CN104502816A - LabVIEW-based cable partial discharge position identification method - Google Patents

Abstract

The invention discloses a LabVIEW-based cable partial discharge position identification method comprising the steps that S100: a cable partial discharge signal is inputted to hardware conditioning equipment for amplification and filtering processing; S200: the signal after amplification and filtering processing is inputted to signal acquisition equipment, and the cable partial discharge signal is acquired; S300: DC filtering, filtering and threshold judgment processing of a LabVIEW are performed on the output signal of the signal acquisition equipment; S400: the signal obtained in the step S300 is smoothly processed by utilizing a TFA Time Varying Filter algorithm of the LabVIEW; S500: source wave and the reflected wave of the partial discharge signal are identified by utilizing an LVQ algorithm of the LabVIEW as for the signal obtained in the step S400: and S600: information of the source wave and the reflected wave acquired in the step S500 is analyzed, and the position of partial discharge is calculated.

Description

Translated fromChinese

基于LabVIEW的电缆局部放电的位置识别方法Location Recognition Method of Cable Partial Discharge Based on LabVIEW

技术领域technical field

本发明涉及LabVIEW高级信号处理和机器识别技术领域，特别是涉及一种基于LabVIEW的电缆局部放电的位置识别方法。The invention relates to the technical field of LabVIEW advanced signal processing and machine identification, in particular to a position identification method of cable partial discharge based on LabVIEW.

背景技术Background technique

随着城市电网电缆化率的程度不断提高,社会发展和进步对供电可靠性的要求也不断提高,如何准确掌握配电电缆的健康状态,制定正确的检修对策,避免因电缆本身质量问题导致的突发性事故的发生,变得尤为重要。研究发现,电缆的局部放电量与其绝缘状况密切相关,局部放电量的变化预示着电缆绝缘可能存在危害电缆安全运行的缺陷。With the continuous improvement of the cable rate of urban power grids, the requirements of social development and progress on the reliability of power supply are also constantly increasing. How to accurately grasp the health status of distribution cables, formulate correct maintenance countermeasures, and avoid accidents caused by the quality problems of the cables themselves The occurrence of unexpected accidents has become particularly important. The study found that the partial discharge of the cable is closely related to its insulation condition, and the change of the partial discharge indicates that there may be defects in the cable insulation that may endanger the safe operation of the cable.

目前，电缆局部放电识别普遍采用滤波后波形峰值提取的方法进行源波和反射波的识别。上述方法存在的一个严重问题就是如果干扰和有用信号的频率几乎在一个频率点附近时有用信号两边会出现很多杂波，从而很容易导致源波或反射波和杂波不容易区分，尤其是峰值比较小的反射波，严重时会出现识别错误的情况。At present, the identification of cable partial discharge generally adopts the method of extracting the peak value of the filtered waveform to identify the source wave and reflected wave. A serious problem with the above method is that if the frequencies of the interference and the useful signal are almost at the same frequency point, there will be a lot of clutter on both sides of the useful signal, which will easily cause the source wave or reflected wave to be difficult to distinguish from the clutter, especially the peak Relatively small reflected waves may cause recognition errors in severe cases.

发明内容Contents of the invention

为了克服上述现有技术的不足，本发明提供了一种基于LabVIEW的电缆局部放电的位置识别方法。In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method for identifying the position of cable partial discharge based on LabVIEW.

本发明采用以下具体技术方案：The present invention adopts following concrete technical scheme:

一种基于LabVIEW的电缆局部放电的位置识别方法，其步骤包括：A method for identifying the location of cable partial discharge based on LabVIEW, the steps of which include:

S100：电缆局部放电信号输入至硬件调理设备中，进行放大和滤波处理；S100: The cable partial discharge signal is input to the hardware conditioning equipment for amplification and filtering;

S200：经过放大和滤波处理后的信号输入至信号采集设备，对电缆局部放电信号进行采集；S200: Input the amplified and filtered signal to the signal acquisition device to collect the partial discharge signal of the cable;

S300：信号采集设备的输出信号经过LabVIEW的去直流、滤波和阀值判断处理；S300: the output signal of the signal acquisition device is processed by LabVIEW for DC removal, filtering and threshold judgment;

S400：对步骤S300所得的信号利用LabVIEW的TFA Time Varying Filter算法进行平滑处理；S400: The signal obtained in step S300 is smoothed using the TFA Time Varying Filter algorithm of LabVIEW;

