Comprehensive online monitoring method and system based on GIS partial discharge

Abstract

The invention discloses a comprehensive online monitoring method and system based on GIS partial discharge, and relates to the technical field of power monitoring, and the method comprises the steps: obtaining the operation state information of GIS equipment, and constructing a partial discharge monitoring model; acquiring a multi-modal sensing signal through the partial discharge monitoring model to obtain fused monitoring data; calculating a filter coefficient based on environment characteristic parameters and equipment operation characteristics, performing interference suppression processing on the fused monitoring data, and outputting filtered monitoring data; performing hierarchical processing on the filtered monitoring data by adopting a cloud edge collaborative architecture, extracting basic feature parameters through edge nodes, and transmitting the basic feature parameters to cloud nodes for deep learning analysis; and a deep learning algorithm is combined to identify a partial discharge fault feature mode, and a GIS equipment comprehensive online monitoring result is generated based on multi-dimensional feature fusion and a consistency verification mechanism. According to the invention, intelligent identification and credibility evaluation of the fault mode are realized, the misjudgment rate is obviously reduced, and the accuracy of maintenance decision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system monitoring, in particular to a comprehensive online monitoring method and system based on GIS partial discharge. BACKGROUND

[0002] Briefly introduce the current development situation, and the second paragraph mainly writes the existing problems and the corresponding technical problems and solutions. For example: supercritical pressure parameter thermal power generation is a new technology for effective utilization of energy, and the temperature and pressure of its steam exceed those of any previous unit, which can greatly improve the thermal efficiency of the unit. However, high parameters also require higher safety and in-service supervision of various components of the unit. In order to prevent early failure of important metal components of thermal power generating units, the current domestic method is mainly offline, indirect and static preventive inspection, that is, using the opportunity of shutdown maintenance, using non-destructive testing to conduct on-site inspection of important components, and timely discovering defects and processing, which also forms a system of inspection and maintenance procedures.

[0003] However, the disadvantages of such treatment methods are: limited inspection opportunity: that is, the inspection can only be carried out during shutdown maintenance, and cannot be carried out online and in real time; high cost and long cycle: in order to ensure the metal inspection requirement of the unit components, the unit needs to be cooled in advance, scaffolding is set up, the insulation layer is removed, and polishing is carried out, which not only consumes a long shutdown time, but also consumes a large amount of maintenance cost; the inspection result has limitations: the current inspection is offline, indirect and static, based on theoretical model calculation with defects, and can only predict the influence of defects on the safety of components, and cannot achieve the purpose of direct and real-time monitoring and prediction. SUMMARY

[0004] The purpose of this part is to summarize some aspects of the embodiments of the present application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this part and the abstract and title of the specification of the present application in order to avoid obscuring the purpose of this part, the abstract and the title, and such simplifications or omissions cannot be used to limit the scope of the present application.

[0005] In view of the above-mentioned existing problems, the present application is proposed.

[0006] Therefore, the present application provides a comprehensive online monitoring method and system based on GIS partial discharge, which can solve the problems mentioned in the background art.

[0007] To solve the above technical problems, the present application provides the following technical solutions: In a first aspect, the present application provides a comprehensive online monitoring method based on GIS partial discharge, which comprises: acquiring GIS device operating state information and constructing a partial discharge monitoring model. The multi-modal sensing signals are collected through the partial discharge monitoring model to obtain fusion monitoring data; The filtering coefficients are calculated based on the environmental characteristic parameters and the equipment operation characteristics, the fusion monitoring data are subjected to interference suppression processing, and filtered monitoring data are outputted; The filtered monitoring data are subjected to hierarchical processing by using a cloud-edge collaborative architecture, basic characteristic parameters are extracted by an edge node and transmitted to a cloud node for deep learning analysis; The partial discharge fault characteristic modes are identified in combination with a deep learning algorithm, and GIS equipment comprehensive online monitoring results are generated based on multi-dimensional feature fusion and consistency verification mechanism.

[0008] As a preferred scheme of the comprehensive online monitoring method based on GIS partial discharge, the GIS equipment operation state information includes voltage grade, rated current, operation load rate and environmental temperature and humidity parameters of the GIS equipment, and a device basic operation state database is established; The construction of the partial discharge monitoring model includes: A device basic operation state database is established according to GIS equipment operation state information, ultrasonic sensors, ultrahigh frequency sensors and pulse current sensors are deployed at key monitoring positions of the GIS equipment, and sampling frequencies and range of each sensor are configured; Based on the geometric structure parameters and material characteristics of the GIS equipment, a multi-physical field coupling model including electric field distribution, magnetic field distribution and acoustic field distribution is constructed; According to historical partial discharge cases and equipment fault records, a typical fault mode library including corona discharge, surface discharge and floating discharge is established; The device operation state database, sensor configuration parameters, multi-physical field coupling model and fault mode library are integrated to construct a partial discharge monitoring model.

