Fault identification method and device for power transmission and transformation equipment of wind power plant and related system

Abstract

The invention discloses a fault identification method and device for power transmission and transformation equipment of a wind power plant and a related system. The method comprises the following steps: acquiring working signals, temperature data and gas state data of a plurality of sensors in a target wind power plant within a preset time period; the working signal comprises a high-frequency current signal and a voltage signal; performing decoupling processing on the high-frequency current signal and the voltage signal to obtain decoupling current data and decoupling voltage data; fusing the decoupling current data, the decoupling voltage data, the temperature data and the gas state data to obtain multi-dimensional fusion data; inputting the multi-dimensional fusion data into a preset fault recognition model to obtain fault recognition information; and the fault identification information is sent to the cloud server. It can be seen that the current sensor, the voltage sensor, the temperature sensor and the gas sensor are used for collecting multi-dimensional monitoring data for modeling, centralized intelligent diagnosis is carried out on multi-source data, and then the fault recognition accuracy and the fault recognition efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of power technology, and in particular to a fault identification method, device and related system for wind farm transmission and transformation equipment. Background Technology

[0002] Wind farm transmission and transformation equipment is a critical link in the transmission of wind power, and its operational stability directly determines the reliability of power supply from the wind farm. For monitoring transmission and transformation equipment in offshore wind farms, existing methods focus on single-dimensional parameters, neglecting the coupling effect of multi-dimensional monitoring data. When using single-dimensional monitoring data for fault identification, the accuracy of fault identification is low. Furthermore, existing data processing methods employ a data acquisition-cloud storage-data processing approach. When identifying faults, data is sent to the cloud for processing and analysis. This process is significantly affected by network latency, leading to substantial delays in fault identification and impacting the efficiency of fault identification for offshore transmission and transformation equipment.

[0003] Therefore, improving the accuracy and efficiency of fault identification in wind farm transmission and transformation equipment is an urgent issue that needs to be addressed. Summary of the Invention

[0004] This application provides a fault identification method, device, and related system for wind farm transmission and transformation equipment. It collects multi-dimensional monitoring data through current, voltage, temperature, and gas sensors for modeling, and adopts a centralized intelligent diagnosis of multi-source data through data acquisition, edge computing, and cloud centralized management, thereby improving the accuracy and efficiency of fault identification.

[0005] In a first aspect, embodiments of this application provide a fault identification method for wind farm transmission and transformation equipment, applied to an edge computing module of a transmission and transformation equipment management system. The transmission and transformation equipment management system includes: wind farm transmission and transformation equipment, multiple sensors installed on the wind farm transmission and transformation equipment, the edge computing module communicatively connected to the multiple sensors, and a cloud server communicatively connected to the edge computing module. The method includes: Acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals; The high-frequency current signal and the voltage signal are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, the decoupled voltage data, the temperature data, and the gas state data are fused to obtain multidimensional fused data; The multidimensional fused data is input into a preset fault identification model to obtain fault identification information; and the fault identification information is sent to the cloud server.

[0006] Secondly, this application provides a fault identification device for wind farm transmission and transformation equipment, applied to an edge computing module of a transmission and transformation equipment management system. The transmission and transformation equipment management system includes: wind farm transmission and transformation equipment, multiple sensors installed on the wind farm transmission and transformation equipment, an edge computing module communicatively connected to the multiple sensors, and a cloud server communicatively connected to the edge computing module. The device includes: The intelligent sensing unit is used to acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals. An edge computing unit is used to decouple the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data; and to fuse the decoupled current data, the decoupled voltage data, the temperature data and the gas state data to obtain multi-dimensional fused data. An edge execution unit is used to input the multi-dimensional fused data into a preset fault identification model to obtain fault identification information, and then send the fault identification information to the cloud server.

[0007] Thirdly, embodiments of this application provide an electronic device, including: a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for performing the steps in the first aspect of embodiments of this application.

[0008] Fourthly, embodiments of this application provide a fault identification system for wind farm transmission and transformation equipment, wherein the fault identification system for wind farm transmission and transformation equipment performs some or all of the steps described in the first aspect of embodiments of this application.

[0009] Fifthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of embodiments of this application. The computer program product may be a software installation package.

[0010] It can be seen that the embodiments of this application have the following beneficial effects: By implementing the embodiments of this application, operating signals, temperature data, and gas state data from multiple sensors in a target wind farm within a preset time period are acquired. The operating signals include high-frequency current signals and voltage signals. The high-frequency current signals and voltage signals are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, decoupled voltage data, temperature data, and gas state data are fused to obtain multi-dimensional fused data. The multi-dimensional fused data is input into a preset fault identification model to obtain fault identification information. The fault identification information is then sent to the cloud server. It is evident that by collecting multi-dimensional monitoring data from current, voltage, temperature, and gas sensors for modeling, and employing data acquisition, edge computing, and centralized cloud management for centralized intelligent diagnosis of multi-source data, the accuracy and efficiency of fault identification are improved. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the background art, the accompanying drawings used in the embodiments of this application or the background art will be described below.

[0012] Figure 1 This is a flowchart illustrating a fault identification method for wind farm transmission and transformation equipment provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a real-time monitoring system for an offshore wind farm provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a high-frequency current sensor provided in an embodiment of this application; Figure 4 This is a schematic diagram of an integration circuit for a transient ground voltage sensor provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an SF6 gas real-time monitoring system provided in an embodiment of this application; Figure 6 This is a schematic diagram of the architecture of a fault identification system for wind farm transmission and transformation equipment provided in an embodiment of this application; Figure 7 This is a functional module block diagram of a fault identification device for wind farm power transmission and transformation equipment provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0014] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0015] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0016] The following describes the relevant content, concepts, meanings, technical issues, technical solutions, and beneficial effects involved in the embodiments of this application.

[0017] Please see Figure 1 , Figure 1 This is a flowchart illustrating a fault identification method for wind farm transmission and transformation equipment provided in an embodiment of this application. The method is applied to the edge computing module of a transmission and transformation equipment management system. The transmission and transformation equipment management system includes: wind farm transmission and transformation equipment; multiple sensors installed on the wind farm transmission and transformation equipment; the edge computing module communicatively connected to the multiple sensors; and a cloud server communicatively connected to the edge computing module. The method includes, but is not limited to, the following steps: S101. Acquire the working signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the working signals include: high-frequency current signals and voltage signals.

[0018] In this embodiment, the preset time period refers to a fixed data acquisition cycle pre-set based on the operation and maintenance needs, equipment operating characteristics, and historical fault data of the wind farm's transmission and transformation equipment. Its duration can be configured in conjunction with equipment type (e.g., switchgear, transformer, surge arrester) and monitoring parameters (e.g., partial discharge signal, temperature, gas state). For example, the preset time period can be set to 15 minutes to ensure that details of equipment status changes are captured and to avoid data redundancy. The target wind farm refers to the specific wind farm for which fault identification of the transmission and transformation equipment is to be performed in this embodiment. This wind farm may include offshore or onshore wind farms, and is not limited here. The communication connection refers to the data interaction link established between the edge computing module and multiple sensors, and between the edge computing module and the cloud server, based on a preset communication protocol and transmission medium. The edge computing module and sensors can communicate via RS485 bus, Ethernet, or wireless radio frequency, while the edge computing module and the cloud server can communicate via fiber optic transmission to ensure data transmission stability, real-time performance, and anti-interference capabilities.

[0019] In a specific embodiment, firstly, the edge computing module of the power transmission and transformation equipment management system sends data acquisition commands to multiple sensors installed on the power transmission and transformation equipment of the target wind farm according to a preset time period. Upon receiving the commands, the multiple sensors collect the corresponding equipment's operating signals, temperature data, and gas state data. Specifically, high-frequency current sensors can collect high-frequency current signals from power transmission and transformation equipment such as cable grounding wires and circuit breaker contact grounding wires; transient ground voltage sensors or ultra-high frequency sensors can collect voltage signals from the equipment; UHF RFID passive wireless temperature sensors can collect temperature data from equipment such as circuit breaker contacts and cable joints; and SF6 gas sensors can collect gas state data from the GIS equipment's gas chambers, such as pressure, dew point, and trace moisture content. Next, the multiple sensors transmit the collected data to the edge computing module in real time through a preset communication connection method. The edge computing module performs preliminary reception and storage of the received data, completing the acquisition of multi-dimensional data within the preset time period. Faults in wind farm transmission and transformation equipment are often caused by a combination of factors. For example, insulation aging may be accompanied by excessive partial discharge and abnormal temperature. Acquiring only a single type of data may not fully reflect the operating status of the equipment. Therefore, it is necessary to acquire three types of multi-dimensional data simultaneously: working signals, temperature data, and gas status data. Among them, working signals can reflect the electrical insulation status of the equipment, temperature data can reflect the heating status of the equipment, and gas status data can reflect the sealing and insulation status of SF6 equipment.

[0020] It is evident that by acquiring multi-dimensional operational data of the target wind farm's transmission and transformation equipment on a regular and stable basis, a comprehensive and continuous data source can be provided for subsequent edge computing modules to perform data preprocessing, multi-dimensional fusion analysis, and fault identification, thus laying the foundation for improving the accuracy and efficiency of fault identification of wind farm transmission and transformation equipment.

