Ring main unit with intelligent online partial discharge detection system

Abstract

According to the ring main unit with the intelligent online partial discharge detection system provided by the invention, the problem of multisource data asynchronization is solved through a four-dimensional space-time synchronous acquisition mechanism, so that the time alignment error of ultrahigh frequency and ultrasonic signals is lower, and the positioning precision is higher; the edge computing unit is used for executing adaptive filtering and pulse extraction, so that the effective data compression rate is greatly increased, the bandwidth requirement is reduced, and quick response within 3 seconds is realized; cross-modal coupling factor quantification electromagnetic-acoustic emission correlation characteristics are innovatively proposed, the contribution degree of a sensor is adaptively adjusted in combination with a dynamic weight distribution strategy, and high discharge type recognition accuracy is still kept in a strong electromagnetic interference environment; the space-time heterogeneous data is cooperatively processed through the two-channel deep learning model, the composite discharge mode is effectively identified, and the false alarm rate is lower compared with a traditional method; in addition, the system automatically and incrementally trains the model per week, continuously adapts to signal attenuation caused by equipment aging, and prolongs the effective monitoring life by more than three times.

Description

TECHNICAL FIELD

[0001] The application relates to the field of power equipment, in particular to a ring main unit with an intelligent online partial discharge detection system. BACKGROUND

[0002] As a key equipment of the power distribution network, the ring main unit bears the task of power distribution and fault isolation. The internal insulation medium is easily caused to have partial discharge (PD) under the influence of electric thermal stress and environmental temperature and humidity changes in long-term operation. According to the power grid fault statistics, about 68% of the ring main unit faults are caused by insulation deterioration, and PD is the primary sign of insulation deterioration. The traditional PD detection relies on artificial periodic inspection, and a handheld TEV or ultrasonic detector is used, which has three defects: 1. Poor timeliness: the inspection period is usually quarterly, and sudden discharge cannot be captured; 2. High missed detection rate: a single sensor is easily affected by electromagnetic interference (such as switch operation), resulting in a signal-to-noise ratio of less than 8dB and a misjudgment rate of more than 40%; 3. Weak positioning capability: the sound source position is determined by experience, and the positioning error is often greater than 50cm.

[0003] Although the existing online monitoring scheme (such as application number CN202111415200.X) realizes current-ultrasonic dual-mode detection, it still has limitations: Insufficient applicability: designed for transformer structure, without considering the signal attenuation difference caused by the compact space of the ring main unit (UHF signal attenuation in the ring main unit is 15-20dB higher than that in the transformer); Data processing lag: the full amount of raw data is uploaded, resulting in bandwidth pressure (single cabinet daily data volume is more than 2TB), and the analysis delay is more than 10 minutes; Anti-interference mechanism is missing: the problem of asynchronous data of multiple sensors (time difference>1ms, positioning error expands to 30cm) is not solved.

[0004] Especially, the ring main unit PD has the characteristics of short duration (duration <200ns) and multi-source (suspension / creepage discharge coexist in the same cabinet), and the existing system has insufficient recognition accuracy of 75% for composite discharge types due to insufficient sampling rate (typical value 2MS / s) and poor model generalization.

[0005] Therefore, it is urgent to develop a ring main unit PD online detection scheme that integrates high-precision sensing, edge computing and intelligent diagnosis to realize real-time early warning and accurate positioning of discharge. SUMMARY

[0006] The purpose of the application is to provide a ring main unit with an intelligent online partial discharge detection system to solve the problems existing in the prior art.

[0007] To achieve the above purpose, the application provides the following scheme: The application provides a ring network cabinet with an intelligent online partial discharge detection system, comprising a ring network cabinet body, further comprising: a multi-sensor array unit, comprising a ground electric wave sensor, an ultrasonic sensor, an ultra-high frequency antenna array and an infrared thermal imaging sensor arranged on the ring network cabinet body; an edge computing unit connected with the ground electric wave sensor, the ultrasonic sensor, the ultra-high frequency antenna array and the infrared thermal imaging sensor through a hardware circuit; a cloud intelligent analysis platform interacting with the edge computing unit through a wireless communication network and performing partial discharge analysis based on a double-channel deep learning model; a management terminal wirelessly communicating with the cloud intelligent analysis platform and used for receiving alarm information output by the cloud intelligent analysis platform.

