Ring main unit intelligent sensing on-line monitoring system and method thereof

Abstract

The invention relates to the technical field of fault detection, in particular to an intelligent sensing online monitoring system and method for a ring main unit, and the system comprises an energy collection module, a sudden change analysis module, an abnormity locking module, a risk partitioning module and a link synchronization module. According to the method, an energy parameter is formed through 50 millisecond sampling period integration, a short-time energy change characteristic is captured, an abnormal segment is extracted based on a sudden change rate threshold value, an energy change direction and sudden change amplitude are extracted to lock an abnormal starting point, a leakage current instantaneous value is continuously collected, and a maximum amplitude and an average amplitude are extracted to calculate offset quantification insulation fluctuation. And calling time window data to calculate an amplitude fluctuation rate, executing multi-threshold judgment and evaluation of a degradation degree, packaging an energy parameter, an amplitude offset and insulation grade information in combination with an offset, executing synchronization consistency to correct node deviation, ensuring reliable data synchronization, and improving insulation trend identification sensitivity and abnormal state monitoring accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to an intelligent sensing online monitoring system for a ring main unit and a method thereof. Background Art

[0002] The field of fault detection technology includes an overall technical system that conducts real-time monitoring and diagnosis of electrical equipment, lines, systems, etc. to detect potential faults and take maintenance measures in a timely manner. The core content of this field includes determining the type and location of the fault by detecting parameters such as current, voltage, temperature, and insulation performance in the operating state of the equipment, and combining information collection, data processing, and remote transmission technology to achieve a comprehensive assessment and early warning of the health status of the equipment. The overall development of fault detection technology is systematically reflected in the evolution from single parameter detection to multi-parameter fusion judgment, and from regular maintenance to condition-based maintenance. It covers multiple specific links such as sensor deployment, state parameter collection, feature extraction, fault discrimination analysis, and information communication support. It is an important basic technology to ensure the safe and stable operation of power systems and other industrial application scenarios.

[0003] Among them, the intelligent sensing online monitoring system of the ring main unit refers to the power distribution ring main unit equipment, which collects the voltage, current, partial discharge signal, temperature and humidity data inside the ring main unit by deploying a variety of sensors, and uses edge computing nodes to perform preliminary processing and feature analysis on the collected real-time data, and identifies abnormal conditions based on specific fault discrimination rules, and transmits the analysis results to the monitoring platform through the wireless communication module. The system uses distributed sensing perception, real-time data collection and analysis, edge intelligent identification and remote data transmission, covering specific matters such as partial discharge monitoring, environmental status detection, and primary circuit status perception, and conducts continuous perception of the operating status of the ring main unit and early identification of potential faults.

[0004] The existing technology collects a single instantaneous parameter and lacks the extraction of continuous energy change rate, making it difficult to quickly locate the onset of local anomalies. Leakage current monitoring is only based on absolute amplitude comparison and ignores the dynamic change process, resulting in slow recognition of insulation performance fluctuations. Anomaly recognition is mainly based on single detection and does not introduce amplitude change statistics in continuous time windows, which reduces the resolution of degradation development trends. Data link synchronization is only aligned with timestamps and does not combine energy and offset parameter verification consistency. There is a risk of asynchronous offset of node data, which affects the accuracy of fault judgment and response timeliness in distributed monitoring scenarios. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and to propose an intelligent sensing online monitoring system and method for a ring main unit.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A ring main unit intelligent sensing online monitoring system comprises: The energy collection module collects the instantaneous values ​​of partial discharge, current and voltage, forms energy parameters through periodic integration, sets the sampling period of 50 milliseconds to dynamically adapt to the mutation rate threshold, binds the time tag, and outputs the energy parameters to the mutation analysis module; The mutation analysis module receives the energy parameters, calculates the mutation rate formed by the energy changes of adjacent time windows based on the sliding window difference algorithm, extracts abnormal fragments based on the mutation rate threshold of 5%-20% determined by the gradient descent method optimization, extracts the energy change direction and mutation amplitude, and outputs abnormal features to be transmitted to the abnormal locking module; An abnormality locking module receives the abnormality feature, triggers the insulation monitoring unit to collect the instantaneous value of the leakage current, extracts the maximum value and the average value and calculates the offset, and outputs the offset to the risk partition module; The risk partition module receives the offset, calls the front and back time window data, calculates the amplitude fluctuation rate, performs multi-threshold judgment based on the offset and the fluctuation rate, and outputs the insulation level to the link synchronization module.

[0007] As a further solution of the present invention, the energy parameters include time synchronization identification, single-window energy integral value, and energy change trend characteristics; the abnormal characteristics include abnormal trigger points, energy mutation directions, and mutation amplitude values; the offsets include maximum amplitude offsets, minimum amplitude offsets, and amplitude average difference; the insulation levels include insulation degradation classification labels, amplitude fluctuation rate levels, and risk level identification; the dynamic adaptation algorithm based on a 50-millisecond sampling period and an energy mutation rate threshold.

[0008] As a further solution of the present invention, the energy harvesting module includes: The instantaneous data acquisition submodule obtains the signal sources of the instantaneous values ​​of partial discharge, current, and voltage, extracts the sampled data synchronously according to the time label sorting, sets the sampling step length and timing verification standard to filter the continuous data segments, removes the time dislocation and abnormal interruption nodes, organizes and archives the sampled stream data, and generates the instantaneous sampling data set; The periodic integration processing submodule extracts the instantaneous value data pairs of current and voltage in each period based on the instantaneous sampling data set, accumulates the energy contribution according to the corresponding product integral formula, normalizes the influence of the sampling time step, organizes and archives the calculation results of each period, and generates a periodic energy integration data set; The energy-time binding submodule extracts the energy value and the corresponding time tag of each cycle according to the periodic energy integral data set, binds the energy data and the time node according to the time sequence, archives the energy parameters and the time information flow, organizes them into a data set for uploading, and generates energy parameters; The periodic integration is to perform an integration operation on the square values ​​of the instantaneous values ​​of partial discharge, current and voltage within a 50 millisecond time window.

[0009] As a further solution of the present invention, the mutation analysis module comprises: The mutation rate calculation submodule receives the energy parameter, calculates the energy change of adjacent time windows based on the sliding window difference algorithm, forms the mutation rate, combines the 5%-20% mutation rate threshold, optimizes the calculation through the gradient descent method, and generates the mutation rate calculation result; The threshold matching submodule performs matching according to the mutation rate calculation result and the optimized mutation rate threshold, extracts the abnormal fragments that meet the conditions, and generates the abnormal fragment identification; The abnormal feature extraction submodule extracts the energy change direction and mutation amplitude based on the abnormal segment identifier, and outputs the abnormal feature to the abnormal locking module; The mutation rate threshold parameters include energy mutation rate 5%-20%, offset threshold set to ±20%, and fluctuation rate ≥±10%; The energy mutation rate is analyzed through historical data, observing the energy change characteristics of normal and abnormal events, and comparing the energy fluctuations under normal conditions with the mutation amplitudes in abnormal events. The mutation rate of normal fluctuations is less than 5%, while the mutation rate of abnormal events is more than 20%; When the fluctuation rate operates normally, the fluctuation rate is between ±1% and ±5%, while the fluctuation rate caused by abnormal conditions exceeds ±10%. The influence of the fluctuation rate on abnormal detection is analyzed by the gradient descent method, and the fluctuation rate threshold is set to ±10%.

