AC/DC LEAKAGE DETECTION METHOD

Abstract

An AC/DC leakage detection method, which supports the detection of DC leakage and AC leakage. A DC leakage current and a low-frequency leakage current are measured by means of magnetic modulation technology, an AC signal is measured in the form of pure induction, and two detection modes are performed in a time-sharing manner. A collected leakage signal is converted into a digital signal through an AD converter. By means of the method of the present invention, the processing of a leakage signal is divided into three channels for respectively processing DC leakage, low-frequency AC leakage, and high-frequency AC leakage. The overall effective value of residual current is calculated by integrating results of DC detection and AC detection. The method of the present invention supports the detection of a suddenly increased current; and when there is a suddenly increased current, detection mode switching is performed by detecting the current sudden change of the current.

Claims

1. A method for detecting an AC/DC leakage current, comprising: detecting a DC current and a low-frequency AC current in a mode of positive and negative square wave excitation, where a magnetic core enters a deep saturation region; detecting the low-frequency AC current and a high-frequency AC current in a pure induction mode; switching, in an AC detection mode, the AC detection mode to a DC detection mode in a case that it is determined that a sudden large current change is detected; re-windowing in the DC detection mode at a current time in a case that it is determined that the sudden large current change is detected; and performing tripping in a case that it is determined that a current is greater than a threshold.

2. The method for detecting an AC/DC leakage current according to claim 1, wherein an H-bridge driving module outputs a three-state voltage of positive excitation, negative excitation and zero excitation by adopting a circuit structure with multiplexing of an excitation coil and a detection coil; wherein the method comprises: detecting, in a case that a positive level and a negative level are outputted, the DC current and the low-frequency AC current, where bidirectional excitation is applied to the magnetic core to make the magnetic core enter a saturation region; and detecting, in a case that a low level is outputted, an AC current in the pure induction mode.

3. The method for detecting an AC/DC leakage current according to claim 1, comprising: performing time-sharing detection on the DC current and the AC current; wherein the low-frequency AC current is detected during DC current detection, and only an AC leakage current and a sudden DC leakage current are detected in the AC detection mode.

4. The method for detecting an AC/DC leakage current according to claim 1, wherein the method comprises: controlling a detection mode to switch immediately by detecting a sudden change of a current, comprising: in the AC detection mode, switching the detection mode in a case that an AC current or a DC current suddenly increases, if it is detected that an amplitude of the current detected within a period of time continuously exceeds a threshold for a number of times; and in the DC detection, windowing at the current time in a case that it is detected that the amplitude of the current detected continuously exceeds the threshold for a number of times, to quickly detect and tripping under a condition of sudden increase of the current.

5. The method for detecting an AC/DC leakage current according to claim 1, wherein in the DC detection mode, a processing method for a detection current comprises the following steps: step 1: adopting an excitation square wave for synchronization, and distinguishing effective data of a sampled signal corresponding to a rising edge of the square wave from effective data of a sampled signal corresponding to a falling edge of the square wave; step 2: extracting, after signal synchronization, effective data corresponding to the rising edge, and effective data corresponding to the falling edge; step 3: calculating a mean of the effective data corresponding to the rising edge and a mean of the effective data corresponding to the falling edge, by the following equations: $\{\begin{array}{c}\stackrel{\_}{x_p}=\frac{1}{N_1}.Math.\underset{i-0}{\overset{N_1-1}{.Math.}}{x}_{\mathrm{pi}}\\ \stackrel{\_}{x_n}=\frac{1}{N_2}.Math.\underset{i=0}{\overset{N_2-1}{.Math.}}{x}_{\mathrm{ni}}\end{array},$ wherein a valid data sequence of the effective data corresponding to the rising edge is x.sub.p=[x.sub.p0, x.sub.p1, x.sub.p2, . . . , x.sub.p(N.sub.1.sub.-1)], and a valid data sequence of the effective data corresponding to the falling edge is x.sub.n=[x.sub.n0, x.sub.n1, x.sub.n2, . . . , x.sub.n(N.sub.2.sub.-1)]; step 4: subtracting a mean difference of a non-leakage template signal, to acquire a demodulation value by an equation: y.sub.i=x.sub.p−x.sub.n; step 5: comparing the demodulation value y.sub.i with the threshold, and counting a comparison result; step 6: re-windowing at a current position in a case that a count is greater than a preset value for a time period; step 7: calculating an estimated DC current in the window, where a demodulation sequence in the window is y=[y.sub.0, y.sub.1, . . . , y.sub.L-1], and the estimated DC current is: $\overline{y}=\frac{1}{L}.Math.\underset{i=0}{\overset{L-1}{.Math.}}{y}_i;$ step 8: calculating an AC current at a frequency of 50 Hz in the DC detection mode by an equation: $y_{50\mathrm{HZ}}=\frac{1}{L}.Math.|\underset{n=0}{\overset{L-1}{.Math.}}x\left(n\right).Math.e^{-j\frac{2\pi n}{N}}.Math.;$ and step 9: correcting the estimated DC current and the AC current at the frequency of 50 Hz.