[0009] As a preferred scheme of the comprehensive online monitoring method based on GIS partial discharge, the fusion monitoring data are obtained by: The ultrasonic sensors, ultrahigh frequency sensors and pulse current sensors are started, and acoustic signals, electromagnetic signals and current pulse signals generated by partial discharge are synchronously collected according to preset sampling frequencies; The collected original sensing signals are preprocessed, including signal amplification, filtering and analog-to-digital conversion, to eliminate noise and deviation of the sensors themselves; Time domain characteristic parameters, frequency domain characteristic parameters and phase domain characteristic parameters of each sensor signal are extracted to construct multi-dimensional feature vectors; Based on the spatial position relationship and signal propagation characteristics of the sensors, the characteristic parameters collected by the multi-sensors are subjected to time synchronization correction and spatial matching; The corrected multi-dimensional feature parameters are fused by using a weighted fusion algorithm to generate fusion monitoring data containing complete partial discharge information.

[0010] As a preferred scheme of the GIS-based partial discharge comprehensive online monitoring method, the output filtered monitoring data comprises: The temperature, humidity and electromagnetic interference intensity in the environmental feature parameters, and the voltage level, load current and running time in the equipment operation features are acquired; According to the acquired environmental feature parameters and equipment operation features, a preset filter coefficient mapping table is queried to determine the filter parameters corresponding to each interference source; An adaptive filter is constructed based on the determined filter parameters, including a combination configuration of a low-pass filter, a high-pass filter and a band-pass filter; The fusion monitoring data are input into the constructed adaptive filter, and the signals are segmented and filtered according to the frequency response characteristics of the filter; The filtered data are subjected to signal-to-noise ratio evaluation and feature retention degree verification, and filtered monitoring data meeting preset quality standards are output.

[0011] As a preferred scheme of the GIS-based partial discharge comprehensive online monitoring method, the deep learning analysis comprises: The filtered monitoring data are input into an edge computing node for basic feature extraction with high real-time requirement, including amplitude statistics, frequency distribution and phase feature calculation; A feature vector is generated based on the extracted basic feature parameters, and the feature vector is compressed and encoded according to a preset data compression algorithm to reduce data transmission volume; The compressed feature vector is transmitted to a cloud computing node through a secure communication protocol to ensure the integrity and security during data transmission; The cloud computing node receives the transmitted feature vector, decompresses and reconstructs the complete feature data set, and constructs the input data of the deep learning model; Based on the reconstructed feature data set, a pre-trained deep learning model is used for partial discharge fault pattern recognition and classification, and the recognition result is output.

[0012] As a preferred scheme of the GIS-based partial discharge comprehensive online monitoring method, the generation of the GIS device comprehensive online monitoring result comprises: The partial discharge fault feature pattern recognized by the deep learning algorithm is matched with a preset fault pattern library to determine the confidence score of the fault type; Based on a multi-dimensional feature fusion algorithm, the time domain feature, frequency domain feature and phase domain feature are weighted and fused to generate comprehensive feature parameters; The consistency checking mechanism is used to cross-verify the multi-sensor identification results, and a consistency coefficient between the results of each sensor is calculated. According to the confidence score, the comprehensive feature parameter, and the consistency coefficient, a monitoring result evaluation model is constructed, and a device state health index is calculated. Based on the output of the monitoring result evaluation model, a GIS device comprehensive online monitoring report containing fault type, severity, development trend, and maintenance suggestion is generated.

[0013] As a preferred scheme of the GIS partial discharge-based comprehensive online monitoring method, the consistency coefficient is calculated using the Pearson correlation coefficient, and when the correlation coefficient is greater than 0.8, it is determined as high consistency, when the correlation coefficient is between 0.5 and 0.8, it is determined as medium consistency, and when the correlation coefficient is less than 0.5, it is determined as low consistency, and the consistency coefficient calculation formula is: ; Wherein, r is the Pearson correlation coefficient, is the identification result of the ultrasonic sensor at the i-th moment, is the average value of the ultrasonic sensor identification result, is the identification result of the ultrahigh frequency sensor at the i-th moment, y is the average value of the ultrahigh frequency sensor identification result, Σ represents the summation operation, and i is the time series index.

[0014] In a second aspect, the present application provides a GIS partial discharge-based comprehensive online monitoring system, which comprises: a data acquisition modeling module for acquiring GIS device operating state information and constructing a partial discharge monitoring model; A multi-modal signal acquisition module is used to acquire multi-modal sensing signals through the partial discharge monitoring model to obtain fusion monitoring data. A filter output processing module is used to calculate a filter coefficient based on environmental feature parameters and device operating characteristics, to perform interference suppression processing on the fusion monitoring data, and to output filtered monitoring data. A cloud-edge collaborative analysis module is used to perform hierarchical processing on the filtered monitoring data using a cloud-edge collaborative architecture, to extract basic feature parameters through an edge node and transmit them to a cloud node for deep learning analysis. A comprehensive monitoring evaluation module is used to identify partial discharge fault feature patterns in combination with a deep learning algorithm, and to generate GIS device comprehensive online monitoring results based on multi-dimensional feature fusion and consistency checking mechanism.

[0015] In a third aspect, the present application provides a computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the GIS partial discharge-based comprehensive online monitoring method when executing the computer program.

[0016] In a fourth aspect, the present application provides a computer readable storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the steps of the GIS-based partial discharge comprehensive online monitoring method.

