Partial discharge detection device and partial discharge position locating method

Abstract

The present invention comprises: a plurality of ultrasonic sensors (2) installed on the surface of an electric apparatus (900); a detection time difference calculation unit (7) for calculating the detection time difference with respect to ultrasonic waves from respective sensor signals (S) of the plurality of ultrasonic sensors (2); and a discharge position calculation unit (8) for calculating, from the results of calculating the detection time difference, the generation position of partial discharge generated inside the electric apparatus (900). The present invention is configured so that either a method for reading a first peak or a cross-correlation method can be selected as a method used for calculating the detection time difference.

Description

Partial discharge detection device and partial discharge location method

[0001] The present disclosure relates to a partial discharge detection device and a partial discharge location method.

[0002] In diagnosing the deterioration of insulation in electrical equipment, it is effective to detect partial discharges, which occur as a sign of impending insulation breakdown. Partial discharges are generally detected by detecting electromagnetic waves or ultrasonic waves generated by partial discharges. Furthermore, in operation, in addition to detecting whether or not a discharge is occurring, it is important to identify the location of the discharge when it is detected. This is because, depending on the identified fault location, it is possible to determine whether operation can be continued, and whether repairs or replacement are required.

[0003] The location of the discharge can be identified by detecting the first peak of the waveform of the ultrasonic wave generated by the discharge with ultrasonic sensors installed at multiple locations and estimating the detection time difference (see, for example, Patent Document 1). However, when the ultrasonic wave propagates through the electrical device, it follows a complex path due to reflections from components inside the device, such as coils and insulators, which causes the signal to rise gradually, making it difficult to estimate the detection time difference. Therefore, it is conceivable to estimate the phase shift of the waveform by applying a cross-correlation method (see, for example, Patent Document 2), which uses not only the rising portion but also the signal waveform after that portion.

[0004] Japanese Patent Publication No. 6-58392 (page 3, left column [Example], Figures 1, 2, and 4); Japanese Patent Publication No. 64-74465 (page 2, lower left column, page 3, upper right column, Figures 1 and 2)

[0005] However, when using the cross-correlation method, when the waveform similarity is low, the detection accuracy may be lower than when detecting the first peak, making it difficult to accurately identify the location where the discharge occurred.

[0006] The present disclosure discloses a technique for solving the above-described problems, and aims to accurately identify the location of discharge occurrence in an electrical device.

[0007] The partial discharge detection device disclosed in the present disclosure comprises a plurality of ultrasonic sensors installed on the surface of an electrical device, a detection time difference calculation unit that calculates the detection time difference of ultrasonic waves from the sensor signals of each of the plurality of ultrasonic sensors, and a discharge position calculation unit that calculates the occurrence position of partial discharge that has occurred inside the electrical device from the calculation result of the detection time difference, and is characterized in that it is configured so that the method used to calculate the detection time difference can be selected from either a method of reading a first peak or a cross-correlation method.

[0008] The partial discharge location method disclosed herein is characterized by including a detection time difference calculation step of calculating the ultrasonic detection time difference from the respective sensor signals of multiple ultrasonic sensors installed on the surface of an electrical device; a discharge position calculation step of calculating the occurrence position of a partial discharge that has occurred inside the electrical device from the calculation result of the detection time difference; and a calculation method selection step of selecting either a first peak reading method or a cross-correlation method as the method to be used for calculating the detection time difference.

[0009] According to the partial discharge detection device or partial discharge location locating method of the present disclosure, the detection time difference can be calculated using a calculation method suited to the situation, so that the location of discharge occurrence in an electrical device can be identified with high accuracy.

[0010] FIG. 5 is a functional block diagram showing the configuration of a partial discharge detection device according to a first embodiment. FIG. 6 is a flowchart showing a discharge position calculation operation in the partial discharge detection device and partial discharge position locating method according to the first embodiment. FIG. 7 is a functional block diagram showing the configuration of a partial discharge detection device according to a second embodiment. FIG. 8 is a functional block diagram showing the configuration of a partial discharge detection device according to a third embodiment. FIG. 5A and FIG. 5B are flowcharts showing the operation of partial discharge detection devices according to a third embodiment and a modification thereof, respectively. FIG. 9 is a functional block diagram showing the configuration of a partial discharge detection device according to a fourth embodiment. FIG. 10 is a block diagram showing the hardware configuration of an arithmetic execution part for executing control in the partial discharge detection device of the present disclosure. FIG. 11 is a flowchart showing a discharge position calculation operation in the partial discharge detection device and partial discharge position locating method according to a fifth embodiment.

[0011] Hereinafter, with reference to the drawings, detailed descriptions will be given of each embodiment of the partial discharge detection device and partial discharge location method of the present disclosure. In the following description of the embodiments, the same or corresponding parts in each drawing are given the same reference numerals, and the description thereof will not be repeated. Furthermore, the embodiments described below are merely examples, and the present disclosure is not limited to these embodiments.

[0012] 1 and 2 are diagrams for explaining the configuration and operation of a partial discharge detection device according to a first embodiment, where Fig. 1 is a functional block diagram showing the relationship between the partial discharge detection device and an electrical device to be diagnosed, and Fig. 2 is a flowchart showing the partial discharge detection device and the discharge position calculation operation in the partial discharge position locating method.

[0013] As shown in Fig. 1, the partial discharge detection device 1 of the present disclosure detects partial discharges in an electrical device 900 and locates the location of the partial discharges. The electrical device 900 is widely applicable to devices that are covered with a metal casing or a resin, making the parts (interiors) where discharges may occur invisible, and whose surfaces are kept at ground potential by metal or conductive coating. Examples of such electric power devices include oil-filled transformers, gas-insulated transformers, molded transformers, gas-insulated switchgears, cubicle-type gas-insulated switchgears, generators, rotating machines, instrument transformers, and instrument transformers.

[0014] When an insulation defect occurs inside the electric device 900 and partial discharge occurs, ultrasonic waves are generated. To detect these ultrasonic waves, a plurality of ultrasonic sensors 2-1, 2-2, 2-3, ..., 2-n (collectively referred to as ultrasonic sensors 2) are fixedly attached to the surface of the electric device 900. In the drawing, a state in which four ultrasonic sensors are attached (n=4) is depicted, but the present invention is not limited to this.

