A Smart Diagnostic Method for High-Voltage Cables

Abstract

This invention discloses an intelligent diagnostic method for high-voltage cables, belonging to the field of cable diagnostic technology. Specifically, it includes: dividing the high-voltage cable line into multiple monitoring sections of varying lengths; installing various data acquisition sensors in each monitoring section to collect cable data; setting up an IoT gateway for each monitoring section; adjusting the sampling frequency of the data acquisition sensors via the IoT gateway; preprocessing the data transmitted from the sensors to obtain integrated data; compressing the integrated data and sending it to a cloud monitoring platform; the cloud monitoring platform determines whether the integrated data is abnormal based on an anomaly threshold; upon detecting an anomaly, it identifies the monitoring section where the anomaly occurs and increases the sensor sampling frequency in that section; confirming the anomaly's authenticity; comprehensively analyzing various anomaly data to determine the anomaly type; determining the alarm level based on the anomaly type; and visually outputting the diagnostic results on the monitoring terminal. This invention achieves long-distance real-time monitoring of the high-voltage cable status and improves the accuracy of anomaly detection.

Description

Technical Field

[0001] This invention relates to the field of cable diagnostic technology, and more specifically to an intelligent diagnostic method for high-voltage cables. Background Technology

[0002] High-voltage cables are a type of power cable used to transmit power between 1kV and 1000kV. They are a crucial component of power transmission and distribution systems, enabling long-distance, high-capacity power transmission and playing a vital role in modern life and production. During operation, partial discharge is a significant cause of insulation breakdown in high-voltage cables and a key indicator of insulation degradation. Therefore, timely detection of partial discharge in high-voltage cables is essential.

[0003] Currently, most high-voltage cables still rely on planned maintenance for partial discharge detection, using various specialized portable partial discharge detection devices to collect and process high-frequency signals such as pulse currents or electromagnetic waves to determine the occurrence of partial discharge.

[0004] However, this planned maintenance method is inefficient and makes it difficult to detect potential faults in high-voltage cables in a timely manner. Furthermore, relying on only one type of detection data makes it difficult to guarantee the accuracy of the detection results. Therefore, this invention proposes an intelligent diagnostic method for high-voltage cables. Through sensors and the Internet of Things, it enables long-distance real-time monitoring of high-voltage cables, can intelligently determine the type of cable abnormality by integrating various detection data, and can visualize the diagnostic results. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent diagnostic method for high-voltage cables, solving the following technical problems:

[0006] Traditional methods for detecting the operational status of high-voltage cables still rely on planned maintenance. This method is inefficient and cannot detect potential faults in high-voltage cables in a timely manner. Furthermore, planned maintenance often uses a single testing device to detect partial discharge phenomena in high-voltage cables, resulting in limited data collection and making it difficult to guarantee the accuracy of the test results.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A smart diagnostic method for high-voltage cables includes the following steps:

[0009] The entire high-voltage cable transmission line is initially divided into M equal-length initial segments, denoted by L (where L is a positive number). The ambient temperature and noise levels of each initial segment are measured, and the environmental factor u is calculated, with u ranging from (0,1). Based on this environmental factor, each initial segment is further divided to obtain monitoring segments. After all initial segments have been divided, all resulting monitoring segments are labeled N1, N2, ..., N... i Where M and i are positive integers;

[0010] In each monitoring section, data acquisition sensors are installed at the cable joints of the high-voltage cable. These sensors include high-frequency pulse current sensors, ultra-high-frequency electromagnetic wave sensors, cable temperature sensors, load current sensors, and vibration sensors.

[0011] An IoT gateway is set up for each monitoring section, and all the data acquisition sensors in the monitoring section can transmit data with the corresponding IoT gateway. The sampling frequency of the data acquisition sensors can be adjusted through the IoT gateway.

[0012] The data collected by the sensor is sent to the IoT gateway. After receiving the data, the IoT gateway preprocesses the data through edge computing to obtain integrated data. The integrated data includes pulse current data, electromagnetic wave data, cable temperature data, load current data, and vibration data. The integrated data is then compressed.

