CN106980051B - A kind of intermittence tandem type fault electric arc recognition methods - Google Patents

Abstract

The invention discloses a kind of linear load intermittence tandem type fault electric arc recognition methods.Method includes the following steps: 1) carry out the experiment of tandem type fault electric arc, the line current signal under three kinds of acquisition load normal work, intermittent defect electric arc, closing operation electric arc states；2) comentropy under three kinds of states after the wavelet package reconstruction of current signal is sought；3) K mean cluster is carried out to comentropy, distinguishes running current signal and arc signal；4) it asks every kind to be supported on adjacent two periodic currents signal derivative maximum value ratio when the intermittent electric arc of generation and closing operation electric arc, and sets adjacent two periods derivative maximum value fractional threshold, distinguish intermittent defect electric arc and closing operation electric arc；5) intermittent tandem type fault electric arc criterion is established.Compared with prior art, energy reliable diagnosis intermittence tandem type fault electric arc of the present invention, and early warning can be made to stable tandem type fault electric arc.

Description

A method for identifying intermittent series arc faults

technical field

The invention relates to a fault arc protection method, in particular to an intermittent alternating current series type fault arc identification method.

Background technique

Arc faults are divided into parallel fault arcs and series fault arcs. Parallel fault arcs are mainly caused by leakage faults; series fault arcs are mainly caused by virtual connection of power lines, aging of cable material insulation, irregular wiring, unqualified device quality and external force damage. The parallel fault arc can be diagnosed by the leakage protection device; since the current amplitude of the series fault arc is less than or equal to the normal load current and is within the rated current range of the circuit breaker, it will not be detected by the traditional protection device. The detection dead zone makes the series arc fault arc the main cause of electrical fires. According to statistics, fire losses caused by electrical faults account for about 33% of the total fire losses in my country. It can be seen that the effective identification of series-type fault arcs has become a key issue to improve the reliability of power supply and prevent electrical fires.

In 1999, Underwriters Laboratories established UL1699 "Safety Standard for Arc Fault Circuit Breakers". The research on arc faults in my country is relatively late. In 2014, China issued the standard GB14287.4-2014 for electrical fire monitoring systems, in which the identification of fault arcs was defined.

In recent years, scholars at home and abroad have done a lot of research work on the identification of stable arc faults: ①According to the arc sound, arc light and other physical effects generated when the arc occurs to diagnose the fault arc; ②According to the sudden change of the arc voltage signal when the arc occurs and The zero-break phenomenon of the arc current is used to diagnose the fault arc; 3. The wavelet transform, short-time Fourier transform, etc. are combined with neural network or support vector machine. Changes to diagnose arc faults. Method ① is not applicable to the detection of fault arcs when the location of occurrence is unknown. When the location of the arc is unknown, the sudden change of the voltage in method ② cannot be detected, and the use of a large number of high-power power electronic devices has made the zero-break cannot be used as the detection basis for the occurrence of the arc. Although method ③ can identify the fault arc to a certain extent, it does not consider the difference between the operating arc and the fault arc.

The inventor found in the research that the series steady state arc fault is often accompanied by intermittent unstable series arc fault, and the duration of the intermittent arc fault is similar to that of the operating arc, generally not exceeding 5 cycles. Therefore, it is of great significance to effectively distinguish the operating arc from the intermittent fault arc, and to use the intermittent fault arc as an early warning that a steady-state fault arc will occur.

SUMMARY OF THE INVENTION

(1) Technical problems to be solved

In order to solve the problem that the arcing time of the operating arc and the intermittent series fault arc is short and difficult to distinguish, the present invention provides a method based on the wavelet packet information entropy and the K-means clustering algorithm to determine whether there is an arc. The method for identifying the operating arc and the intermittent series fault arc by the current signal derivative maximum value ratio threshold is characterized in that the specific steps are as follows:

1) Based on the self-developed AC220V series arc fault test platform, carry out the series arc fault simulation experiment of resistive load and resistive inductive load and the arc experiment of contactor closing operation. The main circuit of fault arc experiment includes AC power supply, circuit breaker, arc generator and load connected in sequence. The main circuit of the closing operation arc experiment includes AC power supply, contactor, circuit breaker, and load. The load mainly includes 800W bulb, 1600W bulb, 2.2KW no-load three-phase asynchronous motor.