[0021] For easier understanding, please refer to Figure 2 , Figure 2 This is a schematic diagram of the structure of a real-time monitoring system for an offshore wind farm provided in an embodiment of this application. As can be seen, the real-time monitoring system 200 for an offshore wind farm includes a cloud platform, a local monitoring backend at the wind farm site, a station-based wireless network system (802.1x certified), and multiple distributed monitoring units and wireless communication units that communicate via on-site WiFi. Through the real-time monitoring system 200, multi-dimensional status monitoring and data interaction of the wind farm's transmission and transformation equipment can be achieved, constructing a three-level intelligent sensing and analysis architecture of "cloud—station—equipment".

[0022] Specifically, the cloud platform is used for centralized management and analysis of offshore wind farm operation data, data storage, fault identification, early warning generation, and visualization. The cloud platform is connected to the on-site monitoring backend at the wind farm via a data link, receiving multi-source monitoring data from the edge and performing data modeling and intelligent diagnostics in the cloud. The on-site monitoring backend is located in the control room of the offshore booster station or wind turbine substation unit, primarily for data acquisition, preliminary processing, and status uploading. This backend system is connected to the lower-level station wireless network system (802.1x authentication), ensuring data transmission security and reliability through a local area network authentication mechanism. The station wireless network system serves as a communication hub, establishing an intra-site WiFi communication channel to achieve wireless interconnection between multiple sensor units, communication modules, and the monitoring host, thereby constructing a local monitoring network for the offshore wind farm. Below the station's wireless network system, multiple monitoring and communication modules with different functions are installed, interacting with corresponding sensing devices via coaxial cable, RS485, or Ethernet. Two distributed monitoring hosts are also installed, each connected to a corresponding HFCT (High-Frequency Current Sensor) via coaxial cable, for real-time acquisition of partial discharge signals, enabling monitoring of the discharge status of cables and circuit breaker contacts. The HFCT sensor can detect MHz-level high-frequency discharge pulse signals, reflecting the characteristics of partial insulation breakdown and providing raw input for edge-end partial discharge analysis algorithms. A wireless communication unit, connected via RS485 or Ethernet, is connected to the core grounding IED (Intelligent Electronic Device) to monitor changes in the transformer core grounding current. When multiple grounding points or abnormal grounding circulation current occur in the core, the IED sends real-time current parameters to the wireless communication unit, which then uploads them to the monitoring backend via the station's WiFi, providing online early warning of transformer grounding anomalies. Furthermore, the wireless communication unit is connected to a surge arrester monitoring unit, which acquires leakage current information from the surge arrester via an RS485 interface. By analyzing the changing trend of the resistive current component in the leakage current, it is possible to determine whether the surge arrester has moisture or aging problems, thereby achieving dynamic assessment of the surge arrester's health status. Simultaneously, an RFID reader and its subordinate RFID temperature sensing unit are also installed. This part constitutes the passive wireless temperature monitoring module of the system. The reader transmits a specific frequency radio frequency signal to the temperature sensing unit to remotely sense and read the temperature of key components such as circuit breaker contacts and cable joints. The RFID temperature sensing unit does not require battery power and can operate stably for a long time in the high humidity and high salt spray environment at sea, significantly improving the system's environmental adaptability and lifespan. The read temperature data is uploaded to the wireless communication unit via RS485 and then transmitted to the monitoring backend via WiFi. On the other hand, the real-time monitoring system 200 for offshore wind farms also includes a connection structure for an SF6 gas monitoring unit. This part includes an SF6 gas sensor and an SF6 temperature sensor, both of which are connected to the wireless communication unit via an RS485 interface.SF6 gas sensors are used to monitor gas pressure, dew point, and trace moisture content in the gas chambers of GIS equipment or circuit breakers to determine gas sealing performance and insulation status. SF6 temperature sensors record gas temperature in real time, providing auxiliary parameters for pressure compensation and condition analysis. These data are uploaded to the on-site monitoring backend at the site via a wireless communication unit, and further analyzed and recorded via a cloud platform to achieve continuous monitoring and safety assessment of the gas system status.

[0023] visible, Figure 2 The offshore wind farm real-time monitoring system 200 shown uses a multi-layer distributed structure to achieve unified acquisition and hierarchical processing of multi-source data. The lower-layer sensor unit is responsible for sensing raw signals such as partial discharge, current, voltage, temperature and gas state; the middle-layer wireless communication unit and distributed monitoring host are responsible for data aggregation and initial decoupling; the upper-layer on-site monitoring backend performs edge computing and anomaly identification; and the cloud platform completes comprehensive diagnosis and intelligent decision-making at the entire farm level.

[0024] S102. Decouple the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data.

[0025] In this embodiment, decoupling processing refers to addressing the characteristics of high-frequency current and voltage signals superimposed and interfering in the operating environment of wind farm transmission and transformation equipment. Signal processing techniques are used to separate the coupled components of the two signals, eliminating cross-interference and ensuring that the decoupled current data reflects only the high-frequency current characteristics related to partial discharge of the equipment. The decoupled voltage data reflects only the true state of the equipment's transient ground voltage or ultra-high frequency voltage, ensuring that both types of signal data can independently and accurately characterize the corresponding operating parameters of the equipment.

[0026] In a specific embodiment, the edge computing module can extract features from high-frequency current and voltage signals to identify interference components caused by electromagnetic coupling in both types of signals. Specifically, interference components in the high-frequency current signal mainly originate from electromagnetic radiation of the voltage signal, while interference components in the voltage signal mainly originate from conduction coupling of the high-frequency current signal. Next, based on the extracted interference features, adaptive filtering is used to remove voltage coupling interference from the high-frequency current signal, resulting in decoupled current data containing only high-frequency partial discharge current information of the device. Simultaneously, current coupling interference is removed from the voltage signal, resulting in decoupled voltage data containing only transient ground voltage or ultra-high frequency voltage information of the device. By decoupling the high-frequency current and voltage signals, mutual coupling interference between them can be effectively eliminated, resulting in decoupled data that accurately reflects the characteristics of the partial discharge current and voltage state of the device, thus avoiding signal distortion caused by coupling interference.

[0027] Optionally, the above step of decoupling the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data specifically includes the following steps: A201. Perform spectrum analysis on the high-frequency current signal to obtain the current spectrum analysis results; A202. Determine the main frequency band and current noise frequency band based on the current spectrum analysis results; A203. Determine the decoupling filter function corresponding to the high-frequency current signal based on the main frequency band of the current; A204. Generate the decoupled current data based on the decoupling filter function, the main current frequency band, and the current noise frequency band; A205. The voltage signal is processed using a preset pulse processing algorithm to obtain a voltage pulse signal; A206. Decouple the voltage pulse signal from noise according to the preset bandwidth parameters to obtain a decoupled voltage signal; A207. Integrate the decoupling voltage signal to obtain the decoupling voltage energy value; A208. Determine the decoupling voltage data based on the decoupling voltage energy value.

[0028] In this embodiment, spectrum analysis refers to converting high-frequency current signals in the time domain to the frequency domain using signal processing techniques such as Fourier transform, obtaining the amplitude distribution characteristics of the signal in different frequency ranges, and thus clarifying the frequency ranges of effective and interference components in the signal. The main current frequency band refers to the primary frequency range in the high-frequency current signal that reflects the partial discharge characteristics of the equipment. The signal amplitude in this range is significantly higher than in other frequency ranges, and it is the core frequency range characterizing the partial discharge state of the equipment. The current noise frequency band refers to the frequency range in the high-frequency current signal generated by non-partial discharge factors such as electromagnetic interference and equipment background noise. The signal in this range does not reflect the actual operating state of the equipment and needs to be removed in subsequent processing. The decoupling filter function is a signal filtering model constructed based on the characteristics of the main current frequency band and the current noise frequency band. It can selectively retain the effective signal in the main current frequency band and filter out the interference signal in the current noise frequency band. The preset pulse processing algorithm refers to an algorithm pre-stored in the edge computing module, used to extract pulse features from continuous voltage signals. It can identify transient pulse components in the voltage signal and eliminate stationary interference. The preset bandwidth parameter refers to the filtering bandwidth pre-set according to the standard frequency range of the voltage signal. This parameter matches the frequency range of the effective components of the voltage signal, enabling precise filtering of noise in the voltage signal. The decoupled voltage energy value refers to the energy quantization value obtained by integrating the decoupled voltage pulse signal. It can intuitively reflect the intensity characteristics of the voltage signal and provide a quantitative basis for subsequent fault diagnosis.