[0010] As a further solution of the present invention, the abnormal locking module includes: The leakage current acquisition submodule obtains the abnormal characteristics, sets the insulation monitoring sampling starting point according to the time tag sequence, sets the continuous sampling step to collect the leakage current instantaneous value, screens out the data nodes with abnormal sampling time intervals, archives the valid time series samples, and generates the leakage current instantaneous sequence; The current feature extraction submodule extracts the maximum and average values ​​of the leakage current in each continuous sampling period based on the leakage current instantaneous sequence, calculates the internal feature variation of the period based on the maximum and average values, marks the abnormal sample segments, archives the valid feature variation value sequence, and generates a current feature variation data set; The offset calculation submodule calls the feature change value extracted in each time segment according to the current feature change data set, calculates the offset value according to the offset formula, removes the time segment samples with abnormal offset, archives the pairing relationship between the offset value and the time tag, and generates the offset; The triggering condition for triggering the insulation monitoring unit to collect the instantaneous value of the leakage current is that when the abnormal characteristic exceeds the mutation rate threshold, the insulation monitoring unit is triggered by a hardware interrupt signal.

[0011] As a further solution of the present invention, the risk zoning module includes: The data window extraction submodule obtains the offset, divides the continuous window segments in chronological order, filters the data segments whose time span meets the requirements, removes the abnormal time interval segments, archives the continuous valid window data, and generates the window continuous data set; The fluctuation rate calculation submodule calls the offset data sequence based on the window continuous data set, calculates the amplitude change rate per unit time of the continuous time window, filters the amplitude change rate abnormal data segment, archives the effective rate and time label data, and generates the amplitude fluctuation rate sequence; The insulation level determination submodule calls the window segment offset data according to the amplitude fluctuation rate sequence, makes a judgment according to the multi-threshold interval standard, determines the corresponding relationship between the insulation level and the time label, archives the judgment result, and generates the insulation level; The fluctuation rate is calculated by sliding window difference method, the offset determines the amplitude abnormality, the fluctuation rate determines the trend stability, a low fluctuation rate and less than the offset indicates an excellent insulation grade, and a high fluctuation rate greater than the offset indicates a poor grade judgment; The sliding window difference method slides the time series data by setting a window of fixed length, calculating the difference between adjacent data points each time, with a window step of 1, and using the mean of the difference results to evaluate the fluctuation rate of the data; The so-called multiple thresholds include an offset threshold and a fluctuation rate threshold. The offset threshold is set to ±20%, and the fluctuation rate threshold is set to ≥±10%.

[0012] As a further solution of the present invention, the system further includes: The link synchronization module receives the insulation level, offset and energy parameters, and calculates the synchronization error based on the comparison between the current cache value of the relay node and the received data. If the error is qualified, the energy parameter, offset and insulation level are packaged and uploaded to the master node. If the link synchronization error threshold is exceeded, the supplementary sampling command is triggered and the data closed-loop processing is completed; The recollection command generates a recollection task and transmits it to the data collection unit, executes recollection and verification of data, and uploads the recollection result to the master node to ensure data accuracy; The synchronization error comparison result specifically refers to a node synchronization error value, a synchronization error determination identifier, package verification data or a supplementary sampling instruction; The link synchronization error threshold is specifically that the difference between the received data and the cache value does not exceed ±3%.

[0013] As a further solution of the present invention, the link synchronization module includes: The synchronization error calculation submodule obtains the insulation level, offset and energy parameters, compares the current cache value of the relay node with the received data, calculates the error, and generates a synchronization error calculation result if the difference is less than the synchronization error threshold; The data upload encapsulation submodule encapsulates the energy parameter, offset and insulation level according to the synchronization error calculation result, and uploads them to the master node to generate the upload data result if the synchronization error is qualified; The data supplementary sampling processing submodule determines whether the synchronization state exceeds the threshold according to the synchronization error calculation result. If the synchronization deviation rate value in the synchronization error calculation result is greater than the link synchronization error threshold, a supplementary sampling command is sent to the energy acquisition module to re-collect energy parameters and offsets, and call the original time label and insulation level for field compounding operations. After performing the same data processing operations as the encapsulation upload process, it is uploaded to the master node to obtain energy parameters, offsets and insulation levels.

