Partial discharge on-line diagnosis method and system for hydroelectric generator neutral point

Abstract

The application discloses a partial discharge neutral point on-line diagnosis method and system of a hydroelectric generating set generator, relates to the technical field of intelligent diagnosis, and comprises the following steps: installing a high-frequency current sensor at the neutral point position of the hydroelectric generating set generator, constructing a multi-conductor transmission line model, extracting multiple partial discharge signal samples and transmission attenuation characteristics thereof, and establishing a line-attenuation fingerprint database; after collecting partial discharge signals in real time, matching attenuation fingerprint characteristics are used to acquire a matched conductor transmission line, and directional multi-source sensing data collection is carried out on the basis of the matched conductor transmission line; partial discharge fault positioning is carried out according to multi-source data analysis, and a fault positioning result is output. The application solves the technical problems of low fault positioning precision and poor timeliness of the existing partial discharge detection method of the hydroelectric generating set generator, achieves the technical effect that the attenuation distortion disadvantage in signal propagation is converted into a fault recognition advantage by extracting transmission channel attenuation characteristics, and realizes high-precision real-time positioning of partial discharge faults.

Description

Technical Field

[0001] This invention relates to the field of intelligent diagnostic technology, specifically to an online diagnostic method and system for the neutral point of partial discharge in hydropower generators. Background Technology

[0002] During long-term operation, hydropower generators are prone to partial discharge due to factors such as insulation aging, mechanical stress, and uneven electromagnetic fields. Partial discharge not only accelerates the aging of insulation materials but can also lead to insulation breakdown, causing equipment failure and seriously affecting the safe and stable operation of hydropower units. As the operating conditions of hydropower generators become increasingly complex, accurate diagnosis of partial discharge is crucial for ensuring their safe and stable operation. However, traditional partial discharge detection methods mainly rely on manual inspection and offline testing, which suffer from low detection sensitivity, inaccurate positioning, and inability to monitor in real time. Summary of the Invention

[0003] This application provides an online diagnostic method and system for partial discharge neutral point of hydropower generators, which solves the technical problems of low fault location accuracy and poor timeliness in existing partial discharge detection methods for hydropower generators.

[0004] The first aspect of this application provides an online diagnostic method for partial discharge neutral point of a hydropower generator. The method includes: setting a high-frequency current sensor according to the neutral point location of the hydropower generator; constructing a multi-conductor transmission line model of the hydropower generator; extracting a set of multiple partial discharge signal samples based on the multi-conductor transmission line model; extracting transmission attenuation features from the set of multiple partial discharge signal samples; constructing a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features; acquiring real-time partial discharge signals through the high-frequency current sensor; extracting real-time transmission attenuation fingerprint features from the real-time partial discharge signals; matching the real-time transmission attenuation fingerprint features using the line-attenuation fingerprint database to obtain a matching conductor transmission line; performing directional multi-source sensor data acquisition based on the matching conductor transmission line; locating partial discharge faults on the matching conductor transmission line based on the directional multi-source sensor data; and outputting the partial discharge fault location result.

[0005] A second aspect of this application provides an online diagnostic system for the neutral point of a hydropower generator's partial discharge. The system includes: a current sensor setting module for setting a high-frequency current sensor according to the neutral point position of the hydropower generator; an attenuation fingerprint database construction module for constructing a multi-conductor transmission line model of the hydropower generator, extracting a set of multiple partial discharge signal samples based on the multi-conductor transmission line model, extracting transmission attenuation features from the set of multiple partial discharge signal samples, and constructing a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features; an attenuation fingerprint feature matching module for acquiring real-time partial discharge signals through the high-frequency current sensor, extracting real-time transmission attenuation fingerprint features of the real-time partial discharge signals, matching the real-time transmission attenuation fingerprint features using the line-attenuation fingerprint database, and obtaining a matching conductor transmission line; and a partial discharge fault location module for performing directional multi-source sensor data acquisition based on the matching conductor transmission line, locating partial discharge faults on the matching conductor transmission line based on the directional multi-source sensor data, and outputting the partial discharge fault location result.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method and system for online diagnosis of partial discharge neutral point of hydropower generator provided in this application belong to the field of intelligent diagnostic technology. By setting up a high-frequency current sensor to monitor the partial discharge signal of the neutral point of the hydropower generator, constructing a multi-conductor transmission line model and extracting transmission attenuation characteristics, establishing a line-attenuation fingerprint database, acquiring signals in real time and matching them with the database fingerprints, and determining the matching line, directional multi-source data acquisition is performed to achieve accurate location of partial discharge faults. This solves the technical problems of low fault location accuracy and poor timeliness in existing hydropower generator partial discharge detection methods. It realizes the technical effect of high-precision real-time location of partial discharge faults by extracting attenuation characteristics of the transmission channel, transforming the disadvantage of attenuation and distortion in signal propagation into the advantage of fault identification. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic flowchart of the online diagnostic method for partial discharge neutral point of a hydroelectric generator provided in an embodiment of this application; Figure 2This is a schematic diagram of the structure of the online diagnostic system for the neutral point of a hydroelectric generator provided in an embodiment of this application.

[0009] Figure labeling: Current sensor setting module 11, attenuation fingerprint database construction module 12, attenuation fingerprint feature matching module 13, partial discharge fault location module 14. Detailed Implementation

[0010] This application provides an online diagnostic method and system for partial discharge neutral point of hydropower generators, which solves the technical problems of low fault location accuracy and poor timeliness in existing partial discharge detection methods for hydropower generators.

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0012] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. 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 server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0013] Example 1, as Figure 1 As shown, this application provides an online diagnostic method for the neutral point of partial discharge in a hydroelectric generator, the method comprising: P10: A high-frequency current sensor is installed based on the neutral point position of the hydropower generator.