[0008] Preferably, the ground electric wave sensor is installed at a metal joint of the ring network cabinet body; the ultrasonic sensor is installed at an insulating sleeve root and a cable terminal of the ring network cabinet body; the ultra-high frequency antenna array is installed around an insulator of the ring network cabinet body; and the infrared thermal imaging sensor is installed at the top of the inner side of the cabinet door of the ring network cabinet body.

[0009] Preferably, the ground electric wave sensor, the ultrasonic sensor, the ultra-high frequency antenna array and the infrared thermal imaging sensor all have a built-in GPS / Beidou dual-mode time service module.

[0010] Preferably, the edge computing unit performs the following preprocessing steps: S101. Adaptive power frequency harmonic filtering based on FPGA, and the filtering algorithm is: ; wherein, is an effective partial discharge signal after filtering, is a raw signal time domain waveform collected by a sensor, is a dynamic amplitude of the kth harmonic, is a power interference harmonic frequency set, when K is 1, the frequency is 50Hz fundamental wave, when K is 2, the frequency is 150Hz third harmonic, when K is 3, the frequency is 250Hz fifth harmonic, and t is a sampling time point, is a phase offset of the kth harmonic; S102. Improved double-threshold pulse detection condition.

[0011] Preferably, the cloud intelligent analysis platform performs the following processing steps: S201. Cross-modal coupling factor calculation, and the calculation formula is: ; wherein, is the amplitude of the ground wave signal of the time-aligned i-th data point, is the amplitude of the ultra-high frequency signal of the time-aligned i-th data point, is the mean of the ground wave signal within the time window, is the mean of the ultra-high frequency signal within the time window, N is the total number of data points within the sliding window, is the standard deviation of the ground wave signal within the time window, is the standard deviation of the ultra-high frequency signal within the time window; S202. Dynamic weight distribution, the calculation formula is: ; wherein, is the fusion weight of the j-th type of sensor, is the signal-to-noise ratio of the j-th type of sensor, is the maximum signal-to-noise ratio among the four types of sensors, alpha is the aging factor, t is the device running time, is the sensor aging inflection point.

[0012] Preferably, in the dual-channel deep learning model, the CNN branch processes the ultra-high frequency spectrum graph and the thermal imaging graph, the LSTM branch processes the ground wave and ultrasonic wave time sequence characteristics, and the attention fusion layer weights and splices the double-branch output feature vectors.

[0013] Preferably, the alarm information includes a discharge type and a severity level.

[0014] Preferably, LoRaWAN or NB-IoT communication is adopted between the edge computing unit and the cloud intelligent analysis platform, and WebSocket is adopted to push alarm information between the cloud intelligent analysis platform and the management terminal.

[0015] The application also provides a partial discharge judgment method of a ring main unit with an intelligent online partial discharge detection system, including the following steps: S1. Time and space synchronous data acquisition of a multi-sensor array unit; S2. The edge computing unit performs pulse extraction and primary feature calculation, and uploads data; S3. The cloud intelligent analysis platform calculates a cross-modal coupling factor and inputs a dual-channel deep learning model; S4. The dual-channel deep learning model outputs a discharge type and a severity level to a management terminal; S5. The dual-channel deep learning model is automatically updated by incremental training every week.