S500：对步骤S400中所得的信号利用LabVIEW的LVQ算法来识别局部放电信号的源波和反射波；S500: Using the LVQ algorithm of LabVIEW to identify the source wave and reflected wave of the partial discharge signal for the signal obtained in step S400;

S600：对步骤S500中获取的源波和反射波的信息进行分析，计算出反射波在电缆中的具体位置：步骤S500所得波形分为三部分，左侧、中间部分和右侧；其中，中间部分的幅值波动平稳，左侧幅值最大的峰值(或谷值)处是源波的位置，右侧幅值最大的峰值(或谷值)处是反射波的位置，则反射波距离源波的位置可以使用以下公式进行计算得出：S600: Analyze the information of the source wave and the reflected wave obtained in step S500, and calculate the specific position of the reflected wave in the cable: the waveform obtained in step S500 is divided into three parts, the left side, the middle part and the right side; Part of the amplitude fluctuation is stable, the peak (or valley) with the largest amplitude on the left is the position of the source wave, and the peak (or valley) with the largest amplitude on the right is the position of the reflected wave, and the distance between the reflected wave and the source wave is The position of the wave can be calculated using the following formula:

L＝(L₂-L₁)/v_sam*vL＝(L₂ -L₁ )/v_sam *v

其中，L为反射波距离源波的距离；L₂为反射波索引位置；L₁为源波索引位置；v_sam为源波和反射波的采样率；v为源波和反射波在电缆中的波速；Among them, L is the distance between the reflected wave and the source wave; L₂ is the index position of the reflected wave; L₁ is the index position of the source wave; v_sam is the sampling rate of the source wave and reflected wave; v is the source wave and reflected wave in the cable wave speed;

所述步骤S400中TFA Time Varying Filter算法对信号进行平滑处理的具体过程为：In the step S400, the specific process of smoothing the signal by the TFA Time Varying Filter algorithm is as follows:

S410：初始化内存；S410: initialize memory;

S420：对原始信号进行双高斯拟合处理；S420: Perform double Gaussian fitting processing on the original signal;

S430：接着将拟合后的信号分成两路，一路进行离散Gabor变换，另一路进行阈值参数操作，以及提取频谱分析参数得到峰值或谷值索引；S430: Then divide the fitted signal into two paths, one path performs discrete Gabor transform, the other path performs threshold parameter operation, and extracts spectrum analysis parameters to obtain peak or valley index;

在步骤S430中获取频谱分析参数的过程为：The process of obtaining spectrum analysis parameters in step S430 is:

使用频率范围在100k～10M的Butterworth带通滤波器对拟合后的信号进行频谱分析，获取频谱幅值的平均值和频谱幅值大于平均值的1.5倍的峰值或谷值索引。Use a Butterworth bandpass filter with a frequency range of 100k to 10M to perform spectrum analysis on the fitted signal, and obtain the average value of the spectrum amplitude and the peak or valley index whose spectrum amplitude is greater than 1.5 times the average value.

获取阈值参数的过程为：The process of obtaining threshold parameters is:

设定信号幅值的平均值乘以±1.2作为阈值，对信号幅值绝对值小于平均值绝对值1.2倍的信号部分进行移动窗口为5的滑动平均处理，然后用上述阈值对处理后的所有值进行二值化处理，得到阈值参数；Set the average value of the signal amplitude multiplied by ±1.2 as the threshold value, and perform a moving average processing with a moving window of 5 on the signal part whose absolute value of the signal amplitude value is less than 1.2 times the absolute value of the average value, and then use the above threshold value for all processed The value is binarized to obtain the threshold parameter;

S440：用提取到的频谱分析参数，峰值或谷值索引和阈值参数对离散Gabor变换后的信号进行离散Gabor展开操作；S440: Use the extracted spectrum analysis parameters, peak or valley index and threshold parameters to perform a discrete Gabor expansion operation on the discrete Gabor transformed signal;

S450：用原始信号减去离散Gabor展开后的信号。S450: Subtract the discrete Gabor expanded signal from the original signal.

