JP2018072304A - Rotary machine diagnostic system and rotary machine diagnostic method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a rotary machine diagnostic system and a rotary machine diagnostic method capable of detecting properly a change of an insulation deterioration state.SOLUTION: A rotary machine diagnostic system 100 comprises: a partial discharge signal distribution acquisition part 111 for acquiring a partial discharge signal distribution with respect to an applied voltage phase of a rotary machine 102; a maximum value extraction processing part 112 for extracting maximum values of positive/negative of a charge amount based on the acquired partial discharge signal distribution; a voltage phase extraction processing part 113 for extracting a voltage phase of the maximum value of any one polarity out of the extracted maximum values, and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction in which the maximum value is a center thereof; and a partial discharge mechanism determination part 114 for determining a partial discharge pattern shape based on magnitude of the maximum values of both positive/negative polarities, and a width from the voltage phase before and after to the voltage phase of the maximum value.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a rotating machine diagnostic system and a rotating machine diagnostic method.

In a rotating machine, deterioration such as cracks and peeling occurs in the insulating layer of the stator winding due to operating stress. A partial discharge occurs in the deteriorated part due to the voltage during operation, and the deterioration state of the insulation is grasped by detecting this partial discharge. Since the partial discharge of the rotating machine is buried in radio noise (hereinafter referred to as noise) accompanying operation, it is necessary to distinguish between partial discharge and noise and detect only partial discharge.
If a motor built in to drive a production facility or industrial machine has a sudden failure, unplanned repairs or parts replacement work is required, and the production facility may be suspended. Although it is possible to grasp the degree of deterioration of the motor to some extent by a planned diagnosis for stopping the production facility, the operation rate of the production facility is reduced by stopping the motor. If it is a diagnosis during operation, it is possible to prevent a reduction in equipment operation rate, and thus there is an increasing need for a motor diagnosis technique during operation. Here, since about 25% of the high-voltage motor failure is caused by insulation deterioration, development of a technique for diagnosing insulation deterioration is in progress. In a high voltage motor, local insulation deterioration inside the motor is diagnosed from a distribution of discharge electric charge with respect to an applied voltage phase called a partial discharge pattern obtained by partial discharge measurement.

  In Patent Document 1, a signal detected by a partial discharge sensor provided in a rotating electrical machine is branched and simultaneously measured in two different frequency bands, and a two-frequency intensity ratio is within a certain range based on a two-frequency intensity correlation of the measured signal. The signal group is separated into a certain signal group, the separated signal group is classified into each occurrence frequency distribution, each signal group indicated in the separated occurrence frequency distribution is processed, and the distribution pattern of each signal group is set to the phase characteristic. A method for detecting partial discharge of a rotating electric machine is described. In the partial discharge detection method described in Patent Document 1, external discharge and internal discharge are determined by comparing the positive and negative polarities of the partial discharge charge amount from the acquired partial discharge pattern.

Japanese Patent Application Laid-Open No. 2001-074742

  In the partial discharge detection method described in Patent Document 1, a partial discharge generation site is roughly determined by comparing which of the positive and negative values of the partial discharge charge amount is larger. However, it is necessary to analyze the partial discharge pattern in order to determine in detail the location where the partial discharge occurs. Until now, the inspector visually discriminates the partial discharge pattern, but there has been a problem that it is necessary to automatically discriminate the partial discharge pattern in order to diagnose the insulation deterioration state.

  This invention is made | formed in view of such a situation, and makes it a subject to provide the rotating machine diagnostic system and rotating machine diagnostic method which can detect the change of an insulation deterioration state appropriately.

  In order to solve the above problems, a rotating machine diagnostic system according to the present invention includes a partial discharge signal distribution acquisition unit that acquires a partial discharge signal distribution with respect to an applied voltage phase of a rotating machine, and a charge based on the acquired partial discharge signal distribution. Obtained before and after the maximum value extraction processing unit that extracts the positive and negative maximum values of the quantity, the voltage phase of the maximum value of at least one of the extracted maximum values, and the phase direction centered on the maximum value Based on the voltage phase extraction processing unit for extracting the voltage phase based on the partial discharge signal distribution, the magnitude of the maximum value of both positive and negative polarities, and the width from the previous and subsequent voltage phases to the voltage phase of the maximum value And a partial discharge mechanism determination unit for determining a partial discharge pattern shape.

  According to the present invention, it is possible to provide a rotating machine diagnosis system and a rotating machine diagnosis method that can appropriately detect a change in an insulation deterioration state.