[0014] Specifically, in the process of diagnosing partial discharge in hydropower generators, a high-frequency current sensor needs to be installed at the generator's neutral point to accurately collect partial discharge signals and monitor them in real time. The neutral point, as the grounding terminal of the motor windings, is often the location where partial discharge occurs. Partial discharge usually originates from defects in the insulation system. Especially in high-voltage motor systems, partial discharge can damage the equipment. Therefore, timely detection and location of partial discharge sources are crucial to ensuring the safe operation of the generator.

[0015] A high-frequency current sensor is a device specifically designed for detecting high-frequency current signals, operating on the principle of electromagnetic induction. Unlike traditional low-frequency sensors, high-frequency current sensors can accurately capture transient signals during partial discharge processes. These signals are characteristic in frequency and typically manifest as short-duration current pulses. Due to the propagation characteristics of electromagnetic signals generated by partial discharge, these high-frequency signals often exhibit high time-varying properties, making them difficult for traditional low-frequency current sensors to capture effectively. Therefore, using a high-frequency current sensor can significantly improve the accuracy and timeliness of signal acquisition.

[0016] When selecting the sensor installation location, special attention should be paid to the generator's neutral point. This is because partial discharge is particularly common at the neutral point of electrical equipment, and the signal attenuation of partial discharge is closely related to distance. By placing the sensor near the neutral point, signal propagation loss can be minimized, thereby improving the signal-to-noise ratio. This not only ensures accurate signal acquisition at the location where partial discharge is most likely to occur but also guarantees the reliability and integrity of the data.

[0017] Furthermore, when installing a high-frequency current sensor, its real-time performance must be considered. To ensure real-time monitoring of partial discharge, the sensor must have a high sampling frequency and be able to respond quickly to signal changes. This means the sensor's sampling frequency should be much higher than the frequency at which the partial discharge signal occurs to capture instantaneous current changes. In addition, to adapt to partial discharge signals of varying intensities, the sensor must also have adaptive sampling and gain adjustment functions. These settings allow for automatic adjustment of sampling parameters when the signal strength is weak or highly variable, ensuring accurate capture of the characteristic signals of partial discharge under various operating conditions.

[0018] P20: Construct a multi-conductor transmission line model for the hydropower generator, extract a set of multiple partial discharge signal samples based on the multi-conductor transmission line model, extract transmission attenuation features from the set of multiple partial discharge signal samples, and construct a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features.

[0019] Optionally, a multi-conductor transmission line model of the hydroelectric generator needs to be constructed first. This model can accurately simulate the propagation characteristics of partial discharge signals in the multi-conductor transmission line inside the generator. The multi-conductor transmission line model is based on the electrical structure and electromagnetic field theory inside the generator. It takes into account the distributed parameters of the conductors, such as inductance, capacitance, and resistance. These parameters together affect the propagation speed and attenuation characteristics of the signal during transmission.

[0020] After establishing the multi-conductor transmission line model, the next step is to extract a set of multiple partial discharge (PD) signal samples based on the model. Partial discharge refers to a local electrical breakdown phenomenon that occurs in an insulating medium, usually accompanied by the generation of high-frequency current pulses. By simulating partial discharge conditions in the laboratory or in the field, partial discharge signals of different locations and intensities are collected to form a set of multiple PD signal samples. These signal samples will be used to analyze the attenuation characteristics of partial discharge signals in conductor transmission lines, thereby extracting representative attenuation fingerprint features.

[0021] Subsequently, transmission attenuation features were extracted from the multi-partial discharge signal sample set. Transmission attenuation features refer to the changes in the amplitude, frequency, and phase of a partial discharge signal as it propagates through a multi-conductor transmission line due to the influence of line parameters and electromagnetic fields. These attenuation features can be extracted using signal processing techniques such as Fourier transform and wavelet transform. These features reflect the propagation path and attenuation degree of the signal in different conductor transmission lines, and are crucial for identifying the signal source and locating the fault.

[0022] Finally, a line-attenuation fingerprint database is constructed, corresponding to multiple sets of transmission attenuation fingerprint features for multi-conductor transmission lines. This database stores the correspondence between multi-conductor transmission lines and their transmission attenuation features. Each conductor transmission line has its unique attenuation fingerprint features, which are recorded in detail in the database. During actual monitoring, by matching the attenuation features of the real-time acquired partial discharge signals with the fingerprint features in the database, the signal propagation path can be quickly determined, thus supporting the accurate location of partial discharge faults.

[0023] Furthermore, in constructing the multi-conductor transmission line model of the hydropower generator, step P20 of this embodiment also includes: P21: Define multiple conductor transmission lines, where each conductor transmission line is the transmission path of the partial discharge signal in the hydropower generator from the discharge source to the neutral point; P22: Identify the model parameters of each conductor transmission line, including conductor length, cross-sectional area, insulating dielectric constant, line shear coupling coefficient, distributed capacitance, inductance parameters, and termination impedance characteristics; P23: Construct a multi-conductor transmission line model based on the multiple conductor transmission lines and the model parameters of each conductor transmission line.

[0024] It should be understood that the process of constructing the multi-conductor transmission line model can be further refined. First, multiple conductor transmission lines within the generator are defined; these lines represent the paths through which partial discharge signals propagate from the discharge source to the neutral point. In hydroelectric generators, partial discharges can occur at various locations, such as winding insulation, lead connection points, or joints. Therefore, these potential discharge sources must be identified, and their corresponding signal propagation paths determined. The definition of each conductor transmission line should include its starting point (discharge source location) and ending point (neutral point location), as well as any possible branches and connections along the path. This process can be accomplished through detailed analysis and mapping of the generator's internal structure, ensuring that every possible signal propagation path is accurately recorded.

