Online monitoring method and device for gas insulation cabinet

Abstract

The invention discloses a gas insulation cabinet on-line monitoring method and device, and relates to the technical field of insulation cabinets, and the method comprises the steps: building a discharge signal category library based on a gas insulation cabinet off-line test; a multi-channel sensing unit is optimally arranged in the key area to collect partial discharge signals; performing time-sharing superposition on the multi-channel signals, performing fast Fourier transform to obtain a frequency domain amplitude spectrum, and further extracting voltage frequency domain envelope of the insulation cabinet and voltage energy distribution of the insulation cabinet; calculating an energy distribution difference value with a category library by dynamically adjusting a frequency weight, calculating a frequency domain envelope similarity value by using a dynamic time warping method, obtaining a discharge signal category matching value by combining the two values, and identifying a discharge type. Based on the discharge type and the dielectric characteristics, an energy gradient estimation method is adopted to realize discharge source positioning; and collecting the time sequence characteristics of the positioning points for trend analysis to obtain a partial discharge signal evolution value, dividing the discharge risk level according to the partial discharge signal evolution value, and completing online monitoring and early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of insulating cabinets, and in particular to an online monitoring method and device for a gas insulating cabinet. Background Art

[0002] Gas-insulated metal-enclosed switchgear (GIS), with its significant advantages such as compact structure, high reliability, and strong environmental adaptability, has become a core device in modern high-voltage and ultra-high-voltage power transmission and transformation systems. It is widely used in substations, urban power grids, and power supply hubs for large industrial and mining enterprises. GIS equipment is filled with SF6 gas, which has excellent insulation and arc-extinguishing properties. Its long-term operation is directly related to the safety and stability of the power grid. Partial discharge is one of the most critical early signs of insulation degradation within GIS. Effective online monitoring and diagnosis of partial discharge is a key technical means to assess the insulation status of equipment, predict potential faults, implement condition-based maintenance, and ensure the safe operation of the power grid.

[0003] Traditional monitoring methods lack accuracy in identifying discharge types and rely solely on simple threshold comparisons of single signal features (such as discharge amplitude and frequency). They lack analysis of discharge signals from different physical mechanisms (such as corona discharge, surface discharge, and internal discharge), resulting in a high misjudgment rate in complex noise environments. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method and device for online monitoring of a gas-insulated cabinet to solve the problems in the background art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for online monitoring of a gas-insulated cabinet comprises the following steps: Step S1: embedding a plurality of partial discharge sensing channels in a key electric field non-uniform area in a gas-insulated cabinet in a reverse interleaving manner, collecting discharge signals of the gas-insulated cabinet through the partial discharge sensing channels, and obtaining a plurality of local voltage signal data of the gas-insulated cabinet; Step S2: performing time-sharing superposition on the plurality of local voltage signal data to obtain an insulation cabinet superimposed voltage signal, performing fast Fourier transform on the insulation cabinet superimposed voltage signal to obtain an insulation cabinet voltage frequency domain amplitude spectrum, smoothing the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage frequency domain envelope; and performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage energy distribution; Step S3: Performing offline testing on the gas-insulated cabinet and constructing a discharge signal category library, the discharge signal category library including: discharge category signal energy distribution and discharge category signal frequency domain envelope; dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, calculating the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, and obtaining the insulation cabinet energy distribution difference value; calculating the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope using a dynamic time warping method, and obtaining the insulation cabinet frequency domain similarity value; Step S4: calculating a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtaining a discharge signal category according to the discharge signal category matching value; Step S5: collecting and obtaining the internal dielectric absorption coefficient of the gas-insulated cabinet, and based on the type of the discharge signal and the internal dielectric absorption coefficient of the gas-insulated cabinet, performing location analysis on the local voltage signal data using an energy gradient estimation method to obtain the discharge source location; Step S6: Collect the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source position within a preset time, and perform trend change analysis to obtain the partial discharge signal evolution value; divide the discharge risk level according to the partial discharge signal evolution value to achieve online monitoring of the gas insulated cabinet.

[0006] Preferably, the time-sharing superposition of the plurality of local voltage signal data to obtain the insulation cabinet superposition voltage signal comprises the following specific steps: Calculate the energy contribution of partial discharge data for each channel:

[0007] in, is the energy contribution value of the local voltage signal data of the i-th channel, n is the index of the n-th sampling point, and N is the total number of samples in the time window; In the time window, the local voltage signal data is time-sharingly superimposed to obtain the insulation cabinet superimposed voltage signal:

[0008] in, Indicates the insulation cabinet superimposed voltage signal, Represents the total number of channels, i represents the index of the local voltage signal data of the i-th channel, is the energy contribution value of the local voltage signal data of the i-th channel, represents the local voltage signal data of the i-th channel, and n represents the index of the n-th sampling point.

[0009] Preferably, the method of performing fast Fourier transform on the insulation cabinet superimposed voltage signal to obtain the insulation cabinet voltage frequency domain amplitude spectrum, and smoothing the insulation cabinet voltage frequency domain amplitude spectrum to obtain the insulation cabinet voltage frequency domain envelope comprises the following steps: Perform fast Fourier transform on the superimposed discharge signal to obtain a frequency domain complex sequence :

[0010] in, ] is the complex amplitude of the kth frequency point, represents the superimposed voltage signal of the insulation cabinet, j is a complex unit, k is the frequency index in the frequency domain, and N is the total number of samples in the time window; Taking the modulus of the complex sequence in the frequency domain, we can obtain the frequency domain amplitude spectrum of the insulation cabinet voltage. |; The frequency domain amplitude spectrum of the insulation cabinet voltage is curve fitted to obtain the frequency domain envelope of the insulation cabinet voltage:

[0011] in, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of ] is the complex amplitude of the kth frequency point, is the Hilbert transform operator, is the frequency corresponding to the kth frequency point.

[0012] Preferably, the step of performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain the insulation cabinet voltage energy distribution comprises the following steps: By performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum, the insulation cabinet voltage energy distribution is obtained:

[0013] in, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, k is the frequency index in the frequency domain, N is the total number of samples in the time window, ] is the complex amplitude of the kth frequency point.

