JP2004328810A - Method and device for partial discharge diagnosis of gas insulated switch - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To remove external noise for improved precision in partial discharge diagnosis. <P>SOLUTION: A signal is taken from a sensor 6, and a central frequency of a bandpass filter 11 is swept over to extract a frequency component (spectral data). A data processor 17 takes the spectral data, a swept frequency, and a system application voltage V through an A/D converter 16; and divides the spectral data into a plurality of frequency divisions for each cycle of the voltage V, generating a cycle pattern for each divisions. After a steady noise region is removed, the similarity between the cycle patterns for each frequency division acquired by repeated sweeping with signal intensity exceeding a threshold value is judged by a similarity judging part 19. A partial discharge abnormality is decided if the similarity is equal to a prescribed level or higher. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for diagnosing the insulation characteristics of gas-insulated equipment by means of partial discharge.
[0002]
[Prior art]
High-voltage switching device to be used in substations etc., a superior gas compressed filling insulating properties such as SF ₆ to the ground in the container, mechanically supports the high voltage division conductors moiety by solid insulator structure Gas insulation equipment is mainly used.
[0003]
In order to confirm the soundness and reliability of the insulation performance of these devices, it is extremely important to check for the occurrence of partial discharges from the main circuit structure such as insulators and metal conductors inside the devices. Therefore, various methods for performing insulation diagnosis of equipment through measurement of partial discharge have been proposed and put into practical use.
[0004]
The sensors for measuring partial discharge are roughly classified into those that capture a mechanical signal associated with the partial discharge and those that capture an electrical signal. As an example of the latter, a method of attaching a sensor (electromagnetic wave detection antenna) to the wall of a container of a high-voltage device and performing internal insulation diagnosis based on the signal of the sensor is used as a method that can detect abnormalities inside the device with high sensitivity Have been.
[0005]
High-voltage equipment in substations, in sealed container is filled with SF ₆ gas of about 0.5 MPa, (hereinafter referred to as GIS) Gas insulated switchgear incorporating circuit breakers and disconnecting switch or the like, in this Is the mainstream. SF ₆ gas in the container has extremely high dielectric strength, and when partial discharge occurs due to some abnormality, it becomes a steep pulse compared to air corona generated outside the container and generates electromagnetic waves with many high frequency components It is known to Therefore, a method of diagnosing the presence or absence of an internal abnormality by analyzing the signal detected by the sensor is effective.
[0006]
As a method for realizing this, for example, there is an invention described in Patent Document 1. Here, a cycle pattern is created by controlling a signal input switch by combining a phase controller linked to the cycle of an applied voltage before inputting a partial discharge signal detected by a sensor (such as an antenna) to a frequency analyzer. are doing.
[0007]
This method will be described with reference to FIG. The detected partial discharge signal 1 is input to the frequency analyzer 3 via the signal input switch 2. On the other hand, the system voltage is taken in as a reference signal 5 using a transformer or the like (not shown) and is selected by the signal changeover switch 4. The output of the signal changeover switch 4 is connected to a phase control device 6, which turns on and off the signal input switch 2 and a signal short-circuit switch 7 for short-circuiting an input signal. The on / off control in the phase control device 6 is performed by, for example, generating a gate signal having a preset delay time and pulse width from a zero point of the system voltage based on the signal of the signal changeover switch 4. In this way, the partial discharge signal in a specific phase range with respect to the system voltage is input to the frequency analyzer 3 and the information of the arbitrary phase and the information of the frequency analysis are changed by changing the setting of the phase controller 6. It can be obtained at the same time, so that partial discharge diagnosis can be advanced.
[0008]
[Patent Document 1]
JP-A-6-2300067 (paragraphs 0013-0017, FIGS. 1 and 2)
[0009]
[Problems to be solved by the invention]
In the above-described related art, a circuit configuration of the phase control device 6 is required. At the same time, since the partial discharge signal 1 is a weak high-frequency signal, it is necessary to take measures to suppress noise caused by the signal input switch 2 and the signal short-circuit switch 7 that are turned on and off by the gate signal. In particular, when there is external noise that is detected in synchronization with the system voltage, the noise is always input to the frequency analysis device 3, which easily affects the detection function.