It is a figure which shows the structure of the rotary machine diagnostic system which concerns on embodiment of this invention. It is a block diagram of the data processor of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the partial discharge signal distribution which the rotary machine diagnostic system which concerns on the said embodiment acquired as a table | surface. It is a figure explaining the kind of partial discharge pattern shape generate | occur | produced with the high voltage motor prescribed | regulated by IEC60034-27-2 of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the partial discharge pattern shape determination table which classify | categorizes the partial discharge pattern shape of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure explaining the outline | summary of the insulation deterioration diagnosis of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the classification example 1 of the partial discharge pattern shape of the rotary machine diagnostic system which concerns on the said embodiment, (a) is a figure which shows a partial discharge pattern shape, (b) is a figure which shows the determination of a partial discharge pattern shape. is there. It is a figure which shows the classification example 2 of the partial discharge pattern shape of the rotary machine diagnostic system which concerns on the said embodiment, (a) is a figure which shows a partial discharge pattern shape, (b) is a figure which shows the determination of a partial discharge pattern shape. is there. It is a flowchart which shows the process of the insulation deterioration diagnosis of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the application example of the partial discharge mechanism determination of the rotary machine diagnostic system which concerns on the said embodiment, (a) is a figure which shows the parameter of the charge amount distribution of the application example 1, (b) (c) is an application example. FIG. 6 is a diagram illustrating a charge amount distribution parameter of 2; It is a figure explaining the application example 3 of the partial discharge mechanism determination of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of a determination of (1-1) Internal void (internal void) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of a determination of (1-2) Internal delamination (internal peeling) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of determination of (1-3) Delamination between conductor and insulation (conductor-insulation interlayer discharge) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of a determination of (2) Slot partial discharges (slot partial discharge) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of a determination of (3-1) Corona activity at the S / C and stress grading coating (corona discharge) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of a determination of (3-2) Surface discharges tracking (creeping discharge) of the rotary machine diagnostic system which concerns on the said embodiment. It is a figure which shows the example of determination of (3-3) Gap type discharges (gap type discharge) of the rotary machine diagnostic system which concerns on the said embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(Embodiment)
FIG. 1 is a diagram showing a configuration of a rotating machine diagnostic system according to an embodiment of the present invention. The rotating machine diagnosis system of this embodiment is an example applied to insulation deterioration diagnosis of a rotating machine.
As shown in FIG. 1, the power source 101 supplies power to the rotating machine 102 via power supply lines 103a, 103b, and 103c. The rotating machine diagnostic system 100 includes a lead wire 104, a coupling capacitor 105, a current sensor 106, a current sensor 107, a voltage sensor 108, and a data processing device 110.
Electric power is supplied from the power source 101 to the rotating machine 102 via power supply lines 103a, 103b, and 103c. A lead line 104 is led out from the power supply line 103 a and is grounded through a coupling capacitor 105.
The current sensor 106 measures a current signal waveform flowing through the feeder line 103a, and the current sensor 107 measures fluctuations in the current flowing through the lead line 104. The two current sensors 106 and 107 detect the current at two locations by changing the installation location, so that it is possible to estimate the insulation degradation location based on the direction of the flowing current and measure with high accuracy (no influence of noise). be able to.
The voltage sensor 108 measures a voltage waveform applied to the rotating machine 102 via the feeder line 103a.
The current sensor 106 and the voltage sensor 108 can be installed in any of the feeder lines 103a, 103b, and 103c. Similarly, the lead line 104, the coupling capacitor 105, and the current sensor 107 can be connected to any of the feeder lines 103a, 103b, and 103c. Current sensors 106 to 108 detect signals of installed feeder lines 103a, 103b, and 103c. Note that the power supply lines 103a, 103b, and 103c are collectively referred to as a power supply line 103.

  In FIG. 1, the current sensor 107 is installed between the feed line 103 and the coupling capacitor 105, but may be installed between the coupling capacitor 105 and the ground line depending on the installation space. In addition, two current sensors are provided for each phase, that is, the power supply line 103 and the lead-out line 104 from the power supply line 103, but the current sensor 106 may be provided only in the power supply line 103. In this case, although weak against noise, the number of current sensors can be reduced and a simple device configuration can be obtained. Also, any type of current sensor may be used. For example, a through-type current sensor, a clamp-type current sensor, a split-type current sensor, an optical fiber sensor using a magneto-optic effect, or the like can be used. The type of the voltage sensor is not limited, and for example, a probe type sensor or a non-contact type sensor can be used.

FIG. 2 is a configuration diagram of a data processing device of the rotating machine diagnosis system according to the embodiment of the present invention.
As shown in FIG. 2, the data processing device 110 includes a partial discharge signal distribution acquisition unit 111, a maximum value extraction processing unit 112, a voltage phase extraction processing unit 113, and a partial discharge mechanism determination unit 114.
The data processing device 110 is realized by a computer into which a predetermined program for realizing each function related to the rotating machine diagnosis system 100 is introduced, and the CPU of the computer reads the predetermined program into a memory and executes it, thereby rotating the rotating machine. It functions as a rotating machine diagnostic method of the diagnostic system 100.

<Partial discharge signal distribution acquisition unit 111>
The partial discharge signal distribution acquisition unit 111 acquires partial discharge signal distribution data (hereinafter referred to as partial discharge signal distribution) with respect to the applied voltage phase (hereinafter referred to as voltage phase) of the rotating machine 102. Specifically, the partial discharge signal distribution acquisition unit 111 acquires the partial discharge signal distribution by synchronizing the measured current fluctuation caused by the partial discharge with respect to the voltage phase. In the present embodiment, the partial discharge signal distribution acquisition unit 111 synchronously extracts the current signal caused by the partial discharge detected by the current sensors 106 and 107 attached to the rotating machine 102 and the voltage signal detected from the voltage sensor 108. Thus, the partial discharge signal distribution is acquired.

The partial discharge signal distribution can be acquired as digital data as shown in FIG. FIG. 3 is a diagram showing the acquired partial discharge signal distribution in a table. As shown in FIG. 3, the partial discharge charge amount [pC] with respect to the voltage phase [deg] is stored.
Here, by integrating the measured current, a charge amount distribution (PRPD pattern: PRPD pattern, Phase Resolved Partial Discharge pattern) released by the partial discharge can be acquired. For example, as indicated by reference symbol e in FIG. 6B described later, when the voltage phase is taken on the horizontal axis and the charge amount is taken on the vertical axis, the charge amount distribution (PRPD pattern) is obtained by applying the applied voltage waveform 101 and the partial discharge signal 102 to each other. Overlaid with the distribution.

In the present embodiment, the rotating machine diagnosis system 100 includes a pass frequency band filter 300 that separates partial discharge signals. The pass frequency band filter 300 is a plurality of filters having different pass frequency bands, and the partial discharge signal for each pass frequency band is obtained from the current signal caused by the partial discharge detected by the current sensors 106 and 107 attached to the rotating machine 102. To separate.
The partial discharge signal distribution acquisition unit 111 uses a plurality of pass frequency band filters 300 having different pass frequency bands for the current signals caused by the partial discharges measured by the current sensors 106 and 107 attached to the rotating machine 102. The partial discharge signals can be separated to obtain partial discharge signals.

<Maximum value extraction processing unit 112>
The maximum value extraction processing unit 112 extracts positive and negative maximum values (maximum values of both positive and negative polarities) of the charge amount based on the acquired partial discharge signal distribution.

<Voltage phase extraction processing unit 113>
The voltage phase extraction processing unit 113 performs the voltage phase of the maximum value of at least one of the extracted maximum values (voltage phase when the maximum value occurs), and before and after the phase direction around the maximum value. The voltage phase based on the acquired partial discharge signal distribution is extracted.
The voltage phase extraction processing unit 113 extracts the voltage phase before and after the extracted maximum value as a predetermined ratio. The voltage phase before and after is a value (half value) when the maximum value is halved. Further, the voltage phase before and after is a voltage phase of a value (3/4 value) when it is set to, for example, 3/4 other than a half value.
In addition, as shown in a modification example described later, the voltage phase extraction processing unit 113 acquires the partial voltage signals obtained before and after the extracted voltage phase of each positive and negative maximum value and the phase direction centered on each positive and negative maximum value. You may make it extract the voltage phase based on distribution.