[0014] Preferably, dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, calculating the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, and obtaining the insulation cabinet energy distribution difference value comprises the following steps: Calculate the average value of the energy distribution of the discharge category signals in the discharge signal category library:

[0015] in, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the total number of discharge category signal energy distributions, For the The index of the energy distribution of the discharge category signal, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point; The weight of the energy ratio of each frequency point in the insulation cabinet voltage energy distribution is calculated by the average value of the discharge category signal energy distribution:

[0016] in, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the adjustment coefficient, >0, used to adjust the impact of the difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, is a constant term, >1; The reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value:

[0017] in, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, k is the frequency point index in the frequency domain, N is the total number of samples in the time window, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point.

[0018] Preferably, the method of calculating the cumulative distance between the frequency domain envelope of the insulation cabinet voltage and the frequency domain envelope of the discharge category signal by the dynamic time warping method to obtain the frequency domain similarity value of the insulation cabinet includes the following specific steps: The cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope is calculated by the dynamic time warping method to obtain the insulation cabinet frequency domain similarity value:

[0019] in, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of For the The frequency envelope of the discharge signal of each category is The envelope amplitude, DTW() is the dynamic time warping function, is the adjustment coefficient, which is used to adjust the impact of the difference between the frequency domain envelope of the insulation cabinet voltage and the frequency domain envelope of the discharge category signal.

[0020] Preferably, the method of calculating the discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value comprises the following specific steps: By combining the energy distribution difference value of the insulation cabinet and the frequency domain similarity value of the insulation cabinet, the discharge signal category matching value is calculated:

[0021] in, The local discharge signal and the discharge signal category matching value of the discharge signal category, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal.

[0022] Preferably, the performing location analysis on the local voltage signal data by an energy gradient estimation method to obtain the discharge source location comprises the following specific steps: The objective function of the energy gradient estimation method is:

[0023] Among them, F is the objective function of the energy gradient estimation method, is the response amplitude of the local voltage signal data of the i-th channel, ' is the theoretical amplitude calculated based on the preset discharge source position, is the actual signal arrival time difference between the ith channel and the reference channel, ' is the theoretical time difference between the i-th channel and the reference channel calculated based on the preset discharge source position, is the energy contribution of the local voltage signal data of the i-th channel, is the theoretical energy contribution value of the ith channel calculated based on the preset discharge source position, i is the index of the ith channel, Indicates the total number of channels; Adjust the preset discharge source candidate points by gradient descent method The coordinates of the energy gradient estimation method are Converges to the minimum value, then That is the discharge source location.

[0024] Preferably, the collecting of the insulation cabinet voltage frequency domain envelope and the insulation cabinet voltage energy distribution at the discharge power source position within a preset time, and performing trend change analysis to obtain the partial discharge signal evolution value comprises the following specific steps: Calculate the difference between the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution in each adjacent time window:

[0025] in, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The index of the time window, is the envelope amplitude difference adjustment coefficient, which is used to adjust the influence of envelope difference on the difference value of partial discharge signal. is the energy ratio difference adjustment coefficient, which is used to adjust the influence of the energy ratio difference adjustment coefficient on the partial discharge signal difference value. To prevent division by zero, the value is ; The partial discharge signal evolution value of the partial discharge signal within a preset time is calculated using the partial discharge signal difference value:

[0026] Among them, XB is the evolution value of the partial discharge signal, MS is the total number of time windows, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the length of the unit time window.

[0027] Preferably, the method of dividing the discharge risk level by the partial discharge signal evolution value to realize online monitoring of the gas insulated cabinet includes the following specific steps: The discharge risk level is divided by the partial discharge signal evolution value, and the online monitoring of the gas insulated cabinet is realized, and the multi-level risk level is set: XB , then the risk level is level 1, which is normal and requires routine inspection; <XB< , then the risk level is Level 2 Caution, and the monitoring frequency needs to be increased; <XB< , then the risk level is Level 3 alert, requiring dedicated monitoring; , then the risk level is Level 4 emergency and requires immediate maintenance.

[0028] An online monitoring device for a gas-insulated cabinet, comprising: Signal acquisition module: A plurality of partial discharge sensing channels are embedded in the key electric field non-uniform area of ​​the gas-insulated cabinet in an inverse interleaving manner. The discharge signals of the gas-insulated cabinet are collected through the partial discharge sensing channels to obtain a plurality of local voltage signal data of the gas-insulated cabinet; The insulation cabinet voltage energy distribution generation module is configured to perform time-sharing superposition on the plurality of local voltage signal data to obtain an insulation cabinet superposition voltage signal, perform fast Fourier transform on the insulation cabinet superposition voltage signal to obtain an insulation cabinet voltage frequency domain amplitude spectrum, smooth the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage frequency domain envelope, and perform energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage energy distribution. Insulation cabinet frequency domain similarity value generation module: By performing offline testing on the gas-insulated cabinet and constructing a discharge signal category library, the discharge signal category library includes: discharge category signal energy distribution and discharge category signal frequency domain envelope; by dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, the reconstruction difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value; using the dynamic time warping method to calculate the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope to obtain the insulation cabinet frequency domain similarity value; A discharge signal category generation module is configured to calculate a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtain a discharge signal category based on the discharge signal category matching value; Discharge source location determination module: collects and obtains the internal dielectric absorption coefficient of the gas-insulated cabinet, and based on the discharge signal type and the internal dielectric absorption coefficient of the gas-insulated cabinet, performs location analysis on the local voltage signal data using an energy gradient estimation method to determine the discharge source location; Monitoring module: collects the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source location within a preset time, and performs trend change analysis to obtain the partial discharge signal evolution value; divides the discharge risk level according to the partial discharge signal evolution value to realize online monitoring of the gas insulated cabinet.

