Distributed remote probe ultrasonic partial discharge test method and system

Abstract

The invention discloses a distributed remote probe ultrasonic partial discharge test method and system, and belongs to the technical field of power systems. The partial discharge test method comprises the following steps: obtaining a plurality of ultrasonic signal groups, obtaining a signal attribute from each ultrasonic signal group, obtaining a basic attribute at the same time, and standardizing the signal attributes to form a standard signal attribute point group; obtaining a core variable according to the basic attribute and a test requirement, wherein the test requirement at least comprises one of test accuracy and test efficiency; according to the core variable, different clusters are obtained after calculation through a BDSCAN algorithm, the different clusters form a clustering result corresponding to the partial discharge probe, and the adjacent distribution intensity of the clusters is obtained; and determining that the adjacent distribution intensity of a certain cluster of a certain clustering result is suddenly changed compared with clustering results obtained by other partial discharge probes, and indicating that partial discharge defects exist.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular, to a distributed remote probe ultrasonic partial discharge testing method and system. Background Art

[0002] Partial discharge in GIS is an early manifestation of the deterioration of the insulation system, and its hazards are progressive and hidden. First, the charged particles generated by partial discharge directly impact the insulating material, resulting in the destruction of the molecular structure. For example, fiber materials may be fragmented due to repeated impacts, accelerating the decline of insulation performance. Second, the energy concentration in the discharge area causes a local temperature rise, and long-term action may lead to carbonization of the insulating material, further reducing the dielectric strength. In addition, ozone and nitrogen oxides generated during the discharge process combine with moisture in the air to form nitric acid, which chemically corrodes organic insulating materials, and this process is particularly significant in a humid environment. For oil-immersed equipment, partial discharge also causes oil decomposition, forming conductive sludge, increasing dielectric loss and hindering heat dissipation, and ultimately may lead to thermal breakdown.

[0003] If partial discharge defects are not dealt with in time, they will gradually expand, and may even lead to insulation breakdown or even equipment explosion. In existing solutions, ultrasonic signals output by partial discharge probes are often used to form spectrograms after signal processing for manual judgment. Each test requires careful viewing of the spectrograms, greatly increasing the labor cost. Sometimes, the staff may make misjudgments. Moreover, within 24 hours after the equipment is put into operation, the test personnel test closely on the equipment, which has a certain safety risk. Summary of the Invention

[0004] To solve the deficiencies in the prior art, the present invention provides a distributed remote probe ultrasonic partial discharge testing method and system.

[0005] The present invention adopts the following technical solutions:

[0006] The present invention discloses a distributed remote probe ultrasonic partial discharge testing method on the one hand, including: an amplifier, a partial discharge probe, a partial discharge detection device, a terminal, and an Ethernet. Each partial discharge probe is configured with an amplifier, the amplifier is connected to the partial discharge detection device, and the partial discharge probe, the partial discharge detection device, and the terminal can be connected via the Ethernet. The method includes the following steps:

[0007] Obtain a plurality of ultrasonic signal groups, obtain signal attributes from each ultrasonic signal group, and at the same time obtain basic attributes. Standardize the signal attributes to form a standard signal attribute point group. The standardization is to perform normalization processing on the information in the signal attributes; the basic attributes include: the density of the standard signal attributes and the quantity order of magnitude of the standard signal attributes;

[0008] Derive the core variables based on the basic attributes and test requirements, where the test requirements include at least one of test accuracy and test efficiency;

[0009] According to the core variables, after performing operations using the BDSCAN algorithm, different clusters are obtained. Different clusters constitute the clustering results corresponding to the partial discharge probe, and the adjacent distribution intensity of the clusters is obtained;

[0010] If it is confirmed that the adjacent distribution intensity of a certain cluster in a certain clustering result mutates compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect.

[0011] According to the described distributed remote probe ultrasonic partial discharge test method, the core variables include the minimum number of neighborhood points m and the neighborhood radius e. m is a non-zero natural number, and e is a positive real number;

[0012] Obtain different clusters through the BDSCAN algorithm;

[0013] Select a certain cluster, select the starting standard signal attribute point one from this cluster, and use the starting standard signal attribute point one as the center and e as the radius to draw a circle to form a neighborhood circle;

[0014] If there are greater than or equal to m standard signal attribute points in the neighborhood circle, then all the standard signal attribute points within this neighborhood circle are called adjacent standard signal attribute points one;

[0015] If there are less than m standard signal attribute points in the neighborhood circle, then this starting standard signal attribute point one is called a noise point;

[0016] Calculate the adjacent distribution intensity value of the starting standard signal attribute point one.