<Partial discharge mechanism determination unit 114>
The partial discharge mechanism determination unit 114 performs partial discharge based on the magnitudes of the maximum values of both positive and negative polarities and the width from the voltage phase before and after the phase value around the maximum value to the voltage phase of the maximum value. The pattern shape is determined. The partial discharge mechanism determination unit 114 determines a partial discharge mechanism indicating a partial discharge occurrence site based on the determined partial discharge pattern shape. That is, the partial discharge mechanism determination unit 114 determines a partial discharge mechanism indicating the degree of imbalance on the left and right of the distribution centered on the maximum value, based on the difference between the voltage phase before and after the voltage phase of the maximum value.

  When the voltage phase extraction processing unit 113 extracts the extracted voltage phases of the positive and negative maximum values and the voltage phases before and after the partial discharge mechanism determination unit 114, the partial discharge mechanism determination unit 114 determines the magnitudes of the positive and negative maximum values and the voltage phases of the front and rear. The partial discharge pattern shape is determined based on the width from the voltage phase to the maximum voltage phase.

In the present embodiment, the partial discharge mechanism determination unit 114 determines the partial discharge pattern shape classified by the magnitude of the maximum value of both positive and negative polarities and the width from the previous and subsequent voltage phases to the voltage phase of the maximum value. A table 200 (storage unit) (see FIG. 5) is stored.
The partial discharge mechanism determination unit 114 refers to the partial discharge pattern shape determination table 200 based on the acquired maximum and minimum values of both positive and negative polarities and the width from the front and rear voltage phases to the maximum voltage phase. The discharge pattern shape is determined.
The partial discharge mechanism determination unit 114 collates the determined partial discharge pattern shape with the partial discharge pattern shape of the partial discharge pattern shape determination table 200, and determines whether or not the rotating machine 102 is appropriate.
The partial discharge mechanism determination unit 114 compares the acquired maximum value with a threshold value determined for each classified partial discharge pattern shape, and performs insulation deterioration diagnosis of the rotating machine 102.
The partial discharge mechanism determination unit 114 acquires a parameter from the determined partial discharge pattern shape and applies machine learning to the parameter to detect an insulation deterioration state of the rotating machine and a change in the state. In machine learning, VQC (Vector Quantization Clustering) may be applied.

  When it is considered that a plurality of partial discharge phenomena occur at the same time, the partial discharge mechanism determination unit 114 uses the pass frequency band filter 300 when acquiring the partial discharge signal, and generates a partial discharge pattern for each frequency band. Acquire and determine the partial discharge mechanism for each. This makes it possible to determine even when partial discharges are generated by a plurality of mechanisms.

  As described above, in the rotating machine diagnosis system 100, the maximum value extraction processing unit 112 extracts the magnitude relationship between the maximum values of both positive and negative polarities with respect to the partial discharge signal distribution acquired by the partial discharge signal distribution acquisition unit 111. The voltage phase extraction processing unit 113 extracts the voltage phase at the time when the maximum value is generated, and two voltage phases at the front and rear in the phase direction centering on the maximum value (four at the front and rear when a 3/4 value is used). Then, the width of the voltage phase with the maximum value is acquired. The partial discharge mechanism determination unit 114 determines the partial discharge pattern based on the magnitude of the maximum value of both positive and negative polarities and the width from the voltage phase before and after the phase value around the maximum value to the voltage phase of the maximum value. The shape of the partial discharge pattern is determined with reference to the shape determination table 200. Since the partial discharge mechanism can be determined, it is possible to select a threshold for determination along the partial discharge mechanism, and it is possible to determine external discharge and internal discharge. In addition, since it is possible to determine whether the distribution has a steep shape, it is possible to determine whether the discharge is a partial discharge caused by a void or a partial discharge caused by peeling.

  In addition, the rotating machine diagnosis method of the present embodiment extracts the positive and negative maximum values of the charge amount based on the process of acquiring the partial discharge signal distribution with respect to the applied voltage phase of the rotating machine 102 and the acquired partial discharge signal distribution. The voltage phase of the maximum value of the polarity of the process and at least one of the extracted maximum values, and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction around the maximum value are extracted. And determining a partial discharge pattern shape based on the magnitude of the maximum value of both positive and negative polarities and the width from the voltage phase before and after the voltage phase to the voltage phase of the maximum value.

[Partial discharge type]
Next, the types of partial discharge pattern shapes generated in the high voltage motor will be described.
FIG. 4 is a diagram for explaining the types of partial discharge pattern shapes generated in a high-voltage motor defined by IEC (International Electrotechnical Commission) 60034-27-2. The horizontal axis represents the voltage phase, and the vertical axis represents the partial discharge signal distribution.
The types of partial discharge pattern shapes generated in a high voltage motor can be classified into the following seven typical types.
(1-1) Internal void (discharge in internal void) (See Fig. 4 (a))
(1-2) Internal delamination (see Fig. 4 (b))
(1-3) Delamination between conductor and insulation (conductor-insulation interlayer discharge) (See Fig. 4 (c))
(2) Slot partial discharges (see Fig. 4 (d))
(3-1) Corona activity at the S / C and stress grading coating (corona discharge) (See Fig. 4 (e))
(3-2) Surface discharges tracking (see Fig. 4 (f))
(3-3) Gap type discharges (See Fig. 4 (g))

[Partial discharge pattern shape classification]
Next, classification of partial discharge pattern shapes will be described.
FIG. 5 is a diagram showing a partial discharge pattern shape determination table 200 for classifying partial discharge pattern shapes.
The partial discharge pattern shapes can be classified into seven types based on the comparison of positive and negative maximum values (Comparison positive and negative maximum charges) and the distribution state of partial discharge patterns.
The classifications (1-1), (1-2), (1-3), (2), (3-1), (3-2), and (3-3) in the partial discharge pattern shape determination table 200 are: Corresponds to the partial discharge types (1-1), (1-2), (1-3), (2), (3-1), (3-2), and (3-3) in FIG. doing.
For example, in the case of Q ^(p) max = Q ⁽ⁿ⁾ max, if p ⁽ⁿ⁾ 5−p ^{(n) 1} << ^{p (p)} 5−p ^(p) 1 then Gap type discharges (3-3 When the distribution state p3-p1 ≒ p5-p3, it is classified as internal void (1-1), and when the distribution state p3-p1 << p5-p3, it is classified as internal delamination (1-2). The In the case of Q ^(p) _max = Q ⁽ⁿ⁾ _max , when the number of positive partial discharges N ^(p) is larger than the number of negative partial discharges N ^(N) , corona discharge is selected.