[0016] The application has the following beneficial technical effects relative to the prior art: The application provides a ring network cabinet with an intelligent online partial discharge detection system, solves the asynchronous problem of multi-source data through a four-dimensional space-time synchronous acquisition mechanism, makes the time alignment error of ultrahigh frequency and ultrasonic wave signals lower, and makes the positioning accuracy higher; adaptive filtering and pulse extraction are performed by using an edge computing unit, the effective data compression rate is greatly improved, the bandwidth demand is reduced, and fast response within 3 seconds is realized; the cross-modal coupling factor is innovatively proposed to quantize the electromagnetic-acoustic emission correlation characteristics, a dynamic weight distribution strategy is combined to adaptively adjust the contribution of sensors, and a higher discharge type recognition accuracy is still maintained in a strong electromagnetic interference environment; through the cooperative processing of a dual-channel deep learning model, time-space heterogeneous data is effectively identified, and the composite discharge mode is effectively identified, and the false alarm rate is lower than that of a traditional method; in addition, the system automatically incrementally trains the model every week, continuously adapts to signal attenuation caused by equipment aging, and prolongs the effective monitoring life by more than 3 times. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 The application provides a ring network cabinet with an intelligent online partial discharge detection system structure diagram; Figure 2 The application provides a ring network cabinet with an intelligent online partial discharge detection system communication relationship diagram; Figure 3 The application provides a partial discharge judgment method flowchart of a ring network cabinet with an intelligent online partial discharge detection system; In the figure: 1: ring network cabinet body, 2: multi-sensor array unit, 21: ground wave sensor, 22: ultrasonic wave sensor, 23: ultrahigh frequency antenna array, 24: infrared thermal imaging sensor, 3: edge computing unit, 4: cloud intelligent analysis platform, 5: management terminal. DETAILED DESCRIPTION

[0019] In order to make the application purpose, features, advantages of the present application more obvious and easy to understand, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the following described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0020] In the description of the embodiments of the present application, it should be understood that the terms "length", "width", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0021] In the embodiments of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection", "fixing" and the like should be understood in a broad sense, for example, it can be fixed connection, or detachable connection, or integral; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, it can be the internal communication of two elements or the interaction relationship of two elements. For those skilled in the art, the specific meaning of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0022] The purpose of the present application is to provide a ring network cabinet with intelligent online partial discharge detection system to solve the problems existing in the prior art.

[0023] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0024] Embodiment 1: The present embodiment provides a ring network cabinet with intelligent online partial discharge detection system, as shown in Figure 1 and 2 , comprising a ring network cabinet body 1, which is a common ring network cabinet in the prior art, and its specific structure and type are not limited, further comprising: A multi-sensor array unit 2, the multi-sensor array unit 2 includes a ground electric wave sensor 21, an ultrasonic sensor 22, an ultra-high frequency antenna array 23 and an infrared thermal imaging sensor 24 disposed on the ring network cabinet body 1, thereby realizing a four-dimensional space-time synchronous acquisition mechanism to realize the acquisition of multi-source data; An edge computing unit 3, the edge computing unit 3 is connected with the ground electric wave sensor 21, the ultrasonic sensor 22, the ultra-high frequency antenna array 23 and the infrared thermal imaging sensor 24 through a hardware circuit; A cloud intelligent analysis platform 4, the cloud intelligent analysis platform 4 interacts with the edge computing unit 3 through a wireless communication network, and performs partial discharge analysis based on a double-channel deep learning model; A management terminal 5, the management terminal 5 is in wireless communication with the cloud intelligent analysis platform 4, and is used to receive alarm information output by the cloud intelligent analysis platform 4.

[0025] As an implementation, please refer to Figure 1 , the ground electric wave sensor 21 can adopt a capacitive coupling sensor, installed at the metal joint of the ring main unit body 1, with a spacing of less than 30 cm, and an adaptive gain circuit built-in to suppress power frequency interference; the ultrasonic sensor 22 can adopt a piezoelectric acoustic emission sensor, installed at the root of the insulating sleeve and the cable terminal of the ring main unit body 1, equipped with a sound wave guide rod to make the signal-to-noise ratio SNR> 20dB; the ultra-high frequency antenna array 23 can adopt a microstrip patch antenna array, installed around the insulator of the ring main unit body 1, to achieve coverage of 3-4 azimuth angles; the infrared thermal imaging sensor 24 can adopt a non-cooled focal plane infrared camera, installed at the top inside of the cabinet door of the ring main unit body 1, to achieve overhead multispectral temperature measurement.

[0026] As an implementation, the ground electric wave sensor 21, the ultrasonic sensor 22, the ultra-high frequency antenna array 23 and the infrared thermal imaging sensor 24 all have a GPS / Beidou dual-mode timing module built-in to achieve μs-level time synchronization, solve the multi-source data alignment problem, and the data sampling rate is configured as: ultra-high frequency antenna array 10 MS / s, ground electric wave sensor, ultrasonic sensor 1 MS / s, infrared thermal imaging sensor 1 Hz.

