A partial discharge on-line detection device based on pulse current method

Abstract

The application discloses a partial discharge on-line detection device based on a pulse current method, belongs to the technical field of partial discharge on-line detection of power distribution network equipment, and aims to solve the problems of low routine inspection sensitivity, insufficient operation efficiency, high online monitoring cost and poor environmental adaptability in the existing power distribution network intelligent operation and maintenance technology. The core is to expand the pulse current method (direct method) from offline test to on-line detection. With the help of the capacitive coupling sensor of the high-voltage on-line display device, the high-frequency pulse current is extracted through the direct connection type or the switching type test tool. The signal processing and digital conversion are completed through the data acquisition device. Then, through the multi-parameter comprehensive early warning model and the BP neuron network model, the data denoising, health early warning and defect automatic identification are realized. The device takes into account low cost and high efficiency, improves the detection accuracy and the field operation efficiency, and maximizes the performance of the power distribution equipment operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of partial discharge detection technology for power distribution network equipment, and specifically to a partial discharge detection device based on the pulse current method. Background Technology

[0002] In power systems, over 80% of power outages originate from distribution network faults. Improving the safety and reliability of the distribution network is crucial for the overall safe and economical operation of the power grid. Intelligent operation and maintenance technology is a key approach to ensuring distribution network safety, reducing operation and maintenance costs, and improving operation and maintenance efficiency. However, its practical implementation requires addressing both economic and technological constraints.

[0003] Cable distribution boxes, switchgear, ring main units, and other power distribution equipment are numerous, diverse, and have complex electrical structures. They are also widely installed and operate in harsh environments, requiring a high degree of universality in operation and maintenance technologies. At the same time, the relatively low unit cost of these devices makes it difficult to widely adopt intelligent online monitoring technologies that rely on advanced techniques.

[0004] After more than a decade of development, intelligent operation and maintenance of domestic power distribution networks has formed two main technical approaches: one is routine inspection technology based on live-line detection, and the other is online health status monitoring technology based on equipment intelligence. However, both have significant drawbacks. Routine inspection technology uses indirect detection methods such as transient ground voltage method and ultrasonic method, which suffers from low sensitivity and applicability, decreased detection effectiveness after equipment sealing is improved, low operation efficiency due to numerous detection points, and chaotic diagnostic standards. Online health status monitoring technology, on the other hand, is constrained by low equipment cost and large scale, resulting in high manufacturing cost, limited functionality, significant impact from harsh operating environments, and high security risks for remote data transmission. To address these issues, a partial discharge live-line detection device based on pulse current method is proposed. Summary of the Invention

[0005] The present invention aims to solve the problems mentioned in the background art by providing a partial discharge charging detection device based on the pulse current method.

[0006] The specific technical solution is as follows:

[0007] A partial discharge energized detection device based on pulse current method includes a testing fixture, a data acquisition device, and a data processing device. One end of the testing fixture is connected to the exposed metal part of the phase detection hole or signal lead of a high-voltage energized display device to extract the high-frequency pulse current generated by partial discharge of the power distribution equipment. The other end of the testing fixture is connected to the input port of the data acquisition device, which performs impedance matching, signal transformation, filtering, amplification, and detection processing on the high-frequency pulse current, and converts it into digital pulse current information via a high-speed acquisition circuit. The data acquisition device is connected to the data processing device through a communication interface, and the data processing device performs noise reduction, characteristic parameter calculation, health status early warning, and automatic defect type identification on the digital pulse current information.

[0008] As a preferred embodiment of the present invention, the testing fixture includes a direct-connection testing fixture and an adapter testing fixture. The direct-connection testing fixture and the adapter testing fixture are selected for use in conjunction with the high-voltage live display device and the data acquisition device to adapt to high-voltage live display devices of different models and structures.

[0009] As a preferred embodiment of the present invention, the direct-connection test fixture is a branched cable structure. The first branch end of the branched cable is provided with a combination terminal that can be adapted to different sizes of phase holes, or an expandable length probe for contacting the exposed metal part of the signal lead; the second branch end of the branched cable is provided with a grounding terminal for connecting to the exposed metal grounding part of the power distribution equipment; the common end of the branched cable is connected to the input port of the data acquisition device.

[0010] As a preferred embodiment of the present invention, the adapter-type test fixture includes a test cable and a current sensing unit. One end of the test cable is provided with a combined terminal, and the other end is provided with a grounding terminal. The current sensing unit is assembled on the test cable and is a radio frequency magnetic field sensor used to couple the high-frequency magnetic field generated by the high-frequency current in the test cable and convert it into a high-frequency current output. The current sensing unit can be deployed on the three-phase signal lead to realize phase detection separately, or deployed on the grounding lead to realize simultaneous measurement of the three phases, and the current sensing unit has a closed or open structure.

[0011] In a preferred embodiment of the present invention, the data acquisition device includes an impedance matching / signal conversion module, a filtering module, an amplification module, a detection module, and a high-speed acquisition module. The input terminal of the impedance matching / signal conversion module is connected to the test fixture, and the output terminal is connected to the high-speed acquisition module in sequence via the filtering module, the amplification module, and the detection module. The impedance matching / signal conversion module separates the high-frequency pulse current signal from the power frequency current signal. The high-frequency pulse current signal is transmitted to the high-speed acquisition module after being processed by filtering, amplification, and detection in sequence. The power frequency current signal is transmitted to the phase synchronization circuit and shaped into a standard square wave signal for extracting the phase information of the partial discharge signal.

[0012] As a preferred embodiment of the present invention, the impedance matching / signal conversion module adopts an active circuit structure and uses the concept of virtual short to reduce the impedance of the input circuit to near zero, so as to reduce the influence of the impedance of the high-voltage live display device on the partial discharge measurement.