[0029] Beneficial effects The present invention provides a method for online monitoring of a gas-insulated cabinet, which relates to a manufacturing technology for distribution switch control equipment and has the following beneficial effects: (1) The significance of dynamically adjusting the frequency weight and calculating the energy distribution difference value of the insulation cabinet. This step can adaptively highlight the frequency points with significant characteristic differences between the voltage signal and the discharge category library by dynamically adjusting the frequency weight, so that the energy distribution difference value of the insulation cabinet can more accurately reflect the degree of deviation between the real-time signal and the standard category. This process enhances the sensitivity to the characteristics of the discharge signal and can effectively identify the uniqueness of the energy distribution of different types of discharges (such as corona, surface discharge, etc.), providing a key quantitative basis for subsequent discharge category matching, and improving the accuracy and reliability of fault identification.

[0030] (2) The cumulative distance between the frequency domain envelope of the insulation cabinet voltage and the frequency domain envelope of the discharge category signal is calculated by the dynamic time warping method, and the frequency domain similarity value of the insulation cabinet is obtained. Its core significance lies in solving the nonlinear alignment problem of the frequency domain feature sequence, thereby accurately capturing the shape similarity of different discharge signals. Since the discharge signal in the gas-insulated cabinet is affected by factors such as the propagation path and dielectric attenuation, its frequency domain envelope may have an offset on the time axis or frequency axis (such as phase difference, frequency broadening, etc.). The traditional Euclidean distance cannot effectively handle such asynchronous features. By finding the optimal time bending path through the dynamic programming algorithm, the two frequency domain envelope sequences are allowed to be locally stretched or compressed in the time or frequency dimension, making the cumulative distance calculation more in line with the actual signal difference. The obtained frequency domain similarity value of the insulation cabinet can quantitatively characterize the degree of morphological matching between the real-time signal and the standard discharge category, and then assist in judging the discharge type (such as corona discharge, surface discharge, etc.), providing a key basis for the accurate classification of the discharge signal. At the same time, it can also identify the evolution law of early fault characteristics through the similarity value change trend in long-term equipment monitoring.

[0031] (3) The significance of collecting time series data and analyzing the evolution value of partial discharge signals By collecting the frequency domain envelope and energy distribution data within a preset time and analyzing the trend, the evolution law of the partial discharge signal can be obtained, such as the rate of change and stability of the characteristic parameters. The evolution value of the partial discharge signal can dynamically reflect the development trend of the discharge, such as the gradual change from normal to alert state, and provide quantitative support for the risk level classification in the time dimension. This mechanism can provide early warning of potential faults and help operation and maintenance personnel formulate differentiated maintenance strategies (such as routine inspections and emergency repairs) according to the risk level, realize forward-looking management of the status of gas-insulated cabinets, and reduce the risk of equipment failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0033] Figure 1 This is a flow chart of the steps of the on-line monitoring method for a gas-insulated cabinet proposed by the present invention; Figure 2 This is a hierarchical diagram of the steps of the online monitoring method for gas-insulated cabinets proposed by the present invention; DETAILED DESCRIPTION

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0035] See also Figure 1-2 The present invention provides a technical solution: an online monitoring method for a gas-insulated cabinet.

[0036] Step S1: embedding a plurality of partial discharge sensing channels in a key electric field non-uniform area in a gas-insulated cabinet in a reverse interleaving manner, collecting discharge signals of the gas-insulated cabinet through the partial discharge sensing channels, and obtaining a plurality of local voltage signal data of the gas-insulated cabinet.

[0037] In the gas insulated switchgear (GIS) casing, identify key areas where the electric field is unevenly distributed and prone to partial discharge, including busbar connection joints, high-voltage cable terminal interfaces, circuit breaker fracture areas, and other areas with weak insulation or complex structures.

[0038] Deploy multi-channel sensing units: A separate partial discharge sensing channel is installed in each selected area. Each channel includes: an ultra-high frequency (UHF) voltage detection module for capturing electromagnetic wave signals generated by partial discharge (usually in the 300MHz–3GHz frequency range); and a high-sensitivity ultrasonic receiving module for detecting mechanical vibration sound wave signals associated with discharge (usually in the 20kHz–200kHz frequency range).

[0039] Optimize channel spatial arrangement: Arrange sensing channels in a reverse interleaved manner, ensuring that signals from the same discharge source are received by multiple channels at different locations (for example, signal source A is detected simultaneously by channels 1, 3, and 5, while signal source B is detected by channels 2, 4, and 6). Consider signal interference from the GIS's internal structure (such as electromagnetic wave reflection and metal shielding attenuation), and adjust sensor angles and spacing to avoid signal distortion caused by overlap.

[0040] Establish a full-area coverage network: Through a multi-channel heterogeneous layout (i.e., using differentiated sensor combinations at different locations), ensure that there are no blind spots in the internal space of the GIS.

[0041] Verify the sensing network: Use a simulated discharge source to test the response consistency of each channel to ensure that the signal can be effectively captured by at least three channels. Collect the discharge signal of the gas-insulated cabinet in real time to obtain local voltage signal data.

[0042] Step S2: Time-sharingly superimposing the plurality of local voltage signal data to obtain an insulating cabinet superimposed voltage signal, performing fast Fourier transform on the insulating cabinet superimposed voltage signal to obtain an insulating cabinet voltage frequency domain amplitude spectrum, smoothing the insulating cabinet voltage frequency domain amplitude spectrum to obtain an insulating cabinet voltage frequency domain envelope; and performing energy distribution analysis on the insulating cabinet voltage frequency domain amplitude spectrum to obtain an insulating cabinet voltage energy distribution.

[0043] In a time window, the local voltage signal data has a sampling frequency of (e.g. 3GHz), the collected partial discharge signal is an analog voltage signal, which is converted into a discrete sequence after being sampled by an analog-to-digital converter (ADC). The local voltage signal data is V[n], where n=1,2,...,N, and N is the total number of samples in the time window.

[0044] Calculate the energy contribution of partial discharge data for each channel:

[0045] in, is the energy contribution value of the local voltage signal data of the i-th channel, n is the index of the n-th sampling point, and N is the total number of samples in the time window.