In addition, when Q ^(p) _max > Q ⁽ⁿ⁾ _max , p ⁽ⁿ⁾ 5−p ^{(n) 1} << ^{p (p)} 5−p ^(p) 1 does not exist, and the distribution state p3−p1 When p5−p3, it is classified as Corona activity at the S / C and stress grading coating (3-1), and when distribution state p3−p1 << p5−p3, it is classified as Slot partial discharges (2).
If Q ^(p) _max <Q ⁽ⁿ⁾ _max , then p ⁽ⁿ⁾ 5−p ^{(n) 1} << ^{p (p)} 5−p ^(p) 1 and Surface discharges tracking (3-2 ) And distribution state p3-p1≈p5-p3, it is classified as Delamination between conductor and insulation (1-3), and there is no distribution state p3-p1 << p5-p3.

Hereinafter, an insulation deterioration diagnosis method of the rotating machine diagnosis system 100 configured as described above will be described.
[Outline of insulation deterioration diagnosis]
First, an outline of the insulation deterioration diagnosis of the rotating machine diagnosis system will be described.
FIG. 6 is a diagram for explaining the outline of the insulation deterioration diagnosis of the rotating machine diagnosis system.

<Measurement of PD signal>
6A, the voltage and current of the rotating machine 102 in operation are detected from the current sensors 106 and 107 and the voltage sensor 108 provided on the power supply line 103 and the lead-out line 104 of the rotating machine 102. Then, it is measured as a PD (Partial Discharge) signal (see symbol b in the figure). In the present embodiment, the partial discharge signal distribution acquisition unit 111 synchronously extracts the current signal caused by the partial discharge detected by the current sensors 106 and 107 attached to the rotating machine 102 and the voltage signal detected from the voltage sensor 108. Thus, the partial discharge signal distribution is acquired. As indicated by reference symbol c in FIG. 6A, a PD signal is separated from noise (Separation of PD signal from noise).

<PD signal to PRPD pattern transformation>
It is assumed that the PD signal indicated by the symbol d in FIG. 6B is acquired and the partial discharge signal distribution indicated by the symbol e in FIG. 6B is acquired. When the voltage phase is taken on the horizontal axis and the charge amount is taken on the vertical axis, the charge amount distribution (PRPD pattern) is shown by superimposing the applied voltage waveform 101 and the distribution of the partial discharge signal 102.
As described above, by integrating the measured current, the charge amount distribution (PRPD pattern) released by the partial discharge can be acquired.

<Extraction of parameters indicating the shape of PRPD pattern>
As shown in FIG. 6C, the positive maximum value Q ^(p) _max and the negative maximum value Q ⁽ⁿ⁾ _max of the partial discharge signal distribution are calculated, and the positive maximum value Q ^(p) _max and the negative value are calculated. The maximum value Q ⁽ⁿ⁾ _max is compared. A plurality of points indicating the outer shape of the partial discharge signal distribution are extracted with respect to the partial discharge signal distribution of the positive maximum value Q ^(p) _max having a large absolute value of the maximum value.
The point indicating the outer shape of the partial discharge signal distribution, with respect to partial discharge signal distribution shown by reference numeral e in FIG. 6 (b), the positive maximum value Q ^(p) point extracted with _max p3, Q ^(p) _max The point p1−p5 extracted by 1/2 × Q ^(p) _max (ie, Q ^(p) _half ) of the above and the above half × Q ^(p) _max on the convex side is further half (1/2) value And points p2−p4 extracted with 3/4 × Q ^(p) _max .
For FIG. 6 (c), the positive maximum value Q ^(p) point extracted with _max p3, and extracted with Q ^(p) _max of the ^{1/2 × Q (p)} max ( i.e. Q ^{(p) _half)} Points p1 to p5 and points p2 to p4 extracted by 3/4 × Q ^(p) _{max obtained} by further halving the above 1/2 × Q ^(p) _max on the convex side are used. Note that it is also possible to use only the point p3 extracted with the positive maximum value Q ^(p) _max and the points p1 to p5 extracted with Q ^(p) _half .

<Diagnosis with VQC (VQC diagnosis)>
As shown in FIG. 6D, the insulation deterioration state of the rotating machine 102 and a change in the state are detected by applying VQC (Vector Quantization Clustering). In the present embodiment, VQC is used as a learning method for machine learning.
As indicated by a symbol f in FIG. 6D, clustering parameters are set and analyzed, and diagnosis is performed based on the analysis result. The symbol h in FIG. 6D is a diagram for explaining vector quantization, and a plurality of sample data is represented by
1478250654997_0
And as a whole
1478250654997_1
To do. VQC first takes N samples N-dimensional vector, and once all samples are taken, the next number of coded K K
1478250654997_2
I do. Thereafter, one “representative vector” is determined from each cluster, and vectors other than the representative vector are replaced with the representative vectors of the assigned clusters. Finally, the representative vector
1478250654997_3
To complete the quantization.

[Example of classification of partial discharge pattern shapes]
Next, an example of classification of partial discharge pattern shapes will be described.
<Classification example 1>
FIG. 7 is a diagram showing classification example 1 of partial discharge pattern shapes, FIG. 7A is a diagram showing partial discharge pattern shapes, and FIG. 7B is a diagram showing determination of partial discharge pattern shapes. The shaded portion in FIG. 7A schematically shows points where the measured charge amount is plotted. Actually, noise points are removed (filtering processing) from the points where the measured charge amounts are plotted, and a charge amount distribution pattern is drawn (contour processing). The partial discharge signal distribution of the drawn charge amount is called a partial discharge pattern shape.
It is assumed that the parameter indicating the partial discharge pattern shape shown in FIG. 7A is acquired by the <PD signal to PRPD pattern transformation> process shown in FIG.
Next, the positive maximum value Q ^(p) _max and the negative maximum value Q ⁽ⁿ⁾ _max of the partial discharge pattern shape are obtained by the <Extraction of parameters indicating the shape of PRPD pattern> process shown in FIG. Calculate and compare the positive and negative maximum values. Here, as shown in FIG. 7A, Q ^(p) _max > Q ⁽ⁿ⁾ _max .

