KR100477215B1 - Method for detecting partly spark of Gas Insulator Switchgear by wavelet transform - Google Patents

Abstract

A partial discharge detection method of a gas insulated switchgear using wavelet transform is disclosed. In the partial discharge detection method of the gas insulated switchgear using the wavelet transform of the present invention, since the external noise affects the measurement of the partial discharge in the gas insulated switchgear, the noise is removed using the wavelet transform. Normalized to compare the size. Although it can be seen that partial discharge occurs even if the noise is removed, the wavelet transform is applied again to clearly detect when the partial discharge occurs. As a result, the partial discharge detection was easy in the high pass filter D1 and the high pass filter D2, and the partial discharge could be accurately detected by the combination of the signals filtered in the high pass filter D1 and the high pass filter D2. According to the present invention, it is possible to continuously and uninterruptedly monitor the state of the gas insulated switchgear, and it is possible to monitor the state of the gas insulated switchgear that detects the accurate partial discharge in each time zone using an algorithm using a wavelet transform. . In particular, the application of an algorithm that is easy to implement hardware is expected to be effective in terms of practical value.

Description

Method for detecting partly spark of gas insulator switchgear by wavelet transform

The present invention relates to a partial discharge detection method of a gas insulated switchgear (GIS) using a wavelet transform, in particular, when the partial discharge occurs in a gas insulated switchgear quickly detects the precise detection and the degree of failure in a quick response and emergency recovery The present invention relates to a partial discharge detection method of a gas insulated switchgear using a wavelet transform that can improve the reliability of power supply.

In general, gas insulator switchgear (GIS) is for substations, and it is a device that protects the system appropriately by opening and closing the line safely under abnormal conditions such as accident, short circuit, etc. ₆ Opening / closing device such as disconnector (DS), ground switch (ES), and breaker (CB) in a gas-enclosed steel enclosure (Enclosure).

In general, the gas insulation switchgear is a water substation facility that uses a SF ₆ gas having excellent insulation performance and arcextinguishing function as an insulating medium in a metal sealed container (tank) to accommodate conductors and various protection devices to improve reliability. The whole system is being built by uniting various devices and connectors.

The gas insulated switchgear is configured by connecting various devices with buses, which are electrical conductors. The buses and devices insulated with gas may be disconnected or disconnected due to capacitance between devices, capacitance of devices themselves, voltage drop, etc. The voltage of about 300 [V] and the current of about 4000 [A] must be opened and closed. As a result, a part of the switching device is fused due to the arc heat generated by the arc, or insulation breakdown is caused by the conductive thermal gas generated at this time.

On the other hand, in order to solve the above problems, conventionally, the method of increasing the opening and closing speed of the disconnecting device is mainly used, the outside of the mechanism is provided with an electric spring operating device or a pneumatic cylinder using a pneumatic cylinder for generating a large power The opening speed of the mover was opened and closed at a high speed. Accordingly, there is a problem in that noise pollution occurs by using an excessive power generator such as the electric spring operating device or the pneumatic cylinder as described above. This is because the existing methods for monitoring the partial discharge of the gas insulated switchgear has many disadvantages that the detection of the partial discharge is difficult due to a lot of noise.

Accordingly, an object of the present invention is to monitor the abnormal symptoms of the gas insulated switchgear on-line at all times and predict the situation to occur in advance in order to provide more economical maintenance and reliable power supply. The present invention provides a partial discharge detection method of a gas insulated switchgear using wavelet transformation to prevent fatal accidents caused by partial discharge.

The present invention for achieving the above object of the present invention is to open and close the system in the normal state, accident and short-circuit state by the built-in switchgear and busbar including a disconnecting device, a grounding switch, a circuit breaker in a gas-sealed container Applied to a gas insulated switchgear for protection, the partial discharge detection method of the gas insulated switchgear using the wavelet transform of the present invention, the signal sensing step of detecting a signal including a partial discharge; Performing a first order wavelet transform on the signal to remove noise; A normalization step of normalizing the wavelet transformed signal to determine the magnitude comparison and the frequency of occurrence of the partial discharges; Performing the low pass filtering and the high pass filtering on the normalized signal and performing the low pass filtering and the high pass filtering on the low pass filtered signal to at least one or more times to an application region and a detail region. A secondary wavelet transform step of performing a secondary wavelet transform to decompose; A detail region combining step of combining the high-pass filtered detail region; And a partial discharge detection step of detecting a partial discharge through the combined detail region. At this time, the number of times of performing low pass filtering and high pass filtering in the second wavelet transform is preferably set within the range of 5 to 10.