[0017] According to the described distributed remote probe ultrasonic partial discharge test method, select the starting adjacent standard signal attribute point one, use the starting adjacent standard signal attribute point one as the center and e as the radius to draw a circle to form a neighborhood circle, and traverse all the adjacent standard signal attribute points one;

[0018] If there are greater than or equal to m standard signal attributes in the neighborhood circle, then all the standard signal attributes within this neighborhood circle are called adjacent standard signal attribute points two;

[0019] If there are less than m standard signal attributes in the neighborhood circle, then this starting adjacent standard signal attribute point one is called a noise point;

[0020] Calculate the adjacent distribution intensity values of each starting adjacent standard signal attribute point one;

[0021] Repeat the above steps until all the adjacent standard signal attribute points have no subordinate adjacent standard signal attribute points, and calculate the adjacent distribution intensity values of all the standard signal attribute points within this cluster.

[0022] According to the described distributed remote probe ultrasonic partial discharge testing method, reselect a cluster, and then select the starting standard signal attribute point 1 in this cluster, and repeat the above steps until all clusters are traversed, and calculate the adjacent distribution intensity values of all standard signal attribute points within each cluster.

[0023] According to the described distributed remote probe ultrasonic partial discharge testing method, the value of m is: 2D - 1, where D is the dimension of the standard signal attribute point.

[0024] According to the described distributed remote probe ultrasonic partial discharge testing method, e can be calculated by the following formula:

[0025]

[0026] In the formula, M is the area surrounded by all standard signal attribute points.

[0027] According to the described distributed remote probe ultrasonic partial discharge testing method, calculate the adjacent distribution intensity of each cluster by the following formula:

[0028]

[0029] In the formula, Apdiv represents the adjacent distribution intensity value of the standard signal attribute points within the cluster, and n represents the number of standard signal attribute points within the cluster.

[0030] According to the described distributed remote probe ultrasonic partial discharge testing method, obtain parameter 1 based on the standard signal attribute density, and parameter 1 is positively correlated with the standard signal attribute density; obtain parameter 2 based on the standard signal attribute quantity order, and parameter 2 is negatively correlated with the standard signal attribute quantity order; obtain parameter 3 based on the test accuracy, and parameter 3 is positively correlated with the test accuracy; obtain parameter 4 based on the test efficiency, and parameter 4 is negatively correlated with the test efficiency; optimize the basic core variable by at least one of parameter 1, parameter 2, parameter 3, and parameter 4 to obtain the core variable.

[0031] According to the described distributed remote probe ultrasonic partial discharge testing method, each partial discharge probe corresponds to a clustering result. When the partial discharge detection device compares the clustering results corresponding to each partial discharge probe and finds that the adjacent distribution intensity of a certain cluster in a certain clustering result changes compared with other clustering results, it indicates that there is partial discharge in the GIS device, alarms the staff, and at the same time transmits the clustering result to the terminal in the form of a picture.

[0032] On the other hand, the present invention discloses a distributed remote probe ultrasonic partial discharge testing system, including:

[0033] A processing unit, which is used to process the obtained ultrasonic signal group, obtain signal attributes from the ultrasonic signal group, and at the same time obtain basic attributes. The signal attributes at least include one of the standard signal attribute density and the standard signal attribute quantity magnitude;

[0034] A selection unit, which is used to select core variables according to the basic attributes and test requirements. The test requirements at least include one of the test accuracy and the test efficiency;

[0035] An operation unit, which is used to calculate different clusters according to the core variables through the BDSCAN algorithm and obtain the adjacent distribution intensity of the clusters;

[0036] A test unit, which is used to confirm that if the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect.

