Equipment defect active early warning method and device based on automated big data

Abstract

The present invention proposes an active early warning method and device for equipment defects based on automated big data, including: obtaining corresponding overcurrent alarm records and recorded waveforms when equipment in the distribution network is running; screening out overcurrent alarm records with flashover defects; determining the distribution line where the equipment involved in the overcurrent alarm record is located, and determining the flashover defect range based on the topological relationship of the distribution line and the generation of the overcurrent alarm record; analyzing whether the recorded waveform meets the characteristics of the flashover defect, combining the defect influencing factors, calculating the risk value, and conducting a partial discharge test on the distribution line based on the size of the risk value to obtain an early warning result. The present invention does not require empirical data. After collecting overcurrent alarm records generated based on real-time automated big data, it filters the characteristics of flashover hazards and can be directly used for analyzing and judging flashover hazards that are difficult to directly discover. It eliminates the need for machine learning or experience accumulation processes and improves the efficiency of hazard warnings.

Description

Technical Field

[0001] The present invention belongs to the field of power fault monitoring, and in particular relates to an active early warning method and device for equipment defects based on automated big data. Background Art

[0002] Flashover is the phenomenon of electrical discharge along the surface of a solid insulator when the gas or liquid dielectric surrounding it breaks down. Because flashovers are difficult to detect in real time, existing methods typically require building a fault case database based on historical electrical data from equipment in the distribution network. Machine learning or manual experience is then used to determine whether the equipment's real-time electrical data matches the fault case database in order to monitor equipment flashover defects.

[0003] When using the above method to analyze flashover defects in electrical equipment, a large amount of historical electrical data needs to be accumulated to build a more accurate set of fault cases. In addition, once the power equipment and line topology change, a large amount of historical electrical data needs to be re-collected. The process is tedious and complicated, which reduces the efficiency of defect analysis of electrical equipment. Summary of the Invention

[0004] In order to address the shortcomings and deficiencies in the prior art, the present invention proposes an active early warning method for equipment defects based on automated big data, comprising:

[0005] Obtaining automated big data during the operation of equipment in the distribution network, the automated big data including instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records;

[0006] Filter out instantaneous overcurrent alarm records with flashover defects;

[0007] Determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determine the scope of the flashover defect based on the topology of the distribution line and the generation of the instantaneous overcurrent alarm record;

[0008] Analyze whether the recorded waveform conforms to the characteristics of the flashover defect. Combined with the influencing factors of the defect, calculate the risk value of the equipment within the flashover defect range. Based on the size of the risk value, conduct partial discharge tests on the distribution line to obtain early warning results.

[0009] Optionally, the obtaining of automated big data during operation of equipment in the distribution network, wherein the automated big data includes instantaneous overcurrent alarm records and recorded waveforms having corresponding relationships, includes:

[0010] The three-phase current and zero-sequence current telemetry values ​​of the equipment are obtained through the three-remote DTU. When the telemetry value reaches the preset 0-second time limit instantaneous overcurrent condition, the fault recorder is triggered to collect the recorded waveform of the equipment until the telemetry value returns to normal;

[0011] Generate instantaneous overcurrent alarm records and send them together with the recorded waveform to the distribution automation master station.

[0012] Optionally, screening out instantaneous overcurrent alarm records with flashover defects includes:

[0013] Eliminate instantaneous overcurrent alarm records that do not conform to the preset topology relationship;

[0014] Obtain the action records and reset records from the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record, and record it as the reset time;

[0015] The instantaneous overcurrent alarm records with a reset time less than a preset threshold are screened out, and the instantaneous overcurrent alarm records are marked as instantaneous overcurrent alarm records with flashover defects.

[0016] Optionally, determining the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determining the flashover defect range according to the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record, includes:

[0017] In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records;

[0018] The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range;

[0019] The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

[0020] Optionally, the analysis of whether the recorded waveform meets the characteristics of the flashover defect, combining the defect influencing factors, calculating the risk value of the equipment within the flashover defect range, and conducting a partial discharge test on the distribution line according to the risk value to obtain an early warning result, including:

[0021] Input the recorded waveform corresponding to the instantaneous overcurrent alarm record into the pre-trained neural network model to analyze whether the characteristics of the recorded waveform meet the characteristics of the flashover fault;

[0022] Whether the recorded waveform conforms to the flashover fault characteristics is combined with weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time as defect influencing factors;

[0023] Based on the preset impact weights, the defect impact factors are weighted and summed, and the weighted calculation result is used as the risk value;

[0024] According to the order of risk values ​​from high to low, partial discharge tests are carried out on each section of the distribution line within the flashover defect range in turn using a partial discharge detection instrument, and the obtained partial discharge value is output as the end of the early warning.