所述步骤S500中LVQ算法识别局部放电信号的源波和反射波的具体过程为：The specific process of identifying the source wave and reflected wave of the partial discharge signal by the LVQ algorithm in the step S500 is as follows:

S510：首先对原始信号，也就是源波信号进行分割，将其分成若干段；S510: First, segment the original signal, that is, the source wave signal, and divide it into several segments;

S520：然后对分割信号中初始化并获取权重系数，用来反映原始信号中源波和放射波位置属性；S520: Then initialize and obtain weight coefficients in the segmented signal, which are used to reflect the position attributes of the source wave and the radiation wave in the original signal;

S530：将权重系数最大的两个值相应的索引值在相应的原始数据的图表中标识出来，幅值最大的一个是源波索引，另外一个是反射波索引。S530: Mark the index values corresponding to the two values with the largest weight coefficients in the corresponding original data chart, the one with the largest amplitude is the source wave index, and the other is the reflected wave index.

其中，TFA Time Varying Filter算法，即为Time Frequency Analysis TimeVarying Filter算法；Among them, the TFA Time Varying Filter algorithm is the Time Frequency Analysis Time Varying Filter algorithm;

LVQ算法，即为Learning Vector Quantization算法；The LVQ algorithm is the Learning Vector Quantization algorithm;

步骤S520中获取权重系数的方法：The method for obtaining the weight coefficient in step S520:

S521：对于个数为N的输入样本，初始化权重系数λ_n＝1/N，以及权重系数的学习率ε，并给定最大迭代次数T，其中，0＜ε＜1，N和T均为大于1的整数；S521: For input samples whose number is N, initialize the weight coefficient λ_n =1/N, and the learning rate ε of the weight coefficient, and give the maximum number of iterations T, where 0<ε<1, N and T are both an integer greater than 1;

S522：判断是否满足迭代停止条件,迭代停止条件为：迭代次数大于最大迭代次数T，如果满足则退出迭代，如果不满足则继续；S522: Judging whether the iteration stop condition is satisfied, the iteration stop condition is: the number of iterations is greater than the maximum number of iterations T, if it is satisfied, exit the iteration, if not, continue;

S523：寻找输入样本x_i的最近邻w_j，满足下列计算公式，其中d_n(x_i,w_j)表示x_i与w_j之间的距离：S523: Find the nearest neighbor w_j of the input sample x_i , which satisfies the following calculation formula, where d_n (_xi , w_j ) represents the distance between x_i and w_j :

d_n(x_i,w_j)＝min(d_n(x_i,w_m),d_n(x_i,w_j)),0＜j＜N-1,0＜m＜N-1；d_n (_xi ,w_j )=min(d_n (xi_, w_m ),d_n (xi_, w_j )), 0<j<N-1, 0<m<N-1;

S524：更新权重系数λ_n：S524: Updating the weight coefficient λ_n :

${\mathrm{<span><span>\&lambda;</span>\&lambda;</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>+</span>+</span>\mathrm{<span><span>1</span>1</span>}<span><span>)</span>)</span><span><span>=</span>=</span>{\mathrm{<span><span>\&lambda;</span>\&lambda;</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>\mathrm{<span><span>\&epsiv;</span>\&epsiv;</span>}<span><span>(</span>(</span>\frac{\mathrm{<span><span>d</span>d</span>}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\mathrm{<span><span>N</span>N</span>}}<span><span>-</span>-</span>{\mathrm{<span><span>d</span>d</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>)</span>)</span>$

其中，0＜t＜N-1,0＜n＜N-1，d表示两个最近邻样本之间的距离；Among them, 0<t<N-1, 0<n<N-1, d represents the distance between two nearest neighbor samples;

S525：返回步骤S522。S525: return to step S522.

与现有技术相比，本发明的有益效果是：Compared with prior art, the beneficial effect of the present invention is:

1)经过硬件和软件滤波后大大减少有用信号附近的杂波信号，并且可以降低识别错误的概率。1) After hardware and software filtering, the clutter signal near the useful signal is greatly reduced, and the probability of recognition error can be reduced.

2)ASPT中的TFA Time Varying Filter算法，可以滤除绝大部分干扰对源波和反射波的识别的影响。2) The TFA Time Varying Filter algorithm in ASPT can filter out most of the influence of interference on the identification of source waves and reflected waves.

3)MLT中的LVQ算法，可以准确的判断出源波和反射波在信号中的位置。3) The LVQ algorithm in the MLT can accurately determine the position of the source wave and the reflected wave in the signal.