[0037] Compared with the prior art, the beneficial effects of the present invention at least include:

[0038] 1. The present invention obtains a plurality of ultrasonic signal groups, obtains signal attributes from each ultrasonic signal group, and at the same time obtains basic attributes, and performs standardization processing on the signal attributes to form a standard signal attribute point group. The standardization is to perform normalization processing on the information in the signal attributes to avoid the influence of the magnitude on clustering, and obtain core variables according to the obtained basic attributes and test requirements; according to the core variables, different clusters are obtained through the BDSCAN algorithm operation. The different clusters constitute the clustering result corresponding to this partial discharge probe, and the adjacent distribution intensity of the clusters is obtained. The adjacent distribution intensity is the degree of aggregation of attribute points in space. If the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect. The partial discharge defect is accurately judged through the adjacent distribution intensity of the standard signal attribute points, improving the test accuracy of the partial discharge fault, thereby preventing the GIS insulation breakdown and delaying the performance attenuation.

[0039] 2. After the present invention tests out the partial discharge defect, it can alarm the staff, and at the same time transmit the clustering result to the terminal in the form of a picture, which is convenient for the staff to analyze the type and location of the partial discharge, and is convenient for timely maintenance. Description of the Drawings

[0040] Figure 1 It is the basic device diagram of the present invention.

[0041] Figure 2 It is the process diagram of the partial discharge test method of the present invention.

[0042] Figure 3 It is the process diagram of the method for obtaining the adjacent distribution intensity values of each standard signal attribute point in the cluster through the BDSCAN algorithm operation of the present invention.

[0043] Figure 4 It is a process diagram of the method for the present invention to obtain core variables based on basic attributes and test requirements.

[0044] Figure 5 It is a diagram of the partial discharge test system of the present invention. Detailed implementation manners

[0045] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in the present invention are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0046] Please refer to Figures 1 - 5 , the present invention provides a distributed remote probe ultrasonic partial discharge test method and system.

[0047] If partial discharge defects are not dealt with in time, they will gradually expand, and may even lead to insulation breakdown or even equipment explosion. The characterization of partial discharge faults indicates that there may be defects in GIS, such as metal particles inside GIS equipment. In the existing solutions, ultrasonic signals output by partial discharge probes are often used to form spectrograms after signal processing for manual judgment. Each test requires careful viewing of the spectrograms, which greatly increases the labor cost, and sometimes the staff may make misjudgments.

[0048] Please refer to Figure 1 , the present invention provides a distributed remote probe ultrasonic partial discharge test device, comprising: an amplifier 1, a partial discharge probe 2, a partial discharge detection device 3, a terminal 4 and an Ethernet 5.

[0049] A plurality of partial discharge probes 2 are provided and arranged on the GIS housing at equal intervals, and the interval is less than 1 m. Each partial discharge probe 2 is configured with an amplifier 1, and the amplifier 1 is connected to the partial discharge detection device 3.

[0050] The partial discharge probe 2 is used to continuously obtain ultrasonic signals, and after the obtained ultrasonic information is amplified and processed by the amplifier 1 to form an ultrasonic information group, the ultrasonic signal group is then transmitted to the partial discharge detection device 3. The partial discharge detection device 3 performs partial discharge detection operations on the ultrasonic signal group and transmits the partial discharge detection results to the terminal 4.

[0051] In a preferred but non-limiting embodiment of the present invention, if partial discharge occurs, after the partial discharge detection device 3 analyzes the partial discharge information based on the ultrasonic signal group, it can give a warning to the terminal 4 to prompt the staff to handle or pay attention to the partial discharge problem. Here, the partial discharge probe 2, the partial discharge detection device 3, and the terminal 4 can transmit information to each other via the Ethernet 5.

[0052] It should be noted that the types and numbers of the amplifier 1, the partial discharge probe 2, the partial discharge detection device 3, the terminal 4, and the Ethernet 5 can be selected according to different scenarios with different combination schemes.

