Partial discharge monitoring device and partial discharge monitoring method

Abstract

This invention further improves the monitoring accuracy of electrical equipment based on partial discharge occurrence. The physical quantity generated when partial discharge occurs in the inspected electrical equipment (1) is detected as an electrical signal. Based on the detector waveform representing the time variation of the amplitude of the detected electrical signal, a differential waveform is calculated, consisting of the difference between the amplitude of the detector waveform and a waveform with m periods offset from the power supply frequency of the electrical equipment (1). The distribution near zero in the amplitude distribution of the differential waveform data is assumed to be a normal distribution. An amplitude threshold (TA) is calculated based on the standard deviation (σ) of this normal distribution and a pre-set multiplier (X). Based on the calculated differential waveform, the moment when the amplitude exceeds the amplitude threshold (TA) is set as the partial discharge occurrence moment, the partial discharge signal is extracted, and the partial discharge status of the inspected electrical equipment (1) is evaluated based on the extracted partial discharge signal.

Description

Technical Field

[0001] This invention relates to technology for preventing malfunctions of AC electrical equipment owned by individuals or companies due to harsh environments or deterioration. Furthermore, this invention relates to technology for evaluating the occurrence of partial discharge in electrical equipment.

[0002] This invention is particularly suitable for use with switchboards above 400V (especially switchboards above 3kV) or transformers, and related electrical equipment such as wires and cables. Background Technology

[0003] Partial discharge refers to the phenomenon of localized bridging in insulators near conductors. In high-voltage insulated circuits, partial discharge can occur as a precursor to accidents. Therefore, by evaluating the occurrence of partial discharge, it is possible to detect the deterioration of insulation materials in advance. Furthermore, based on this detection, repair measures can be implemented, thereby preventing accidents from occurring.

[0004] Therefore, there is a need for monitoring technologies for electrical equipment, such as those used in operation, that can detect and diagnose abnormalities such as insulation degradation by detecting partial discharge.

[0005] As is well known, in alternating current equipment, partial discharge occurs periodically with a power supply cycle or twice that cycle. Furthermore, partial discharge is detected by extracting this periodicity from the waveform measured by a sensor capable of detecting partial discharge. A known method for detecting partial discharge uses, for example, a discharge component ratio calculated using the following formula. This discharge component ratio is derived from the power supply frequency component (fv component) and twice that power supply frequency component (2fv component) contained in the signal waveform.

[0006] Discharge component ratio [%)

[0007] ={(fv component + 2fv component) / total data} × 100

[0008] Here, the aforementioned fv component and 2fv component can be used to perform frequency analysis on waveforms composed solely of partial discharge signals, determining the component intensity as the power supply frequency or twice that frequency. Furthermore, in calculating the discharge component ratio, only the 2fv component can be used, or components that are integer multiples of the power supply frequency (three times or more) can be added.

[0009] On the other hand, in environments where AC equipment operates, in addition to partial discharge, interference noise such as electromagnetic vibration and magnetostrictive sound sometimes occurs. Furthermore, this interference noise sometimes exhibits periodicity equal to or an integer multiple of the power supply cycle. Therefore, there is a concern that interference noise might be misdetected as partial discharge if the periodicity of the signal received by a sensor for partial discharge coincides with the power supply cycle or an integer multiple thereof.

[0010] Therefore, for example, Patent Document 1 discloses a technique for detecting the sound generated during partial discharge, utilizing interference noise that occurs strictly at power cycle intervals relative to the power cycle, where partial discharge causes a timing deviation in the signal. Specifically, Patent Document 1 calculates the correlation coefficient between the signal waveform of the high-frequency component extracted from the detection signal and a signal waveform that is delayed by one power cycle relative to the signal waveform of the high-frequency component. Furthermore, Patent Document 1 discloses a discharge detection / identification device for detecting / identifying the discharge sound based on the aforementioned correlation coefficient and a frequency component twice the power frequency.

[0011] Existing technical documents

[0012] Patent documents

[0013] Patent Document 1: Japanese Patent No. 5092878 Summary of the Invention

[0014] The problem that the invention aims to solve

[0015] However, the method described in Patent Document 1 observes the overall correlation of the waveform obtained through a single measurement. Therefore, in Patent Document 1, even if the signal caused by the partial discharge to be detected overlaps with the interference noise signal to be excluded, it does not distinguish or separate the two signals within the waveform. Therefore, in Patent Document 1, for example, when the partial discharge signal is relatively small compared to the overlapping interference noise signal, there is a concern that the partial discharge signal may be missed. Furthermore, while Patent Document 1 can determine the presence or absence of partial discharge in the case of overlapping interference noise, it cannot evaluate the degree or frequency of partial discharge.

[0016] This invention addresses the aforementioned points, aiming to enable the evaluation of the degree and frequency of partial discharge even under interference noise by distinguishing or separating the signals of two components within a waveform obtained in a measurement. Furthermore, this invention aims to improve the accuracy of determining the presence or absence of partial discharge or the monitoring accuracy of electrical devices based on partial discharge detection.

[0017] Methods for solving problems

[0018] The inventors made the following discovery.

[0019] That is, the inventors considered the situation where interference noise with a power cycle as shown in Patent Document 1 occurs strictly at power cycle intervals, and partial discharge causes a timing deviation in the signal occurrence. Furthermore, the inventors derived the insight that by judging each signal rather than a whole measurement, the extraction accuracy of the partial discharge signal can be improved. Specifically, the inventors discovered that if the difference between the signal and the detected signal, which is offset by an integer period, is taken, the amplitude of the interference noise signal, which is strictly synchronized with the power cycle, is relatively smaller than that of the partial discharge signal. Based on this insight, the present invention was completed.