[0017] Compared with the prior art, the present application has the following beneficial effects: by acquiring GIS device operating state information and constructing a partial discharge monitoring model, the monitoring system is accurately matched with the actual working condition of the device; by multi-modal sensor signal synchronous acquisition and weighted fusion algorithm, the sound, electric and magnetic signals are cooperatively processed and information is complemented, significantly improving the comprehensiveness and accuracy of fault detection; by adaptive filter coefficient calculation based on environmental and operating characteristics, dynamic interference suppression is realized for specific working conditions, effectively improving the signal-to-noise ratio and data quality; by the hierarchical processing mechanism of the cloud-edge collaborative architecture, the organic combination of edge real-time processing and cloud deep analysis is realized, ensuring real-time while fully utilizing the advantages of big data analysis; by the deep learning algorithm combined with multi-dimensional feature fusion and consistency verification mechanism, intelligent identification and reliability evaluation of fault modes are realized, significantly reducing the misjudgment rate and improving the accuracy of maintenance decisions. Overall, the present application solves the technical problems of low accuracy, weak anti-interference ability and low processing efficiency of traditional GIS partial discharge monitoring methods, and provides an efficient and reliable technical solution for the state monitoring and preventive maintenance of power equipment. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 A method flowchart of a GIS-based partial discharge comprehensive online monitoring method and system provided by an embodiment of the present application; Figure 2 An internal structure diagram of a computer device of a GIS-based partial discharge comprehensive online monitoring method and system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned purposes, features and advantages of the present application more apparent, the specific embodiments of the present application will be described in detail below with reference to the drawings in the specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0021] In the following description, numerous specific details are set forth to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without the specific details. In other instances, well-known methods have not been described in detail in order not to unnecessarily obscure aspects of the present application.

[0022] The relative arrangement of parts and steps, numerical expressions, and numerical values set forth in these embodiments are not intended to limit the scope of the present disclosure unless otherwise specifically stated.

[0023] It should be understood, of course, that the dimensions of the various parts illustrated in the various drawings are shown for simplicity and the actual dimensions can depend on the specific application.

[0024] Techniques, methods, and apparatus known to the relevant skilled person can not be discussed in detail, but should be considered part of the specification where appropriate.

[0025] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as a limitation. Thus, other examples of the example embodiments can have different values.

[0026] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, or characteristic under discussion. The appearance of the phrase "in one embodiment" in various places in the specification is not meant to refer to the same embodiment, nor is it meant to refer to only one embodiment or a particular embodiment. Rather, the phrase "in one embodiment" is meant to refer to at least one embodiment.

[0027] In addition, in the description of the present disclosure, the terms "first", "second", "third", and the like are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance and sequence. Similarly, although the operations are depicted in a specific order in the drawings, this should not be understood as requiring the specific order or sequential order shown to perform such operations, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multi-task processing and parallel processing can be advantageous.

[0028] Embodiment 1, refer to Figure 1 For the first embodiment of the present application, the embodiment provides a GIS-based comprehensive online monitoring method of partial discharge, comprising: Figure 1 A method flowchart of a GIS-based comprehensive online monitoring method and system of partial discharge is shown, comprising: S1: Obtain GIS device operating state information and construct a partial discharge monitoring model; S2: Collecting multi-modal sensing signals through the partial discharge monitoring model to obtain fusion monitoring data; S3: Calculating filtering coefficients based on environmental characteristic parameters and equipment operation characteristics, performing interference suppression processing on the fusion monitoring data, and outputting filtered monitoring data; S4: Performing hierarchical processing on the filtered monitoring data using a cloud-edge collaborative architecture, extracting basic characteristic parameters through an edge node and transmitting them to a cloud node for deep learning analysis; S5: Identifying partial discharge fault feature patterns in combination with deep learning algorithms, and generating GIS equipment comprehensive online monitoring results based on multi-dimensional feature fusion and consistency verification mechanisms.

[0029] It should be noted that during the operation of the GIS equipment, the partial discharge signal has the characteristics of strong instantaneousness, wide frequency range, and complex interference signal, making it difficult to accurately extract and identify the partial discharge characteristics. At the same time, the GIS equipment operates in a complex environment, and factors such as electromagnetic interference, temperature changes, and humidity fluctuations can significantly affect the monitoring signal, resulting in high false alarm rates and large missed alarm rates for traditional monitoring methods. In addition, the partial discharge fault mode is diverse, and a single sensor cannot fully reflect the fault characteristics, while the simple superposition of multi-sensor data can easily produce data redundancy and conflicts.

[0030] Therefore, to solve the above technical problems of low accuracy, weak anti-interference ability, and difficult fault identification of partial discharge monitoring, steps S1-S5 are used to construct a multi-dimensional monitoring model to achieve accurate collection and effective filtering of partial discharge signals. A cloud-edge collaborative hierarchical processing architecture is used to improve data processing efficiency and real-time performance. Deep learning algorithms and multi-sensor fusion technology are combined to accurately identify and comprehensively evaluate partial discharge fault modes, providing reliable technical support for the condition monitoring and preventive maintenance of GIS equipment.

[0031] Embodiment 2, refer to Figures 1-2 For the second embodiment of the present application, the present embodiment also provides a comprehensive online monitoring method based on GIS partial discharge, comprising: In the present application, the GIS equipment operation state information is obtained and the partial discharge monitoring model is constructed in step S1 by the following steps, including: S11: Collecting voltage levels, rated currents, operating load rates, and environmental temperature and humidity parameters of the GIS equipment, and establishing a basic equipment operation state database; In an optional embodiment, the voltage levels include 110kV, 220kV, and 500kV levels, the rated current range is 630A to 4000A, the operating load rate is obtained by real-time measurement of the current transformer, and the environmental temperature and humidity parameters are collected by the temperature and humidity sensor.