6. The method for detecting an AC/DC leakage current according to claim 1, wherein in a low-frequency AC detection mode, a processing method for a detection signal comprises the following steps: step 1: passing by an AD sampled signal through a low-pass filter; step 2: performing down sampling on the AD sampled signal, to reduce a sampling rate of a low-frequency signal; step 3: comparing a value of the sampled signal with a threshold Thr1 ; counting a comparison result; switching, in a case that a count is greater than a preset value, the low-frequency AC detection mode to the DC detection mode; and analyzing, in a case that a count is less than or equal to the preset value, in the AC detection mode; step 4: windowing in a case of analyzing in the AC detection mode; step 5: performing FFT analysis on data in the window, dividing an absolute value of a result of the FFT analysis by N, where N represents the number of FFT points; and step 6: correcting an amplitude of a signal at each frequency point.

7. The method for detecting an AC/DC leakage current according to claim 1, wherein in a high-frequency AC detection mode, a processing method for a detection signal comprises the following steps: passing by an AD sampled signal through a band-pass filter; windowing; and performing FFT analysis on data in the window.

8. The method for detecting an AC/DC leakage current according to claim 1, comprising: calculating an overall leakage current by synthesizing the DC current, the current at each low-frequency point and the current at each high-frequency point; and determining whether to performing tripping according to the current.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] FIG. 1 is a block diagram of an algorithm for detecting an AC/DC leakage current according to the present disclosure;

[0029] FIG. 2 is a diagram of a solution of an H-bridge driving circuit for detecting AC/DC current according to the present disclosure;

[0030] FIG. 3 is a schematic diagram showing synchronization using an excitation square wave according to the present disclosure;

[0031] FIG. 4 is a diagram showing difference of a signal with DC leakage and a signal without DC leakage according to the present disclosure;

[0032] FIG. 5 is a schematic diagram showing extraction of effective data at a rising edge and effective data at a falling edge respectively in a DC detection channel according to the present disclosure; and

[0033] FIG. 6 is a diagram showing a relationship between an estimated DC leakage current and an actual DC leakage current according to the present disclosure.

DETAILED DESCRIPTION

[0034] A solution for detecting an AC/DC leakage current and a signal processing method according to the present disclosure are described below in combination with a solution of an H-bridge driving circuit in FIG. 2 and a block diagram of an algorithm for detecting an AC/DC leakage current in FIG. 1.

[0035] As shown in FIG. 2, a coil N1 represents a coil winding for excitation and detection, a coil N2 represents a winding for a leakage current. A current signal is acquired through a resistor Rs. During DC current detection, a magnetic core is excited by a bi-directional square wave through an H-bridge, and the magnetic core enters a saturation region in both directions to detect a DC current and a low-frequency AC current. During AC current detection, the H-bridge outputs at a low level. An AC leakage signal is detected in a pure induction mode. A signal on the sampling resistor Rs is amplified, the signal is converted into a digital signal through AD, results of the DC current detection and the AC current detection are processed to acquire a residual current.