[0015] Ultrasonic waves are also called acoustic emissions, and ultrasonic sensors are also called AE sensors. When partial discharge occurs in the main circuit of the electrical device 900, ultrasonic waves are generated and travel through space, propagating through the housing of the electrical device 900 and generating a voltage in the piezoelectric element on the sensing surface of the ultrasonic sensor 2 attached to the frame.

[0016] The signal output from the ultrasonic sensor 2 is a weak, high-frequency signal, but may be superimposed with external surges, etc., so the surges are absorbed by arresters 3-1, 3-2, 3-3, ..., 3-n (collectively, arresters 3). Arrestors 3 may be made of zinc oxide elements, or alternatively, diodes or gap arresters may be used.

[0017] After that, bandpass filters 4-1, 4-2, 4-3, ..., 4-n (collectively referred to as bandpass filters 4) are installed to reduce noise and extract the ultrasonic signals. Generally, the frequency components of ultrasonic signals are between 10 kHz and 1 MHz, so a filter that only passes this band is used. If the frequency components of the noise can be identified, it is desirable to use a bandpass filter that does not pass that band. Furthermore, the extracted weak signals are amplified by amplifiers 5-1, 5-2, 5-3, ..., 5-n (collectively referred to as amplifiers 5). An amplification factor of 20 dB to 150 dB is often used.

[0018] The sensor signals S1, S2, S3, ..., Sn (collectively sensor signals S) after being amplified by the amplifier 5 are sent to the discharge determination unit 6, and the discharge determination unit 6 determines that a partial discharge has occurred when the strength (signal strength) of the sensor signal S exceeds a first threshold value.

[0019] It is important to understand the relationship between the ultrasonic sensor output and the magnitude of the partial discharge in order to determine the magnitude of the partial discharge from the signal strength and judge the reliability of the electrical device 900. The attenuation rate of the ultrasonic signal strength varies depending on the size of the electrical device 900, the layout of the main circuit, the configuration of the frame, and the installation state of the ultrasonic sensor 2. The threshold value for determining the discharge is selected based on these characteristics.

[0020] The sensor signal S sent to the discharge determination unit 6 is also sent to the detection time difference calculation unit 7, and if the discharge determination unit 6 determines that a discharge has occurred, the detection time difference calculation unit 7 calculates the time at which each of the multiple ultrasonic sensors 2 detected the discharge. There are two methods for calculating the detection time difference. The first method is to read the peak value (first peak), also called the pulse, immediately after the waveform of the sensor signal S exceeds a second threshold (different from the first threshold) that is set above the noise level, and is the method described in the background art using Patent Document 1 as an example.

[0021] The second is a method using a cross-correlation coefficient, which is explained in the background art using Patent Document 2 as an example. This is a method for calculating the time difference at which the cross-correlation coefficient between the outputs of a plurality of ultrasonic sensors 2 becomes maximum.

[0022] For example, if the digital data waveforms of two signals are xi (i = 1, 2, ... N) and yi (1, 2, ... N), the cross-correlation coefficient Rxy(k) of these two waveforms is expressed by equation (1), where k is the amount of shift. Rxy(k) = Σ [i = 0, ..., N-1] (x(i) * y(i+k)) / (Nk) ... (1)

[0023] The time difference τ can be calculated using equation (2): τ = arg[m](max(Rxy)) (m=1, 2, ..., N) ... (2)

[0024] Rxy indicates the similarity between waveforms x(i) and y(i+k), and reaches its maximum value when the two waveforms are most similar. This method calculates the time difference from the k at which Rxy reaches its maximum. For the first peak, the detected time difference is obtained from only the rising portion of the signal, but this method is characterized by the fact that it calculates the detected time difference using the entire waveform, not just the rising portion.

[0025] The detection time difference calculated by the detection time difference calculation unit 7 is sent to the discharge position calculation unit 8, which calculates the discharge position based on the sent detection time difference. Assume that there are four ultrasonic sensors 2 (n=4), the detection times of each sensor are t1, t2, t3, and t4, and the discharge occurrence time is td. The installation positions of each sensor in the x, y, and z coordinates are (x1, y1, z1), (x2, y2, z2), (x3, y3, z3), and (x4, y4, z4), respectively. If the speed of sound is C and the discharge position is (xd, yd, zd) in the x, y, and z coordinates, these relationship equations can be expressed by the following four equations (Equation (3A) to Equation (3D)).

[0026] √((x1-xd) 2 + (y1-yd) 2 + (z1-zd) 2 ) = C(t1-td) ・・・(3A) √((x2-xd) 2 + (y2-yd) 2 + (z2-zd) 2 ) = C(t2-td) ・・・(3B) √((x3-xd) 2 + (y3-yd) 2 + (z3-zd) 2 ) = C(t3-td) ・・・(3C) √((x4-xd) 2 + (y4-yd) 2 + (z4-zd) 2 ) = C(t4-td) ... (3D)

[0027] The discharge position (xd, yd, zd) can be calculated by numerical calculation from these four equations (Equation (3A) to Equation (3D)). The calculation procedure will be explained with reference to the flowchart in FIG.

[0028] The detection time difference calculation unit 7 calculates the detection time difference and outputs the calculation result to the discharge position calculation unit 8 (step S100). Then, the discharge position calculation unit 8 sets appropriate initial values ​​xd0, yd0, zd0, and td0 for xd, yd, zd, and td. The initial values ​​input here do not need to be strictly assumed, and for simplicity, there is no problem in using 0 for all of them (if the setting is such that 0 is used) (step S110).

[0029] When the discharge position and detection time (xd, yd, zd, td) are the initial values ​​xd0, yd0, zd0, td0 set in step S110, the values ​​to be measured as pseudoranges (estimated pseudoranges) are calculated (step S120). The pseudoranges are the distances from the discharge source to each sensor, and the estimated pseudoranges are r10, r20, r30, and r40. The estimated pseudoranges r10, r20, r30, and r40 can be calculated by solving the following four simultaneous equations (Equations (4A) to (4D)).