[0032] In an optional embodiment, the voltage level is 220 kV, the rated current is 2500 A, the operating load rate is determined by measuring the ratio of the average value of the three-phase current to the rated current, the environmental temperature sensor has an accuracy of ±0.5℃, the humidity sensor has an accuracy of ±3%RH, and the sampling period is 1 minute.

[0033] In an optional embodiment, data quality verification is further performed on the collected operating state parameters, abnormal data points are removed, and linear interpolation is used to complete missing data, so as to ensure the integrity and accuracy of the data in the database.

[0034] S12: Deploying ultrasonic sensors, ultrahigh frequency sensors and pulse current sensors at key monitoring positions of the GIS device, and configuring the sampling frequency and range of each sensor; In an optional embodiment, the working frequency range of the ultrasonic sensor is 20 kHz-300 kHz, the working frequency range of the ultrahigh frequency sensor is 300 MHz-1.5 GHz, and the measurement range of the pulse current sensor is 1 pC-10000 pC. Each sensor is connected to the data acquisition unit by a shielded cable.

[0035] In an optional embodiment, the ultrasonic sensor is installed near the basin-type insulator of the GIS device, the ultrahigh frequency sensor is installed on the grounding downlead of the device shell, and the pulse current sensor is sleeved on the grounding wire of the last screen of the device. The sampling frequency is set to 50 MHz.

[0036] S13: Based on the geometric structure parameters and material characteristics of the GIS device, a multi-physical field coupling model including electric field distribution, magnetic field distribution and acoustic field distribution is constructed; In an optional embodiment, the geometric structure parameters include the diameter of the device shell, the diameter of the conductor, the thickness of the insulator and the size of the shielding cover, and the material characteristics include the relative permittivity, the magnetic permeability and the acoustic impedance. A three-dimensional multi-physical field coupling model is established by using the finite element method.

[0037] S14: According to historical partial discharge cases and device failure records, a typical fault mode library including corona discharge, surface discharge and floating discharge is established; In an optional embodiment, the characteristics of corona discharge are small discharge pulse amplitude and high frequency, the characteristics of surface discharge are that the discharge pulses are concentrated near the voltage peak value, and the characteristics of floating discharge are large discharge pulse amplitude and wide phase distribution. Each fault mode includes characteristic parameters in time domain, frequency domain and phase domain.

[0038] S15: Integrating the device operating state database, the sensor configuration parameters, the multi-physical field coupling model and the fault mode library, a partial discharge monitoring model is constructed.

[0039] In the embodiment of the present application, the multi-modal sensor signals in step S2 are collected by the following steps to obtain the fusion monitoring data, comprising: S21: Start the ultrasonic sensor, the ultrahigh frequency sensor and the pulse current sensor, and synchronously collect the acoustic signals, electromagnetic signals and current pulse signals generated by partial discharge according to the preset sampling frequency; In an optional embodiment, the ultrasonic sensor adopts a center frequency of 200 kHz, the ultrahigh frequency sensor adopts a center frequency of 1.2 GHz, the pulse current sensor adopts a range of 1000 pC, and the sampling frequency is uniformly set to 50 MHz. The microsecond-level synchronous collection of the three signals is realized through the GPS synchronous clock.

[0040] In an optional embodiment, the sampling duration is 10 seconds, the signal data of the three sensors is continuously collected in each sampling period, the number of ultrasonic signal sampling points is 1000000 points, the number of ultrahigh frequency signal sampling points is 1000000 points, and the number of pulse current signal sampling points is 1000000 points, so as to ensure the integrity and synchronism of the signals.

[0041] S22: Preprocess the collected original sensor signals, including signal amplification, filtering and analog-to-digital conversion, to eliminate the noise and deviation of the sensor itself; In an optional embodiment, the signal amplification multiple is 100 times, a Butterworth low-pass filter is used for anti-aliasing filtering, the cutoff frequency is set to 0.4 times the sampling frequency, the analog-to-digital converter resolution is 16 bits, and the sampling accuracy is ±0.1 mV.

[0042] In an optional embodiment, the preprocessing further includes baseline correction and DC component removal. The signal baseline is calculated by using the moving average method, and the DC component is removed by using a high-pass filter. The filter cutoff frequency is 50 Hz.

[0043] S23: Extract the time domain feature parameters, frequency domain feature parameters and phase domain feature parameters of the signals of each sensor, and construct a multi-dimensional feature vector; In an optional embodiment, the time domain feature parameters include signal amplitude, pulse width and rise time, the frequency domain feature parameters include spectral peak, barycenter frequency and frequency band energy, and the phase domain feature parameters include phase distribution, phase concentration and number of pulses in the phase window.

[0044] In an optional embodiment, the time domain feature parameter is specifically: signal peak 1.5V-5.0V, pulse width 10ns-100ns, rise time 1ns-10ns; the frequency domain feature parameter is specifically: the spectral peak appears in the range of 300MHz-800MHz, the center of gravity frequency is 400MHz-600MHz, and the frequency band energy calculation range is 200MHz-1000MHz; the phase domain feature parameter is specifically: the phase distribution is counted in the range of 0°-360°, the phase concentration is calculated by using the standard deviation, and the phase window is set to 30°.