[0029] In a specific embodiment, the edge computing module can call a spectrum analysis tool to perform frequency domain conversion on the acquired high-frequency current signal, obtaining the amplitude distribution of the signal at different frequencies, i.e., the current spectrum analysis result. Based on the current spectrum analysis result, the frequency range with the highest amplitude and conforming to the frequency characteristics of the device's partial discharge signal can be selected as the main current frequency band. Simultaneously, the frequency range with lower amplitude and not conforming to the characteristics of the partial discharge signal is determined as the current noise frequency band. Based on the frequency range and amplitude characteristics of the main current frequency band, a decoupling filter function can be constructed. The passband of this function matches the main current frequency band, and the stopband matches the current noise frequency band. Next, the high-frequency current signal is input into the decoupling filter function. The filter function retains the effective signal within the main current frequency band and filters out the interference signal within the current noise frequency band, generating decoupling current data. The edge computing module then activates a preset pulse processing algorithm to extract pulses from the voltage signal, separating the transient pulse component from the voltage signal to obtain the voltage pulse signal. Based on preset bandwidth parameters, the voltage pulse signal is bandwidth filtered to remove noise components exceeding the bandwidth range, achieving noise decoupling and obtaining a decoupled voltage signal. Simultaneously, integration is used to quantize the energy of the decoupled voltage signal, calculating the cumulative energy value per unit time, i.e., the decoupled voltage energy value. This decoupled voltage energy value is then compared with a preset voltage energy threshold, and combined with the frequency characteristics of the voltage pulse signal, quantitative data characterizing the device voltage state is determined—the decoupled voltage data. For high-frequency current signals, the difference between the effective and interference components mainly lies in the frequency range. Therefore, spectral analysis is used to determine the main frequency band and noise band, and a filtering function is constructed based on this to achieve accurate decoupling of the current signal. Voltage signals often exist in pulse form, and interference noise is mostly distributed outside a specific bandwidth. Therefore, pulse characteristics are first extracted using a pulse processing algorithm, then noise is filtered out according to preset bandwidth parameters, and finally, the energy value is obtained through integration, ensuring that the decoupled voltage signal data accurately reflects the device voltage state.

[0030] It is evident that through decoupling, high-frequency current signals and voltage signals can be accurately filtered out and effective signals can be extracted, ensuring that decoupled current data and decoupled voltage data can truly and accurately characterize the partial discharge current characteristics and voltage status of the equipment.

[0031] Optionally, the above step of generating the decoupled current data based on the decoupling filter function, the main current frequency band, and the current noise frequency band specifically includes the following steps: B201. Determine the basis functions and the number of decomposition layers of the decoupling filter function; B202. Decompose the high-frequency current signal according to the basis function and the number of decomposition layers to obtain multiple frequency band decomposition coefficients; B203. Select the frequency band decomposition coefficients corresponding to the main frequency band of the current from the plurality of frequency band decomposition coefficients to obtain the main frequency decomposition coefficients; B204. Select the frequency band decomposition coefficients corresponding to the current noise frequency band from the plurality of frequency band decomposition coefficients to obtain the first noise decomposition coefficient; B205. Determine the noise decomposition coefficients that are greater than the preset high-frequency coefficient threshold, and obtain the second noise decomposition coefficients. B206. Wavelet reconstruction is performed based on the second noise decomposition coefficient and the main frequency decomposition coefficient to obtain the decoupled current data.

[0032] In this embodiment, the basis function refers to the wavelet basis function used when constructing the decoupling filter function. It needs to be adapted to the waveform attributes of partial discharge characteristics in the high-frequency current signal. In the wind farm transmission and transformation equipment monitoring scenario corresponding to this application, the db4 wavelet basis function is preferred. This basis function has good time-frequency localization characteristics and can accurately capture the pulse characteristics of partial discharge in the high-frequency current signal. The number of decomposition levels refers to the number of levels used for frequency domain division of the high-frequency current signal using wavelet decomposition. It needs to be determined in conjunction with the frequency range of the current main frequency band and is set to 3 levels to ensure that the frequency band components corresponding to the current main frequency band and the current noise frequency band can be clearly distinguished after decomposition. The frequency band decomposition coefficients refer to the signal coefficients obtained at different decomposition levels and different frequency intervals after the high-frequency current signal is decomposed by wavelet decomposition. Their amplitude reflects the energy intensity of the signal within the corresponding frequency interval. The high-frequency coefficient threshold is a coefficient threshold pre-set based on historical monitoring data and typical noise characteristics of the wind farm transmission and transformation equipment. It is used to determine whether the frequency band decomposition coefficients correspond to strong interference noise. This threshold needs to be dynamically calibrated according to the noise level of the equipment type (such as switchgear, transformer). Wavelet reconstruction refers to the process of recombining the filtered effective frequency band decomposition coefficients with the processed noise frequency band decomposition coefficients, and then recovering the time-domain signal through inverse wavelet transform to obtain a pure current signal after eliminating invalid interference.

[0033] In a specific embodiment, the edge computing module, based on the partial discharge characteristics and historical data of the high-frequency current signal from the wind farm's transmission and transformation equipment, determines that the basis function of the decoupling filter function is the db4 wavelet basis function. Combining this with the frequency range of the current main frequency band (2MHz-100MHz), the number of wavelet decomposition layers is set to 3. The edge computing module calls the wavelet decomposition algorithm to perform wavelet decomposition on the high-frequency current signal using the determined basis function and number of decomposition layers. This decomposes the original time-domain signal into frequency band decomposition coefficients corresponding to 3 high-frequency levels and 1 low-frequency level, resulting in multiple frequency band decomposition coefficients covering different frequency ranges. Based on the determined current main frequency band range, coefficients matching this frequency range are selected from the multiple frequency band decomposition coefficients as the main frequency decomposition coefficients that reflect the partial discharge characteristics. Based on the frequency range of the current noise frequency band, coefficients corresponding to the frequency range are selected from the multiple frequency band decomposition coefficients to obtain the first noise decomposition coefficients containing only noise components. Then, the first noise decomposition coefficients are compared with a preset high-frequency coefficient threshold. First noise decomposition coefficients with amplitudes greater than this threshold are selected; these coefficients correspond to strong interference noise components and are determined as the second noise decomposition coefficients. Simultaneously, the edge computing module performs inverse wavelet transform on the primary frequency decomposition coefficients and the second noise decomposition coefficients. Through wavelet reconstruction, the two types of coefficients are fused into a time-domain signal. This signal has eliminated weak interference noise with amplitudes below the high-frequency coefficient threshold and retains complete partial discharge characteristics, i.e., decoupled current data.

[0034] It is evident that wavelet decomposition and reconstruction can accurately remove interference noise from high-frequency current signals, especially the targeted suppression of strong interference noise, and obtain decoupled current data that can truly and completely reflect the partial discharge characteristics of wind farm transmission and transformation equipment. This effectively solves the problem that traditional filtering methods are prone to signal distortion or noise residue.

[0035] S103. The decoupled current data, the decoupled voltage data, the temperature data, and the gas state data are fused to obtain multidimensional fused data.

[0036] In this embodiment of the application, multidimensional fusion data refers to a unified data form that can comprehensively characterize the overall operating status of wind farm transmission and transformation equipment after the correlation analysis and information integration of four different dimensions of equipment operating parameters: decoupled current data, decoupled voltage data, temperature data, and gas state data.

[0037] In a specific embodiment, the edge computing module first standardizes the decoupling current data, decoupling voltage data, temperature data, and gas state data, converting all types of data into a unified numerical range to eliminate the influence of differences in the dimensions of different parameters on the fusion result. Then, the original standardized values ​​of each parameter are fused to obtain multidimensional fused data.

[0038] Optionally, the above step of fusing the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain multidimensional fused data specifically includes the following steps: A301. Obtain historical detection data from the multiple sensors in the target wind farm during a historical period; A302. Determine the fault information in the historical detection data; A303. Determine the confidence set corresponding to the fault information using DS theory; A304. Determine the weight parameters corresponding to the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data based on the confidence set, and obtain the current weight parameter, voltage weight parameter, temperature weight parameter, and gas weight parameter; A305. Weighted fusion is performed based on the current weight parameter, the voltage weight parameter, the temperature weight parameter, the gas weight parameter, the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain multidimensional fused data.

[0039] In this embodiment, historical monitoring data refers to the operational data of power transmission and transformation equipment collected and stored by the same type of sensors before the monitoring period of this embodiment. This includes historical high-frequency current signals, historical voltage signals, historical temperature data, and historical gas state data, and the data must cover multiple scenarios such as normal equipment operation, abnormal conditions, and fault occurrence. Fault information refers to the feature information related to power transmission and transformation equipment faults extracted from historical monitoring data, including the numerical range, trend of change, and correlation between parameters at the time of fault occurrence, such as the historical partial discharge intensity threshold and abnormal temperature range corresponding to cable insulation aging faults. DS theory, or evidence theory, is a multi-source information fusion method for handling uncertain information. It can synthesize evidence from different sources by defining a confidence function, such as synthesizing the correlation between various monitoring parameters and faults, to achieve decision-making on uncertain issues. The confidence set refers to the set formed by quantifying the correlation between various monitoring parameters and fault information in historical monitoring data based on DS theory. Each element corresponds to the degree of support of a certain monitoring parameter for a specific fault type, i.e., the confidence value. A higher confidence value indicates a greater contribution of the parameter to fault identification.