[0013] All IoT gateways are connected to a single cloud monitoring platform. The IoT gateways send compressed, integrated data to the cloud monitoring platform, which decompresses the data, sets anomaly thresholds for different types of integrated data, and determines whether the integrated data is abnormal. If any type of integrated data exceeds the anomaly threshold, it is marked as abnormal. After detecting abnormal data, the cloud monitoring platform locates the monitoring segment N from which the abnormal data originates. i and improve N i The sampling frequency of the sensor in the middle is used to continuously detect N. i Whether any abnormal data reappears in the integrated data from various sources, and determine N based on the analysis results of all integrated data. i The system monitors the operating status of the internal high-voltage cable and displays the diagnostic results on the terminal equipment.

[0014] As a further aspect of the present invention, the specific steps for secondary division of the initial segment are as follows:

[0015] The lengths of the initial section and monitoring section refer to the length of the high-voltage cable body within the corresponding section. The ambient temperature value P and ambient noise value H of each initial section are measured. The ranges of ambient temperature and ambient noise values ​​that allow the high-voltage cable and data acquisition sensors to operate normally are set as (P...).min , P max ), and (H min , H max ), and normalize P and H. The calculation formula is:

[0016] P j = (P - P min + 1) / (P max - P min + 1)

[0017] H j = (H - H min + 1) / (H max - H min + 1)

[0018] Where P j and H j represent the normalized values of the environmental temperature value P and the environmental noise value H corresponding to the initial section, j = 1, 2,..., M;

[0019] Use P j and H j and the sigmoid function to calculate the environmental factor u, and constrain the value of u to the range of (0, 1). The calculation formula is:

[0020] u = sigmoid(a · P j + b · H j + ε)

[0021] sigmoid(x) = 1 / (1 + e(-x))

[0022] Where a and b are the weight coefficients of the environmental temperature value and the environmental noise value respectively, a + b < 1, ε is the error term, and its value range is [-0.1, 0.1]. sigmoid(x) represents the sigmoid function, and e is the base of the natural logarithm;

[0023] When 0 < u ≤ 0.5, divide the corresponding initial section into m monitoring sections of the same length. The calculation formula of m is:

[0024] m = [1 / u]

[0025] Where [1 / u] represents rounding down the value of 1 / u;

[0026] When 0.5 < u ≤ 0.8, divide the corresponding initial section into two monitoring sections, and the lengths are marked as L u1 and L u2 , and the calculation formula is:

[0027] L u1 = L · u

[0028] L u2 = L·(1 - u)

[0029] When 0.8 < u < 1, the value of u is uniformly taken as 1, and the corresponding initial section is directly set as the monitoring section. If the u values of two adjacent initial sections are both 1, they are merged into one monitoring section.

[0030] As a further solution of the present invention: The IoT gateway will adjust the sampling frequency of the acquisition sensors in the corresponding section according to the environmental factor u. Mark the sampling frequency of the acquisition sensors as f, and the calculation formula is:

[0031] f1 = f0·(1 + K / u)

[0032] f = min(f1, f max )

[0033] where f0 is the default initial sampling frequency of the acquisition sensor, K is a positive constant, f1 is the reference sampling frequency obtained by adjusting f0 according to the environmental factor u, and f max is the maximum sampling frequency that the sampling sensor can support. min means taking the minimum value between f1 and f max ;

[0034] The IoT gateway will detect the operating status of the acquisition sensors. When it detects that an acquisition sensor fails in the corresponding monitoring section, the IoT gateway will report the fault information to the cloud monitoring platform and increase the sampling frequency of the remaining normally operating acquisition sensors in the monitoring section.

[0035] As a further solution of the present invention: The specific process of installing the acquisition sensors is as follows:

[0036] The cable joints of the high-voltage cable include terminal joints and intermediate joints. When installing the high-frequency pulse current sensor, connect the sensor to the grounding wire of the high-voltage cable;

[0037] The ultra-high-frequency electromagnetic wave sensor adopts a non-contact working mode and does not directly contact or access the cable. The ultra-high-frequency electromagnetic wave sensor is installed at the terminal joints and intermediate joints of the high-voltage cable;

[0038] The vibration sensor selects a piezoelectric acceleration sensor. When installing, the sensing surface of the sensor is in full contact with the surface of the high-voltage cable.