[0020] To address the problem, one aspect of the partial discharge monitoring device of the present invention is characterized by comprising: a sensor unit capable of detecting a physical quantity generated when a partial discharge occurs in the inspected electrical device as an electrical signal; a detection processing unit capable of extracting a detection processing signal as a time-varying amplitude of the signal waveform generated by the partial discharge from the electrical signal detected by the sensor unit; a partial discharge signal extraction unit capable of determining, based on the signal measured at each time moment and a signal that is offset from the power supply frequency of the electrical device by m periods (m: an integer greater than or equal to 1) relative to that time moment, whether the signal at that time moment is a partial discharge signal originating from the partial discharge, thereby extracting a partial discharge signal from the electrical signal or the detection processing signal; and a partial discharge evaluation unit capable of evaluating the occurrence of partial discharge in the inspected electrical device based on the partial discharge signal extracted by the partial discharge signal extraction unit.

[0021] Furthermore, the partial discharge monitoring method of another aspect of the present invention is characterized in that the physical quantity generated when partial discharge occurs in the inspected electrical equipment is detected as an electrical signal; from the detected electrical signal, the time change of the amplitude of the signal waveform generated by partial discharge is extracted as a detection processing signal; based on the signal measured at each time moment and the signal that is offset from the power supply frequency of the electrical equipment at that time moment by m periods (m: an integer greater than or equal to 1), it is determined whether the signal at that time moment is a partial discharge signal originating from partial discharge; the partial discharge signal is extracted from the electrical signal or the detection processing signal; and based on the extracted partial discharge signal, the condition of partial discharge occurrence in the inspected electrical equipment is evaluated.

[0022] Invention Effects

[0023] According to the method of the present invention, the amount of interference noise synchronized with the power supply cycle can be significantly removed from the electrical signal detected by the sensor, thus improving the extraction accuracy of the partial discharge signal. As a result, according to the method of the present invention, the monitoring accuracy of electrical devices based on the occurrence of partial discharge is further improved. Furthermore, even in the case of overlapping interference noise, the degree and frequency of partial discharge can be evaluated. Attached Figure Description

[0024] Figure 1 This is a diagram illustrating the structure of a partial discharge monitoring device based on an embodiment of the present invention.

[0025] Figure 2 This is a diagram showing an example of a detector waveform in the event of partial discharge.

[0026] Figure 3 This is an example diagram showing the correlation of the detected waveform.

[0027] Figure 4 This is an example of the frequency distribution of the absolute values ​​of the amplitude in a one-period differential waveform of the detector waveform.

[0028] Figure 5 This is a diagram showing an example of the frequency distribution in a one-period differential waveform of a detector waveform.

[0029] Figure 6 This is a diagram showing an example of a detected waveform and a waveform (partial discharge signal) from which only the partial discharge signal contained is extracted.

[0030] Figure 7 This is a diagram showing an example of a detected waveform and a waveform (partial discharge signal) from which only the partial discharge signal contained is extracted.

[0031] Figure 8 It is used to illustrate the frequency distribution of the detector waveform in Modified Example 2.

[0032] Figure 9 This is a diagram showing the region (shaded area) in Variation 2 where the contained data is considered as a partial discharge signal. Detailed Implementation

[0033] The inventors discovered that in differential waveform data, the amount of interference noise synchronized with the power supply cycle near zero can be approximated by a normal distribution with an average of zero. Furthermore, the aforementioned differential waveform data is composed of the amplitude difference between the waveform of a measured signal that has undergone detection processing and the amplitude of a signal offset from the power supply frequency by an integer number of cycles relative to the detected waveform. Based on this insight, in the following embodiment, the data in the aforementioned differential waveform near zero with small amplitude is approximated as a normal distribution. This allows for the determination of thresholds for interference noise and partial discharge signals that are strictly synchronized with the power supply cycle. Then, by extracting moments exceeding these thresholds, only the signal based on partial discharge is extracted from the measured signal waveform or its detected waveform.

[0034] In this embodiment, the identification of interference noise signals is improved by comparing them with signal waveforms that are offset by integer periods. In this respect, this embodiment is the same as Patent Document 1. However, in this embodiment, interference noise signals synchronized with the power supply cycle and partial discharge signals are separated and identified for each data point contained in the observed waveform. As a result, even when power supply cycle synchronization noise and partial discharge signals are mixed, only the partial discharge signal can be extracted in this embodiment. Therefore, this embodiment improves the monitoring accuracy of partial discharge occurrence.

[0035] Hereinafter, embodiments of the invention will be described in more detail with reference to the accompanying drawings.

[0036] However, the present invention can be embodied in various ways and is not limited to the embodiments described in this specification. These embodiments are provided so that those skilled in the art can fully understand the scope of the invention by making the disclosure of the specification sufficient.

[0037] (structure)

[0038] <Partial Discharge Monitoring Device 10>

[0039] like Figure 1 As shown, the partial discharge monitoring device 10 of this embodiment includes a sensor unit 10A, a detection processing unit 10B, a partial discharge signal extraction unit 10C, and a partial discharge evaluation unit 10D.

[0040] In addition, the partial discharge monitoring device 10 can operate continuously or intermittently in order to monitor the occurrence of partial discharge.

[0041] <Sensor Section 10A>

[0042] The sensor unit 10A is a measuring device capable of detecting the physical quantity generated when a partial discharge occurs in the electrical device 1 being inspected as an electrical signal.

[0043] In this embodiment, the sensor unit 10A is configured as a TEV sensor that detects changes in TEV (transient ground voltage) occurring in the inspected electrical equipment 1. The sensor unit 10A is not limited to a TEV sensor. The sensor unit 10A may also be an electromagnetic wave sensor that detects electromagnetic waves generated from the inspected electrical equipment 1 accompanied by partial discharge, or an ultrasonic sensor that detects ultrasound. Furthermore, the sensor unit 10A may also be an HFCT (high-frequency converter), an AE sensor (sound sensor), or the like.

