WO2021134737A1 - 一种基于矢量分析计算的电弧串扰信号识别方法 - Google Patents

一种基于矢量分析计算的电弧串扰信号识别方法 Download PDF

Info

Publication number
WO2021134737A1
WO2021134737A1 PCT/CN2020/000333 CN2020000333W WO2021134737A1 WO 2021134737 A1 WO2021134737 A1 WO 2021134737A1 CN 2020000333 W CN2020000333 W CN 2020000333W WO 2021134737 A1 WO2021134737 A1 WO 2021134737A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
crosstalk
arc
voltage
sampling
Prior art date
Application number
PCT/CN2020/000333
Other languages
English (en)
French (fr)
Inventor
马越
王建华
刘振
江泽安
王华荣
Original Assignee
青岛鼎信通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 青岛鼎信通讯股份有限公司 filed Critical 青岛鼎信通讯股份有限公司
Priority to US17/784,601 priority Critical patent/US11733286B2/en
Priority to EP20909189.1A priority patent/EP4063876B1/en
Priority to CA3162848A priority patent/CA3162848C/en
Publication of WO2021134737A1 publication Critical patent/WO2021134737A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/08—Locating faults in cables, transmission lines, or networks
    • G01R31/088—Aspects of digital computing
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
    • G01R31/14—Circuits therefor, e.g. for generating test voltages, sensing circuits

