Signal processor and signal processing method

Abstract

To provide a signal processor and signal processing method that can discriminate a partial discharge signal from noise.SOLUTION: A signal processor includes a clustering section and an identification section. The clustering section distributes a plurality of pulse signals measured from a target device into a plurality of clusters by clustering based on a waveform of each pulse signal. The identification section identifies, on the basis of a distribution of power supply phase with respect to the target device when the pulse signal distributed to each of the plurality of clusters is generated, from the plurality of clusters, a partial discharge cluster to which a pulse signal related to partial discharge is distributed or a noise cluster to which a pulse signal related to noise is distributed.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] Embodiments of the present invention relate to a signal processing apparatus and a signal processing method.

Background Art

[0002] In power equipment such as a switchgear, when the insulation part deteriorates, partial discharge occurs from the power equipment, and when the deterioration further progresses, insulation breakdown occurs. Therefore, it is known to diagnose power equipment by detecting partial discharge, which is a sign of insulation breakdown. On the other hand, since noise exists in the installation environment of power equipment, the measurement signal measured by a sensor or the like includes a partial discharge signal emitted from the power equipment and noise.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The problem to be solved by the present invention is to provide a signal processing apparatus and a signal processing method capable of discriminating between a partial discharge signal and noise.

Means for Solving the Problems

[0005] The signal processing apparatus according to the embodiment has a clustering unit and an identifying unit. The clustering unit distributes a plurality of pulse signals measured from a target device into a plurality of clusters by clustering based on the waveforms of the respective pulse signals. The identifying unit identifies, from the plurality of clusters, a partial discharge cluster in which a pulse signal related to partial discharge is distributed or a noise cluster in which a pulse signal related to noise is distributed, based on the distribution of the power supply phase with respect to the target device when the pulse signals distributed to the respective plurality of clusters are generated.

Brief Description of the Drawings

[0006]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Embodiments for Carrying Out the Invention

[0007] Hereinafter, the signal processing apparatus and the signal processing method of the embodiment will be described with reference to the drawings.

[0008] (First Embodiment) FIG. 1 is a diagram showing the appearance of a signal processing apparatus 100 according to a first embodiment. The signal processing apparatus 100 according to the first embodiment monitors the occurrence of partial discharge in a target power device 200, which is an electrical device to be monitored, and diagnoses the deterioration of the target power device 200. The signal processing apparatus 100 diagnoses the deterioration of the target power device 200 using a machine learning model. Noise exists in the installation environment of the target power device 200. The noise varies depending on the load conditions and the like. Therefore, the signal processing apparatus 100 monitors the partial discharge of the target power device 200 after learning the noise in the installation environment.

[0009] The target power device 200 includes a box body and a device main body. The device main body is housed in the grounded box body. The device main body is a device that may generate partial discharge, such as a switchgear, a power transformer, a gas-insulated switch, a generator, a motor, a capacitor, or a reactor. The device main body is composed of devices such as a disconnector, a circuit breaker, a current transformer, or a transformer. An electrode 111 is attached to the box body of the target power device 200. The electrode 111 and the signal processing apparatus 100 are connected by a cable. The signal processing apparatus 100 measures the voltage of the electrode 111 to measure the potential formed on the surface of the box body through the stray capacitance between the box body and the device main body. That is, the electrode 111 according to the first embodiment is a TEV (Transient Earth Voltage) sensor. On the other hand, the electrode 111 according to other embodiments may be other electrodes 111 that measure electromagnetic signals derived from the operation of electrical devices such as a CT (Current Transformer) sensor, an AE (Acoustic Emission) sensor, or an antenna. The electrode 111 may be installed inside the box body of the target power device 200.

[0010] FIG. 2 is a block diagram showing the software configuration of the signal processing apparatus 100 according to the first embodiment. The signal processing apparatus 100 includes an acquisition unit 101, a smoothing unit 102, a signal storage unit 103, a pulse extraction unit 104, a clustering unit 105, a feature amount calculation unit 106, an identification unit 107, a diagnosis unit 108, and an output control unit 109.