[0030] r10 = √((x1-x0) 2 + (y1-y0) 2 + (z1-z0) 2 ) + td0...(4A) r20 = √((x2-x0) 2 + (y2-y0) 2 + (z2-z0) 2 ) + td0...(4B) r30 = √((x3-x0) 2 + (y3-y0) 2 + (z3-z0) 2 ) + td0...(4C) r40 = √((x4-x0) 2 + (y4-y0) 2 + (z4-z0) 2 ) + td0 ... (4D)

[0031] Next, the residual Δri defined by equation (5) for the estimated pseudoranges (r10, r20, r30, r40) relative to the actually measured pseudoranges (r1, r2, r3, r4) is calculated (step S130): Δri = ri - ri0 (5) where i is any value from 1 to 4.

[0032] Using equations (6A) to (6D), partial derivatives ∂ri / ∂x, ∂ri / ∂y, ∂ri / ∂z, and ∂ri / ∂td of the residual ri with respect to x, y, z, and td are calculated (step S140).

[0033] ∂ri / ∂x = -(xi-x o ) / ri0...(6A) ∂ri / ∂y = -(yi-y o) / ri0...(6B) ∂ri / ∂z = -(zi-x o ) / ri0...(6C) ∂ri / ∂td = 1...(6D)

[0034] The following four simultaneous equations (Equations (7A) to (7D)) are solved to calculate the errors Δx, Δy, Δz, and Δtd (step S150).

[0035] Δr1 = Δx(∂r1 / ∂x) + Δy(∂r1 / ∂y) +Δz(∂r1 / ∂z) +Δtd(∂r1 / ∂td) ...(7A) Δr2 = Δx(∂r2 / ∂x) + Δy(∂r2 / ∂y) +Δz(∂r2 / ∂z) +Δtd(∂r2 / ∂td) ...(7B) Δr3 = Δx(∂r3 / ∂x) + Δy(∂r3 / ∂y) + Δz(∂r3 / ∂z) + Δtd(∂r3 / ∂td) ・・・(7C) Δr4 = Δx(∂r4 / ∂x) + Δy(∂r4 / ∂y) +Δz(∂r4 / ∂z) +Δtd(∂r4 / ∂td) ... (7D)

[0036] Based on the errors Δx, Δy, Δz, and Δtd obtained in step S150, corrected values ​​(xd1, yd1, zd1, td1) of the initial values ​​xd0, yd0, zd0, and td0 are calculated using equations (8A) to (8D) (step S160).

[0037] xd1 = xd0 + Δx ... (8A) yd1 = yd0 + Δy ... (8B) zd1 = zd0 + Δz ... (8C) td1 = td0 + Δtd ... (8D)

[0038] It is determined whether Δx, Δy, Δz, and Δtd have converged (step S200). If they have not converged ("No"), the corrected values ​​are substituted for the initial values ​​and the process proceeds to step S120. The above procedure is repeated until Δx, Δy, Δz, and Δtd have converged. Convergence often occurs after a few times. The operational termination condition for the calculation can be set by specifying the number of calculations in advance, or by observing the transition of the initial values ​​xd0, yd0, zd0, and td0 during the repeated calculations to determine whether convergence has occurred.

[0039] If Δx, Δy, Δz, and Δtd converge ("Yes" in step S200), the corrected values ​​(xd1, yd1, zd1, td1) at this point in time are displayed as the calculated discharge position in the discharge position display unit 9 (step S300). By knowing the discharge position, i.e., the fault position, it is possible to identify the faulty part and determine whether operation can be continued and whether repair or replacement is required.

[0040] In this example, the detection time difference calculation unit 7 employs the method using a cross-correlation coefficient out of two methods for calculating the detection time. However, the present invention is not limited to this, and the method of reading the first peak may also be employed. It is conceivable to configure the application operating as the detection time difference calculation unit 7 to be interchangeable so that either method can be selected. Furthermore, the detection time difference calculation unit 7 may calculate the detection time difference using both of the two methods, the discharge position calculation unit 8 may calculate the discharge position based on each of the detection time differences, and the discharge position display unit 9 may display the calculation results of the discharge position obtained by each of the two methods together with a distinction between the methods.

[0041] To replace the application, a method suited to the installation target may be selected, and an application for executing the selected method may be installed from a terminal, etc. Alternatively, the selection may be made by inserting into a slot either a chip containing an application for executing a cross-correlation function or a chip containing an application for executing a method for reading the first peak.

[0042] This allows the application to be installed to be selected according to the state of the electrical equipment 900 to be diagnosed and the installation status of the ultrasonic sensor 2, making it possible to accurately locate the partial discharge position in the electrical equipment 900 to be diagnosed.

[0043] After the above-described application replacement, only the selected method will be used until the next replacement, but since only one license is required, costs can be reduced. However, it is also possible to store applications that execute two methods and select one as appropriate depending on the situation at the time, as will be described in the second embodiment below.

[0044] Second Embodiment In the first embodiment described above, an example was described in which one of the two methods for calculating the detection time is set in advance when or before the partial discharge detection device is installed. In the second embodiment, an example is described in which the method for calculating the detection time difference can be changed as needed even after installation. Figure 3 is a block diagram corresponding to Figure 1 for explaining the configuration and operation of the partial discharge detection device according to the second embodiment. Note that the basic configuration and operation relating to ultrasonic wave detection and discharge position calculation are the same in the second embodiment as in the first embodiment, and therefore explanations of similar parts will be omitted, and Figure 2 of the first embodiment will be used.

[0045] 3, the partial discharge detection device 1 according to the second embodiment is configured by adding a detection time difference calculation method switching unit 10 that switches between calculation methods for the detection time difference to the partial discharge detection device 1 described in the first embodiment. For example, the detection time difference calculation method switching unit 10 or the detection time difference calculation unit 7 stores programs for executing two methods for calculating the detection times in advance.

[0046] The detection time difference calculation unit 7 is configured to calculate the detection time for the sensor signal S using the selected method in response to a command from the detection time difference calculation method switching unit 10. Alternatively, the detection time difference calculation unit 7 is configured to calculate both the detection time difference due to the first peak and the detection time difference using the cross-correlation method based on the sensor signal S, and to switch the calculation results to be output to the discharge position calculation unit 8 in response to a command from the detection time difference calculation method switching unit 10.

