Optical push-scanning positioning device for reactor insulation fault
An optical push-broom positioning device combining a SiC ultraviolet solar-blind sensor array and the KNN algorithm solves the electromagnetic interference problem in reactor insulation fault monitoring, achieving high-precision online monitoring and fault early warning, and improving power grid stability and equipment utilization.
Patent Information
- Application Number
- CN202210062879.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-01-19
AI Technical Summary
Existing technologies are insufficient for effectively monitoring insulation faults in reactors, especially in environments with electromagnetic interference, leading to decreased grid stability and equipment utilization.
By combining a SiC ultraviolet solar-blind sensor array with the KNN algorithm, the ultraviolet signal of the reactor is monitored in real time through an optical push-broom positioning device to locate key discharge points, avoid electromagnetic interference, and improve monitoring accuracy.
This enables online monitoring of reactors, avoids electromagnetic interference, improves monitoring accuracy, detects potential faults early, reduces unnecessary power outages for maintenance, and improves grid stability and equipment utilization.
Smart Images

Figure CN114578191B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of online monitoring of reactors, in particular to an optical push-scanning positioning device for insulation faults of reactors. Background Art
[0002] Reactors are important equipment in power systems, which play the roles of compensating stray capacitive current, limiting closing inrush current, limiting short-circuit current, filtering, and blocking waves. In current power systems, the utilization rate of reactors is very high, but due to their structural characteristics, they are prone to partial discharge, insulation damage, turn-to-turn short circuit and other faults, as well as threats caused by loose installation of fasteners, which affect the stable operation of the power grid.
[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the invention and therefore it may contain information that does not form the prior art that is already known in this country to a person of ordinary skill in the art. Summary of the invention
[0004] In view of the shortcomings or defects of the above-mentioned prior art, an optical push-scanning positioning device for reactor insulation fault is provided, which can conveniently detect potential fault threats of dry-type reactors during inspection, while avoiding electromagnetic interference on site and meeting sensitivity requirements. It can detect whether the reactor is discharged and locate the position of key discharge points.
[0005] The purpose of the present invention is achieved through the following technical solutions.
[0006] The optical push-sweep positioning device for reactor insulation fault includes:
[0007] A base, which is arranged in the reactor and the center of the base is colinear with the central axis of the reactor;
[0008] A transmission device, which is vertically rotationally connected to the base for periodic rotation;
[0009] A support beam connected to the transmission device to perform periodic sweeping under the actuation of the transmission device, wherein the sweeping period of the periodic sweeping is an integer multiple of the power frequency period;
[0010] A plurality of SiC ultraviolet solar-blind sensor probes are arranged linearly and non-uniformly on the support beam, and the linear non-uniform arrangement collects ultraviolet signals emitted by the reactor in real time and converts them into voltage signals;
[0011] A processing unit is connected to the plurality of SiC ultraviolet solar-blind sensor probes. In response to the voltage signal, the processing unit synchronously generates the amplitude of the ultraviolet light irradiance and converts it into a voltage signal for output.
[0012] The optical push-scanning positioning device for reactor insulation fault also includes:
[0013] A data transmission module connected to the processing unit to transmit the voltage signal output in real time;
[0014] A server is communicatively connected to the data transmission module, and the server generates reactor insulation fault information based on the voltage signal output, and the server includes:
[0015] A database storing data matching fault locations with light intensity data;
[0016] A data analysis module, which is connected to the database and the data transmission module;
[0017] The data analysis module outputs the spatial clustering relationship between the voltage signal matrix and the fault information based on the voltage signal matrix output by the sensor through the KNN algorithm, and obtains the insulation fault information of the reactor through the KNN algorithm.
[0018] In the optical push-scanning positioning device for reactor insulation fault, the processing unit includes a sampling unit with a sampling rate greater than 100kHz and an amplification factor greater than 20dB, a detection unit for obtaining the envelope peak of the voltage signal, and a calculation unit for obtaining the light irradiation intensity through proportional transformation.
[0019] In the optical push-scanning positioning device for reactor insulation fault, the base is arranged at the bottom of the reactor or the lower side of the rainproof umbrella.