Definitions

  • the invention belongs to the field of fault arc detection, and mainly relates to a method for identifying crosstalk signals introduced by a bypass arc.
  • Arc faults are usually caused by the aging and damage of the insulation of lines and equipment, or poor electrical connections. When arc faults occur, the occurrence of arcs will release high temperatures, which is extremely easy to cause fires.
  • Arcs are mainly divided into two categories: parallel arcs and series arcs. Among them, when a parallel arc occurs, the effective value of the current in the circuit generally exceeds 75A. The existing circuit protection devices can identify over-current faults and disconnect the circuit; therefore, the series arc is the main cause of electrical fires.
  • the current signal in the line will be significantly distorted and rich in high frequency components.
  • the arc signal can be detected by the arc fault detection device in this circuit. Due to the high similarity between the two in nature, it is easy to be identified as a local fault arc, leading to the false tripping operation of the local fault arc detection device. Traditional arc fault detection equipment is often unable to distinguish between arc and crosstalk signals. Frequent false tripping also makes it difficult for arc detection equipment to be applied and promoted in a large area, laying a serious safety hazard for the vast majority of electrical equipment and power lines in the society.
  • the bypass crosstalk arc has essentially the same characteristics as the current arc, and it is difficult to distinguish between the arc and the crosstalk directly.
  • the present invention specifically proposes a new sampling circuit scheme. Through the Y-type circuit, the voltage signals at both ends of the two sampling resistors can be obtained, and the amplitude ratio and phase difference characteristics of the two can be extracted through the vector analysis calculation tool. Timely and accurately identify the arc and bypass crosstalk signals of this circuit, avoid the accidental tripping of the arc detection device, and lay a solid foundation for the application and promotion of the arc detection equipment.
  • the present invention proposes a new sampling circuit structure, sampling from the zero live line, outputting two sampling signals, and sending the digital signal Processing unit.
  • the vector analysis method the amplitude ratio and phase difference of the two-way signal are extracted, and sent to the neural network for classification decision in real time, to distinguish whether the arc signal originates from the current circuit or the bypass.
  • Figure 2 and Figure 3 are respectively equivalent circuit diagrams of the sampling circuit and load in the present invention when arc occurs locally and crosstalk occurs in bypass.
  • the signal between the zero and live wires is sampled.
  • the resistors R 1 and R 2 are divided into the current sampling resistor and the bypass sampling resistor.
  • Figure 2 shows the equivalent circuit analysis in the case of local arcing under ideal conditions.
  • the resistors R 1 , R 2 , R 3 , the capacitor C and the inductance L constitute the Y-type sampling circuit proposed by the present invention; bypass and local access
  • the load is equivalent to Z 1 and Z 2 respectively .
  • the arc signal is equivalent to a high-frequency small-signal current source.
  • the current reference direction of each branch is shown in the figure. According to Kirchhoff's law of voltage and current, the vector can be derived The formula is as follows:
  • Figure 3 shows the equivalent analysis of the crosstalk circuit in the case of bypass arcing under ideal conditions, and the vector is derived
  • the formula is as follows:
  • Figure 4 shows when an arc occurs at the bypass and this road position with The amplitude ratio varies with the inductance of the sampling circuit.
  • the arc amplitude ratio in this circuit is less than 1; the bypass arc amplitude ratio is greater than 1.
  • the present invention uses a multi-channel narrow-band filter circuit to extract multiple frequency components from two sampled signals for data analysis, according to the law of amplitude, the partial pressure of the inductive reactance (j ⁇ l) is proportional to the frequency, so Different frequencies are corresponding to different voltage divisions of different amplitudes, so that the amplitude volatility of the two resistance signals is different.
  • the present invention uses the amplitude signal to take the absolute value and then compare the two resistance signals. , Distinguish between arcing on this road and arcing by bypass crosstalk.
  • Figure 5 shows when an arc occurs in this circuit and the bypass position with The phase difference varies with the inductance of the sampling circuit.
  • the hindering effect of inductive reactance (j ⁇ l) on current the higher the frequency, the more obvious the hindering effect.
  • the phase difference of the arc in this circuit is positive and distributed between 90° and 180°; when the arc occurs in the bypass, the phase difference is negative and distributed between -90° and 0° .
  • the arc crosstalk signal identification method based on vector analysis proposed by the present invention extracts the characteristic difference of the crosstalk signal from two angles of signal amplitude and phase, and the judgment result is more reliable.
  • the present invention adopts a data time-sharing processing method to segment the two signals in time, and extract the characteristic value every 20us segment.
  • the effective value of the dual-channel signal sequence is extracted to characterize the amplitude characteristics of the signal during the period, and the result is very representative.
  • the invention extracts the phase characteristics of the signal at multiple frequency points on the basis of the traditional digital phase-locked amplification.
  • This method uses the principle that the noise and the reference signal are not correlated with each other, which can greatly suppress the noise and accurately extract the amplitude and phase information of the weak signal.
  • the phase characteristics and phase fluctuation characteristics of multiple frequency points are integrated to enhance the anti-interference ability of the algorithm and make the extracted phase characteristics more reliable.
  • the system counts the characteristic values of the amplitude ratio and phase difference of each channel within 20ms.
  • Fig. 1 is a flow chart of the crosstalk feature extraction system used in the present invention.
  • Figure 2 shows the equivalent circuit analysis of the arc in this way under ideal conditions.
  • Figure 3 shows the equivalent circuit analysis of the bypass arc under ideal conditions.
  • Figure 4 is when an arc occurs at the bypass and this road position with The amplitude ratio varies with the inductance of the sampling circuit.
  • Figure 5 is when arcing in this circuit and bypass position with The phase difference varies with the inductance value of the sampling circuit.
  • Fig. 6 is a block diagram of the amplitude ratio calculation process used in the present invention.
  • Fig. 7 is a block diagram of the phase difference calculation process used in the present invention.
  • Fig. 8 is a block diagram of the calculation flow of amplitude volatility used in the present invention.
  • Fig. 9 is a block diagram of the phase volatility calculation process used in the present invention.
  • the circuit sampling circuit includes: a first resistor R 1 , a second resistor R 2 , a third resistor R 3 , an inductance L and a capacitor C.
  • the inductance C is connected in series between the first measurement point 1 and the second measurement point 2 in the live wire
  • the third resistor R 3 is connected in parallel with the inductance C
  • the first resistor R 1 is connected in series between the first measurement point 1 and the third measurement point 1
  • the second resistance R 2 is connected in series between the second measuring point 2 and the third measuring point 3
  • the third measuring point 3 is connected to the neutral line through the capacitor C, wherein the A load is connected between the live wire and the neutral wire; when sampling, the inductance voltage signal between the first measurement point 1 and the second measurement point 2 and the first measurement point between the first measurement point 1 and the third measurement point 3 are collected respectively.
  • a resistance voltage signal is used to analyze whether the sampled circuit has a fault arc.
  • the system is mainly based on the processing of hardware digital signal processing system. It includes the following steps:
  • Step 1 Build the Y-type sampling circuit proposed in the present invention, perform continuous AD sampling on the signals at both ends of the resistor R 1 and the inductor L, with a sampling rate of 200 MHz, to obtain digital signals y R1 (n) and y L (n), and send them Into the hardware digital signal processing system, real-time vector analysis is performed to calculate the amplitude ratio, amplitude volatility, phase difference, and phase volatility characteristic quantities.
  • Step 2 The signals y R1 (n) and y L (n) are respectively filtered by band-pass digital filters.
  • the filter order can be designed to 64 steps, and the pass bands are 5MHz ⁇ 10MHz, 15MHz ⁇ 20MHz, 25MHz ⁇ 30MHz, 35MHz ⁇ 40MHz, 45MHz ⁇ 50MHz.
  • the unit impulse response of the digital filter is h(n), and the filtered signal:
  • the adaptive gain can be adjusted according to the signal amplitude to amplify the weak arc signal, while preventing the digital signal from overflowing, and ensuring the reliability of arc crosstalk feature extraction.
  • Step 3 Time-sharing processing of the filtered two-way data, performing a vector analysis on the signal within every 20 us, and respectively calculating the amplitude ratio feature quantity and the phase difference feature quantity proposed by the present invention.
  • the method of calculating the characteristic quantity of amplitude ratio is to obtain the ratio of the effective value of the signal amplitude at both ends of the resistance R 1 and the resistance R 2 in each period of time. As shown in the system processing flow chart in Figure 6, suppose the signal sequences at both ends of the resistor R 1 and the inductor L are respectively
  • the resistance R 1 , the resistance R 2 and the inductance L form a closed-loop triangle.
  • the signal sequence of the resistance R 2 can be calculated as
  • y R2_FIR ⁇ b 1 -a 1 , b 2 -a 2 ,..., b N -a N ⁇
  • the method of calculating the characteristic quantity of amplitude volatility difference is to obtain the proportion of the absolute value of the signal amplitude difference between the two ends of the resistance R 1 and the resistance R 2 in each period of time. As shown in process flow chart, let the resistance R 1 and an inductor L of FIG. 8 at both ends of the signal sequence system are
  • the resistance R 1 , the resistance R 2 and the inductance L form a closed-loop triangle.
  • the signal sequence of the resistance R 2 can be calculated as
  • the calculation method of the phase volatility characteristic quantity is to obtain the difference of the signal phase between the resistance R and the inductance L in each period of time. As shown in the system processing flow chart in Figure 9, suppose the signal sequences at both ends of the resistor R and the inductor L are respectively
  • the phase difference characteristic quantity calculation method is to obtain the difference of the signal phase between the resistance R and the inductance L in each period of time. As shown in the system processing flowchart in Figure 7, suppose the signal sequences at both ends of the resistance R and the inductance L are respectively
  • Step 4 The hardware digital signal processing unit performs vector analysis on the two-channel sampled signal, and outputs the amplitude ratio, amplitude volatility characteristic value and phase difference, phase volatility characteristic value in real time, and sends them to the MCU system for statistical processing.
  • Step 5 The system queries the state of the hardware zero-crossing detection circuit, and every time the zero-crossing signal arrives, the amplitude ratio, amplitude volatility, phase difference, and phase volatility are counted, and the eigenvalues in the half-wave are spliced. Become a 10*500-dimensional feature matrix, send it to the neural network for calculation, and give the discrimination result.
  • the arc crosstalk signal identification method proposed in the present invention performs vector analysis on the high-frequency components of the arc signal from the angles of amplitude and phase, and performs vector analysis on the arc.
  • the high-frequency characteristics of the signal are processed in a targeted manner, which can accurately identify the arc crosstalk signal, and has strong anti-interference ability in various complex power environments, ensuring stable and reliable discrimination results.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Measurement Of Resistance Or Impedance (AREA)
  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)