[0011] The acquisition unit 101 acquires the measurement signal measured by the electrode 111. The measurement signal may include a partial discharge signal generated from the target power equipment 200 and noise in the environment of the target power equipment 200. The smoothing unit 102 smooths the measurement signal acquired by the acquisition unit 101. For example, the smoothing unit 102 smooths the measurement signal by applying it to a low-pass filter. The signal storage unit 103 stores the smoothed measurement signal.

[0012] The pulse extraction unit 104 extracts a pulse signal from the measurement signal stored in the signal storage unit 103. Specifically, the pulse extraction unit 104 extracts the pulse signal according to the following procedure. First, the pulse extraction unit 104 obtains the average of the absolute value of the voltage related to the smoothed measurement signal as a threshold value. The pulse extraction unit 104 identifies, from the smoothed measurement signal, the maximum value of the portion where the absolute value of the voltage exceeds the threshold value as the peak of the pulse. Then, the pulse extraction unit 104 extracts the signal within a predetermined period centered on the identified peak as the pulse signal. Note that the method of extracting the pulse is not limited to this.

[0013] The clustering unit 105 clusters the plurality of pulse signals extracted by the pulse extraction unit 104 based on their waveforms. Examples of the clustering method include the k-means method. The number of clusters related to the clustering unit 105 may be determined in advance or may be specified by the user.

[0014] The feature quantity calculation unit 106 calculates the feature quantity of the cluster based on the pulse signals classified into the cluster. Specifically, the feature quantity calculation unit 106 calculates, as the feature quantity of the cluster, the phase dispersion degree, the envelope area, and the envelope area ratio. The phase dispersion degree is obtained by the standard deviation of the power supply phase with respect to the target power equipment 200 when the pulse signal assigned to the cluster is generated. The power supply phase with respect to the target power equipment 200 may be the phase of the power supply voltage applied to the target power equipment 200, or may be the phase of the power supply voltage obtained from an outlet (wiring plug connector) connected to the target power equipment 200 or an outlet in the vicinity thereof. The envelope area and the envelope area ratio are parameters representing the characteristics of the waveform of the pulse signal representing the class. In the first embodiment, the pulse signal representing the class is the average waveform of a plurality of pulse signals belonging to the class. The envelope area is a parameter representing the magnitude of the pulse signal, and is the integral value of the envelope line of the pulse signal representing the class. The envelope area ratio is a parameter representing the distortion of the pulse signal, and is obtained by dividing the integral value of the envelope line of the first half when the pulse signal representing the cluster is divided into the first half and the second half with the peak as the center by the integral value of the entire envelope line (envelope area).

[0015] FIG. 3 is a diagram showing an example of the envelope line of the pulse signal according to the first embodiment. That is, when the integral value of the envelope line of the first half of the pulse signal representing the class is set as A and the integral value of the envelope line of the second half is set as B, the envelope area is represented by (A + B), and the envelope area ratio is represented by (A / (A + B)).

[0016] Based on the phase dispersion degree calculated by the feature quantity calculation unit 106, the envelope area, and the envelope area ratio, the specific unit 107 classifies each of the plurality of clusters into a partial discharge cluster that is a set of partial discharge signals and a noise cluster that is a set of noises. That is, the specific unit 107 identifies the partial discharge cluster from among the plurality of clusters and also identifies the noise cluster from among the plurality of clusters. Specifically, the specific unit 107 classifies the clusters according to the following procedure. The specific unit 107 identifies a cluster whose phase dispersion degree exceeds a predetermined threshold value (for example, 20 degrees) as a noise cluster. The specific unit 107 identifies a cluster that satisfies an identification inequality, which is a linear inequality for identifying a noise cluster, in the relationship between the envelope area and the envelope area ratio, as a noise cluster. The specific unit 107 identifies a cluster that is not classified as a noise cluster as a partial discharge cluster. That is, the specific unit 107 identifies a cluster whose phase dispersion degree does not exceed a predetermined threshold value and whose relationship between the envelope area and the envelope area ratio does not satisfy the identification inequality as a partial discharge cluster. FIG. 4 is a diagram showing an example of the identification inequality according to the first embodiment. Note that the identification inequality is calculated in advance based on the partial discharge data acquired in the past. When the envelope area is set as X and the envelope area ratio is set as Y, the identification inequality is represented by (Y > -aX + b: a > 0, b > 0).