Next, a plurality of points indicating the outer shape of the partial discharge pattern shape are extracted from the positive partial discharge pattern shape having a large absolute value of the maximum value by the above-described <Extraction of parameters indicating the shape of PRPD pattern> process. More specifically, as shown in FIG. 6 (c), the maximum positive value Q ^(p) point extracted with _max p3, Q ^(p) _max of the ^{1/2 × Q (p)} max ( i.e. Q ^{(p) The} points p1 to p5 extracted in _half ) are used. Note that the point p2−p4 extracted by 3/4 × Q ^(p) _{max in} which the above 1/2 × Q ^(p) _max is further halved (1/2) on the convex side may be used. .

As shown in FIG. 7B, the partial discharge pattern shape shown in FIG. 7A is such that Q ^(p) _max > Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^{(p) 1} << p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), it can be determined that the partial discharge pattern shape shown in FIG. 7A is Delamination between conductor and insulation (1-3). It can be determined that the partial discharge corresponds to the slot partial discharge (2) shown in FIG.

<Classification example 2>
8A and 8B are diagrams showing classification example 2 of partial discharge pattern shapes. FIG. 8A is a diagram showing partial discharge pattern shapes, and FIG. 8B is a diagram showing determination of partial discharge pattern shapes.
It is assumed that the parameter indicating the partial discharge pattern shape shown in FIG. 8A is acquired by the <PD signal to PRPD pattern transformation> process shown in FIG.
By the <Extraction of parameters indicating the shape of PRPD pattern> process shown in FIG. 6C, the positive maximum value Q ^(p) _max and the negative maximum value Q ⁽ⁿ⁾ _max of the partial discharge pattern shape are calculated, Compare positive and negative maximum values. Here, as shown in FIG. 8A, Q ^(p) _max <Q ⁽ⁿ⁾ _max .

Then, by the <Extraction of parameters indicating the shape of PRPD pattern> process, a plurality of points indicating the outer shape of the partial discharge pattern shape are extracted with respect to the negative partial discharge pattern shape having the largest absolute value. Specifically, the p3 point extracted by the maximum negative value _{^{Q (n) max, Q (}} n) max of ^{1/2 × Q (n)} max ( i.e. Q ^{(n) _half)} point extracted in p1- p5 is used.
As shown in FIG. 8B, the partial discharge pattern shape shown in FIG. 8A is Q ^(p) max <Q ⁽ⁿ⁾ max and p ⁽ⁿ⁾ 3-p ⁽ⁿ⁾ 1≈ p ⁽ⁿ⁾ 5−p ⁽ⁿ⁾ 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), it can be determined that the partial discharge pattern shape shown in FIG. 8A is Slot partial discharges (1-3). It can be determined that the partial discharge corresponds to the conductor-insulating interlayer discharge (1-3) shown in FIG.
In this way, external discharge and internal discharge are distinguished by comparing the maximum values of the extracted bipolar values, and partial discharge at voids or partial discharge at peeling by checking whether the distribution is steep or not. Can be determined.

[Processing flow for insulation deterioration diagnosis]
FIG. 9 is a flowchart showing a process of insulation deterioration diagnosis. This flow is executed by the data processing device 110 (see FIG. 2) of the rotating machine diagnosis system 100.
When the insulation deterioration diagnosis is started, in step S101, the partial discharge signal distribution acquisition unit 111 detects the current signal resulting from the partial discharge detected by the current sensors 106 and 107 (see FIG. 1) attached to the rotating machine 102 and the voltage sensor 108. The partial discharge signal distribution is acquired by synchronously extracting the voltage signal detected from.
In step S102, the partial discharge signal distribution acquisition unit 111 determines whether or not the partial discharge signal. Since the partial discharge of the rotating machine 102 is buried in the noise accompanying the operation, the partial discharge signal and the noise are discriminated and only the partial discharge signal is detected.
When there is a partial discharge signal (step S102: Yes), the process proceeds to step S103, and when there is no partial discharge signal (step S102: No), the process returns to step S101.

In step S103, the local maximum value extraction processing unit 112 extracts positive and negative local maximum values (maximum values of both positive and negative polarities) based on the acquired partial discharge signal distribution.
In step S104, the voltage phase extraction processing unit 113 acquires the partial discharge signal obtained before and after the voltage phase of the maximum value of at least one of the extracted maximum values and the phase direction centered on the maximum value. A voltage phase based on the distribution is extracted.
In step S105, the partial discharge mechanism determination unit 114 is based on the magnitude of the maximum value of both positive and negative polarities and the width from the voltage phase before and after the front and back in the phase direction around the maximum value to the voltage phase of the maximum value. Then, the partial discharge pattern shape is determined.

In step S106, the partial discharge mechanism determination unit 114 compares the determined partial discharge pattern shape with the partial discharge pattern shape of the partial discharge pattern shape determination table 200, and the degree of deterioration abnormality increases. It is determined whether or not.
When the degree of deterioration abnormality is increasing (step S106: Yes), the process proceeds to step S107. When the degree of deterioration abnormality is not increasing (step S106: No), the process returns to step S101.
In step S107, the partial discharge mechanism determination unit 114 calculates the magnitude relationship between the shape parameters from the determined partial discharge pattern shape. By calculating the magnitude relationship of the shape parameters, it is possible to select a threshold for determination along the partial discharge mechanism.
In step S108, the partial discharge mechanism determination unit 114 determines the degradation site by the determination along the partial discharge mechanism.
In step S109, the partial discharge mechanism determination unit 114 determines whether an emergency measure is necessary.
If an emergency measure is necessary (step S109: Yes), the process proceeds to step S110. If an emergency measure is not required (step S109: No), the process returns to step S101.
In step S110, the partial discharge mechanism determination unit 114 plans maintenance in accordance with the deteriorated part and ends the insulation deterioration diagnosis process according to this flow.