附图说明Description of drawings

图1为本发明的电缆局部放电识别的流程图；Fig. 1 is the flowchart of cable partial discharge identification of the present invention;

图2为本发明的TFA Time Varying Filter算法的流程图；Fig. 2 is the flowchart of TFA Time Varying Filter algorithm of the present invention;

图3为本发明的LVQ算法流程图；Fig. 3 is the LVQ algorithm flowchart of the present invention;

图4为原始波形；Figure 4 is the original waveform;

图5为经贝塞尔带通滤波器的效果；Fig. 5 is the effect through Bessel band-pass filter;

图6为经TFA Time Varying Filter平滑后的效果，阈值系数为0.9；Figure 6 shows the smoothing effect of TFA Time Varying Filter, the threshold coefficient is 0.9;

图7为经TFA Time Varying Filter平滑后的效果，阈值系数为0.8；Figure 7 shows the smoothing effect of TFA Time Varying Filter, the threshold coefficient is 0.8;

图8为经LVQ算法识别的效果。Fig. 8 is the effect identified by the LVQ algorithm.

具体实施方式Detailed ways

下面结合附图对本发明进一步说明。The present invention will be further described below in conjunction with the accompanying drawings.

如图1所示，一种基于LabVIEW的电缆局部放电的位置识别方法，其步骤包括：As shown in Figure 1, a method for identifying the location of cable partial discharge based on LabVIEW, the steps include:

S100：电缆局部放电信号输入至硬件调理设备中，进行放大和滤波处理；S100: The cable partial discharge signal is input to the hardware conditioning equipment for amplification and filtering;

S200：经过放大和滤波处理后的信号输入至信号采集设备；S200: input the amplified and filtered signal to the signal acquisition device;

S300：信号采集设备的输出信号经过LabVIEW的去直流、滤波、阀值判断操作；S300: the output signal of the signal acquisition device is subjected to DC removal, filtering, and threshold judgment operations of LabVIEW;

S400：对S300所得的信号使用LabVIEW的ASPT中TFA Time Varying Filter算法进行平滑处理；S400: Use the TFA Time Varying Filter algorithm in ASPT of LabVIEW to smooth the signal obtained by S300;

S500：对S400中所得的信号使用LabVIEW的MLT中LVQ算法来识别局部放电信号的源波和反射波；S500: Use the LVQ algorithm in the MLT of LabVIEW to identify the source wave and reflected wave of the partial discharge signal for the signal obtained in S400;

S600：根据S500中所得源波和反射波在信号中的位置、波速以及线缆长度计算产生局部放电的电缆位置。S600: Calculate the position of the cable where the partial discharge occurs according to the positions of the source wave and the reflected wave in the signal, the wave velocity and the length of the cable obtained in S500.

如图1所示，触发信号是一个脉冲，当触发脉冲产生时，振荡开始产生衰减并立即采集振荡波信号；振荡波信号是从高压电缆经过分压器采集到的；局部放电信号是振荡波信号经耦合电容和检测阻抗滤波后的高频信号。As shown in Figure 1, the trigger signal is a pulse. When the trigger pulse is generated, the oscillation begins to attenuate and the oscillatory wave signal is collected immediately; the oscillatory wave signal is collected from the high-voltage cable through a voltage divider; the partial discharge signal is an oscillatory wave The signal is a high-frequency signal filtered by a coupling capacitor and a detection impedance.

如图2所示，TFA Time Varying Filter算法的流程为：As shown in Figure 2, the process of the TFA Time Varying Filter algorithm is:

S410：初始化内存；S410: initialize memory;

S420：对原始信号进行双高斯拟合处理；S420: Perform double Gaussian fitting processing on the original signal;

S430：接着将拟合后的信号分成两路，一路进行离散Gabor变换，另一路进行阈值参数操作，以及提取频谱分析参数得到峰值或谷值索引；S430: Then divide the fitted signal into two paths, one path performs discrete Gabor transform, the other path performs threshold parameter operation, and extracts spectrum analysis parameters to obtain peak or valley index;

在步骤S430中获取频谱分析参数的过程为：The process of obtaining spectrum analysis parameters in step S430 is:

使用频率范围在100k～10M的Butterworth带通滤波器对拟合后的信号进行频谱分析，获取频谱幅值的平均值和频谱幅值大于平均值的1.5倍的峰值或谷值索引。Use a Butterworth bandpass filter with a frequency range of 100k to 10M to perform spectrum analysis on the fitted signal, and obtain the average value of the spectrum amplitude and the peak or valley index whose spectrum amplitude is greater than 1.5 times the average value.