[0047] The advantage of using the first peak in calculating the detection time difference is that it increases the accuracy of identifying the position because it is possible to focus on the signal component that has propagated via the shortest path even when the structure inside electrical device 900 is complex and the propagation path of the ultrasonic waves is complicated. However, if the signal strength is low, it may be difficult to identify the first peak, which may reduce the accuracy.

[0048] On the other hand, the advantage of using the cross-correlation method is that it is easy to achieve high accuracy in locating the signal by calculating the detection time difference by looking at the entire waveform even when the signal strength is low. Furthermore, if there are restrictions on the placement of multiple ultrasonic sensors 2 and the sensors cannot be placed to surround the expected discharge source, it is necessary to calculate the detection time difference with higher accuracy, and this method is more likely to achieve higher accuracy than the first peak method. However, if the structure inside the electrical device 900 is complex and the ultrasonic propagation path is complicated, the waveform shapes observed by each ultrasonic sensor 2 will differ significantly, which may reduce the accuracy in locating the signal.

[0049] As described above, calculation of the detection time difference using the first peak should be used when the internal structure of the electrical device 900 is complex and the propagation path of the ultrasonic waves is complicated. Also, it is appropriate to use this method when there are restrictions on the placement of multiple ultrasonic sensors 2 and the sensors can be placed to surround the expected discharge source. Oil-filled transformers, gas-insulated transformers, gas-insulated switchgears, and cubicle-type gas-insulated switchgears meet the conditions for using the first peak.

[0050] Calculation of the detection time difference using the cross-correlation method should be used when noise is large or the signal to be detected is small, or when the structure inside the electrical device 900 is relatively simple and the ultrasonic propagation path is simple. It is also appropriate to use this method when sensors cannot be arranged to surround the expected discharge source. Among the electrical devices 900, the following meet the conditions for using the cross-correlation method: a molded transformer, a generator, a rotating machine, a voltage transformer, a voltage transformer, a power cable, a power cable connection, and a power cable termination.

[0051] The detection time difference calculation method switching unit 10 switches the calculation method by operating a selector switch (not shown) that can be accessed from the outside, or by a user instruction via wireless or wired communication. This allows the detection time difference calculation method to be selected depending on the above-mentioned situation, thereby improving the accuracy of locating. Note that the detection time difference calculation method switching unit 10 may be configured in a terminal separate from the partial discharge detection device 1.

[0052] Embodiment 3 In the above-described embodiments 1 and 2, examples have been described in which the detection time difference calculation method is switched by a user operation, regardless of whether the device is installed or in operation. In this embodiment 3, an example will be described in which the partial discharge detection device automatically switches to the optimal calculation method based on the waveform of the obtained sensor signal.

[0053] 4, 5A, and 5B are diagrams for explaining the configuration and operation of a partial discharge detection device according to a third embodiment. Fig. 4 is a functional block diagram showing the relationship between the partial discharge detection device and the electrical equipment to be diagnosed. Fig. 5A is a flowchart showing the partial discharge detection device and the operation of switching between detection time difference calculation methods in the partial discharge location method. Fig. 5B is a flowchart showing the operation of switching between detection time difference calculation methods in the partial discharge detection device and the partial discharge location method according to a modified example. In the third embodiment, the basic configuration and operation related to ultrasonic wave detection and discharge location calculation are the same as in the second embodiment, and therefore a description of similar parts will be omitted. Also, as in the second embodiment, Fig. 2 of the first embodiment is used.

[0054] As shown in Fig. 4, the partial discharge detection device 1 according to the third embodiment has additional components compared to the partial discharge detection device 1 described in the second embodiment. The additional components are a waveform similarity calculation unit 11 that calculates the similarity between the waveforms detected by the sensors, a waveform similarity display unit 12 that displays the calculated waveform similarity, and a detection time difference calculation method selection unit 13 that receives the output of the waveform similarity calculation unit 11 and selects a method for calculating the detection time difference.

[0055] The waveform similarity calculation unit 11 can calculate waveform similarity using Euclidean distance, cosine similarity, dynamic time warping, cross-correlation, etc. These methods can compare not only the time domain but also the frequency domain obtained by Fourier transforming the sensor signal S (signal waveform).

[0056] The operation of switching the time difference calculation method in locating the partial discharge position will be described with reference to the flowchart in Fig. 5A. Upon receiving the sensor signals S from the plurality of ultrasonic sensors 2 (step S10), the waveform similarity calculation unit 11 calculates the waveform similarity between the plurality of sensor signals S (step S20).

[0057] The waveform similarity calculation unit 11 outputs the calculated waveform similarity to the waveform similarity display unit 12, where it is displayed so that the diagnostician (user) can recognize it (step S30). This allows the user to refer to the displayed waveform similarity information and determine whether to use the first peak or the cross-correlation method to calculate the detection time difference.

[0058] Furthermore, upon receiving information from the waveform similarity calculation unit 11 that the waveform similarity has been calculated, the detection time difference calculation method selection unit 13 accepts an input operation for selecting a method via, for example, the waveform similarity display unit 12, in order to allow the user to select a detection time difference (step S40). When the user performs an input operation for selection, the detection time difference calculation method selection unit 13 outputs the selected information (selection information) to the detection time difference calculation method switching unit 10. The detection time difference calculation method switching unit 10 adopts a method for calculating the detection time difference based on the selection information (step S50), and the detection time difference calculation unit 7 calculates the detection time difference using the adopted method.

[0059] Thereafter, the process proceeds to the steps of calculating the detection time difference and locating the discharge position (steps S100 to S300) as explained in Fig. 2 of embodiment 1. Note that the waveform similarity display unit 12 may also be configured in a terminal separate from the partial discharge detection device 1, similar to the detection time difference calculation method switching unit 10 explained in embodiment 2.

[0060] When the waveforms are highly similar, it is desirable to use the cross-correlation method, but in addition, as described above, the decision is made taking into consideration the complexity of the ultrasonic wave propagation path inside the electrical device 900 and the installation status of the sensor.

[0061] Modification: Alternatively, a judgment index may be provided so that the partial discharge detection device 1 automatically selects the method. As a modification, the operation in the case of automatic selection will be described with reference to the flowchart in FIG. 5B. Note that the operation of steps S10 to S20 by the waveform similarity calculation unit is the same as that described in FIG. 5A.