[0044] There is no particular problem as long as the sensor unit 10A is a measuring device that can detect physical quantities related to partial discharge released from the inspected electrical equipment 1 as electrical signals.

[0045] The sensor unit 10A continuously measures the physical quantities generated during partial discharge as electrical signals.

[0046] <Detection Processing Unit 10B>

[0047] The detection processing unit 10B removes low-frequency noise components from the electrical signals continuously detected by the sensor unit 10A, for example, by using a high-pass filter.

[0048] Furthermore, in the detection processing unit 10B, peak hold detection processing based on holding the maximum value for a certain period of time is performed, and A / D conversion and quantification are performed at a preset sampling frequency. Additionally, the detection processing unit 10B does not need to use all the data detected by the sensor unit 10A.

[0049] <Partial Discharge Signal Extraction Unit 10C>

[0050] The partial discharge signal extraction unit 10C, based on the detected signal processed by the detection processing unit 10B, determines whether a partial discharge has occurred at a given time by using the signal measured at each time and the signal that is offset from the power supply frequency of the electrical device 1 by m periods (m: an integer greater than or equal to 1) relative to that time. Then, the partial discharge signal extraction unit 10C performs processing to extract the partial discharge signal from the electrical signal or the detected signal.

[0051] The determination of whether or not the aforementioned partial discharge occurs is based, for example, on the amplitude difference between the signal measured at each time and the signal that is offset from the power supply frequency of the aforementioned electrical device 1 by m periods (m: an integer greater than or equal to 1) relative to that time.

[0052] The partial discharge signal extraction unit 10C in this embodiment includes a differential waveform calculation unit 10CA, an amplitude threshold setting unit 10CB, and a signal extraction main unit 10CC.

[0053] <Differential Waveform Arithmetic Unit 10CA>

[0054] The differential waveform calculation unit 10CA performs differential waveform calculation based on the signal waveform (detection processing signal) processed by the detection processing unit 10B. The differential waveform is composed of the difference between the amplitude of the detection processing signal and the amplitude of a signal whose amplitude is offset from the power supply frequency of the aforementioned electrical device 1 by m periods (m: an integer greater than or equal to 1).

[0055] Through this process, the amplitude of the signal that is strictly synchronized with the power cycle becomes zero or smaller.

[0056] In the differential waveform calculation unit 10CA of this embodiment, a differential waveform is calculated, which is formed by the difference between the amplitude of the signal waveform processed in the detection processing unit 10B and the amplitude of the signal waveform that is offset from the power supply frequency by one cycle.

[0057] Here, the power frequency of the electrical device 1 is usually 60Hz or 50Hz, but the present invention is not limited to this.

[0058] <Amplitude threshold setting unit 10CB>

[0059] In this embodiment, it is assumed that the amplitude distribution of the differential waveform data calculated by the differential waveform calculation unit 10CA, near zero, follows a normal distribution with a mean of zero. Then, the amplitude threshold setting unit 10CB calculates the standard deviation σ in this normal distribution and sets the amplitude threshold TA based on the standard deviation σ.

[0060] In a normal distribution, the percentage of points whose difference from the mean is less than z times the standard deviation can be determined using a standard normal distribution table. For example, points whose difference from the mean is less than half the standard deviation account for 38.3% of the total.

[0061] Using this situation, for example, the following process can be performed: For the differential waveform calculated by the differential waveform calculation unit 10CA, an amplitude value that exactly contains the lower 38.3% of the absolute value of its amplitude is obtained. Then, by doubling this amplitude value, the standard deviation σ of the amplitude distribution of the differential waveform data calculated by the differential waveform calculation unit 10CA near zero can be obtained, showing that the distribution follows a normal distribution.

[0062] In this embodiment, the value of z is increased in increments of 0.1, starting from 0.1, and the standard deviation is estimated successively. Then, it is confirmed that if z is small enough, the estimated standard deviation σ takes a certain value, and this value is set as the standard deviation σ of the normal distribution followed by the distribution near the zero amplitude of the difference waveform.

[0063] The z mentioned above is used to calculate the standard deviation σ of the normal distribution followed by the distribution near zero amplitude of the difference waveform. z can be chosen from 0 to 1 to perform the above calculation. In this case, z is preferably set to 0.5.

[0064] Regarding the amplitude threshold TA, it can be set as the value obtained by multiplying the standard deviation σ by a multiplier X. The multiplier X is preferably selected from a range of 2 to 4. More preferably, the multiplier X is set to 3 or around 3 (e.g., 2.5 to 3.5).

[0065] In this embodiment, the case where the amplitude threshold TA is set to 3σ is illustrated.

[0066] <Signal Extraction Main Unit 10CC>

[0067] The signal extraction main unit 10CC extracts the moments exceeding the aforementioned amplitude threshold TA from the differential waveform calculated by the differential waveform calculation unit 10CA. Then, the signal extraction main unit 10CC identifies the signal corresponding to that moment in the detected signal as a partial discharge signal, and identifies other signals as signals based on interference noise. Thus, the signal extraction main unit 10CC performs zero-permutation on the interference noise-based signals, thereby performing processing to extract only the partial discharge signal waveform.

[0068] <Partial Discharge Evaluation Section 10D>

[0069] The partial discharge evaluation unit 10D performs a process to evaluate the presence and extent of partial discharge in the inspected electrical equipment 1 based on the partial discharge signal extracted by the signal extraction main unit 10CC.