[0017] Explain the reason why the noise cluster can be identified by the phase dispersion degree. It is known that the polarity of partial discharge corresponds to the polarity of the power supply voltage. Therefore, the polarity of the pulse signal related to partial discharge switches every half cycle of the power supply voltage. Here, since the clustering unit 105 performs clustering based on the waveform, pulse signals with different polarities are assigned to different clusters. Therefore, the pulse signals belonging to the partial discharge cluster have a small variation in the phase of the power supply voltage (within at least 180 degrees). On the other hand, since noise does not depend on the power supply voltage, the polarity of noise is independent of the period of the power supply voltage. Therefore, the pulse signals belonging to the noise cluster can have a large variation in the phase of the power supply voltage.

[0018] Explain the reason why the noise cluster can be identified by the envelope area and the envelope area ratio. Generally, it is known that the pulse signal related to partial discharge tends to have a fast rising edge in the first half and a gentle falling edge in the second half. On the other hand, noise seldom shows a difference between the rising edge and the falling edge. From this, the envelope area ratio of the pulse signal related to partial discharge is smaller than that of the pulse signal related to noise. Also, noise often has a gentle rising and a gentle falling waveform. Therefore, the envelope area of the pulse signal related to noise is larger than that of the pulse signal related to partial discharge. Thus, when the envelope area or the envelope area ratio is large, there is a high possibility of partial discharge. That is, the specific part 107 can identify the noise cluster based on the envelope area, the envelope area ratio, and the discrimination inequality.

[0019] The diagnosis unit 108 diagnoses the state of the target power equipment 200 based on the pulse signals assigned to the partial discharge cluster. For example, the diagnosis unit 108 may estimate the presence or absence of partial discharge, or may estimate the deterioration state and the cause of deterioration of the target power equipment 200 based on the ΦQN pattern calculated from the pulse signals belonging to the partial discharge cluster.

[0020] The output control unit 109 outputs the diagnosis result by the diagnosis unit 108 to an external device, a display device, etc. The display device may be provided in the signal processing device 100.

[0021] FIG. 5 is a flowchart showing the diagnosis process of the signal processing device 100 according to the first embodiment. The acquisition unit 101 of the signal processing device 100 acquires a measurement signal from the electrode 111 while the target power equipment 200 is operating. The smoothing unit 102 smooths the acquired measurement signal and records it in the signal storage unit 103. As a result, the signal storage unit 103 stores the pre-smoothed measurement signal in advance.

[0022] When the signal processing device 100 starts the diagnosis process according to an instruction from the administrator or the like, the pulse extraction unit 104 extracts a plurality of pulse signals from the measurement signals stored in the signal storage unit 103 (step S1).

[0023] The clustering unit 105 clusters the plurality of pulse signals extracted in step S1 based on the waveform (step S2). The signal processing apparatus 100 selects each of the clusters obtained as a result of the clustering (step S3), and executes the following steps S4 to S10.

[0024] The feature amount calculation unit 106 obtains the average waveform of the cluster selected in step S3 (step S4). The feature amount calculation unit 106 passes the obtained average waveform through a band-pass filter and performs a Hilbert transform to obtain the envelope of the average waveform (step S5). The band of the band-pass filter may be from 5 MHz to 30 MHz. The feature amount calculation unit 106 calculates the phase dispersion degree, the envelope area, and the envelope area ratio for the cluster selected in step S3 (step S6).