[0062] When the detection time difference calculation method selection unit 13 receives the waveform similarity information from the waveform similarity calculation unit 11, it determines whether the similarity is equal to or greater than a predetermined third threshold (different from the first threshold and the second threshold) (step S60). If it determines that the similarity is equal to or greater than the third threshold ("Yes" in step S60), it selects the cross-correlation method (step S70A) and causes the detection time difference calculation unit 7 to calculate the detection time difference using the cross-correlation method. On the other hand, if it determines that the similarity is less than the third threshold ("No" in step S60), it selects the first peak (step S70B) and causes the detection time difference calculation unit 7 to calculate the detection time difference using the first peak.

[0063] In either case, the waveform similarity display unit 12 displays the automatically selected method in addition to the waveform similarity (step S80).Then, the process proceeds to the steps of calculating the detection time difference and locating the discharge position (steps S100 to S300).

[0064] In this way, if the partial discharge detection device 1 automatically determines the calculation method, it is possible to eliminate the dependency of the selection result on individual users. Furthermore, there is no need for communication for input operations by the user, which also leads to a reduction in the amount of communication during diagnosis.

[0065] Fourth Embodiment In this fourth embodiment, an example will be described in which an ultrasonic signal is detected in accordance with the cycle of the AC voltage of a system connected to an electrical device. Figure 6 is a block diagram corresponding to Figure 1 for explaining the configuration and operation of a partial discharge detection device according to the fourth embodiment. Note that the fourth embodiment is also similar to the first embodiment except for detecting an ultrasonic signal in accordance with the cycle of the AC voltage, so a description of similar parts will be omitted and Figure 2 of the first embodiment will be used. Furthermore, the configurations described in the second and third embodiments may be added to enable the functions thereof to be exerted.

[0066] 6, the partial discharge detection device 1 according to the fourth embodiment is configured by adding a current transformer 14 for detecting the AC voltage of the system to the partial discharge detection device 1 described in the first embodiment. The current transformer 14 measures the waveform of the AC voltage of the system, for example, 50 Hz or 60 Hz, and outputs the waveform to the detection time difference calculation unit 7.

[0067] The detection time difference calculation unit 7 acquires a signal in one cycle of the AC waveform from the sensor signals S from each of the multiple ultrasonic sensors 2. The ultrasonic signal in one cycle is also called a φ-q pattern, and the detection time difference is calculated based on this signal using a cross-correlation method.

[0068] If the structure inside the electrical device 900 is complex, the propagation path of the ultrasonic waves is complex, and the sensor signal is unclear due to the local noise environment, it becomes difficult to determine the detection time difference using the first peak. Furthermore, ultrasonic waves are repeatedly reflected inside the electrical device 900, and the entire detected ultrasonic signal differs significantly from sensor to sensor, which may make it difficult to calculate the detection time difference using the cross-correlation method.

[0069] In contrast, when the sensor signal S is observed in one cycle of the system voltage, ultrasonic signals due to multiple discharges are often observed within one cycle, and by using the intervals between these signals as a feature, the accuracy of calculating the detection time difference using the cross-correlation coefficient can be improved. Also, by observing the sensor signal S for multiple cycles in one input cycle and performing averaging processing, noise can be removed and fluctuations in the timing of discharge occurrence can be eliminated, thereby further improving accuracy.

[0070] As shown in FIG. 7 , the discharge determination unit 6, the detection time difference calculation unit 7, the discharge position calculation unit 8, the detection time difference calculation method switching unit 10, the waveform similarity calculation unit 11, and the detection time difference calculation method selection unit 13 constituting the partial discharge detection device 1 may be configured as a single piece of hardware 100 including a processor 101 and a storage device 102. Although not shown, the storage device 102 includes a volatile storage device such as a random access memory and a non-volatile auxiliary storage device such as a flash memory. Alternatively, a hard disk auxiliary storage device may be used instead of the flash memory. The processor 101 executes a program input from the storage device 102. In this case, the program is input from the auxiliary storage device to the processor 101 via the volatile storage device. The processor 101 may output data such as calculation results to the volatile storage device of the storage device 102, or may store the data in the auxiliary storage device via the volatile storage device.

[0071] Although various exemplary embodiments and examples are described in this disclosure, the various features, aspects, and functions described in one or more embodiments are not limited to the application of a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless modifications not illustrated are contemplated within the scope of the technology disclosed in this specification. For example, this includes cases where at least one component is modified, added, or omitted, or where at least one component is extracted and combined with components of another embodiment.

[0072] For example, in each of the above embodiments, the number of sensors (n) is described as 4, but this is not limiting and may be 5 or more, as shown in each of the following embodiments. In that case, for example, instead of using all the sensor signals S, the top four sensor signals S with the highest similarity may be selected from the five or more sensor signals S and used to calculate the detection time difference.

[0073] Furthermore, for example, the frequency with which the similarity becomes low may be stored, and sensors with a high frequency of low similarity compared to other sensors may be displayed to the user. If the user uses this information to change the installation location of sensors with a high frequency of low similarity, the frequency with which the similarity becomes low can be reduced, and the accuracy of locating the discharge position can be further improved.

[0074] Fifth Embodiment In this fifth embodiment, a partial discharge location method will be described when five or more ultrasonic sensors are installed. Fig. 8 is a flowchart corresponding to Fig. 2 for explaining the configuration and operation of a partial discharge detection device according to the fifth embodiment, and the partial discharge location method. Note that the fifth embodiment is also similar to the first embodiment except for the method of calculating the partial discharge location, so the description of similar parts will be omitted and Fig. 1 of the first embodiment will be used. Furthermore, the configurations described in the second, third, and fourth embodiments may be added to enable the functions thereof to be exerted.

[0075] As shown in Fig. 8, the partial discharge detection device 1 according to the fifth embodiment includes step S150V instead of step S150 in the operation of the partial discharge detection device 1 described in the first embodiment. In step S150V, weighting is performed using the residual calculated in S130 in the calculation of Δx, Δy, Δz, and Δtd. The steps before step S140 and the steps after step S160 are the same as those in the first embodiment.