[0027] As an implementation, the edge computing unit 3 performs the following preprocessing steps: S101. Adaptive power harmonic filtering based on FPGA, the filtering algorithm is: ; wherein, is the effective partial discharge signal after filtering, is the original signal time waveform collected by the sensor, is the dynamic amplitude of the kth harmonic, is the power interference harmonic frequency set, when K takes the value of 1, the frequency is 50Hz fundamental wave, when K takes the value of 2, the frequency is 150Hz third harmonic, when K takes the value of 3, the frequency is 250Hz fifth harmonic, t is the sampling time point, is the phase shift of the kth harmonic; S102. Improved double-threshold pulse detection condition.

[0028] As an implementation, the cloud intelligent analysis platform 4 performs the following processing steps: S201. Cross-modal coupling factor calculation, used to quantify electromagnetic-acoustic emission correlation, the calculation formula is: ; wherein, is the ground electric wave signal amplitude of the ith data point after time alignment, is the ultra-high frequency signal amplitude of the ith data point after time alignment, is the mean value of the ground electric wave signal in the time window, is the mean value of the ultra-high frequency signal in the time window, N is the total number of data points in the sliding window, is the standard deviation of the ground electric wave signal in the time window, is the standard deviation of the ultra-high frequency signal in the time window; S202. Dynamic weight distribution, the calculation formula is: wherein, is the fusion weight of the jth sensor, is the signal-to-noise ratio of the jth sensor, is the maximum signal-to-noise ratio among the four types of sensors, a is the aging factor, t is the device running time, is the sensor aging inflection point.

[0029] As an embodiment, in the dual-channel deep learning model, the CNN branch processes the ultra-high frequency spectrum graph and the thermal imaging graph, the ResNet-18 network based on transfer learning, the LSTM branch processes the ground electric wave and ultrasonic time sequence features, the input dimension is [pulse density, rise time, RMS value], and the attention fusion layer weights and splices the double-branch output feature vectors.

[0030] As an embodiment, the alarm information includes discharge type and severity level, the discharge type includes floating discharge, surface discharge or internal discharge, and the severity level is divided into three levels.

[0031] As an embodiment, LoRaWAN or NB-IoT communication is used between the edge computing unit 3 and the cloud intelligent analysis platform 4, encrypted data packets are transmitted, the structure is [32-bit timestamp][16-bit sensor ID][16-bit data length][32-bit CRC32][AES-256 encrypted data], WebSocket is used between the cloud intelligent analysis platform 4 and the management terminal 5 to push alarm information, and at the same time, the management terminal 5 can issue control instructions through HTTPS.

[0032] Embodiment 2: The embodiment also provides a partial discharge judgment method of the ring main unit with the intelligent online partial discharge detection system, as shown in Figure 3 , comprising the following steps: S1. The multi-sensor array unit 2 synchronously collects data in time and space; S2. The edge computing unit 3 performs pulse extraction and primary feature calculation, and uploads data; S3. The cloud intelligent analysis platform 4 calculates the cross-modal coupling factor and inputs the dual-channel deep learning model; ​S4. The dual-channel deep learning model outputs the discharge type and severity level to the management terminal 5; S5. The dual-channel deep learning model is automatically incrementally trained and updated every week.

[0033] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the described device embodiments are merely schematic. For example, the division of the units is only a logical function division. There can be another division manner for the actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0034] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0035] In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0036] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application, essentially or the part that contributes to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0037] The above-described embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalent replacements; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A ring main unit with an intelligent online partial discharge detection system, comprising a ring main unit body (1), characterized in that: Also includes: A multi-sensor array unit (2), the multi-sensor array unit (2) comprising a ground wave sensor (21), an ultrasonic sensor (22), an ultra-high frequency antenna array (23), and an infrared thermal imaging sensor (24) deployed on the ring network cabinet body (1); An edge computing unit (3), the edge computing unit (3) being connected to the ground wave sensor (21), the ultrasonic sensor (22), the ultra-high frequency antenna array (23), and the infrared thermal imaging sensor (24) via a hardware circuit; A cloud-based intelligent analysis platform (4), wherein the cloud-based intelligent analysis platform (4) interacts with the edge computing unit (3) via a wireless communication network and performs partial discharge analysis based on a dual-channel deep learning model; A management terminal (5), the management terminal (5) is in wireless communication with the cloud-based intelligent analysis platform (4), and is used to receive alarm information output by the cloud-based intelligent analysis platform (4).