[0040] In a specific embodiment, the edge computing module retrieves all sensor detection data of the target wind farm's transmission and transformation equipment from a cloud server or local historical database within a historical period via a data interface. This ensures the retrieved data spans at least one year and contains at least 1000 sets of fault-related data to guarantee data representativeness. Next, the edge computing module performs anomaly identification and fault labeling on the historical detection data. Combined with wind farm operation and maintenance records, it extracts the characteristics of each parameter corresponding to the fault occurrence time, determines the fault information corresponding to different fault types, and establishes a mapping relationship between fault types and parameter characteristics. Fault types can include insulation aging, SF6 leakage, etc. Based on DS theory, a confidence function is constructed, matching the historical data of each monitoring parameter with the corresponding fault information. The confidence value of each parameter for different fault types is calculated, forming a confidence set covering all parameters and fault types. Then, based on the proportion of confidence values ​​for each parameter in the confidence set, the weight parameters corresponding to decoupling current data, decoupling voltage data, temperature data, and gas state data are determined. Among these parameters, those with higher reliability values ​​correspond to larger weight parameters. For example, if historical data shows that decoupling current data has the highest reliability value for partial discharge faults, then the current weight parameter is set to the maximum, resulting in current weight parameters, voltage weight parameters, temperature weight parameters, and gas weight parameters. Finally, the edge computing module first standardizes the decoupling current data, decoupling voltage data, temperature data, and gas state data to eliminate dimensional differences. Then, it multiplies each standardized data point by its corresponding weight parameter to obtain a weighted value for each parameter. Subsequently, it sums all weighted values ​​and integrates the results with the original standardized values ​​and weight parameter information to form multidimensional fused data. Furthermore, since different monitoring parameters of wind farm transmission and transformation equipment contribute differently to fault identification, using fixed weights for data fusion can easily weaken the role of key parameters or amplify interference from secondary parameters, affecting the accuracy of the fusion results. Therefore, by retrieving historical detection data and extracting fault information, the correlation between parameters and faults can be determined based on actual equipment operating experience rather than theoretical assumptions, improving the rationality of weight allocation. The DS theory has the ability to handle uncertain information, effectively integrate the fault characterization capabilities of various parameters in historical data, quantify the contribution of parameters through the confidence set, and then determine the weight parameters.

[0041] It is evident that the scientific determination of weight parameters based on historical fault data and DS theory has been achieved, avoiding the bias caused by subjective weight setting. At the same time, the role of key monitoring parameters has been highlighted through weighted fusion, enabling the obtained multi-dimensional fused data to more accurately correlate equipment operating status and fault characteristics, significantly improving the rationality and effectiveness of data fusion.

[0042] S104. Input the multi-dimensional fusion data into a preset fault identification model to obtain fault identification information; and send the fault identification information to the cloud server.

[0043] In this embodiment, the preset fault identification model refers to an intelligent diagnostic model that is pre-built and trained based on historical fault data, operation and maintenance records, and equipment characteristics of the target wind farm's transmission and transformation equipment. The model is preferably a random forest model, which learns from multi-dimensional data by integrating multiple decision trees, enabling the classification and severity determination of various fault types. Furthermore, the model training samples contain at least 1000 sets of labeled data including different fault types and severity levels, ensuring diagnostic accuracy. Fault identification information refers to the result information output by the fault identification model after analyzing the multi-dimensional fused data. It includes at least whether the equipment has a fault, the fault type (partial discharge exceeding the standard fault, abnormal temperature fault), the fault severity (minor, moderate, severe, etc.), and the probability of fault occurrence.

[0044] In a specific embodiment, the edge computing module invokes a locally deployed preset fault identification model, inputting multi-dimensional fused data into the model. The model uses an internally integrated decision tree algorithm to perform feature matching and probability calculation on the data, outputting fault identification information including the presence or absence of a fault, fault type, severity, and probability of occurrence. Then, the edge computing module encrypts and transmits the fault identification information to the cloud server via a fiber optic communication link. Upon receiving the fault identification information, the cloud server associates it with the corresponding device number and collection time, stores it, updates the device status monitoring database, and simultaneously triggers a refresh of the visualization interface for real-time viewing by maintenance personnel.

[0045] Optionally, the power transmission and transformation equipment management system further includes an early warning module, and the method specifically includes the following steps: A401. Extract the discharge current data corresponding to the target power transmission and transformation equipment from the decoupling current data to obtain the target discharge current data; the target power transmission and transformation equipment is any one of the power transmission and transformation equipment in the wind farm; A402. Determine the target discharge intensity value based on the target discharge current data; A403. Determine the target temperature corresponding to the target power transmission and transformation equipment based on the temperature data; A404. Extract the gas pressure value from the gas state data corresponding to the target power transmission and transformation equipment to obtain the target gas pressure value; A405. Based on a preset weight allocation algorithm, the weight coefficients corresponding to the target discharge intensity value, the target temperature, and the target gas pressure value are used to obtain the equipment weight coefficients corresponding to the power transmission and transformation equipment. A406. The target power transmission and transformation equipment value is obtained by calculating based on the equipment weighting coefficient, the target discharge intensity value, the target temperature and the target gas pressure value; A407. Based on the preset mapping relationship between power transmission and transformation equipment values ​​and equipment health indices, determine the health index corresponding to the target power transmission and transformation equipment value to obtain the target health index; A408. When the target health index is less than or equal to a preset index threshold, the abnormality level of the target power transmission and transformation equipment is determined based on the target health index. A409. The early warning module executes the corresponding early warning event according to the anomaly level and sends the early warning event to the cloud server.

[0046] In this embodiment, the preset weight allocation algorithm is set based on equipment failure mechanisms and historical operation and maintenance data. For example, gas pressure has a greater impact on insulation in GIS equipment, resulting in a higher weight coefficient. The equipment weight coefficient is a quantitative coefficient calculated by the weight allocation algorithm, representing the importance of target discharge intensity, target temperature, and target gas pressure in equipment health assessment. The sum of the coefficients is 1, with higher coefficients corresponding to parameters of higher importance. The preset index threshold is a critical health index value set based on the target power transmission and transformation equipment's factory standards, industry operation and maintenance specifications, and historical failure data. When the health index is below this threshold, it indicates an abnormal risk in the equipment. The anomaly level refers to the severity level of equipment anomaly classified according to the difference between the target health index and the index threshold, typically categorized as minor, moderate, and severe anomalies, with different emergency response priorities for each level. A warning event is a warning message generated by the warning module based on the anomaly level, containing the equipment number, abnormal parameters, anomaly level, suggested handling measures, and trigger time. This message can be pushed via email or system pop-ups.

[0047] In a specific embodiment, any one of the transmission and transformation equipment in the wind farm is analyzed to determine the target transmission and transformation equipment. Then, the discharge current record corresponding to this equipment is selected from the decoupling current data to obtain the target discharge current data. The edge computing module performs statistical analysis on the target discharge current data, calculates the peak-to-average value of the discharge current within a preset time period (e.g., 1 hour), and determines this average value as the target discharge intensity value. The edge computing module extracts the temperature values ​​of each monitoring point of the target transmission and transformation equipment from the temperature data, and takes the average temperature of each monitoring point as the target temperature. If the target transmission and transformation equipment is a GIS device, the edge computing module selects the SF6 gas pressure data corresponding to the gas chamber of the equipment from the gas state data, removes instantaneous fluctuation values ​​to obtain the target gas pressure value. If the equipment does not need to monitor the gas state, this parameter weight allocation is skipped. The edge computing module calls the weight allocation algorithm to assign weight coefficients according to the type of the target transmission and transformation equipment. For example, the weight coefficient for the target gas pressure value of the GIS device is set to 0.35, the target discharge intensity value to 0.4, and the target temperature to 0.25. The equipment weight coefficient is calculated through the algorithm. Then, the target discharge intensity value, target temperature, and target gas pressure value are multiplied by the corresponding equipment weight coefficients, and the products are summed to obtain the target power transmission and transformation equipment value. The edge computing module retrieves a preset mapping table between the target power transmission and transformation equipment value and the equipment health index, and looks up the corresponding health index based on the calculated target power transmission and transformation equipment value to determine the target health index. The target health index is compared with a preset index threshold (e.g., 60 points). If the target health index is ≤60 points, the abnormality level is classified according to the difference range (e.g., 50-60 points for minor abnormality, 30-50 points for moderate abnormality, and <30 points for severe abnormality). Finally, the early warning module generates an early warning event based on the abnormality level. Minor abnormalities only push a system pop-up notification, moderate abnormalities send an email to the operations and maintenance manager, and severe abnormalities send both SMS and email and trigger an audible and visual alarm. The early warning event is then encrypted and transmitted to the cloud server for storage, facilitating traceability and handling by operations and maintenance personnel.

[0048] It is evident that by accurately assessing equipment and providing differentiated early warnings, personalized health monitoring can be achieved for different types of transmission and transformation equipment in wind farms, avoiding the limitations of uniform assessments. At the same time, tiered early warnings ensure efficient handling of abnormal events, reducing waste of operation and maintenance resources. Furthermore, cloud storage of early warning events facilitates traceability and management, effectively improving the accuracy and response efficiency of wind farm transmission and transformation equipment operation and maintenance, and reducing the risk of equipment failure escalation.