[0076] As explained in the first embodiment, partial discharge location can be determined with four ultrasonic sensors, and a calculation method for this case has been shown. On the other hand, when there are five or more sensors, it has also been shown that four signals to be used for location are selected and discarded based on the similarity in the cross-correlation method. In this case, information from the unselected sensors is discarded, but the partial discharge detection device 1 and partial discharge location method according to the fifth embodiment utilize this information to further improve location accuracy.

[0077] A specific calculation method is described below. The basic calculation is the same as that described in embodiment 1, and the number of variables and equations in the simultaneous equations changes depending on the number of sensors. If the number of sensors is n, the installation positions of each sensor in the x, y, and z coordinates are (x1, y1, z1), (x2, y2, z2), (x3, y3, z3), (x4, y4, z4), ..., (xn, yn, zn). Furthermore, the number of simultaneous equations in the simultaneous equations of equations (3A) to (3D), the simultaneous equations of equations (4A) to (4D), and the simultaneous equations of equations (7A) to (7D) is n.

[0078] First, to simplify the handling of the simultaneous equations of equations (4A) to (4D), the equations are expressed as matrices. Assuming that vectors Δx = [Δx, Δy, Δz, Δtd]^T and Δr = [Δr1, Δr2, Δr3, Δr4, ..., Δrn]^T, where ^T represents transpose, the simultaneous equations of equations (4A) to (4D) can be simply expressed as equation (9): GΔx = Δr ... (9) Furthermore, G is expressed by the following matrix (equation (10)).

[0079]

[0080] When n is 5 or greater, the number of equations is greater than the number of unknowns, so in this case it is common to calculate the solution using the least squares method, and the solution to equation (9) is obtained by equation (11): Δx = (G^T G^(-1))^(-1)G^TΔr ... (11) where ^T represents the transpose and ^(-1) represents the inverse matrix. Δx calculated in step S150V in Figure 8 is obtained in this way, and is used to carry out steps S160 and onwards.

[0081] After revisiting the calculation method for when n is greater than 4 as described above, the sensors used in the calculation are weighted based on the accuracy of the sensor information. The accuracy is determined using the residual Δri obtained in step S130 of FIG. 8. The weighting matrix W is a matrix whose diagonal components are Δr1, ..., Δrn and the other components are zero. This weighting matrix can be used to weight the equation and find a solution that minimizes the expected error by calculating Δx using equation (12): Δx = (G^TWG^(-1))^(-1)G^TWΔr ... (12) If n = 4, the effect of weighting cannot be obtained, but equation (12) can still be calculated.

[0082] In the position location calculation, steps S120 to S200 in FIG. 8 are repeated, but empirically it is effective to use the weighting calculation in equation (12) from the fifth time onwards.

[0083] Sixth Embodiment In this sixth embodiment, a partial discharge location method will be described for a case where five or more ultrasonic sensors are installed. Note that the sixth embodiment is also similar to the fifth embodiment except for the weighting method used in the calculation of partial discharge location, and therefore the description of similar parts will be omitted and FIG. 1 will be used in conjunction with FIG. 8 of the fifth embodiment. Furthermore, the configurations described in the second, third, fourth, and fifth embodiments may be added to enable the functions thereof to be exerted.

[0084] Ultrasonic waves generated by discharges travel complex paths as they propagate through electrical equipment, being reflected by components inside the equipment such as coils and insulators. As a result, the signal is attenuated, and as the detected signal becomes unclear due to being buried in noise, the accuracy of calculating the partial discharge location also decreases. This effect becomes particularly significant when the sensor is far from the discharge location, or when there are components between the sensor and the discharge location that obstruct the propagation of ultrasonic waves.

[0085] Therefore, in calculating the discharge position, the accuracy of locating the discharge position can be improved by assigning a larger weight to sensor information with a stronger detected signal strength. Specifically, if the detection strengths of the sensors are V1, V2, ..., Vn, the weight matrix W is a matrix whose diagonal components are V1, ..., Vn and the other components are zero. By calculating equation (12) using this weight matrix, the position location calculation is performed with priority given to information from a reliable sensor with a stronger signal strength, which ultimately improves the accuracy of the position location.

[0086] Seventh Embodiment In this seventh embodiment, a partial discharge location method will be described in which five or more ultrasonic sensors are installed. Note that the seventh embodiment is also the same as the sixth embodiment except for the weighting criteria in the calculation of partial discharge location, and therefore, description of similar parts will be omitted and FIG. 1 will be used, and FIG. 8 of the fifth embodiment and FIG. 4 of the third embodiment will be used for the calculation of the similarity that serves as the weighting criteria. Furthermore, the configurations described in the second, fourth, fifth, and sixth embodiments may be added to enable the functions thereof to be fulfilled.

[0087] If the structure inside the electrical device 900 is complex, the propagation path of the ultrasonic waves is complex, and the sensor signal is unclear due to the local noise environment, it becomes difficult to determine the detection time difference using the first peak. Furthermore, ultrasonic waves are repeatedly reflected inside the electrical device 900, and the entire detected ultrasonic signal differs significantly from sensor to sensor, which may make it difficult to calculate the detection time difference using the cross-correlation method.

[0088] Therefore, in calculating the discharge position, the accuracy of discharge position location can be improved by assigning a larger weight to sensor information with a higher similarity in the cross-correlation method. Specifically, the waveform similarity calculation step of step S20 described with reference to FIG. 5A or 5B is executed before at least step S150V in FIG. 8. Then, if the similarities of the sensors obtained by equation (1) are R1, R2, ..., Rn, the weight matrix W is set to a matrix whose diagonal components are R1, ..., Rn and whose other components are zero. By calculating equation (11) using this weight matrix, the position location calculation is performed with priority given to information from more reliable sensors, thereby improving the position location accuracy.

[0089] As described above, the partial discharge detection device 1 of the present disclosure includes a plurality of ultrasonic sensors 2 installed on the surface of the electric device 900, a detection time difference calculation unit 7 that calculates a detection time difference of ultrasonic waves from the sensor signals S of the plurality of ultrasonic sensors 2, and a discharge position calculation unit 8 that calculates the occurrence position of partial discharge that has occurred inside the electric device 900 from the calculation result of the detection time difference, and is configured so that either a method of reading a first peak or a cross-correlation method can be selected as a method used to calculate the detection time difference. This makes it possible to select a method appropriate for the situation and to accurately identify the occurrence position of discharge in the electric device 900.