[0070] The partial discharge evaluation unit 10D evaluates the occurrence of partial discharge in the device 1, for example, by using the PD pulse ratio (%). The PD pulse ratio (%) is the proportion of signals other than zero in the partial discharge signal waveform extracted by the signal extraction main unit 10CC. Regarding the evaluation, for example, the higher the PD pulse ratio, the more partial discharges are diagnosed. Furthermore, if the PD pulse ratio is above a preset threshold value, it is determined that maintenance of the device 1 is required.

[0071] Regarding the PD pulse ratio, for example, if it is several percent to more than 10%, it can be determined that partial discharge has reliably occurred.

[0072] The threshold PT for determining the PD pulse ratio is preferably set based on (a) below. When setting the multiplier X in the range of less than 2, (b) below is also considered, and it is preferred to set it by the larger of the two.

[0073] (a) Threshold for determining the proportion of PD pulses PT [%)

[0074] = (Power supply frequency [Hz] / Sampling frequency [Hz]) × 100

[0075] This value represents the proportion of PD pulses occurring exactly once per power supply cycle (1 data point). For example, with a power supply frequency of 60Hz and a sampling frequency of 2.56kHz, the PD pulse proportion is 2.34%.

[0076] (b) The probability q of errors that may occur when the frequency distribution of the waveform follows a normal distribution in the absence of partial discharge, i.e., after detection processing.

[0077] In addition, the partial discharge evaluation unit 10D can monitor the occurrence of partial discharge in the device 1 by performing conventional extraction power periodic processing on the partial discharge signal waveform extracted by the signal extraction main unit 10CC.

[0078] In this case, the discharge component ratio can also be used as an indicator of the periodicity of the power supply.

[0079] Furthermore, for the electrical equipment 1 under inspection, monitoring processing can be performed, for example, based on the increase in the ratio of PD pulses measured over time or the ratio of discharge components relative to the partial discharge signal waveform extracted by the signal extraction main unit 10CC.

[0080] <Regarding the sampling size of the measurement data used for evaluation>

[0081] The measurement time in the sensor unit 10A used for evaluating a single partial discharge is, for example, set to at least three times the power supply cycle. Preferably, in order to observe the periodicity of the extracted partial discharge signal, this measurement time is preferably at least ten times the power supply cycle.

[0082] Furthermore, the waveform data after detection processing used in the evaluation of a single partial discharge preferably consists of at least 6 points. More preferably, the waveform data after detection processing has at least 400 points. There is no upper limit to the amount of waveform data.

[0083] For example, if the length of a power cycle is 10 cycles and the data is 400 points, then each cycle contains 40 points (e.g., 0.4 to 0.5 ms per point), thus allowing sufficient observation of the discharge signal fluctuations. Alternatively, for example, with a power cycle of approximately 100 cycles, approximately 40 points of data can be acquired per cycle.

[0084] Here, the amplitude threshold TA is set to 3σ. In this case, if more than 400 points are generated following a normal distribution N(μ, σ) with a mean of μ and a standard deviation of σ, it can be expected that there is more than one point that is more than 3σ away from the mean. Therefore, it can be expected that the estimation of σ can be performed with excellent accuracy.

[0085] <Preferred range for the preset magnification X>

[0086] Here, we assume that there is data with more than 3 cycles relative to the power supply cycle τ.

[0087] At this point, if the measured value f(t) at time t after detection processing has both the difference f(t)-f(t-τ) with the previous cycle and the difference f(t)-f(t+τ) with the next cycle that are both above the amplitude threshold TA, then it is estimated to be a partial discharge signal.

[0088] Furthermore, consider the case where no partial discharge occurs at any of the times t-τ, t, and t+τ. Also, let p be the probability that the difference f(t)-f(t-τ) and the difference f(t)-f(t+τ) accidentally exceed the amplitude threshold TA. In this case, with probability p... 2 Under (=q), the signal f(t) that is not a partial discharge will be treated as a partial discharge.

[0089] The probability p that the difference between signals that are not partial discharges is the amplitude threshold TA is multiplied by the factor X.

[0090] Therefore, the range of the multiplier X is examined as follows.

[0091] For a given σ, it follows a normal distribution N(0, σ) with a mean of zero. 2 The probability p of a signal entering a range above nσ and the value of p at that time. 2 As shown in Table 1.

[0092] [Table 1]

[0093] nσ p <![CDATA[p 2 (=q)]]> 1σ 0.3174 0.1007 2σ 0.0456 0.002079 3σ 0.0026 0.00000676 4σ 0.00006334 <![CDATA[4.0119556*10 -9 ]]>

[0094] According to Table 1, if it is 1σ, then p 2 =10%. If a 10% error rate can be tolerated, then the multiplier X can also be set to 1. However, considering a 10% error rate as noise removal is insufficient.

[0095] On the other hand, the error at 2σ is less than 0.2%, so if the multiplier X is set to 2 or higher, noise removal can be sufficiently achieved. Conversely, if n of nσ is excessively increased, the possibility of missing partial discharge signals increases.

[0096] Thus, the multiplier X is preferably in the range of 2 or higher and 4 or lower.

[0097] (Actions and others)

[0098] For the measured waveform measured by the sensor unit 10A, the waveform after detection processing in the detection processing unit 10B becomes, for example, Figure 2 That way. Figure 2 The horizontal axis represents time, and the vertical axis represents amplitude.

[0099] Figure 2 The measured waveform shown is an example of a case with partial discharge data.

[0100] about Figure 2 The waveform f(t) shown is plotted with data points of f(t) on the horizontal axis and data points of the measured waveform f(t-τ) staggered by one period on the vertical axis, as shown in the following graph. Figure 3 That's how it's recorded.