[0013] In a preferred embodiment of the present invention, the data processing device includes a tablet computer and professional application software running thereon. The professional application software includes a detection control and data display module, a data recording and query module, a diagnostic analysis module, and a typical case module. The detection control and data display module is used to display phase-resolved pulse sequence information and perform diagnostic analysis and health status early warning. The data recording and query module is used to store and query detection data and export it to a typical case library. The diagnostic analysis module is used to perform data noise reduction, feature parameter calculation, and defect type identification. The typical case module is used to classify, manage, and replay typical discharge defect detection data.

[0014] As a preferred embodiment of the present invention, the diagnostic analysis module includes a comprehensive early warning model for partial discharge status. The input parameters of the comprehensive early warning model for partial discharge status include probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient. The probability intensity is the maximum discharge pulse value when the cumulative probability in the detection data sample reaches a specified limit. The average intensity is the ratio of the sum of effective discharge pulse intensities to the total number of pulses within a selected time interval. The discharge frequency is the ratio of the total number of effective discharge pulses within the selected time interval to the time interval itself. The background noise is the average value of the signal level in the measurement channel. The power frequency correlation coefficient is calculated using discrete Fourier transform and is used to reflect the periodic characteristics of discharge activity. The comprehensive early warning model for partial discharge status outputs a health index by weighting the input parameters and performing activation function operations. The calculation expression for the health index is: , where HI0 is the initial health index, Wᵢ is the weighting coefficient, Kᵢ is the input parameter, and b is the bias coefficient.

[0015] As a preferred embodiment of the present invention, the diagnostic analysis module further includes an automatic discharge defect identification model, which is a BP neural network model with one hidden layer. The basic data source of the model is phase-resolved pulse sequence information, and the derived data sources include Qmax-Phase maps, N-Phase maps, S-Phase maps, and NQ maps. The input parameters of the model are 26 statistical feature parameters, which are extracted based on the derived maps and include total number of discharges, quadrant discharge frequency ratio, map symmetry, skewness, kurtosis, iso-phase center, and Weibull distribution fitting parameters. The first layer of neurons in the BP neural network model integrates the statistical parameters and outputs a logical quantity, and the second layer of neurons combines the logical output of the first layer to output a logical quantity that matches the target defect type.

[0016] As a preferred embodiment of the present invention, the high-speed acquisition module adopts a small-capacity storage space that can be read and written repeatedly, and is combined with analog or digital watchdog technology. When the input signal exceeds the threshold, the digital information of the discharge pulse is automatically transferred. At the same time, the phase timer count information is read to obtain the phase information of the discharge pulse, thereby forming phase-resolved pulse sequence information.

[0017] As a preferred embodiment of the present invention, the combined terminal is a combined banana head, which can be matched with adapters of different specifications and sizes to adapt to high-voltage live display devices with different core phase hole sizes; the grounding terminal is an alligator clip, which is used for detachable connection to the exposed metal grounding parts of high-voltage switchgear and cable branch boxes.

[0018] As a preferred embodiment of the present invention, for a high-voltage live display device without a core phase hole, the current sensing unit adopts a fixed installation method and is combined with the input signal lead of the high-voltage live display device to form a fixed installation adapter test fixture, so that detection can be achieved without disconnecting the input signal lead.

[0019] As a preferred embodiment of the present invention, the partial discharge state integrated early warning model optimizes the diagnostic accuracy through five derived parameters: intensity level, significance level, signal-to-noise ratio, repetition rate, and power frequency correlation coefficient; the intensity level is the ratio of probability intensity to evaluation benchmark, the significance level is the ratio of probability intensity to average intensity, the signal-to-noise ratio is the ratio of average intensity to background noise, and the repetition rate is the ratio of discharge frequency to benchmark frequency.

[0020] As a preferred embodiment of the present invention, the communication interface is a high-speed communication interface, used to transmit the digitized pulse current information converted by the data acquisition device to the data processing device in real time. The high-speed communication interface supports the complete transmission of phase-resolved pulse sequence information.

[0021] As a preferred embodiment of the present invention, the Qmax-Phase spectrum is a spectrum showing the relationship between the maximum discharge intensity and the discharge phase in a specified phase interval per unit time; the N-Phase spectrum is a spectrum showing the relationship between the number of discharges and the discharge phase in a specified phase interval per unit time; the S-Phase spectrum is a spectrum showing the relationship between the cumulative sum of discharge pulse amplitudes in a specified phase interval per unit time and the discharge phase; and the NQ spectrum is a spectrum showing the relationship between the total number or percentage of discharges in a specified amplitude interval and the discharge intensity.

[0022] The present invention has the following beneficial effects:

[0023] 1. The test fixture offers two options: direct connection and adapter. It is equipped with combination terminals, expandable probes, and current sensing units with different deployment methods. It can be adapted to high-voltage live display devices of different models and structures (with / without core phase holes), which greatly improves the universality of the device and lowers the application threshold in diverse power distribution equipment scenarios.

[0024] 2. The data acquisition device adopts an impedance matching / signal conversion module with active circuit design to minimize the impact of the impedance of the high-voltage live display device on the measurement. At the same time, it achieves effective separation of high-frequency pulse signals and power frequency signals, reduces signal interference, ensures the accuracy of digital data, and provides a high-quality foundation for subsequent analysis.

[0025] 3. The comprehensive early warning model for partial discharge status abandons the traditional single threshold judgment method and integrates multi-dimensional feature parameters to effectively suppress the influence of isolated interference pulses, power electronic switch interference and background noise, significantly reduce the probability of false alarms, improve the reliability of health status early warning, and provide a scientific basis for operation and maintenance decisions.