[0025] Based on the definition of the conductor transmission line, it is necessary to further identify the model parameters of each conductor transmission line. These parameters are crucial for accurately simulating the behavior of signals during transmission. Specific parameters include conductor length, cross-sectional area, dielectric constant, inter-line coupling coefficient, distributed capacitance, inductance parameters, and terminating impedance characteristics. Conductor length and cross-sectional area directly affect the propagation delay and attenuation of the signal; the dielectric constant determines the propagation speed and attenuation characteristics of the signal in the insulating medium; the inter-line coupling coefficient assesses the degree of electromagnetic coupling between adjacent conductors, which affects the signal propagation path and strength; distributed capacitance and inductance parameters measure the distributed capacitance and inductance of the conductor, determining the frequency response and phase change of the signal during transmission; terminating impedance characteristics determine the reflection characteristics of the signal at the end of the transmission line, and a suitable terminating impedance can minimize signal reflection and improve signal transmission efficiency. The identification of these parameters can be accomplished through experimental measurement, theoretical calculation, or a combination of both. For example, conductor length and cross-sectional area can be obtained through direct measurement, while the dielectric constant and distributed parameters may need to be determined through experimental testing or by consulting the characteristic data of relevant materials.

[0026] After defining the conductor transmission lines and identifying the model parameters, a complete multi-conductor transmission line model is constructed based on this information. This model can simulate the propagation characteristics of partial discharge signals along different paths, including signal amplitude attenuation, frequency variation, and phase delay. The model can be constructed by establishing a set of mathematical equations based on electromagnetic field theory and transmission line theory, taking into account the parameters of each conductor transmission line. By solving these equations, the signal propagation behavior along different paths can be obtained, providing theoretical support for subsequent signal analysis and fault location. In practice, computer-aided design and simulation software can be used to assist in the construction and verification of the multi-conductor transmission line model. These tools provide intuitive visualization interfaces, helping engineers to more accurately define conductor paths and parameters, and verify the accuracy of the model through simulation. Furthermore, by comparing with actual measurement data, the model parameters can be further optimized to ensure that the model truly reflects the signal propagation situation inside the generator.

[0027] Furthermore, based on the multi-conductor transmission line model, a set of multiple partial discharge signal samples is extracted. In this embodiment, step P20 further includes: P24: Determine multiple discharge source locations through historical partial discharge sample events. The multiple discharge source locations include at least the winding outlet, terminal, stator core location, and neutral point location. P25: Based on the multi-conductor transmission line model, each conductor transmission line is subjected to at least two pulse injection excitations at the multiple discharge source locations. The neutral point response waveform corresponding to each discharge source location is used as a partial discharge signal sample output.

[0028] Specifically, the process of extracting a set of multiple partial discharge signal samples based on the multi-conductor transmission line model can be further refined to ensure the diversity and representativeness of the samples, providing reliable data support for subsequent transmission attenuation feature extraction and fault location.

[0029] First, multiple discharge source locations are identified through historical partial discharge sample events. These locations can be determined based on previous partial discharge monitoring data and fault records, covering key areas within the generator where partial discharge may occur, including at least the winding outlet, terminals, stator core, and neutral point. The winding outlet and terminals are high-incidence areas for partial discharge because the insulation materials at these locations are susceptible to mechanical stress and electric field concentration effects; the stator core may experience discharge due to insulation aging or local defects; and the neutral point, as the signal convergence point, is also an important monitoring location. The location of each discharge source may have different effects on the propagation and characteristics of the partial discharge signal; therefore, determining these discharge source locations is crucial for subsequent signal analysis. By reviewing historical partial discharge sample events, the locations of these discharge sources can be determined, thus providing corresponding physical location calibrations for each discharge source in the multi-conductor transmission line model.

[0030] Next, based on the multi-conductor transmission line model, at least two pulse injection excitation experiments were conducted at each determined discharge source location. This process simulates the pulse signals generated by partial discharge and injects them into the multi-conductor transmission line model to obtain the neutral point response waveform corresponding to each discharge source location. Pulse injection excitation is a commonly used signal simulation method. By injecting pulse signals with known characteristics at specific locations, the changes in the signal during transmission can be observed, thereby obtaining the signal propagation characteristics. After each pulse injection, the response waveform at the neutral point was recorded. These waveforms contain information such as the amplitude, frequency, and phase of the signal during transmission, which are important manifestations of partial discharge characteristics. The number and method of injection excitation can help to comprehensively understand the impact of different discharge sources on the neutral point. The response waveforms generated by multiple pulse injection excitations at each discharge source location can provide multiple partial discharge signal samples. These samples contain information on the transmission attenuation characteristics of different discharge source locations and their corresponding transmission attenuation characteristics, which will be used for subsequent transmission attenuation feature extraction and fault location algorithm training and verification, ensuring the accuracy and reliability of the diagnostic system.

[0031] Furthermore, transmission attenuation features are extracted from the multi-partition discharge signal sample set to construct a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features. Step P20 in this embodiment of the application further includes: P26: Extract transmission attenuation features from the multi-partition discharge signal sample set, including time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features; P27: Obtain transmission attenuation fingerprints for each conductor transmission line by performing feature standardization processing on the time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features corresponding to each conductor transmission line, resulting in multiple sets of transmission attenuation fingerprint features; P28: Generate a line-attenuation fingerprint database based on the correspondence between the multi-conductor transmission lines and the multiple sets of transmission attenuation fingerprint features.

[0032] Optionally, the process of extracting transmission attenuation features from multiple partial discharge signal sample sets and constructing a line-attenuation fingerprint database can be further refined.