[0101] Here, interference noise such as magnetostrictive tone is noise that is strictly synchronized with the power supply cycle (power supply cycle synchronization noise). Therefore, the signal of this interference noise in... Figure 3 The position, like the symbol A, becomes a straight line close to 45 degrees. On the other hand, the signal originating from partial discharge occurs with fluctuations relative to the power supply cycle. Therefore, in Figure 3 In this diagram, as indicated by symbol B, the signal originating from partial discharge oscillates outwards along a 45-degree straight line. Furthermore, the position of this 45-degree straight line represents the point where the amplitude difference between the measured waveform and a measured waveform offset by one cycle is zero. The further away from this 45-degree straight line, the greater the amplitude difference.

[0102] Thus, regarding the measured waveform, it can be seen that in the differential waveform data, the amplitude of the power supply cycle synchronization noise is concentrated near zero, while the signal originating from partial discharge exists at an amplitude position above the specified value. Here, the differential waveform data is the difference in amplitude between the original waveform and the waveform offset by one cycle.

[0103] That is, if the amplitude of the differential waveform is taken as its absolute value and its frequency distribution is obtained, then it becomes Figure 4 That way. In Figure 4 In this diagram, the horizontal axis is set to the absolute value of the amplitude, and the vertical axis is set to the frequency.

[0104] As can be seen from the above explanation, Figure 4 The data near zero in the distribution are power cycle synchronization noise data, and the data above the specified amplitude are estimated to be signals originating from partial discharge. This specified amplitude corresponds to the amplitude threshold TA in this embodiment.

[0105] In addition, the frequency distribution of the amplitude data of the differential waveform becomes Figure 5 That way.

[0106] To determine the amplitude threshold TA, in this embodiment, based on Figure 4 We assume that the interference noise in the frequency distribution of the differential waveform can approximate a steep normal distribution. Then, we treat the distribution near zero in the differential waveform data as a normal distribution. Furthermore, using the standard deviation σ of this normal distribution and the set multiplier X, we calculate the amplitude threshold TA.

[0107] In this embodiment, for example, the multiplier X is set to 3. In this case, the amplitude equivalent to 3σ in the differential waveform data, where the distribution near zero is considered a normal distribution, is set as the amplitude threshold TA (absolute value).

[0108] Then, for the waveform f(t) before the difference is taken, it is determined that partial discharge was observed at a time t that satisfies both f(t) - f(t+τ) and f(t-τ) above the calculated amplitude threshold TA, and this time is extracted as a partial discharge signal. Thus, in this embodiment, a partial discharge signal with significantly reduced interference noise can be obtained.

[0109] Therefore, in this embodiment, signals originating from partial discharge can be extracted with superior accuracy compared to the past.

[0110] Furthermore, in this embodiment, the occurrence of partial discharge in the device 1 is monitored by extracting signals originating from partial discharge with excellent accuracy. As a result, the accuracy of the monitoring is improved.

[0111] Figure 6 as well as Figure 7 Examples are given for obtaining the observed waveform after detection processing, the periodic analysis results of the observed waveform, the correlation diagram of the observed waveform, the extraction of only the waveform of the discharge (the extracted waveform of the partial discharge signal), and the PD pulse ratio as an evaluation example. Figure 6 This is an example of a partial discharge exceeding the specified limit. Figure 7 This is an example of a situation where partial discharge occurs on a small scale.

[0112] Such as Figure 6 (In cases where partial discharge exceeding the specified level occurs) and Figure 7 (In cases where partial discharge occurs only slightly), the results of periodic analysis (using conventional methods for observing power supply periodicity in waveform observation) show periodicity that is an integer multiple of the power supply frequency (60Hz). Therefore, it can be seen that even when the presence or absence of partial discharge cannot be distinguished, it is possible to distinguish the presence or absence of partial discharge by using the PD pulse ratio. Thus, it can be seen that by using the PD pulse ratio, the occurrence of partial discharge in the inspected electrical equipment 1 can be monitored with excellent accuracy.

[0113] (Effect)

[0114] This implementation method, for example, achieves the following effects.

[0115] (1) The partial discharge monitoring method of this embodiment detects the physical quantity generated when partial discharge occurs in the inspected electrical equipment as an electrical signal. From the detected electrical signal, the time change of the amplitude of the signal waveform generated by partial discharge is extracted as a detection processing signal. Based on the signal measured at each time and the signal that is offset from the power supply frequency of the electrical equipment at that time by m periods (m: an integer greater than or equal to 1), it is determined whether the signal at that time is a partial discharge signal originating from partial discharge. The partial discharge signal is extracted from the electrical signal or the detection processing signal. Based on the extracted partial discharge signal, the partial discharge status of the inspected electrical equipment is evaluated.

[0116] For example, the partial discharge monitoring device 10 of this embodiment includes: a sensor unit capable of detecting physical quantities generated when partial discharge occurs in the inspected electrical equipment as electrical signals; a detection processing unit capable of extracting the time change of the amplitude of the signal waveform generated by partial discharge from the electrical signals detected by the sensor unit as a detection processing signal; a partial discharge signal extraction unit capable of determining, based on the signals measured at each time and the signals that are offset from the power supply frequency of the electrical equipment by m periods (m: an integer greater than or equal to 1) at that time, whether the signal at that time is a partial discharge signal originating from partial discharge, thereby extracting a partial discharge signal from the electrical signals or the detection processing signals; and a partial discharge evaluation unit capable of evaluating the occurrence of partial discharge in the inspected electrical equipment based on the partial discharge signal extracted by the partial discharge signal extraction unit.

[0117] According to this structure, the amount of interference noise is separated from the measured waveform, thereby enabling the extraction of partial discharge data with significantly high accuracy. As a result, in this embodiment, partial discharge signals that are easily mistaken for interference noise such as magnetostrictive sound that occurs synchronously with the power supply can be detected with high accuracy. Therefore, it becomes possible to monitor the occurrence of partial discharge.