[0046] In the time window, the local voltage signal data is time-sharingly superimposed to obtain the insulation cabinet superimposed voltage signal:

[0047] in, Indicates the insulation cabinet superimposed voltage signal, Represents the total number of channels, i represents the index of the local voltage signal data of the i-th channel, is the energy contribution value of the local voltage signal data of the i-th channel, represents the local voltage signal data of the i-th channel, and n represents the index of the n-th sampling point.

[0048] Perform fast Fourier transform on the superimposed discharge signal to obtain a frequency domain complex sequence :

[0049] in, ] is the complex amplitude of the kth frequency point, represents the superimposed voltage signal of the insulation cabinet, j is a complex unit, k is the frequency index in the frequency domain, and N is the total number of samples in the time window.

[0050] Taking the modulus of the complex sequence in the frequency domain, we can obtain the frequency domain amplitude spectrum of the insulation cabinet voltage. |.

[0051] The frequency domain amplitude spectrum of the insulation cabinet voltage is curve fitted to obtain the frequency domain envelope of the insulation cabinet voltage:

[0052] in, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of ] is the complex amplitude of the kth frequency point, is the Hilbert transform operator, is the frequency corresponding to the kth frequency point.

[0053] By performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum, the insulation cabinet voltage energy distribution is obtained:

[0054] in, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, k is the frequency index in the frequency domain, N is the total number of samples in the time window, ] is the complex amplitude of the kth frequency point.

[0055] Step S3: Perform offline testing on the gas-insulated cabinet and construct a discharge signal category library, wherein the discharge signal category library includes: discharge category signal energy distribution and discharge category signal frequency domain envelope; dynamically adjust the frequency weight in the insulation cabinet voltage energy distribution, calculate the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, and obtain the insulation cabinet energy distribution difference value; calculate the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope through the dynamic time warping method, and obtain the insulation cabinet frequency domain similarity value.

[0056] Offline testing of the gas-insulated cabinet (GIC) was performed using an offline testing method, and a discharge signal library was constructed. Ultrasonic sensors, ultra-high frequency sensors, and other devices were used to collect the GIC's discharge signals offline. The acquisition process covered various operating conditions and typical discharge types to ensure signal integrity and representativeness. Environmental parameters and equipment operating status were also recorded. The collected raw signals were preprocessed, using bandpass filtering and denoising to remove environmental interference and invalid data, improving signal quality and laying the foundation for subsequent analysis. Next, feature extraction was performed. First, the normalized energy distribution of the discharge signal was calculated. This was done by converting the preprocessed signal to the frequency domain using a Fourier transform, determining the proportion of energy at each frequency point to the total energy, reflecting the energy distribution characteristics in the frequency domain. Second, the frequency domain envelope was extracted (by smoothing the frequency domain amplitude spectrum) to obtain the spectral profile characteristics and characterize the distribution trend of the signal's frequency components. The extracted features are then classified according to the discharge signal type (e.g., corona discharge, creeping discharge, internal discharge, etc.). The energy distribution and frequency domain envelope data of similar discharge signals are integrated to construct a discharge signal category library. The library clearly labels the characteristic parameter ranges and typical characteristic curves of each category to facilitate subsequent matching and comparison. Finally, the constructed category library is verified and optimized. Test signals of known discharge types are input into the library for matching tests. The energy distribution difference values ​​and frequency domain similarity values ​​of the insulation cabinets are calculated (e.g., using a dynamic time warping algorithm). The recognition accuracy of the category library is evaluated. For categories with large recognition errors, the feature extraction process is re-examined or sample data is supplemented until the category library meets practical application requirements. The constructed category library can be used for online monitoring of gas-insulated cabinet discharge signals and fault type identification. By acquiring signals in real time and comparing them with features in the library, the discharge type can be quickly determined.

[0057] By dynamically adjusting the frequency weight, the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value.

[0058] Calculate the average value of the energy distribution of the discharge category signals in the discharge signal category library:

[0059] in, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the total number of discharge category signal energy distributions, For the The index of the energy distribution of the discharge category signal, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point.

[0060] The weight of the energy ratio of each frequency point in the insulation cabinet voltage energy distribution is calculated by the average value of the discharge category signal energy distribution:

[0061] in, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the adjustment coefficient, >0, used to adjust the impact of the difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, is a constant term, >1.

[0062] It should be noted that dynamically adjusting the frequency weights in the insulation cabinet voltage energy distribution is intended to adaptively enhance sensitivity to key frequency characteristics, thereby more accurately capturing the differences between the real-time discharge signal and the discharge signal library. Different types of discharges (such as corona discharge, surface discharge, and internal discharge) have unique frequency-domain energy distributions. Furthermore, energy distribution can shift during signal propagation due to the influence of the GIS's internal structure (such as metal shielding and electromagnetic wave reflection) and dielectric properties (such as SF6 gas attenuation). Dynamically adjusting the frequency weights assigns higher weights to frequencies with significant differences based on the degree of deviation between the real-time signal and the discharge signal library, highlighting the role of these features in the matching process.

[0063] The reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value:

[0064] in, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, k is the frequency point index in the frequency domain, N is the total number of samples in the time window, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point.

[0065] It should be noted that the insulation cabinet energy distribution difference value quantifies the degree of deviation between the real-time collected insulation cabinet voltage energy distribution and the pre-established discharge category signal energy distribution. By dynamically adjusting frequency weights, this process adaptively highlights frequency points with significant differences, effectively capturing the unique frequency domain energy distribution characteristics of different discharge types (such as tip discharge and air gap breakdown). The smaller the difference value, the closer the match between the real-time signal and the energy characteristics of a particular discharge type, providing a key quantitative basis for accurately identifying discharge types. The trend of the difference value can also be used to determine the severity of discharge development during equipment operating status monitoring.

[0066] The cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope is calculated by the dynamic time warping method to obtain the insulation cabinet frequency domain similarity value:

[0067] in, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of For the The frequency envelope of the discharge signal of each category is The envelope amplitude, DTW() is the dynamic time warping function, is the adjustment coefficient, which is used to adjust the impact of the difference between the frequency domain envelope of the insulation cabinet voltage and the frequency domain envelope of the discharge category signal.