[0014] A ring main unit intelligent sensing online monitoring method, the network cabinet intelligent sensing online monitoring method is performed based on the above-mentioned ring main unit intelligent sensing online monitoring system, comprising the following steps: S1: Collect the instantaneous value of partial discharge, instantaneous value of current and instantaneous value of voltage through the monitoring device, set 50 milliseconds as the period for integration processing to generate energy parameters, bind the time tag, output the energy parameters, and perform subsequent analysis and processing; S2: Based on the energy parameter, extract the difference in the integral values ​​of adjacent time windows, call the differential change rate function to calculate the mutation rate, make interval judgment based on the mutation rate and the set threshold, extract the change direction and change amplitude, and output the abnormal characteristics; S3: Based on the abnormal characteristics, collect the instantaneous value of the leakage current in a continuous time period, extract the maximum amplitude and the average amplitude of each group, calculate the difference between the two to construct an amplitude offset, and output the amplitude offset; S4: Based on the amplitude offset, call the amplitude ratio of two consecutive time windows, input it into the continuous change rate function to calculate the amplitude fluctuation rate, perform multi-threshold judgment according to the amplitude offset and the amplitude fluctuation rate, and output the insulation level; S5: Based on the insulation level, combined with the amplitude offset and the energy parameter, call the relay node cache value to perform comparison processing, input it into the mean square error function to calculate the synchronization error, and determine whether the error meets the upload conditions. If it meets the conditions, package it and upload it. If not, generate a supplementary sampling instruction.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, energy parameters are formed by integrating through a fixed sampling period of 50 milliseconds, short-term energy change characteristics are captured, abnormal fragments are extracted based on the mutation rate threshold, the direction and amplitude of energy change are refined, and the abnormal starting point at the energy level is accurately locked; the instantaneous value of the leakage current is continuously collected, the maximum and average amplitudes are extracted, the offset is calculated, and the local fluctuation trend of the insulation is quantified in real time; the continuous time window data is called to calculate the amplitude fluctuation rate, and multi-dimensional threshold judgment is performed in combination with the offset to comprehensively measure the degree of insulation degradation; the synchronous packaging energy, offset and insulation level information are verified for synchronization consistency, and the node data deviation is corrected to ensure that the link data upload is accurate and stable, and the overall timeliness of insulation degradation trend identification and monitoring data consistency are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a system flow chart of the present invention; Figure 2 This is a system framework diagram of the present invention; Figure 3 It is a schematic diagram of the steps of the method of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0019] See also Figure 1 The present invention provides a technical solution: a ring main unit intelligent sensing online monitoring system comprising: The energy collection module collects the instantaneous values ​​of partial discharge, current and voltage, integrates them periodically to form energy parameters, sets the sampling period of 50 milliseconds to dynamically adapt to the mutation rate threshold, binds the time label, and outputs the energy parameters to the mutation analysis module; Periodic integration: Integrate the square values ​​of the instantaneous values ​​of partial discharge, current and voltage within a 50 millisecond time window; The mutation analysis module receives energy parameters, calculates the mutation rate formed by the energy change of adjacent time windows based on the sliding window difference algorithm, extracts abnormal fragments based on the 5%-20% mutation rate threshold determined by the gradient descent method optimization, extracts the energy change direction and mutation amplitude, and outputs abnormal features to pass to the abnormal locking module; Mutation rate threshold: parameters include energy mutation rate 5%-20%, amplitude shift ratio ≥1.3 or ≤0.7, and fluctuation rate ≥±10%; The abnormal locking module receives the abnormal characteristics, triggers the insulation monitoring unit to collect the instantaneous value of the leakage current, extracts the maximum value and the average value, calculates the offset, and outputs the offset to the risk partition module; The risk partition module receives the offset, calls the data of the previous and next time windows, calculates the amplitude fluctuation rate, makes multi-threshold judgments based on the offset and fluctuation rate, and outputs the insulation level to the link synchronization module; Fluctuation rate: Calculated by sliding window difference method, the offset determines the amplitude abnormality, and the fluctuation rate determines the trend stability. The two jointly determine the insulation level; The link synchronization module receives the insulation level, offset and energy parameters, and calculates the synchronization error based on the comparison between the current cache value of the relay node and the received data. If the error is qualified, the energy parameter, offset and insulation level are packaged and uploaded to the master node. If the link synchronization error threshold is exceeded, the supplementary sampling command is triggered and the data closed-loop processing is completed; Synchronization error comparison result: refers to the node synchronization error value, synchronization error judgment mark, package verification data or supplementary sampling instruction; Link synchronization error threshold: The difference between the received data and the cache value does not exceed ±3%; The link synchronization error threshold is in a very small range in actual data transmission. When designing the system, setting a smaller error threshold of ±3% can ensure system stability and accuracy. Through statistical analysis of historical data, it can be observed that the error fluctuation under normal link conditions will remain in the range of ±1%-±2%. In a few cases, the error will increase slightly and approach ±3%. In the event of link failure, transmission interference or other abnormal conditions, the difference between the received data and the cached value will increase. Based on historical data, it can be analyzed that in the case of differentiated link problems, the error will increase to a certain threshold. In an environment with large signal attenuation, the error will rise to ±3%. Through statistics of abnormal data, it can be confirmed whether ±3% is a reasonable tolerance range.

[0020] See also Figure 2 , the energy harvesting module includes: The instantaneous data acquisition submodule obtains the signal sources of the instantaneous values ​​of partial discharge, current, and voltage, extracts the sampled data synchronously according to the time label sorting, sets the sampling step length and timing verification standard to filter the continuous data segments, removes the time dislocation and abnormal interruption nodes, organizes and archives the sampled stream data, and generates the instantaneous sampling data set; In the instantaneous data acquisition submodule, the system first obtains the instantaneous signal source data such as local discharge, current, voltage, etc. through sensors to ensure that these signals are accurately collected within the set sampling time interval. The collected data includes current, voltage and time tags. After sampling, the system sorts the data according to the time tag to ensure the time consistency of the data. Assume that in one sampling, the current signals are 0.5A, 0.8A, and 1.2A, the voltage signals are 220V, 230V, and 240V, and the time tags are t1, t2, and t3 respectively. The system matches the current, voltage, and time tags one by one, and arranges them in chronological order to ensure that the time sequence of the data is correct. Then, the system performs a time sequence check based on the set sampling step to ensure that there is no misalignment or interruption between the data. If it is found that the timestamp interval of a data segment exceeds the set maximum allowable time difference, or the current and voltage data fluctuate abnormally, and the current suddenly jumps from 0.8A to 5A, the system will automatically remove the abnormal segment. In actual applications, assuming that the maximum time interval is set to 0.1 seconds, if the time interval between two consecutive data points is greater than 0.1 seconds, the data is considered to have time misalignment and is removed. The system organizes and archives the filtered data to generate a complete instantaneous sampling data set for subsequent energy calculations. The periodic integration processing submodule extracts the instantaneous value data pairs of current and voltage in each period based on the instantaneous sampling data set, accumulates the energy contribution according to the corresponding product integral formula, normalizes the influence of the sampling time step, organizes and archives the calculation results of each period, and generates a periodic energy integration data set; In the cycle integration processing submodule, the system extracts the current and voltage data pairs in each cycle from the instantaneous sampling data set, and calculates the energy contribution of each cycle according to the product integral formula. Assuming that the sampling time of each cycle is 50ms, the system extracts the current and voltage data from the data set for energy calculation. For example, in a certain cycle, the current data is 0.5A, 0.8A, 1.2A, the voltage data is 220V, 230V, 240V, and the sampling step is 0.01 seconds. According to the energy calculation formula: ; in, and are the instantaneous values ​​of current and voltage, respectively, is the time step of each sampling. Assuming the sampling step is 0.01 seconds, the energy is calculated as: ; The system will perform similar energy calculations for each cycle, obtain the energy values ​​of each cycle, and perform normalization processing to ensure the consistency of the calculation results under different sampling steps. The role of normalization processing is to eliminate the influence of sampling frequency on the results, so that the energy values ​​under different sampling frequencies are comparable; In the periodic integration processing submodule, different sampling time steps will lead to "estimation" errors of the time period in the energy calculation, especially when the sampling frequency is low or high, the error is amplified. The purpose of normalizing the influence of the sampling time step is to eliminate the influence of different sampling time intervals on the energy calculation results, to ensure that the calculated energy contribution can be standardized and consistent under different sampling steps. This process can ensure that the energy calculation results under different time steps are comparable and reasonable. In practical applications, the normalization method is often implemented by the following formula: ; is the normalized energy; is the energy of the original calculation; is the reference time step (choose a standard value, such as 0.01 seconds); is the actual sampling time step; Through this formula, the system can ensure the comparability and accuracy of calculation results at different time steps; The energy-time binding submodule extracts the energy value and corresponding time tag of each cycle according to the periodic energy integral data set, binds the energy data and time nodes according to the time sequence, archives the energy parameters and time information flow, organizes them into a data set for uploading, and generates energy parameters; In the energy-time binding submodule, the system binds the energy values ​​in the periodic energy integral data set to the corresponding time tags. Assuming that in period 1, the calculated energy is 58.2mJ, the time tag is t1, the energy of period 2 is 64.5mJ, the time tag is t2, and the energy of period 3 is 82.0mJ, the time tag is t3, the system will bind these energy values ​​to the corresponding time tags in chronological order. In this way, the system generates a mapping relationship between time and energy. The energy value of each period will correspond to its corresponding time tag one by one, thus forming a complete time-energy data stream. These bound data will be organized into data sets and archived as energy parameter and time tag data sets, which are suitable for uploading or subsequent analysis. The data set will contain the energy value of each period and its corresponding time tag, which is convenient for equipment energy consumption analysis, trend monitoring, and equipment operation status diagnosis. This process ensures the strict correspondence between energy data and time, and can provide a reliable data basis for subsequent analysis and processing.