[0020] In the optical push-scanning positioning device for reactor insulation fault, the multiple SiC ultraviolet solar-blind sensor probes are ultraviolet solar-blind SiC avalanche diode APD sensor arrays.
[0021] In the optical push-scanning positioning device for reactor insulation fault, the waveband of the SiC avalanche diode APD sensor array includes 200nm-400nm, the lower limit of its response energy is 1nW / cm2, and the gain is 105-106.
[0022] In the optical push-scanning positioning device for reactor insulation fault, the single pulse response time of the SiC avalanche diode APD sensor array to MHz-level corona discharge is 40ns.
[0023] In the optical push-scanning positioning device for reactor insulation fault, the SiC ultraviolet solar-blind sensor probe includes an ultraviolet signal preprocessing unit for preprocessing and collecting, a high-speed analog-to-digital sampling unit, and an ultraviolet sensor for converting ultraviolet light intensity signals into voltage signals, and the signal preprocessing unit includes an amplifier.
[0024] In the optical push-scanning positioning device for reactor insulation fault, the normal line of the detection plane of the SiC ultraviolet solar-blind sensor probe is parallel to the center axis of the reactor.
[0025] Beneficial Effects
[0026] The present invention realizes online monitoring of dry-type air-core reactors. Compared with the traditional electrical monitoring method, it can avoid electromagnetic interference to improve monitoring accuracy, track and observe the equipment, discover potential failures of the equipment early, improve the judgment of poor working conditions of the equipment, avoid unnecessary power outages for maintenance and testing, improve the stability of the power grid and the utilization rate of the equipment, and has the characteristics of strong real-time performance and high sensitivity. Using solar-blind ultraviolet SiC arrays to realize discharge positioning greatly reduces the threshold for the application of ultraviolet sensing technology in terms of cost.
[0027] The above description is only an overview of the technical solution of the present invention. In order to make the technical means of the present invention clearer and to enable those skilled in the art to implement it according to the contents of the specification, and to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described below by way of example. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] By reading the detailed description of the preferred specific embodiments below, various other advantages and benefits of the present invention will become clear to those of ordinary skill in the art. The drawings in the specification are only for the purpose of illustrating the preferred embodiments and are not considered to be limitations of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work. Moreover, the same reference numerals are used to represent the same components throughout the drawings.
[0029] In the attached picture:
[0030] Figure 1 is a schematic diagram of optical push-scanning positioning of reactor insulation fault according to an embodiment of the present invention;
[0031] Figure 2 A schematic structural diagram of optical push-scanning positioning of reactor insulation fault according to an embodiment of the present invention;
[0032] Figure 3 A simulation model diagram of a reactor for optical push-scanning location of reactor insulation fault according to an embodiment of the present invention;
[0033] Figure 4 A cross-sectional view of the internal structure of a reactor model according to an embodiment of the present invention;
[0034] Figure 5 Light intensity distribution diagram of the bottom plane of the reactor according to one embodiment of the present invention;
[0035] Figure 6A schematic diagram of a circular data result converted into a rectangular data result through data transformation according to an embodiment of the present invention;
[0036] Figure 7 A schematic diagram of a data conversion process according to an embodiment of the present invention;
[0037] Figure 8 A schematic diagram of a labeled matrix after conversion according to an embodiment of the present invention;
[0038] Fig. 9 A schematic diagram of dividing the five parts by principal component analysis according to an embodiment of the present invention;
[0039] Fig.10 A schematic diagram of data eigenvalues of a first principal component of a first part divided by principal component analysis according to an embodiment of the present invention;
[0040] Fig.11 A schematic diagram of data eigenvalues of the first principal component of the second part divided by principal component analysis according to an embodiment of the present invention;
[0041] Fig.12 A schematic diagram of data eigenvalues of the first principal component of the third part divided by principal component analysis according to an embodiment of the present invention;
[0042] Fig.13 A schematic diagram of data eigenvalues of the first principal component of the fourth part divided by principal component analysis according to an embodiment of the present invention;
[0043] Fig.14 A schematic diagram of data eigenvalues of the first principal component of the fifth part divided by principal component analysis according to an embodiment of the present invention.
[0044] The present invention is further explained below in conjunction with the accompanying drawings and embodiments. DETAILED DESCRIPTION
[0045] The following will refer to the accompanying drawings Figures 1 to 14 Specific embodiments of the present invention are described in more detail. Although specific embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.