[0010]
Further, in order to obtain the output of the frequency analyzer 3 over all phases, it is necessary to control the delay time, and a means for recognizing which phase is currently being measured is required.
[0011]
Further, for a partial discharge that occurs randomly with respect to the phase of the system voltage, if the timing of the gate and the timing of the occurrence of the partial discharge are different, the partial discharge cannot be detected. In particular, in the initial stage where the level of the partial discharge is small, it is likely to occur intermittently, which may affect the improvement of the detection sensitivity. In order to avoid this, it is necessary to perform measurement for a relatively long time while changing the delay time to prevent detection omission.
[0012]
SUMMARY OF THE INVENTION An object of the present invention is to overcome the above-described problems of the prior art, and to create a pattern by partial discharge even under a condition intermittently affected by external noise, and to solve a problem such as a timing shift. It is an object of the present invention to provide a method of diagnosing partial discharge of an insulation device. It is another object of the present invention to simplify the circuit configuration to reduce the size and weight of the diagnostic device and reduce the cost.
[0013]
[Means for Solving the Problems]
In order to achieve the above object, the present invention provides a method for diagnosing internal insulation abnormality by detecting an electromagnetic wave signal accompanying partial discharge by a sensor attached to a gas insulated switchgear, and sweeping a center frequency of a narrow band filter. While taking in the frequency spectrum of the electromagnetic wave signal and simultaneously taking in the cycle information of the system applied voltage, the spectrum band of the electromagnetic wave signal is frequency-divided based on the cycle information, and the spectrum of the same frequency division obtained by multiple sweeps A pattern is generated a plurality of times, and a partial discharge is determined when the first similarity between the spectral patterns exceeds a first predetermined value.
[0014]
When the frequency band is a specific frequency section in which a signal is constantly detected in the frequency band by the sweep, a spectrum pattern corresponding to the frequency section is excluded from a diagnosis area.
[0015]
Further, a spectrum pattern of the same frequency division is accumulated for a predetermined time to form a synthesized pattern, and a second similarity between the synthesized pattern and a previously registered abnormal factor pattern is a second predetermined value. Is determined to be a partial discharge when the value exceeds the threshold value.
[0016]
It is determined whether or not the partial discharge exceeding the second threshold value of the second similarity lasts for a predetermined time, and if it does not continue, it is notified as a sign of the partial discharge.
[0017]
A partial discharge diagnostic device for a gas insulated switchgear according to the present invention includes a narrow band filter that captures a frequency spectrum of the electromagnetic wave signal, a frequency shifter that repeats and sweeps a center frequency of the narrow band filter from a lower limit to an upper limit, A reference voltage generation unit that captures cycle information of a voltage applied to the switchgear; an A / D converter that captures the frequency spectrum, the center frequency, and the cycle information at the same time and digitally converts each; and the A / D conversion Frequency band of the spectrum band of the electromagnetic wave signal based on the cycle information by the output of the device, to generate a plurality of times the spectrum pattern of the same frequency division obtained by a plurality of sweeps, to determine the similarity between the spectrum pattern And a data processing unit for determining the presence or absence of partial discharge. That.
[0018]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a configuration diagram of a partial discharge diagnosis device of a gas insulated switchgear. The high-voltage conductor 3 is supported by an insulating spacer 4 made of a solid insulator, and is housed in a metal grounding container 1. The insulating spacers 4 are fixed to each other by a flange 2 provided at an end of the grounding container 1. Further, by combining the connections in the axial direction, the opening / closing device is configured in a required shape, and is installed as a whole on a stand (not shown).
[0019]
Built-in electrode sensors 6 are provided at various places in the grounding container 1. This sensor is an antenna designed to detect an electromagnetic wave propagating in a conduit caused by partial discharge with high sensitivity, and is installed in a structure that does not disturb the electric field distribution in the container.