[0026] 4. The automatic discharge defect identification model introduces S-Phase spectrum and two-parameter Weibull distribution parameters to capture partial discharge characteristics from multiple dimensions such as discharge intensity, number, amplitude summation, and distribution pattern. This breaks through the limitation of traditional models in not fully characterizing features and improves the accuracy and comprehensiveness of defect type identification.

[0027] 5. The entire device extends the pulse current method from offline to live detection, eliminating the need for large-scale installation of intelligent online monitoring equipment. While controlling costs, it significantly improves on-site detection efficiency by simplifying detection operations and enhancing data processing efficiency. It balances low cost and high efficiency, effectively solving the problems of low sensitivity and insufficient efficiency of traditional routine inspection techniques, as well as high cost and poor environmental adaptability of online monitoring techniques, thereby maximizing the performance of power distribution equipment operation and maintenance. Attached Figure Description

[0028] Figure 1The equivalent circuit diagram of the direct method for partial discharge detection shows the connection relationship of the three-capacitor model (equivalent capacitor Cg at the defect, capacitor Cb of the intact insulation part, and coupling capacitor Cc), the detection impedance Zm, and the shunting path of the high-frequency pulse current. It is used to illustrate the core principle of the direct method for capturing partial discharge pulse signals through the detection impedance or high-frequency current transformer.

[0029] Figure 2 The block diagram of the high-voltage live display device shows the parallel connection relationship of the high-voltage capacitor CH, the low-voltage capacitor CL, the terminal blocks a to c, the indicators La to c, and the core phase holes Ha to c. It is used to explain the working mechanism of the capacitively coupled sensor of the high-voltage live display device and the signal output characteristics of the core phase holes.

[0030] Figure 3 The schematic diagram of the principle of testing partial discharge current through the high-voltage live display phase hole shows the low-impedance connection branch Zm added between the phase hole Ha~c and the ground circuit, as well as the transmission path of the high-frequency pulse current Ia1 flowing into the detector. It is used to illustrate the wiring logic of the present invention to realize the live detection of partial discharge by using the high-voltage live display device.

[0031] Figure 4 The main components of the device of this invention are shown in the block diagram, which illustrates the sequential connection relationship of the test fixture, the data acquisition device, and the data processing device, as well as the core functional direction of each device, so as to clearly present the overall architecture of the device and the signal flow path.

[0032] Figure 5 : A schematic diagram of the direct-connection test fixture structure, showing the structure of the branched cable, combined banana head (including adapter C), alligator clip and expandable probe, used to illustrate the specific implementation of the direct-connection fixture to adapt to different core phase aperture sizes and equipment without core phase apertures;

[0033] Figure 6 : Schematic diagram of the adapter test fixture structure, showing the assembly relationship of test cables, combined banana plugs, alligator clips, RF terminals and current sensing unit (CSD), used to illustrate the structural design of the adapter fixture to couple high-frequency signals through RF magnetic field sensor;

[0034] Figure 7 The diagram shows the internal module connection of the data acquisition device, illustrating the connection sequence of the impedance matching / signal conversion module, filtering module, amplification module, detection module, high-speed acquisition module, and phase synchronization circuit. This diagram is used to explain the digital processing flow of high-frequency pulse current.

[0035] Figure 8 Example of a phase-resolved pulse sequence (PRPS) map, with the horizontal axis representing the power frequency phase, the vertical axis representing the discharge pulse amplitude, and the point density representing the discharge frequency, used to demonstrate the amplitude-phase-frequency distribution characteristics of partial discharge pulses;

[0036] Figure 9 The diagram shows the architecture of the comprehensive early warning model for partial discharge status, illustrating the input paths for probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient, as well as the logical flow from weighted processing and activation function calculation to health index output.

[0037] Figure 10 The diagram shows the architecture of the BP neural network automatic defect identification model, illustrating the neuron connections in the input layer (26 statistical parameters), hidden layer, and output layer, as well as the signal processing logic of each layer, to explain the network implementation principle of defect type identification.

[0038] Figure 11 A schematic diagram of the current sensing unit deployment shows two forms: one deployed on the three-phase signal lead (phase-by-phase detection) and the other on the ground lead (simultaneous measurement of the three phases), as well as the structural differences between the closed and open types.

[0039] Figure 12 Example of a multidimensional derived graph set, including coordinate definitions and typical forms of Qmax-Phase, N-Phase, S-Phase, and NQ graphs, used to illustrate the specific representation of the data source required by the defect identification model. Detailed Implementation

[0040] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0041] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual images. They should not be construed as limiting the scope of this application. To better illustrate the embodiments of the present invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0042] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present application. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0043] In the description of this invention, unless otherwise explicitly specified and limited, the term "connection" or similar designation indicating a connection between components should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0044] Example

[0045] The partial discharge charge detection device based on the pulse current method provided in this embodiment, such as... Figures 1-12 As shown, the system includes: a testing fixture, a data acquisition device, and a data processing device. One end of the testing fixture is connected to the phase hole or exposed metal part of the signal lead of the high-voltage live display device to extract the high-frequency pulse current generated by partial discharge of the power distribution equipment. The other end of the testing fixture is connected to the input port of the data acquisition device, which performs impedance matching, signal transformation, filtering, amplification, and detection processing on the high-frequency pulse current, and converts it into digital pulse current information through a high-speed acquisition circuit. The data acquisition device is connected to the data processing device through a communication interface, which performs noise reduction, characteristic parameter calculation, health status early warning, and automatic defect type identification on the digital pulse current information.