[0039] As a further aspect of the present invention: the installed acquisition sensor has a built-in DSP data acquisition function and supports the IuC60870-5-104 or IuC61850 IoT communication protocol. The communication methods between the IoT gateway and the acquisition sensor include wired communication and wireless communication. At the same time, the high-frequency pulse current sensor and the ultra-high frequency electromagnetic wave sensor have built-in automatic gain controllers to perform gain processing on the acquired data before sending it to the IoT gateway.

[0040] As a further aspect of the present invention: the preprocessing includes data standardization, data alignment, and data denoising;

[0041] The data standardization method is to convert the storage format of data collected by various acquisition sensors into a unified XML or JSON format;

[0042] The data alignment method is as follows: within any monitoring segment, align the data collected by all sensors according to a common time point;

[0043] The data denoising method is as follows: the IoT gateway uses Kalman filtering to denoise the received data.

[0044] As a further aspect of the present invention: the specific process of compressing and decompressing the integrated data utilizes compressed sensing theory, and the compression and decompression steps are as follows:

[0045] S1: In IoT gateways, Fourier transform or wavelet transform is used as a sparse basis to obtain sparse signals from the collected data by solving the L1 norm minimization problem.

[0046] S2: In the IoT gateway, the Hadamard matrix or Gaussian matrix is ​​used as a random matrix to measure sparse signals, obtain the measured values, and complete the compression process.

[0047] S3: In the cloud monitoring platform, the orthogonal matching pursuit algorithm is used to reconstruct the signal based on the measured values, thus completing the decompression process.

[0048] As a further aspect of the present invention: the diagnostic results include: abnormality type and alarm level, the abnormality type includes abnormal cable temperature, abnormal load current, abnormal cable vibration, abnormal pulse current and abnormal partial discharge, and the alarm level is divided into level one alarm, level two alarm and level three alarm;

[0049] Among the abnormality types, abnormal cable temperature, abnormal load, and abnormal cable vibration are judged individually based on the corresponding integrated data, while abnormal pulse current and abnormal partial discharge are judged comprehensively based on the integrated data of all types.

[0050] The alarm level is determined as follows: when only one of the abnormal types of cable temperature, load current and cable vibration occurs, it is judged as a level one alarm; when two or more of the abnormal types occur at the same time, it is judged as a level two alarm; when only pulse current or partial discharge occurs, it is also judged as a level two alarm; when partial discharge and other abnormal types occur, it is judged as a level three alarm.

[0051] The diagnostic results output by the cloud monitoring platform to the monitoring terminal show all types of anomalies detected, provide the alarm level, and identify the monitoring section where the anomaly was found.

[0052] As a further aspect of the present invention, the specific process for determining the abnormality type of a high-voltage cable is as follows:

[0053] When the cloud monitoring platform detects abnormal data, it locates the monitoring segment N from which the abnormal data originates. i Set a continuous monitoring period T, and the IoT gateway will monitor N during the monitoring period T. i The sampling frequency of the internal acquisition sensor is simultaneously increased by a factor of G (G>1), and N is continuously detected. i If abnormal data appears continuously or periodically within a given period of time, then the abnormal data is determined to be true.

[0054] When the cloud monitoring platform determines that any abnormal data in the cable temperature data, load current data, or vibration data is true, it judges that the high-voltage cable has experienced the corresponding abnormality type.

[0055] When the cloud monitoring platform determines that the abnormal pulse current data is true and the electromagnetic wave data is normal, it will only determine that the high-voltage cable has a partial discharge abnormality if other abnormal types are also determined at the same time; otherwise, it will be determined that the pulse current is abnormal.

[0056] When the cloud monitoring platform determines that both pulse current data and electromagnetic wave data are true, or only electromagnetic wave data is true, it directly judges that the high-voltage cable has a partial discharge abnormality.

[0057] The cloud monitoring platform records historical diagnostic results, counts the frequency of abnormal types in each monitoring section, and displays the statistical results on the monitoring terminal.