[0025] Based on the same idea, the present invention also proposes an active early warning device for equipment defects based on automated big data, comprising:

[0026] Generating unit: used to obtain automation big data of equipment in the distribution network during operation, wherein the automation big data includes instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records;

[0027] Screening unit: used to screen out instantaneous overcurrent alarm records with flashover defects;

[0028] Positioning unit: used to determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determine the flashover defect range based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record;

[0029] Calculation unit: used to analyze whether the recorded waveform meets the characteristics of the flashover defect, calculate the risk value of the equipment within the flashover defect range based on the defect influencing factors, and conduct partial discharge tests on the distribution line according to the size of the risk value to obtain early warning results.

[0030] Optionally, the generating unit is specifically configured to:

[0031] The three-phase current and zero-sequence current telemetry values ​​of the equipment are obtained through the three-remote DTU. When the telemetry value reaches the preset 0-second time limit instantaneous overcurrent condition, the fault recorder is triggered to collect the recorded waveform of the equipment until the telemetry value returns to normal;

[0032] Generate instantaneous overcurrent alarm records and send them together with the recorded waveform to the distribution automation master station.

[0033] Optionally, the screening unit is specifically used to:

[0034] Eliminate instantaneous overcurrent alarm records that do not conform to the preset topology relationship;

[0035] Obtain the action records and reset records from the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record, and record it as the reset time;

[0036] The instantaneous overcurrent alarm records with a reset time less than a preset threshold are screened out, and the instantaneous overcurrent alarm records are marked as instantaneous overcurrent alarm records with flashover defects.

[0037] Optionally, the positioning unit is specifically configured to:

[0038] In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records;

[0039] Determine the first and last load switches that generate instantaneous overcurrent alarm records respectively;

[0040] The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range;

[0041] The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

[0042] Optionally, the computing unit is specifically configured to:

[0043] Input the recorded waveform corresponding to the instantaneous overcurrent alarm record into the pre-trained neural network model to analyze whether the characteristics of the recorded waveform meet the characteristics of the flashover fault;

[0044] Whether the recorded waveform conforms to the flashover fault characteristics is combined with weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time as defect influencing factors;

[0045] Based on the preset impact weights, the defect impact factors are weighted and summed, and the weighted calculation result is used as the risk value;

[0046] According to the order of risk values ​​from high to low, partial discharge tests are carried out on each section of the distribution line within the flashover defect range in turn using a partial discharge detection instrument, and the obtained partial discharge value is output as the end of the early warning.

[0047] The beneficial effects brought about by the technical solution provided by the present invention are:

[0048] (1) The active warning method for equipment defects proposed in the present invention does not require empirical data. After collecting the instantaneous overcurrent alarm records generated based on real-time automated big data, it is filtered according to the characteristics of flashover defects. It can be directly used to analyze and judge the defect faults such as flashover that are difficult to directly detect, eliminating the machine learning or experience accumulation process and improving the efficiency of defect warning.

[0049] (2) The present invention uses the ability to accurately locate equipment defects based on the uploaded overcurrent alarm information combined with the line topology, narrowing the scope of further judging the flashover risk, reducing the amount of calculation, and facilitating operators to quickly determine the impact range of the flashover fault.

[0050] (3) The present invention screens out fault data without instantaneous characteristics caused by loop reverse load and existing faults during normal operation before data analysis, thereby ensuring the accuracy of flashover defect analysis.

[0051] (4) After determining the flashover defect range of the suspected flashover fault, the present invention automatically triggers the fault recorder to record the waveform, and then compares whether it is consistent with short-term discharges such as flashover, thereby increasing the credibility of the data analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 This is a flowchart of an active early warning method for equipment defects based on automated big data proposed in one embodiment of the present invention;

[0054] Figure 2 This is a structural block diagram of an active early warning device for equipment defects based on automated big data proposed in another embodiment of the present invention. DETAILED DESCRIPTION

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0056] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in sequences other than those illustrated or described herein.

[0057] It should be understood that in various embodiments of the present invention, the size of the sequence number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0058] It should be understood that in the present invention, "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0059] It should be understood that in the present invention, "multiple" refers to two or more. "And / or" is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "Contains A, B and C", "Contains A, B, C" means that A, B, and C are all included, "Contains A, B or C" means that one of A, B, and C is included, and "Contains A, B and / or C" means that any one, any two, or any three of A, B, and C are included.

[0060] It should be understood that, in the present invention, "B corresponding to A," "B corresponding to A," "A corresponds to B," or "B corresponds to A" means that B is associated with A and B can be determined based on A. Determining B based on A does not mean determining B based solely on A; B can also be determined based on A and / or other information. A and B match when the similarity between A and B is greater than or equal to a preset threshold.

[0061] Depending on the context, "if" as used herein may be interpreted as "when" or "when" or "in response to determining" or "in response to detecting."