[0033] First, transmission attenuation features are extracted from a multi-partial discharge signal sample set. To ensure a comprehensive understanding of the signal attenuation characteristics, different attenuation features need to be extracted from multiple perspectives, including time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features. Time-domain attenuation features mainly reflect the amplitude changes of the signal on the time axis. By analyzing parameters such as amplitude attenuation, rise time, and fall time, we can understand the energy loss of the signal during propagation. Frequency-domain attenuation features transform the signal from the time domain to the frequency domain using Fourier transform, analyzing the attenuation of the signal at different frequency components. Frequency-domain features can reveal how the frequency components of the signal change with increasing transmission distance, especially the rate and pattern of frequency attenuation during high-frequency signal propagation. Wavelet time-frequency domain features combine time and frequency information. Through wavelet transform, we can analyze the changes of the signal at different time and frequency scales more precisely, which is particularly suitable for analyzing non-stationary signals, such as partial discharge signals. By extracting these three attenuation features, we can comprehensively capture the changing characteristics of the signal during transmission.

[0034] Next, feature standardization is performed on the time-domain, frequency-domain, and wavelet time-frequency-domain attenuation features corresponding to each conductor transmission line. Since these different features may have different units, scales, and dimensions, standardization is necessary to ensure their uniformity and comparability in subsequent analyses. The standardization process includes a uniform scale transformation for each feature, such as zero-mean unit variance standardization, or mapping each feature value to a fixed range through normalization. In this way, scale differences between different features are eliminated, making the influence of each feature relatively balanced in data analysis, avoiding some features from having an excessively large value that overly affects the results, or some features from having too small a value that is ignored. After standardization, the transmission attenuation features corresponding to each conductor transmission line are integrated into a unique transmission attenuation fingerprint, resulting in multiple sets of transmission attenuation fingerprint features. These transmission attenuation fingerprints are accurate representations of signal attenuation characteristics and can effectively capture subtle changes in the signal during propagation.

[0035] Finally, the correspondence between multi-conductor transmission lines and multiple sets of transmission attenuation fingerprint features is combined to generate a line-attenuation fingerprint database. In this database, the attenuation fingerprint of each transmission line is actually a "fingerprint" of the signal during its propagation along that line, containing important information such as attenuation rate, frequency variation, and time-domain and frequency-domain characteristics. By associating multiple sets of transmission attenuation fingerprint features with corresponding conductor transmission lines, the database can provide a fast and accurate reference for real-time signal matching and fault location. In practical applications, when a partial discharge signal is detected, its transmission attenuation features can be extracted and compared with the fingerprint features in the database to quickly determine the signal propagation path and possible discharge source location, thereby achieving precise fault location.

[0036] P30: Acquire the real-time partial discharge signal through the high-frequency current sensor, extract the real-time transmission attenuation fingerprint features of the real-time partial discharge signal, call the line-attenuation fingerprint database to match the real-time transmission attenuation fingerprint features, and obtain the matching conductor transmission line.

[0037] Furthermore, step P30 in this embodiment of the application also includes: P31: Process the real-time transmission attenuation fingerprint features according to the fingerprint feature storage template of the line-attenuation fingerprint database to obtain the processed real-time transmission attenuation fingerprint features; P32: Calculate the similarity between the processed real-time transmission attenuation fingerprint features and the transmission attenuation fingerprints corresponding to each conductor transmission line in the line-attenuation fingerprint database based on feature similarity, and output multiple similarities; P33: Obtain multiple confidence intervals corresponding to the multiple similarities; P34: Obtain matching conductor transmission lines from the multiple conductor transmission lines based on the multiple confidence intervals and the multiple similarities.

[0038] It should be understood that by acquiring real-time partial discharge signals through a high-frequency current sensor and comparing them with a pre-built line-attenuation fingerprint database, the precise location of the partial discharge source can be achieved.

[0039] First, a high-frequency current sensor is used to acquire real-time partial discharge signals. High-frequency current sensors can capture transient current changes caused by partial discharge, especially at the generator neutral point or other critical locations, where the waveform and amplitude of the signal reflect the health status of the electrical equipment. Real-time partial discharge signals are typically affected by multiple factors, including equipment operating status, the location of the partial discharge source, and its propagation path; therefore, further processing and analysis of these signals are necessary.

[0040] Subsequently, the real-time transmitted attenuated fingerprint features are processed according to the fingerprint feature storage template of the line-attenuation fingerprint database to ensure that the extracted features are consistent with the fingerprint features in the database in terms of format and structure, thus facilitating subsequent matching operations. Specifically, the processing may include normalizing, standardizing, or converting the real-time features to a specific format to conform to the fingerprint feature storage specifications in the database. This processing can eliminate feature deviations caused by factors such as measurement conditions and differences in sensor performance, thereby improving the accuracy of matching.

[0041] Next, based on feature similarity calculation, the processed real-time transmission attenuation fingerprint features are compared with the attenuation fingerprints corresponding to each conductor transmission line in the line-attenuation fingerprint database. Various mathematical methods, such as Euclidean distance, cosine similarity, and correlation coefficient, can be used to quantify the similarity between the real-time transmission attenuation fingerprint features and the database fingerprint features. Each conductor transmission line has a unique representation in the database; through this similarity calculation, it can be determined which transmission line is closest to the real-time signal.

[0042] Then, multiple confidence intervals are obtained corresponding to multiple similarities. A confidence interval is a quantitative indicator that assesses the reliability of the matching result based on the similarity value. Each similarity value corresponds to a confidence interval, which reflects the credibility of the matching result. The determination of confidence intervals can be achieved through statistical analysis, empirical formulas, or machine learning methods. For example, a mapping relationship between similarity and confidence can be established based on historical matching data and known fault cases, thereby assigning a reasonable confidence interval to each similarity value. For instance, if the attenuation fingerprint of a transmission line has a very high similarity to the real-time signal, the corresponding confidence interval is large, indicating a high degree of matching between the transmission line and the real-time signal; if the similarity is low, the confidence interval is small. By introducing confidence intervals, the reliability of the matching result can be assessed more comprehensively, avoiding misjudgments caused by the randomness of a single similarity value.