[0021] See also Figure 2 , mutation analysis module includes: The mutation rate calculation submodule receives the energy parameters and calculates the energy changes of adjacent time windows based on the sliding window difference algorithm to form the mutation rate. It combines the 5%-20% mutation rate threshold and optimizes the calculation through the gradient descent method to generate the mutation rate calculation result. In the mutation rate calculation submodule, the system first receives the energy parameter and the corresponding time tag, calculates the energy change between adjacent time windows through the sliding window difference algorithm, and calculates the mutation rate based on this. Specifically, the system divides the time window according to the set size (for example, 50ms), and calculates the energy change between every two adjacent time windows. For example, assuming that the energy in the first time window is 100mJ and the energy in the second time window is 150mJ, the energy change is 50mJ. According to the mutation rate formula: ; in, is the energy of the first time window, is the energy of the second time window. In this example, the mutation rate is calculated as follows: ; In this way, the system will calculate the mutation rate for each pair of adjacent time windows to obtain the energy mutation situation in each time period. Next, the system will compare the mutation rate with the set threshold. The set threshold range is 5% to 20%. If the mutation rate exceeds the set upper limit (for example, 20%), the energy fluctuation during this period is considered to be an abnormal fluctuation, and the system will further identify and process the abnormality. In addition, the system will optimize the threshold through the gradient descent method to automatically adjust the sensitivity of the mutation rate according to different actual application scenarios and device status; Energy mutation rate is 5%-20%. Energy mutation rate refers to the ratio of the energy mutation change amplitude to the total energy within a specific time window. The threshold is set to 5%-20%, which is based on the working environment of the equipment, error tolerance and system response capability. A lower mutation rate (such as 5%) indicates that the system can operate smoothly, while a high mutation rate (close to 20%) indicates that the system has a fault or extreme working conditions. Therefore, the range of 5%-20% is set to limit the system's excessively fast energy changes to reduce the risk of equipment damage, while ensuring that in some abnormal situations (such as energy mutations in a very short time), alarms can be triggered or measures can be taken, and the energy changes in each time period can be recorded, and the normal fluctuation range of energy changes can be counted. For example, during the continuous operation of the equipment, the experiment can collect energy data at regular intervals (such as 1 second or 0.5 seconds), calculate the mutation rate, and set the energy mutation rate range of 5%-20%. For example, if in an experiment, the equipment changes rapidly from 100mJ of energy to 120mJ, and the mutation rate of this change is 20%, it means that such changes belong to the upper limit of the threshold, and the system needs to issue an alarm or take protective measures; The amplitude offset ratio is ≥1.3 or ≤0.7. The amplitude offset ratio reflects the amplitude offset of the signal during the operation of the device. If the amplitude offset ratio is greater than 1.3 or less than 0.7, it means that the system has experienced excessive signal changes, which is caused by external interference, equipment failure or unstable working conditions. The setting of the offset ratio threshold is based on the study of the stable operation characteristics of the equipment and its tolerance to external changes. If the offset ratio exceeds 1.3 or is less than 0.7, it means that the signal has abnormal fluctuations, causing equipment problems or inaccurate measurements. Corrective measures must be taken. By running the device in normal and abnormal environments, measuring the amplitude changes of its output signal, and comparing the amplitude offsets under different test conditions, the setting of thresholds 1.3 and 0.7 is evaluated. For example, when the device is working in the experiment, when its amplitude increases from 1.0 to 1.4, the amplitude offset ratio is 1.4, which exceeds the range of 1.3, indicating that corrective measures need to be taken. On the contrary, if the amplitude decreases from 1.0 to 0.6, the offset ratio is 0.6, which is lower than the lower limit of 0.7, and a warning must also be triggered; Fluctuation rate ≥ ±10%. Fluctuation rate indicates the speed at which the amplitude of the device output signal changes, measured in percentage. The ±10% fluctuation rate threshold is set to prevent frequent and large fluctuations in the system and ensure the smooth operation of the system. When the fluctuation rate exceeds ±10%, it indicates that the signal changes too much, the system is disturbed or there is a risk of failure. This threshold is set based on experimental observations of the long-term stable operation of the device, aiming to control the overreaction of the device and avoid unnecessary losses caused by frequent adjustments. The system signal will be continuously monitored during stable operation and the fluctuation rate will be calculated. For example, assuming that the voltage or current signal output by the device changes from 10V to 11V (change of 1V) within 10 seconds, the fluctuation rate is: ; If the fluctuation rate exceeds the ±10% range (for example, the actual change rate is 0.15V / s), it is considered abnormal. At this time, the system will take supplementary sampling, adjustment or other intervention measures; The displacement threshold is set to ±20%. The setting of the displacement threshold (±20%) is determined based on the signal variation range and external interference tolerance when the device is operating normally. This threshold is mainly used to identify abnormal offsets or changes in the system output signal and to detect whether the system is affected by changes in the external environment or failures. It is set to ±20% to trigger an alarm or take measures in time when the signal changes to a certain extent to avoid major equipment failures. This threshold is set based on the device design parameters, working environment, and long-term stable operation data. Through experiments, the amplitude changes of the device in a normal working environment are analyzed, and the tolerable signal variation range is observed. The threshold is then determined based on actual application requirements. The ±20% displacement threshold can ensure that the signal changes will not be too large when the device is operating normally, and can quickly identify and take measures when a device failure occurs. When the device is operating normally, the output signal is 100mA. After simulating different experimental scenarios (such as voltage instability, current load fluctuations, etc.), the following data is collected; Table 1: Current amplitude change statistics As shown in the table, in scenario 1 and scenario 2, the output current changes are +20% and -20% respectively. Within this range, the device can operate normally. However, when the output current of the device changes by more than ±20%, the system will trigger an abnormal flag and make necessary repairs. Through multiple experiments and statistical data, the system confirms that in most cases, the output signal of the device will not change by more than ±20%. Therefore, the offset threshold is set to ±20% as a standard to respond to abnormal signals in a timely manner and ensure stable operation of the device. The threshold matching submodule performs matching based on the mutation rate calculation results and the optimized mutation rate threshold, extracts qualified abnormal fragments, and generates abnormal fragment identifiers; In the threshold matching submodule, the system matches the mutation rate calculation result with the optimized mutation rate threshold, and filters out the abnormal fragments that meet the conditions. For example, the mutation rate threshold is set to 20%. If the mutation rate calculated within a certain time period is 25%, the segment of data meets the abnormal condition. The system marks this segment of data as abnormal and generates an abnormal segment identifier. In this way, the system can effectively extract the time period with abnormalities from a large amount of sampled data, identify the problematic areas, and ensure that these anomalies can be monitored and processed in a timely manner. In addition, the system will classify and manage data within different mutation rate ranges. If the mutation rate is within the threshold range (such as 15%), it is marked as a mild abnormality; if the mutation rate exceeds the upper limit of the threshold (such as 25%), it is marked as a serious abnormality. Through this refined matching, the system can more accurately identify and classify abnormal data; The abnormal feature extraction submodule extracts the energy change direction and mutation amplitude based on the abnormal fragment identification, and outputs the abnormal features to the abnormal locking module; In the abnormal feature extraction submodule, the system extracts the energy change direction and mutation amplitude of the abnormal segment based on the abnormal segment identifier generated in the previous step. Assuming that in a certain segment of data, the mutation rate is 25%, and the energy change direction changes from positive to negative, the system will extract the features of this segment of data and mark it as "energy mutation and reverse change". These features are passed to the abnormal locking module together with the corresponding time tag for further analysis and processing. By extracting the features of the abnormal segment, the system can identify the type and severity of the abnormal data in detail, which helps the system to diagnose and lock abnormal events. For example, in a certain period of time, assuming that the mutation rate calculation value is 30%, and the energy suddenly drops from 200mJ to 150mJ, this change is identified as an abnormality, and through the feature extraction process, the system identifies it as "energy sudden drop". These identifiers will be passed to the abnormal locking module, which will further evaluate the impact of the abnormality on the system operation. Through this process, the system can analyze each abnormal time segment in detail, respond in time and make appropriate processing to ensure the stability and security of the system. Table 2: Examples of identification of fragments with abnormal mutation rates As shown in Table 2, the system identifies the time segments with abnormalities by calculating the mutation rate of each time period. Through these identifications, the system can promptly pay attention to the areas with large energy fluctuations and affecting the stability of the system in the subsequent processing process; See also Figure 2 , the exception locking module includes: The leakage current acquisition submodule obtains the abnormal characteristics, sets the insulation monitoring sampling starting point according to the time tag sequence, sets the continuous sampling step to collect the leakage current instantaneous value, screens out the data nodes with abnormal sampling time intervals, archives the valid time series samples, and generates the leakage current instantaneous sequence; In the leakage current acquisition submodule, the system first receives the abnormal characteristics and sets the starting point of the insulation monitoring sampling according to the time tag sequence. The specific way to set the sampling starting point is to select a known time mark (for example, time t0) as the start of sampling, and set the sampling step backward according to this time point. The sampling step can be set according to actual needs, for example, the leakage current is sampled every 50ms. Assuming that the instantaneous value of the sampled current is 0.1mA, 0.2mA, and 0.15mA, the system will record the corresponding current values ​​at timestamps such as t1, t2, and t3. Next, the system will filter out nodes with abnormal time intervals during the sampling process. For example, when the time interval between two consecutive sampling points exceeds the set maximum allowable time (for example, 1 second), the data will be considered abnormal data and eliminated. The valid data after elimination will be organized into a time series format and a leakage current instantaneous series will be generated.