[Application example]
FIG. 10 is a diagram illustrating an application example of partial discharge mechanism determination.
<Application example 1>
It is assumed that the charge amount distribution (PRPD pattern) parameters shown in FIG.
The maximum positive value Q ^(p) _max and the negative maximum value Q ⁽ⁿ⁾ _max of the partial discharge pattern shape are calculated, and the positive and negative maximum values are compared. And the point which shows the external shape of a partial discharge pattern shape with respect to the positive partial discharge pattern shape with a large absolute value of a maximum value is extracted. In this case, the positive maximum value Q ^(p) point extracted with _max p3, Q ^(p) _max of the 1/2 × Q ^(p) point extracted with _max using the p1-p5.
As shown in FIG. 10A, with respect to the point p3 extracted with the positive maximum value Q ^(p) _max which is the maximum charge amount, the voltage phase width of the left point p1−p3 is the right point. It can be seen that the voltage phase width is smaller than p3-p5. Therefore, it can be seen that the voltage phase width at the left point p1-p3 shown in FIG. 10A is smaller than the voltage phase width at the right point p3-p5.

The <Application Example 1> For convenience of description, the maximum positive value Q ^(p) point extracted with _max p3, Q ^(p) point and extracted with ^{1/2 × Q (p)} max of the _max p1-p5 That is, this is an example in which the voltage phase width at the points p1 to p3 is compared with the voltage phase width at the points p3 to p5. Thereby, the partial discharge pattern shape shown to Fig.10 (a) has been determined to be slot partial discharge. However, the partial discharge pattern shape in which the voltage phase width of the left point p1-p3 is smaller than the voltage phase width of the right point p3-p5 is not limited to the slot partial discharge. For example, in the case of the corona discharge shown in FIG. 4E, the voltage phase width at the left point p1-p3 is smaller than the voltage phase width at the right point p3-p5. In other words, it is not possible to immediately determine that the discharge is a slot discharge only from the voltage phase width on the left side and the voltage phase width on the right side of the maximum charge amount.

<Application example 2>
FIGS. 10B and 10C are diagrams showing an application example 2 of the partial discharge mechanism determination. FIG. 10B is the same as the charge amount distribution (PRPD pattern) shown in FIG. 10A, and FIG. FIG. 10C shows parameters of another acquired charge amount distribution (PRPD pattern).
In the present embodiment, as shown in FIG. 10 (b) (c), the maximum positive value Q ^(p) point extracted with _max p3, with Q ^(p) _max of the ^{1/2 × Q (p)} max in addition to the extracted points p1-p5, further use of the Q ^(p) _max of the ^{1/2 × Q (p)} max was ^{1/2 3/4 × Q (p} ) point extracted with _max p2-p4 . Then, in addition to the comparison between the voltage phase width at the point p1-p3 and the voltage phase width at the point p3-p5, the voltage phase width at the point p1-p2 and the voltage phase width at the point p2-p3, and further the point p3 The voltage phase width at −p4 and the voltage phase width at points p4−p5 are also compared. Accordingly, the partial discharge pattern shape shown in FIG. 10B is obtained by comparing the voltage phase width at the point p1-p2 with the voltage phase width at the point p2-p3. It can be determined that it is larger than the voltage phase width of p1-p2, that is, it is convex outward on the left side of the maximum charge amount. When the width of the voltage phase on the left side of the maximum charge amount is smaller than the width of the voltage phase on the right side and the voltage phase on the left side is an outwardly convex shape, from the partial discharge pattern shape determination table 200 (see FIG. 5), It can be determined that the slot partial discharge shown in FIG.

  On the other hand, in the case of the partial discharge pattern shape shown in FIG. 10C, the voltage phase width at the point p1-p2 is larger than the voltage phase width at the point p2-p3, that is, indented on the left side of the maximum charge amount. It can be determined that there is. When the width of the voltage phase on the left side of the maximum charge amount is smaller than the width of the voltage phase on the right side and the voltage phase on the left side is a concave shape, from the partial discharge pattern shape determination table 200 (see FIG. 5), It can be determined that the corona discharge shown in FIG.

  Thus, in addition to the voltage phase at the time of half the maximum value of the partial discharge signal, the voltage phase at the time of the value of 3/4 of the maximum value is used, so that the partial discharge mechanism can be accurately determined. It becomes possible. Here, if the partial discharge signal distribution is acquired in an environment without noise, the voltage phase at the time of a small signal value can be used. By increasing the voltage phase to be extracted, the partial discharge pattern shape can be accurately grasped, and the partial discharge mechanism can be accurately determined.

<Application example 3>
In the above <Application Example 1> and <Application Example 2>, the maximum values of the partial discharge charges are compared, and the partial discharge pattern shape is determined for the signal distribution of one polarity of the positive and negative maximum values. The shape of the polar signal distribution may be used for discrimination. Hereinafter, a case where the shapes of the signal distributions of both polarities are used for determination in <Application Example 3> will be described.
FIG. 11 is a diagram for explaining an application example 3 of partial discharge mechanism determination. In FIG. 11, the vertical axis represents the left phase width of the maximum charge amount, and the horizontal axis represents the right phase width of the maximum charge amount.
By applying machine learning, it is possible to discriminate between a change in charge amount distribution shape due to an increase in the maximum discharge charge amount (see the solid line in FIG. 11) and a change in charge amount distribution shape due to a change in the discharge mechanism (see the solid line in FIG. 11). can do.

[Example]
12-18 is a figure explaining an Example. 12 to 18, (a) shows the maximum positive or negative charge amount extracted as digital data from the partial discharge pattern of (b), the width of the voltage phase on the left side of this maximum charge amount, and the voltage phase on the right side. (B) is a figure which shows the shape of a partial discharge pattern, (c) is a figure which shows the determination of a partial discharge pattern.

FIG. 12 is a diagram illustrating a determination example of (1-1) Internal void (internal void).
The partial discharge pattern shape parameters shown in FIG. 12A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 12C, the partial discharge pattern shape shown in FIG. 12B is Q ^(p) _max ≈Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^(p) 1 ≈ p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), it was determined that the partial discharge pattern shape shown in FIG. 12B was an internal void (1-1).

FIG. 13 is a diagram illustrating a determination example of (1-2) Internal delamination.
The partial discharge pattern shape parameters shown in FIG. 13A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 13C, the partial discharge pattern shape shown in FIG. 13B is Q ^(p) _max ≈Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^{(p) 1} << p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), the partial discharge pattern shape shown in FIG. 13B was determined to be internal delamination (1-2).