[0046] By coordinating testing fixtures, data acquisition devices, and data processing devices, the pulse current method is extended to live-line testing scenarios for power distribution equipment. This effectively solves the problems of low sensitivity and applicability, and insufficient operational efficiency of traditional routine inspection techniques, while avoiding the drawbacks of high cost and environmental constraints of online health status monitoring technologies. It achieves accurate extraction, digital conversion, and in-depth analysis of partial discharge signals, balancing detection cost and efficiency. It comprehensively covers data noise reduction, characteristic parameter calculation, health early warning, and defect identification functions, significantly improving the overall performance of power distribution equipment operation and maintenance.

[0047] Specifically, in this embodiment, the testing fixture includes a direct-connection testing fixture and an adapter testing fixture. Either the direct-connection or adapter testing fixture can be used in conjunction with the high-voltage live-line display device and the data acquisition device to adapt to different models and structures of high-voltage live-line display devices. By providing both direct-connection and adapter testing fixtures, the testing device can be adapted to different models and structures of high-voltage live-line display devices, eliminating the need for separate fixture design for specific equipment. This significantly improves the device's versatility, lowers the application threshold in diverse power distribution equipment scenarios, and enhances the flexibility of on-site testing.

[0048] Specifically, in this embodiment, the direct-connection test fixture has a branched cable structure. The first branch of the branched cable is equipped with a combination terminal that can adapt to different sizes of phase-connection apertures, or an expandable probe for contacting the exposed metal part of the signal lead. The second branch of the branched cable is equipped with a grounding terminal for connecting to the exposed metal grounding part of the power distribution equipment. The common end of the branched cable is connected to the input port of the data acquisition device. The branched cable structure of the direct-connection test fixture, combined with the combination terminal and expandable probe, can flexibly adapt to high-voltage live display devices with different sizes of phase-connection apertures and those without phase-connection apertures. The connection method between the grounding terminal and the exposed metal grounding part of the power distribution equipment is simple and reliable. This simplifies the on-site connection operation, allows signal extraction without complex equipment modifications, ensures the stability of high-frequency pulse current transmission, and improves the convenience of the testing operation.

[0049] Specifically, in this embodiment, the adapter-type test fixture includes a test cable and a current sensing unit. One end of the test cable has a combined terminal, and the other end has a grounding terminal. The current sensing unit is assembled on the test cable and is a radio frequency magnetic field sensor used to couple the high-frequency magnetic field generated by the high-frequency current in the test cable and convert it into a high-frequency current output. The current sensing unit can be deployed on the three-phase signal leads to achieve phase-by-phase detection, or deployed on the grounding leads to achieve simultaneous three-phase measurement. The current sensing unit has a closed or open / closed structure. The current sensing unit of the adapter-type test fixture uses a radio frequency magnetic field sensor, which achieves signal conversion by coupling a high-frequency magnetic field. There is no direct electrical coupling, which effectively improves the anti-interference capability. Its design, which allows for phase-by-phase deployment or simultaneous three-phase measurement, as well as the choice of closed or open / closed structure, can meet different detection requirements. At the same time, it is not necessary to disconnect the input signal lead of the high-voltage live display device, avoiding the safety hazards caused by removing and restoring the lead, and ensuring detection accuracy and operational safety.

[0050] Specifically, in this embodiment, the data acquisition device includes an impedance matching / signal conversion module, a filtering module, an amplification module, a detection module, and a high-speed acquisition module. The input of the impedance matching / signal conversion module is connected to the test fixture, and its output is connected to the high-speed acquisition module sequentially via the filtering module, amplification module, and detection module. The impedance matching / signal conversion module separates the high-frequency pulse current signal from the power frequency current signal. The high-frequency pulse current signal is transmitted to the high-speed acquisition module after being filtered, amplified, and detected sequentially. The power frequency current signal is transmitted to the phase synchronization circuit and shaped into a standard square wave signal for extracting the phase information of the partial discharge signal. Through the orderly cooperation of each module, the data acquisition device effectively separates the high-frequency pulse current signal from the power frequency current signal, and performs targeted processing on the two types of signals respectively. This ensures the accurate extraction of the partial discharge signal intensity information and provides a stable standard square wave signal for phase information acquisition, reducing mutual interference between different signals and providing high-quality digital foundation data for subsequent data processing.

[0051] Specifically, in this embodiment, the impedance matching / signal conversion module adopts an active circuit structure and utilizes the concept of virtual short to reduce the impedance of the input circuit to near zero, thereby reducing the influence of the impedance of the high-voltage live display device on the partial discharge measurement. By employing an active circuit and utilizing the concept of virtual short to reduce the input circuit impedance to near zero, the impedance of the high-voltage live display device itself is minimized, reducing attenuation and distortion during signal transmission, ensuring the accuracy of partial discharge signal extraction, and improving the reliability of the measurement data.

[0052] Specifically, in this embodiment, the data processing device includes a tablet computer and professional application software running on it. The professional application software includes a detection control and data display module, a data recording and query module, a diagnostic analysis module, and a typical case module. The detection control and data display module is used to display phase-resolved pulse sequence information and perform diagnostic analysis and health status warnings. The data recording and query module is used to store and query detection data and export it to a typical case library. The diagnostic analysis module is used to perform data noise reduction, feature parameter calculation, and defect type identification. The typical case module is used to classify, manage, and replay typical discharge defect detection data. Using a tablet computer as the hardware carrier, coupled with fully functional professional application software, the data processing device leverages its powerful computing, storage, and display capabilities to achieve multi-dimensional management and analysis of detection data. The detection control and data display module facilitates real-time viewing and warnings; the data recording and query module supports traceability of detection results and case accumulation; the diagnostic analysis module enables in-depth data mining; and the typical case module helps improve the capabilities of testing personnel, comprehensively enhancing the intelligence level and operational efficiency of on-site testing.