[0046] It should be noted that certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that technicians may use different nouns to refer to the same component. This specification and claims do not use the difference in nouns as a way to distinguish components, but use the difference in the functions of the components as the criterion for distinction. As mentioned throughout the specification and claims, "including" or "comprising" is an open term, so it should be interpreted as "including but not limited to". The subsequent description of the specification is a preferred embodiment of the present invention, but the description is based on the general principles of the specification and is not intended to limit the scope of the present invention. The scope of protection of the present invention shall be determined by the attached claims.
[0047] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each of the accompanying drawings does not constitute a limitation on the embodiments of the present invention.
[0048] like Figure 1 to Figure 2 As shown, the optical push-scanning positioning device for reactor insulation fault includes:
[0049] A base 2 is provided in the reactor 1 and the center of the base 2 is colinear with the central axis of the reactor 1.
[0050] A transmission device 3, which is vertically connected to the base 2 for periodic rotation,
[0051] The supporting beam 4 is connected to the transmission device 3 so as to perform periodic sweeping under the actuation of the transmission device 3, wherein the sweeping period of the periodic sweeping is an integer multiple of the power frequency period.
[0052] A plurality of SiC ultraviolet solar-blind sensor probes 5 are arranged linearly and non-uniformly on the support beam 4. The linear non-uniform arrangement collects the ultraviolet signal emitted by the reactor 1 in real time and converts it into a voltage signal.
[0053] A processing unit is connected to the plurality of SiC ultraviolet solar-blind sensor probes 5. In response to the voltage signal, the processing unit synchronously generates the amplitude of the ultraviolet light irradiance and converts it into a voltage signal output.
[0054] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, a data transmission module is connected to the processing unit to transmit the voltage signal output in real time, and the data transmission module is communicatively connected to a server, which generates reactor insulation fault information based on the voltage signal output, and the server includes:
[0055] Database, which stores data matching fault location and light intensity data,
[0056] A data analysis module, which is connected to the database and the data transmission module;
[0057] The data analysis module outputs the spatial clustering relationship between the voltage signal matrix and the fault information based on the voltage signal matrix output by the sensor through the KNN algorithm, and obtains the insulation fault information of the reactor through the KNN algorithm.
[0058] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the processing unit includes a sampling unit with a sampling rate greater than 100kHz and an amplification factor greater than 20dB, a detection unit for obtaining the envelope peak of the voltage signal, and a calculation unit for obtaining the light irradiation intensity through proportional transformation.
[0059] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the base 2 is placed at the bottom of the reactor 1 or at the lower side of a rainproof umbrella.
[0060] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the plurality of SiC ultraviolet solar-blind sensor probes 5 are ultraviolet solar-blind SiC avalanche diode APD sensor arrays.
[0061] In the preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the band of the SiC avalanche diode APD sensor array includes 200nm-400nm, the lower limit of its response energy is lnW / cm2, and the gain is 10 5 -10 6 .
[0062] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the single pulse response time of the SiC avalanche diode APD sensor array to MHz-level corona discharge is 40ns.
[0063] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the front surface of the support beam 4 has holes arranged side by side so that the ultraviolet sensor probe array faces the reactor 1; the ultraviolet sensor probe is welded to the processing unit, and the signal processing unit is connected to the data transmission module via a cable.
[0064] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the SiC ultraviolet solar-blind sensor probe 5 includes an ultraviolet signal preprocessing unit for preprocessing and collecting, a high-speed analog-to-digital sampling unit, and an ultraviolet sensor for converting ultraviolet light intensity signals into voltage signals, and the signal preprocessing unit includes an amplifier.
[0065] In a preferred embodiment of the optical push-scanning positioning device for reactor insulation fault, the normal line of the detection plane of the SiC ultraviolet solar-blind sensor probe 5 is parallel to the central axis of the reactor 1 .
[0066] In one embodiment, the plurality of SiC ultraviolet solar-blind sensor probes 5 are gradually distributed and denser in a direction away from the center of the periodic push-scan.
[0067] In one embodiment, the reactor 1 is a dry-type air-core reactor 1 .