2) Perform 4-layer wavelet packet decomposition on each 5-cycle current signal before and after arcing, and obtain the information entropy value of 16 frequency bands after the four-layer wavelet packet decomposition.

3) K-means clustering analysis is performed on the information entropy values during normal operation, intermittent fault arcs, and closing operation arcs.

4) Identify the normal current signal and the arc current signal through the K-means clustering results.

5) The derivation of the current signal in the K-means clustering result is carried out, and the threshold value of the current signal derivative value ratio of the adjacent two cycles of the operating arc and the intermittent fault arc is obtained, which is used as the diagnosis condition of the operating arc and the intermittent arc.

6) Establish the diagnostic criteria of intermittent arc fault.

The invention provides an intermittent series fault arc identification method based on the combination of wavelet packet information entropy, K-means clustering and the maximum threshold value of the derivative value ratio of adjacent two-cycle current signals, which solves the problem of operation arc and intermittent series arc. The problem of misjudgment improves the accuracy of the identification of the intermittent series arc fault, and can give an early warning for the stable arc fault to occur. The invention has universal applicability to the diagnosis of the intermittent arc fault of the linear load.

Description of drawings

Fig. 1 is the fault arc experimental circuit diagram of the present invention;

Fig. 2 is the circuit diagram of the closing operation arc experiment of the present invention;

Fig. 3 is the LabWIEW data acquisition program block diagram of the present invention;

Fig. 4 is the normal working current waveform of the 4×200W light bulb of the present invention;

5 is a waveform diagram of the normal working current of the 8×200W light bulb of the present invention;

Fig. 6 is the waveform diagram of the normal working current of the 2.2KW three-phase asynchronous motor of the present invention;

Fig. 7 is the arc current waveform diagram of 4 × 200W bulb closing operation of the present invention;

Fig. 8 is the arc current waveform diagram of the 8 × 200W bulb closing operation of the present invention;

Fig. 9 is a 2.2KW three-phase asynchronous motor closing operation arc current waveform diagram of the present invention;

10 is a waveform diagram of the intermittent series arc fault current of the 4×200W bulb of the present invention;

11 is a waveform diagram of the intermittent series arc fault current of the 8×200W bulb of the present invention;

Fig. 12 is the waveform diagram of the intermittent series arc fault current of the 2.2KW three-phase asynchronous motor of the present invention;

13 is a waveform diagram of arc fault current of 4×200W bulb series type of the present invention;

14 is a waveform diagram of arc fault current of 8×200W bulb series type of the present invention;

Fig. 15 is the 2.2KW asynchronous motor series arc fault current waveform diagram of the present invention;

Figure 16 shows the information entropy of each frequency band after the four-layer decomposition of the wavelet packet when the closing operation arc and the intermittent series fault arc occur;

Fig. 17 is the information entropy situation of each frequency band after the four-layer decomposition of the wavelet packet when the two rows of lamps according to the present invention work normally, the closing operation arc occurs, and the intermittent series fault arc occurs;

Fig. 18 is the information entropy of each frequency band after the four-layer decomposition of the wavelet packet when the 2.2KW no-load motor of the present invention works normally, the closing operation arc occurs, and the intermittent series fault arc occurs;

Fig. 19 is the information entropy clustering result of the wavelet packet when the typical load of the present invention works normally and the closing operation arc occurs;

Fig. 20 is the clustering result of wavelet packet information entropy when a typical load of the present invention works normally, a closing operation arc occurs, and an intermittent series fault arc occurs;

Fig. 21 is the detection criterion for the intermittent series arc fault of the resistive and resistive-inductive loads of the present invention.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings.

(1) Experimental circuit design

1. Main circuit

According to the UL1699 standard, the invention develops a fault arc simulation experiment device. The main circuit of the fault arc simulation experiment is shown in Figure 1. The main circuit consists of an AC power source, a circuit breaker, an arc generator, and a load. The arc generator consists of a stationary electrode and a moving electrode. The stationary electrode is a cylindrical copper rod, and the moving electrode is a conical copper rod. Good contact between the two electrodes is first achieved, then the moving electrode is separated from the stationary electrode by translation of the stepping motor, and the moving electrode stops when a stable arc occurs in the arc generator. The current transformer and the voltage transformer send the current signal collected in real time to the data acquisition card through the signal conditioning circuit.