[0045] S24: based on the spatial position relationship of the sensor and the signal propagation characteristics, the feature parameters collected by the multiple sensors are time-synchronized and corrected and spatially matched; In an optional embodiment, the time synchronization correction adopts a cross-correlation algorithm to calculate the time delay between the signals of each sensor, the spatial matching is based on the signal arrival time difference and the speed of sound and the speed of electromagnetic wave for positioning calculation, and the correction accuracy reaches the nanosecond level.

[0046] S25: the corrected multi-dimensional feature parameters are fused by using a weighted fusion algorithm to generate fusion monitoring data containing complete partial discharge information.

[0047] In an optional embodiment, the weighted fusion algorithm allocates weights based on the signal-to-noise ratio and detection sensitivity of each sensor, the weight coefficient is determined through historical data training, and the fusion formula is as follows: ; ; Wherein, F is the fused comprehensive feature parameter, is the weight coefficient of the ultrasonic sensor, F1 is the feature parameter extracted by the ultrasonic sensor, is the weight coefficient of the ultrahigh frequency sensor, F2 is the feature parameter extracted by the ultrahigh frequency sensor, is the weight coefficient of the pulse current sensor, and F3 is the feature parameter extracted by the pulse current sensor.

[0048] In an optional embodiment, the fusion monitoring data includes comprehensive amplitude characteristics, comprehensive frequency spectrum characteristics and comprehensive phase characteristics, the data format adopts a JSON structure for storage, and includes a time stamp, a sensor identifier, a feature parameter value and a confidence score.

[0049] In the embodiment of the application, the interference suppression processing in step S3 includes the following steps: S31: acquiring temperature, humidity and electromagnetic interference intensity in the environmental feature parameters, and voltage level, load current and running time in the equipment running features; In an optional embodiment, the environmental temperature range is -10℃ to 50℃, the humidity range is 10%RH to 90%RH, the electromagnetic interference intensity is measured by a spectrum analyzer, the voltage level includes 110kV, 220kV, 500kV, the load current is 20%-100% of the rated current, and the operating time is the number of hours of continuous operation of the equipment.

[0050] In an optional embodiment, the temperature sensor accuracy is ±0.5℃, the humidity sensor accuracy is ±3%RH, the electromagnetic interference intensity measurement frequency range is 0.1MHz-100MHz, the voltage level is 220kV, the load current is measured in real time by a current transformer, and the operating time is calculated by a system clock.

[0051] S32: According to the obtained environmental characteristic parameters and equipment operation characteristics, a preset filter coefficient mapping table is queried to determine the filter parameters corresponding to each interference source; In an optional embodiment, the preset filter coefficient mapping table contains a three-dimensional correspondence relationship of temperature-filter coefficient, humidity-filter coefficient, and electromagnetic interference intensity-filter coefficient. When the temperature is 25℃, the humidity is 60%RH, and the electromagnetic interference intensity is -60dBm, the corresponding low-pass filter cutoff frequency is 10MHz, and the high-pass filter cutoff frequency is 50Hz.

[0052] In an optional embodiment, the filter parameters include a low-pass filter cutoff frequency range of 1MHz-50MHz, a high-pass filter cutoff frequency range of 10Hz-200Hz, a band-pass filter passband range of 100kHz-500kHz, and a filter order of 2-10 orders.

[0053] S33: An adaptive filter is constructed based on the determined filter parameters, including a combination configuration of a low-pass filter, a high-pass filter, and a band-pass filter; In an optional embodiment, the adaptive filter adopts a FIR digital filter structure, the low-pass filter is used to eliminate high-frequency noise, the high-pass filter is used to remove the direct current component, the band-pass filter is used to extract the partial discharge characteristic frequency band, and the three filters are combined in a series mode.

[0054] S34: The fusion monitoring data is input into the constructed adaptive filter, and the signal is subjected to segmented filter processing according to the frequency response characteristics of the filter; In an optional embodiment, the segmented filter processing includes: first removing the 0-50Hz direct current and power frequency components through a high-pass filter, then extracting the 100kHz-500kHz partial discharge characteristic signal through a band-pass filter, and finally suppressing the 50MHz and above high-frequency noise through a low-pass filter.

[0055] S35: Perform signal-to-noise ratio evaluation and feature retention verification on the filtered data, and output the filtered monitoring data that meets the preset quality standard.

[0056] In an optional embodiment, the signal-to-noise ratio evaluation adopts a signal power-to-noise power ratio calculation, and the feature retention verification is performed by comparing the change rate of the feature parameters before and after filtering. When the signal-to-noise ratio is greater than 10 dB and the feature parameter change rate is less than 5%, it is determined that the preset quality standard is met.

[0057] In an optional embodiment, the preset quality standard includes: the signal-to-noise ratio is not less than 8 dB, the feature parameter retention rate is not less than 95%, and the data integrity reaches 100%. The filtered monitoring data that meets the standard is output in a digital signal format and stored in a database.

[0058] In the embodiments of the present application, the basic feature parameters are extracted in step S4 and transmitted to the cloud node for deep learning analysis by the following steps, including: S41: Input the filtered monitoring data into the edge computing node, and perform basic feature extraction with high real-time requirement on the data, including amplitude statistics, frequency distribution and phase feature calculation; In an optional embodiment, the edge computing node adopts an ARM Cortex-A53 processor, and the basic feature extraction includes: calculating the signal peak value in the range of 0.1V-10V, the root mean square value in the range of 0.01V-5V, the frequency distribution is calculated by fast Fourier transform, and the phase feature is the pulse distribution in the range of 0°-360°.