[0053] Specifically, in this embodiment, the diagnostic analysis module includes a comprehensive early warning model for partial discharge status. The input parameters of this model include probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient. Probability intensity is the maximum discharge pulse value when the cumulative probability in the detection data sample reaches a specified limit. Average intensity is the ratio of the sum of effective discharge pulse intensities to the total number of pulses within a selected time interval. Discharge frequency is the ratio of the total number of effective discharge pulses within a selected time interval to that time interval. Background noise is the average value of the signal level in the measurement channel. The power frequency correlation coefficient is calculated using discrete Fourier transform and is used to reflect the periodic characteristics of discharge activity. The comprehensive early warning model for partial discharge status outputs a health index by weighting the input parameters and performing activation function operations. The expression for calculating the health index is: Where HI0 is the initial health index, Wᵢ is the weighting coefficient, Kᵢ is the input parameter, and b is the bias coefficient. The comprehensive early warning model for partial discharge status abandons the drawbacks of traditional simple threshold models. By introducing multi-dimensional input parameters such as probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient, and combining weighted processing and activation function calculations, the health index is obtained. This effectively suppresses the impact of isolated interference pulses, power electronic device switching interference, and background noise on the early warning results, significantly reducing the probability of false alarms, improving the accuracy and reliability of health status early warning, and providing a more scientific decision-making basis for equipment operation and maintenance.

[0054] Specifically, in this embodiment, the diagnostic analysis module also includes an automatic discharge defect identification model, which is a BP neural network model with one hidden layer. The basic data source of the model is phase-resolved pulse sequence information, and the derived data sources include Qmax-Phase maps, N-Phase maps, S-Phase maps, and NQ maps. The model's input parameters are 26 statistical feature parameters, which are extracted based on the derived maps, including total number of discharges, quadrant discharge frequency ratio, map symmetry, skewness, kurtosis, iso-phase center, and Weibull distribution fitting parameters. The first layer of neurons in the BP neural network model synthesizes the statistical parameters and outputs a logical quantity, and the second layer of neurons combines the logical output of the first layer to output a logical quantity that matches the target defect type. The automatic discharge defect identification model is based on a BP neural network architecture, uses multi-dimensional derived maps as the data source, and extracts 26 comprehensive statistical feature parameters, including newly added S-Phase map-related parameters and two-parameter Weibull distribution parameters. It can more comprehensively capture the statistical characteristics and amplitude distribution patterns of partial discharge activities, breaking through the limitations of traditional identification models in terms of incomplete feature characterization, significantly improving the accuracy and comprehensiveness of defect type identification, and helping inspection personnel to quickly and accurately determine the type of equipment defects.

[0055] The mathematical expression for the neural network model can be described as follows:

[0056] ;

[0057] ;

[0058] in:

[0059] W i (Weighting coefficients): These are coefficients used in the integrated early warning model for partial discharge status to adjust the importance of each input parameter. They need to be determined in advance through training according to the evaluation rules to ensure the accuracy of the model output.

[0060] X i The input parameters of the BP neural network model, namely the 26 statistical feature parameters extracted based on various derived graphs, are the basic data of the defect identification model.

[0061] Y i The intermediate output value of the first layer of the BP neural network model is determined by the input parameter X. i The product of the weight coefficient Wi and the bias coefficient Bi is added to calculate the result, which is then passed to the hidden layer for further processing.

[0062] B i The bias coefficients of the first layer in a BP neural network model are used to correct the computational results of the first layer and ensure the intermediate output value Y. i The rationality of it.

[0063] X h The input values ​​of the hidden layers in the BP neuron network model (i.e., the output values ​​of the first layer) are obtained by the activation function σ from the intermediate output value Yi and are used to pass them to the second layer for combination processing.

[0064] Y h The intermediate output value of the second layer of the BP neural network model is calculated by summing the product of the hidden layer input value Xh and the corresponding weight coefficient Wh, and adding the bias coefficient Bh. It is used to pass to the output layer to generate the final result.

[0065] B h The bias coefficients of the second layer in the BP neural network model are used to correct the calculation results of the second layer and ensure the rationality of the intermediate output value Yh.

[0066] W h The weights of the second layer in a backpropagation (BP) neural network model are used to adjust the input values ​​X of the hidden layers. h The importance of this factor is to ensure the model's accuracy in identifying defect types.

[0067] X o The output value of the BP neural network model is determined by the intermediate output value Y. hThe result obtained by the activation function σ is a logical quantity that matches the target defect type and is used to directly output the defect identification result.

[0068] σ (activation function): A non-linear transformation function in the BP neural network model, used to process the intermediate output value Y. i Y h The calculations are performed to achieve nonlinear mapping of parameters, which helps the model accurately identify defect types.

[0069] Specifically, in this embodiment, the high-speed acquisition module employs a small-capacity storage space with cyclic read / write capability, combined with analog or digital watchdog technology. When the input signal exceeds a threshold, it triggers the automatic transfer of digitized discharge pulse information, while simultaneously reading phase timer count information to obtain the phase information of the discharge pulse, thereby forming phase-resolved pulse sequence information. The high-speed acquisition module's use of a small-capacity storage space with cyclic read / write capability, combined with analog or digital watchdog technology, enables accurate capture of short-duration discharge pulse signals without relying on large-capacity storage hardware. The threshold-triggered automatic transfer function ensures the complete acquisition of discharge pulse amplitude information, while simultaneously reading phase timer count information to ensure no loss of phase information. This approach controls hardware costs while guaranteeing data integrity and validity.