[0068] It should be noted that the dynamic time warping method calculates the cumulative distance between the frequency-domain envelope of the insulation cabinet voltage and the frequency-domain envelope of the discharge-classified signal, and obtains the insulation cabinet frequency-domain similarity value. This effectively addresses the time-axis or frequency-axis offset (e.g., phase difference, frequency broadening) caused by factors such as the propagation path and dielectric attenuation in the frequency-domain characteristic sequence. Traditional distance calculation methods struggle to handle such asynchronous features. However, the dynamic time warping method uses a dynamic programming algorithm to find the optimal time-warping path, allowing the two sequences to be locally stretched or compressed in the time or frequency dimension, making the cumulative distance calculation more accurate to the actual signal differences. The resulting similarity value quantitatively represents the degree of match between the real-time signal and the standard discharge-classified frequency-domain morphology, assisting in accurately determining the discharge type (e.g., corona discharge, creeping discharge, etc.). Furthermore, during long-term monitoring, the similarity value trend can be used to identify the evolution of early-stage fault characteristics, providing a key basis for accurately assessing the discharge status of gas-insulated cabinets.

[0069] Step S4: calculating a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtaining a discharge signal category according to the discharge signal category matching value.

[0070] By combining the energy distribution difference value of the insulation cabinet and the frequency domain similarity value of the insulation cabinet, the discharge signal category matching value is calculated:

[0071] in, The local discharge signal and the discharge signal category matching value of the discharge signal category, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal.

[0072] It should be noted that the insulation cabinet energy distribution difference value is inversely proportional to the discharge signal category matching value, while the insulation cabinet frequency domain similarity value is directly proportional to the discharge signal category matching value. Specifically, the insulation cabinet energy distribution difference value reflects the degree of deviation between the real-time insulation cabinet voltage energy distribution and the standard energy distribution in the discharge category library. The smaller the value, the closer the energy characteristics of the real-time signal are to a certain discharge category, and the easier it is to match the corresponding discharge category. Therefore, its contribution to the discharge signal category matching value is negative, that is, the smaller the difference value, the greater the matching value. The insulation cabinet frequency domain similarity value measures the similarity between the real-time signal frequency domain envelope and the standard frequency domain envelope using the dynamic time warping method. The larger the value, the higher the frequency domain morphology match between the two, and the more certain it is that the discharge belongs to a certain category. It contributes positively to the discharge signal category matching value, that is, the larger the similarity value, the greater the matching value. Through this inverse relationship, the two factors work together to accurately determine the discharge signal category.

[0073] By calculating the discharge signal category matching values ​​between the discharge signal and all discharge signal categories in the discharge category library, the discharge signal category (such as corona discharge, creeping discharge, and internal discharge) corresponding to the maximum value among the discharge signal category matching values ​​is selected as the discharge signal category of the partial discharge signal. For example, if the discharge signal category matching value between the partial discharge signal and the discharge signal category corresponding to "corona discharge" in the discharge category library is the largest, the discharge signal category of the partial discharge signal is "corona discharge."

[0074] Step S5: The internal dielectric absorption coefficient of the gas-insulated cabinet is collected and obtained. Based on the discharge signal type and the internal dielectric absorption coefficient of the gas-insulated cabinet, the local voltage signal data is positioned and analyzed by an energy gradient estimation method to obtain the discharge source location.

[0075] Extract the response amplitude of the local voltage signal data of each channel , i=1,2,3,... , represents the response amplitude of the local voltage signal data of the i-th channel, Indicates the total number of channels.

[0076] Core input parameters: Response amplitude: the response amplitude of the local voltage signal data received by the i-th channel (Unit: mV), reflects the distance between the signal source and the channel and the propagation attenuation. Response time delay: is the signal arrival time difference between the local voltage signal data of the i-th channel and the reference channel, which is caused by the reflection of the metal shell inside the GIS and the difference in path length.

[0077] Backward iterative energy gradient estimation method: Initial assumption: preset discharge source candidate points in the three-dimensional space of the gas-insulated cabinet .

[0078] Forward calculation: Calculate candidate points based on the gas-insulated cabinet structure Theoretical distance to the i-th channel ', Theoretical time difference , theoretical energy contribution value .

[0079] Structural parameters: Input the three-dimensional coordinates of other insulation cabinet shells, the position of internal metal parts, and the internal dielectric absorption coefficient (gas medium absorption coefficient), electromagnetic wave propagation speed , = , ' is the theoretical time difference between the local voltage signal data of the ith channel and the reference channel calculated based on the preset discharge source position, where 'Indicates the distance from the preset discharge source location to the reference channel.

[0080] It should be noted that, when a discharge event is detected, the reference channel is the channel that detects the discharge earliest among all channels as the reference channel.

[0081] Calculate the theoretical amplitude based on a preset discharge source location:

[0082] in, The theoretical amplitude of the i-th channel calculated based on the preset discharge source position, is the initial amplitude of the discharge source, The discharge source to channel is calculated based on the preset discharge source position distance, α is the geometric attenuation factor, is the Fresnel loss function.

[0083] Calculate the theoretical energy contribution based on the preset discharge source location:

[0084] in, is the theoretical energy contribution value of the ith channel calculated based on the preset discharge source position, i is the index of the ith channel, Indicates the total number of channels, ' indicates candidate points The theoretical distance to the i-th channel, α is the geometric attenuation factor.

[0085] It should be noted that, for the discharge signal category, when the discharge signal category is tip discharge, the signal propagation is mainly electromagnetic waves, with a fast attenuation speed and a larger geometric attenuation factor; when the discharge signal category is surface creepage, the signal propagation is mainly ultrasonic waves, with stable propagation and a larger dielectric absorption coefficient; when the discharge signal category is air gap breakdown, the signal propagation is a mixture of electromagnetic waves and ultrasonic waves, the path is complex, and the reflection step length has a higher loss function.

[0086] The objective function of the energy gradient estimation method is:

[0087] Among them, F is the objective function of the energy gradient estimation method, is the response amplitude of the local voltage signal data of the i-th channel, ' is the theoretical amplitude calculated based on the preset discharge source position, is the actual signal arrival time difference between the ith channel and the reference channel, ' is the theoretical time difference between the i-th channel and the reference channel calculated based on the preset discharge source position, is the energy contribution of the local voltage signal data of the i-th channel, is the theoretical energy contribution value of the ith channel calculated based on the preset discharge source position, i is the index of the ith channel, Indicates the total number of channels.