[0058] The beneficial effects of this invention are:

[0059] This invention provides an intelligent diagnostic method for high-voltage cables. By combining data acquisition sensors, an IoT gateway, and a cloud monitoring platform, it achieves long-distance real-time monitoring of high-voltage cables, overcoming the shortcomings of traditional planned maintenance methods that cannot detect potential cable faults in a timely manner. This invention divides a high-voltage cable transmission line into different monitoring sections based on environmental factors, setting different sensor sampling frequencies for each section to maximize the reliability of the monitoring data. Furthermore, in this invention, the cloud monitoring platform comprehensively judges the partial discharge phenomenon of the high-voltage cable based on multiple data points, and can also diagnose other types of high-voltage cable anomalies, intelligently diagnosing the alarm level of high-voltage cable anomalies. This improves the accuracy of judging partial discharge phenomena and provides relevant personnel with visualized diagnostic results. Attached Figure Description

[0060] The invention will now be further described with reference to the accompanying drawings.

[0061] Figure 1 This is a flowchart illustrating the present invention;

[0062] Figure 2 This is a schematic diagram of the process for dividing monitoring sections according to the present invention;

[0063] Figure 3 This is a schematic diagram illustrating the process of the cloud monitoring platform of the present invention outputting diagnostic results for high-voltage cables. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Please see Figures 1-3 As shown, this invention provides an intelligent diagnostic method for high-voltage cables, comprising the following steps:

[0066] Step 1: Initially divide the entire high-voltage cable transmission line into M initial segments of equal length, denoted as L (where L is a positive number). Measure the ambient temperature and noise levels of the initial segments, and calculate the environmental factor u for each initial segment. The value of u ranges from (0,1). The specific calculation method is as follows:

[0067] The ambient temperature value P and ambient noise value H of each initial section were measured. The range of ambient temperature and ambient noise values ​​that allow the high-voltage cable and data acquisition sensor to operate normally were set as (P...). min ,P max ) and (H min,H max ), and normalize P and H. The calculation formula is:

[0068] P j =(P - P min + 1) / (P max - P min + 1)

[0069] H j =(H - H min + 1) / (H max - H min + 1)

[0070] where P j and H j represent the normalized values of the environmental temperature value P and the environmental noise value H corresponding to the initial section. j = 1, 2,..., M. The purpose of normalization is to convert data with different dimensions and dimension units into the same range, thereby eliminating the influence of dimensions between different features.

[0071] Use P j and H j and the sigmoid function to calculate the environmental factor u, and constrain the value of u to the range (0, 1). The calculation formula is:

[0072] u = sigmoid(a·P j + b·H j + ε)

[0073] sigmoid(x)= 1 / (1 + e(-x))

[0074] where a and b are the weight coefficients of the environmental temperature value and the environmental noise value respectively, a + b < 1, ε is the error term, and its value range is [-0.1, 0.1]. sigmoid(x) represents the sigmoid function, and e is the base of the natural logarithm.

[0075] Step 2: After obtaining the environmental factors e corresponding to each initial section, perform a secondary division on a single initial section according to the range of e. The specific division steps are as follows:

[0076] When 0 < u ≤ 0.5, divide the corresponding initial section into m monitoring sections of the same length. The calculation formula for m is:

[0077] m = [1 / u]

[0078] where [1 / u] represents rounding down the value of 1 / u;

[0079] When 0.5 < u ≤ 0.8, divide the corresponding initial section into two monitoring sections, and the lengths are marked as L<00000�2>and Lu2 , the calculation formula is:

[0080] L u1 = L·u

[0081] L u2 = L·(1 - u)

[0082] When 0.8 < u < 1, it indicates that the ambient temperature and ambient noise of the environment where the high - voltage cable is located are within the normal range. Therefore, the value of u is uniformly taken as 1, and the corresponding initial section is directly set as the monitoring section. If the u values of two adjacent initial sections are both 1, it means that the environments of these two sections are relatively stable, so they can be merged as one monitoring section.

[0083] After all the initial sections have completed the secondary division, all the obtained monitoring sections are respectively marked as N1, N2,..., N i , where M and i are positive integers. The lengths of the initial section and the monitoring section refer to the length of the high - voltage cable body within the corresponding section, rather than the straight - line distance between the starting points of the sections.