[0020]
Note that the grounding container 1 has a shielding effect against external noise, but the insulating spacer 4 has no shielding effect, and electromagnetic waves such as air corona and broadcast / communication waves enter the grounding container 1 from this portion. . Conversely, the electromagnetic waves associated with the partial discharge generated inside the grounded container 1 leak from the insulating spacer 4 to the outside. Therefore, the sensor for partial discharge diagnosis can be combined with the external antenna 7 provided on the outer peripheral surface of the insulating spacer 4 in addition to the type of the built-in electrode sensor 6. However, the built-in electrode sensor 6 is excellent in detection sensitivity.
[0021]
The partial discharge signal detected by the built-in electrode sensor 6 is drawn out of the grounding container 1 through the sealed terminal 31 and is input to the partial discharge diagnostic device 5 using the coaxial cable 8.
[0022]
The partial discharge diagnostic device 5 converts the partial discharge into a partial discharge through a wide band filter 9 that passes a predetermined frequency band, an amplifier 10 that amplifies a high frequency signal, and a narrow band filter 11 whose center frequency (resonant frequency) is swept by a frequency shifter 21. The accompanying frequency component signal is detected. The center frequency is the resonance frequency of the narrow band filter 11.
[0023]
The detection signal is input to the A / D converter 16 via the signal line 15. At the same time, the signal of the frequency shifter 21 is also taken into the A / D converter 16 by the signal line 13, and the detected frequency information is obtained based on the signal.
[0024]
The reference voltage generator 12 (not shown in detail) generates a reference signal having the same phase as the main circuit applied voltage of the gas insulated switchgear, and the reference voltage V is transmitted to the A / D converter 16 via the signal line 14. It is captured. The input signals 13, 14, and 15 of the A / D converter 16 are digitally converted at the same time and taken into the data processing device 17 as instantaneous value data.
[0025]
The CPU 18 of the data processing device 17 fetches the digitized partial discharge signal strength, frequency information and reference voltage, and edits, processes, determines, and saves it by a predetermined program. When the detected partial discharge signal intensity satisfies a preset condition, the similarity determination unit 19 is activated and performs a comparison determination of the cycle pattern. If there is a similarity in the cycle patterns, accumulation is performed for a predetermined time, and the cause of the partial discharge is determined with reference to the diagnostic pattern unit 20.
[0026]
If the result of the determination is a partial discharge abnormality, a command is issued from the CPU 18 to the DI / O unit 22, and an alarm contact 24 is output to the outside via the input / output I / F unit 23. The diagnostic information / detected waveform data is displayed on the display unit 26 in a summarized form, but is transmitted to a higher-level personal computer (PC) (not shown) via the communication control unit 25 as necessary, and is used for diagnosis. Provided for detailed display and storage of results.
[0027]
FIG. 2 shows the relationship between the sweep characteristic, the phase of the reference voltage, and the detection signal. The frequency shifter 21 repeatedly sweeps between a lower limit (f _L ) and an upper limit (f _U ) of the frequency at a predetermined time (t). The data processing unit 17 measures the signal strength (Q) of the electromagnetic wave at the corresponding frequency via the narrow band filter 11. At the same time, the system reference voltage (v) is also taken in. This reference voltage indicates the position of the center frequency of the narrow band filter in the sweep cycle t, that is, the phase. Focusing on the signal spectrum, the waveform of the signal strength (Q) in FIG. 2 is observed at each time (t).
[0028]
If the signal strength (Q) in correspondence with the cycle of the system voltage, the data between f _n from the signal frequency component f _m is the frequency division Fn data. Although FIG. 2 shows one sweep cycle, a plurality of cycles are actually detected.