[0043] Finally, matching conductor transmission lines are obtained from multiple conductor transmission lines based on multiple confidence intervals and multiple similarity scores. This process is the final step in the matching decision-making process, and its purpose is to select the conductor transmission line that best matches the real-time signal characteristics from multiple candidate lines. Specifically, the conductor transmission line with the highest similarity and the most reliable confidence interval can be selected as the matching result based on a comprehensive evaluation of similarity values ​​and confidence intervals. In practice, a threshold can be set; a match is considered successful only when the similarity value exceeds the threshold and the confidence interval is within an acceptable range. If multiple candidate lines meet the conditions, a final decision can be made based on further analysis or auxiliary information, such as sensor location and signal strength.

[0044] Furthermore, step P30 in this embodiment of the application also includes: P35: The line-attenuation fingerprint database includes multiple line-attenuation fingerprint databases of the hydropower generator under multiple operating conditions, and the multiple operating conditions include at least no-load condition, rated load condition and overload condition; P36: The multiple line-attenuation fingerprint databases are switched and called according to the real-time operating conditions of the hydropower generator.

[0045] In one possible embodiment of this application, the application scope of the line-attenuation fingerprint database can be further expanded, and a database switching mechanism based on different operating conditions can be introduced to ensure that the diagnostic system can adapt to the changes in partial discharge characteristics of the generator under different operating conditions, thereby improving the accuracy and adaptability of the diagnosis.

[0046] First, the line-attenuation fingerprint database does not merely contain line-attenuation fingerprint features under a single operating condition. Instead, it constructs multiple distinct line-attenuation fingerprint databases based on the generator's status under various operating conditions. These databases correspond to different operating conditions, with common conditions including at least no-load, rated load, and overload. Under different operating conditions, the electromagnetic field distribution, temperature distribution, and electrical properties of the insulation materials within the generator all change, thus affecting the propagation characteristics of partial discharge signals. For example, under no-load conditions, the generator's electromagnetic field distribution is relatively uniform, and the propagation path and attenuation characteristics of partial discharge signals are relatively stable. However, under overload conditions, the increased current and electromagnetic field strength may accelerate the local aging of the insulation materials, altering the characteristics of the partial discharge signals. By establishing independent databases for each operating condition, the signal propagation characteristics under different conditions can be more accurately reflected, providing a more reliable reference for subsequent fault diagnosis.

[0047] Next, based on the real-time operating conditions of the hydropower generator, the system switches between multiple line-attenuation fingerprint databases. Since the operating conditions of the hydropower unit may change at different times, it is necessary to dynamically select the appropriate database for matching based on real-time operating condition information. This switching process can be triggered by real-time collected operating condition data, such as load, speed, and temperature. Specifically, the system needs to collect generator operating parameters in real time, such as current, voltage, and power, and use these parameters to determine the current operating condition. For example, when the current is close to the rated value, the system switches to the database under rated load conditions; when the current exceeds the rated value, it switches to the database under overload conditions. Through this dynamic switching mechanism, the diagnostic system can ensure that the most suitable fingerprint features are used for matching under different operating conditions, thereby improving the accuracy and reliability of fault diagnosis.

[0048] P40: Based on the matching conductor transmission line, perform directional multi-source sensing data acquisition, locate partial discharge faults on the matching conductor transmission line according to the directional multi-source sensing data, and output the partial discharge fault location result.

[0049] Furthermore, based on the matched conductor transmission line, directional multi-source sensing data acquisition is performed. In this embodiment, step P40 further includes: P41: The directional multi-source sensing data is acquired through a multi-source data acquisition module; P42: The multi-source data acquisition module includes multiple sensing data storage blocks, which are used to store multi-source sensing data corresponding to multiple conductor transmission lines; P43: The directional multi-source sensing data is read by matching the sensing data storage blocks based on the storage encoding mapping of the matching conductor transmission lines.

[0050] Specifically, directional multi-source sensor data is acquired based on matched conductor transmission lines, and this data is used for partial discharge fault location. The directional multi-source sensor data is acquired through a multi-source data acquisition module. This module coordinates the operation of multiple sensors, collecting and initially processing sensor data from different locations. It can simultaneously acquire different types of data from multiple sensors, such as current, voltage, temperature, and vibration. The multi-source data acquisition module consists of multiple sensor units, which can be distributed at different locations on the generator, such as the winding outlet, terminals, and neutral point, to obtain signals from different angles and paths when partial discharge occurs. Through the multi-source data acquisition module, comprehensive information related to partial discharge can be obtained, providing a rich data foundation for subsequent fault diagnosis.

[0051] This multi-source data acquisition module includes multiple sensor data storage blocks. These storage blocks are used to store multi-source sensor data corresponding to multiple conductor transmission lines. Each conductor transmission line has its own unique sensor data storage block, which can store all sensor data related to that line. This block-based storage method facilitates rapid retrieval and processing of data related to a specific conductor transmission line, improving the efficiency and accuracy of data management. For example, when a conductor transmission line is identified as a matching line, the system can directly access the storage block corresponding to that line to obtain the relevant multi-source sensor data without having to search through a large amount of irrelevant data.

[0052] Based on a storage-encoded mapping mechanism for matched conductor transmission lines, the system reads directional multi-source sensor data from corresponding sensor data blocks. Each conductor transmission line has a unique storage code in the database, allowing the system to accurately locate and access the multi-source data storage block associated with that specific transmission line. In practice, when the system identifies a matched conductor transmission line, it automatically reads the corresponding sensor data block through the storage-encoded mapping mechanism, retrieving directional multi-source sensor data from it. This method not only improves data retrieval speed but also ensures data accuracy and relevance. The directional multi-source sensor data obtained in this way can be directly used for partial discharge fault location analysis, thereby improving the efficiency and accuracy of fault diagnosis.