[0053] The present invention also provides a distributed remote probe ultrasonic partial discharge test method. A number of ultrasonic signal groups are obtained by the cooperation of each partial discharge probe 2 and the amplifier 1, and the number of the ultrasonic signal groups is the same as the number of the partial discharge probes 2. Signal attributes are obtained from each ultrasonic signal group, and at the same time, basic attributes are obtained. Since the information dimensions in the signal attributes are different, the signal attributes need to be standardized to form a standard signal attribute point group. The standardization is to normalize the information in the signal attributes to avoid the influence of magnitude on clustering, and the core variables are obtained according to the obtained basic attributes and test requirements. According to the core variables, different clusters are obtained after the operation of the DBSCAN algorithm. Different clusters constitute the clustering result corresponding to the partial discharge probe 2, and the adjacent distribution intensity of the clusters is obtained. The adjacent distribution intensity is the degree of aggregation of attribute points in space. If the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes 2, it indicates that there is a partial discharge defect. The partial discharge defect is accurately judged through the adjacent distribution intensity of the standard signal attribute points, improving the accuracy of the partial discharge fault test, thus preventing the GIS insulation breakdown and delaying the performance decay.

[0054] Here, the cluster may be an isolated point.

[0055] Please refer to Figure 2 , the specific steps of the distributed remote probe ultrasonic partial discharge test method include:

[0056] Step 1-1: Obtain a number of ultrasonic signal groups, obtain signal attributes from each ultrasonic signal group, and at the same time obtain basic attributes, and standardize the signal attributes to form a standard signal attribute point group. The standardization is to normalize the information in the signal attributes.

[0057] The signal attributes include: amplitude, frequency, phase, peak value, and effective value, etc. At least the first three items are included in the present invention. The basic attributes include: the density of the standard signal attributes and the quantity order of the standard signal attributes.

[0058] Step 1-2: Obtain the core variables according to the basic attributes and test requirements.

[0059] The test requirements shall include at least one of test accuracy and test efficiency.

[0060] Steps 1-3: According to the core variables, different clusters are obtained through the BDSCAN algorithm operation. The different clusters constitute the clustering result corresponding to the partial discharge probe 2, and the adjacent distribution intensity of the clusters is obtained.

[0061] Step 1-4: If it is confirmed that the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes 2, it indicates that there is a partial discharge defect.

[0062] Here, the above method is executed by the partial discharge detection device 3.

[0063] The partial discharge detection device 3 can obtain each ultrasonic signal group. Each ultrasonic signal group is obtained by each partial discharge probe 2 continuously collecting the GIS device. In another embodiment, the obtained ultrasonic signals can be denoised first. The denoising method can adopt the wavelet transform denoising method, and then the denoised ultrasonic signals are formed into ultrasonic signal groups.

[0064] After the ultrasonic signal group is obtained, the signal attributes can be obtained from each ultrasonic signal group, and the basic attributes are obtained at the same time. The signal attributes shall include at least amplitude, frequency and phase. The basic attributes include: the density of the standard signal attributes and the quantity order of the standard signal attributes. The signal attributes can be obtained by the Fourier transform method and other methods.

[0065] The test requirements can also be obtained. The test requirements shall include at least one of test accuracy and test efficiency. The core variables can be obtained according to the obtained basic attributes and test requirements.

[0066] After the core variables are obtained, different clusters can be obtained through the BDSCAN algorithm operation according to the core variables, and the adjacent distribution intensity of the clusters is obtained. The ultrasonic signal group is obtained through the cooperation of each partial discharge probe 2 and the amplifier 1. The signal attributes are obtained from each ultrasonic signal group, and the basic attributes are obtained at the same time. Then the signal attributes are standardized to form a standard signal attribute point group. The core variables are obtained according to the obtained basic attributes and test requirements. Different clusters are obtained through the BDSCAN algorithm operation according to the core variables. The different clusters constitute the clustering result corresponding to the partial discharge probe 2, and the adjacent distribution intensity of the clusters is obtained. The adjacent distribution intensity is the degree of aggregation of attribute points in space. If the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes 2, it indicates that there is a partial discharge defect. The partial discharge defect is accurately judged through the adjacent distribution intensity of the standard signal attribute points, improving the test accuracy of the partial discharge fault, thus preventing the GIS insulation breakdown and delaying the performance degradation.

[0067] Here, the core variables include the minimum number of neighborhood points m and the neighborhood radius e. Here, m is a non-zero natural number and e is a positive real number.

[0068] Different clusters are obtained through the BDSCAN algorithm.

[0069] Please refer to Figure 3 , the present invention also provides a method for obtaining the proximity distribution intensity values of each standard signal attribute point in a cluster through the operation of the BDSCAN algorithm, specifically including:

[0070] Step 2-1: Select a certain cluster, select the first starting standard signal attribute point from this cluster, and use the first starting standard signal attribute point as the center and e as the radius to draw a circle to form a neighborhood circle.