[0029]
FIG. 3 shows a cycle pattern of the detection signal. One cycle of the system voltage, for example, a detection signal corresponding to the _Fn section is displayed with reference to the time resolution of the apparatus. The signal intensity is represented by a cycle pattern as shown in FIG. 3, and this pattern becomes a characteristic pattern according to the factor of the partial discharge.
[0030]
Conventionally, the presence / absence of an abnormality in the insulation diagnosis apparatus is determined by comparing the measured signal strength with a preset threshold value and determining the abnormality based on the length of time during which the threshold value is exceeded. However, even when external noise is detected, if the signal intensity is equal to or higher than a predetermined level, it is erroneously determined that the partial discharge is abnormal.
[0031]
In order to avoid this, the present embodiment incorporates in advance a function of not performing abnormality determination in a frequency band in which stationary noise is occurring. However, when a temporary or intermittent noise other than the stationary noise is detected, a malfunction is likely to occur.
[0032]
In the present embodiment, the data of the signal strength in the repetitive sweep is generated a plurality of times as a cycle pattern for each frequency section (Fn), and the similarity between the cycle patterns in the section is determined. That is, the signal intensity detected by the sensor is affected by the structure of the GIS and the distance between the location where the partial discharge occurs and the sensor, and changes according to the detection frequency. Under the situation where the partial discharge is continuously generated, the cycle pattern shows the same aspect because the partial discharge pulse depends on the phase of the sweep cycle associated with the system voltage.
[0033]
Therefore, the similarity between a plurality of cycle patterns in the same phase is obtained, and if the similarity is equal to or higher than a predetermined level, it can be determined that the partial discharge is abnormal. Further, when the pattern is similar to the database of cycle patterns for each partial discharge factor stored in advance, the discharge factor can be specified.
[0034]
FIG. 4 shows the flow of the partial discharge diagnosis process. In ST (step) 1, a spectrum pattern of a signal is obtained by sweeping the center frequency of the bandpass filter 11. In ST2, the spectrum pattern is divided into a plurality of frequency divisions based on the reference voltage taken in at the same time, and a cycle pattern including a peak point of the signal intensity is created for each frequency division.
[0035]
In recent years, radio waves from mobile phones, transceivers, aeronautical radios, and the like have a spectrum intensity exceeding the partial discharge abnormality determination level, and particularly, the carrier frequency signal of the mobile phone is constantly detected regardless of the presence or absence of a call. Since this signal is generated in a known frequency band and has a characteristic pattern in terms of time width, it can be set in the determination exclusion area in advance. However, there are some types, such as aeronautical radios, which cannot be set to the judgment exclusion in advance because they are detected according to the arrival of an aircraft.
[0036]
Therefore, in ST3, when the signal time width of the cycle pattern approximates to a known noise pattern registered in advance in the frequency domain of the signal exceeding the detection level and continuing for a predetermined time, this frequency domain is set as a stationary noise domain. Excluded from judgment.
[0037]
Next, in ST4, the intensity of the signal outside the steady noise region is compared with a preset abnormality determination level. The signal strength is compared with the signal strength threshold in ST5, and if the signal strength threshold is exceeded, the first similarity determination of the cycle pattern is performed in ST6.
[0038]
FIG. 5 is an explanatory diagram of the similarity determination of the cycle pattern. Since the partial discharge is an electromagnetic wave having a wide spectrum, the partial discharge generally has a spectrum pattern having large and small peaks at a plurality of frequencies. Since the pattern appearing here is in a region other than the known noise region, there is a possibility that it is a signal accompanying partial discharge.
[0039]
For each of the frequency divisions F ₁ , F ₂ ,..., F _n ,..., Cycle patterns in the peak signal strength therebetween are sequentially created as shown in (a), (b), (c). You. Here, in the section (Fn) where the peak signal strength exceeds the signal strength threshold, the similarity of the cycle pattern in the section is determined. Depending on the degree of abnormality, there may be many frequency divisions (Fn) exceeding the threshold. For this reason, it is also possible to provide a restriction such as selecting the top five sections from the section having the highest signal strength.