[0022] Assume that within a certain period of time, the instantaneous current values ​​are 0.1mA, 0.2mA, and 0.15mA, and the time labels are t1, t2, and t3 respectively, and the sampling step is 50ms. After acquiring the data, the system will sort them in the order of time labels and ensure that the time interval between data does not exceed the set maximum time interval (such as 0.5 seconds). If it is found that the data with too large an interval will be eliminated, the generated leakage current instantaneous sequence will provide basic data for subsequent feature extraction; The current feature extraction submodule extracts the maximum and average values ​​of the leakage current in each continuous sampling period based on the leakage current instantaneous sequence, calculates the internal feature variation of the period based on the maximum and average values, marks the abnormal sample segments, archives the valid feature variation value sequence, and generates a current feature variation data set; In the current feature extraction submodule, the system will extract the maximum and average values ​​of the leakage current in each continuous sampling period based on the leakage current instantaneous sequence. Assuming that the instantaneous values ​​of the current in a certain period are 0.1mA, 0.2mA, and 0.15mA, the system will calculate the maximum value of 0.2mA and the average value of 0.15mA in the period. Next, the system will calculate the internal characteristic variation of the period based on these maximum and average values. For example, assuming that the instantaneous values ​​of the current in the next period are 0.18mA, 0.22mA, and 0.16mA, the maximum value is 0.22mA, and the average value is 0.18mA. The variation is calculated as follows: ; In this example, the change is: ; When the change exceeds the set change threshold (for example, 0.1 mA), the data in this period will be discarded, and all qualified data will be archived as a current characteristic change data set for subsequent processing; The offset calculation submodule calls the feature change value extracted in each time segment according to the current feature change data set, calculates the offset value according to the offset formula, removes the time segment samples with abnormal offset, archives the pairing relationship between the offset value and the time tag, and generates the offset; In the offset calculation submodule, the system calls the feature change value extracted in each time segment according to the current feature change data set, and calculates the offset according to the offset formula. Assume that the offset formula used by the system is: ; Set the reference value to 0.1mA. If the change in the current time segment is 0.08mA, the offset is: ; If the offset is less than the set offset threshold (e.g. 0.05mA), the time segment data will be discarded, and all valid offset values ​​and corresponding time tags will be archived to generate offsets, which will be used for subsequent abnormal monitoring and positioning; Table 3: Power data offset statistics As shown in Table 3, the system calculates the maximum value, average value and its variation based on the current value in each time period, and further calculates the offset based on the set offset threshold. After the offset calculation, if the offset of a certain time segment is abnormal, the system will remove the data of this segment. These results will be used as data sets for monitoring and analysis.