[0084] Step three: Install acquisition sensors for the cable joint part of the high - voltage cable in each monitoring section. The acquisition sensors include high - frequency pulse current sensors, ultra - high - frequency electromagnetic wave sensors, cable temperature sensors, load current sensors, and vibration sensors.

[0085] When installing the sensors, connect the high - frequency pulse current sensor to the grounding wire of the high - voltage cable. The ultra - high - frequency electromagnetic wave sensor adopts a non - contact working mode and does not directly contact or connect to the cable. The vibration sensor selects a piezoelectric acceleration sensor, and ensure that the sensing surface of the sensor is in complete contact with the surface of the high - voltage cable during installation. The cable temperature sensor should completely fill the gap between the sensor and the cable outer skin with a heat - conducting pad or heat - conducting paste during installation.

[0086] Step four: Set an Internet of Things gateway for each monitoring section. All the acquisition sensors within the monitoring section have built - in DSP data acquisition functions and support the IuC60870 - 5 - 104 or IuC61850 Internet of Things communication protocol. The communication methods between the Internet of Things gateway and the acquisition sensors include wired communication and wireless communication. At the same time, an automatic gain controller is built into the high - frequency pulse current sensor and the ultra - high - frequency electromagnetic wave sensor. The collected data is subjected to gain processing and then sent to the Internet of Things gateway. The sampling frequency of the acquisition sensors can be adjusted through the Internet of Things gateway. The specific adjustment process is as follows:

[0087] The Internet of Things gateway will adjust the sampling frequency of the acquisition sensors within the corresponding section according to the environmental factor u. Mark the sampling frequency of the acquisition sensors as f, and the calculation formula is:

[0088] f1 = f0 * (1 + K / u)

[0089] f = min(f1, f max )

[0090] Where f0 is the default initial sampling frequency of the sensor, K is a positive constant, and f1 is the reference sampling frequency obtained after adjusting f0 according to the environmental factor u. max It is the maximum sampling frequency that the sampling sensor can support; min represents the frequency between f1 and f2. max Take the minimum value.

[0091] Meanwhile, the IoT gateway will detect the operating status of the acquisition sensors. When a failure is detected in the corresponding monitoring section, the IoT gateway will report the failure information to the cloud monitoring platform and increase the sampling frequency of the remaining normally operating acquisition sensors in the monitoring section. This can temporarily compensate for the amount of monitoring data in that monitoring section and ensure the richness of the data.

[0092] Step 5: The data acquisition sensors send the collected data to the IoT gateway. After receiving the data, the IoT gateway performs preprocessing on the data through edge computing. The preprocessing operations include data standardization, data alignment, and data denoising. The specific preprocessing process is as follows: data standardization converts the storage format of the data collected by various acquisition sensors into a unified XML or JSON format; data alignment aligns the data collected by all acquisition sensors in any monitoring segment according to a common time point; and data denoising uses the Kalman filter method to denoise the received data.

[0093] Step Six: After preprocessing, integrated data is obtained, including pulse current data, electromagnetic wave data, cable temperature data, load current data, and vibration data. The integrated data is then compressed using compressed sensing theory. The specific compression process is as follows: First, Fourier transform or wavelet transform is used as a sparse basis to obtain sparse signals by solving the L1 norm minimization problem in the collected data. Then, Hadamard matrix or Gaussian matrix is ​​used as a random matrix to measure the sparse signals and obtain the measured values, thus completing the compression process.

[0094] Step 7: Connect all IoT gateways to the same cloud monitoring platform. The IoT gateways send the compressed and integrated data to the cloud monitoring platform. The cloud monitoring platform uses the orthogonal matching pursuit algorithm to reconstruct the signal and complete the decompression process of the integrated data.