Hereinafter, with reference to the accompanying drawings will be described a preferred embodiment of the present invention.

First, before the description of the present invention, let's briefly look at the wavelet transform and the state monitoring method using the same.

In general, performing multiple wavelet transforms can be problematic because a large amount of data is calculated and analysis time is lengthened. Therefore, if we choose scale and shift based on powers of two, the analysis is more efficient, and this analysis is implemented through discrete wavelet transform. Discrete Wavelet Transform Can be represented by the following [Formula 1].

[Equation 1]

Here, the variable representing the scale Where the variable representing the shift to be. Is an energy normalization component to maintain energy of the same size as the mother wavelet.

In general, the signal measured in the gas insulated switchgear is greatly affected by the ambient noise. Due to the influence of noise, it is difficult to check the presence or absence of partial discharge, so the noise is removed by using the wavelet transform. By the way, it can be seen that the generation of the partial discharge signal in the signal from which the noise is removed, but because the magnitude is different, it is difficult to determine the degree of severe partial discharge because the data cannot be compared with each other. In order to solve this problem, the present invention undergoes a normalization process. Through this normalization process, the magnitude of the partial discharge signal can be compared, and the wavelet transform is performed again. The occurrence of the partial discharge signal in the high pass filters D1 and D2 can be confirmed, and in the present invention, the partial discharge is detected by a combination of the signals filtered by the high pass filters D1 and D2.

1 is a view showing a partial discharge measurement position in the gas insulated switchgear. Referring to FIG. 1, the gas insulated switchgear buses (# 1BUS, # 2BUS), breakers 6100CB, upper disconnectors (6101DS), and lower disconnectors (6102DS) for each direct gas insulation switchgear preventive diagnosis check. Shows the measurement position of partial discharge using partial discharge measuring equipment. The acquired data has information of time, phase, charge amount and voltage, and the measurement is performed for 40 [sec]. In the present invention, as the partial discharge measuring equipment, a model name LDS-5, which is a computer-aided complete partial discharge measuring system manufactured and sold by LEMKE DIAGNOSTICS GmbH, Germany, was used.

Table 1 below shows the characteristics of the mother wavelet.

 Feature kind Morl Meyr Haar DbN SymN CoifN  Infinitely regular ⊙ ⊙  Compactly supported orthogonal ⊙ ⊙ ⊙ ⊙  Symmetry ⊙ ⊙ ⊙  Asymmetry ⊙  Near symmetry ⊙ ⊙  Arbitrary regularity ⊙ ⊙ ⊙  Existence of Φ ⊙ ⊙ ⊙ ⊙ ⊙  Orthogonal analysis ⊙ ⊙ ⊙ ⊙ ⊙  Biorthogonal analysis ⊙ ⊙ ⊙ ⊙ ⊙  Exact reconstruction ⊙ ⊙ ⊙ ⊙ ⊙  FIR filters ⊙ ⊙ ⊙ ⊙  Continuous transform ⊙ ⊙ ⊙ ⊙ ⊙ ⊙  Discrete transform ⊙ ⊙ ⊙ ⊙ ⊙  Fast algorithm ⊙ ⊙ ⊙ ⊙

Table 1 shows the characteristics of the mother wavelet. In the experiment for the present invention, the mother wavelet is selected as db4. The db4 mother wavelet is a scaling filter with four taps of the Dowchiechies mother wavelet, an eight-sample wavelet length, and transient transients occurring over a short period of time. It can be said to be a mother wavelet suitable for detecting the partial discharge.

2 is a graph showing a normalized signal after removing noise. Referring to FIG. 2, a result of removing a noise from a measured signal including a lot of noise using a wavelet transform is shown. The mother wavelet uses 'db4', which is good for nonlinear interpretation, and the level goes up to eight levels. Among the wavelet toolbox functions provided by Matlab, noise reduction uses various functions to remove noise. The occurrence of the partial discharge signal can be known using the sqtwolog threshold among the threshold methods. This signal is normalized to distinguish and compare multiple data and magnitudes.