[0053] Furthermore, step P42 in this embodiment of the application also includes: P42-1: Based on the line configuration attributes of the multiple conductor transmission lines, the multi-source data acquisition module performs sensor data source complementarity identification to obtain multiple sensor data sources for each conductor transmission line that are greater than a preset complementarity threshold; P42-2: Configure the data source acquisition items for each sensor data storage block according to the multiple sensor data sources.

[0054] It should be understood that the optimization process of the sensor data storage blocks of the multi-source data acquisition module can be further refined to ensure that the acquired data has higher complementarity and relevance, thereby improving the accuracy and efficiency of fault diagnosis.

[0055] First, based on the line configuration attributes of multiple conductor transmission lines, the multi-source data acquisition module performs sensor data source complementarity identification. Line configuration attributes include the physical location, length, and insulation material characteristics of the conductor transmission lines, which determine the complementarity between different sensor data. The purpose of complementarity identification is to identify the sensor data sources that provide the most additional information, thereby enhancing the reliability of fault diagnosis. For example, by calculating the complementarity index between each sensor data source and other data sources and comparing it with a preset complementarity threshold, multiple sensor data sources for each conductor transmission line with a complementarity value greater than the preset threshold can be selected. These data sources have high spatial and functional complementarity and can provide more comprehensive fault information.

[0056] Next, based on the spatial distribution and functional characteristics of the selected multiple sensor data sources, the sensor data storage blocks of the multi-source data acquisition module are partitioned and optimized. Spatial distribution refers to the physical location of the sensors inside the generator, while functional characteristics include the sensor's measurement type, such as current, voltage, temperature, and vibration. The purpose of partition optimization is to ensure that the sensor data storage block corresponding to each conductor transmission line can efficiently store and manage the sensor data related to that line. By rationally allocating storage resources, data redundancy can be reduced, data reading speed can be improved, and data integrity and consistency can be ensured. For example, spatially adjacent and functionally related sensor data can be stored in the same storage block for subsequent joint analysis. Through the complementary identification and optimized configuration of data sources, the system can more efficiently acquire necessary partial discharge signals while reducing unnecessary data redundancy, thus improving the quality and efficiency of data acquisition.

[0057] Furthermore, based on the directional multi-source sensing data, partial discharge fault location is performed on the matched conductor transmission line, and the partial discharge fault location result is output. In this embodiment, step P40 further includes: P44: Identify the arrival time difference of the directional multi-source sensor data based on the multi-sensor time difference algorithm; P45: Calculate multiple candidate partial discharge fault location results according to the arrival time difference; P46: Perform weighted fusion of the multiple candidate partial discharge fault location results according to the complementarity of the directional multi-source sensor data, and output the partial discharge fault location result.

[0058] Optionally, the process of partial discharge fault location based on directional multi-source sensor data can be further refined to ensure the accuracy and reliability of fault location.

[0059] First, a multi-sensor time difference algorithm is used to identify the arrival time differences of directional multi-source sensor data. This algorithm utilizes the differences in signal arrival times detected by multiple sensors to determine the location of the signal source. In partial discharge fault location, the approximate location of the discharge source can be inferred by analyzing the time differences of partial discharge signals received by different sensors. Specifically, when a partial discharge occurs, the generated electromagnetic signal propagates at the speed of light, and sensors at different locations will receive the signal at different times due to their varying distances from the discharge source. Accurately measuring these time differences provides crucial timing information for subsequent fault location.

[0060] Next, multiple candidate partial discharge fault location results are calculated based on the arrival time difference. That is, based on the time difference information, combined with the generator's geometry and signal propagation model, multiple possible discharge locations are calculated. Since partial discharge can occur at multiple locations on the generator, and the signals received by the sensors may be affected by various factors, calculating multiple candidate locations is necessary. Each candidate location is calculated based on different assumptions and models, providing multiple possibilities for subsequent fusion processing.

[0061] Finally, the candidate partial discharge fault location results are weighted and fused according to the complementarity of the directional multi-source sensor data to output the final partial discharge fault location result. Complementarity refers to the degree of information complementarity between different sensor data. During the fusion process, weights are assigned based on the complementarity of each sensor data point. Data sources with high complementarity provide more accurate location information; therefore, these data sources are given higher weights during weighted fusion, and data with higher weights have a greater influence in the fusion process. Ultimately, through weighted fusion, the advantages and disadvantages of multiple candidate locations can be combined to obtain an optimal partial discharge fault location result. This result can reflect the accurate location of the partial discharge source to the greatest extent, providing precise support for equipment fault diagnosis and maintenance.

[0062] In summary, the embodiments of this application have at least the following technical effects: This application utilizes a high-frequency current sensor and a multi-conductor transmission line model to acquire partial discharge signals in real time and accurately extract transmission attenuation characteristics, achieving efficient monitoring and precise location of partial discharge faults. It transforms the attenuation distortion during signal propagation into an advantage for fault location, effectively improving the accuracy and reliability of fault location through database matching based on attenuation fingerprint characteristics. By optimizing data processing through directional multi-source data acquisition and fusion, the credibility and accuracy of fault location results are enhanced, reducing errors and redundant data in traditional methods. Database matching and real-time data analysis based on transmission attenuation characteristics significantly improve the timeliness and automation level of partial discharge fault diagnosis, reducing the need for manual intervention.

[0063] This technology achieves the technical effect of transforming the disadvantage of attenuation and distortion in signal propagation into the advantage of fault identification by extracting the attenuation characteristics of the transmission channel, thus realizing high-precision real-time positioning of partial discharge faults.