[0049] Optionally, the multiple sensors include UHF RFID passive wireless temperature sensors disposed at a predetermined distance on the wind farm's transmission and transformation equipment. The UHF RFID passive wireless temperature sensors are connected to a UHF RFID reader via wireless radio frequency communication, wherein: The UHF RFID reader is controlled to transmit a radio frequency signal in a preset frequency band to the UHF RFID passive wireless temperature sensor; The UHF RFID passive wireless temperature sensor receives the radio frequency signal and performs temperature detection on the wind farm power transmission and transformation equipment. The temperature signal is processed to obtain the temperature data.

[0050] In this embodiment, the preset distance refers to the installation distance between the sensor and the monitoring point of the equipment, pre-set according to the signal transmission capability of the UHF RFID passive wireless temperature sensor, the structural dimensions of the wind farm power transmission and transformation equipment, and the monitoring accuracy requirements. This distance is usually controlled within a range that ensures the sensor can stably collect temperature signals without affecting the normal operation of the equipment. For example, the preset distance is set to 1-3 cm when monitoring cable joints. The preset frequency band refers to the radio frequency signal frequency range that conforms to the UHF RFID communication standard and avoids the strong electromagnetic interference frequency band of the wind farm. In this application, the 902-928MHz industrial frequency band is preferred. This frequency band has strong anti-interference and penetration capabilities, which can ensure stable communication between the sensor and the reader. The UHF RFID reader refers to a device with radio frequency signal transmission, reception, and data processing functions. It can activate the passive sensor by transmitting radio frequency signals in the preset frequency band and receive the temperature-related signals fed back by the sensor, thereby realizing batch data reading and management of multiple UHF RFID passive wireless temperature sensors.

[0051] In a specific embodiment, the edge computing module sends control commands to the UHF RFID reader, which include preset frequency band parameters and data acquisition cycles. Upon receiving the commands, the UHF RFID reader continuously transmits radio frequency signals according to the preset frequency band. These radio frequency signals are used not only to establish a communication link with the sensor but also to provide power to the UHF RFID passive wireless temperature sensor. The UHF RFID passive wireless temperature sensor, installed at a monitoring point of the wind farm's power transmission and transformation equipment within a preset distance, activates its internal temperature detection module using the signal energy after receiving the radio frequency signal. This module performs real-time detection of the surface or internal temperature of the equipment, generating an analog temperature signal. Next, the sensor's internal signal processing unit performs analog-to-digital conversion and encoding on the temperature signal, converting the analog signal into a digital signal and adding unique sensor identification information, such as the equipment number and monitoring point location. Finally, the sensor feeds back the processed digital temperature signal to the UHF RFID reader via wireless radio frequency communication. The reader decodes and verifies the received digital signal, removing invalid or erroneous data to obtain accurate temperature data characterizing the equipment's temperature state. This data is then transmitted to the edge computing module for subsequent fusion processing.

[0052] Wind farm transmission and transformation equipment is often located in environments with high salt spray, strong electromagnetic interference, or dispersed deployments. Traditional wired temperature sensors suffer from wiring corrosion and insulation degradation, while ordinary battery-powered wireless sensors require frequent battery replacements, resulting in high maintenance costs. UHF RFID passive wireless temperature sensors eliminate the need for built-in batteries, obtaining energy from the radio frequency signals emitted by a UHF RFID reader, effectively solving the power supply problem and extending sensor lifespan. Employing preset frequency band radio frequency communication avoids electromagnetic interference within the wind farm, ensuring stable data transmission. Controlling the reader to read sensor data in batches enables centralized temperature monitoring of multiple devices and monitoring points, adapting to the dispersed nature of wind farm equipment.

[0053] Optionally, the multiple sensors include a high-frequency current sensor and a transient ground voltage sensor; the high-frequency current sensor is used to acquire the high-frequency current signal within a preset first frequency range; the magnetic core of the high-frequency current sensor is nickel-zinc ferrite; the inner diameter of the magnetic core is 60 mm; and the outer diameter of the magnetic core is 90 mm. The coupling electrode of the transient ground voltage sensor is a ball electrode; the transient ground voltage sensor includes an integrating circuit; the integrating circuit is used to receive the ground voltage signal of the coupling electrode, and to integrate and restore the ground voltage signal to obtain the voltage signal.

[0054] In this embodiment, the preset first frequency range refers to a high-frequency current signal acquisition interval pre-set according to the typical frequency characteristics of partial discharge signals from wind farm transmission and transformation equipment. In this application, this range is set to 2MHz-100MHz. This range can completely cover the high-frequency signals of partial discharge caused by equipment insulation defects, while avoiding the frequency range of power frequency interference signals. Nickel-zinc ferrite is a magnetic material with excellent high-frequency permeability and low loss characteristics, suitable for the sensing and transmission of high-frequency signals. As the core material of the high-frequency current sensor, it can ensure that the sensor has a high sensitivity acquisition capability for high-frequency current signals in the range of 2MHz-100MHz. The coupling electrode is a conductive component in the transient ground voltage sensor used to sense the transient ground voltage signal generated by partial discharge of the equipment. In this step, a ball electrode structure is adopted. This structure has uniform electric field sensing characteristics and can stably capture the transient ground voltage signal leaked on the surface of the equipment. The integrating circuit is a circuit unit inside the transient voltage sensor used to integrate the instantaneous voltage signal collected by the coupling electrode. It consists of resistors and capacitors and can convert the instantaneously changing voltage signal into a stable signal that reflects the accumulation of signal energy, thereby realizing the restoration and quantization of the transient voltage signal.

[0055] For easier understanding, please refer to Figure 3 , Figure 3This is a schematic diagram of a high-frequency current sensor provided in an embodiment of this application. As can be seen, the high-frequency current sensor 300 mainly consists of a magnetic core, a winding coil, and a corresponding equivalent circuit. This structure is used to detect high-frequency partial discharge current signals in wind farm transmission and transformation equipment, thereby achieving real-time monitoring and evaluation of the equipment's insulation status.

[0056] Specifically, the physical structure of the high-frequency current sensor adopts a toroidal magnetic core structure. Several turns of induction coil are evenly wound around the outer circumference of the magnetic core. The conductor being measured (such as a cable grounding wire or the grounding terminal of a circuit breaker) passes through the center of the magnetic core to form a primary circuit. When a partial discharge pulse current exists in the conductor... At this time, a time-varying magnetic flux is generated in the magnetic core, thereby inducing an induced current in the winding coil according to the principle of electromagnetic induction. The induced current outputs an induced voltage through the signal output terminal. ( The output voltage is the original signal reflecting the intensity of partial discharge activity. This structure enables non-contact high-frequency current detection, avoiding interference with the primary circuit while ensuring the safety and stability of signal acquisition. The high-frequency current sensor 300 can be equivalent to a mutual inductance voltage source. = / dt, sensor coil inductance coil internal resistance Parasitic capacitance And the circuit formed by the load resistance Ra. Mutual inductance coefficient. This indicates the degree of electromagnetic coupling between the primary circuit and the secondary winding, and its magnitude is closely related to the core material, number of turns, and geometry. When the current in the measured conductor undergoes a high-frequency change, the mutual inductance voltage source induces a current in the secondary circuit. The induced current passes through the inductor With capacitor The generated induced current The resulting resonant network transmits power to the output. In this equivalent circuit, the inductor... The self-inductance characteristics of the sensor winding were characterized, playing a dominant role in the high-frequency response of the signal; resistance The internal resistance of the winding wire will cause signal amplitude attenuation and energy loss; capacitor This reflects the influence of the distributed capacitance of the insulation layers between windings and between wires on the high-frequency cutoff characteristics of the sensor. At lower frequencies, the sensor exhibits inductive characteristics; however, in the high-frequency region, a resonance effect forms between the inductance and capacitance, causing the output voltage to... The system exhibits a more sensitive response to partial discharge pulse signals. The output voltage can be acquired via the load resistor Ra and transmitted to the edge computing unit for subsequent analysis. The high-frequency current sensor 300 is typically installed in the cable grounding wire or circuit breaker grounding loop to capture high-frequency partial discharge signals in the frequency range of 2MHz to 100MHz. The sensor core is preferably made of nickel-zinc ferrite material, which has high permeability and low loss characteristics, effectively improving signal sensitivity and anti-interference performance. Its typical structural parameters include an inner diameter of 60mm, an outer diameter of 90mm, a thickness of 20mm, 6 winding turns, and an integrating resistance of 50Ω, achieving a detection sensitivity of at least 1pC discharge level. After the signal is acquired by the high-frequency current sensor, it can be denoised and feature extracted using an integrating circuit or wavelet filtering algorithm, and then used to determine the insulation degradation degree and discharge mode (such as intermittent, continuous, or surface discharge) of the power transmission and transformation equipment. When abnormal changes occur in the amplitude or spectral characteristics of the high-frequency current signal, the system can immediately trigger an edge-end early warning mechanism, providing key feature input to the cloud-based intelligent diagnostic module.

[0057] For easier understanding, please refer to Figure 4 , Figure 4 This is a schematic diagram of the integrating circuit of a transient ground voltage sensor provided in an embodiment of this application. As can be seen, the integrating circuit 400 of the transient ground voltage sensor is used to integrate the transient ground voltage signal to recover the original transient ground voltage waveform of the device under test. This circuit mainly includes an input resistor. Feedback capacitor Operational amplifier A and its input terminals With output terminal It consists of components such as [list of components], and is used to convert the high-frequency transient voltage signal from the sensor coupling electrode into an output voltage signal that is proportional to the time integral.