[0040]
Same frequency division multiple times obtained in the sweep repetition similarity cycle patterns in (Fn) (referred to as a first degree of similarity it), i.e., (a) F ₁ classification, (b) F ₂ division, (c) F ₃ segment, in the cycle pattern in each group of ..., there will be those that exhibit high similarity to each other.
[0041]
In the example of FIG. 5, (a) group: low similarity, (b) group: high similarity, (c) group: low similarity... Conversely, if the similarity of the cycle pattern is not very high, it means that an extraneous signal due to another factor has been detected, which is considered to be due to intermittent external noise or the like. The determination of the similarity of the cycle pattern can be obtained by a practical accuracy by combining the area difference method of the waveform and the pulse count.
[0042]
When the first similarity is equal to or more than the first set value in ST7, an alarm may be generated as a partial discharge abnormality. In the present embodiment, the process further proceeds to ST8 to generate a combined cycle pattern and determine the second similarity.
[0043]
Partial discharge is a phenomenon that occurs stochastically due to insulation abnormality. Therefore, by accumulating data over a predetermined period of time and synthesizing each cycle pattern, it is possible to capture characteristics more. The combined cycle pattern is typically represented by a pattern using the maximum value of each cycle pattern in a predetermined number of sweep repetitions. When the partial discharge continues, a characteristic cycle pattern corresponding to the factor is recognized. In the present embodiment, a fixed diagnostic pattern is registered in the diagnostic pattern unit 20 in advance for each partial discharge whose factor is known.
[0044]
FIG. 6 shows a diagnostic pattern. Diagnosis patterns due to protrusions on the high-voltage conductors and foreign substances on the surface of the spacer are configured as combined cycle patterns for each factor of the partial discharge, and are stored in the diagnosis pattern unit 20.
[0045]
In ST9, it is determined whether or not the similarity (referred to as a second similarity) between the combined cycle pattern generated in ST8 and the combined cycle pattern for diagnosis is equal to or greater than a second set value. If there is, a persistence determination of exceeding the threshold is performed in ST10. On the other hand, when the second similarity is less than the set value, it may be regarded as a noise, not a partial discharge abnormality, and registered as a noise pattern in ST11.
[0046]
For the second similarity determination, the same method as the first similarity determination or a technique such as a hierarchical neuro determination can be applied. Thus, when the second similarity is determined to be equal to or greater than the set value, the factor of the detected partial discharge can be guided by the factor of the registered cycle pattern.
[0047]
In addition, even if the first and second similarities are high, if the occurrence is not persistent, it may not be possible to immediately determine that the abnormality is abnormal. For this reason, the presence or absence of the continuity of the cycle pattern is determined in ST10 and ST12, and when the continuity of the signal is small, a notification of the occurrence of the partial discharge sign in a sense of urging the operator to continue monitoring is performed in ST14. If the continuity is recognized, it is determined that the partial discharge is abnormal in ST13, and the data is stored and stored together with an external alarm.
[0048]
According to the present embodiment, by combining the first and second similarity determinations, malfunction due to external noise can be reliably prevented, and the cause of the partial discharge can be determined. Diagnosis becomes possible.
[0049]
【The invention's effect】
According to the signal processing of the present invention, a cycle pattern for each frequency component of the sensor signal associated with the phase of the frequency sweep cycle is obtained, and partial discharge is determined based on the similarity between the cycle patterns of the same phase. Even for a partial discharge, there is an effect that a cycle pattern can be reliably obtained and an initial sign of abnormality can be detected. In addition, by determining the first and second similarities, there is an effect that the determination accuracy of the partial discharge is increased and the factor determination is sure even in an environment where there is a lot of external noise. Furthermore, there is an effect that a low-cost diagnostic device can be provided with simplification of hardware that does not require a frequency analyzing device.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a partial discharge diagnosis device for a gas-insulated switch according to an embodiment of the present invention.
FIG. 2 is an explanatory diagram showing a relationship between a frequency sweep, a detection signal, and a system reference voltage.