[0070] Specifically, in this embodiment, the combined terminal is a combined banana plug, which can be used with adapters of different sizes to adapt to high-voltage live display devices with different phase hole sizes; the grounding terminal is an alligator clip, used for detachable connection to the exposed metal grounding parts of high-voltage switchgear and cable branch boxes. The combined banana plug can be used with adapters of different sizes to adapt to various phase hole sizes, and the alligator clip can securely connect to the exposed metal grounding parts of power distribution equipment. This enhances the compatibility range of direct-connection test fixtures and equipment, ensures connection reliability, simplifies the on-site plugging and unplugging operation process, reduces connection time, and further improves the efficiency of on-site testing.

[0071] Specifically, in this embodiment, for high-voltage live-line display devices without a core-hole, the current sensing unit adopts a fixed installation method, combining with the input signal leads of the high-voltage live-line display device to form a fixed-installation adapter test fixture, enabling detection without disconnecting the input signal leads. For high-voltage live-line display devices without a core-hole, the fixed-installation current sensing unit combined with the input signal leads to form a fixture allows for detection without disconnecting the signal leads. This avoids potential safety hazards caused by removing and restoring the leads, simplifies the testing process, reduces operational difficulty, and ensures the stability and accuracy of the detection, expanding the device's application capabilities in scenarios involving devices without a core-hole.

[0072] Specifically, in this embodiment, the comprehensive early warning model for partial discharge status optimizes diagnostic accuracy through five derived parameters: intensity level, significance level, signal-to-noise ratio (SNR), repetition rate, and power frequency correlation coefficient. The intensity level is the ratio of probability intensity to the evaluation benchmark; the significance level is the ratio of probability intensity to average intensity; the SNR is the ratio of average intensity to background noise; and the repetition rate is the ratio of discharge frequency to the benchmark frequency. By introducing these five derived parameters, the diagnostic logic of the comprehensive early warning model for partial discharge status is optimized. This enables more accurate differentiation between partial discharge signals and various interference signals, further enhancing the model's ability to identify signals in different scenarios, making the health status early warning results more targeted and reliable.

[0073] Specifically, in this embodiment, the communication interface is a high-speed communication interface used to transmit the digitized pulse current information converted by the data acquisition device to the data processing device in real time. The high-speed communication interface supports the complete transmission of phase-resolved pulse sequence information. The high-speed communication interface ensures the real-time and complete transmission of digitized pulse current information between the data acquisition device and the data processing device, especially the lossless transmission of phase-resolved pulse sequence information. This provides timely data support for the real-time analysis, diagnosis, and early warning of the data processing device, avoiding the impact of data transmission delays or losses on the detection results and improving the real-time performance of the entire detection process.

[0074] Specifically, in this embodiment, the Qmax-Phase spectrum is the relationship between the maximum discharge intensity and the discharge phase within a specified phase interval per unit time; the N-Phase spectrum is the relationship between the number of discharges and the discharge phase within a specified phase interval per unit time; the S-Phase spectrum is the relationship between the cumulative sum of discharge pulse amplitudes and the discharge phase within a specified phase interval per unit time; and the NQ spectrum is the relationship between the total number or percentage of discharges within a specified amplitude interval and the discharge intensity. The definitions of the four derived spectrums—Qmax-Phase, N-Phase, S-Phase, and NQ—are clarified, providing a multi-dimensional and comprehensive data source for the automatic discharge defect identification model. This enables the model to capture partial discharge characteristics from multiple dimensions, such as maximum discharge intensity, number of discharges, cumulative amplitude sum, and discharge intensity distribution, providing a rich data foundation for statistical parameter extraction and further supporting the accuracy of defect type identification.

[0075] In summary, the partial discharge charge detection device based on the pulse current method provided in this embodiment has the following advantages:

[0076] 1. The test fixture offers two options: direct connection and adapter. It is equipped with combination terminals, expandable probes, and current sensing units with different deployment methods. It can be adapted to high-voltage live display devices of different models and structures (with / without core phase holes), which greatly improves the universality of the device and lowers the application threshold in diverse power distribution equipment scenarios.

[0077] 2. The data acquisition device adopts an impedance matching / signal conversion module with active circuit design to minimize the impact of the impedance of the high-voltage live display device on the measurement. At the same time, it achieves effective separation of high-frequency pulse signals and power frequency signals, reduces signal interference, ensures the accuracy of digital data, and provides a high-quality foundation for subsequent analysis.

[0078] 3. The comprehensive early warning model for partial discharge status abandons the traditional single threshold judgment method and integrates multi-dimensional feature parameters to effectively suppress the influence of isolated interference pulses, power electronic switch interference and background noise, significantly reduce the probability of false alarms, improve the reliability of health status early warning, and provide a scientific basis for operation and maintenance decisions.

[0079] 4. The automatic discharge defect identification model introduces S-Phase spectrum and two-parameter Weibull distribution parameters to capture partial discharge characteristics from multiple dimensions such as discharge intensity, number, amplitude summation, and distribution pattern. This breaks through the limitation of traditional models in not fully characterizing features and improves the accuracy and comprehensiveness of defect type identification.

[0080] 5. The entire device extends the pulse current method from offline to live detection, eliminating the need for large-scale installation of intelligent online monitoring equipment. While controlling costs, it significantly improves on-site detection efficiency by simplifying detection operations and enhancing data processing efficiency. It balances low cost and high efficiency, effectively solving the problems of low sensitivity and insufficient efficiency of traditional routine inspection techniques, as well as high cost and poor environmental adaptability of online monitoring techniques, thereby maximizing the performance of power distribution equipment operation and maintenance.