[0095] Anomaly thresholds are set for various types of integrated data to determine whether the integrated data is abnormal. When any type of integrated data exceeds the anomaly threshold, the integrated data is marked as abnormal. When the cloud monitoring platform detects abnormal data, it locates the monitoring segment N from which the abnormal data originates.i Set a continuous monitoring period T, and the IoT gateway will monitor N during the monitoring period T. i The sampling frequency of the internal acquisition sensor is simultaneously increased by a factor of G (G>1), and N is continuously detected. i If outliers appear consistently or periodically within a given timeframe (T), they are considered true. Increasing the sampling frequency is to prevent sampling failures caused by outliers occurring periodically and potentially deviating from the existing sampling frequency.

[0096] Step 8: When the cloud monitoring platform determines that any abnormal data in the cable temperature data, load current data, or vibration data is true, it determines that the high-voltage cable has experienced the corresponding abnormality type.

[0097] When the cloud monitoring platform determines that the abnormal pulse current data is true and the electromagnetic wave data is normal, it will only judge that the high-voltage cable has a partial discharge abnormality if other abnormal types are also determined at the same time. Otherwise, it will be judged as a pulse current abnormality, because the cause of the abnormal pulse current data is not necessarily partial discharge. Abnormal current inside the cable may also cause data abnormality.

[0098] When the cloud monitoring platform determines that both pulse current data and electromagnetic wave data are true, or only electromagnetic wave data is true, it directly judges that a partial discharge abnormality has occurred in the high-voltage cable. This is because pulse current data and electromagnetic wave data have high weight in judging partial discharge phenomena, and the detection accuracy of the ultra-high frequency electromagnetic wave sensor is higher than that of the high frequency pulse current sensor. When its data is confirmed to be abnormal, it can be directly judged that a partial discharge phenomenon has occurred.

[0099] Step 9: The cloud monitoring platform displays the diagnostic results on the terminal device. The diagnostic results include: anomaly type and alarm level. The anomaly types include abnormal cable temperature, abnormal load current, abnormal cable vibration, abnormal pulse current, and abnormal partial discharge. The alarm levels are divided into Level 1 alarm, Level 2 alarm, and Level 3 alarm.

[0100] Among the aforementioned anomaly types, cable temperature anomaly, load anomaly, and cable vibration anomaly are judged individually based on their respective integrated data, while pulse current anomaly and partial discharge anomaly are judged comprehensively based on the integrated data of all types.

[0101] The alarm level is determined as follows: when only one of the following abnormal types occurs, namely abnormal cable temperature, abnormal load current, or abnormal cable vibration, it is judged as a Level 1 alarm; when two or more of these abnormal types occur simultaneously, it is judged as a Level 2 alarm; when only pulse current abnormality or partial discharge abnormality occurs, it is also judged as a Level 2 alarm; and when partial discharge abnormality and other abnormal types occur, it is judged as a Level 3 alarm.

[0102] Step 10: The cloud monitoring platform outputs diagnostic results to the monitoring terminal, displaying all detected anomalies and their corresponding alarm levels. It also identifies the monitoring sections where anomalies were detected. The cloud monitoring platform also records historical diagnostic results, calculates the frequency of anomaly types in each monitoring section, and displays the statistical results on the monitoring terminal. Relevant personnel can then use these results to strengthen cable maintenance in monitoring sections prone to anomalies.