FIG. 3 is an explanatory diagram showing a cycle pattern of a detection signal.
FIG. 4 is a flowchart showing a process of determining an abnormality of the partial discharge diagnostic device.
FIG. 5 is an explanatory diagram showing a combination of cycle patterns.
FIG. 6 is an explanatory diagram showing an example of a diagnostic pattern.
FIG. 7 is a configuration diagram schematically showing a conventional partial discharge diagnosis device.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Grounding container, 2 ... Flange part, 3 ... High voltage conductor, 4 ... Insulating spacer, 5 ... Partial discharge diagnosis device, 6 ... Built-in electrode (sensor), 7 ... External antenna (sensor), 8 ... Coaxial cable , 9 broadband filter, 10 high frequency amplifier, 11 narrow band filter, 12 reference voltage generator, 13 frequency component signal, 14 reference voltage signal, 15 partial discharge spectrum intensity signal, 16 A / D conversion Device 17 data processing unit 18 CPU 19 similarity determination unit 20 diagnostic pattern unit 21 frequency shifter 22 DI / O unit 23 input / output I / F unit 24 external output Contacts, 25: communication control unit, 26: display unit, 31: sealed terminal.

Claims (6)

In a method of diagnosing an internal insulation abnormality by detecting an electromagnetic wave signal accompanying partial discharge by a sensor attached to a gas insulated switchgear,
While the frequency spectrum of the electromagnetic wave signal is captured while sweeping the center frequency of the narrow band filter, the cycle information of the system applied voltage is simultaneously captured, and the spectrum band of the electromagnetic wave signal is frequency-divided based on the cycle information, and a plurality of sweeps are performed. Generating a spectrum pattern of the same frequency division obtained a plurality of times, and determining a partial discharge when the first similarity between the spectrum patterns exceeds a first predetermined value. Partial discharge diagnosis method. In claim 1,
Partial discharge of the gas insulated switchgear, wherein, if the frequency band is a specific frequency segment in which a signal is steadily detected, the spectrum pattern corresponding to the frequency segment is excluded from a diagnostic area. Diagnostic method. In claim 1,
A case where a spectrum pattern of the same frequency division is accumulated for a predetermined time to create a synthesis pattern, and a second similarity between the synthesis pattern and a pre-registered abnormal factor pattern exceeds a second predetermined value. And a partial discharge diagnosis method for the gas insulated switchgear, wherein the partial discharge is determined. In claim 3,
Determining whether the partial discharge exceeding the second predetermined value of the second similarity lasts for a predetermined time, and notifying the partial discharge as a sign of the partial discharge when the partial discharge does not continue for a predetermined time, characterized in that: . From a sensor attached to the gas insulated switchgear, in a partial discharge diagnostic device that detects an electromagnetic wave signal accompanying partial discharge and diagnoses internal insulation abnormality,
A narrow-band filter that captures the frequency spectrum of the electromagnetic wave signal; a frequency shifter that repeats and sweeps the center frequency of the narrow-band filter from a lower limit to an upper limit; and a reference voltage that captures cycle information of a voltage applied to the gas-insulated switchgear. A creating unit, an A / D converter that simultaneously takes in the frequency spectrum, the center frequency, and the cycle information and converts them into digital signals; and an A / D converter that converts the electromagnetic wave signal based on the cycle information output from the A / D converter. A data processing unit that frequency-divides the spectrum band, generates a plurality of spectrum patterns of the same frequency division obtained by performing a plurality of sweeps, and determines the presence / absence of partial discharge by calculating the similarity between the spectrum patterns. A partial discharge diagnostic device for a gas insulated switchgear, characterized in that: In claim 5,
A partial discharge diagnostic device for a gas insulated switchgear, comprising: a diagnostic pattern unit for storing a spectral pattern by a known partial discharge for each factor, and comparing a spectral pattern by a detected electromagnetic wave signal with the diagnostic pattern for each factor.

Images