[0025] The specifying unit 107 determines whether or not the phase dispersion degree calculated in step S5 exceeds a predetermined threshold value (step S7). When the phase dispersion degree exceeds the threshold value (step S7: YES), the specifying unit 107 specifies the cluster selected in step S3 as a noise cluster (step S8).

[0026] When the phase dispersion degree does not exceed the threshold value (step S7: NO), the specifying unit 107 determines whether or not the relationship between the envelope area and the envelope area ratio satisfies the discrimination inequality (step S9). When the relationship between the envelope area and the envelope area ratio satisfies the discrimination inequality (step S9: YES), the specifying unit 107 specifies the cluster selected in step S3 as a noise cluster (step S8).

[0027] On the other hand, when the phase dispersion degree does not exceed the predetermined threshold value and the relationship between the envelope area and the envelope area ratio does not satisfy the discrimination inequality (step S9: NO), the specifying unit 107 specifies the cluster selected in step S3 as a partial discharge cluster (step S10).

[0028] When the signal processing device 100 identifies whether each cluster is a partial discharge cluster or a noise cluster, based on the pulse signals assigned to the clusters identified as partial discharge clusters in step S10, the state of the target power equipment 200 is diagnosed (step S11). The output control unit 109 outputs the diagnosis result of step S11 (step S12).

[0029] As described above, the signal processing device 100 according to the first embodiment performs the following processing. The clustering unit 105 of the signal processing device 100 distributes a plurality of pulse signals measured from the target power equipment 200 to a plurality of clusters by clustering based on the waveforms of the respective pulse signals. The specifying unit 107 specifies a partial discharge cluster or a noise cluster from the plurality of clusters based on the distribution of the power supply phases of the target power equipment 200 when the pulse signals distributed to each of the plurality of clusters are generated. Thereby, the signal processing device 100 can discriminate between partial discharge signals and noise.

[0030] FIG. 6 is a diagram showing the clustering result of pulse signals by the experiment according to the first embodiment. The inventors discriminated the pulse signals of the switchgear, which is a power equipment, using the signal processing device 100 according to the first embodiment. In the experiment, the signal processing device 100 extracted 2266 pulse signals from the measurement signals of the power equipment and executed clustering to assign them to 20 clusters. As a result, the signal processing device 100 divided the pulse signals into clusters as shown in FIG. 6. As shown in FIG. 6, it can be seen that the average waveforms of the respective clusters have different characteristics.

[0031] FIG. 7 is a diagram showing the relationship between the clusters and the discrimination inequality in the experiment according to the first embodiment. In the experiment, the signal processing device 100 obtained the envelope area X and the envelope area ratio Y from the average waveforms of the respective clusters and determined whether or not the discrimination inequality was satisfied. As shown in FIG. 7, the clusters that did not satisfy the discrimination inequality (clusters that may be partial discharge clusters) were cluster 1, cluster 4, cluster 11, and cluster 19.

[0032] FIG. 8 is a diagram showing the relationship between the power supply phase of the pulse signal and the clusters in the experiment according to the first embodiment. FIG. 9 is a diagram showing the phase dispersion degree for each cluster in the experiment according to the first embodiment. In the experiment, the signal processing device 100 obtained the phase dispersion degree of each cluster and determined whether the phase dispersion degree exceeded a threshold value (20 degrees). As shown in FIG. 9, the clusters with a phase dispersion degree of 20 degrees or less (clusters that may be partial discharge clusters) were cluster 12 and cluster 19. From these facts, in the experiment, cluster 19 was identified as a partial discharge cluster. When the inventor confirmed the pulse signal belonging to cluster 19, it was confirmed that the partial discharge signal was actually included in cluster 19.