[0068] In one embodiment, the transmission device 3 supports the array to perform periodic push-sweeping along the radial direction with the center of the circle as the fulcrum, and the push-sweeping period is an integer multiple of the power frequency period.
[0069] In one embodiment, the ultraviolet sensor probe collects in real time the optical signal of the ultraviolet solar-blind frequency band emitted by the internal discharge of the reactor 1 .
[0070] In one embodiment, the optical push-scanning positioning device for reactor insulation fault comprises:
[0071] The mechanical transmission device 3 supports the array to perform periodic sweeping along the radial direction with the center of the circle as the fulcrum.
[0072] SiC UV solar-blind sensor probe 5, a sensor probe for real-time collection of UV signals emitted by reactor 1, is placed on the upper side of the support rod in a linear non-uniform arrangement, facing reactor 1. The solar-blind UV sensor probe converts the collected UV light intensity signal into a voltage signal.
[0073] A processing unit is connected to the solar-blind ultraviolet sensor probe, and in response to the voltage signal, the processing unit synchronously generates ultraviolet light intensity data.
[0074] The data transmission module realizes real-time data transmission through the WiFi module, transmits the signal to the data center, and performs subsequent data analysis and diagnosis.
[0075] Among them, the ultraviolet sensor probe includes a signal preprocessing unit for preprocessing the collected ultraviolet signal, an analog-to-digital sampling unit and an ultraviolet sensor for converting the ultraviolet signal into a voltage signal, the signal preprocessing unit includes an amplifier, the processing unit includes a sampling unit with a sampling rate greater than 100kHz and an amplification factor greater than 100 times, a detection unit for the envelope peak of the voltage signal, and a calculation unit for obtaining the sound intensity of the optical signal through proportional transformation.
[0076] The processing unit includes a digital signal processor, an application specific integrated circuit ASIC or a field programmable gate array FPGA, and the processing unit includes one or more read-only memories ROM, random access memories RAM, flash memories or electronically erasable programmable read-only memories EEPROM.
[0077] To facilitate understanding of the present invention, the positioning device and the positioning method thereof will be further explained below by taking several specific embodiments as examples in combination with the accompanying drawings, and each of the accompanying drawings does not constitute a limitation on the embodiments of the present invention.
[0078] like Figure 1 to Figure 2 As shown, the present invention discloses an optical push-scanning positioning method for reactor insulation fault, comprising the following steps:
[0079] The base 2 is arranged in the reactor 1 and the center of the base 2 is colinear with the central axis of the reactor 1.
[0080] The transmission device 3 is vertically rotated and connected to the base 2 to rotate periodically. The support beam 4 is connected to the transmission device 3 to perform periodic sweeping under the actuation of the transmission device 3. The sweeping period of the periodic sweeping is an integer multiple of the power frequency period.
[0081] The light intensity distribution of the circular area on the bottom plane of the reactor under different light sources is collected, the grayscale image of the circular area is extracted, the blank data outside the circular area is eliminated, and the data matrix of the circular area is obtained.
[0082] Based on the image transformation, the data matrix of the circular area is expanded along the radial direction, and converted into a rectangular data matrix through data filling. After the expansion, the circumference of the circle becomes the length of the rectangle, the radius of the circle becomes the width of the rectangle, and the data of the center of the circle becomes the length of the other side of the rectangle through data replication. The data matrix of the circular area is converted into a rectangular data matrix, and each column of the rectangular data matrix is the data of the data matrix of the circular area along the radius;
[0083] Based on the principal component analysis method, a data dimension reduction analysis is performed on the rectangular data moment to extract feature points. A plurality of SiC ultraviolet solar-blind sensor probes 5 are arranged linearly and non-uniformly on the support beam 4 according to the positions of the feature points. The linear non-uniform arrangement collects the ultraviolet signal emitted by the reactor in real time and converts it into a voltage signal.
[0084] In a preferred embodiment of the method, the principal component analysis method uses the length extension direction of the supporting beam as the radial direction and divides the beam into multiple parts, and principal component analysis is performed on each part to obtain characteristic points.
[0085] In a preferred embodiment of the method, according to the physical structure of the reactor, the length extension direction of the support beam is used as the radial direction and the coordinates are divided into five parts, and each part is subjected to principal component analysis to obtain the data characteristics of the corresponding first principal component.