[0036] When DC leakage occurs, change of the signal on the sampling resistor is shown in FIG. 3. A dotted line represents a waveform of a signal in a case that a DC leakage current is 100 mA, and a solid line represents a waveform of a signal without leakage current. It can be seen that the signal deviates up and down when current leaks. When a low-frequency leakage current is detected, an AC signal may be formed by superposing DC signals for different time periods.

[0037] According to the above phenomena and principles, a processing method in a DC leakage current detection channel according to the present disclosure includes the following steps.

[0038] In step 1, an excitation square wave is used for signal synchronization. As shown in FIG. 4, a response signal when the magnetic core is positively excited is distinguished from a response signal when the magnetic core is negatively excited by using a rising edge and a falling edge of the excitation square wave. A horizontal axis represents a data sequence number after AD sampling, and a vertical axis represents a voltage of a resistor R.

[0039] In step 2, as shown in FIG. 5, effective data in a detection signal corresponding to a rising edge of the square wave and effective data in the detection signal corresponding to a falling edge of the square wave are extracted, where a flat area with slow change is an effective data area, a horizontal axis represents a data sequence number after AD sampling, and a vertical axis represents voltage amplitude.

[0040] In step 3, a mean of the effective data corresponding to the rising edge and a mean of the effective data corresponding to the falling edge are calculated.

[0041] It is assumed that the effective data corresponding to the rising edge extracted in step 2 is expressed by an equation: x.sub.p=[x.sub.p0, x.sub.p1, x.sub.p2, . . . , X.sub.p(N.sub.1.sub.-1)], and the effective data corresponding to the falling edge extracted in step 2 is expressed by an equation: x.sub.n=[x.sub.n0, x.sub.n1, x.sub.n2, . . . , x.sub.n(N.sub.2.sub.-1)]. The mean of the effective data corresponding to the rising edge, and the mean of the effective data corresponding to the falling edge are respectively calculated by the following equations:

[00001]$\{\begin{array}{c}\stackrel{\_}{x_p}=\frac{1}{N_1}.Math.\underset{i-0}{\overset{N_1-1}{.Math.}}{x}_{\mathrm{pi}}\\ \stackrel{\_}{x_n}=\frac{1}{N_2}.Math.\underset{i=0}{\overset{N_2-1}{.Math.}}{x}_{\mathrm{ni}}\end{array}.$

[0042] After the above operation, two means are outputted for each excitation square wave period, which respectively are the mean x.sub.p of the effective data corresponding to the rising edge and the mean x.sub.n of the effective data corresponding to the falling edge.

[0043] In step 4, a demodulation value is equal to a difference between the mean corresponding to the rising edge and the mean corresponding to the falling edge by an equation: y.sub.1=x.sub.p−x.sub.n, where x.sub.p represents the mean of the effective data corresponding to the rising edge, and x.sub.n represents the mean of the effective data corresponding to the falling edge.

[0044] In step 5, the demodulation value y.sub.i is compared with a preset threshold, and a comparison result is counted.

[0045] In a case that the demodulation value y.sub.i exceeds the threshold continuously by an amount, it is determined that a current suddenly increases, and windowing is performed directly at a current time instant, and data in the window is analyzed or tripped.

[0046] In step 6, windowing is performed on an output sequence.

[0047] The above demodulation point is a current corresponding to each square wave period. Windowing is performed on the demodulated sequence. It is assumed that Ts represents a time length of a window and f represents a frequency of the square wave, amount of data in the 20 ms window is expressed by an equation:

L=┌f.sub.square wave.Math.T.sub.s ┐.

[0048] In step 7, an estimated DC current is calculated in the window.

[0049] It is assumed that a sequence of demodulation points in the window is expressed by an equation: y=[y.sub.0, y.sub.1, . . . , y.sub.L-1]. L represents a data length in the window. The DC current is acquired by averaging the sequence of demodulation points in the window by an equation:

[00002]$\overline{y}=\frac{1}{L}.Math.\underset{i=0}{\overset{L-1}{.Math.}}{y}_i.$

[0050] In a case that N1 represents the number of turns of the winding coil and R represents resistance of the sampling resistor, an initial estimated DC current is expressed by an equation:

[00003]${\tilde{y}}_{\mathrm{DC}}=\frac{\stackrel{\_}{y}}{R}.Math.N_1.$

[0051] In step 8, an AC current at a frequency of 50 Hz in a DC detection mode is calculated.