获取阈值参数的过程为：The process of obtaining threshold parameters is:

设定信号幅值的平均值乘以±1.2作为阈值，对信号幅值绝对值小于平均值绝对值1.2倍的信号部分进行移动窗口为5的滑动平均处理，然后用上述阈值对处理后的所有值进行二值化处理，得到阈值参数；Set the average value of the signal amplitude multiplied by ±1.2 as the threshold value, and perform a moving average processing with a moving window of 5 on the signal part whose absolute value of the signal amplitude value is less than 1.2 times the absolute value of the average value, and then use the above threshold value for all processed The value is binarized to obtain the threshold parameter;

S440：用提取到的频谱分析参数，峰值或谷值索引和阈值参数对离散Gabor变换后的信号进行离散Gabor展开操作；S440: Use the extracted spectrum analysis parameters, peak or valley index and threshold parameters to perform a discrete Gabor expansion operation on the discrete Gabor transformed signal;

S450：用原始信号减去离散Gabor展开后的信号。S450: Subtract the discrete Gabor expanded signal from the original signal.

如图3所示，LVQ算法流程为：As shown in Figure 3, the LVQ algorithm flow is:

S510：首先对原始信号，也就是源波信号进行分割，将其分成若干段；S510: First, segment the original signal, that is, the source wave signal, and divide it into several segments;

S520：然后对分割信号中初始化并获取权重系数，用来反映原始信号中源波和放射波位置属性；S520: Then initialize and obtain weight coefficients in the segmented signal, which are used to reflect the position attributes of the source wave and the radiation wave in the original signal;

S530：将权重系数最大的两个值相应的索引值在相应的原始数据的图表中标识出来，幅值最大的一个是源波索引，另外一个是反射波索引。S530: Mark the index values corresponding to the two values with the largest weight coefficients in the corresponding original data chart, the one with the largest amplitude is the source wave index, and the other is the reflected wave index.

其中，TFA Time Varying Filter算法，即为Time Frequency Analysis TimeVarying Filter算法；Among them, the TFA Time Varying Filter algorithm is the Time Frequency Analysis Time Varying Filter algorithm;

LVQ算法，即为Learning Vector Quantization算法；The LVQ algorithm is the Learning Vector Quantization algorithm;

步骤S520中获取权重系数的方法：The method for obtaining the weight coefficient in step S520:

S521：对于个数为N的输入样本，初始化权重系数λ_n＝1/N，以及权重系数的学习率ε，并给定最大迭代次数T，其中，0＜ε＜1，N和T均为大于1的整数；S521: For input samples whose number is N, initialize the weight coefficient λ_n =1/N, and the learning rate ε of the weight coefficient, and give the maximum number of iterations T, where 0<ε<1, N and T are both an integer greater than 1;

S522：判断是否满足迭代停止条件,迭代停止条件为：迭代次数大于最大迭代次数T，如果满足则退出迭代，如果不满足则继续；S522: Judging whether the iteration stop condition is satisfied, the iteration stop condition is: the number of iterations is greater than the maximum number of iterations T, if it is satisfied, exit the iteration, if not, continue;

S523：寻找输入样本x_i的最近邻w_j，满足下列计算公式，其中d_n(x_i,w_j)表示x_i与w_j之间的距离：S523: Find the nearest neighbor w_j of the input sample x_i , which satisfies the following calculation formula, where d_n (_xi , w_j ) represents the distance between x_i and w_j :

d_n(x_i,w_j)＝min(d_n(x_i,w_m),d_n(x_i,w_j)),0＜j＜N-1,0＜m＜N-1；d_n (_xi ,w_j )=min(d_n (xi_, w_m ),d_n (_xi ,w_j )), 0<j<N-1, 0<m<N-1;

S524：更新权重系数λ_n：S524: Updating the weight coefficient λ_n :

${\mathrm{<span><span>\&lambda;</span>\&lambda;</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>+</span>+</span>\mathrm{<span><span>1</span>1</span>}<span><span>)</span>)</span><span><span>=</span>=</span>{\mathrm{<span><span>\&lambda;</span>\&lambda;</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span><span><span>+</span>+</span>\mathrm{<span><span>\&epsiv;</span>\&epsiv;</span>}<span><span>(</span>(</span>\frac{\mathrm{<span><span>d</span>d</span>}<span><span>(</span>(</span>\mathrm{<span><span>t</span>t</span>}<span><span>)</span>)</span>}{\mathrm{<span><span>N</span>N</span>}}<span><span>-</span>-</span>{\mathrm{<span><span>d</span>d</span>}}_{\mathrm{<span><span>n</span>no</span>}}<span><span>)</span>)</span>$

其中，0＜t＜N-1,0＜n＜N-1，d表示两个最近邻样本之间的距离；Among them, 0<t<N-1, 0<n<N-1, d represents the distance between two nearest neighbor samples;

S525：返回步骤S522。S525: return to step S522.