[0103] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0104] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0105] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for intelligent diagnosis of high-voltage cables, characterized in that, It includes the following steps: The entire high-voltage cable transmission line is initially divided into M equal-length initial segments, denoted by L (where L is a positive number). The ambient temperature and noise levels of each initial segment are measured, and the environmental factor u is calculated, with u ranging from (0,1). Based on this environmental factor, each initial segment is further divided to obtain monitoring segments. After all initial segments have been divided, all resulting monitoring segments are labeled N1, N2, ..., N... i Where M and i are positive integers; Install acquisition sensors for the cable joint parts of high-voltage cables in each monitoring section. The acquisition sensors include high-frequency pulse current sensors, ultra-high-frequency electromagnetic wave sensors, cable temperature sensors, load current sensors, and vibration sensors; Set up an Internet of Things gateway for each monitoring section. All acquisition sensors within the monitoring section can transmit data to the Internet of Things gateway corresponding to the monitoring section, and the sampling frequency of the acquisition sensors can be adjusted through the Internet of Things gateway; The acquisition sensors send the acquired data to the Internet of Things gateway. After receiving the data, the Internet of Things gateway preprocesses the data through edge computing to obtain integrated data. The integrated data includes: pulse current data, electromagnetic wave data, cable temperature data, load current data, and vibration data, and compresses the integrated data; All IoT gateways are connected to a single cloud monitoring platform. The IoT gateways send compressed, integrated data to the cloud monitoring platform, which decompresses the data, sets anomaly thresholds for different types of integrated data, and determines whether the integrated data is abnormal. If any type of integrated data exceeds the anomaly threshold, it is marked as abnormal. After detecting abnormal data, the cloud monitoring platform locates the monitoring segment N from which the abnormal data originates. i and improve N i The sampling frequency of the sensor in the middle is used to continuously detect N. i Whether any abnormal data reappears in the integrated data, and determine N based on the analysis results of the integrated data. i The system monitors the operating status of the internal high-voltage cable and displays the diagnostic results on the terminal equipment. The specific steps for secondary division of the initial section are as follows: The lengths of the initial section and monitoring section refer to the length of the high-voltage cable body within the corresponding section. The ambient temperature value P and ambient noise value H of each initial section are measured. The ranges of ambient temperature and ambient noise values ​​that allow the high-voltage cable and data acquisition sensors to operate normally are set as (P...). min ,P max ) and (H min H max The formula is as follows: P and H are normalized. P j =(P-P min +1) / (P max -P min +1) H j =(H-H min +1) / (H max -H min +1) Where P j and H j Represents the normalized values ​​of the ambient temperature P and ambient noise H for the corresponding initial section, j=1,2,...,M; Using P j and H j The sigmoid function is used to calculate the environmental factor u, constraining the value of u to the range of (0,1). The calculation formula is as follows: u=sigmoid(a·P j +b·H j +ε) sigmoid(x)=1 / (1+e(-x)) where a and b are the weight coefficients of the ambient temperature value and the ambient noise value respectively, a + b < 1, ε is the error term, and its value range is [-0.1, 0.1]. sigmoid(x) represents the sigmoid function, and e is the base of the natural logarithm; When 0 < u ≤ 0.5, the corresponding initial section is divided into m monitoring sections with the same length. The calculation formula for m is: m=[1 / u] where [1 / u] represents rounding down the value of 1 / u; When \(0.5\lt u\leq0.8\), the corresponding initial section is further divided into two monitoring sections, and their lengths are marked as \(L\) u1 and \(L\) u2 , and the calculation formula is as follows: L u1 =L·u L u2 =L·(1-u) When 0.8 < u < 1, the value of u is uniformly taken as 1, and the corresponding initial section is directly set as the monitoring section. If the u values of two adjacent initial sections are both 1, they are merged into one monitoring section; The Internet of Things gateway will adjust the sampling frequency of the acquisition sensors within the corresponding section according to the environmental factor u. Mark the sampling frequency of the acquisition sensors as f, and the calculation formula is: f1=f0·(1+K / u) f=min(f1,f max ) Where f0 is the default initial sampling frequency of the sensor, K is a positive constant, and f1 is the reference sampling frequency obtained after adjusting f0 according to the environmental factor u. max It is the maximum sampling frequency that the sampling sensor can support; min represents the frequency between f1 and f2. max Take the minimum value from the middle. The Internet of Things gateway will detect the operating status of the acquisition sensors. When it detects that an acquisition sensor fails within the corresponding monitoring section, the Internet of Things gateway will report the fault information to the cloud monitoring platform and increase the sampling frequency of the remaining normally operating acquisition sensors within the monitoring section; The installed acquisition sensors have built-in DSP data acquisition functions and support IuC60870-5-104 or IuC61850 Internet of Things communication protocols. The communication methods between the Internet of Things gateway and the acquisition sensors include wired communication and wireless communication. At the same time, automatic gain controllers are built into the high-frequency pulse current sensors and ultra-high-frequency electromagnetic wave sensors, and the acquired data is subjected to gain processing before being sent to the Internet of Things gateway.