[0023] See also Figure 2 , the risk zoning module includes: The data window extraction submodule obtains the offset, divides the continuous window segments in chronological order, filters the data segments whose time span meets the requirements, removes the abnormal time interval segments, archives the continuous valid window data, and generates the window continuous data set; In the data window extraction submodule, the system first receives the offset and divides the continuous time window segments according to the order of time tags. The system segments the data according to the preset window size (for example, 100ms). If the time tags are t1=0ms, t2=100ms, t3=200ms, and t4=300ms, the system divides the data into four window segments of 100ms, namely [t1, t2], [t2, t3], and [t3, t4]. Next, the system will screen the data segments whose time span meets the requirements. Specifically, if the time interval of a segment exceeds the set maximum span (such as 150ms), the data segment will be considered abnormal and eliminated. For example, if t1=0ms, t2=160ms, the system will remove the [t1, t2] window segment because its time span exceeds 150ms. All data segments with time intervals that meet the requirements will be sorted and archived to generate a continuous and valid window data set. Assuming that during the sampling process, the data time intervals are t1=0ms, t2=100ms, t3=250ms, and t4=350ms, the system will take [t1, t2] and [t3, t4] as valid data segments and remove the [t2, t3] segment because its time interval is greater than 150ms. Through these processes, the system ensures that the window data set meets the normative requirements of the time interval and provides reliable input data for subsequent analysis. The fluctuation rate calculation submodule calls the offset data sequence based on the window continuous data set, calculates the amplitude change rate per unit time of the continuous time window, filters the amplitude change rate abnormal data segment, archives the effective rate, and generates the amplitude fluctuation rate sequence; In the fluctuation rate calculation submodule, the system calls the offset data sequence based on the obtained window continuous data set to calculate the amplitude change rate per unit time of the continuous time window. For example, in a certain time window, the offset data is 0.05mA, 0.06mA, and 0.07mA. The system will calculate the change rate of these data at each sampling point. First, calculate the amplitude change between every two consecutive sampling points, such as 0.06mA-0.05mA=0.01mA, and then divide it by the corresponding time interval (for example, 100ms) to obtain the amplitude change rate per unit time. Assuming the time interval is 100ms, the rate calculation is as follows: ; If the calculated change rate exceeds the preset rate threshold (e.g. 0.2mA / s), the system will mark the data of this time period as abnormal and remove it. For example, in the subsequent window segment, if the change rate is 0.3mA / s, which exceeds the set 0.2mA / s threshold, then this section of data will be excluded from the valid data. All amplitude change rates that meet the requirements and the corresponding time tags will be archived to generate an amplitude fluctuation rate sequence. The insulation level determination submodule calls the window segment offset data according to the amplitude fluctuation rate sequence, makes a judgment according to the multi-threshold interval standard, determines the corresponding relationship between the insulation level and the time label, archives the judgment result, and generates the insulation level; In the insulation level judgment submodule, the system calls the offset data of the window segment according to the amplitude fluctuation rate sequence, makes a judgment based on the multi-threshold interval standard, and determines the corresponding relationship between the insulation level and the time label. The system sets multiple threshold intervals to judge the insulation level. Assume that the set rate threshold is: Normal: 0mA / s≤rate≤0.2mA / s; Warning: 0.2mA / s<rate≤0.5mA / s; Severe abnormality: rate>0.5mA / s; If the amplitude fluctuation rate is 0.15mA / s within a certain time window, the system will mark the data as "normal", and if the amplitude fluctuation rate is 0.4mA / s, the system will mark it as "warning". For the time period when the amplitude fluctuation rate exceeds 0.5mA / s, such as 0.6mA / s, the system will mark it as "severe abnormality". These judgment results will correspond to the time tags one by one to generate the insulation level; Assuming that the amplitude fluctuation rate is 0.6 mA / s within a certain period of time, the system will classify it as "serious abnormality" according to the set threshold and record the corresponding time tag. The system will generate a database containing all time periods and their corresponding insulation levels for subsequent fault diagnosis and equipment monitoring; Table 4: Insulation grade table As shown in Table 4, the system determines the insulation level of each time period by calculating the amplitude fluctuation rate. According to the different rates, the system specifies "normal", "warning" or "serious" levels for each time period and establishes a mapping relationship with the corresponding time label. These determination results provide key data for subsequent evaluation and monitoring of the equipment.

[0024] See also Figure 2 , the link synchronization module includes: The synchronization error calculation submodule obtains the insulation level, offset and energy parameters, compares the current cache value of the relay node with the received data, calculates the error, and generates the synchronization error calculation result if the difference is less than the synchronization error threshold; In the synchronization error calculation submodule, the system first obtains the insulation level, offset and energy parameters, and compares the received data with the current cache value to calculate the synchronization error. Assuming that the offset received by the system is 0.05mA, and the offset stored in the cache is 0.03mA, the calculated synchronization error is: ; The system compares the calculated synchronization error with the preset synchronization error threshold. If the synchronization error is less than the threshold (for example, set to 0.05mA), the synchronization is considered qualified, the synchronization error calculation result is generated, and the next step is performed. If the error is 0.02mA, which is less than 0.05mA, the system considers the synchronization qualified, and then performs data upload or supplementary sampling operations based on these calculation results; The data upload encapsulation submodule calculates the synchronization error and, if the synchronization error is qualified, encapsulates the energy parameter, offset and insulation level and uploads them to the master node to generate the uploaded data result; In the data upload and packaging submodule, the system determines whether the synchronization error is qualified based on the synchronization error calculation results. If the synchronization error meets the requirements, the system will upload the package energy parameters, offset and insulation level. For example, if the system has verified that the synchronization error is 0.02mA, which is lower than the set 0.05mA threshold, the system will then package the following data: Energy parameter: 50mJ; Offset: 0.03mA; Insulation grade: normal; Time tag: t1=10:00:00; The data will be packaged into a data packet and prepared to be uploaded to the master node for storage and further processing. After the data is uploaded, the system will generate an upload data result, indicating whether the upload is successful and saving relevant information; The data supplementary sampling processing submodule determines whether the synchronization state exceeds the threshold according to the synchronization error calculation result. If the synchronization deviation rate value in the synchronization error calculation result is greater than the link synchronization error threshold, a supplementary sampling command is sent to the energy acquisition module to re-collect energy parameters and offsets, and the original time tag and insulation level are called for field compound operations. After performing the same data processing operations as the package upload process, it is uploaded to the master node to obtain energy parameters, offsets and insulation levels. The data acquisition processing submodule determines whether the synchronization state exceeds the threshold according to the synchronization error calculation result, receives the synchronization deviation rate value δ obtained from the synchronization error calculation result, and compares it with the link synchronization error threshold δ 0 For numerical comparison, the threshold δ 0The setting is based on the system synchronization tolerance requirements and is set based on the historical data fluctuation range. The setting interval is between 0.25 and 0.30. The measurement point P001 in Table 5 is selected as the current processing object. Its receiving offset is 12.5mA, the buffer offset is 11.9mA, the receiving energy parameter is 5.82J, and the buffer energy parameter is 5.90J. The unified conversion rules for parameters of different units are set to directly retain the values ​​of mA-type electrical parameter units and unify J-type energy parameters with 100 times. The weight is used to balance its value, that is, the energy difference is multiplied by the weight and then participates in the calculation. Since the energy parameter has a higher contribution to the system stability, this weight conversion is used to reasonably reflect its influence. Under this rule, the energy difference is calculated as |5.82-5.90|=0.08, which is 0.08×100=8.0 after weighted conversion, and the offset difference is |12.5-11.9|=0.6. After integration, the synchronization deviation rate value δ is (0.6+8.0) / 2=4.3, which is greater than the set threshold δ 0 =0.3, the data is judged to be in an over-limit state, and a supplementary sampling command is immediately generated based on this judgment, and sent to the energy acquisition module to re-acquire the energy parameters and offset values ​​of the measuring point. After the acquisition is completed, the updated offset is 12.1mA, the updated energy parameter is 5.88J, and its time tag is kept as 2025-04-3014:00:00. The corresponding insulation level parameter is maintained as level II. Then the four data are called as encapsulation input fields and integrated into an upload data structure. The upload fields are energy parameter 5.88J, offset 12.1mA, insulation level II, and time tag 2025-04-30 14:00:00. The field processing order is standardized and formatted according to the encapsulation interface field protocol. After the data is arranged, the interface performs the upload operation to the main node system to form the uploaded data result, that is, the energy parameter, offset and insulation level are obtained; In the above formula, the offset difference is: ; The energy difference is converted to: ; The synchronization deviation rate is: ; Link synchronization error threshold: ; For comparison, , it is judged as synchronization over limit, which requires re-collection and uploading of data; Table 5: Monitoring point synchronization error data table As shown in Table 5, the data lists the initial parameter values ​​of multiple measurement points, including receiving offset, cache offset, receiving energy parameter and cache energy parameter, and is equipped with time tags and insulation levels to meet the input requirements of error calculation and judgment. The data are all derived from the measured input of the relay node and the current receiving device. The subsequent re-sampling process needs to complete the difference evaluation and packaging retransmission based on this type of data.