[0071] Step 2-1-1: If there are greater than or equal to m standard signal attribute points in the neighborhood circle, then all the standard signal attribute points in this neighborhood circle are called the first neighboring standard signal attribute points.

[0072] Step 2-1-2: If there are less than m standard signal attribute points in the neighborhood circle, then this first starting standard signal attribute point is called a noise point.

[0073] Here, the first starting standard signal attribute point is any standard signal attribute point in the standard signal attribute point group.

[0074] Step 2-2: Calculate the proximity distribution intensity value of the first starting standard signal attribute point.

[0075] Here, the number of the first neighboring standard signal attribute points is called the proximity distribution intensity value of the first starting standard signal attribute point, and the proximity distribution intensity value of a noise point is 0.

[0076] When calculating the proximity distribution intensity value of the first starting standard signal attribute point, the first starting standard signal attribute point can be selected first, and a circle is drawn with the first starting standard signal attribute point as the center and e as the radius to form a neighborhood circle. If there are greater than or equal to m standard signals in the neighborhood circle, then all the standard signal attributes in this neighborhood circle are called the first neighboring standard signal attribute points; otherwise, this first starting standard signal attribute point is called a noise point.

[0077] Step 2-3: Select the first starting neighboring standard signal attribute point, use the first starting neighboring standard signal attribute point as the center and e as the radius to draw a circle to form a neighborhood circle, and traverse all the first neighboring standard signal attribute points.

[0078] Step 2-3-1: If there are greater than or equal to m standard signal attributes in the neighborhood circle, then all the standard signal attributes in this neighborhood circle are called the second neighboring standard signal attribute points.

[0079] Step 2-3-2: If there are less than m standard signal attributes in the neighborhood circle, then the starting adjacent standard signal attribute point 1 is called a noise point.

[0080] Step 2-4: Calculate the neighborhood distribution intensity value of each starting adjacent standard signal attribute point 1.

[0081] Here, the number of adjacent standard signal attribute points 2 is called the neighborhood distribution intensity value of the starting standard signal attribute point 1. If the starting standard signal attribute point 1 is a noise point, then its neighborhood distribution intensity value is 0.

[0082] Step 2-5: Repeat the above steps until there are no subordinate adjacent standard signal attributes for all adjacent standard signal attribute points, and calculate the neighborhood distribution intensity values of all standard signal attributes in this cluster.

[0083] Step 2-6: Re-select a cluster, and then select the starting standard signal attribute point 1 in this cluster. Repeat the above steps until all clusters are traversed, and calculate the neighborhood distribution intensity values of all standard signal attributes in each cluster.

[0084] The value of the above m is: 2D - 1, where D is the dimension of the standard signal attribute point. Of course, the value of m can also be adjusted adaptively.

[0085] The above e can be calculated by the following formula:

[0086]

[0087] In the formula, M is the area surrounded by all standard signal attribute points.

[0088] Calculate the neighborhood distribution intensity of each cluster through the following formula:

[0089]

[0090] In the formula, Apdiv represents the neighborhood distribution intensity value of the standard signal attribute points in the cluster, and n represents the number of standard signal attribute points in this cluster.

[0091] Each partial discharge probe 2 corresponds to a clustering result. When the partial discharge detection device 3 compares the clustering results corresponding to each partial discharge probe 2 and finds that the neighborhood distribution intensity of a certain cluster in a certain clustering result suddenly increases, suddenly decreases, suddenly appears or disappears compared with other clustering results, it indicates that there is partial discharge in the GIS device, and alarms the staff. At the same time, the clustering result is transmitted to the terminal in the form of a picture, which is convenient for the staff to analyze the type and location of the partial discharge and facilitate timely maintenance.

[0092] Please refer to Figure 4 , the present invention also provides a method for obtaining core variables based on basic attributes and test requirements, specifically including:

[0093] Step 3-1: Obtain Parameter 1 according to the density of standard signal attributes.

[0094] Here, Parameter 1 is positively correlated with the density of standard signal attributes.

[0095] Step 3-2: Obtain Parameter 2 according to the quantity magnitude of standard signal attributes.