[0081] Working principle:

[0082] The main logic of this device is to extend the pulse current method (direct method) traditionally used for offline high-voltage testing to the scenario of live detection of power distribution equipment. By using the capacitively coupled sensor of the existing high-voltage live display device, the device can accurately extract and analyze partial discharge signals.

[0083] 1. Signal Extraction Principle

[0084] The high-voltage live display device is essentially a special type of high-voltage capacitor, with its core phase hole connected in parallel with internal low-voltage capacitors and indicators. This device adds a low-impedance connection branch between the core phase hole and the ground circuit. When partial discharge occurs inside the power distribution equipment, a high-frequency pulse current is generated. Under high-frequency conditions, most of this current flows into the detection equipment through the testing fixture, avoiding the installation and deployment difficulties of capacitively coupled sensors in traditional direct methods.

[0085] 2. Core Component Working Logic

[0086] Test fixtures are available in direct-connect and adapter types, used to extract high-frequency pulse current. The direct-connect type uses a branched cable, with one end connected to the phase detection aperture via a combination terminal or contacting the exposed signal area via a probe, and the other end grounded, directly transmitting the signal to the data acquisition device. The adapter type uses a test cable paired with an RF magnetic field sensor (current sensing unit) to couple a high-frequency magnetic field into a high-frequency current output. No direct electrical connection is required; it can be deployed in phases or measured in three phases simultaneously, adapting to different equipment structures.

[0087] Data acquisition device: First, the high-frequency pulse current signal and the power frequency current signal are separated through an impedance matching / signal conversion module. The high-frequency pulse signal is filtered, amplified, and detected, and then converted into digital information by a high-speed acquisition circuit. The power frequency signal is shaped into a standard square wave for extracting phase information. The impedance matching module uses an active circuit, leveraging the concept of virtual short to reduce the input circuit impedance to near zero, minimizing the impact of the high-voltage live display device's impedance on the measurement. Simultaneously, through a small-capacity, cyclically readable and writable storage space and watchdog technology, the amplitude and phase information of the discharge pulse are dynamically captured, forming phase-resolved pulse sequence information (PRPS).

[0088] Data processing device: Based on a ruggedized tablet and professional application software, it performs in-depth processing of PRPS information. Through a comprehensive early warning model for partial discharge status, it combines five types of parameters—probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient—and outputs a health index through weighted calculations and activation functions. Through an automatic discharge defect identification model (containing a BP neural network with one hidden layer), it uses PRPS and its derived Qmax-Phase, N-Phase, S-Phase, and NQ maps as data sources to extract 26 statistical feature parameters, achieving automatic defect type identification.

[0089] How to use:

[0090] 1. Tooling selection and connection: Select direct-connection or adapter-type test tooling according to the model and structure of the high-voltage live display device.

[0091] 1.1 When using a direct connection, connect the combined terminal of the branched cable to the phase-connection port (compatible with different sized phase-connection ports), or contact the exposed signal part of the device without a phase-connection port using an extension probe. Connect the other end to the exposed metal grounding part of the device using an alligator clip, and connect the common end to the input port of the data acquisition device.

[0092] 1.2 When using the adapter type, connect the combined terminals of the test cable to the phase detection hole, ground the ground terminal, and deploy the current sensing unit on the three-phase signal lead (phase detection) or ground lead (simultaneous measurement of three phases) according to the testing requirements. The output end is connected to the data acquisition device through the radio frequency terminal. For equipment without a phase detection hole, a fixed current sensing unit can be used, which can be connected without disconnecting the signal lead.

[0093] 2. Data Acquisition Startup: The data acquisition device is turned on, and it automatically completes impedance matching, signal separation, filtering, amplification, detection and high-speed digital conversion, captures the amplitude and phase information of the discharge pulse, forms PRPS, and transmits it to the data processing device through a high-speed communication interface.

[0094] 3. Data Processing and Result Application: Follow-up work will be carried out using specialized application software from the data processing equipment.

[0095] 3.1 Use the detection, control, and data display module to view PRPS information and health status warning results;

[0096] 3.2 Utilize the data recording and query module to store the test data, and export it to the typical case library when necessary;

[0097] 3.3 The diagnostic analysis module performs data noise reduction, feature parameter calculation, and defect type identification.

[0098] 3.4 Refer to the categorized data in the typical case module to assist in quickly judging the on-site test results.

[0099] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A partial discharge charge detection device based on the pulse current method, characterized in that, The device includes a testing fixture, a data acquisition device, and a data processing device. One end of the testing fixture is connected to the phase detection port or exposed metal part of the signal lead of the high-voltage live display device to extract the high-frequency pulse current generated by partial discharge of the power distribution equipment. The other end of the testing fixture is connected to the input port of the data acquisition device, which performs impedance matching, signal transformation, filtering, amplification, and detection processing on the high-frequency pulse current, and converts it into digital pulse current information through a high-speed acquisition circuit. The data acquisition device is connected to the data processing device through a communication interface, which performs noise reduction, characteristic parameter calculation, health status early warning, and automatic defect type identification on the digital pulse current information.