[0088] Adjust the preset discharge source candidate points by gradient descent method The coordinates of the energy gradient estimation method are Converges to the minimum value, then That is the discharge source location.

[0089] Step S6: Collect the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source position within a preset time, and perform trend change analysis to obtain the partial discharge signal evolution value; divide the discharge risk level according to the partial discharge signal evolution value to achieve online monitoring of the gas insulated cabinet.

[0090] Collect the insulation cabinet superimposed voltage signal and insulation cabinet voltage energy distribution at the power supply location within the preset time, divide the preset time into MS time windows, and perform trend change analysis.

[0091] Calculate the difference between the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution in each adjacent time window:

[0092] in, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The index of the time window, is the envelope amplitude difference adjustment coefficient, which is used to adjust the influence of envelope difference on the difference value of partial discharge signal. is the energy ratio difference adjustment coefficient, which is used to adjust the influence of the energy ratio difference adjustment coefficient on the partial discharge signal difference value. To prevent division by zero, the value is .

[0093] It should be noted that the envelope amplitude difference adjustment coefficient and energy ratio difference adjustment coefficient , can be determined according to the discharge signal category. For example, when the discharge signal category is "corona discharge", the change of the frequency domain envelope amplitude is more sensitive. It can be 0.7, It can be 0.3; when the discharge signal type is "surface discharge", the energy distribution offset is more critical, then It can be 0.4, It can be 0.6; when the discharge signal category is "internal discharge", it is necessary to balance the characteristics of the two. It can be 0.5, 0.5 can be taken.

[0094] The partial discharge signal evolution value of the partial discharge signal within a preset time is calculated using the partial discharge signal difference value:

[0095] Among them, XB is the evolution value of the partial discharge signal, MS is the total number of time windows, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the length of the unit time window.

[0096] The discharge risk level is divided by the evolution value of the partial discharge signal, and the online monitoring of the gas insulated cabinet is realized. , then the risk level is level 1, which is normal and requires routine inspection; <XB< , then the risk level is Level 2 Caution, and the monitoring frequency needs to be increased; <XB< , then the risk level is Level 3 alert, requiring dedicated monitoring; , then the risk level is Level 4 emergency and requires immediate maintenance.

[0097] This paper proposes a method and device for online monitoring of gas-insulated cabinets. By constructing a discharge signal category library offline, local voltage signals are collected in real time and processed through time-sharing superposition and fast Fourier transform. The frequency domain envelope and energy distribution characteristics are obtained. Then, the difference and similarity values ​​are calculated through algorithms such as dynamic adjustment of frequency weights and dynamic time warping. Discharge signal category matching and discharge source location are achieved. Finally, the discharge risk level is divided through trend analysis to complete online monitoring.

[0098] Dynamically adjusting frequency weights and calculating the energy distribution difference values ​​for the insulation cabinets is crucial. This step adaptively highlights frequency points where the voltage signal significantly differs from the characteristics in the discharge category library, allowing the energy distribution difference values ​​for the insulation cabinets to more accurately reflect the degree of deviation between the real-time signal and the standard category. This process enhances sensitivity to discharge signal characteristics and effectively identifies the unique energy distribution characteristics of different discharge types (such as corona and creeping discharge). This provides a key quantitative basis for subsequent discharge category matching, improving the accuracy and reliability of fault identification.

[0099] The dynamic time warping method calculates the cumulative distance between the frequency-domain envelope of the insulation cabinet voltage and the frequency-domain envelope of the discharge classification signal, generating a frequency-domain similarity value for the insulation cabinet. Its core significance lies in resolving the nonlinear alignment problem of frequency-domain feature sequences, thereby accurately capturing the shape similarity of different discharge signals. Because discharge signals within gas-insulated cabinets are affected by factors such as the propagation path and dielectric attenuation, their frequency-domain envelopes may exhibit offsets in time or frequency (e.g., phase differences and frequency broadening). Traditional Euclidean distance cannot effectively handle these non-synchronous features. However, a dynamic programming algorithm is used to find the optimal time-warping path, allowing the two frequency-domain envelope sequences to be locally stretched or compressed in time or frequency, making the cumulative distance calculation more accurate for actual signal differences. The resulting frequency-domain similarity value for the insulation cabinet quantifies the degree of morphological match between the real-time signal and the standard discharge classification, thereby assisting in determining the discharge type (e.g., corona discharge, creeping discharge), providing a key basis for accurate discharge signal classification. Furthermore, the similarity value trend can be used to identify the evolution of early-stage fault characteristics during long-term equipment monitoring.

[0100] The significance of collecting time series data and analyzing the evolution values ​​of partial discharge signals. By collecting frequency domain envelope and energy distribution data over a preset time period and analyzing trends, the evolution patterns of partial discharge signals, such as the rate of change and stability of characteristic parameters, can be determined. The evolution values ​​of partial discharge signals dynamically reflect the development of discharges, such as the gradual transition from normal to alert status, providing quantitative support for risk classification in the time dimension. This mechanism can provide early warning of potential failures, helping operations and maintenance personnel develop differentiated maintenance strategies (such as routine inspections and emergency repairs) based on risk levels. This enables proactive management of the status of gas-insulated cabinets and reduces the risk of equipment failure.