[0059] In an optional embodiment, the processing delay of the basic feature extraction is not more than 100ms, the amplitude statistics includes maximum value, minimum value, average value and standard deviation, the frequency distribution calculation includes the frequency spectrum energy distribution within 500MHz bandwidth, and the phase feature calculation includes the pulse count within each 30° phase window.

[0060] S42: Generate a feature vector based on the extracted basic feature parameters, and compress and encode the feature vector according to a preset data compression algorithm to reduce the data transmission amount; In an optional embodiment, the feature vector has a dimension of 128, adopts a wavelet transform compression algorithm, has a compression ratio of 4:1, and after compression, the data amount is reduced from 1MB to 256KB, and the compression distortion rate is less than 1%.

[0061] S43: Transmit the compressed feature vector to the cloud computing node through a secure communication protocol to ensure the integrity and security of the data transmission process; In an optional embodiment, the secure communication protocol adopts TLS 1.3 encrypted transmission, the transmission bandwidth is 10 Mbps, the single transmission delay is less than 500 ms, the data integrity is verified by CRC32 check code, and the security is guaranteed by AES-256 encryption algorithm.

[0062] S44: The cloud computing node receives the transmitted feature vector, reconstructs the complete feature data set after decompression, and constructs the input data of the deep learning model; In an optional embodiment, the cloud computing node adopts a GPU cluster server, the decompression algorithm corresponds to the compression algorithm, the reconstruction error is less than 0.5%, the input data format is a 128-dimensional floating-point number vector, and the data set includes a training set, a validation set, and a test set.

[0063] S45: Based on the reconstructed feature data set, a pre-trained deep learning model is used for partial discharge fault pattern recognition and classification, and an identification result is output.

[0064] In an optional embodiment, the deep learning model adopts a convolutional neural network architecture, includes 3 convolutional layers, 2 pooling layers, and 1 fully connected layer, the model parameter quantity is 1 million, the fault recognition accuracy is not less than 95%, and the identification response time is less than 2 seconds.

[0065] In an optional embodiment, the partial discharge fault pattern includes corona discharge, surface discharge, floating discharge, and internal discharge, the classification confidence is calculated using a softmax function, and the output result includes a fault type identifier and a corresponding confidence score, the confidence score is a floating-point number between 0 and 1.

[0066] In the embodiments of the present application, the GIS device comprehensive online monitoring result is generated in step S5 by the following steps, including: S51: The partial discharge fault feature pattern recognized by the deep learning algorithm is matched with the preset fault pattern library to determine the confidence score of the fault type; In an optional embodiment, the fault pattern library includes four standard patterns of corona discharge, surface discharge, floating discharge, and internal discharge, the confidence score is calculated by Euclidean distance to calculate the similarity, when the distance is less than 0.1, the confidence score is 0.95, when the distance is between 0.1 and 0.3, the confidence score is 0.8, and when the distance is greater than 0.3, the confidence score is 0.5.

[0067] In an optional embodiment, the matching algorithm adopts a K-nearest neighbor algorithm, the K value is set to 5, the feature distance of the test sample and the training sample is calculated, the 5 nearest samples are selected for voting decision, and the confidence score is the statistical probability of the voting result.

[0068] S52: Based on the multi-dimensional feature fusion algorithm, the time domain feature, the frequency domain feature and the phase domain feature are weighted and fused to generate a comprehensive feature parameter; In an optional embodiment, the multi-dimensional feature fusion algorithm adopts principal component analysis method, the time domain feature weight is 0.4, the frequency domain feature weight is 0.35, the phase domain feature weight is 0.25, and the comprehensive feature parameter dimension is 32 dimensions, containing the main component information of each dimension.

[0069] S53: A consistency checking mechanism is used to cross-verify the multi-sensor recognition results, and a consistency coefficient between the results of each sensor is calculated; In an optional embodiment, the consistency coefficient is calculated by Pearson correlation coefficient, when the correlation coefficient is greater than 0.8, it is determined as high consistency, when the correlation coefficient is between 0.5 and 0.8, it is determined as medium consistency, and when the correlation coefficient is less than 0.5, it is determined as low consistency, and the consistency coefficient calculation formula is: ; Wherein, r is the Pearson correlation coefficient, is the recognition result of the ultrasonic sensor at the i-th moment, is the average value of the ultrasonic sensor recognition result, is the recognition result of the ultrahigh frequency sensor at the i-th moment, y is the average value of the ultrahigh frequency sensor recognition result, ∑ represents summation operation, and i is the time series index.

[0070] S54: According to the confidence score, the comprehensive feature parameter and the consistency coefficient, a monitoring result evaluation model is constructed, and a device state health index is calculated; In an optional embodiment, the monitoring result evaluation model adopts weighted summation algorithm, and the health index calculation formula is: ; Wherein, HI is the health index, CS is the confidence score, CF is the comprehensive feature parameter matching degree, CC is the consistency coefficient, and the health index range is 0-100.

[0071] S55: Based on the output of the monitoring result evaluation model, a GIS device comprehensive online monitoring report containing fault type, severity, development trend and maintenance suggestion is generated.