[0064] Example 2, based on the same inventive concept as the online diagnostic method for the neutral point of the partial discharge of the hydropower generator in the foregoing examples, such as... Figure 2 As shown, this application provides an online diagnostic system for the neutral point of partial discharge in hydropower generators. The system and method embodiments in this application are based on the same inventive concept. The system includes: The current sensor setting module 11 is used to set a high-frequency current sensor according to the neutral point position of the hydropower generator.

[0065] The attenuation fingerprint database construction module 12 is used to construct a multi-conductor transmission line model of the hydropower generator, extract a set of multiple partial discharge signal samples based on the multi-conductor transmission line model, extract transmission attenuation features from the set of multiple partial discharge signal samples, and construct a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features.

[0066] The attenuation fingerprint feature matching module 13 is used to acquire a real-time partial discharge signal through the high-frequency current sensor, extract the real-time transmission attenuation fingerprint feature of the real-time partial discharge signal, call the line-attenuation fingerprint database to match the real-time transmission attenuation fingerprint feature, and obtain a matching conductor transmission line.

[0067] The partial discharge fault location module 14 is used to collect directional multi-source sensing data based on the matching conductor transmission line, locate partial discharge faults on the matching conductor transmission line according to the directional multi-source sensing data, and output the partial discharge fault location result.

[0068] Furthermore, the attenuation fingerprint database construction module 12 is also used to perform the following steps: Multiple conductor transmission lines are defined, each of which is the transmission path of the partial discharge signal in the hydropower generator from the discharge source to the neutral point. Model parameters of each conductor transmission line are identified, including conductor length, cross-sectional area, insulating dielectric constant, line shear coupling coefficient, distributed capacitance, inductance parameters, and termination impedance characteristics. A multi-conductor transmission line model is constructed based on the multiple conductor transmission lines and the model parameters of each conductor line.

[0069] Furthermore, the attenuation fingerprint database construction module 12 is also used to perform the following steps: Multiple discharge source locations are determined by historical partial discharge sample events. These multiple discharge source locations include at least the winding outlet, terminal, stator core, and neutral point. Based on the multi-conductor transmission line model, each conductor transmission line is subjected to at least two pulse injection excitations at these multiple discharge source locations. The neutral point response waveform corresponding to each discharge source location is then output as a partial discharge signal sample.

[0070] Furthermore, the attenuation fingerprint database construction module 12 is also used to perform the following steps: Transmission attenuation features are extracted from the multi-partial discharge signal sample set, including time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features. By performing feature standardization processing on the time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features corresponding to each conductor transmission line, a transmission attenuation fingerprint corresponding to each conductor transmission line is obtained, resulting in multiple sets of transmission attenuation fingerprint features. Based on the correspondence between the multi-conductor transmission lines and the multiple sets of transmission attenuation fingerprint features, a line-attenuation fingerprint database is generated.

[0071] Furthermore, the attenuated fingerprint feature matching module 13 is also used to perform the following steps: The real-time transmission attenuation fingerprint features are processed according to the fingerprint feature storage template of the line-attenuation fingerprint database to obtain the processed real-time transmission attenuation fingerprint features; based on feature similarity, the processed real-time transmission attenuation fingerprint features are compared with the transmission attenuation fingerprints corresponding to each conductor transmission line in the line-attenuation fingerprint database to calculate the similarity and output multiple similarity scores; multiple confidence intervals corresponding to the multiple similarity scores are obtained; and based on the multiple confidence intervals and the multiple similarity scores, matching conductor transmission lines are obtained from the multiple conductor transmission lines.

[0072] Furthermore, the attenuated fingerprint feature matching module 13 is also used to perform the following steps: The line-attenuation fingerprint database includes multiple line-attenuation fingerprint databases based on multiple operating conditions of the hydropower generator. The multiple operating conditions include at least no-load condition, rated load condition, and overload condition. The databases are switched and called according to the real-time operating conditions of the hydropower generator.

[0073] Furthermore, the partial discharge fault location module 14 is also used to perform the following steps: Directional multi-source sensing data is acquired based on the matched conductor transmission line, and the directional multi-source sensing data is obtained through a multi-source data acquisition module; wherein, the multi-source data acquisition module includes multiple sensing data storage blocks, which are used to store multi-source sensing data corresponding to multiple conductor transmission lines; the directional multi-source sensing data is read based on the storage encoding mapping of the matched conductor transmission line and the matching sensing data storage blocks.

[0074] Furthermore, the partial discharge fault location module 14 is also used to perform the following steps: Based on the line configuration attributes of the multiple conductor transmission lines, the multi-source data acquisition module performs sensor data source complementarity identification to obtain multiple sensor data sources for each conductor transmission line that are greater than a preset complementarity threshold; and configures the data source acquisition items for each sensor data storage block according to the multiple sensor data sources.

[0075] Furthermore, the partial discharge fault location module 14 is also used to perform the following steps: The arrival time difference of the directional multi-source sensor data is identified based on the multi-sensor time difference algorithm; multiple candidate partial discharge fault location results are calculated according to the arrival time difference; the multiple candidate partial discharge fault location results are weighted and fused according to the complementarity of the directional multi-source sensor data, and the partial discharge fault location result is output.

[0076] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0077] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0078] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for online diagnosis of partial discharge neutral point in a hydropower generator, characterized in that, The method includes: A high-frequency current sensor is installed based on the neutral point position of the hydropower generator. A multi-conductor transmission line model of the hydropower generator is constructed. A set of multiple partial discharge signal samples is extracted based on the multi-conductor transmission line model. Transmission attenuation features are extracted from the set of multiple partial discharge signal samples. A line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features is constructed. The high-frequency current sensor acquires the real-time partial discharge signal, extracts the real-time transmission attenuation fingerprint features of the real-time partial discharge signal, and calls the line-attenuation fingerprint database to match the real-time transmission attenuation fingerprint features to obtain the matching conductor transmission line. Based on the matching conductor transmission line, directional multi-source sensing data is acquired, and partial discharge fault location is performed on the matching conductor transmission line according to the directional multi-source sensing data, and the partial discharge fault location result is output.