[0101] An online monitoring device for a gas-insulated cabinet, comprising: Signal acquisition module: A plurality of partial discharge sensing channels are embedded in the key electric field non-uniform area of ​​the gas-insulated cabinet in an inverse interleaving manner. The discharge signals of the gas-insulated cabinet are collected through the partial discharge sensing channels to obtain a plurality of local voltage signal data of the gas-insulated cabinet; The insulation cabinet voltage energy distribution generation module is configured to perform time-sharing superposition on the plurality of local voltage signal data to obtain an insulation cabinet superposition voltage signal, perform fast Fourier transform on the insulation cabinet superposition voltage signal to obtain an insulation cabinet voltage frequency domain amplitude spectrum, smooth the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage frequency domain envelope, and perform energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage energy distribution. Insulation cabinet frequency domain similarity value generation module: By performing offline testing on the gas-insulated cabinet and constructing a discharge signal category library, the discharge signal category library includes: discharge category signal energy distribution and discharge category signal frequency domain envelope; by dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, the reconstruction difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value; using the dynamic time warping method to calculate the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope to obtain the insulation cabinet frequency domain similarity value; A discharge signal category generation module is configured to calculate a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtain a discharge signal category based on the discharge signal category matching value; Discharge source location determination module: collects and obtains the internal dielectric absorption coefficient of the gas-insulated cabinet, and based on the discharge signal type and the internal dielectric absorption coefficient of the gas-insulated cabinet, performs location analysis on the local voltage signal data using an energy gradient estimation method to determine the discharge source location; Monitoring module: collects the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source location within a preset time, and performs trend change analysis to obtain the partial discharge signal evolution value; divides the discharge risk level according to the partial discharge signal evolution value to realize online monitoring of the gas insulated cabinet.

[0102] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprise," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations. The phrase "includes an element defined by..." does not exclude the presence of other identical elements in the process, method, article, or device that includes the element.

[0103] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for online monitoring of a gas-insulated cabinet, characterized by: The following steps are involved: Step S1: embedding a plurality of partial discharge sensing channels in a key electric field non-uniform area in a gas-insulated cabinet in a reverse interleaving manner, collecting discharge signals of the gas-insulated cabinet through the partial discharge sensing channels, and obtaining a plurality of local voltage signal data of the gas-insulated cabinet; Step S2: performing time-sharing superposition on the plurality of local voltage signal data to obtain an insulation cabinet superimposed voltage signal, performing fast Fourier transform on the insulation cabinet superimposed voltage signal to obtain an insulation cabinet voltage frequency domain amplitude spectrum, smoothing the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage frequency domain envelope; and performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage energy distribution; Step S3: Performing offline testing on the gas-insulated cabinet and constructing a discharge signal category library, the discharge signal category library including: discharge category signal energy distribution and discharge category signal frequency domain envelope; dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, calculating the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, and obtaining the insulation cabinet energy distribution difference value; calculating the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope using a dynamic time warping method, and obtaining the insulation cabinet frequency domain similarity value; Step S4: calculating a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtaining a discharge signal category according to the discharge signal category matching value; Step S5: collecting and obtaining the internal dielectric absorption coefficient of the gas-insulated cabinet, and based on the type of the discharge signal and the internal dielectric absorption coefficient of the gas-insulated cabinet, performing location analysis on the local voltage signal data using an energy gradient estimation method to obtain the discharge source location; Step S6: Collect the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source position within a preset time, and perform trend change analysis to obtain the partial discharge signal evolution value; divide the discharge risk level according to the partial discharge signal evolution value to achieve online monitoring of the gas insulated cabinet.

2. The method for online monitoring of a gas-insulated cabinet according to claim 1, characterized in that: The time-sharing superposition of the plurality of local voltage signal data to obtain the insulation cabinet superposition voltage signal comprises the following specific steps: Calculate the energy contribution of partial discharge data for each channel: ； in, is the energy contribution value of the local voltage signal data of the i-th channel, n is the index of the n-th sampling point, and N is the total number of samples in the time window; In the time window, the local voltage signal data is time-sharingly superimposed to obtain the insulation cabinet superimposed voltage signal: ； in, Indicates the insulation cabinet superimposed voltage signal, Represents the total number of channels, i represents the index of the local voltage signal data of the i-th channel, is the energy contribution value of the local voltage signal data of the i-th channel, represents the local voltage signal data of the i-th channel, and n represents the index of the n-th sampling point.

3. The method for online monitoring of a gas-insulated cabinet according to claim 2, characterized in that: The method comprises the following steps: performing fast Fourier transform on the superimposed voltage signal of the insulation cabinet to obtain a frequency domain amplitude spectrum of the insulation cabinet voltage; and performing smoothing on the frequency domain amplitude spectrum of the insulation cabinet voltage to obtain a frequency domain envelope of the insulation cabinet voltage. Perform fast Fourier transform on the superimposed discharge signal to obtain a frequency domain complex sequence : ； in, ] is the complex amplitude of the kth frequency point, represents the superimposed voltage signal of the insulation cabinet, j is a complex unit, k is the frequency index in the frequency domain, and N is the total number of samples in the time window; Taking the modulus of the complex sequence in the frequency domain, we can obtain the frequency domain amplitude spectrum of the insulation cabinet voltage. |; The frequency domain amplitude spectrum of the insulation cabinet voltage is curve fitted to obtain the frequency domain envelope of the insulation cabinet voltage: ； in, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of ] is the complex amplitude of the kth frequency point, is the Hilbert transform operator, is the frequency corresponding to the kth frequency point.

4. The method for online monitoring of a gas-insulated cabinet according to claim 3, characterized in that: The method of performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain the insulation cabinet voltage energy distribution includes the following steps: By performing energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum, the insulation cabinet voltage energy distribution is obtained: ； in, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, k is the frequency index in the frequency domain, N is the total number of samples in the time window, ] is the complex amplitude of the kth frequency point.

5. The method for online monitoring of a gas-insulated cabinet according to claim 4, characterized in that: The method of dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution and calculating the reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution to obtain the insulation cabinet energy distribution difference value includes the following steps: Calculate the average value of the energy distribution of the discharge category signals in the discharge signal category library: ； in, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the total number of discharge category signal energy distributions, For the The index of the energy distribution of the discharge category signal, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point; The weight of the energy ratio of each frequency point in the insulation cabinet voltage energy distribution is calculated by the average value of the discharge category signal energy distribution: ； in, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, is the average energy ratio of the discharge category signal energy distribution at the kth frequency point, is the adjustment coefficient, >0, used to adjust the impact of the difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution, is a constant term, >1; The reconstructed difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value: ； in, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, k is the frequency point index in the frequency domain, N is the total number of samples in the time window, is the energy proportion weight of the insulation cabinet voltage energy distribution at the kth frequency point, is the energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point, Indicates the The energy ratio of the discharge category signal energy distribution at the kth frequency point.