Abstract

一种基于矢量分析计算的电弧串扰信号识别方法。方法提出一种新的取样电路方式，对零火线上电流信号取样，并通过双路ADC，将信号转换为两路采样率200MHz的数字信号，送入硬件数字信号处理单元。选取5个通频段，分别对两路信号进行带通滤波，滤除通频带外的干扰信号。对滤波后信号进行分时处理，进行矢量分析，提取两路电阻端电压的幅值比及波动特性，以及旁路电阻与电感端电压信号的相位差，作为串扰特征量，系统根据过零信号对硬件处理模块提取的特征量进行分段，送入神经网络进行分类判决，即可对本路电弧信号与旁路串扰信号进行有效区分。

Description

一种基于矢量分析计算的电弧串扰信号识别方法
本申请要求于2020年01月02日提交中国专利局、申请号为202010000311.3、发明名称为“一种基于矢量分析计算的电弧串扰信号识别算法”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明属于故障电弧检测领域，主要涉及一种旁路电弧引入的串扰信号的识别方法。
背景技术
电气火灾在当今社会中的火灾事故中占有很高的比例，而电弧故障是导致电气火灾的重要原因之一。故障电弧通常是由于线路、设备绝缘老化和破损，或不良的电气连接造成的，当发生故障电弧时，电弧的发生会释放高温，极其容易引起火灾。电弧主要分为并联电弧与串联电弧两大类。其中，并联电弧发生时电路中电流有效值一般超过75A，目前已有的电路保护装置均能够识别过流故障，断开电路；因此，串联电弧才是引起电气火灾的主要原因。发生串联电弧时，由于电路其他负载的影响，电路中电流有效值与正常情况下的有效值区别不大，此时，传统的电路保护装置不能有效的发现电路中的电弧故障，更易引发火灾事故的发生。
当发生故障电弧时，线路中的电流信号会发生较为明显的畸变，富含高频成分。当旁路发生故障电弧时，电弧信号可被本路的故障电弧检测装置检测。由于二者本质上存在很高的相似性，因此很容易被识别为本地的故障电弧，导致本地故障电弧检测装置的误脱扣操作。传统故障电弧检测设备往往无法区分电弧与串扰信号，频繁的误脱扣动作也使得电弧检测设备难以大面积应用推广，为社会上绝大多数电器设备及电力线路埋下严重的安全隐患。
传统的电弧检测设备受电子技术发展水平限制，对电弧信号的检测以及串扰信号的识别手段比较单一，仅通过低频电流波形考察电弧特征，且受电器负载干扰严重。随着IC技术发展脚步的加快，在各类设备上应用也越来越广泛。IC技术水平的提高，为各类复杂功能算法的实现提供了 技术支撑，基于IC技术基础上的电路系统，普遍具有功耗低，量产成本低以及抗干扰能力强等优点，因此本发明提出的串扰信号识别方法得以实现及应用。
旁路串扰电弧与本路电弧本质上具有一致的特征，直接进行电弧与串扰的区分难度较大。本发明中针对性的提出了一种新的取样电路方案，通过Y型电路，可以获取两个取样电阻两端电压信号，通过矢量分析计算工具，提取二者的幅值比与相位差特征，及时准确的识别本路电弧与旁路串扰信号，避免电弧检测装置误脱扣，为电弧检测设备的应用推广打下坚实的基础。
发明内容
针对传统电弧检测方法不能准确识别旁路串扰信号，从而导致设备频繁地误脱扣动作，本发明提出一种新的采样电路结构，从零火线上采样，输出两路采样信号，送入数字信号处理单元。通过矢量分析方法，提取出双路信号的幅值比和相位差，实时送入神经网络分类决策，区分电弧信号来源于本路还是旁路。
本发明的原理是，图2、图3分别为在本地发生电弧和旁路发生串扰时，本发明中取样电路及负载的等效电路图。在Y型取样电路的基础上对零火线间信号取样，根据取样位置，将电阻R 1与R 2区分为本路取样电阻与旁路取样电阻。
图2为理想情况下，本地打电弧情况下的等效电路分析，电阻R 1，R 2，R 3，电容C与电感L构成本发明提出的Y型采样电路；旁路和本路接入负载分别等效为Z 1，Z 2。电弧信号等效为一个高频小信号电流源，各支路电流参考方向如图中标注所示，根据基尔霍夫电压和电流定律，可推导向量
Figure PCTCN2020000333-appb-000001
公式如下：
Figure PCTCN2020000333-appb-000002
Figure PCTCN2020000333-appb-000003
Figure PCTCN2020000333-appb-000004
其中，
Figure PCTCN2020000333-appb-000005
Figure PCTCN2020000333-appb-000006
同理，图3为理想情况下，旁路打电弧情况下的串扰电路等效分析，推导向量
Figure PCTCN2020000333-appb-000007
公式如下：
Figure PCTCN2020000333-appb-000008
Figure PCTCN2020000333-appb-000009
Figure PCTCN2020000333-appb-000010
其中，
Figure PCTCN2020000333-appb-000011
Figure PCTCN2020000333-appb-000012
设定电路参数如下：R 1＝10Ω，R 2＝10Ω，R 3＝20Ω，电容C＝20nf，旁路和本路分别串联接入阻性，感性，容性负载，设定
Figure PCTCN2020000333-appb-000013
在取样电路参数选定情况下，观察通频段内信号。以上参数均为本发明一个实施例设定参数，本发明不局限于以上电路参数。在理想情况下，以不同频率的信号进行分析：
图4给出在旁路和本路位置发生电弧时
Figure PCTCN2020000333-appb-000014
和
Figure PCTCN2020000333-appb-000015
的幅值比随取样电路电感变化曲线。按照以上实施例的电路参数，本路发生电弧幅值比小于1；旁路发生电弧幅值比大于1。
由本路打弧和旁路串扰电弧的等效电路，
Figure PCTCN2020000333-appb-000016
为对应采样位置的电压值，可以得到：
本路打弧时：
Figure PCTCN2020000333-appb-000017
旁路打弧时：
Figure PCTCN2020000333-appb-000018
由上述两个公式，可以看出本路打弧时
Figure PCTCN2020000333-appb-000019
旁路串扰打弧时，
Figure PCTCN2020000333-appb-000020
两种情况下的幅值差异为电感L的分压
Figure PCTCN2020000333-appb-000021
由于本发明采用多通道窄带滤波电路对两路采样信号进行多个频率成分的提取，用于数据分析，故根据幅值的规律，感抗(jωl)的分压和频率成正比的关系，所以不同频率是对应的不同幅值大小的分压，从而引起两路电阻信号的幅值波动性大小是不同的，本发明采用幅值信号的进行取绝对值的再对比两路电阻信号大小的方式，进行本路打弧和旁路串扰打弧的区分。
图5给出在本路和旁路位置发生电弧时
Figure PCTCN2020000333-appb-000022
和
Figure PCTCN2020000333-appb-000023
的相位差异随取样电路电感变化曲线。感抗(jωl)的对电流的阻碍作用，频率越高，阻碍作用越明显。。按照以上实施例的电路参数，本路发生电弧相位差为正值，分布于90°到180°之间；旁路发生电弧时，相位差为负值，分布于-90°到0°之间。
本发明提出的基于矢量分析的电弧串扰信号识别方法，从信号幅值和相位两个角度，提取串扰信号特征差异，判决结果更加可靠。
本发明采用数据分时处理方法，对两路信号按时间分段，每20us一段进行特征值提取。提取双路信号序列的有效值表征该段时间内信号的幅度特征，结果具有很强的代表性。本发明在传统数字锁相放大基础上，提取信号在多个频点上的相位特征。该方法利用噪声与参考信号互不相关的原理，能够极大地抑制噪声从而准确提取出微弱信号的幅值和相位信息。同时综合多个频点的相位特征以及相位波动性特征，增强算法的抗干扰能力，使提取的相位特征更加可靠。系统根据过零检测电路输出的过零信号，统计各通道20ms内的幅值比、与相位差的特征值。