2. The partial discharge charge detection device based on pulse current method according to claim 1, characterized in that, The testing fixture includes a direct-connection testing fixture and an adapter testing fixture. One of the direct-connection testing fixture and the adapter testing fixture can be used in conjunction with the high-voltage live display device and the data acquisition device to adapt to high-voltage live display devices of different models and structures. The direct-connection test fixture has a branched cable structure. The first branch end of the branched cable is provided with a combination terminal that can be adapted to different sizes of phase holes, or an expandable length probe for contacting the exposed metal part of the signal lead. The second branch end of the branched cable is provided with a grounding terminal for connecting to the exposed metal grounding part of the power distribution equipment. The common end of the branched cable is connected to the input port of the data acquisition device. The adapter-type test fixture includes a test cable and a current sensing unit. One end of the test cable is provided with a combined terminal, and the other end is provided with a grounding terminal. The current sensing unit is assembled on the test cable and is a radio frequency magnetic field sensor used to couple the high-frequency magnetic field generated by the high-frequency current in the test cable and convert it into a high-frequency current output. The current sensing unit can be deployed on the three-phase signal lead to realize phase detection, or deployed on the grounding lead to realize simultaneous three-phase measurement, and the current sensing unit has a closed or open structure.

3. The partial discharge charge detection device based on pulse current method according to claim 1, characterized in that, The data acquisition device includes an impedance matching / signal conversion module, a filtering module, an amplification module, a detection module, and a high-speed acquisition module. The input of the impedance matching / signal conversion module is connected to the test fixture, and the output is connected to the high-speed acquisition module in sequence via the filtering module, the amplification module, and the detection module. The impedance matching / signal conversion module separates the high-frequency pulse current signal from the power frequency current signal. The high-frequency pulse current signal is transmitted to the high-speed acquisition module after being filtered, amplified, and detected in sequence. The power frequency current signal is transmitted to the phase synchronization circuit and shaped into a standard square wave signal for extracting the phase information of the partial discharge signal. The impedance matching / signal conversion module adopts an active circuit structure and uses the concept of virtual short to reduce the impedance of the input circuit to near zero, thereby reducing the influence of the impedance of the high-voltage live display device on the partial discharge measurement.

4. The partial discharge charge detection device based on pulse current method according to claim 1, characterized in that, The data processing device includes a tablet computer and professional application software running on it. The professional application software includes a detection control and data display module, a data recording and query module, a diagnostic analysis module, and a typical case module. The detection control and data display module is used to display phase-resolved pulse sequence information and perform diagnostic analysis and health status early warning. The data recording and query module is used to store and query detection data and export it to a typical case library. The diagnostic analysis module is used to perform data noise reduction, feature parameter calculation, and defect type identification. The typical case module is used to classify, manage, and replay typical discharge defect detection data.

5. The partial discharge charge detection device based on pulse current method according to claim 4, characterized in that, The diagnostic analysis module includes a comprehensive early warning model for partial discharge status. The input parameters of the comprehensive early warning model for partial discharge status include probability intensity, average intensity, discharge frequency, background noise, and power frequency correlation coefficient. The probability intensity is the maximum discharge pulse value when the cumulative probability in the detection data sample reaches a specified limit; the average intensity is the ratio of the sum of effective discharge pulse intensities to the total number of pulses within a selected time interval; the discharge frequency is the ratio of the total number of effective discharge pulses within a selected time interval to that time interval; the background noise is the average value of the signal level of the measurement channel; the power frequency correlation coefficient is calculated through discrete Fourier transform and is used to reflect the periodic characteristics of discharge activity; the partial discharge state comprehensive early warning model outputs a health index by weighting the input parameters and performing activation function operations, and the calculation expression of the health index is: , where HI0 is the initial health index, Wᵢ is the weighting coefficient, Kᵢ is the input parameter, and b is the bias coefficient.

6. The partial discharge charged detection device based on pulse current method according to claim 5, characterized in that, The diagnostic analysis module also includes an automatic discharge defect identification model, which is a BP neural network model with one hidden layer. The basic data source of the model is phase-resolved pulse sequence information, and the derived data sources include Qmax-Phase maps, N-Phase maps, S-Phase maps, and NQ maps. The input parameters of the model are 26 statistical feature parameters, which are extracted based on the derived maps and include total number of discharges, quadrant discharge frequency ratio, map symmetry, skewness, kurtosis, iso-phase center, and Weibull distribution fitting parameters. The first layer of neurons in the BP neural network model integrates the statistical parameters and outputs a logical quantity, and the second layer of neurons combines the logical output of the first layer to output a logical quantity that matches the target defect type.

7. The partial discharge charge detection device based on pulse current method according to claim 2, characterized in that, The high-speed acquisition module uses a small-capacity storage space that can be read and written repeatedly. Combined with analog or digital watchdog technology, when the input signal exceeds the threshold, it triggers the automatic transfer of the digitized information of the discharge pulse. At the same time, it reads the phase timer count information to obtain the phase information of the discharge pulse, thereby forming phase-resolved pulse sequence information.

8. The partial discharge charge detection device based on pulse current method according to claim 2, characterized in that, The combined terminal is a combined banana head, which can be matched with adapters of different specifications and sizes to adapt to high-voltage live display devices with different core phase hole sizes; the grounding terminal is an alligator clip, which is used for detachable connection to the exposed metal grounding parts of high-voltage switchgear and cable branch boxes.

9. The partial discharge charged detection device based on pulse current method according to claim 2, characterized in that, For high-voltage live display devices without core phase holes, the current sensing unit adopts a fixed installation method and is combined with the input signal lead of the high-voltage live display device to form a fixed installation adapter test fixture, which can realize detection without disconnecting the input signal lead.

10. The partial discharge charge detection device based on pulse current method according to claim 5, characterized in that, The comprehensive early warning model for partial discharge status optimizes diagnostic accuracy through five derived parameters: intensity level, significance level, signal-to-noise ratio, repetition rate, and power frequency correlation coefficient. The intensity level is the ratio of probability intensity to evaluation benchmark, the significance level is the ratio of probability intensity to average intensity, the signal-to-noise ratio is the ratio of average intensity to background noise, and the repetition rate is the ratio of discharge frequency to benchmark frequency.

Images