[0118] In the above embodiment, the determination of whether or not the partial discharge occurs is based on the amplitude difference between the signal measured at the determination time and the signal that is offset from the power supply frequency of the electrical device by m periods (m: an integer greater than or equal to 1) relative to that time.

[0119] For example, the partial discharge signal extraction unit described above may be configured as follows: it includes: a differential waveform calculation unit that calculates a differential waveform for the detected signal processed by the detection processing unit, which is composed of the difference between the amplitude of the detected signal and the amplitude of a signal whose amplitude is offset from the power supply frequency of the electrical equipment by m periods (m: an integer greater than or equal to 1); an amplitude threshold setting unit that sets an amplitude threshold based on the differential waveform calculated by the differential waveform calculation unit, which is used to divide the amplitude of the differential waveform into small-amplitude interference noise signals and partial discharge signals whose amplitude is larger than the interference noise signals; and a signal extraction main unit that extracts the moment exceeding the amplitude threshold from the differential waveform calculated by the differential waveform calculation unit, and extracts the signal corresponding to the extracted moment as a partial discharge signal from the electrical signal detected by the sensor unit or the detected signal processed by the detection processing unit.

[0120] (2) The amplitude threshold setting unit regards the distribution near zero in the amplitude distribution of the differential waveform data calculated by the differential waveform calculation unit as a normal distribution, and sets the amplitude threshold based on the standard deviation σ of the normal distribution.

[0121] Based on this structure, noise components can be separated with greater accuracy and quality.

[0122] (3) The amplitude threshold setting unit sets the value of the above standard deviation σ multiplied by a multiplier X selected in the range of 2 or more and 4 or less as the above amplitude threshold.

[0123] Based on this structure, more accurate and superior partial discharge signals can be extracted.

[0124] Furthermore, even when the aforementioned magnification X is outside the range of 2 to 4, partial discharge signals can still be extracted with excellent accuracy compared to the past.

[0125] (4) The aforementioned physical quantities are, for example, changes in TEV (transient ground voltage) caused by partial discharge in the inspected electrical equipment 1, current, electromagnetic waves, or ultrasonic waves. That is, the aforementioned sensor unit 10A is, for example, a TEV sensor that detects changes in TEV (transient ground voltage) caused by partial discharge in the inspected electrical equipment 1, an HFCT sensor that detects current caused by partial discharge in the inspected electrical equipment 1, an electromagnetic wave sensor that detects electromagnetic waves, or an ultrasonic sensor that detects ultrasonic waves caused by partial discharge in the inspected electrical equipment 1.

[0126] According to this structure, the physical quantities generated when partial discharge occurs in the inspected electrical equipment 1 can be reliably detected as electrical signals.

[0127] (Modified Example)

[0128] <Variation Example 1>

[0129] In the above description, the amplitude threshold TA is set to a value determined based on a preset multiplier X and the standard deviation σ of the normal distribution, but it is not limited to this.

[0130] For example, in Figure 4 In the frequency distribution of the absolute values ​​of the data, the amplitude value that moderates the steepness of the gradient of the estimated distribution can be simply set as the amplitude threshold TA.

[0131] <Variation Example 2>

[0132] Alternatively, partial discharge signals can be extracted as follows.

[0133] In the above implementation, the differential waveform is obtained, and the amplitude distribution near zero of the differential waveform is assumed to be a normal distribution. Based on this normal distribution, the standard deviation σ and the amplitude threshold TA are determined, and the partial discharge signal is extracted.

[0134] In contrast, in Modified Example 2, the partial discharge signal is extracted as follows.

[0135] That is, the extraction of the partial discharge signal is performed as follows: For the detection signal, a differential waveform is calculated, consisting of the difference between the amplitude of the detection signal and the amplitude of a signal whose amplitude is offset from the power supply frequency of the electrical equipment by m periods (m: an integer greater than or equal to 1). Based on the calculated differential waveform, an amplitude threshold is set to distinguish the amplitudes of small-amplitude interference noise signals and partial discharge signals with larger amplitudes in the differential waveform. In the calculated differential waveform, moments exceeding the amplitude threshold are extracted, and the signal corresponding to the extracted moment is extracted from the electrical signal or the detection signal as a partial discharge signal.

[0136] For example, the partial discharge signal extraction unit 10C described above can be configured such that, for the detected signal processed by the detection processing unit, data or a function can be obtained expressing the signal amplitude measured at each time and the signal amplitude of a period m (m: an integer greater than or equal to 1) that is offset from the power supply frequency of the electrical equipment relative to that time. Then, based on the data or function expressing the scatter plot, if it is determined that the signal at the time of determination is in a predetermined region of the scatter plot, the signal is determined to be a partial discharge signal.

[0137] At this point, regarding the aforementioned predetermined region, for example, the frequency distribution of the detected signal processed by the aforementioned detection processing unit 10B is calculated, and the frequency distribution within the amplitude range below the maximum frequency amplitude μ is approximated as a normal distribution. Then, an amplitude threshold is set based on the standard deviation σ of this normal distribution. Then, the following regions are defined: regions where the signal amplitude measured at each time is above the amplitude threshold and the signal amplitude after a period of m is below the amplitude threshold; and regions where the signal amplitude measured at each time is below the amplitude threshold and the signal after a period of m is above the amplitude threshold.