说明书附图
下面结合附图对本发明作进一步说明：
图1是本发明中使用的串扰特征提取系统流程图。
图2为理想情况下，本路打电弧情况下的等效电路分析。
图3为理想情况下，旁路打电弧情况下的等效电路分析。
图4是在旁路和本路位置发生电弧时
Figure PCTCN2020000333-appb-000024
和
Figure PCTCN2020000333-appb-000025
的幅值比随取样电路电感变化曲线。
图5是在本路和旁路位置电弧时
Figure PCTCN2020000333-appb-000026
和
Figure PCTCN2020000333-appb-000027
的相位差异随取样电路电感值变化曲线。
图6是本发明中使用的幅值比计算流程框图。
图7是本发明中使用的相位差计算流程框图。
图8是本发明中使用的幅值波动性大小计算流程框图。
图9是本发明中使用的相位波动性计算流程框图。
具体实施方式
下面结合图1至图9对本发明所提供的电弧串扰识别方法进行说明。
本发明基于矢量分析计算的串扰信号特征识别流程如图1所示，采用电路取样电路包括：第一电阻R 1、第二电阻R 2、第三电阻R 3、电感L以及电容C，所述电感C串联于火线中的第一测量点1与第二测量点2之间， 所述第三电阻R 3与所述电感C并联，第一电阻R 1串联于第一测量点1与第三测量点3之间，所述第二电阻R 2串联于第二测量点2与第三测量点3之间，所述第三测量点3通过所述电容C与零线连接，其中，所述火线与所述零线之间连接有负载；取样时，分别采集第一测量点1与第二测量点2之间的电感电压信号以及第一测量点1与第三测量点3之间的第一电阻电压信号，用以分析被取样线路是否发生故障电弧。系统主要基于硬件数字信号处理系统处理实现。包括以下步骤：
步骤1：搭建本发明提出的Y型取样电路，对电阻R 1和电感L两端信号进行连续的AD采样，采样率达200MHz，得到数字信号y R1(n)和y L(n)，送入硬件数字信号处理系统，实时地进行矢量分析，计算幅值比、幅值波动性大小和相位差、相位波动性的特征量。
步骤2：信号y R1(n)和y L(n)分别经带通数字滤波器滤波，滤波器阶数可设计为64阶，通频段分别为5MHz～10MHz，15MHz～20MHz，25MHz～30MHz，35MHz～40MHz，45MHz～50MHz。数字滤波器单位冲激响应分别为h(n)，滤波后信号：
Figure PCTCN2020000333-appb-000028
滤波后可根据信号幅值大小进行自适应增益调节，放大微弱电弧信号，同时防止数字信号溢出，保证电弧串扰特征提取的可靠性。
步骤3：对滤波后的双路数据分时处理，对每20us内信号进行一次矢量分析，分别计算本发明所提出的幅值比特征量和相位差特征量。
幅值比特征量计算方法是求取每段时间内电阻R 1，电阻R 2两端信号幅值有效值的比值。如图6系统处理流程图所示，设电阻R 1和电感L两端信号序列分别为
Figure PCTCN2020000333-appb-000029
根据Y型电路，电阻R 1，电阻R 2和电感L构成闭环三角形，根据基尔霍夫电压定律可计算电阻R 2信号序列为
y R2_FIR＝{b 1-a 1，b 2-a 2，...，b N-a N}
(1)分别对电阻R 1，电阻R 2信号序列求平方得：
Figure PCTCN2020000333-appb-000030
(2)分别对双路信号序列的平方求平均值，得：
Figure PCTCN2020000333-appb-000031
(3)分别对双路信号的平方和的平均值开方，得：
Figure PCTCN2020000333-appb-000032
(4)求电阻R 1，电阻R 2信号的幅值比，得：
Figure PCTCN2020000333-appb-000033
分别按如上方法，计算每个通道上信号的幅值比特征量Amp_R 1，Amp_R 2，Amp_R 3，Amp_R 4，Amp_R 5。
幅值波动性大小差异特征量计算方法是求取每段时间内电阻R 1，电阻R 2两端信号幅值差分绝对值的大小占比。如 图8系统处理流程图所示，设电阻R 1和电感L两端信号序列分别为
Figure PCTCN2020000333-appb-000034
根据Y型电路，电阻R 1，电阻R 2和电感L构成闭环三角形，根据基尔霍夫电压定律可计算电阻R 2信号序列为
{(b 1-a 1)，(b 2-a 2)，...，(b N-a N)}
(1)对两路电阻信号R 1和R 2求取差分的绝对值：
Figure PCTCN2020000333-appb-000035
(2)对每个时间段内的两路电阻信号R 1和R 2的幅值大小做统计，这里取R 1小于R 2做为特征量输出，对20ms的数据输出一组特征量送入神经网络。
相位波动性特征量计算方法是求取每段时间内电阻R，电感L两端信号相位的差异。如图9系统处理流程图所示，设电阻R和电感L两端信号序列分别为
Figure PCTCN2020000333-appb-000036
(1)对两路电阻信号R和L求取差分序列：
Figure PCTCN2020000333-appb-000037
(2)对每个时间段内的电阻和电感上信号的波动性做统计，这里取R 1*L＜0做为特征量输出，对20ms的数据输出一组特征量送入神经网络。
相位差特征量计算方法是求取每段时间内电阻R，电感L两端信号相位的差值。如图7系统处理流程图所示，设电阻R和电感L两端信号序列分别为
Figure PCTCN2020000333-appb-000038
(1)选定滤波器通频段内频率f 0＝30MHz，设为参考频率，计算该频点上标准复信号为序列：
Figure PCTCN2020000333-appb-000039
(2)电阻R 1和电感L两端信号序列分别与标准复信号序列做乘法运算，提高待测频率上信号的信噪比，极大地改善对微弱信号提取相位特征的效果。计算结果如下：
Figure PCTCN2020000333-appb-000040
(3)对上一步得到的乘积序列做积分运算，根据离散信号积分转换为求和原理，得到积分结果：
Figure PCTCN2020000333-appb-000041
(4)根据二者积分结果，分别计算该段时间内双路信号序列在参考频率点上的初始相位为：
Figure PCTCN2020000333-appb-000042
(5)计算二者的差值，得到一次矢量分析中双路信号在频率f 0＝30MHz的相位差特征量
Figure PCTCN2020000333-appb-000043
(6)在每个通频段内选取各自的参考频率f n＝f 1，f 2，f 3，f 4，f 5，重复以上步骤(1)～步骤(5)，计算每个通道信号在该段时间内电阻R 1，电感L两端信号在的相位差异
Figure PCTCN2020000333-appb-000044
步骤4：硬件数字信号处理单元对双路采样信号进行矢量分析，实时地输出幅值比、幅值波动性大小特征值和相位差、相位波动性特征值，送入MCU系统统计处理。
步骤5：系统查询硬件过零检测电路的状态，每次过零信号到来，对幅值比、幅值波动性大小和相位差、相位波动性的特征量进行统计，将半波内特征值拼接成10*500维的特征矩阵，送入神经网络计算，给出判别结果。
与传统设备单一检测故障电弧信号，无法区分本路电弧与旁路电弧相比，本发明提出的电弧串扰信号识别方法分别从幅度和相位的角度对电弧信号的高频成分进行矢量分析，对电弧信号的高频特性进行针对性的处理，能准确识别电弧串扰信号，在各种复杂的用电环境中，具有较强的抗干扰能力，保证判别结果稳定可靠。
以上所述只是本发明的优选实施方式，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以作出若干改进和变化。凡在本发明的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (17)