[0072] In an optional embodiment, the monitoring report adopts five-level evaluation standard: health index 90-100 is normal state, 75-89 is attention state, 60-74 is abnormal state, 40-59 is serious state, and 0-39 is critical state, each state corresponds to different maintenance suggestions and processing time limit.

[0073] Embodiment 3 is an embodiment of the application, and the embodiment also provides a GIS partial discharge-based comprehensive online monitoring system, comprising: a data acquisition and modeling module configured to acquire GIS device operation state information and construct a partial discharge monitoring model; a multi-modal signal acquisition module configured to acquire multi-modal sensing signals through the partial discharge monitoring model to obtain fusion monitoring data; a filtering and output processing module configured to calculate a filtering coefficient based on environmental characteristic parameters and device operation characteristics, perform interference suppression processing on the fusion monitoring data, and output filtered monitoring data; a cloud-edge collaborative analysis module configured to perform hierarchical processing on the filtered monitoring data using a cloud-edge collaborative architecture, extract basic characteristic parameters through an edge node and transmit the basic characteristic parameters to a cloud node for deep learning analysis; a comprehensive monitoring and evaluation module configured to identify partial discharge fault feature patterns in combination with a deep learning algorithm, and generate GIS device comprehensive online monitoring results based on multi-dimensional feature fusion and consistency verification mechanisms.

[0074] The embodiment also provides a computer device, which can be a terminal, and an internal structure diagram of the computer device can be as shown in Figure 2 The computer device comprises a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement a GIS partial discharge-based comprehensive online monitoring method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or can be a key, trackball or touchpad arranged on the computer device shell, or can be an external keyboard, touchpad or mouse, etc.

[0075] The embodiment also provides a computer-readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the following steps: acquire GIS device operation state information and construct a partial discharge monitoring model; acquire multi-modal sensing signals through the partial discharge monitoring model to obtain fusion monitoring data; The filter coefficient is calculated based on the environmental characteristic parameters and the equipment operation characteristics, interference suppression processing is performed on the fusion monitoring data, and filtered monitoring data is output; The filtered monitoring data is processed in layers using a cloud-edge collaborative architecture, basic characteristic parameters are extracted by an edge node and transmitted to a cloud node for deep learning analysis; The deep learning algorithm is combined to identify the partial discharge fault characteristic mode, and the GIS equipment comprehensive online monitoring result is generated based on the multi-dimensional feature fusion and consistency verification mechanism.

[0076] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and they should be covered in the scope of the claims of the present application.

[0077] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0078] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The device that implements the functions specified in one flow or multiple flows and / or blocks Figure 1 The device that implements the functions specified in one flow or multiple flows and / or blocks

[0079] These computer program instructions can also be stored in a computer-readable storage medium that can guide the computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable storage medium produce a manufactured product including instruction devices that implement the functions specified in the flowcharts and / or block diagrams. Figure 1one or more processes and / or blocks Figure 1 the function specified in the one or more blocks.

[0080] These computer program instructions can also be loaded into computer or other programmable data processing devices, so that a series of operation steps are performed on the computer or other programmable data processing devices to generate computer-implemented processes, thus the instructions executed on the computer or other programmable data processing devices provide a process for implementing the flow Figure 1 one or more processes and / or blocks Figure 1 the function specified in the one or more blocks.

[0081] Although preferred embodiments of the application have been described, those skilled in the art will be able to make additional modifications and variations to these embodiments without departing from the spirit and scope of the application. Accordingly, the appended claims are intended to encompass all such modifications and variations as falling within the scope of the application.

[0082] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A comprehensive online monitoring method for partial discharge based on GIS, characterized in that: include, Acquire GIS equipment operating status information and construct a partial discharge monitoring model; Multimodal sensing signals are acquired using the partial discharge monitoring model to obtain fused monitoring data; Based on environmental characteristic parameters and equipment operating characteristics, the filtering coefficients are calculated, and interference suppression processing is performed on the fused monitoring data to output filtered monitoring data. A cloud-edge collaborative architecture is adopted to perform hierarchical processing on the filtered monitoring data. Basic feature parameters are extracted through edge nodes and transmitted to cloud nodes for deep learning analysis. By combining deep learning algorithms to identify partial discharge fault characteristic patterns, and based on multi-dimensional feature fusion and consistency verification mechanisms, comprehensive online monitoring results of GIS equipment are generated.

2. The integrated online monitoring method for partial discharge based on GIS as described in claim 1, characterized in that: The GIS equipment operating status information includes: the voltage level, rated current, operating load rate, and environmental temperature and humidity parameters of the GIS equipment, and a basic operating status database of the equipment is established. The construction of the partial discharge monitoring model includes: Establish a basic operating status database for the equipment based on the operating status information of the GIS equipment, deploy ultrasonic sensors, ultra-high frequency sensors and pulse current sensors at key monitoring locations of the GIS equipment, and configure the sampling frequency and range of each sensor. Based on the geometric parameters and material properties of GIS equipment, a multi-physics coupling model including electric field distribution, magnetic field distribution and sound field distribution is constructed. Based on historical partial discharge cases and equipment failure records, a library of typical failure modes including corona discharge, surface discharge, and floating discharge was established. By integrating equipment operation status database, sensor configuration parameters, multi-physics coupling model and fault mode library, a partial discharge monitoring model is constructed.