[0096] Here, Parameter 2 is negatively correlated with the quantity magnitude of standard signal attributes.

[0097] Step 3-3: Obtain Parameter 3 according to the test accuracy.

[0098] Here, Parameter 3 is positively correlated with the test accuracy.

[0099] Step 3-4: Obtain Parameter 4 according to the test efficiency.

[0100] Here, Parameter 4 is negatively correlated with the test efficiency.

[0101] Step 3-5: Optimize the basic core variable by at least one of Parameter 1, Parameter 2, Parameter 3, and Parameter 4 to obtain the core variable.

[0102] When obtaining the core variable according to the basic attributes and test requirements, Parameter 1 can be obtained first according to the density of standard signal attributes. Here, Parameter 1 is positively correlated with the density of standard signal attributes. In other words, if the standard signal attributes are dense, the value of e is small; if the standard signal attributes are sparse, the value of e needs to be increased.

[0103] Parameter 2 can be obtained according to the quantity magnitude of standard signal attributes. Here, Parameter 2 is negatively correlated with the quantity magnitude of standard signal attributes. For example, if the quantity magnitude of standard signal attributes is large, the operation cost can be reduced by reducing the value of e.

[0104] Parameter 3 can also be obtained according to the test accuracy, and Parameter 4 can be obtained according to the test efficiency. Here, Parameter 3 is positively correlated with the test accuracy, and Parameter 4 is negatively correlated with the test efficiency. In other words, Parameter 3 and Parameter 4 can be obtained according to the required test efficiency.

[0105] The basic core variable can also be obtained. Here, the basic core variable can be a core variable set according to experience, or a core variable used in the past, or a core variable obtained at the start of the algorithm operation. Then, by optimizing the basic core variable with at least one of Parameter 1, Parameter 2, Parameter 3, and Parameter 4, the core variable for the current operation of BDSCAN can be obtained. In other words, in the BDSCAN of the present invention, the core variable can be continuously fine-tuned according to the basic attributes and test requirements, thereby optimizing the operation accuracy and operation efficiency of BDSCAN, and increasing the accuracy of partial discharge detection and test efficiency.

[0106] Please refer to Figure 5 , the present invention also provides a distributed remote probe ultrasonic partial discharge test system, specifically including:

[0107] A processing unit 402, configured to process the obtained ultrasonic signal group, obtain signal attributes from the ultrasonic signal group, and simultaneously obtain basic attributes. The signal attributes at least include one of the standard signal attribute density and the standard signal attribute quantity magnitude.

[0108] A selection unit 403, configured to select core variables according to the basic attributes and test requirements. The test requirements at least include one of test accuracy and test efficiency.

[0109] An operation unit 404, configured to obtain different clusters through BDSCAN algorithm operation based on the core variables, and obtain the adjacent distribution intensity of the clusters.

[0110] A test unit 405, configured to confirm that if the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes 2, it indicates that there is a partial discharge defect.

[0111] In this embodiment, by obtaining a plurality of ultrasonic signal groups, obtaining signal attributes from each ultrasonic signal group, and simultaneously obtaining basic attributes, the signal attributes are standardized to form a standard signal attribute point group. The standardization is to normalize the information in the signal attributes to avoid the influence of magnitude on clustering, and obtain core variables based on the obtained basic attributes and test requirements; based on the core variables, different clusters are obtained through BDSCAN algorithm operation. The different clusters constitute the clustering result corresponding to this partial discharge probe, and the adjacent distribution intensity of the clusters is obtained. The adjacent distribution intensity is the degree of aggregation of attribute points in space. If the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect. The partial discharge defect is accurately judged through the adjacent distribution intensity of the standard signal attribute points, improving the test accuracy of the partial discharge fault, thereby preventing the GIS insulation breakdown and delaying the performance decay.