[0090] Furthermore, if an input device (detection time difference calculation method selection unit 13) for selecting the method to be used for calculation is provided, it is possible to select the most suitable method in response to (changes in) the situation.

[0091] By providing a waveform similarity calculation unit 11 that calculates the similarity of the waveforms of the sensor signals S of the plurality of ultrasonic sensors 2, an index for selecting a method can be obtained, and automatic selection becomes possible.

[0092] If a display unit (waveform similarity display unit 12) that displays the calculation results of the similarity is provided, an index for selecting a method can be presented to the user, enabling an appropriate method to be selected.

[0093] If the detection time difference calculation method switching unit 10 is configured to select the cross-correlation method as the calculation method when the calculated similarity is equal to or greater than a threshold (third threshold), and select the method of reading the first peak when the calculated similarity is less than the threshold (third threshold), the appropriate method is switched depending on the waveform similarity, so that the position where discharge occurs in the electrical device 900 can be accurately identified depending on the situation.

[0094] If the number of ultrasonic sensors 2 is five or more (n≧5), and the detection time difference calculation unit 7 is configured to calculate the detection time difference using the signals of the top four sensors in descending order of similarity when using the cross-correlation method, the discharge occurrence position can be identified more accurately.

[0095] When the electrical device 900 is provided with a current transformer 14 that measures the AC voltage of the system that supplies power to the electrical device 900 and the cross-correlation method is used, the detection time difference calculation unit 7 can be configured to calculate the detection time difference using the signal (φ-q pattern) in one cycle of the AC voltage from the sensor signals S of each of the multiple ultrasonic sensors 2, thereby further improving the calculation accuracy by, for example, using the interval between signals due to multiple discharges as a feature.

[0096] The number of ultrasonic sensors 2 is five or more (n≧5), and the discharge position calculation unit 8 calculates the difference between the value (estimated pseudo distance: r10, r20, r30, r40) set from the initial value set as the generation position used in the repeated value calculation and the calculation result based on the detection time difference (pseudo distance: r1, r2, r3, r4) as the estimated error Δri. If the information from five or more sensors is weighted more heavily so that the smaller the estimated error Δri, the more the generation position is calculated, thereby further improving the position location accuracy.

[0097] The position location accuracy can be further improved by having five or more (n≧5) sensors as the multiple ultrasonic sensors 2 and having the discharge position calculation unit 8 calculate the generation position by assigning a higher weight to the information from sensors with higher detection signal strength for the five or more sensors.

[0098] The position location accuracy can be further improved by having five or more (n≧5) sensors as the multiple ultrasonic sensors 2 and having the discharge position calculation unit 8 calculate the generation position by weighting the information from sensors with higher similarity more highly.

[0099] Furthermore, the partial discharge position locating method disclosed herein includes a detection time difference calculation step (step S100) of calculating a detection time difference of ultrasonic waves from the respective sensor signals S of the multiple ultrasonic sensors 2 installed on the surface of the electric device 900, a discharge position calculation step (steps S110 to S200) of calculating the occurrence position of partial discharge occurring inside the electric device 900 from the calculation result of the detection time difference, and a calculation method selection step (steps S60, S70A, S70B) of selecting either a method of reading a first peak or a cross-correlation method as a method to be used for calculating the detection time difference. This makes it possible to select a method appropriate for the situation, and to accurately identify the discharge occurrence position in the electric device 900.

[0100] If the method is configured to include a waveform similarity calculation step (step S20) for calculating the similarity of the waveforms of the sensor signals S of the multiple ultrasonic sensors 2, an index for selecting a method can be obtained, and automatic selection becomes possible.

[0101] If the method is configured to include a display step (step S30) of displaying the calculation results of the similarity, an index for selecting a method can be presented to the user, enabling an appropriate method to be selected.

[0102] In the calculation method selection step (steps S60, S70A, S70B), if the calculated similarity is equal to or greater than a threshold (third threshold), the cross-correlation method is selected as the method to be used for calculation, and if the calculated similarity is less than the threshold (third threshold), the calculation method is switched to select a method to read the first peak.Since the appropriate method is switched depending on the similarity of the waveforms, the position where discharge occurs in electrical device 900 can be accurately identified depending on the situation.

[0103] In the detection time difference calculation step (step S100), when the cross-correlation method is used, if five or more signals are input as the sensor signal S, the detection time difference can be calculated using the top four signals in order of highest similarity, thereby making it possible to more accurately identify the location of the discharge occurrence.

[0104] When the cross-correlation method is used and the step of measuring the AC voltage of the system that supplies power to the electrical device 900 is included, the detection time difference calculation step may be configured to calculate the detection time difference using the signal (φ-q pattern) in one cycle of the AC voltage from the sensor signals S of each of the multiple ultrasonic sensors 2. This allows for further improvement in the calculation accuracy, for example, by using the interval between signals due to multiple discharges as a feature.

[0105] In the detection time difference calculation step (step S100), the detection time difference of ultrasonic waves is calculated from the sensor signals S of five or more sensors as the plurality of ultrasonic sensors 2, and in the discharge position calculation step (steps S110 to S200), the difference between the value set from the initial value set as the generation position used in the repeated calculation (estimated pseudo distances: r10, r20, r30, r40) and the calculation result based on the detection time difference (pseudo distances: r1, r2, r3, r4) is calculated as the estimated error Δri, and for the five or more sensor signals, the smaller the estimated error Δri, the higher the weight is assigned to the information of the sensor signal S to calculate the generation position, thereby further improving the position location accuracy.

[0106] In the detection time difference calculation step (step S100), the detection time difference of ultrasonic waves is calculated from the sensor signals S of five or more sensors as the multiple ultrasonic sensors 2, and in the discharge position calculation step (steps S110 to S200), the information of the five or more sensor signals S is weighted more heavily so that the stronger the detection signal strength is, and the generation position is calculated, thereby further improving the position location accuracy.

[0107] In the detection time difference calculation step (step S100), the detection time difference of ultrasonic waves is calculated from the sensor signals S of five or more sensors as the multiple ultrasonic sensors 2, and in the discharge position calculation step (steps S110 to S200), the position location accuracy can be further improved by weighting the information of the sensor signals S with higher similarity to calculate the generation position.