如图4所示，该信号是经过硬件调理设备进行放大、滤波操作后，再经过信号采集设备到计算机内存中，并且使用LabVIEW进行一些基本的信号处理，比如去直流、滤波和阈值判断操作后的信号。As shown in Figure 4, the signal is amplified and filtered by the hardware conditioning equipment, and then sent to the computer memory through the signal acquisition equipment, and uses LabVIEW to perform some basic signal processing, such as removing DC, filtering and threshold judgment operations. signal of.

如图5所示，该信号是对图4中的信号经过贝塞尔带通滤波器后的信号，信号采样率为100MHz，设置参数如下：下限截止频率为5MHz、上限截止频率为20MHz、贝塞尔带通滤波器阶数为9。As shown in Figure 5, the signal is the signal in Figure 4 after the Bessel band-pass filter, the signal sampling rate is 100MHz, the setting parameters are as follows: the lower limit cut-off frequency is 5MHz, the upper limit cut-off frequency is 20MHz, the Bessel The order of the Searle bandpass filter is 9.

如图6所示，该信号是对图5中的信号经过TFA Time Varying Filter处理后的信号，设置参数如下：阈值大小为0.9。As shown in Figure 6, this signal is the signal processed by the TFA Time Varying Filter on the signal in Figure 5, and the setting parameters are as follows: the threshold value is 0.9.

如图7所示，该信号是对图5中的信号经过TFA Time Varying Filter处理后的信号，设置参数如下：阈值大小为0.8。As shown in Figure 7, this signal is the signal processed by the TFA Time Varying Filter on the signal in Figure 5, and the setting parameters are as follows: the threshold value is 0.8.

由图6和图7可以看出：不同的阈值系数对信号的平滑处理效果还是有所差异的，优选阈值范围为0.8～0.95。It can be seen from FIG. 6 and FIG. 7 that different threshold coefficients still have different smoothing effects on the signal, and the preferred threshold range is 0.8-0.95.

如图8所示，该信号是对图6中的信号进行LVQ算法识别后的信号。根据采样率、波速等信息，对图5中所得的信号进行分析，即可计算出反射波在电缆中的具体位置：步骤S500所得波形分为三部分，左侧、中间部分和右侧；其中，中间部分的幅值波动平稳，左侧幅值最大的峰值(或谷值)处是源波的位置，右侧幅值最大的峰值(或谷值)处是反射波的位置，则反射波距离源波的位置可以使用以下公式进行计算得出：As shown in FIG. 8 , the signal is a signal obtained by performing LVQ algorithm identification on the signal in FIG. 6 . According to sampling rate, wave velocity and other information, the signal obtained in Fig. 5 is analyzed to calculate the specific position of the reflected wave in the cable: the waveform obtained in step S500 is divided into three parts, the left side, the middle part and the right side; , the amplitude fluctuation in the middle part is stable, the peak (or valley) with the largest amplitude on the left is the position of the source wave, and the peak (or valley) with the largest amplitude on the right is the position of the reflected wave, then the reflected wave The position from the source wave can be calculated using the following formula:

L＝(L₂-L₁)/v_sam*vL＝(L₂ -L₁ )/v_sam *v

其中，L为反射波距离源波的距离；L₂为反射波索引位置；L₁为源波索引位置；v_sam为源波和反射波的采样率；v为源波和反射波在电缆中的波速。Among them, L is the distance between the reflected wave and the source wave; L₂ is the index position of the reflected wave; L₁ is the index position of the source wave; v_sam is the sampling rate of the source wave and reflected wave; v is the source wave and reflected wave in the cable wave speed.