3 is a graph showing a wavelet transform result of a detail region. Referring to FIG. 3, the signal of FIG. 2 is selected as a 'db4' mother wavelet and shows a detail region of a signal decomposed into a wavelet filter bank. The partial discharge signal is divided into a signal passing through the low pass filter (A1) and a signal passing through the high pass filter (D1) by performing a discrete wavelet, and the signal passing through the high pass filter (A1) is again divided into two low pass filters. (A2) and the high pass filter (D2), and the signal passing through the low pass filter was further decomposed into two low pass filters (A3 (not shown) and a high pass filter (D3) (not shown). This process has been completed up to eight steps, and the reason is to accurately detect partial discharge. Here, the number of times of performing the low pass filtering and the high pass filtering is preferably performed at least five times. Since the number of times of performing the low pass filtering and the high pass filtering exceeds ten times, the change in precision is not large. Preference is given to performing at 10 times or less.

FIG. 3 shows detail components from 1 to 8, wherein in (a), 10 partial discharges are detected, and the maximum value of the partial discharges is -0.7 to 0.6 at 32 seconds. In (b), five partial discharges were detected, and partial discharges were detected at 10 seconds, 12 seconds, 32 seconds, and 36 seconds which were not detected at D1, but 1 second, 3 seconds, 4 seconds, and 17 seconds detected at D1. A partial discharge of 27 seconds, 33 seconds, 37 seconds, 38 seconds was not detected. At 36 seconds, the coefficient value at which the partial discharge was the maximum value was detected as -0.9 to 1. In (c), partial discharge is not detected. Also in (d), partial discharge is not detected. In (e), partial discharge is detected at 33 seconds. The coefficient value of the partial discharge is -0.4 to 0.2. In Figure (f), the partial discharge signal is not detected. Also in (g), partial discharge is not detected. In (h), a distorted signal occurs for the first few seconds and around 40 seconds, but it is not a partial discharge because it is not a quantity of electric charge flowing at a moment for a short time.

FIG. 4 is a graph showing wavelet transform results in an approximation region. Referring to Figure 4, it shows an apoxy component (Approximation), and analyzed up to 8 steps as described above (Detail) component. In (a), the signal passing through the low pass filter A1 is reconstructed. Partial discharge is detected at 10 seconds, 12 seconds, 30 seconds, and 35 seconds because of the higher frequency range of the relatively low pass filter. The coefficient value of maximum partial discharge is a value of 0.6-1. In (b), the distorted signal is seen at 33 seconds but is not a partial discharge because it is not a signal flowing for a short time. In (c), a distorted signal is detected similarly at 33 seconds, but not partial discharge. In (d), only the distorted signal is continuously detected without any special change, and partial discharge is not detected. In (e), only the inherent characteristics of the measured signal are shown. In (f), although the inherent characteristics of the signal are well represented, detection of partial discharge is not observed. In Figure (g), we only detect distortion-free signals that show the inherent characteristics of the signal, but do not detect partial discharges. In (h), no partial discharge is detected at all.

5 is a flowchart illustrating a partial discharge detection process. Specifically, the partial discharge detection algorithm of the gas insulated switchgear using wavelet transform. Referring to FIG. 5, as described above, an arbitrary signal measurement is performed in the gas insulated switchgear (S10). Since the measured signal is affected by the surrounding noise, the partial discharge is not affected by the noise. The noise is removed using a wavelet transform to easily check the presence (S21). It can be seen that there is the occurrence of the partial discharge signal in the signal from which the noise is removed as shown in FIG. 2, but since the magnitude is different, it is difficult to compare several data with each other, and it is difficult to determine the degree of severe partial discharge. In order to solve this problem, the normalization process is performed (S22). Through the normalization, the magnitudes of the partial discharge signals can be compared, and the signal is subjected to wavelet transformation again (S31). It is possible to confirm the generation of the partial discharge signal in the high-pass filters D1 and D2, and in the present invention, the partial discharge is detected by the combination of the signals filtered by the high-pass filters D1 and D2 (S40) for more accurate detection (S50). ). As such, after the first wavelet transform and normalization process for noise removal is performed (S20), the wavelet transform is performed once again to proceed with the decomposition process up to level 8 (S30).

6 is a graph output by a combination result of the signal filtered by the high pass filters D1 and D2. Specifically, the result of applying the partial discharge detection algorithm to the partial discharge signal measured in the real system. Referring to FIG. 6, the maximum count value of the partial discharge was -0.8 to 0.9 and was detected at 36 seconds. In comparison with the detection of the partial discharge 14 times in FIG. 2, it can be seen that the number of occurrences of the partial discharge coincides.