[0086] In a preferred embodiment of the method, a pseudo-color image is extracted based on the light intensity distribution, and the pseudo-color image is converted into a grayscale image.
[0087] In a preferred embodiment of the method, the data matrix of the circular area is converted into a rectangular data matrix and then normalized.
[0088] In a preferred implementation of the method, data filling is performed when the radius image of the data matrix of the circular area is transformed, and the filling value is the average of the data on both sides.
[0089] In a preferred embodiment of the method, the plurality of SiC ultraviolet solar-blind sensor probes 5 are ultraviolet solar-blind SiC avalanche diode APD sensor arrays.
[0090] In a preferred embodiment of the method, the waveband of the SiC avalanche diode APD sensor array includes 200nm-400nm, and its lower limit of response energy is lnW / cm 2 , the gain is 10 5 -10 6 .
[0091] In a preferred embodiment of the method, the normal line of the detection plane of the SiC ultraviolet solar-blind sensor probe 5 is parallel to the center axis of the reactor.
[0092] In one embodiment, the plurality of SiC ultraviolet solar-blind sensor probes 5 are gradually distributed and denser in a direction away from the center of the periodic push-scan.
[0093] In one embodiment, the reactor 1 is a dry-type air-core reactor.
[0094] In one embodiment, the ultraviolet sensor probe collects in real time the optical signal of the ultraviolet solar-blind frequency band emitted by the internal discharge of the reactor 1 .
[0095] In one embodiment, the processing unit is connected to the solar-blind ultraviolet sensor probe, and in response to the voltage signal, the processing unit synchronously generates ultraviolet light intensity data. The data transmission module realizes real-time data transmission through the wifi module, transmits the signal to the data center, and performs subsequent data analysis and diagnosis.
[0096] Among them, the ultraviolet sensor probe includes a signal preprocessing unit for preprocessing the collected ultraviolet signal, an analog-to-digital sampling unit and an ultraviolet sensor for converting the ultraviolet signal into a voltage signal, the signal preprocessing unit includes an amplifier, the processing unit includes a sampling unit with a sampling rate greater than 100kHz and an amplification factor greater than 100 times, a detection unit for the envelope peak of the voltage signal, and a calculation unit for obtaining the sound intensity of the optical signal through proportional transformation.
[0097] In one embodiment, the light intensity distribution characteristics of the linear array push-scan at different discharge positions of the reactor 1 are obtained through Tracepro simulation; the SiC ultraviolet solar-blind sensor probe 5 collects the light intensity data of the reactor 1, and a database matching the fault position and the light intensity data is established. The correspondence between the position of the discharge fault and the sensor data is determined through deep learning and other methods, and the ultraviolet signal is detected in the running reactor 1 and converted into a voltage signal output. Based on the voltage signal output, deep learning is used to determine whether the reactor 1 has a fault and give a diagnosed fault position.
[0098] In one embodiment, the method comprises,
[0099] For reactor 1, the light intensity distribution characteristics of the linear array push-scan interface at different discharge positions are obtained through Tracepro simulation;
[0100] The light intensity distribution diagram is normalized and image transformed. The data distribution characteristics along the radial direction are analyzed, and the data dimension is reduced by principal component analysis to determine the characteristic points in the radial direction for placing the ultraviolet sensor probe, thus completing the design of the discharge positioning push-scan array device of the reactor 1.
[0101] Light intensity data is collected through ultraviolet sensor probes, and a database matching the fault location and light intensity data is established. Through deep learning and other methods, the corresponding relationship between the location of the discharge fault and the sensor data is determined.
[0102] For the operating dry-type air-core reactor 1, the push-sweep device is used to monitor the ultraviolet solar-blind intensity signal of the reactor 1, and through the deep learning model, it is determined whether the dry-type reactor 1 has a fault, and the fault location of the diagnosis is given.
[0103] Abnormal discharge of reactor 1 can indicate that the equipment has electrical problems. Through partial discharge detection, we can find defects in the equipment early and deal with them in time to avoid accidents. Ultraviolet light is emitted during the discharge process. Based on the method of locating the partial discharge position using the ultraviolet solar-blind sensor push-scan array, we can predictably find signs of reactor 1 failure and deal with it in advance to ensure safe and reliable operation of the equipment.