[0033] FIG. 10 is a schematic block diagram showing the computer configuration of the signal processing device 100 according to the first embodiment. The computer 500 includes a processor 510, a main memory 520, a storage 530, and an interface 540. The above-described signal processing device 100 is implemented in the computer 500. And the operations of the above-described respective processing units are stored in the storage 530 in the form of a program. The processor 510 reads the program from the storage 530, expands it in the main memory 520, and executes the above processing according to the program. Also, the processor 510 secures a storage area corresponding to each of the above-described storage units in the main memory 520 according to the program. Examples of the processor 510 include a CPU (Central Processing Unit), a GPU (Graphic Processing Unit), a microprocessor, and the like.

[0034] The program may be for realizing a part of the functions to be exerted on the computer 500. For example, the program may exert functions in combination with other programs already stored in the storage, or in combination with other programs implemented in other devices. In other embodiments, the computer 500 may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) in addition to or instead of the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, part or all of the functions realized by the processor 510 may be realized by the integrated circuit. Such an integrated circuit is also included as an example of a processor. In other embodiments, the computer 500 may be virtualized on one or more computers.

[0035] Examples of the storage 530 include a magnetic disk, a magneto-optical disk, an optical disk, a semiconductor memory, etc. The storage 530 may be an internal medium directly connected to the bus of the computer 500, or an external medium connected to the computer 500 via the interface 540 or a communication line. Also, when this program is distributed to the computer 500 via a communication line, the computer 500 that has received the distribution may expand the program in the main memory 520 and execute the above processing. In at least one embodiment, the storage 530 is a non-transitory tangible storage medium.

[0036] Also, the program may be for realizing a part of the functions described above. Furthermore, the program may be a so-called differential file (differential program) that realizes the above-described functions in combination with other programs already stored in the storage 530.

[0037] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and its equivalent scope.

[0038] The signal processing apparatus 100 according to the first embodiment classifies a plurality of clusters into partial discharge clusters and noise clusters based on the phase dispersion degree, the envelope area, and the envelope area ratio, but is not limited thereto.

[0039] For example, the signal processing apparatus 100 according to another embodiment may classify a plurality of clusters into partial discharge clusters and noise clusters based on the phase dispersion degree without using the envelope area and the envelope area ratio. Specifically, the signal processing apparatus 100 according to another embodiment may identify a cluster with a phase dispersion degree exceeding a threshold as a noise cluster, and identify a cluster with a phase dispersion degree not exceeding the threshold as a partial discharge cluster. Also in this case, the accuracy may be inferior compared to the first embodiment, but a plurality of clusters can be classified into partial discharge clusters and noise clusters. Further, the phase dispersion degree according to another embodiment does not necessarily have to be a standard deviation, and may be represented by another dispersion degree.

[0040] For example, the signal processing apparatus 100 according to another embodiment may classify a plurality of clusters into partial discharge clusters and noise clusters based on the envelope area and the envelope area ratio without using the phase dispersion degree. Specifically, the signal processing apparatus 100 according to another embodiment may identify a cluster whose relationship between the envelope area and the envelope area ratio satisfies the discrimination inequality as a noise cluster, and identify a cluster whose relationship between the envelope area and the envelope area ratio does not satisfy the discrimination inequality as a partial discharge cluster. Also in this case, although the accuracy may be inferior compared to the first embodiment, a plurality of clusters can be classified into partial discharge clusters and noise clusters.

[0041] For example, the signal processing apparatus 100 according to another embodiment may identify a noise cluster using only one of the envelope area and the envelope area ratio. Specifically, the signal processing apparatus 100 according to another embodiment may identify a cluster whose envelope area exceeds a predetermined threshold as a noise cluster regardless of the envelope area ratio, or may identify a cluster whose envelope area ratio exceeds a predetermined threshold as a noise cluster regardless of the envelope area. Also, the envelope area ratio may be, for example, the ratio of the integral value of the first half to the integral value of the second half, or the ratio of the integral value of the second half to the integral value of the entire waveform, instead of the ratio of the integral value of the first half to the integral value of the entire waveform. These envelope area ratios are all values indicating the relationship between the integral value of the envelope line of the waveform before the peak of the pulse signal and the integral value of the envelope line of the waveform after the peak.