[0108] 1: Partial discharge detection device, 2: Ultrasonic sensor, 3: Arrester, 4: Bandpass filter, 5: Amplifier, 6: Discharge determination unit, 7: Detection time difference calculation unit, 8: Discharge position calculation unit, 9: Discharge position display unit, 10: Detection time difference calculation method switching unit, 11: Waveform similarity calculation unit, 12: Waveform similarity display unit, 13: Detection time difference calculation method selection unit, 14: Current transformer, 900: Electrical equipment, S: Sensor signal.

Claims

1. A plurality of ultrasonic sensors installed on the surface of an electrical device, a detection time difference calculation unit that calculates a detection time difference of ultrasonic waves from the sensor signals of each of the plurality of ultrasonic sensors, and a discharge position calculation unit that calculates the occurrence position of partial discharge generated inside the electrical device from the calculation result of the detection time difference, and is provided with, A partial discharge detection device characterized in that either a method of reading a first peak or a cross-correlation method can be selected as a method used for calculating the detection time difference.

2. The partial discharge detection device according to claim 1, further comprising an input device for selecting a method used for the calculation.

3. The partial discharge detection device according to claim 1, further comprising a waveform similarity calculation unit that calculates the similarity of waveforms of sensor signals of each of the plurality of ultrasonic sensors.

4. The partial discharge detection device according to claim 3, further comprising a display unit that displays the calculation result of the similarity.

5. A detection time difference calculation method switching unit that switches the calculation method so that when the calculated similarity is equal to or greater than a threshold value, the cross-correlation method is selected as the method used for the calculation, and when it is less than the threshold value, the method of reading the first peak is selected. The partial discharge detection device according to claim 3, characterized in that it is provided.

6. Having 5 or more sensors as the plurality of ultrasonic sensors, When using the cross-correlation method, the detection time difference calculation unit calculates the detection time difference using signals of the top 4 sensors in descending order of the similarity. The partial discharge detection device according to any one of claims 3 to 5.

7. Comprising a current transformer that measures the AC voltage of a system that supplies power to the electrical device, When using the cross-correlation method, the detection time difference calculation unit calculates the detection time difference using signals in one cycle of the AC voltage among the sensor signals of each of the plurality of ultrasonic sensors. The partial discharge detection device according to any one of claims 1 to 5.

8. Having 5 or more sensors as the plurality of ultrasonic sensors, The discharge position calculation unit calculates the difference between the value calculated from the initial value set as the generation position used in the iterative calculation, the value calculated from the initial value set as the generation position used in the iterative calculation, and the calculation result based on the detected time difference as the estimation error, and for the five or more sensors, performs a higher weighting on the information of the sensors with smaller estimation errors to calculate the generation position, and is characterized in that it is the partial discharge detection device according to any one of claims 1 to 5.

9. having five or more sensors as the plurality of ultrasonic sensors, The discharge position calculation unit calculates, for the five or more sensors, to perform a higher weighting on the information of the sensors with higher detection signal intensity to calculate the generation position, and is characterized in that it is the partial discharge detection device according to any one of claims 1 to 5.

10. having five or more sensors as the plurality of ultrasonic sensors, The discharge position calculation unit performs a higher weighting on the information of the sensors with higher similarity for the five or more sensors to calculate the generation position, and is characterized in that it is the partial discharge detection device according to any one of claims 3 to 5.

11. a detection time difference calculation step of calculating the detection time difference of ultrasonic waves from the sensor signals of each of the plurality of ultrasonic sensors installed on the surface of the electrical equipment, a discharge position calculation step of calculating the generation position of partial discharge occurring inside the electrical equipment from the calculation result of the detection time difference, and a calculation method selection step of selecting either a method of reading the first peak or a cross-correlation method as the method used for calculating the detection time difference, characterized in that it includes a partial discharge position calibration method.

12. characterized in that it includes a waveform similarity calculation step of calculating the similarity of the waveforms of the sensor signals of each of the plurality of ultrasonic sensors, as described in claim 11.

13. characterized in that it includes a display step of displaying the calculation result of the similarity, as described in claim 12.

14. In the calculation method selection step, when the calculated similarity is equal to or greater than the threshold value, the cross-correlation method is selected as the method used for the calculation, and when it is less than the threshold value, the method of reading the first peak is selected to switch the calculation method, and is characterized in that it is the partial discharge position calibration method according to claim 12.

15. In the detection time difference calculation step, when using the cross-correlation method, When five or more signals are input as the sensor signals, the method for calibrating the partial discharge position according to any one of claims 12 to 14, characterized in that the detection time difference is calculated using the top four signals in descending order of the similarity.

16. including a step of measuring an AC voltage of a system that supplies power to the electrical equipment, When using the cross-correlation method, in the detection time difference calculation step, the detection time difference is calculated using the signals in one cycle of the AC voltage among the sensor signals of each of the plurality of ultrasonic sensors. The method for calibrating the partial discharge position according to any one of claims 11 to 14.

17. In the detection time difference calculation step, the detection time difference of the ultrasonic wave is calculated from the sensor signals of each of five or more sensors as the plurality of ultrasonic sensors, In the discharge position calculation step, the difference between the value calculated from the initial value set as the generation position used in the iterative calculation and the value calculated from the initial value set as the generation position used in the iterative calculation and the calculation result based on the detection time difference is calculated as the estimation error, and for the five or more sensor signals, the information of the sensor signals with a smaller estimation error is weighted more highly to calculate the generation position. The method for calibrating the partial discharge position according to any one of claims 11 to 14.

18. In the detection time difference calculation step, the detection time difference of the ultrasonic wave is calculated from the sensor signals of each of five or more sensors as the plurality of ultrasonic sensors, In the discharge position calculation step, for the five or more sensor signals, the information of the sensor signals with a higher detection signal intensity is weighted more highly to calculate the generation position. The method for calibrating the partial discharge position according to any one of claims 11 to 14.

19. In the detection time difference calculation step, the detection time difference of the ultrasonic wave is calculated from the sensor signals of each of five or more sensors as the plurality of ultrasonic sensors, In the discharge position calculation step, the information of the sensor signals with a higher similarity is weighted more highly to calculate the generation position. The method for calibrating the partial discharge position according to any one of claims 12 to 14.