[0104] The key is to determine the position of the sensor on this long pole, because the placement of the sensor directly determines the final detection effect. In order to achieve this goal, Figure 3 is the designed reactor simulation model, Figure 4 This is a cross-sectional view of the internal structure of the reactor model.
[0105] Light sources at different positions are placed in the designed model. The coordinates of the light source positions are shown in Table 1:
[0106] Table 1 Positions and coordinates of different partial discharge light sources in reactors
[0107]
[0108] Through optical path simulation, the light intensity distribution on the bottom plane of the reactor under different light source distributions can be obtained, such as Figure 5 As shown in the figure, when the light source position is different, the light intensity distribution on the bottom plane of the reactor is different. Therefore, the fault location of the reactor can be determined by different distributions. However, if the entire bottom plane is monitored, too many sensors are required. Therefore, the linear array push-scan mode is used as an equivalent replacement. That is, Figure 2 The device described.
[0109] Determine where to place the sensor on the bar. Need to analyze Figure 5 The light intensity distribution diagram. However, the light intensity distribution is a circular data. We want to know which positions along the radius of the circle have more important data, so that we can place sensors at more important positions. In other words, we need to reduce the dimension of the data along the radius. If the data is processed directly, it will be difficult to operate. Therefore, the circular data results are transformed into rectangles through data transformation, so that the principal component correlation analysis method can be used for data dimension reduction. The specific conversion method is as follows: Figure 6 and Figure 7 As shown. Finally, the light intensity data is converted into Figure 8 The matrix shown in Figure 1 is labeled, and the data in each column are collected along the radial direction. By reducing the dimension of the first 150 rows of the matrix, the optimal sensor coordinates can be obtained from the perspective of global optimization.
[0110] In one embodiment, Figure 7 As shown in , based on the light intensity map, the grayscale image of the circular area is extracted, and the blank data outside the circular area is eliminated to obtain a set of circular data. Figure 6 As shown, using the image transformation algorithm, the circle is expanded along the radius direction and converted into a rectangle by filling in data. After expansion, the circumference of the circle becomes the length of the rectangle, and the radius of the circle becomes the width of the rectangle. The data at the center of the circle is copied and becomes the length of the other side of the rectangle. The radius data also needs to be filled in during image transformation, and the filling value is the average of the original data on both sides. After filling, the data matrix of the circular area is converted into a rectangle, and each column of the rectangle is the data of the original circle along the radius square. Figure 2 Taking the reactor as an example, for a grayscale image of a certain fault, after data transformation and image processing, the result is as follows Figure 8As shown. After the above image transformation, a new data matrix is obtained, the width of the matrix is 150 and the length is 942. For each test, 18 sets of test data can be obtained. After transposing and splicing the 18 matrices, a new data matrix can be obtained, which has 150 rows and 18*942 columns. Finally, we add a row to the matrix and add labels by marking numbers. The labeling method is shown in the figure below. In the end, we get a set of 151*16956 matrices. The principal component analysis method can be used to perform data dimensionality reduction analysis on the 150 rows of the matrix and extract the key number of rows as data feature points.
[0111] For optical fault detection of reactors, if the physical structure of the reactor is not considered and the mathematical dimension reduction algorithm is directly used to optimize the sensor position coordinates, it is easy to miss or misselect feature points. Therefore, considering the physical structure of the reactor, the matrix is spatially reduced by the segmented principal component analysis method to determine the specific placement coordinates of the sensor. In the principal component analysis, the data corresponding to the local peak of the first principal component is selected as the feature data. We first divide the coordinates in the radial direction into five parts according to the physical structure of the reactor, such as Fig. 9 In each part, principal component analysis is performed to obtain the data eigenvalue of the first principal component as shown in Figures 10 to 14 Eight local peaks are selected as feature point coordinates, and the obtained coordinates are shown in Table 2.