FIG. 14 is a diagram showing a determination example of (1-3) Delamination between conductor and insulation (peeling between conductor and insulator: (conductor-insulation interlayer discharge)).
The partial discharge pattern shape parameters shown in FIG. 14A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 14C, the partial discharge pattern shape shown in FIG. 14B is Q ^(p) _max <Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^(p) 1≈ p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), the partial discharge pattern shape shown in FIG. 14 (b) is a Delamination between conductor and insulation: (conductor-insulator interlayer discharge) ( 1-3).

FIG. 15 is a diagram illustrating a determination example of (2) Slot partial discharges.
The partial discharge pattern shape parameters shown in FIG. 15A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 15C, the partial discharge pattern shape shown in FIG. 15B is such that Q ^(p) _max > Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^{(p) 1} << p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), it was determined that the partial discharge pattern shape shown in FIG. 15B is Slot partial discharges (2).

FIG. 16 is a diagram showing a determination example of (3-1) Corona activity at the S / C and stress grading coating (corona discharge).
The partial discharge pattern shape parameters shown in FIG. 16A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 16C, the partial discharge pattern shape shown in FIG. 16B is Q ^(p) _max ≈Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^(p) 1 < p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), the partial discharge pattern shape shown in FIG. 16B is Corona activity at the S / C and stress grading coating (corona discharge) (3-1). Judged.

FIG. 17 is a diagram illustrating a determination example of (3-2) Surface discharges tracking.
The partial discharge pattern shape parameters shown in FIG. 17A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 17C, the partial discharge pattern shape shown in FIG. 17B is Q ^(p) _max <Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^{(p) 1} << p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), the partial discharge pattern shape shown in FIG. 17B was determined to be surface discharges tracking (3-2).

FIG. 18 is a diagram illustrating a determination example of (3-3) Gap type discharges (gap type discharges).
The partial discharge pattern shape parameters shown in FIG. 18A were obtained from the partial discharge pattern shown in FIG.
As shown in FIG. 18C, the partial discharge pattern shape shown in FIG. 18B is Q ^(p) _max ≈Q ⁽ⁿ⁾ _max and p ^(p) 3−p ^{(p) 1} << p ^(p) 5−p ^(p) 3 From the partial discharge pattern shape determination table 200 (see FIG. 5), it was determined that the partial discharge pattern shape shown in FIG. 18B was Gap type discharges (3-3).

  As described above, the rotating machine diagnosis system 100 (see FIG. 1) of the present embodiment includes the partial discharge signal distribution acquisition unit 111 that acquires the partial discharge signal distribution with respect to the applied voltage phase of the rotating machine 102, and the acquired partial discharge. Based on the signal distribution, a maximum value extraction processing unit 112 that extracts the positive and negative maximum values of the charge amount, the voltage phase of the maximum value of at least one of the extracted maximum values, and the predetermined maximum value of the maximum value In this case, the voltage phase extraction processing unit 113 extracts the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction centered on the maximum value, and the maximum and minimum values of the positive and negative polarities. A partial discharge mechanism determination unit 114 that determines the partial discharge pattern shape based on the width from the voltage phase to the maximum voltage phase.

  In addition, the rotating machine diagnosis method of the present embodiment extracts the positive and negative maximum values of the charge amount based on the process of acquiring the partial discharge signal distribution with respect to the applied voltage phase of the rotating machine 102 and the acquired partial discharge signal distribution. The voltage phase of the maximum value of the polarity of the process and at least one of the extracted maximum values, and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction around the maximum value are extracted. And determining a partial discharge pattern shape based on the magnitude of the maximum value of both positive and negative polarities and the width from the voltage phase before and after the voltage phase to the voltage phase of the maximum value.

  By doing in this way, the partial discharge type can be automatically classified by acquiring the partial discharge pattern shape as a parameter. Further, since the pattern shape is extracted in a form suitable for machine learning, it is possible to detect a change in the insulation deterioration state that cannot be detected only by the threshold determination of the partial discharge signal.

  In the present embodiment, the front and rear voltage phases are voltage phases based on the partial discharge signal distribution of values when the maximum values are 1/2, 3/4. In addition to the voltage phase at the time of half the maximum value of the partial discharge signal, by using the voltage phase at the time of 3/4 of the maximum value, the partial discharge can be accurately performed by increasing the voltage phase to be extracted. The pattern shape can be grasped, and the partial discharge mechanism can be accurately determined. If the partial discharge signal distribution is acquired in an environment without noise, the voltage phase at the half value can be used.

  In the present embodiment, the voltage phase extraction processing unit 113 performs the voltage phase based on the acquired partial discharge signal distribution before and after the extracted voltage phase of each positive and negative maximum value and the phase direction centered on each positive and negative maximum value. The partial discharge mechanism determination unit 114 determines the partial discharge pattern shape based on the magnitudes of the positive and negative local maximum values and the width from the voltage phase before and after the voltage phase to the local maximum voltage phase. Thereby, although the number of voltage phases extracted by using both polarities increases, it becomes possible to determine a partial discharge pattern more accurately.

  In the present embodiment, a partial discharge pattern shape determination table 200 classified according to the maximum value of both positive and negative polarities and the width from the preceding and following voltage phases to the maximum voltage phase is stored, and the partial discharge mechanism determination unit 114 is stored. The partial discharge pattern shape is determined with reference to the partial discharge pattern shape determination table 200 based on the obtained maximum and minimum values of both positive and negative polarities and the width from the preceding and following voltage phases to the maximum voltage phase. . Thereby, the determination of the shape of the partial discharge pattern can be advanced along the partial discharge pattern shape determination table 200. It is possible to determine external discharge and internal discharge by comparing the magnitudes of the extracted polarities, and to determine whether it is a partial discharge due to voids or a partial discharge due to peeling by seeing if the distribution is steep. it can. By accumulating the partial discharge pattern shape determination table 200 as a database, it is possible to add knowledge obtained by actual machine diagnosis.

  In the present embodiment, the partial discharge mechanism determination unit 114 compares the determined partial discharge pattern shape with the partial discharge pattern shape of the partial discharge pattern shape determination table 200, and determines whether the rotating machine is normal or not by suitability. To do. Thereby, the insulation deterioration diagnosis can be accurately determined.