3. The integrated online monitoring method for partial discharge based on GIS as described in claim 2, characterized in that: The obtained fusion monitoring data includes: The ultrasonic sensor, ultra-high frequency sensor, and pulse current sensor are activated to synchronously acquire the acoustic signal, electromagnetic signal, and current pulse signal generated by partial discharge according to the preset sampling frequency. The raw sensor signals are preprocessed, including signal amplification, filtering and analog-to-digital conversion, to eliminate the noise and bias of the sensor itself; Extract the time-domain, frequency-domain, and phase-domain feature parameters of each sensor signal to construct a multi-dimensional feature vector; Based on the spatial positional relationship and signal propagation characteristics of the sensors, time synchronization correction and spatial matching are performed on the feature parameters collected by multiple sensors. A weighted fusion algorithm is used to fuse the corrected multi-dimensional feature parameters to generate fused monitoring data containing complete partial discharge information.

4. The integrated online monitoring method for partial discharge based on GIS as described in claim 3, characterized in that: The filtered output monitoring data includes: Obtain environmental characteristic parameters such as temperature, humidity, and electromagnetic interference intensity, as well as equipment operating characteristics such as voltage level, load current, and operating time; Based on the acquired environmental characteristic parameters and equipment operating characteristics, the preset filter coefficient mapping table is queried to determine the filter parameters corresponding to each interference source; An adaptive filter is constructed based on the determined filtering parameters, including a combination of low-pass, high-pass, and band-pass filters. The adaptive filter constructed by integrating monitoring data is used to perform segmented filtering of the signal according to the frequency response characteristics of the filter. The signal-to-noise ratio is evaluated and the feature preservation is verified on the filtered data, and the filtered monitoring data that meets the preset quality standards is output.

5. The integrated online monitoring method for partial discharge based on GIS as described in claim 4, characterized in that: The deep learning analysis includes: The filtered monitoring data is input into the edge computing node to perform real-time basic feature extraction, including amplitude statistics, frequency distribution and phase feature calculation. Feature vectors are generated based on the extracted basic feature parameters, and the feature vectors are compressed and encoded according to a preset data compression algorithm to reduce the amount of data transmission. The compressed feature vector is transmitted to the cloud computing node through a secure communication protocol to ensure the integrity and security of data transmission. The cloud computing node receives the transmitted feature vectors, decompresses them, reconstructs the complete feature dataset, and builds the input data for the deep learning model. Based on the reconstructed feature dataset, a pre-trained deep learning model is used to identify and classify partial discharge fault modes, and the identification results are output.

6. The integrated online monitoring method for partial discharge based on GIS as described in claim 5, characterized in that: The generated comprehensive online monitoring results of GIS equipment include: The partial discharge fault feature patterns identified by the deep learning algorithm are matched with a pre-set fault pattern library to determine the confidence score of the fault type. Based on a multi-dimensional feature fusion algorithm, time-domain features, frequency-domain features, and phase-domain features are weighted and fused to generate comprehensive feature parameters; A consistency verification mechanism is used to cross-validate the recognition results of multiple sensors and calculate the consistency coefficient between the results of each sensor. Based on confidence scores, comprehensive feature parameters, and consistency coefficients, a monitoring result evaluation model is constructed to calculate the equipment health index. Based on the output of the monitoring result evaluation model, a comprehensive online monitoring report for GIS equipment is generated, which includes fault type, severity, development trend, and maintenance recommendations.

7. The integrated online monitoring method for partial discharge based on GIS as described in claim 6, characterized in that: The consistency coefficient is calculated using the Pearson correlation coefficient. A correlation coefficient greater than 0.8 is considered high consistency, a correlation coefficient between 0.5 and 0.8 is considered moderate consistency, and a correlation coefficient less than 0.5 is considered low consistency. The formula for calculating the consistency coefficient is as follows: ； Where r is the Pearson correlation coefficient. Let be the recognition result of the ultrasonic sensor at time i. This represents the average value of the ultrasonic sensor's identification results. Let denot be the identification result of the UHF sensor at time i, ȳ be the average value of the UHF sensor identification result, Σ represent the summation operation, and i be the time series index.

8. A comprehensive online monitoring system for partial discharge based on GIS, based on the comprehensive online monitoring method for partial discharge based on GIS as described in any one of claims 1 to 7, characterized in that: include, The data acquisition and modeling module is used to acquire GIS equipment operating status information and build a partial discharge monitoring model; The multimodal signal acquisition module is used to acquire multimodal sensing signals through the partial discharge monitoring model to obtain fused monitoring data; The filtering output processing module is used to calculate the filtering coefficients based on environmental characteristic parameters and equipment operating characteristics, perform interference suppression processing on the fused monitoring data, and output the filtered monitoring data. The cloud-edge collaborative analysis module is used to perform hierarchical processing on filtered monitoring data using a cloud-edge collaborative architecture. It extracts basic feature parameters through edge nodes and transmits them to cloud nodes for deep learning analysis. The comprehensive monitoring and evaluation module is used to identify partial discharge fault characteristic patterns by combining deep learning algorithms, and to generate comprehensive online monitoring results of GIS equipment based on multi-dimensional feature fusion and consistency verification mechanisms.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the integrated online monitoring method for partial discharge based on GIS as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the integrated online monitoring method for partial discharge based on GIS as described in any one of claims 1 to 7.

Images