2. The ring main unit with an intelligent online partial discharge detection system according to claim 1 is characterized in that: The ground wave sensor (21) is installed at the metal joint of the ring network cabinet body (1); the ultrasonic sensor (22) is installed at the root of the insulating sleeve and the cable terminal of the ring network cabinet body (1); the ultra-high frequency antenna array (23) is installed around the insulator of the ring network cabinet body (1); and the infrared thermal imaging sensor (24) is installed at the top inside the cabinet door of the ring network cabinet body (1).

3. The ring main unit with an intelligent online partial discharge detection system according to claim 1 is characterized in that: The ground wave sensor (21), the ultrasonic sensor (22), the ultra-high frequency antenna array (23), and the infrared thermal imaging sensor (24) all have built-in GPS / Beidou dual-mode timing modules.

4. The ring main unit with an intelligent online partial discharge detection system according to claim 1 is characterized in that: The edge computing unit (3) performs the following pre-processing steps: S101. Adaptive power frequency harmonic filtering based on FPGA. The filtering algorithm is: ； in, is the effective PD signal after filtering, is the time domain waveform of the original signal collected by the sensor, is the dynamic amplitude of the kth harmonic, is the frequency set of power frequency interference harmonics. When K is 1, the frequency is 50 Hz fundamental wave. When K is 2, the frequency is 150 Hz third harmonic. When K is 3, the frequency is 250 Hz fifth harmonic. t is the sampling time point. is the phase shift of the kth harmonic; S102. Improve dual-threshold pulse detection conditions.

5. The ring main unit with an intelligent online partial discharge detection system according to claim 1 is characterized in that: The cloud-based intelligent analysis platform (4) performs the following processing steps: S201. Calculation of cross-modal coupling factor, the calculation formula is: ； in, is the amplitude of the ground wave signal at the time-aligned i-th data point, is the UHF signal amplitude of the time-aligned i-th data point, is the mean value of the ground wave signal in the time window, is the mean value of the UHF signal in the time window, N is the total number of data points in the sliding window, is the standard deviation of the ground wave signal in the time window, is the standard deviation of the UHF signal in the time window; S202. Dynamic weight allocation, the calculation formula is: ； in, is the fusion weight of the j-th sensor, is the signal-to-noise ratio of the j-th sensor, is the maximum signal-to-noise ratio among the four types of sensors, α is the aging factor, t is the operating time of the device, This is the sensor aging inflection point.

6. The ring main unit with an intelligent online partial discharge detection system according to claim 1, characterized in that: In the dual-channel deep learning model, the CNN branch processes ultra-high frequency spectrograms and thermal images, the LSTM branch processes geomagnetic waves and ultrasonic time series features, and the attention fusion layer weightedly splices the dual-branch output feature vectors.

7. The ring main unit with an intelligent online partial discharge detection system according to claim 1, characterized in that: The alarm information includes the discharge type and severity level.

8. The ring main unit with an intelligent online partial discharge detection system according to claim 1, characterized in that: The edge computing unit (3) and the cloud-based intelligent analysis platform (4) communicate using LoRaWAN or NB-IoT, and the cloud-based intelligent analysis platform (4) and the management terminal (5) push alarm information using WebSocket.

9. The method for determining partial discharge of a ring main unit with an intelligent online partial discharge detection system according to any one of claims 1 to 8, characterized in that: The following steps are involved: S1. Multi-sensor array unit (2) collects data in a spatiotemporal synchronous manner; S2. The edge computing unit (3) performs pulse extraction and primary feature calculation, and uploads the data; S3. The cloud-based intelligent analysis platform (4) calculates the cross-modal coupling factor and inputs it into the dual-channel deep learning model; S4. The dual-channel deep learning model outputs the discharge type and severity level to the management terminal (5); S5. Update the dual-channel deep learning model through automatic incremental training every week.

Images