2. The method as described in claim 1, characterized in that, The method for constructing the multi-conductor transmission line model of the hydropower unit generator includes: Define multiple conductor transmission lines, where each conductor transmission line is the transmission path of the partial discharge signal in the hydropower generator from the discharge source to the neutral point location. Identify the model parameters of each conductor transmission line in the plurality of conductor transmission lines, including conductor length, cross-sectional area, insulating dielectric constant, line shear coupling coefficient, distributed capacitance, inductance parameters and termination impedance characteristics; A multi-conductor transmission line model is constructed based on the multiple conductor transmission lines and the model parameters of each conductor transmission line.

3. The method as described in claim 2, characterized in that, The method for extracting a set of multiple partial discharge signal samples based on the multi-conductor transmission line model includes: Multiple discharge source locations are determined by historical partial discharge sample events. These multiple discharge source locations include at least the winding outlet end, the terminal end, the stator core location, and the neutral point location. Based on the multi-conductor transmission line model, each conductor transmission line is subjected to at least two pulse injection excitations at the multiple discharge power source locations, and the neutral point response waveform corresponding to each discharge power source location is output as a partial discharge signal sample.

4. The method as described in claim 3, characterized in that, The method involves extracting transmission attenuation features from the set of multiple partial discharge signal samples and constructing a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features. Transmission attenuation features are extracted from the multi-partial discharge signal sample set, including time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features; By performing feature standardization on the time-domain attenuation features, frequency-domain attenuation features, and wavelet time-frequency domain features corresponding to each conductor transmission line, the transmission attenuation fingerprint corresponding to each conductor transmission line is obtained, resulting in multiple sets of transmission attenuation fingerprint features. Based on the correspondence between the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features, a line-attenuation fingerprint database is generated.

5. The method as described in claim 1, characterized in that, The method includes calling the line-attenuation fingerprint database to match the real-time transmission attenuation fingerprint features to obtain matching conductor transmission lines. The real-time transmission attenuation fingerprint features are processed according to the fingerprint feature storage template of the line-attenuation fingerprint database to obtain the processed real-time transmission attenuation fingerprint features. Based on feature similarity, the processed real-time transmission attenuation fingerprint features are compared with the transmission attenuation fingerprints corresponding to each conductor transmission line in the line-attenuation fingerprint database to calculate the similarity and output multiple similarity scores. Obtain multiple confidence intervals corresponding to the multiple similarities; Based on the multiple confidence intervals and the multiple similarities, a matching conductor transmission line is obtained from multiple conductor transmission lines.

6. The method as described in claim 1, characterized in that, Directional multi-source sensing data acquisition is performed based on the matched conductor transmission line, and the directional multi-source sensing data is acquired through a multi-source data acquisition module; The multi-source data acquisition module includes multiple sensor data storage blocks, which are used to store multi-source sensor data corresponding to multiple conductor transmission lines. Based on the storage encoding mapping of the matched conductor transmission line, the directional multi-source sensor data is read from the matched sensor data storage block.

7. The method as described in claim 6, characterized in that, The multi-source data acquisition module includes multiple sensor data storage blocks, and the method further includes: Based on the line configuration attributes of the multiple conductor transmission lines, the multi-source data acquisition module performs sensor data source complementarity identification to obtain multiple sensor data sources for each conductor transmission line that are greater than a preset complementarity threshold. Configure the data source acquisition items for each sensor data storage block according to the multiple sensor data sources.

8. The method as described in claim 7, characterized in that, The method includes locating partial discharge faults on the matched conductor transmission line based on the directional multi-source sensing data and outputting the partial discharge fault location results. The arrival time difference of the directional multi-source sensor data is identified based on a multi-sensor time difference algorithm. Calculate the location results of multiple candidate partial discharge faults according to the arrival time difference; The partial discharge fault location results are weighted and fused according to the complementarity of the directional multi-source sensor data, and the partial discharge fault location result is output.

9. The method as described in claim 1, characterized in that, The line-attenuation fingerprint database includes multiple line-attenuation fingerprint databases of the hydropower generator under multiple operating conditions, and the multiple operating conditions include at least no-load condition, rated load condition and overload condition; The system switches between and calls upon multiple line-attenuation fingerprint databases based on the real-time operating conditions of the hydropower generator.

10. An online diagnostic system for the neutral point of a hydroelectric generator, characterized in that, The system includes: The current sensor setting module is used to set a high-frequency current sensor according to the neutral point position of the hydropower generator. The attenuation fingerprint database construction module is used to construct a multi-conductor transmission line model of the hydropower generator, extract a set of multiple partial discharge signal samples based on the multi-conductor transmission line model, extract transmission attenuation features from the set of multiple partial discharge signal samples, and construct a line-attenuation fingerprint database corresponding to the multi-conductor transmission line and multiple sets of transmission attenuation fingerprint features. The attenuation fingerprint feature matching module is used to acquire real-time partial discharge signals through the high-frequency current sensor, extract real-time transmission attenuation fingerprint features of the real-time partial discharge signals, call the line-attenuation fingerprint database to match the real-time transmission attenuation fingerprint features, and obtain matching conductor transmission lines. The partial discharge fault location module is used to collect directional multi-source sensor data based on the matching conductor transmission line, locate partial discharge faults on the matching conductor transmission line according to the directional multi-source sensor data, and output the partial discharge fault location result.

Images