[0112] Table 2 Feature point coordinates
[0113] serial number <![CDATA[X1]]> <![CDATA[X2]]> <![CDATA[X3]]> <![CDATA[X4]]> <![CDATA[X5]]> <![CDATA[X6]]> <![CDATA[X7]]> <![CDATA[X8]]> Serial number 6 30 48 103 119 128 139 149 Coordinates(cm) 4 20 32 68.7 79.3 85.3 92.7 99.3
[0114] The KNN algorithm and convolutional neural network algorithm are used to verify the validity of the feature point coordinates. After selecting the sensor placement location, the model is verified and the accuracy of partial discharge fault location is diagnosed using the fifth-order KNN, seventh-order KNN and convolutional neural network models.
[0115] The steps of the KNN model are as follows:
[0116] 1. Calculate the distance between the test data and each training data.
[0117] Existing training set sample X t and the corresponding label Y t , where X t Contains n data points X t -{x t1 , x t2 , x t3 , x t4 , ..., t n}, each data point is an 8-dimensional vector, representing the light intensity information collected by 8 sensors, where the light intensity value uses the mean of 942 data corresponding to the sensor. t Represents the light source position, which is divided into 18 categories as mentioned above. This paper uses Euclidean distance as the basis for judging the similarity, and the specific formula is as follows:
[0118]
[0119] Among them, dis(x pi , x tj ) is the test data x tj With training data x pi The Euclidean distance between pil and x tll are the i-th dimension data of the data point respectively.
[0120] 2. Sort in ascending order of distance.
[0121] 3. Select K points with the smallest distance.
[0122] 4. Determine the frequency of occurrence of the categories of the first K points.
[0123] 5. Return the category with the highest frequency among the first K points as the predicted category of the test data.
[0124] The input data is the light intensity distribution information collected by 8 sensors, and the output data is the classification probability. The category with the largest probability is selected as the predicted classification of the input data.
[0125] The accuracy of the local discharge coordinate positioning was diagnosed using a data set with known discharge source coordinates and sensor data. 20 repeated test simulations were performed, and the spatial position of the light source was randomly adjusted within the range of ±10cm each time, and a total of 360 sets of light intensity distribution data were obtained. The light intensity distribution curves under several characteristic coordinates were used as the input feature vectors, and the light source position was used as the label and output. The K nearest neighbor classifier and convolutional neural network were used for training, and the cross-validation algorithm was used to calculate the fault location accuracy under different classification models. The recognition accuracy of the segmented PCA dimensionality reduction was obtained, as shown in Table 3. The results show that the fault location based on segmented PCA dimensionality reduction has a high positioning accuracy when verified by the fifth-order, seventh-order KNN and convolutional neural network algorithms, indicating that the sensor position optimization is effective.
[0126] Table 3
[0127]
[0128] In one embodiment, a data set with known discharge source coordinates and sensor data is used to diagnose the accuracy of partial discharge coordinate positioning. 20 repeated test simulations are performed, and the spatial position of the light source is randomly adjusted within the range of ±10cm each time, and a total of 360 sets of light intensity distribution data are obtained. The light intensity distribution curves under several characteristic coordinates are used as input feature vectors, and the light source position is used as the label and output. The K nearest neighbor classifier is used for training and classification, and the cross-validation algorithm is used to calculate the fault location accuracy under different classification models.
[0129] by Figure 1 Taking the reactor model of size as an example, 20 sets of simulation test data were obtained by simulation. Each set of data had 18 test results according to the fault location, with a total of 360 data. Among them, 270 sets were used for model training and spatial clustering, and 90 sets were used for model verification. The 270 sets of training results obtained the average response strength of 8 sensors under various fault conditions, and formed a set of strength coordinates or a set of 8*842 dimensional matrices. The specific process has been described in detail above. There are 18 types of faults: upper fault inside envelope 1, lower fault inside envelope 1, upper fault inside envelope 2, lower fault inside envelope 2, upper fault outside envelope 2, lower fault outside envelope 2, upper fault inside envelope 3, lower fault inside envelope 3, upper fault outside envelope 3, lower fault outside envelope 3, upper fault inside envelope 4, lower fault inside envelope 4, upper fault outside envelope 4, lower fault outside envelope 4, upper fault inside envelope 5, lower fault inside envelope 5, upper fault outside envelope 5, lower fault outside envelope 5. During training, the fault type and the 8*842 dimensional response intensity matrix of the sensor are used as input. During verification, the intensity coordinate matrix of the sensor is imported, the Euclidean distance between all fault coordinates in the space and the sensor intensity coordinate is calculated, and the fault coordinate closest to the coordinate is solved, which is considered to be the actual fault coordinate.