The operating arc main circuit is shown in Figure 2. The main circuit consists of AC power supply, contactor, circuit breaker and load. Among them, the arc generated when the main contact of the contactor is closed is used as the operating arc, and the arc current and arc voltage are collected by the signal conditioning circuit and finally sent to the data acquisition card.

The load includes 800W bulb, 1600W bulb, 2.2KW no-load three-phase asynchronous motor.

2. Data acquisition software design

In the present invention, the collected data is uploaded to a PC equipped with LabVIEW 2013 professional edition software by a data acquisition card produced by NI Company for display and storage. The data acquisition LabVIEW software program is shown in Figure 3. Among them, 1 is the current signal time series, and 2 is the voltage signal time series. Through this program, the acquisition and waveform display of current and voltage signals can be realized.

(2) The specific experimental plan

The specific experimental scheme is shown in Table 1:

Table 1 Experimental scheme

Experimental group load type Working voltage/V Working current/A 1 Bulb 200W*4 220 3.6 2 Bulb 200W*8 220 7.2 3 Three-phase motor 2KW (no load) 380 5.0

(3) Criteria for intermittent series arc fault and operating arc

1. Experimental current signal acquisition

Figures 4-15 are the experimental current signals collected by LabVIEW. Figures 4-6 show the current signals of the 800W light bulb, 1600W light bulb, and 2.2KW no-load motor that were collected in real time through the data acquisition card during the experiment during normal operation. Figures 7-9 are the current signals of a typical load when a closing operation arc occurs. Figure 10-12 shows the current signal for a typical load with intermittent arc faults. Figures 13-15 show the current signals for a typical load with a stable arc fault. From Figure 4-15, it can be found that the durations of both the operating arc and the unstable arc fault are short, both within 5 cycles; the load has intermittent series arc fault before the stable arc fault occurs, so it can be The unstable arc fault is used as an early warning signal for the occurrence of a stable arc fault.

2. Four-layer decomposition and reconstruction of wavelet packet to obtain information entropy of each frequency band

The four-layer decomposition of the sym5 wavelet packet is carried out on the current signal of the experimental load current signal in the occurrence of the closing operation arc, the intermittent fault arc, and the normal operation after the arc, and the information entropy value of each frequency band is obtained.

Wavelet packet analysis is a signal analysis method with high time-frequency resolution, which overcomes the disadvantages of poor time resolution in low frequency band and poor frequency resolution in high frequency band of wavelet analysis. Let the signal be y(t), the recursive formula is as follows:

h(k) is a high-pass filter bank, and g(k) is a low-pass filter bank. {y _n (t)} is called the orthogonal wavelet packet, which is the result of all the decomposition of all frequency bands of the original signal at various scales. Let k=n- ^2j , then is the decomposition result of the signal on the frequency band k for scale j.

The information entropy formula used in the present invention is:

In formula ( ₂ ), m=1, 2, .

The inventor performs wavelet packet information entropy calculation on 10 sets of data for each typical load under normal operation, intermittent fault arcs, and closing operation arcs. The wavelet packet information entropy of the 5-cycle current signal when the intermittent fault arc occurs and the closing operation arc occurs is shown in Figure 16-Figure 18. By calculating the information entropy, it can be seen that the information entropy value has changed when the load is in normal operation, intermittent arc fault occurs, and the closing operation arc occurs, that is, the amplitude and energy of the current signal in each frequency band have changed.

3. Wavelet packet information entropy K-means cluster analysis

The inventor performed K-means cluster analysis on the information entropy of 800W light bulb, 1600W light bulb, and 2.2KW no-load motor under three conditions: normal operation, intermittent fault arc, and closing operation arc. The analysis results are shown in Figures 19-20, where the horizontal axis is the number of samples, and the vertical axis is the classification result. As can be seen from Figure 19, K-means clustering can realize the classification of arc signals and normal operation signals, and the recognition rate reaches 97%; as can be seen from Figure 20, K-means clustering cannot realize the classification of intermittent operation arcs and closing operation arcs .