[0052] In order to deal with the sudden increase of current, the amplitude is calculated only at the frequency of 50 Hz during the DC current detection, and the initial estimated current is corrected by an equation:

[00004]$y_{50\mathrm{HZ}}=\frac{1}{L}.Math.|\underset{n=0}{\overset{L-1}{.Math.}}x\left(n\right).Math.e^{-j\frac{2\pi n}{N}}.Math..$

[0053] In step 9, the estimated AC/DC current is corrected.

[0054] The estimated AC and DC current in the steps 7 and 8 are linearly corrected to acquire an actual DC leakage current and an actual AC leakage current at the frequency of 50 Hz. FIG. 6 is a diagram showing a relationship between an estimated DC leakage current {tilde over (y)}.sub.DC and an actual DC leakage current. Therefore, the actual DC leakage current may be acquired by linear correction.

[0055] In a low-frequency AC detection channel, a processing flow of the channel includes the following steps.

[0056] In step 1, an AD sampled signal passes through a low-pass filter to filter out high-frequency component, severing as anti-aliasing filter of down sampling.

[0057] In step 2, in the embodiment, a sampling rate of the AD signal is 1 Msps, and down sampling is performed on the AD sampled signal.

[0058] In step 3, when the DC current or the AC current suddenly increases during the AC current detection, a voltage of the sampled signal is compared with a threshold Thr1 set by software. In a case that the voltage of the sampled signal is greater than the threshold Thr1, counting starts. In a case that a count is greater than a value for a time period, an interrupt signal is generated by hardware to a CPU, and a detection mode is controlled by software to change or is directly tripped.

[0059] In step 4, windowing is performed. The time length of the window is 20 ms, such as a Hamming window, a Hanning window or a rectangle window.

[0060] In step 5, FFT analysis is performed on data in the window, an absolute value of a result of the FFT analysis for each channel is calculated, and the calculated absolute value of each channel is divided by N, to acquire an amplitude of a leakage signal at each frequency point, where N represents the number of FFT points.

[0061] For an extracted channel, the following equation may be acquired:

[00005]$y_h\left(k\right)=\frac{1}{N}.Math.|\underset{n=0}{\overset{N-1}{.Math.}}x\left(n\right).Math.e^{-j\frac{2\pi kn}{N}}.Math.,$

[0062] where k ∈{0,1,2, . . . , N-1}.

[0063] In step 6, a signal at the frequency of 50 Hz is corrected, and then the signal at each frequency point is corrected, to acquire the amplitude of the signal at each frequency point.

[0064] In a high-frequency detection channel, a processing flow of the channel includes the following steps.

[0065] In step 1, the sampled signal passes through a band-pass filter;

[0066] In step 2, windowing is performed. In order to multiplex low-frequency FFT, the length of the window is 2 ms to ensure that the number of data points in the window is consistent.

[0067] In step 3, FFT analysis is performed on data in the window, an absolute value of a result of the FFT analysis is calculated, to acquire an amplitude of a leakage signal at each frequency point.

[0068] For an extracted channel, the following equation may be acquired:

[00006]$y_h\left(k\right)=\frac{1}{N}.Math.|\underset{n=0}{\overset{N-1}{.Math.}}x\left(n\right).Math.e^{-j\frac{2\pi kn}{N}}.Math.,$

[0069] where N represents the number of data points in the window, and k ∈{0,1,2, . . . , N-1}.

[0070] For the DC leakage current and AC leakage current calculated in the above detection modes, in a case that a current at each frequency point is greater than a threshold or an effective value is greater than a threshold, it may be determined that current leakage fault occurs.

[0071] The foregoing are merely preferred embodiments of the present disclosure. Those skilled in the art can make various modifications and variations to the present disclosure without departing from the principle of the present disclosure. Any modifications, equivalent substitutions and improvements made within the spirit and the principle of the present disclosure are within the protection scope of the present disclosure.