[0058] Specifically, the input end This is used to receive transient voltage signals to ground collected by a transient ground voltage sensor from the device under test (such as the inner wall of a switch cabinet or a metal casing). This input signal is transmitted via an input resistor. The current enters the inverting input terminal of the operational amplifier, forming the input current. A capacitor is placed in the negative feedback loop of the operational amplifier. As a feedback element, it is used to transmit the feedback current. Stored inside the capacitor, and a voltage is generated across the capacitor. As the input signal changes over time, the charge in the capacitor continuously accumulates or is released, thus generating an output voltage at the output terminal that is inversely proportional to the time integral of the input signal. Under normal operating conditions, the positive input terminal of operational amplifier A is grounded, and the negative input terminal maintains a virtual ground potential, causing the input current to... Almost all of the flow goes through the feedback capacitor. This enables accurate integration calculations.

[0059] In this embodiment, the integrating circuit 400 is mainly used in a Transient Earth Voltage (TEV) sensor to reconstruct the waveform of the transient signal acquired by the coupling electrode. Because TEV signals have extremely short pulse times and a high frequency range (typically 1MHz to 50MHz), the directly detected voltage waveform often contains high-frequency noise and parasitic interference, making it difficult to accurately reflect the energy changes of partial discharge within the device. Figure 4 The integrator circuit shown can convert these high-frequency pulse signals into corresponding low-frequency integrated signals, thereby recovering their true transient voltage waveform to ground and realizing quantitative analysis of the amplitude, energy and discharge cycle of partial discharge pulses.

[0060] It is worth noting that this integrator circuit structure features high input impedance and low output impedance, effectively avoiding the loading effect on the original signal while enhancing the response sensitivity to high-frequency signals. Operational amplifier A is typically selected as a high-bandwidth, low-noise model to ensure stable gain and linear response at MHz-level signal frequencies. The feedback capacitor C is generally in the range of 100pF to match the system's high-frequency characteristic impedance, achieving fast integration and accurate output.

[0061] In a specific embodiment, during the high-frequency current signal acquisition stage, a high-frequency current sensor with a nickel-zinc ferrite core is installed on the grounding wire of the cable or the grounding wire of the circuit breaker contact of the wind farm transmission and transformation equipment using a clamp-type structure (60mm inner diameter, 90mm outer diameter, 20mm thickness), ensuring a tight fit between the sensor and the grounding wire. During equipment operation, the sensor utilizes the high sensitivity of the nickel-zinc ferrite core to high-frequency signals to acquire high-frequency current signals within a preset first frequency range in real time. This signal directly corresponds to the current characteristics generated by partial discharge in the equipment. The core size design during acquisition is adaptable to most grounding wire specifications, while ensuring the integrity of signal acquisition. During the transient ground voltage signal acquisition stage, a transient ground voltage sensor with a ball electrode as the coupling electrode is fixed to the inner wall of the switch cabinet using a magnetic structure, with the ball electrode facing the area where partial discharge may occur. When partial discharge occurs in the equipment, a transient ground voltage signal is generated. The ball electrode senses this signal and transmits it to the integration circuit inside the sensor. The integrator circuit integrates the received instantaneous voltage signal to ground, restores the energy characteristics of the signal through the charging and discharging of resistors and capacitors, and converts the instantaneous fluctuating voltage signal into a stable and quantifiable voltage signal. The voltage signal is then transmitted to the edge computing module to participate in subsequent data processing together with the high-frequency current signal.

[0062] It is evident that by combining high-frequency current sensors and transient ground voltage sensors, the high-frequency current signals and transient ground voltage signals corresponding to partial discharge of equipment can be obtained completely and accurately. The two types of signals complement each other for verification, effectively avoiding the limitations of single signal acquisition. At the same time, the structural design and material selection of the sensors are adapted to the installation scenario and signal characteristics of wind farm equipment, ensuring the universality, sensitivity and stability of signal acquisition, thereby improving the accuracy of fault identification.

[0063] For easier understanding, please refer to Figure 5 , Figure 5 This is a schematic diagram of the structure of an SF6 gas real-time monitoring system provided in an embodiment of this application. As can be seen, the SF6 gas real-time monitoring system 500 mainly consists of three parts: a primary unit, a local data acquisition cabinet, and a back-end cabinet. It is used to realize real-time monitoring, data acquisition, and remote transmission of the gas insulation system in wind farm power transmission and transformation equipment, thereby ensuring the safety and reliability of equipment operation.

[0064] Specifically, the SF6 gas real-time monitoring system 500 includes a primary body containing multiple SF6 gas sensors, which are installed at the gas chamber interface positions of circuit breakers, switchgear, or gas-insulated metal-enclosed switchgear (GIS) to monitor parameters such as SF6 gas pressure, temperature, humidity, and trace moisture content in real time. Each SF6 sensor is electrically connected to a hub via an inner core cable, enabling parallel acquisition of gas status information from multiple points. The SF6 sensors feature high accuracy and long-term stability, with a typical measurement range of 0–1 MPa pressure and dew point. With a temperature range of 50℃~50℃Td, it can operate stably in the high-humidity and high-salt-fog environment of offshore wind farms. The local data acquisition cabinet includes a hub and an IED (Intelligent Electronic Device) unit. The hub collects electrical signals from multiple SF6 gas sensors and performs preliminary filtering and standardization to suppress noise interference and ensure data consistency. The processed signals are then transmitted to the IED unit via cable. The IED unit, as an intelligent data acquisition and edge computing unit, has multi-channel data input interfaces and data synchronization processing capabilities, enabling real-time analysis and anomaly identification of gas state data from different sensor points. The IED unit uses built-in algorithms to determine gas tightness and leakage trends. When it detects excessive pressure or dew point, it automatically generates an alarm signal and records an anomaly timestamp for further diagnosis by the backend system. The backend cabinet includes a switch and a backend system. The IED unit communicates with the switch via a dual-core optical fiber, forming a high-speed, low-latency data transmission channel. The switch is responsible for aggregating the data uploaded from the local data acquisition cabinet and sending it to the backend monitoring system, achieving centralized data management across multiple devices and regions. The backend system is typically deployed in the wind farm's operation and maintenance center or on a cloud server. Its functions include data storage, trend analysis, alarm management, and visualization. Through the backend system, maintenance personnel can view the monitoring curves of each SF6 sensor, as well as the trends in gas pressure and temperature changes, in real time. They can also predict gas leak development trends using built-in models, thereby enabling early warning and maintenance decisions. In the specific signal transmission path, the analog signal output from the SF6 sensor is first transmitted via cable to the hub, then output from the hub to the IED device via a standardized interface. The IED then uploads the data via fiber optic cable to the switch, and finally the switch transmits the signal to the backend system. The entire signal chain forms a closed-loop structure from "front-end acquisition—edge processing—backend analysis," ensuring the real-time performance and accuracy of the monitoring data. Furthermore, the SF6 gas real-time monitoring system 500 adopts a modular design, facilitating flexible deployment at different monitoring points within the wind farm. Both the hub and the IED device have an IP55 protection rating or higher, allowing for... It operates stably in environments ranging from 25℃ to +70℃, meeting the requirements for use in complex offshore environments. The backend system supports standard communication protocols (such as IEC61850 and Modbus RTU), enabling data interconnection with existing monitoring platforms in wind farms.

[0065] As can be seen, the SF6 gas real-time monitoring system 500 structure enables continuous online detection and intelligent diagnosis of gas conditions at multiple points. The system can automatically identify potential hazards such as gas leaks and abnormal humidity, and achieve remote alarm and trend prediction via network. This system plays an important role in the safe operation and intelligent maintenance of power transmission and transformation equipment in offshore wind farms. It not only significantly improves the automation level of equipment insulation monitoring, but also provides reliable technical support for unmanned operation and refined maintenance of wind farms.

[0066] For easier understanding, please refer to Figure 6 , Figure 6 This is a schematic diagram of the architecture of a fault identification system for wind farm transmission and transformation equipment provided in an embodiment of this application. As can be seen, the fault identification system 600 for wind farm transmission and transformation equipment consists of three main parts: an intelligent sensing layer, an edge computing layer, and a cloud application layer, forming a multi-level integrated structure from on-site data acquisition and edge intelligent processing to cloud-based diagnostic management.