[0025] A ring main unit intelligent sensing online monitoring method, based on the above-mentioned ring main unit intelligent sensing online monitoring system, comprises the following steps: S1: Collect the instantaneous value of partial discharge, instantaneous value of current and instantaneous value of voltage through the monitoring device, set 50 milliseconds as the period for integration processing to generate energy parameters, bind the time tag, output the energy parameters, and perform subsequent analysis and processing; S2: Based on the energy parameter, extract the difference in the integral values ​​of adjacent time windows, call the differential change rate function to calculate the mutation rate, make interval judgments based on the mutation rate and the set threshold, extract the change direction and change amplitude, and output abnormal features; S3: Based on the abnormal characteristics, collect the instantaneous value of the leakage current in the continuous time period, extract the maximum amplitude and the average amplitude of each group, calculate the difference between the two to construct the amplitude offset, and output the amplitude offset; S4: Based on the amplitude offset, call the amplitude ratio of two consecutive time windows, input it into the continuous change rate function to calculate the amplitude fluctuation rate, perform multi-threshold judgment according to the amplitude offset and the amplitude fluctuation rate, and output the insulation level; S5: Based on the insulation level, combined with the amplitude offset and the energy parameter, call the relay node cache value to perform comparison processing, input it into the mean square error function to calculate the synchronization error, and determine whether the error meets the upload conditions. If it meets the conditions, package it and upload it. If not, generate a supplementary sampling instruction.

[0026] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent sensing online monitoring system for ring main units, characterized in that: The system comprises: The energy collection module collects the instantaneous values ​​of partial discharge, current and voltage, forms energy parameters through periodic integration, sets the sampling period of 50 milliseconds to dynamically adapt to the mutation rate threshold, binds the time tag, and outputs the energy parameters to the mutation analysis module; The mutation analysis module receives the energy parameters, calculates the mutation rate formed by the energy changes of adjacent time windows based on the sliding window difference algorithm, extracts abnormal fragments based on the mutation rate threshold of 5%-20% determined by the gradient descent method optimization, extracts the energy change direction and mutation amplitude, and outputs abnormal features to be transmitted to the abnormal locking module; An abnormality locking module receives the abnormality feature, triggers the insulation monitoring unit to collect the instantaneous value of the leakage current, extracts the maximum value and the average value and calculates the offset, and outputs the offset to the risk partition module; The risk partition module receives the offset, calls the data of the previous and next time windows, calculates the amplitude fluctuation rate, performs multi-threshold judgment based on the offset and the fluctuation rate, outputs the insulation level and transmits it to the link synchronization module.

2. The intelligent sensing online monitoring system for ring main unit according to claim 1 is characterized in that: The energy parameters include time synchronization identification, single-window energy integral value, and energy change trend characteristics; the abnormal characteristics include abnormal trigger points, energy mutation directions, and mutation amplitude values; the offsets include maximum amplitude offsets, minimum amplitude offsets, and average amplitude differences; the insulation levels include insulation degradation classification labels, amplitude fluctuation rate levels, and risk level identifications.

3. The intelligent sensing online monitoring system for ring main unit according to claim 2 is characterized in that: The energy harvesting module comprises: The instantaneous data acquisition submodule obtains the signal sources of the instantaneous values ​​of partial discharge, current, and voltage, extracts the sampled data synchronously according to the time label sorting, sets the sampling step length and timing verification standard to filter the continuous data segments, removes the time dislocation and abnormal interruption nodes, organizes and archives the sampled stream data, and generates the instantaneous sampling data set; The periodic integration processing submodule extracts the instantaneous value data pairs of current and voltage in each period based on the instantaneous sampling data set, accumulates the energy contribution according to the corresponding product integral formula, normalizes the influence of the sampling time step, organizes and archives the calculation results of each period, and generates a periodic energy integration data set; The energy-time binding submodule extracts the energy value and the corresponding time tag of each cycle according to the periodic energy integral data set, binds the energy data and the time node according to the time sequence, archives the energy parameters and the time information flow, organizes them into a data set for uploading, and generates energy parameters; The periodic integration is to perform an integration operation on the square values ​​of the instantaneous values ​​of partial discharge, current and voltage within a 50 millisecond time window.

