统的过流保护无法识别。因此,本文提出基于深度学习和证据理论(D-S)的方法来识别串联电弧故障,该方法基于并联电容器电流和电压信号,采用一维卷积神经网络(1DCNN)对检测数据进行电弧识别;在此基础上将基于单个传感数据的识别结果作为证据,运用D-S多信息合成法则计算得到信度分配,最后利用决策规则判断是否发生串联电弧故障。搭建多参数可调模型获取数据进行测试,结果表明:使用1DCNN识别方法,基于并联电容器电流和电压信号的串联电弧识别准确率分别为97.19%和94.98%,而基于1DCNN和D-S多信息融合的光伏系统直流串联电弧故障检测的识别准确率可提升至99%以上。

基于可控串联电抗的新能源送出线路故障性质识别方法

针对不带并联电抗器的新能源送出线,能否正确区分故障性质是确保新能源系统安全可靠运行的重中之重。为降低新能源送出线故障对电力设备和系统的冲击,提高重合闸成功率,减少新能源发电系统脱网的发生,提出基于可控串联电抗限流信息的新能源送出线路故障性质识别方法。在送出线路上设计串联可控限流电抗器,在线路故障相跳闸时基于等值阻抗的预判投入可控电抗,在串联电抗限流特性下分析瞬时性故障和永久性故障时的线路电流差异,建立电流突变量比较判据,并能够确定故障电弧熄灭时刻,进而构建新能源场站送出线路基于故障性质识别的自适应重合闸策略。仿真结果表明,所提方法能快速、可靠地识别结果,应用在新能源场站送出线路工程上能有效提高线路的重合率。

典型神经网络串联型电弧故障检测及选线方法研究

为解决传统特征分析加机器学习的电弧故障检测方法的准确率和实时性受到特征参数选取的主观性及特征分析过程的影响问题,搭建了三相多负载并联的串联型电弧故障实验系统,对不同支路、不同相发生电弧故障时的干路电流信号时间序列进行分析。将电流信号进行分类、分段、标准化处理并作为检测模型样本;对深度卷积神经网络模型、长短期记忆网络模型、普通神经网络模型进行架构及训练;通过差分处理对网络模型在线分类结果进行优化分析;以准确度和损失函数值、在线测试速度、优化后多分类识别准确率为评价指标,对比分析了3种模型故障检测及选线效果。研究结果表明:基于深度卷积神经网络的串联型电弧故障检测及选线模型对电机类负载故障检测及选线准确率可达96.77%,对变频器类负载故障检测及选线准确率可到98%,准确率高于近几年其他三相回路电弧故障检测模型。

基于全息希尔伯特频谱分析的交流串联电弧故障电流特征研究

针对交流串联电弧故障信号特征提取和分析的难题,提出采用循环掩码经验模态分解(itEMD)进行信号分解以获取多层内涵模态分量(IMF),结合全息希尔伯特变换(HHT)展现高维幅度调制频率调制(AM-FM)谱图的分析方法,从而获取高区分度的电弧故障信号特征谱图。采用基于GB/T 31143—2014的电弧故障实验平台,采集7种屏蔽负载模式下电弧故障电流信号,分析研究不同环境下电弧故障特征。所提方法对于电弧故障检测技术的开发具有一定的理论指导意义。