  1. 一种基于矢量分析计算的电弧串扰信号识别方法，其特征在于：搭建特定的采样电路，对零火线间电流进行连续取样，输出两路取样信号；分别经ADC完成模数转换，送入数字信号处理单元；信号经5个通频段的带通滤波处理，分别进行矢量分析计算，提取幅度比、幅值波动性大小与相位差、相位波动性特征值，系统根据硬件过零检测电路输出的过零信号，统计每半波内5个通道的幅值比、幅值波动性大小与相位差、相位波动性的特征向量，送入神经网络进行分类判决。
  2. 根据权利要求1所述的旁路串扰信号的识别方法，其特征在于：
    所述识别方法提出一种专用取样电路，用来区分本路故障电弧和旁路串扰信号；电路结构为：电阻R 1，R 2与电容C构成字母‘Y’形状，连接火线和零线，以下简称Y型电路；电阻R 3与电感L并联，串联在火线上，同时并联在R 1与R 2之间；
    分别对R 1与R 2两端的信号取样，进行矢量分析，可从幅值与相位两个方面分别分析本路电弧与旁路串扰的差别。
  3. 根据权利要求1所述的旁路串扰信号的识别方法，其特征在于：
    所述识别方法从火线上取样下来的双路信号，分别经200MHz的双路ADC转换成数字信号，经5个带通数字滤波器滤波，通频段分别为5MHz～10MHz，15MHz～20MHz，25MHz～30MHz，35MHz～40MHz，45MHz～50MHz；滤波后数据分段处理，每20us内数据为1段，对双路信号进行矢量分析。
  4. 根据权利要求1所述的旁路串扰信号的识别方法，其特征在于：
    所述识别方法基于对取样电路的矢量分析，每20us时间，分别对R 1与R 2两端电压幅值求平方，再对每段数据求平均，再开方，得到R 1与R 2两端电压取样信号
    Figure PCTCN2020000333-appb-100001
    与
    Figure PCTCN2020000333-appb-100002
    的有效值(即均方根值)U 1与U 2，计算二者比值；按照本发明的一种实施例的电路参数计算得，幅值比如果大于1，则为旁路串扰信号，小于1，则为本路故障电弧；具体阈值的设定应根据电路中各元器件参数，用电设备不同及温湿度等实验环境不同而灵活处理。
  5. 根据权利要求1所述的旁路串扰信号检测的识别方法，其特征在于：
    当发生打弧信号时，在电阻R 1通道和电阻R 2通道上采样的电压信号幅值，由于电感L的存在，感抗(jωl)的分压和频率成正相关的关系，所以不同频率是对应的不同幅值大小的分压，从而引起两路电阻信号的幅值波动差异是不同的，且串扰信号通过时R 1通道上的电压幅值波动小于R 2通道上的电压幅值波动，本路打弧信号通过时R 1通道上的电压幅值波动大于R 2通道上的电压幅值波动。
  6. 根据权利要求2所述的旁路故障电弧的串扰信号检测的采样电路，其特征在于：
    当产生打弧的信号时，根据矢量分析，感抗(jωl)的对电流的阻碍作用，频率越高，阻碍作用越明显，根据电感对电流信号会产生滞后作用，在电阻R 1两侧采样到的电压信号和电感L路两侧采样到的电压信号之间存在相位波动性差异，电阻R 1两侧的电压信号滞后于电感L两侧的电压信号，本路打弧电阻R 1两侧的电压信号超前于电感L两侧的电压信号。
  7. 根据权利要求1所述的旁路串扰信号的识别方法，其特征在于：
    所述识别方法基于对取样电路的矢量分析，每20us时间，对双路ADC采样的信号通过数字锁相放大求相位的方法，计算一次R 1与L两端电压取样信号
    Figure PCTCN2020000333-appb-100003
    与
    Figure PCTCN2020000333-appb-100004
    的相位
    Figure PCTCN2020000333-appb-100005
    与
    Figure PCTCN2020000333-appb-100006
    计算二者的差值；按照本发明的一种实施例的电路参数计算得，相位差小于0°且分布于-90°到0°，则为旁路串扰信号；大于0°且分布于0°到180°，则为本路电弧信号；具体阈值的设定应根据电路中各元器件参数，用电设备不同及温湿度等实验环境不同而灵活处理；
    其中，数字锁相放大求相位差原理为：
    设取样信号分别为y 1(n)与y 2(n)，为方便获取信号在某一频率点信号的相位，设定该频率为参考频率ω 0，用该频率上的标准复信号
    Figure PCTCN2020000333-appb-100007
    与滤波后信号做乘法运算，然后积分，由积分结果可以得到该段信号的初始相位
    Figure PCTCN2020000333-appb-100008
    与θ 0，考虑进行故障电弧信号的带宽，非单频信号，因此分别在5个通频带内选取参考频率，分析信号在参考频点上的相位特征，即设定参考频率ω n(n＝1，2，...，5)，使用不同频点的标准复信号
    Figure PCTCN2020000333-appb-100009
    利用数字锁相原理，计算电阻R 1两端信号在ω n(n＝1，2，...，5)频率上的相位
    Figure PCTCN2020000333-appb-100010
    计算电感L两端电压信号在ω n(n＝1，2，...，5)频率上的相位
    Figure PCTCN2020000333-appb-100011
    得到一次矢量分析的相位差特征值
    Figure PCTCN2020000333-appb-100012
    本发明中对信号进行矢量分析中计算双路信号幅值比和相位差的步骤，可按上述方法实现，但不局限于上述方法。
  8. 根据权利要求1所述的旁路串扰信号的识别方法，其特征在于：
    所述识别方法硬件数字系统提取幅值比和相位差特征量，根据过硬检测电路输出的过零信号，按半波时间分段统计5个通道计算得到的幅度比、幅值波动性大小与相位差、相位波动性的特征量，拼接为一个10*500的特征矩阵，送入神经网络，进行串扰信号识别。