[0067] Specifically, the intelligent sensing layer mainly consists of various industrial sensors installed on wind farm power equipment (such as switchgear, transformers, circuit breakers, etc.). These industrial sensors include high-frequency current sensors, ultra-high-frequency sensors, temperature sensors, and gas sensors, used to synchronously collect multi-dimensional physical quantity signals during equipment operation, such as partial discharge signals, current, voltage, temperature, and SF6 gas pressure. These sensors are connected to the power equipment in a non-contact or embedded manner, enabling stable operation in the high-voltage, high-humidity, and strong electromagnetic interference environment of offshore wind power, ensuring the accuracy and reliability of the collected data. The raw signals collected by the intelligent sensing layer are filtered and anti-interference processed at the front end before being transmitted to the upper-level edge computing unit via wired or wireless communication. Above the intelligent sensing layer is the edge computing layer, which includes edge software for partial discharge monitoring and evaluation, an edge computing host, and connected communication modules. The edge computing host has a built-in high-performance processor and multi-channel acquisition interfaces for performing tasks such as signal processing, data integration, feature extraction, and preliminary diagnosis. The system extracts fault-related characteristic parameters by performing wavelet denoising and spectrum analysis on high-frequency signals. Simultaneously, it combines temperature and gas data to construct an edge-side health assessment model for real-time determination of equipment operating status. The edge computing layer supports local visual access through a browser-based B / S architecture management interface, allowing maintenance personnel to directly view equipment health indices, discharge intensity curves, and abnormal alarm information on the edge terminal. The edge computing layer also features data upload and protocol conversion capabilities. Through a built-in communication module, the system supports 4G / 5G or Ethernet transmission of edge analysis results and raw data to the cloud, achieving low-latency data interaction. Furthermore, the system supports multiple industrial communication protocols (such as IEC61850, Modbus RTU, etc.) to ensure data compatibility and integration between devices from different manufacturers. The cloud application layer is primarily an intelligent application cloud platform. This platform undertakes multiple functions, including fault diagnosis, health management, equipment file management, and graph viewing, constructing an intelligent decision-making center from data to knowledge. Specifically, the cloud platform utilizes embedded machine learning and big data analysis models to centrally model and intelligently diagnose multi-source monitoring data from various wind farms. By training on historical fault samples, the platform can identify different types of equipment faults, such as aging cable insulation, damp surge arresters, and abnormal transformer core grounding, and generate corresponding health levels and maintenance recommendations based on the diagnostic results. The cloud application layer also provides a remote visual management interface based on a B / S architecture, allowing maintenance personnel to access the platform via a browser to view the real-time operating status and health trends of all equipment. Furthermore, the cloud platform supports model distribution and remote configuration, enabling updated fault identification models or parameters to be distributed to the edge computing layer for system self-learning and dynamic optimization. Through this two-way linkage mechanism of "edge computing-cloud collaboration," the system can maintain local real-time response while continuously improving identification accuracy by leveraging the computing and analysis capabilities of the cloud.

[0068] In summary, by implementing the embodiments of this application, the operating signals, temperature data, and gas state data of multiple sensors in a target wind farm within a preset time period are obtained. The operating signals include high-frequency current signals and voltage signals. The high-frequency current signals and voltage signals are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, decoupled voltage data, temperature data, and gas state data are fused to obtain multi-dimensional fused data. The multi-dimensional fused data is input into a preset fault identification model to obtain fault identification information. The fault identification information is then sent to the cloud server. It is evident that by collecting multi-dimensional monitoring data from current, voltage, temperature, and gas sensors for modeling, and employing data acquisition, edge computing, and centralized cloud management for centralized intelligent diagnosis of multi-source data, the accuracy and efficiency of fault identification are improved.

[0069] Please see Figure 7 , Figure 7 This is a functional module block diagram of a fault identification device for wind farm transmission and transformation equipment provided in this application embodiment. The fault identification device 700 for wind farm transmission and transformation equipment is applied to the edge computing module of the transmission and transformation equipment management system. The transmission and transformation equipment management system includes: wind farm transmission and transformation equipment, multiple sensors installed on the wind farm transmission and transformation equipment, an edge computing module connected to the multiple sensors via communication, and a cloud server connected to the edge computing module via communication. The fault identification device 700 for wind farm transmission and transformation equipment includes: The intelligent sensing unit 701 is used to acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals. Edge computing unit 702 is used to decouple the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data; and to fuse the decoupled current data, the decoupled voltage data, the temperature data and the gas state data to obtain multi-dimensional fused data. The edge execution unit 703 is used to input the multi-dimensional fused data into a preset fault identification model to obtain fault identification information, and send the fault identification information to the cloud server.

[0070] Optionally, in the process of decoupling the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data, the edge computing unit 702 is further specifically used for: The high-frequency current signal is subjected to spectral analysis to obtain the current spectral analysis results; The dominant current frequency band and the current noise frequency band are determined based on the current spectrum analysis results. Determine the decoupling filter function corresponding to the high-frequency current signal based on the main frequency band of the current. The decoupled current data is generated based on the decoupling filter function, the main current frequency band, and the current noise frequency band. The voltage signal is processed using a preset pulse processing algorithm to obtain a voltage pulse signal; The voltage pulse signal is decoupled from noise according to the preset bandwidth parameters to obtain a decoupled voltage signal; The decoupling voltage signal is integrated to obtain the decoupling voltage energy value; The decoupling voltage data is determined based on the decoupling voltage energy value.

[0071] Optionally, in generating the decoupled current data based on the decoupling filter function, the current main frequency band, and the current noise frequency band, the edge computing unit 702 is specifically used for: Determine the basis functions and the number of decomposition layers of the decoupling filter function; The high-frequency current signal is decomposed according to the basis function and the number of decomposition layers to obtain multiple frequency band decomposition coefficients; The frequency band decomposition coefficients corresponding to the main frequency band of the current are selected from the plurality of frequency band decomposition coefficients to obtain the main frequency decomposition coefficients; The first noise decomposition coefficient is obtained by selecting the frequency band decomposition coefficient corresponding to the current noise frequency band from the plurality of frequency band decomposition coefficients. The noise decomposition coefficients that are greater than the preset high-frequency coefficient threshold are determined to obtain the second noise decomposition coefficients; Wavelet reconstruction is performed based on the second noise decomposition coefficients and the main frequency decomposition coefficients to obtain the decoupled current data.

[0072] Optionally, in fusing the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain multidimensional fused data, the edge computing unit 702 is specifically used for: Acquire historical detection data from the multiple sensors in the target wind farm within a historical time period; Determine the fault information in the historical detection data; The confidence set corresponding to the fault information is determined using DS theory; Based on the confidence set, determine the weight parameters corresponding to the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain the current weight parameter, voltage weight parameter, temperature weight parameter, and gas weight parameter; The current weight parameter, voltage weight parameter, temperature weight parameter, gas weight parameter, decoupled current data, decoupled voltage data, temperature data, and gas state data are weighted and fused to obtain multidimensional fused data.

[0073] Optionally, the power transmission and transformation equipment management system further includes an early warning module, and the fault identification device 700 of the wind farm power transmission and transformation equipment is specifically used for: The discharge current data corresponding to the target power transmission and transformation equipment is extracted from the decoupled current data to obtain the target discharge current data; the target power transmission and transformation equipment is any one of the power transmission and transformation equipment in the wind farm; Determine the target discharge intensity value based on the target discharge current data; The target temperature corresponding to the target power transmission and transformation equipment is determined based on the temperature data. The gas pressure value corresponding to the target power transmission and transformation equipment is obtained by extracting the gas pressure value from the gas state data; Based on the preset weight allocation algorithm, the weight coefficients corresponding to the target discharge intensity value, the target temperature and the target gas pressure value, the equipment weight coefficients corresponding to the power transmission and transformation equipment are obtained; The target power transmission and transformation equipment value is calculated based on the equipment weighting coefficient, the target discharge intensity value, the target temperature, and the target gas pressure value. The target health index is obtained by determining the health index corresponding to the target power transmission and transformation equipment value based on the preset mapping relationship between the power transmission and transformation equipment value and the equipment health index. When the target health index is less than or equal to a preset index threshold, the abnormality level of the target power transmission and transformation equipment is determined based on the target health index. The early warning module executes corresponding early warning events based on the anomaly level and sends the early warning events to the cloud server.

[0074] Optionally, the plurality of sensors includes a UHF RFID passive wireless temperature sensor disposed at a predetermined distance on the wind farm transmission and transformation equipment. The UHF RFID passive wireless temperature sensor is connected to a UHF RFID reader via wireless radio frequency communication, wherein: The UHF RFID reader is controlled to transmit a radio frequency signal in a preset frequency band to the UHF RFID passive wireless temperature sensor; The UHF RFID passive wireless temperature sensor receives the radio frequency signal and performs temperature detection on the wind farm power transmission and transformation equipment. The temperature signal is processed to obtain the temperature data.

[0075] Optionally, the plurality of sensors include a high-frequency current sensor and a transient ground voltage sensor; the high-frequency current sensor is used to acquire the high-frequency current signal within a preset first frequency range; the magnetic core of the high-frequency current sensor is nickel-zinc ferrite; the inner diameter of the magnetic core is 60 mm; and the outer diameter of the magnetic core is 90 mm. The coupling electrode of the transient ground voltage sensor is a ball electrode; the transient ground voltage sensor includes an integrating circuit; the integrating circuit is used to receive the ground voltage signal of the coupling electrode, and to integrate and restore the ground voltage signal to obtain the voltage signal.

[0076] The fault identification device 700 for wind farm transmission and transformation equipment described in this application can acquire operating signals, temperature data, and gas state data from multiple sensors in a target wind farm within a preset time period. The operating signals include high-frequency current signals and voltage signals. The high-frequency current signals and voltage signals are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, decoupled voltage data, temperature data, and gas state data are fused to obtain multi-dimensional fused data. The multi-dimensional fused data is input into a preset fault identification model to obtain fault identification information. The fault identification information is then sent to the cloud server. Therefore, by collecting multi-dimensional monitoring data from current, voltage, temperature, and gas sensors for modeling, and employing data acquisition, edge computing, and centralized cloud management for centralized intelligent diagnosis of multi-source data, the accuracy and efficiency of fault identification are improved.