[0112] In this embodiment, different clusters are obtained through the BDSCAN algorithm. One cluster is selected, and a starting standard signal attribute point 1 is selected from this cluster. Taking the starting standard signal attribute point 1 as the center and e as the radius to draw a circle, an adjacent circle is formed. If there are m or more standard signal attribute points in the adjacent circle, all the standard signal attribute points in this adjacent circle are called adjacent standard signal attribute points to the starting standard signal attribute point 1. If there are less than m standard signal attribute points in the adjacent circle, this starting standard signal attribute point 1 is called a noise point. Here, the starting standard signal attribute point 1 is any standard signal attribute point in the standard signal attribute point group. Calculate the adjacent distribution intensity value of the starting standard signal attribute point 1. Here, the number of adjacent standard signal attribute points to the starting standard signal attribute point 1 is called the adjacent distribution intensity value of the starting standard signal attribute point 1, and the adjacent distribution intensity value of the noise point is 0. When calculating the adjacent distribution intensity value of the starting standard signal attribute point 1, the starting standard signal attribute point 1 can be selected first. Taking the starting standard signal attribute point 1 as the center and e as the radius to draw a circle, an adjacent circle is formed. If there are m or more standard signals in the adjacent circle, all the standard signal attributes in this adjacent circle are called adjacent standard signal attribute points to the starting standard signal attribute point 1. Otherwise, this starting standard signal attribute point 1 is called a noise point.

[0113] In this embodiment, a starting adjacent standard signal attribute point 1 is selected. Taking the starting adjacent standard signal attribute point 1 as the center and e as the radius to draw a circle, an adjacent circle is formed, and all the adjacent standard signal attribute points to the starting standard signal attribute point 1 are traversed. If there are m or more standard signal attributes in the adjacent circle, all the standard signal attributes in this adjacent circle are called adjacent standard signal attribute points 2. If there are less than m standard signal attributes in the adjacent circle, this starting adjacent standard signal attribute point 1 is called a noise point. Calculate the adjacent distribution intensity values of each starting adjacent standard signal attribute point 1. The number of adjacent standard signal attribute points 2 is called the adjacent distribution intensity value of the starting standard signal attribute point 1. If the starting standard signal attribute point 1 is a noise point, its adjacent distribution intensity value is 0. Repeat the above steps until all the adjacent standard signal attribute points have no subordinate adjacent standard signal attribute points, and calculate the adjacent distribution intensity values of all the standard signal attribute points in this cluster. Select a new cluster, and then select a starting standard signal attribute point 1 in this cluster, and repeat the above steps until all the clusters are traversed, and calculate the adjacent distribution intensity values of all the standard signal attribute points in each cluster.

[0114] In this embodiment, core variables are obtained based on basic attributes and test requirements, including: parameter one obtained according to the density of standard signal attributes, and parameter one is positively correlated with the density of standard signal attributes; parameter two obtained according to the quantity order of standard signal attributes, and parameter two is negatively correlated with the quantity order of standard signal attributes; parameter three obtained according to test accuracy, and parameter three is positively correlated with test accuracy; parameter four obtained according to test efficiency, and parameter four is negatively correlated with test efficiency; basic core variables are obtained; the basic core variables are optimized by at least one of parameter one, parameter two, parameter three, and parameter four to obtain core variables.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the specific implementation manners of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement without departing from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. Distributed remote probe ultrasonic partial discharge test method, comprising: An amplifier, a partial discharge probe, a partial discharge detection device, a terminal, and an Ethernet. Each partial discharge probe is configured with an amplifier, and the amplifier is connected to the partial discharge detection device. The partial discharge probe, the partial discharge detection device, and the terminal can be connected via the Ethernet. It is characterized in that: The method includes the following steps: Obtain a number of ultrasonic signal groups, obtain signal attributes from each ultrasonic signal group, and at the same time obtain basic attributes. Standardize the signal attributes to form a group of standard signal attribute points. The basic attributes include: standard signal attribute density and standard signal attribute quantity order of magnitude; Derive core variables based on the basic attributes and test requirements, where the test requirements include at least one of test accuracy and test efficiency; According to the core variables, different clusters are obtained through BDSCAN algorithm operation. Different clusters constitute the clustering result corresponding to the partial discharge probe, and the adjacent distribution intensity of the clusters is obtained; If it is confirmed that the adjacent distribution intensity of a certain cluster in a certain clustering result changes compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect.