  9. 根据权利要求1或8所述的一种基于矢量分析的电弧串扰信号识别方法，其特征在于所述数字信号处理单元的实现可以是分立器件，模拟集成电路，数字逻辑电路，单片机，微处理器，可编程逻辑器件，数字信号处理器或者专用集成电路ASIC中的任一种。
  10. 一种取样电路，其特征在于，包括：第一电阻R 1、第二电阻R 2、第三电阻R 3、电感L以及电容C，所述电感C串联于火线中的第一测量点与第二测量点之间，所述第三电阻R 3与所述电感C并联，第一电阻R 1串联 于第一测量点与第三测量点之间，所述第二电阻R 2串联于第二测量点与第三测量点之间，所述第三测量点通过所述电容C与零线连接，其中，所述火线与所述零线之间连接有负载；取样时，分别采集第一测量点与第二测量点之间的电感电压信号以及第一测量点与第三测量点之间的第一电阻电压信号，用以分析被取样线路是否发生故障电弧。
  11. 一种基于矢量分析计算的电弧串扰信号识别方法，其特征在于，包括：
    采用权利要求1所述的取样电路进行取样，得到电感电压信号以及第一电阻电压信号；
    基于所述电感电压信号以及所述第一电阻电压信号，计算第二电阻两端的电压信号，将所述第一电阻电压信号记为第一取样信号，将第二电阻两端的电压信号记为第二取样信号，将所述电感电压信号记为第三取样信号；
    基于幅度比特征量、幅值波动性特征量、相位差特征量和/或相位波动性特征量识别被取样线路出现的信号是电弧信号还是串扰信号，其中，所述幅度比特征量为所述第一取样信号的电压有效值和所述第二取样信号的电压有效值的比值，所述幅值波动性特征量为所述第一取样信号的电压幅值波动性与所述第二取样信号的电压幅值波动性的大小关系，所述相位差特征向量为所述第一取样信号与所述第三取样信号的相位差，所述相位波动性特征量为所述第一取样信号与所述第三取样信号的相位波动性差异。
  12. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，基于幅度比特征量、幅值波动性特征量、相位差特征量和相位波动性特征量识别被取样线路出现的信号是电弧信号还是串扰信号，具体包括：
    基于幅度比特征量、幅值波动性特征量、相位差特征向量和相位波动性特征量，采用训练好的神经网络模型识别被取样线路出现的信号是电弧信号还是串扰信号。
  13. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，在识别被取样线路出现的信号是电弧信号还是串扰信号之前，还包括：
    对所述电感电压信号以及所述第一电阻电压信号进行模数转换；
    采用若干带通数字滤波器对模数转换后的电感电压信号以及第一电阻电压信号进行滤波；
    根据幅值大小对滤波后的电感电压信号以及第一电阻电压信号进行自适应增益调节，其中，电感电压信号以及第一电阻电压信号对应频段的增益倍数相同。
  14. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，基于幅度比特征量识别被取样线路出现的信号是电弧信号还是串扰信号，具体包括：
    当所述第一取样信号的电压有效值和所述第二取样信号的电压有效值的比值小于设定阈值时，确定被取样线路出现的信号为电弧信号，当所述第一取样信号电压的有效值和所述第二取样信号的电压有效值的比值大于设定阈值时，确定被取样线路出现的信号为串扰信号。
  15. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，基于幅值波动性特征量识别被取样线路出现的信号是电弧信号还是串扰信号，具体包括：
    当所述第一取样信号的电压幅值波动大于所述第二取样信号的电压幅值波动时，确定被取样线路出现的信号为电弧信号，当所述第一取样信号的电压幅值波动小于所述第二取样信号的电压幅值波动时，确定被取样线路出现的信号为串扰信号。
  16. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，基于相位差特征量识别被取样线路出现的信号是电弧信号还是串扰信号，具体包括：
    当所述第一取样信号和所述第三取样信号的相位差大于0°且分布于0°到180°时，确定被取样线路出现的信号为电弧信号，当所述第一取样信号和所述第三取样信号的相位差小于0°且分布于-90°到0°时，确定被取样线路出现的信号为串扰信号。
  17. 根据权利要求11所述的基于矢量分析计算的电弧串扰信号识别方法，其特征在于，基于相位波动性特征量识别被取样线路出现的信号是电弧信号还是串扰信号，具体包括：
    当所述第一取样信号的相位超前于所述第三取样信号的相位时，确定被取样线路出现的信号为电弧信号，当所述第一取样信号的相位滞后于所述第三取样信号的相位时，确定被取样线路出现的信号为串扰信号。
PCT/CN2020/000333 2020-01-02 2020-12-31 一种基于矢量分析计算的电弧串扰信号识别方法 WO2021134737A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
US17/784,601 US11733286B2 (en) 2020-01-02 2020-12-31 Vector analysis calculation-based arc crosstalk signal identification method
EP20909189.1A EP4063876B1 (en) 2020-01-02 2020-12-31 Vector analysis calculation-based arc crosstalk signal identification method
CA3162848A CA3162848C (en) 2020-01-02 2020-12-31 Vector analysis calculation-based arc crosstalk signal identification method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010000311.3 2020-01-02
CN202010000311 2020-01-02