4. Calculation of the ratio of the maximum value of the derivative of the adjacent two cycles of the current signal time series

In order to solve the problem that the wavelet packet information entropy K-means clustering cannot realize the classification of the intermittent operation arc and the closing operation arc, the present invention selects 5 groups when the closing operation arc and the intermittent unstable fault arc occur in the typical load in the experimental scheme. Current signal, analyze the 10-cycle time series of each group of signals, and find the ratio of the maximum value of the derivative of the adjacent two cycles. The results are shown in Table 2.

Table 2 The ratio of the maximum value of the derivative value of the 10-cycle current signal time series

5. Find the threshold value of intermittent series fault arc and closing operation arc

From Table 2, it can be seen that the ratio of the maximum value of the derivative value of the current signal in the adjacent two cycles of the intermittent series arc fault arc is very small, and the maximum value is 1.46; is 84.06. Therefore, the ratio of the maximum value of the derivative value of the current signal in the adjacent two cycles can be used as the diagnostic standard for the intermittent series fault arc and the closing operation arc, that is, if the ratio of the maximum value of the derivative value of the current signal in the adjacent two cycles is less than 3, it is intermittent If the ratio of the maximum value of the derivative value of the arc current signal in the adjacent two cycles is greater than 80, it is an operating arc.

6. Establishment of intermittent series arc fault criterion

According to the above items 1-5, the present invention establishes the intermittent series arc fault criterion, as shown in FIG. 21 . That is, according to the current signal time series, the four-layer decomposition of the wavelet packet is performed to obtain the information entropy, and then the K-means clustering analysis is performed on the information entropy value of the 16 frequency bands to diagnose whether an arc occurs, and then the maximum derivative of the adjacent two cycles is obtained for the arc current signal. The value ratio is used to judge whether an intermittent arc fault occurs through the threshold value. Finally, if an intermittent arc fault occurs, a fault arc warning signal will occur, which is ready for diagnosing the occurrence of a series-type arc fault.

Those skilled in the art should understand that any load identifying intermittent fault arcs and operating arcs through wavelet packet information entropy and K-means clustering and the ratio of the maximum value of the adjacent two-cycle derivative values all fall within the protection scope of the present invention.

Claims (3)

1. Judging whether there is an arc based on wavelet packet information entropy and K-means clustering algorithm, by seeking the ratio of the maximum value of the current signal derivative of adjacent two cycles to identify the closing operation arc and the intermittent series fault arc method, it is characterized in that , including the following steps: 1) Obtain the current signal time series; 2) Perform four-layer decomposition and reconstruction of wavelet packet sym5 on the current signal, and obtain the information entropy of each frequency band; 3) K-means clustering analysis is performed on the wavelet packet information entropy, which is divided into two types: arc current signal and normal operation current signal. If it is an arc current signal, then perform step 4), if it is a normal operation current signal, then return to step 1 ); 4) Take the derivation of the adjacent 10-cycle arc current signal time series, and obtain the ratio of the maximum value of the current signal derivative of the adjacent two cycles. If the maximum value of the ratio is greater than the closing operation arc threshold, it is determined as the closing operation arc. If the ratio is the largest If the value is less than the intermittent arc fault threshold, it is determined as intermittent arc fault, and an early warning signal of arc fault is issued. 2. The method according to claim 1 based on wavelet packet information entropy and K-means clustering algorithm to judge whether there is an arc, by seeking the ratio of the maximum value of the current signal derivative of the adjacent two cycles to identify the closing operation arc and the intermittent series fault arc. method, is characterized in that, described step 2) is specifically: Taking sym5 as the wavelet packet base, the four-layer wavelet packet decomposition and reconstruction are carried out on the five-period signals of the normal operation current, the closing operation arc current, and the intermittent fault arc current, and the information entropy value of the reconstructed signal of each frequency band is obtained. Information entropy is calculated by the following formula: In the formula, m=1,2,...,16 is the number of frequency bands of the reconstructed signal, x(t _i ) is the time series of the current signal, and n is the number of sampling points. 3. according to claim 2 based on wavelet packet information entropy and K-means clustering algorithm to judge whether there is an arc, by seeking the ratio of the maximum value of the current signal derivative of adjacent two cycles to identify the closing operation arc and the intermittent series fault arc. The method is characterized in that, in step 4), the maximum value ratio threshold of the current signal derivative value of the adjacent two cycles of the closing operation arc is 80, and the maximum value ratio threshold of the current signal derivative value of the two adjacent cycles of the intermittent fault arc is 3.