6. The method for online monitoring of a gas-insulated cabinet according to claim 5, characterized in that: The method of calculating the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope by the dynamic time warping method to obtain the insulation cabinet frequency domain similarity value includes the following specific steps: The cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope is calculated by the dynamic time warping method to obtain the insulation cabinet frequency domain similarity value: ； in, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal, is the frequency domain envelope of the insulation cabinet voltage at frequency The envelope amplitude of For the The frequency envelope of the discharge signal of each category is The envelope amplitude, DTW() is the dynamic time warping function, is the adjustment coefficient, which is used to adjust the impact of the difference between the frequency domain envelope of the insulation cabinet voltage and the frequency domain envelope of the discharge category signal.

7. The method for online monitoring of a gas-insulated cabinet according to claim 6, characterized in that: The method of calculating the discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value includes the following specific steps: By combining the energy distribution difference value of the insulation cabinet and the frequency domain similarity value of the insulation cabinet, the discharge signal category matching value is calculated: ； in, The local discharge signal and the discharge signal category matching value of the discharge signal category, The voltage energy distribution of the insulation cabinet and the The energy distribution difference value of the insulation cabinet of the discharge category signal energy distribution, The frequency domain envelope of the insulation cabinet voltage and the The frequency domain similarity value of the insulation cabinet of the frequency domain envelope of the discharge category signal.

8. The method for online monitoring of a gas-insulated cabinet according to claim 7, characterized in that: The method of performing location analysis on the local voltage signal data by using an energy gradient estimation method to obtain the discharge source location includes the following specific steps: The objective function of the energy gradient estimation method is: ； Among them, F is the objective function of the energy gradient estimation method, is the response amplitude of the local voltage signal data of the i-th channel, ' is the theoretical amplitude calculated based on the preset discharge source position, is the actual signal arrival time difference between the ith channel and the reference channel, ' is the theoretical time difference between the i-th channel and the reference channel calculated based on the preset discharge source position, is the energy contribution of the local voltage signal data of the i-th channel, is the theoretical energy contribution value of the ith channel calculated based on the preset discharge source position, i is the index of the ith channel, Indicates the total number of channels; Adjust the preset discharge source candidate points by gradient descent method The coordinates of the energy gradient estimation method are Converges to the minimum value, then That is the discharge source location.

9. The method for online monitoring of a gas-insulated cabinet according to claim 8, characterized in that: The collecting of the insulation cabinet voltage frequency domain envelope and the insulation cabinet voltage energy distribution at the discharge power source position within a preset time, and performing trend change analysis to obtain the partial discharge signal evolution value includes the following specific steps: Calculate the difference between the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution in each adjacent time window: ； in, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The frequency domain envelope of the insulation cabinet voltage in a time window is The envelope amplitude of Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The energy ratio of the insulation cabinet voltage energy distribution at the kth frequency point in a time window is: Indicates the The index of the time window, is the envelope amplitude difference adjustment coefficient, which is used to adjust the influence of envelope difference on the difference value of partial discharge signal. is the energy ratio difference adjustment coefficient, which is used to adjust the influence of the energy ratio difference adjustment coefficient on the partial discharge signal difference value. To prevent division by zero, the value is ; The partial discharge signal evolution value of the partial discharge signal within a preset time is calculated using the partial discharge signal difference value: ； Among them, XB is the evolution value of the partial discharge signal, MS is the total number of time windows, represents the difference value of the partial discharge signal in the ms-th time window, Indicates the length of the unit time window.

10. A gas-insulated cabinet online monitoring device, characterized in that: include: Signal acquisition module: A plurality of partial discharge sensing channels are embedded in the key electric field non-uniform area of ​​the gas-insulated cabinet in an inverse interleaving manner. The discharge signals of the gas-insulated cabinet are collected through the partial discharge sensing channels to obtain a plurality of local voltage signal data of the gas-insulated cabinet; The insulation cabinet voltage energy distribution generation module is configured to perform time-sharing superposition on the plurality of local voltage signal data to obtain an insulation cabinet superposition voltage signal, perform fast Fourier transform on the insulation cabinet superposition voltage signal to obtain an insulation cabinet voltage frequency domain amplitude spectrum, smooth the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage frequency domain envelope, and perform energy distribution analysis on the insulation cabinet voltage frequency domain amplitude spectrum to obtain an insulation cabinet voltage energy distribution. Insulation cabinet frequency domain similarity value generation module: by performing offline testing on the gas insulated cabinet and building a discharge signal category library, the discharge signal category library includes: discharge category signal energy distribution, discharge category signal frequency domain envelope; By dynamically adjusting the frequency weight in the insulation cabinet voltage energy distribution, the reconstruction difference between the insulation cabinet voltage energy distribution and the discharge category signal energy distribution is calculated to obtain the insulation cabinet energy distribution difference value; the cumulative distance between the insulation cabinet voltage frequency domain envelope and the discharge category signal frequency domain envelope is calculated using the dynamic time warping method to obtain the insulation cabinet frequency domain similarity value; A discharge signal category generation module is configured to calculate a discharge signal category matching value by combining the insulation cabinet energy distribution difference value and the insulation cabinet frequency domain similarity value, and obtain a discharge signal category based on the discharge signal category matching value; Discharge source location determination module: collects and obtains the internal dielectric absorption coefficient of the gas-insulated cabinet, and based on the discharge signal type and the internal dielectric absorption coefficient of the gas-insulated cabinet, performs location analysis on the local voltage signal data using an energy gradient estimation method to determine the discharge source location; Monitoring module: collects the insulation cabinet voltage frequency domain envelope and insulation cabinet voltage energy distribution at the discharge power source location within a preset time, and performs trend change analysis to obtain the partial discharge signal evolution value; divides the discharge risk level according to the partial discharge signal evolution value to realize online monitoring of the gas insulated cabinet.

Images