Publications (1)

Publication Number Publication Date
WO2021134737A1 true WO2021134737A1 (zh) 2021-07-08

Family

ID=75413662

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/000333 WO2021134737A1 (zh) 2020-01-02 2020-12-31 一种基于矢量分析计算的电弧串扰信号识别方法

Country Status (5)

Country Link
US (1) US11733286B2 (zh)
EP (1) EP4063876B1 (zh)
CN (1) CN112666436B (zh)
CA (1) CA3162848C (zh)
WO (1) WO2021134737A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113960397B (zh) * 2021-10-25 2024-04-16 矽力杰半导体技术(杭州)有限公司 一种故障检测方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130038971A1 (en) * 2010-06-03 2013-02-14 Shakira Limited Arc fault detector for ac or dc installations
CN208953612U (zh) * 2018-09-03 2019-06-07 湖北窗口科技有限公司 一种基于电流互感器的电流采样电路及四接线柱智能插座
CN110007197A (zh) * 2019-05-15 2019-07-12 合肥工业大学 串联电弧故障检测系统及检测方法
CN110082635A (zh) * 2019-05-06 2019-08-02 江苏格澜得智能电气有限公司 一种可自学习的多功能故障电弧探测器
CN110542835A (zh) * 2018-05-29 2019-12-06 上海海拉电子有限公司 一种用于车辆电弧故障的检测方法、检测系统及试验系统