[0077] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 800 may include a processor 810, a memory 820, a communication interface 830, and one or more programs 821. The processor 810, the memory 820, and the communication interface 830 can be interconnected via a bus. The one or more programs 821 are stored in the memory 820 and configured to be executed by the processor 810. In this embodiment, the programs include instructions for performing the following steps: Acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals; The high-frequency current signal and the voltage signal are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, the decoupled voltage data, the temperature data, and the gas state data are fused to obtain multidimensional fused data; The multidimensional fused data is input into a preset fault identification model to obtain fault identification information; and the fault identification information is sent to the cloud server.

[0078] The electronic device 800 described in this application can acquire operating signals, temperature data, and gas state data from multiple sensors in a target wind farm within a preset time period. The operating signals include high-frequency current signals and voltage signals. The high-frequency current signals and voltage signals are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, decoupled voltage data, temperature data, and gas state data are fused to obtain multi-dimensional fused data. The multi-dimensional fused data is input into a preset fault identification model to obtain fault identification information. The fault identification information is then sent to the cloud server. Therefore, by collecting multi-dimensional monitoring data from current, voltage, temperature, and gas sensors for modeling, and employing data acquisition, edge computing, and centralized cloud management for centralized intelligent diagnosis of multi-source data, the accuracy and efficiency of fault identification are improved.

[0079] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include an electronic device.

[0080] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

[0081] The steps of the methods or algorithms described in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in RAM, flash memory, ROM, EPROM, electrically erasable programmable read-only memory (EEPROM), registers, hard disk, portable hard disk, read-only optical disk (CD-ROM), or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. Alternatively, the processor and storage medium can exist as discrete components in a terminal device or management device.

[0082] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the embodiments of this application. It should be understood that the above descriptions are merely specific embodiments of the embodiments of this application and are not intended to limit the protection scope of the embodiments of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solutions of the embodiments of this application should be included within the protection scope of the embodiments of this application.

Claims

1. A fault identification method for wind farm transmission and transformation equipment, characterized in that, An edge computing module is applied to a power transmission and transformation equipment management system, the power transmission and transformation equipment management system including: wind farm power transmission and transformation equipment, multiple sensors installed on the wind farm power transmission and transformation equipment, the edge computing module communicating with the multiple sensors, and a cloud server communicating with the edge computing module, the method including: Acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals; The high-frequency current signal and the voltage signal are decoupled to obtain decoupled current data and decoupled voltage data. The decoupled current data, the decoupled voltage data, the temperature data, and the gas state data are fused to obtain multidimensional fused data; The multidimensional fused data is input into a preset fault identification model to obtain fault identification information; and the fault identification information is sent to the cloud server.

2. The method as described in claim 1, characterized in that, The decoupling process of the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data includes: The high-frequency current signal is subjected to spectral analysis to obtain the current spectral analysis results; The dominant current frequency band and the current noise frequency band are determined based on the current spectrum analysis results. Determine the decoupling filter function corresponding to the high-frequency current signal based on the main frequency band of the current. The decoupled current data is generated based on the decoupling filter function, the main current frequency band, and the current noise frequency band. The voltage signal is processed using a preset pulse processing algorithm to obtain a voltage pulse signal; The voltage pulse signal is decoupled from noise according to the preset bandwidth parameters to obtain a decoupled voltage signal; The decoupling voltage signal is integrated to obtain the decoupling voltage energy value; The decoupling voltage data is determined based on the decoupling voltage energy value.

3. The method as described in claim 2, characterized in that, The step of generating the decoupled current data based on the decoupling filter function, the current main frequency band, and the current noise frequency band includes: Determine the basis functions and the number of decomposition layers of the decoupling filter function; The high-frequency current signal is decomposed according to the basis function and the number of decomposition layers to obtain multiple frequency band decomposition coefficients; The frequency band decomposition coefficients corresponding to the main frequency band of the current are selected from the plurality of frequency band decomposition coefficients to obtain the main frequency decomposition coefficients; The first noise decomposition coefficient is obtained by selecting the frequency band decomposition coefficient corresponding to the current noise frequency band from the plurality of frequency band decomposition coefficients. The noise decomposition coefficients that are greater than the preset high-frequency coefficient threshold are determined to obtain the second noise decomposition coefficients; Wavelet reconstruction is performed based on the second noise decomposition coefficients and the main frequency decomposition coefficients to obtain the decoupled current data.

4. The method as described in claim 1, characterized in that, The process of fusing the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain multidimensional fused data includes: Acquire historical detection data from the multiple sensors in the target wind farm within a historical time period; Determine the fault information in the historical detection data; The confidence set corresponding to the fault information is determined using DS theory; Based on the confidence set, determine the weight parameters corresponding to the decoupled current data, the decoupled voltage data, the temperature data, and the gas state data to obtain the current weight parameter, voltage weight parameter, temperature weight parameter, and gas weight parameter; The current weight parameter, voltage weight parameter, temperature weight parameter, gas weight parameter, decoupled current data, decoupled voltage data, temperature data, and gas state data are weighted and fused to obtain multidimensional fused data.

5. The method as described in claim 1, characterized in that, The power transmission and transformation equipment management system further includes an early warning module, and the method further includes: The discharge current data corresponding to the target power transmission and transformation equipment is extracted from the decoupled current data to obtain the target discharge current data; the target power transmission and transformation equipment is any one of the power transmission and transformation equipment in the wind farm; Determine the target discharge intensity value based on the target discharge current data; The target temperature corresponding to the target power transmission and transformation equipment is determined based on the temperature data. The gas pressure value corresponding to the target power transmission and transformation equipment is obtained by extracting the gas pressure value from the gas state data; Based on the preset weight allocation algorithm, the weight coefficients corresponding to the target discharge intensity value, the target temperature and the target gas pressure value, the equipment weight coefficients corresponding to the power transmission and transformation equipment are obtained; The target power transmission and transformation equipment value is calculated based on the equipment weighting coefficient, the target discharge intensity value, the target temperature, and the target gas pressure value. The target health index is obtained by determining the health index corresponding to the target power transmission and transformation equipment value based on the preset mapping relationship between the power transmission and transformation equipment value and the equipment health index. When the target health index is less than or equal to a preset index threshold, the abnormality level of the target power transmission and transformation equipment is determined based on the target health index. The early warning module executes corresponding early warning events based on the anomaly level and sends the early warning events to the cloud server.

6. The method according to any one of claims 1-5, characterized in that, The plurality of sensors include UHF RFID passive wireless temperature sensors disposed at a predetermined distance on the wind farm's power transmission and transformation equipment. The UHF RFID passive wireless temperature sensors are connected to a UHF RFID reader via wireless radio frequency communication, wherein: The UHF RFID reader is controlled to transmit a radio frequency signal in a preset frequency band to the UHF RFID passive wireless temperature sensor; The UHF RFID passive wireless temperature sensor receives the radio frequency signal and performs temperature detection on the wind farm power transmission and transformation equipment. The temperature signal is processed to obtain the temperature data.

7. The method as described in claim 1, characterized in that, The plurality of sensors include a high-frequency current sensor and a transient ground voltage sensor; the high-frequency current sensor is used to acquire the high-frequency current signal within a preset first frequency range; the magnetic core of the high-frequency current sensor is nickel-zinc ferrite; the inner diameter of the magnetic core is 60 mm; the outer diameter of the magnetic core is 90 mm. The coupling electrode of the transient ground voltage sensor is a ball electrode; the transient ground voltage sensor includes an integrating circuit; the integrating circuit is used to receive the ground voltage signal of the coupling electrode, and to integrate and restore the ground voltage signal to obtain the voltage signal.

8. A fault identification device for wind farm transmission and transformation equipment, characterized in that, An edge computing module is applied to a power transmission and transformation equipment management system, the power transmission and transformation equipment management system including: wind farm power transmission and transformation equipment, multiple sensors installed on the wind farm power transmission and transformation equipment, an edge computing module communicatively connected to the multiple sensors, and a cloud server communicatively connected to the edge computing module, the device comprising: The intelligent sensing unit is used to acquire the operating signals, temperature data, and gas state data of the multiple sensors in the target wind farm within a preset time period; the operating signals include: high-frequency current signals and voltage signals. An edge computing unit is used to decouple the high-frequency current signal and the voltage signal to obtain decoupled current data and decoupled voltage data; and to fuse the decoupled current data, the decoupled voltage data, the temperature data and the gas state data to obtain multi-dimensional fused data. An edge execution unit is used to input the multi-dimensional fused data into a preset fault identification model to obtain fault identification information, and then send the fault identification information to the cloud server.

9. An electronic device, characterized in that, include: Processor, memory, communication interface, and one or more programs; The one or more programs are stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps of the method as described in any one of claims 1-7.

10. A fault identification system for wind farm transmission and transformation equipment, characterized in that, The fault identification system for the wind farm transmission and transformation equipment performs the method described in any one of claims 1-7.

Images