7 is a graph illustrating signals measured in a real system gas insulated switchgear. Referring to FIG. 7, since the gas insulated switchgear is composed of several devices, a lot of noise is detected and detection of partial discharge is very difficult.

8 is a graph illustrating a signal in which noise is removed in a real system gas insulated switchgear. Referring to FIG. 8, a signal from which noise is removed using a wavelet transform is detected with respect to a signal detected in comparison with FIG. 7.

9 is a graph illustrating a partial discharge detection result in a real system gas insulated switchgear. Referring to FIG. 9, in order to facilitate detection of partial discharge in each time zone, the signal in FIG. 8 is once again combined with D1 and D2 using a wavelet transform to facilitate detection. Through this process, since the partial discharge of the real system gas insulation switch is detected, it can be said that it is a severe failure.

In summary, in the present invention, the partial discharge data was acquired by using a partial discharge measuring device (Model LDS-5 of LEMKE DIAGNOSTICS GmbH) in a 154 ㎸ GIS, and the measured signal was obtained from the gas insulated switchgear. Due to the noise of various devices, it is difficult to detect partial discharges. Therefore, the obtained data was removed by using wavelet transform, and the combination of the signals filtered by the high-pass filters D1 and D2 showed excellent results in the detection of partial discharges. .

Verification of the present invention is that the data acquired using the partial discharge measuring equipment (model LDS-5 of LEMKE DIAGNOSTICS GmbH) is stored in the form of ASCII code in the amount of charge over time, and the data thus stored is Matlab The wavelet transform provided in the analysis was used.

As described above, conventional gas insulation switchgear state monitoring methods have a disadvantage that the detection of partial discharge due to a lot of noise is difficult because the gas insulation switchgear is composed of many devices, the wavelet transform according to the present invention The partial discharge detection method of the gas insulated switchgear has the advantage that the state of the gas insulated switchgear can be monitored continuously and uninterrupted compared with other methods, and the noise is effectively removed by using the algorithm using wavelet transform. The detection of the partial discharge can monitor the state of the gas insulation switchgear. In particular, the application of an algorithm that is easy to implement hardware is expected to be effective in terms of practical value. Through this, when a failure occurs in the gas insulated switchgear, it is possible to quickly detect a failure, accurately determine it, and to accurately cope with and recover the power supply supply of the power system.

The present invention is not limited to the above-described embodiment, and it will be apparent that many modifications are possible by those skilled in the art within the technical spirit of the present invention.

1 is a view showing a partial discharge measurement position in the gas insulated switchgear;

2 is a graph showing a normalized signal after removing noise;

3 is a graph showing a wavelet transform result of a detail region;

4 is a graph showing wavelet transform results of an application region;

5 is a flowchart illustrating a partial discharge detection process;

6 is a graph output by a combination result of a signal filtered by a high pass filter D1 and a high pass filter D2;

7 is a graph showing a signal measured in a real system gas insulated switchgear,

FIG. 8 is a graph illustrating a signal in which noise is removed in a real system gas insulated switchgear;

9 is a graph illustrating a partial discharge detection result in a real system gas insulated switchgear.

Explanation of symbols on the main parts of the drawings

# 1BUS, # 2BUS: Gas Insulated Switchgear Bus

6100CB: Breaker

6101DS: Upper Disconnector

6102DS: Lower Disconnector

Claims (2)

In the partial discharge detection method of the gas insulated switchgear which protects the system by opening and closing the line in a normal state, accident and short circuit state by embedding a switchgear including a disconnector, a grounding switch, and a breaker in a gas-tight sealed container. , A signal sensing step of sensing a signal including a partial discharge; Performing a first order wavelet transform on the signal to remove noise; A normalization step of normalizing the wavelet transformed signal to determine the magnitude comparison and the frequency of occurrence of the partial discharges; Performing the low pass filtering and the high pass filtering on the normalized signal and performing the low pass filtering and the high pass filtering on the low pass filtered signal to at least one or more times to an application region and a detail region. A secondary wavelet transform step of performing a secondary wavelet transform to decompose; A detail region combining step of combining the high-pass filtered detail region; And a partial discharge detection step of detecting a partial discharge through the combined detail region; Partial discharge detection method of the gas insulated switchgear using a wavelet transform characterized in that it comprises a. The method of claim 1, wherein the number of times of performing low pass filtering and high pass filtering in the second wavelet transform is set within a range of 5 to 10. .