  In the present embodiment, the partial discharge mechanism determination unit 114 determines the detailed partial discharge mechanism specific to the partial discharge mechanism by determining the partial discharge mechanism indicating the partial discharge generation site based on the determined partial discharge pattern shape. Discrimination becomes possible. In addition, in the case where the maximum value extraction processing unit 112 to the partial discharge mechanism determination unit 114 are configured by one circuit, can seven partial discharge patterns, which are typical partial discharge patterns, be input and discriminated? Can confirm that it is working properly.

  In the present embodiment, the partial discharge mechanism determination unit 114 compares the acquired maximum value with a threshold value determined for each classified partial discharge pattern shape, and performs an insulation deterioration diagnosis of the rotating machine, whereby the partial discharge mechanism is determined. It is possible to use a threshold value adapted to the above, and to accurately determine the insulation deterioration diagnosis.

  In the present embodiment, the partial discharge mechanism determination unit 114 acquires a parameter from the determined partial discharge pattern shape and applies machine learning including VQC to the parameter to detect an insulation deterioration state of the rotating machine and a change in the state. By doing so, it is possible to detect a minute change in the partial discharge pattern by applying machine learning to the parameter representing the partial discharge pattern shape. In addition, since it is possible to detect an insulation deterioration failure sign by machine learning, it is possible to determine the progress of insulation deterioration that cannot be clarified by threshold determination.

  In the present embodiment, a pass frequency band filter 300 that separates the partial discharge signal is provided, and the partial discharge signal distribution acquisition unit 111 acquires the partial discharge signal distribution for each pass frequency band, thereby obtaining a partial discharge pattern for each frequency band. Is obtained, and partial discharge mechanism determination is performed for each of them, so that it is possible to determine even when partial discharges are generated by a plurality of mechanisms (insulation deterioration with different generation sites). Further, it is possible to detect the progress of deterioration from the change in pattern shape.

  Each of the above-described exemplary embodiments has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to one having all the configurations described. Further, a part of the configuration of an embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of an embodiment. . Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each exemplary embodiment.

DESCRIPTION OF SYMBOLS 1 Vacuum tube part 100 Rotating machine diagnostic system 101 Power supply 102 Rotating machine 103a, 103b, 103c Feeding line 104 Leader line 105 Coupling capacitor 106,107 Current sensor 108 Voltage sensor 110 Data processing apparatus 111 Partial discharge signal distribution acquisition part 112 Maximum value extraction process Unit 113 voltage phase extraction processing unit 114 partial discharge mechanism determination unit 200 partial discharge pattern shape determination table (storage unit)

Claims (11)

A partial discharge signal distribution acquisition unit for acquiring a partial discharge signal distribution with respect to the applied voltage phase of the rotating machine;
Based on the acquired partial discharge signal distribution, a maximum value extraction processing unit that extracts positive and negative maximum values of the charge amount;
Voltage phase extraction that extracts the voltage phase of the maximum value of at least one of the extracted maximum values and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction centered on the maximum value A processing unit;
A partial discharge mechanism determination unit that determines a partial discharge pattern shape based on the magnitude of the maximum value of positive and negative polarities and the width from the previous and subsequent voltage phases to the voltage phase of the maximum value. Rotating machine diagnostic system. The voltage phase extraction processing unit
The rotating machine diagnosis system according to claim 1, wherein the voltage phases before and after the maximum value are set to a predetermined ratio are extracted. The voltage phase based on the partial discharge signal distribution of the value when the local maximum is ½ and / or 3/4 is the voltage phase before and after the front. Rotating machine diagnostic system. The voltage phase extraction processing unit
Extracting the voltage phase of each of the extracted positive and negative maximum values and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction centered on each of the positive and negative maximum values,
The partial discharge mechanism determination unit is
2. The rotating machine according to claim 1, wherein the partial discharge pattern shape is determined based on magnitudes of positive and negative maximum values and a width from the preceding and following voltage phases to the voltage phase of the maximum value. Diagnostic system. A storage unit for storing a partial discharge pattern shape determination table classified by the magnitude of the maximum value of positive and negative polarity and the width from the voltage phase before and after the voltage phase to the voltage phase of the maximum value;
The partial discharge mechanism determination unit is
The partial discharge pattern shape is determined with reference to the partial discharge pattern shape determination table based on the obtained maximum and minimum values of both positive and negative polarities and the width from the voltage phase before and after the voltage phase to the voltage phase of the maximum value. The rotating machine diagnosis system according to claim 1, wherein the determination is made. The partial discharge mechanism determination unit is
The said partial discharge pattern shape and the partial discharge pattern shape of the said partial discharge pattern shape determination table are collated, and it is determined whether the rotating machine is normal by the suitability. Rotating machine diagnostic system. The partial discharge mechanism determination unit is
The rotating machine diagnosis system according to claim 1, wherein a partial discharge mechanism indicating a partial discharge occurrence site is determined based on the determined partial discharge pattern shape. The partial discharge mechanism determination unit is
2. The rotating machine diagnosis system according to claim 1, wherein the obtained local maximum value is compared with a threshold value determined for each of the classified partial discharge pattern shapes to perform insulation deterioration diagnosis of the rotating machine. The partial discharge mechanism determination unit is
A parameter is obtained from the determined partial discharge pattern shape, and machine learning including VQC (Vector Quantization Clustering) is applied to the parameter to detect an insulation deterioration state of the rotating machine and a change in the state. Item 4. The rotating machine diagnosis system according to Item 1. It has a pass frequency band filter that separates the partial discharge signal,
The rotating machine diagnosis system according to claim 1, wherein the partial discharge signal distribution acquisition unit acquires the partial discharge signal distribution for each pass frequency band. Obtaining a partial discharge signal distribution for the applied voltage phase of the rotating machine;
Based on the acquired partial discharge signal distribution, extracting the positive and negative maximum values of the charge amount;
Extracting the voltage phase of the maximum value of the polarity of at least one of the extracted maximum values and the voltage phase based on the acquired partial discharge signal distribution before and after the phase direction centered on the maximum value;
And a step of determining a partial discharge pattern shape on the basis of the magnitude of the maximum value of both positive and negative polarities and the width from the preceding and following voltage phases to the voltage phase of the maximum value. Diagnostic method.