2. The intelligent diagnostic method for high-voltage cables according to claim 1, characterized in that, The specific process of installing the acquisition sensors is as follows: The cable joints of the high-voltage cable include terminal joints and intermediate joints. When installing the high-frequency pulse current sensor, connect the sensor to the grounding wire of the high-voltage cable; The ultra-high-frequency electromagnetic wave sensor adopts a non-contact working method and does not directly contact or access the cable. The ultra-high-frequency electromagnetic wave sensor is installed at the terminal joints and intermediate joints of the high-voltage cable; Select a piezoelectric acceleration sensor for the vibration sensor. When installing, the sensing surface of the sensor is in complete contact with the surface of the high-voltage cable.

3. The intelligent diagnostic method for high-voltage cables according to claim 1, characterized in that, The preprocessing includes data standardization, data alignment, and data denoising. The data standardization method is to convert the storage format of data collected by various acquisition sensors into a unified XML or JSON format; The data alignment method is as follows: within any monitoring segment, align the data collected by all sensors according to a common time point; The data denoising method is as follows: the IoT gateway uses Kalman filtering to denoise the received data.

4. The intelligent diagnostic method for high-voltage cables according to claim 1, characterized in that, The specific process of compressing and decompressing the integrated data utilizes compressed sensing theory. The compression and decompression steps are as follows: S1: In IoT gateways, Fourier transform or wavelet transform is used as a sparse basis to obtain sparse signals from the collected data by solving the L1 norm minimization problem. S2: In the IoT gateway, the Hadamard matrix or Gaussian matrix is ​​used as a random matrix to measure sparse signals, obtain the measured values, and complete the compression process. S3: In the cloud monitoring platform, the orthogonal matching pursuit algorithm is used to reconstruct the signal based on the measured values, thus completing the decompression process.

5. The intelligent diagnostic method for high-voltage cables according to claim 1, characterized in that, The diagnostic results include: abnormality type and alarm level. The abnormality type includes abnormal cable temperature, abnormal load current, abnormal cable vibration, abnormal pulse current, and abnormal partial discharge. The alarm level is divided into level one alarm, level two alarm, and level three alarm. Among the abnormality types, abnormal cable temperature, abnormal load, and abnormal cable vibration are judged individually based on the corresponding integrated data, while abnormal pulse current and abnormal partial discharge are judged comprehensively based on the integrated data of all types. The alarm level is determined as follows: when only one of the abnormal types of cable temperature, load current and cable vibration occurs, it is judged as a level one alarm; when two or more of the abnormal types occur at the same time, it is judged as a level two alarm; when only pulse current or partial discharge occurs, it is also judged as a level two alarm; when partial discharge and other abnormal types occur, it is judged as a level three alarm. The diagnostic results output by the cloud monitoring platform to the monitoring terminal show all types of anomalies detected, provide the alarm level, and identify the monitoring section where the anomaly was found.

6. The intelligent diagnostic method for high-voltage cables according to claim 5, characterized in that, The specific process for determining the type of abnormality in a high-voltage cable is as follows: When the cloud monitoring platform detects abnormal data, it locates the monitoring segment N from which the abnormal data originates. i Set a continuous monitoring period T, and the IoT gateway will monitor N during the monitoring period T. i The sampling frequency of the internal acquisition sensor is simultaneously increased by a factor of G (G>1), and N is continuously monitored. i If abnormal data appears continuously or periodically within a given period of time, then the abnormal data is determined to be true. When the cloud monitoring platform determines that any abnormal data in the cable temperature data, load current data, or vibration data is true, it judges that the high-voltage cable has experienced the corresponding abnormality type. When the cloud monitoring platform determines that the abnormal pulse current data is true and the electromagnetic wave data is normal, it will only determine that the high-voltage cable has a partial discharge abnormality if other abnormal types are also determined at the same time; otherwise, it will be determined that the pulse current is abnormal. When the cloud monitoring platform determines that both pulse current data and electromagnetic wave data are true, or only electromagnetic wave data is true, it directly judges that the high-voltage cable has a partial discharge abnormality. The cloud monitoring platform records historical diagnostic results, counts the frequency of abnormal types in each monitoring section, and displays the statistical results on the monitoring terminal.

Images