Claims (6)

1. based on a location recognition method for the cable local discharge of LabVIEW, it is characterized in that, its step comprises:

S100: cable local discharge signal inputs in hardware conditioning device, carries out amplification and filtering process;

S200: the signal after amplification and filtering process inputs to signal collecting device, gathers cable local discharge signal;

S300: the output signal of signal collecting device judges process through go direct current, filtering and the threshold values of LabVIEW;

S400: the smoothing process of the TFA Time Varying Filter algorithm signal of step S300 gained being utilized to LabVIEW;

S500: utilize the LVQ algorithm of LabVIEW to identify source ripple and the reflection wave of local discharge signal to the signal of gained in step S400;

S600: the information of the source ripple obtained in step S500 and reflection wave is analyzed, calculates reflection wave particular location in the cable: step S500 gained waveform is divided into three parts, left side, center section and right side; Wherein, the amplitude fluctuations of center section is steady, the peak value of left side amplitude maximum or valley place are the positions of source ripple, and the peak value of right side amplitude maximum or valley place are the positions of reflection wave, then the position of reflection wave distance sources ripple can use following formula to calculate:

L＝(L₂-L₁)/v_sam*v

Wherein, L is the distance of reflection wave distance sources ripple; L₂for reflection wave index position; L₁for source ripple index position; v_samfor the sampling rate of source ripple and reflection wave; V is source ripple and reflection wave velocity of wave in the cable.

2. the location recognition method of a kind of cable local discharge based on LabVIEW as claimed in claim 1, it is characterized in that, in described step S400, the detailed process of TFA Time Varying Filter algorithm to the smoothing process of signal is:

S410: initialization internal memory;

S420: double gauss process of fitting treatment is carried out to original signal;

S430: then the signal after matching is divided into two-way, VARIANT FILTER OF DISCRETE GABOR TRANSFORM is carried out on a road, and threshold parameter operation is carried out on another road, and extraction spectrum analysis parameter obtains peak value or valley index;

S440: by the spectrum analysis parameter extracted, peak value or valley index and threshold parameter carry out Discrete Gabor Expansion operation to the signal after VARIANT FILTER OF DISCRETE GABOR TRANSFORM;

S450: deduct the signal after Discrete Gabor Expansion by original signal, obtains the signal after smoothing processing.

3. the location recognition method of a kind of cable local discharge based on LabVIEW as claimed in claim 1, it is characterized in that, in described step S500, the source ripple of LVQ algorithm identification local discharge signal and the detailed process of reflection wave are:

S510: to original signal, namely ripple signal in source is split, and is divided into some sections;

S520: obtain weight coefficient to initialization in splitting signal, is used for source ripple and radioactive wave position attribution in reflection original signal;

S530: identify out by maximum for weight coefficient two corresponding index values of value in the chart of corresponding raw data, one of amplitude maximum is source ripple index, and another one is reflection wave index.

4. the location recognition method of a kind of cable local discharge based on LabVIEW as claimed in claim 2, it is characterized in that, the process obtaining spectrum analysis parameter in step S430 is:

Frequency of utilization scope carries out spectrum analysis in the Butterworth bandpass filter of 100k ~ 10M to the signal after matching, and the mean value of acquisition spectral magnitude and spectral magnitude are greater than peak value or the valley index of 1.5 times of mean value.

5. the location recognition method of a kind of cable local discharge based on LabVIEW as claimed in claim 2, it is characterized in that, the process obtaining threshold parameter in step S430 is:

The mean value of setting signal amplitude is multiplied by ± 1.2 as threshold value, signal section signal amplitude absolute value being less than to mean value absolute value 1.2 times carries out the running mean process that moving window is 5, then by above-mentioned threshold value, binary conversion treatment is carried out to all values after process, obtain threshold parameter.

6. the location recognition method of a kind of cable local discharge based on LabVIEW as claimed in claim 3, is characterized in that, obtain the method for weight coefficient in step S520:

S521: be the input amendment of N for number, initializes weights coefficient lambda_n=1/N, and the learning rate ε of weight coefficient, and given maximum iteration time T, wherein, 0 < ε < 1, N and T is the integer being greater than 1;

S522: judge whether to meet iteration stopping condition, iteration stopping condition is: iterations is greater than maximum iteration time T, if met, exits iteration, if do not met, continues;

S523: find input amendment x_iarest neighbors w_j, meet following computing formula, wherein d_n(x_i, w_j) represent x_iwith w_jbetween distance:

d_n(x_i,w_j)＝min(d_n(x_i,w_m),d_n(x_i,w_j)),0＜j＜N-1,0＜m＜N-1；

S524: upgrade weight coefficient λ_n:

Wherein, 0 < t < N-1,0 < n < N-1, d represents the distance between two nearest samples;

S525: return step S522.