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6313641B1 (en) 1995-03-13 2001-11-06 Square D Company Method and system for detecting arcing faults and testing such system
US6798211B1 (en) * 1997-10-30 2004-09-28 Remote Monitoring Systems, Inc. Power line fault detector and analyzer
US5986860A (en) 1998-02-19 1999-11-16 Square D Company Zone arc fault detection
US7136265B2 (en) 2001-10-17 2006-11-14 Square D Company Load recognition and series arc detection using bandpass filter signatures
WO2004034070A1 (en) * 2002-10-10 2004-04-22 Hanyang Hak Won Co., Ltd. Hybrid type sensor for detecting high frequency partial discharge
CN100507583C (zh) * 2007-03-16 2009-07-01 长沙理工大学 配电网对地绝缘参数测量和控制方法
CN201408209Y (zh) * 2009-03-03 2010-02-17 厦门红相电力设备股份有限公司 一种能同时提供基频试验电压和高频局部放电信号输出的复用型检测阻抗
US8395391B2 (en) * 2009-08-19 2013-03-12 Astronics Advanced Electronic Systems Corp. Method and apparatus for locating a parallel arc fault
WO2013098906A1 (ja) * 2011-12-26 2013-07-04 エヌ・ティ・ティ・データ先端技術株式会社 アーク放電阻止回路及びアーク放電阻止装置
US9634479B2 (en) * 2012-11-16 2017-04-25 Sensata Technologies, Inc. Noise propagation immunity of a multi-string arc fault detection device
US9726697B2 (en) * 2014-12-31 2017-08-08 Texas Instruments Incorporated Coupled inductor current sensing apparatus and systems
CN105282071B (zh) * 2015-11-06 2019-04-09 哈尔滨工业大学 加权分数傅里叶变换域扩频信号并行干扰消除差错修正方法
CN206876813U (zh) * 2017-06-02 2018-01-12 北京英瑞来科技有限公司 一种考虑雷电流隔离设备影响的故障电弧检测装置
CN107505517A (zh) * 2017-07-19 2017-12-22 国网湖南省电力公司 一种避雷器直流续流试验装置及方法
CN208623691U (zh) * 2018-09-13 2019-03-19 杭州石匍丁科技有限公司 一种防干扰式路由器

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130038971A1 (en) * 2010-06-03 2013-02-14 Shakira Limited Arc fault detector for ac or dc installations
CN110542835A (zh) * 2018-05-29 2019-12-06 上海海拉电子有限公司 一种用于车辆电弧故障的检测方法、检测系统及试验系统
CN208953612U (zh) * 2018-09-03 2019-06-07 湖北窗口科技有限公司 一种基于电流互感器的电流采样电路及四接线柱智能插座
CN110082635A (zh) * 2019-05-06 2019-08-02 江苏格澜得智能电气有限公司 一种可自学习的多功能故障电弧探测器
CN110007197A (zh) * 2019-05-15 2019-07-12 合肥工业大学 串联电弧故障检测系统及检测方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHEN YONGZHI: "Study on detection technology for electrical fire of arc and development of Arc Fault Detection Device", SCIENCE & ENGINEERING (B), CHINA MASTER’S THESES FULL-TEXT DATABASE, 1 January 2017 (2017-01-01), XP055827838 *

Also Published As

Publication number Publication date
EP4063876A4 (en) 2023-01-18
CA3162848C (en) 2024-02-13
US20230022120A1 (en) 2023-01-26
EP4063876C0 (en) 2023-10-25
CA3162848A1 (en) 2021-07-08
CN112666436A (zh) 2021-04-16
US11733286B2 (en) 2023-08-22
EP4063876B1 (en) 2023-10-25
EP4063876A1 (en) 2022-09-28
CN112666436B (zh) 2022-06-21

Similar Documents

Publication Publication Date Title
CN108957244B (zh) 一种配网主站单相接地故障选线定位方法
CN105738764B (zh) 基于暂态信息全频带的配电网故障区段定位方法
CN109298287B (zh) 基于故障指示器数据的小电流接地配网故障类型识别方法
Jafarian et al. High-speed superimposed-based protection of series-compensated transmission lines
CN112485595B (zh) 一种配电网接地故障选线保护方法及装置
CN110058157B (zh) 基于对称分量法和多维指标融合的电机监测方法及系统
CN112698103B (zh) 一种提升高过渡电阻接地故障判断准确度的方法
WO2021134737A1 (zh) 一种基于矢量分析计算的电弧串扰信号识别方法
CN113300343B (zh) 一种基于余弦相似度的柔性直流电网故障线路识别方法
CN118226202B (zh) 一种行波定位型一二次融合开关及故障定位方法
CN110058156A (zh) 基于瞬时三相电压电流不平衡度电机在线监测方法及系统
Gafoor et al. Transmission line protection scheme using wavelet based alienation coefficients
CN107526006B (zh) 一种小电流系统单相接地故障定位方法
CN106026040A (zh) 一种防止采样异常大数的母线差动保护方法
CN112345890B (zh) 一种故障电弧及电流检测的复合检测电路装置
CN109507533B (zh) 一种高压直流输电线路的单端速动主保护方法
CN114720905B (zh) 一种灵活接地系统单相高阻接地故障识别、保护方法及装置
Rao et al. Pattern recognition approach for fault identification in power transmission lines
CN116125208A (zh) 基于数据采集设备的配电网单相接地故障定位方法
CN112485715B (zh) 一种基于电流零休暂态特征的可靠选线方法及装置
CN108872747A (zh) 基于相关系数的电涌保护器阻性电流提取装置及其方法
CN114609435A (zh) 一种电压暂降检测和分类辨识方法
CN204116427U (zh) 一种避雷器末端信号在线监测装置
CN112485713B (zh) 一种基于接地故障电流零休特性的选线装置及方法
CN118501620B (zh) 基于连续滤波与峭度变化率的配电网高阻故障识别方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20909189

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 3162848

Country of ref document: CA

ENP Entry into the national phase

Ref document number: 2020909189

Country of ref document: EP

Effective date: 20220622

NENP Non-entry into the national phase

Ref country code: DE