4. The intelligent sensing online monitoring system for ring main unit according to claim 3 is characterized in that: The mutation analysis module comprises: The mutation rate calculation submodule receives the energy parameter, calculates the energy change of adjacent time windows based on the sliding window difference algorithm, forms the mutation rate, combines the 5%-20% mutation rate threshold, optimizes the calculation through the gradient descent method, and generates the mutation rate calculation result; The threshold matching submodule performs matching according to the mutation rate calculation result and the optimized mutation rate threshold, extracts the abnormal fragments that meet the conditions, and generates the abnormal fragment identification; The abnormal feature extraction submodule extracts the energy change direction and mutation amplitude based on the abnormal segment identifier, and outputs the abnormal feature to the abnormal locking module; The mutation rate threshold parameters include energy mutation rate 5%-20%, offset threshold set to ±20%, and fluctuation rate ≥±10%; The energy mutation rate is determined by analyzing historical data, observing the energy change characteristics of normal and abnormal events, and comparing the energy fluctuations under normal conditions with the mutation amplitudes in abnormal events. The mutation rate of normal fluctuations is less than 5%, while the mutation rate of abnormal events is more than 20%; When the fluctuation rate operates normally, the fluctuation rate is between ±1% and ±5%, while the fluctuation rate caused by abnormal conditions exceeds ±10%. The influence of the fluctuation rate on abnormality detection is analyzed by the gradient descent method, and the fluctuation rate threshold is set to ±10%.

5. The intelligent sensing online monitoring system for ring main unit according to claim 4 is characterized in that: The abnormal locking module includes: The leakage current acquisition submodule obtains the abnormal characteristics, sets the insulation monitoring sampling starting point according to the time tag sequence, sets the continuous sampling step to collect the leakage current instantaneous value, screens out the data nodes with abnormal sampling time intervals, archives the valid time series samples, and generates the leakage current instantaneous sequence; The current feature extraction submodule extracts the maximum and average values ​​of the leakage current in each continuous sampling period based on the leakage current instantaneous sequence, calculates the internal feature variation of the period based on the maximum and average values, marks the abnormal sample segments, archives the valid feature variation value sequence, and generates a current feature variation data set; The offset calculation submodule calls the feature change value extracted in each time segment according to the current feature change data set, calculates the offset value according to the offset formula, removes the time segment samples with abnormal offset, archives the pairing relationship between the offset value and the time tag, and generates the offset; The triggering condition for triggering the insulation monitoring unit to collect the instantaneous value of the leakage current is that when the abnormal characteristic exceeds the mutation rate threshold, the insulation monitoring unit is triggered by a hardware interrupt signal.

6. The intelligent sensing online monitoring system for ring main unit according to claim 5 is characterized in that: The risk zoning module includes: The data window extraction submodule obtains the offset, divides the continuous window segments in chronological order, filters the data segments whose time span meets the requirements, removes the abnormal time interval segments, archives the continuous valid window data, and generates the window continuous data set; The fluctuation rate calculation submodule calls the offset data sequence based on the window continuous data set, calculates the amplitude change rate per unit time of the continuous time window, filters the amplitude change rate abnormal data segment, archives the effective rate, and generates the amplitude fluctuation rate sequence; The insulation level determination submodule calls the window segment offset data according to the amplitude fluctuation rate sequence, makes a judgment according to the multi-threshold interval standard, determines the corresponding relationship between the insulation level and the time label, archives the judgment result, and generates the insulation level; The fluctuation rate is calculated by sliding window difference method, the offset determines the amplitude abnormality, the fluctuation rate determines the trend stability, a low fluctuation rate and less than the offset indicates an excellent insulation grade, and a high fluctuation rate greater than the offset indicates a poor grade judgment; The sliding window difference method slides the time series data by setting a window of fixed length, calculating the difference between adjacent data points each time, with a window step of 1, and using the mean of the difference results to evaluate the fluctuation rate of the data; The so-called multiple thresholds include an offset threshold and a fluctuation rate threshold. The offset threshold is set to ±20%, and the fluctuation rate threshold is set to ≥±10%.

7. The intelligent sensing online monitoring system for ring main unit according to claim 6 is characterized in that: The system further comprises: The link synchronization module receives the insulation level, offset and energy parameters, calculates the synchronization error based on the comparison between the current cache value of the relay node and the received data, and if the error is qualified, encapsulates the energy parameter, offset and insulation level and uploads it to the master node; if it exceeds the link synchronization error threshold, triggers the supplementary sampling command and completes the data closed-loop processing; The recollection command generates a recollection task and transmits it to the data collection unit, executes recollection and verification of data, and uploads the recollection result to the master node to ensure data accuracy; The synchronization error comparison result specifically refers to a node synchronization error value, a synchronization error determination identifier, package verification data or a supplementary sampling instruction; The link synchronization error threshold is specifically that the difference between the received data and the cache value does not exceed ±3%.

8. The intelligent sensing online monitoring system for ring main unit according to claim 7 is characterized in that: The link synchronization module comprises: The synchronization error calculation submodule obtains the insulation level, offset and energy parameters, compares the current cache value of the relay node with the received data, calculates the error, and generates a synchronization error calculation result if the difference is less than the synchronization error threshold; The data upload encapsulation submodule calculates the synchronization error result, and if the synchronization error is qualified, the energy parameter, offset and insulation level are uploaded to the master node to generate the upload data result; The data supplementary sampling processing submodule determines whether the synchronization state exceeds the threshold according to the synchronization error calculation result. If the synchronization deviation rate value in the synchronization error calculation result is greater than the link synchronization error threshold, a supplementary sampling command is sent to the energy acquisition module to re-collect energy parameters and offsets, and call the original time label and insulation level for field compounding operations. After performing the same data processing operations as the encapsulation upload process, it is uploaded to the master node to obtain energy parameters, offsets and insulation levels.

9. A ring main unit intelligent sensing online monitoring method, characterized in that: The method is used to implement the intelligent sensing online monitoring system for the ring main unit according to any one of claims 1 to 8, comprising the following steps: S1: Collect the instantaneous value of partial discharge, instantaneous value of current and instantaneous value of voltage through the monitoring device, set 50 milliseconds as the period for integration processing to generate energy parameters, bind the time tag, output the energy parameters, and perform subsequent analysis and processing; S2: Based on the energy parameter, extract the difference in the integral values ​​of adjacent time windows, call the differential change rate function to calculate the mutation rate, make interval judgment based on the mutation rate and the set threshold, extract the change direction and change amplitude, and output the abnormal characteristics; S3: Based on the abnormal characteristics, collect the instantaneous value of the leakage current in a continuous time period, extract the maximum amplitude and the average amplitude of each group, calculate the difference between the two to construct an amplitude offset, and output the amplitude offset; S4: Based on the amplitude offset, call the amplitude ratio of two consecutive time windows, input it into the continuous change rate function to calculate the amplitude fluctuation rate, perform multi-threshold judgment according to the amplitude offset and the amplitude fluctuation rate, and output the insulation level; S5: Based on the insulation level, combined with the amplitude offset and the energy parameter, call the relay node cache value to perform comparison processing, input it into the mean square error function to calculate the synchronization error, and determine whether the error meets the upload conditions. If it meets the conditions, package it and upload it. If not, generate a supplementary sampling instruction.

Images