2. The distributed remote probe ultrasonic partial discharge test method according to claim 1, characterized in that: The core variables include the minimum number of neighborhood points m and the neighborhood radius e. m is a non-zero natural number, and e is a positive real number; Obtain different clusters through the BDSCAN algorithm; Select a certain cluster, select the starting standard signal attribute point one from this cluster, and use the starting standard signal attribute point one as the center and e as the radius to draw a circle to form a neighborhood circle; If there are greater than or equal to m standard signal attribute points in the neighborhood circle, then all the standard signal attribute points in this neighborhood circle are called adjacent standard signal attribute points one; If there are less than m standard signal attribute points in the neighborhood circle, then this starting standard signal attribute point one is called a noise point; Calculate the adjacent distribution intensity value of the starting standard signal attribute point one.

3. The distributed remote probe ultrasonic partial discharge test method according to claim 2, characterized in that: Select the starting adjacent standard signal attribute point one, use the starting adjacent standard signal attribute point one as the center and e as the radius to draw a circle to form a neighborhood circle, and traverse all the starting adjacent standard signal attribute points one; If there are greater than or equal to m standard signals in the neighborhood circle, then all the standard signals in this neighborhood circle are called adjacent standard signal attribute points two; If there are less than m standard signals in the neighborhood circle, then this starting adjacent standard signal attribute point one is called a noise point; Calculate the adjacent distribution intensity values of each starting adjacent standard signal attribute point one; Repeat the above steps until there are no subordinate adjacent standard signal attribute points for all the adjacent standard signal attribute points, and calculate the adjacent distribution intensity values of all the standard signal attribute points in this cluster.

4. The distributed remote probe ultrasonic partial discharge test method according to claim 3, characterized in that: Re-select a cluster, and then select the starting standard signal attribute point one in this cluster, and repeat the above steps until all the clusters are traversed, and calculate the adjacent distribution intensity values of all the standard signal attribute points in each cluster.

5. The distributed remote probe ultrasonic partial discharge test method according to claim 4, characterized in that: The value of m is: 2D - 1, where D is the dimension of the standard signal attribute points.

6. The distributed remote probe ultrasonic partial discharge test method according to claim 5, characterized in that: e can be calculated by the following formula: In the formula, M is the area surrounded by all standard signal attribute points.

7. The distributed remote probe ultrasonic partial discharge test method according to claim 4, characterized in that: Calculate the adjacent distribution intensity of each cluster through the following formula: In the formula, Apdiv represents the adjacent distribution intensity value of the standard signal attribute points within the cluster, and n represents the number of standard signal attribute points within the cluster.

8. The distributed remote probe ultrasonic partial discharge test method according to claim 1, characterized in that: Obtain Parameter 1 based on the density of standard signal attributes, and Parameter 1 is positively correlated with the density of standard signal attributes; Obtain Parameter 2 based on the quantity order of standard signal attributes, and Parameter 2 is negatively correlated with the quantity order of standard signal attributes; Obtain Parameter 3 based on the test accuracy, and Parameter 3 is positively correlated with the test accuracy; Obtain Parameter 4 based on the test efficiency, and Parameter 4 is negatively correlated with the test efficiency; Optimize the basic core variables through at least one of Parameter 1, Parameter 2, Parameter 3, and Parameter 4 to obtain the core variables.

9. The distributed remote probe ultrasonic partial discharge test method according to claim 5, characterized in that: Each partial discharge probe corresponds to a clustering result. When the partial discharge detection device compares the clustering results corresponding to each partial discharge probe and finds that the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with other clustering results, it indicates that there is a partial discharge in the GIS device, alarms the staff, and at the same time transmits the clustering result to the terminal in the form of a picture.

10. Distributed remote probe ultrasonic partial discharge test system, characterized in that: Including: A processing unit for processing the obtained ultrasonic signal group, obtaining signal attributes from the ultrasonic signal group, and at the same time obtaining basic attributes, where the signal attributes include at least one of the density of standard signal attributes and the quantity order of standard signal attributes; A selection unit for selecting core variables according to the basic attributes and test requirements, where the test requirements include at least one of test accuracy and test efficiency; An operation unit for obtaining different clusters through BDSCAN algorithm operation based on the core variables and obtaining the adjacent distribution intensity of the clusters; A test unit for confirming that if the adjacent distribution intensity of a certain cluster in a certain clustering result changes suddenly compared with the clustering results obtained by other partial discharge probes, it indicates that there is a partial discharge defect.

Images