[0138] For example, consider the following structure: if the combination (a(t), b(t)) of the signal amplitude a(t) measured at each time moment and the signal amplitude b(t) offset by m periods (=a(t+τ)) satisfies a predetermined condition, then the signal at that time is identified as a partial discharge signal. Furthermore, for example, regarding the aforementioned predetermined condition, the frequency distribution of the detected signal processed by the aforementioned detection processing unit is calculated, and the frequency distribution within the amplitude range below the maximum frequency amplitude μ is approximately a normal distribution. Then, based on the standard deviation of this normal distribution, an amplitude threshold is set. Then, it is defined as a combination where the signal amplitude a(t) measured at each time moment is above the amplitude threshold and the signal amplitude b(t) offset by m periods is less than the amplitude threshold, and a combination where the signal amplitude a(t) measured at each time moment is less than the amplitude threshold and the signal amplitude b(t) offset by m periods is above the amplitude threshold.

[0139] In addition, for example, the amplitude threshold can be set as the value obtained by multiplying the standard deviation σ by a multiplier selected from a range of 2 to 4, and adding the amplitude μ, which is the maximum frequency.

[0140] That is, in this variation 2, the processing is performed as follows.

[0141] First, the physical quantity generated when partial discharge occurs in the inspected electrical equipment 1 is used as an electrical signal for detection. For the detected electrical signal, the low-frequency components are removed by a high-pass filter, and then the detection waveform is extracted.

[0142] Next, the amplitude distribution of the detected waveform is determined, and the amplitude distribution of amplitudes lower than the most frequent amplitude is approximated as one side of a normal distribution (refer to...). Figure 8 ).

[0143] Next, the amplitude calculated based on the obtained approximate normal distribution using a pre-defined setting method (μ+nσ) is set as the amplitude threshold.

[0144] Next, considering the signal after detection processing and the signal delayed by one cycle of the power supply frequency, consider as follows: Figure 9 Such a scatter plot. In this case, the points (data) contained in any region where the original waveform is above the amplitude threshold and the waveform is below the amplitude threshold after a 1-period shift, and region 1 where the original waveform is below the amplitude threshold and the waveform is above the amplitude threshold after a 1-period shift, are extracted as the partial discharge signal contained in the original waveform. Regions 1 and 2 are in Figure 9 The area in the middle is indicated by a shaded line.

[0145] Compared to the partial discharge signals extracted in the embodiment, the proportion of partial discharge signals misidentified as interference noise increased in Modification 2. However, the partial discharge signals extracted in Modification 2 also removed the main interference noise, and the extraction accuracy was significantly improved compared to the prior art.

[0146] The entire contents of Japanese Patent Application 2020-150801 (filed September 8, 2020), which claims priority in this application, constitute a part of this disclosure by reference. A limited number of embodiments have been described herein, but the scope of the claims is not limited thereto, and modifications to the embodiments disclosed above will be readily apparent to those skilled in the art.

[0147] Explanation of reference numerals in the attached figures

[0148] 1: Electrical equipment; 10: Partial discharge monitoring device; 10A: Sensor unit; 10B: Detection and processing unit; 10C: Partial discharge signal extraction unit; 10CA: Differential waveform calculation unit; 10CB: Amplitude threshold setting unit; 10CC: Signal extraction main unit; 10D: Partial discharge evaluation unit; TA: Amplitude threshold; X: Multiplier.

Claims

1. A partial discharge monitoring device, characterized in that, have: The sensor unit is capable of detecting the physical quantities generated when partial discharge occurs in the inspected electrical equipment as electrical signals. The detection processing unit performs a process to extract the time change of the amplitude of the signal waveform caused by partial discharge from the electrical signal detected by the sensor unit as a detection processing signal. The partial discharge signal extraction unit, taking the detected signal processed by the aforementioned detection processing unit, determines whether the signal at that moment is a partial discharge signal originating from partial discharge, based on the signal measured at each time and a signal with m periods offset from the power supply frequency of the aforementioned electrical equipment relative to that time. It then extracts the partial discharge signal from the aforementioned electrical signal or the aforementioned detected signal, where m is an integer of 1 or more. The partial discharge evaluation unit evaluates the partial discharge status of the inspected electrical equipment based on the partial discharge signal extracted by the partial discharge signal extraction unit. The aforementioned partial discharge signal extraction unit includes: The differential waveform calculation unit calculates a differential waveform for the detection processing signal processed by the detection processing unit, which is composed of the difference between the amplitude of the detection processing signal and the amplitude of a signal with m periods offset from the power supply frequency of the electrical equipment, where m is an integer greater than or equal to 1. The amplitude threshold setting unit sets an amplitude threshold based on the differential waveform calculated by the differential waveform calculation unit, which serves to divide the amplitude of small-amplitude interference noise signals and partial discharge signals with larger amplitudes than the interference noise signals in the differential waveform; and The signal extraction main unit extracts the moment that exceeds the amplitude threshold from the differential waveform calculated by the differential waveform calculation unit, and extracts the signal corresponding to the extracted moment as a partial discharge signal from the electrical signal detected by the sensor unit or the detection processing signal processed by the detection processing unit.

2. The partial discharge monitoring device according to claim 1, characterized in that, The amplitude threshold setting unit considers the distribution near zero in the amplitude distribution of the differential waveform calculated by the differential waveform calculation unit as a normal distribution, and sets the amplitude threshold based on the standard deviation of the normal distribution.

3. The partial discharge monitoring device according to claim 2, characterized in that, The amplitude threshold setting unit sets the value obtained by multiplying the standard deviation by a multiplier selected in the range of 2 to 4 as the amplitude threshold.