[0130] Although the embodiments of the present invention are described above in conjunction with the accompanying drawings, the present invention is not limited to the above specific embodiments and application fields, and the above specific embodiments are only illustrative and instructive, rather than restrictive. A person of ordinary skill in the art can also make many forms under the guidance of this specification and without departing from the scope of protection of the claims of the present invention, all of which belong to the protection of the present invention.
Claims
1. An optical push-scanning positioning device for reactor insulation fault, characterized in that: These include, A base, which is arranged in the reactor and the center of the base is colinear with the central axis of the reactor; A transmission device, which is vertically rotationally connected to the base for periodic rotation; A support beam connected to the transmission device to perform periodic sweeping under the actuation of the transmission device, wherein the sweeping period of the periodic sweeping is an integer multiple of the power frequency period; A plurality of SiC ultraviolet solar-blind sensor probes are arranged linearly and non-uniformly on the support beam, and the linear non-uniform arrangement collects ultraviolet signals emitted by the reactor in real time and converts them into voltage signals; A processing unit is connected to the multiple SiC ultraviolet solar-blind sensor probes. In response to the voltage signal, the processing unit synchronously generates the amplitude of the ultraviolet light irradiance and converts it into a voltage signal output. The normal of the detection plane of the SiC ultraviolet solar-blind sensor probe is parallel to the central axis of the reactor. The SiC ultraviolet solar-blind sensor probe includes a preprocessing unit for preprocessing and collecting ultraviolet signals, a high-speed analog-to-digital sampling unit, and an ultraviolet sensor for converting ultraviolet light intensity signals into voltage signals. The signal preprocessing unit includes an amplifier. The multiple SiC ultraviolet solar-blind sensor probes are gradually distributed and denser in a direction away from the center of the periodic push-scan.
2. The optical push-scanning positioning device for reactor insulation fault according to claim 1, characterized in that: Also includes, A data transmission module connected to the processing unit to transmit the voltage signal output in real time; A server, which is communicatively connected to the data transmission module, and generates reactor insulation fault information based on the voltage signal output, the server comprising: A database storing data matching fault locations with light intensity data; A data analysis module, which is connected to the database and the data transmission module; The data analysis module outputs the spatial clustering relationship between the voltage signal matrix and the fault information based on the voltage signal matrix output by the sensor through the KNN algorithm, and obtains the insulation fault information of the reactor through the KNN algorithm.
3. The optical push-scanning positioning device for reactor insulation fault according to claim 1, characterized in that: The processing unit comprises a sampling unit with a sampling rate greater than 100 kHz and an amplification factor greater than 20 dB, a detection unit for obtaining an envelope peak value of a voltage signal, and a calculation unit for obtaining light irradiation intensity through proportional transformation.
4. The optical push-scanning positioning device for reactor insulation fault according to claim 1, characterized in that: The base is arranged at the bottom of the reactor or the lower side of the rainproof umbrella.
5. The optical push-scanning positioning device for reactor insulation fault according to claim 1, characterized in that: The multiple SiC ultraviolet solar-blind sensor probes are ultraviolet solar-blind SiC avalanche diode APD sensor arrays.
6. The optical push-scanning positioning device for reactor insulation fault according to claim 5, characterized in that: The SiC avalanche diode APD sensor array has a wavelength range of 200nm-400nm, a lower limit of response energy of 1nW / cm2, and a gain of 10 5 -10 6 .
7. The optical push-scanning positioning device for reactor insulation fault according to claim 5, characterized in that: The single pulse response time of the SiC avalanche diode APD sensor array to MHz-level corona discharge is 40ns.
Citation Information
Patent Citations
Path detection system based on solar blind ultraviolet light signal
CN105466437A
Discharge fault positioning system and method for electrical equipment
CN110673002A
Multi-spectral solar blind narrow-band ultraviolet imager and method for detecting different discharge states by using same
CN112285504A
High -pressure discharge testboard
CN206773131U
Ultraviolet pulse sensing node
CN212083613U