[0042] Also, the envelope area and the envelope area ratio according to the first embodiment are calculated from the average waveform of the pulse signals belonging to the cluster, but are not limited thereto. For example, the envelope line area ratio according to another embodiment may be calculated from one of the pulse signals belonging to the cluster (for example, the one closest to the cluster center).

[0043] The signal processing device 100 according to the first embodiment diagnoses power equipment, but is not limited thereto. For example, the signal processing device 100 according to other embodiments may monitor abnormal vibrations of rotating machines such as motors and generators. That is, the target signal may be a signal representing the vibration of a rotating machine. Also, for example, the signal processing device 100 according to other embodiments may monitor abnormal vibrations of a reactor. In this case, the target signal may be a signal representing the vibration of the reactor or an acoustic wave signal emitted from the reactor.

Explanation of Signs

[0044] 100…Signal processing device 101…Acquisition unit 102…Smoothing unit 103…Signal storage unit 104…Pulse extraction unit 105…Clustering unit 106…Feature quantity calculation unit 107…Specification unit 108…Diagnosis unit 109…Output control unit 111…Electrode 200…Target power equipment

Claims

1. A clustering unit that distributes a plurality of pulse signals measured from a target device into a plurality of clusters by clustering based on the waveforms of the respective pulse signals; A specifying unit that specifies, from the plurality of clusters, a partial discharge cluster into which pulse signals related to partial discharge are distributed or a noise cluster into which pulse signals related to noise are distributed, based on the distribution of the power supply phase with respect to the target device when the pulse signals distributed to the respective plurality of clusters are generated; A signal processing apparatus comprising the above.

2. The specifying unit specifies the partial discharge cluster or the noise cluster based on the characteristics regarding the waveforms of the pulse signals distributed to the respective plurality of clusters and the distribution of the power supply phase. The signal processing apparatus according to Claim 1.

3. The plurality of pulse signals are a plurality of signals related to a predetermined period centered on each of a plurality of peaks included in the signal measured from the target device, The characteristics include a value indicating the relationship between the integral value of the envelope of the waveform before the peak of the pulse signal and the integral value of the envelope of the waveform after the peak. The signal processing apparatus according to Claim 2.

4. The characteristics include the integral value of the envelope of the pulse signal. The signal processing apparatus according to Claim 2.

5. The plurality of pulse signals are a plurality of signals related to a predetermined period centered on each of a plurality of peaks included in the signal measured from the target device, The characteristics include an envelope area that is the integral value of the envelope of the pulse signal and an envelope area ratio that indicates the relationship between the integral value of the envelope of the waveform before the peak of the pulse signal and the integral value of the envelope of the waveform after the peak. including The specifying unit specifies, as the noise cluster, a cluster in which the relationship between the envelope area and the envelope area ratio satisfies an identification inequality that is a linear inequality for identifying the noise cluster. The identification inequality is represented as Y > -aX + b (where a > 0, b > 0) when the envelope area is set as X and the envelope area ratio is set as Y. The signal processing apparatus according to Claim 2.

6. The specifying unit specifies, as the noise cluster, a cluster in which the degree of dispersion of the power supply phase of the target device when the distributed pulse signal is generated exceeds a predetermined threshold. The signal processing apparatus according to Claim 1.

7. The specifying unit specifies the partial discharge cluster from the plurality of clusters. A diagnostic unit that diagnoses the target device based on the pulse signal sorted into the specified partial discharge cluster The signal processing device according to any one of claims 1 to 6, comprising:

8. Sorting a plurality of pulse signals measured from a target device into a plurality of clusters by clustering based on the waveform of each pulse signal; Based on the distribution of the power phase with respect to the target device when the pulse signals sorted into each of the plurality of clusters are generated, from the plurality of clusters, a partial discharge cluster into which a pulse signal related to partial discharge is sorted or a noise cluster into which a pulse signal related to noise is sorted is identified; A signal processing method comprising:

Images