4. A partial discharge monitoring device, characterized in that, have: The sensor unit is capable of detecting the physical quantities generated when partial discharge occurs in the inspected electrical equipment as electrical signals. The detection processing unit performs a process to extract the time change of the amplitude of the signal waveform caused by partial discharge from the electrical signal detected by the sensor unit as a detection processing signal. The partial discharge signal extraction unit, taking the detected signal processed by the aforementioned detection processing unit, determines whether the signal at that moment is a partial discharge signal originating from partial discharge, based on the signal measured at each time and a signal with m periods offset from the power supply frequency of the aforementioned electrical equipment relative to that time. It then extracts the partial discharge signal from the aforementioned electrical signal or the aforementioned detected signal, where m is an integer of 1 or more. The partial discharge evaluation unit evaluates the partial discharge status of the inspected electrical equipment based on the partial discharge signal extracted by the partial discharge signal extraction unit. The partial discharge signal extraction unit calculates data or a function for the detected signal processed by the detection processing unit, which expresses the signal amplitude measured at each time and the signal amplitude of the signal amplitude with m periods offset from the power supply frequency of the electrical equipment relative to that time. If, based on the data or function expressing the scatter plot, it is determined that the signal at the time of determination is in a predetermined region of the scatter plot, then the signal is determined to be a partial discharge signal, where m is an integer of 1 or more.

5. The partial discharge monitoring device according to claim 4, characterized in that, The aforementioned predetermined region is defined as follows: the frequency distribution of the detected signal processed by the aforementioned detection processing unit is calculated, and the frequency distribution within the amplitude range below the maximum frequency amplitude μ is approximately a normal distribution. An amplitude threshold is set based on the standard deviation of this normal distribution. The regions where the signal amplitude measured at each time is above the amplitude threshold and is offset by m periods are regions where the signal amplitude is less than the amplitude threshold, and the regions where the signal amplitude measured at each time is less than the amplitude threshold and is offset by m periods are regions where the signal amplitude is above the amplitude threshold.

6. The partial discharge monitoring device according to claim 5, characterized in that, The amplitude threshold is set as the sum of the value obtained by multiplying the standard deviation by a factor selected from the range of 2 to 4 and the amplitude μ that becomes the maximum frequency.

7. The partial discharge monitoring device according to any one of claims 1 to 6, characterized in that, The aforementioned sensor unit is any one of the following: a TEV sensor that detects transient ground voltage changes (TEVs) generated from the electrical equipment being inspected; a high-frequency CT sensor that measures current pulses flowing to the grounding wire; an electromagnetic wave sensor that detects electromagnetic waves generated along with partial discharge; or an ultrasonic sensor that detects ultrasonic waves generated from the electrical equipment being inspected.

8. A method for monitoring partial discharge, characterized in that, The physical quantities generated when partial discharge occurs in the inspected electrical equipment are used as electrical signals for detection. From the detected electrical signals, the time variation of the amplitude of the signal waveform caused by partial discharge is extracted as a detection processing signal. Based on the extracted detection processing signal, and considering the signal measured at each time point and a signal with m periods offset from the power supply frequency of the electrical equipment relative to that time point, it is determined whether the signal at that time is a partial discharge signal originating from partial discharge. The partial discharge signal is then extracted from the electrical signals or the detection processing signal, where m is an integer greater than or equal to 1. Based on the extracted partial discharge signals, the partial discharge status of the inspected electrical equipment is evaluated. The determination of whether or not partial discharge occurs is based on the amplitude difference between the signal measured at each moment and the signal with m periods offset from the power supply frequency of the aforementioned electrical equipment relative to that moment, where m is an integer greater than or equal to 1. The extraction of the above partial discharge signals is performed as follows: For the aforementioned detection signal, a differential waveform is calculated, which is formed by the difference between the amplitude of the detection signal and the amplitude of a signal with m periods that are offset from the power supply frequency of the aforementioned electrical equipment. Based on the calculated differential waveform, an amplitude threshold is set to divide the amplitude of the interference noise signal with a small amplitude and the amplitude of the partial discharge signal with a larger amplitude than the interference noise signal in the differential waveform. In the calculated differential waveform, the moment that exceeds the amplitude threshold is extracted. From the aforementioned electrical signal or the aforementioned detection signal, the signal corresponding to the extracted moment is extracted as the partial discharge signal, where m is an integer greater than or equal to 1.

9. A method for monitoring partial discharge, characterized in that, The physical quantities generated when partial discharge occurs in the inspected electrical equipment are used as electrical signals for detection. From the detected electrical signals, the time variation of the amplitude of the signal waveform caused by partial discharge is extracted as a detection processing signal. Based on the extracted detection processing signal, and considering the signal measured at each time point and a signal with m periods offset from the power supply frequency of the electrical equipment relative to that time point, it is determined whether the signal at that time is a partial discharge signal originating from partial discharge. The partial discharge signal is then extracted from the electrical signals or the detection processing signal, where m is an integer greater than or equal to 1. Based on the extracted partial discharge signals, the partial discharge status of the inspected electrical equipment is evaluated. The above discrimination is performed as follows: For the above-mentioned detection processed signal, data or functions are obtained to express the signal amplitude measured at each time and the signal amplitude of the signal amplitude that is offset from the power supply frequency of the above-mentioned electrical equipment by m periods relative to that time. If, based on the data or functions expressing the scatter plot, the signal amplitude at the time of discrimination is in a predetermined region of the above-mentioned scatter plot, the signal at that time is discriminated as a partial discharge signal, where m is an integer greater than or equal to 1.

10. The partial discharge monitoring method according to claim 9, characterized in that, The aforementioned predetermined region is defined as follows: the frequency distribution of the above-mentioned detection processed signal is calculated, and the frequency distribution within the amplitude range below the maximum frequency amplitude μ is approximately a normal distribution. Based on the standard deviation of this normal distribution, an amplitude threshold is set. The regions where the signal amplitude measured at each time is above the amplitude threshold and is offset by m periods are regions where the signal amplitude is less than the amplitude threshold, and the regions where the signal amplitude measured at each time is less than the amplitude threshold and is offset by m periods are regions where the signal amplitude is above the amplitude threshold.

Images