[0062] The following specific embodiments are used to describe the technical solution of the present invention in detail. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0063] Example 1

[0064] like Figure 1 As shown, this embodiment proposes an active early warning method for equipment defects based on automated big data, including:

[0065] S1: Obtaining automation big data during the operation of equipment in the distribution network, wherein the automation big data includes instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records;

[0066] S2: Filter out instantaneous overcurrent alarm records with flashover defects;

[0067] S3: Determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determine the flashover defect range based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record;

[0068] S4: Analyze whether the recorded waveform conforms to the characteristics of the flashover defect. Combined with the factors affecting the defect, calculate the risk value of the equipment within the flashover defect range. According to the size of the risk value, conduct partial discharge test on the distribution line to obtain early warning results.

[0069] This embodiment does not require the accumulation of a large amount of experience data, but can directly use the real-time automated big data of the equipment to analyze fault defects, eliminating the tedious process of machine learning or experience accumulation for different distribution networks.

[0070] In this embodiment, the automatic big data of the equipment in the distribution network during operation is obtained, and the automatic big data includes instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, including:

[0071] The three-phase current and zero-sequence current telemetry values ​​of the device are obtained through the three-remote DTU. The three-remote DTU refers to a data transfer unit (DTU) with three functions: telemetry, telesignaling, and remote control. Telemetry means remotely measuring the three-phase voltage, three-phase current, battery voltage, capacitor voltage, and zero-sequence current of the device. In this embodiment, telemetry data is used to monitor whether the device has faults in real time.

[0072] When the telemetry value reaches the preset 0-second instantaneous overcurrent condition, the fault recorder is triggered to collect the device's recorded waveform until the telemetry value returns to normal. In this embodiment, the 0-second instantaneous overcurrent condition is achieved when the telemetry value reaches 700A and persists for 0ms, triggering an alarm. In practical applications, this occurs when the telemetry value reaches 700A and persists for 5ms (5ms is considered the device's accuracy time). This demonstrates that this embodiment achieves indiscriminate maximum quantization to capture overcurrent signals, sending an alarm regardless of the duration of the overcurrent. Existing systems typically only send an alarm when the telemetry value persists for 100ms. Compared to existing systems, this generates more instantaneous overcurrent alarm records, improving the accuracy of subsequent risk analysis. Simultaneously, the three-remote DTU generates an instantaneous overcurrent alarm record, recording the duration of the recorded waveform as the reset time for the instantaneous overcurrent alarm record. The reset time and the recorded waveform are output together as the instantaneous overcurrent alarm record. In this embodiment, the uploaded instantaneous overcurrent alarm record is displayed in a table format, which respectively displays the action record and reset record of the instantaneous overcurrent alarm record, the name of the switch station where the equipment is located, the name of the transmission line where the equipment is located, and the weather. A hyperlink to the recorded waveform image is attached in the last column of the table to establish a correspondence between the instantaneous overcurrent alarm record and the recorded waveform. The operator can click on the hyperlink to view the recorded waveform during the overcurrent alarm. The recorded waveform will then be compared to see if it is consistent with the short-term discharge of the flashover fault, thereby increasing the credibility of the data analysis results.

[0073] Then, relying on the remote signaling function of the three remote DTUs, the instantaneous overcurrent alarm record and the recorded waveform are sent to the distribution automation master station. The distribution automation master station in this embodiment is set in the Supervisory Control And Data Acquisition (SCADA) system. The SCADA system is widely used in power systems and is used to collect big data from the equipment running on the distribution network. It usually consists of three parts: the master station, the communication system and the remote terminal unit.

[0074] The main purpose of this embodiment is to capture hidden flashover faults within the equipment. These faults will not produce external manifestations in the overall operation of the distribution network, but will gradually worsen as the flashover fault lasts longer. The power company analyzed the pre-alarm signals of multiple actual cable head, busbar, and cable intermediate joint burnout faults and found that they all matched abnormal spike waveforms. The earliest alarm was captured ranged from 10 days to 7 months before the fault, and only one alarm signal occurred 15 minutes before the accident. This shows that most of the time there is sufficient time to correct the fault after the signal is captured. Therefore, this embodiment uses the overcurrent alarm signal and its corresponding recorded waveform to realize the function of pre-determining equipment flashover faults.

[0075] In order to effectively monitor the flashover fault of the equipment, it is necessary to eliminate the signals generated in normal operation and the signals corresponding to the fault. Specifically, it includes:

[0076] Eliminate instantaneous overcurrent alarm records that do not conform to the preset topological relationship; obtain the action records and reset records in the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record and record it as the reset time; screen out instantaneous overcurrent alarm records with a reset time less than a preset threshold, and mark the instantaneous overcurrent alarm record as an instantaneous overcurrent alarm record with a flashover defect.

[0077] In this embodiment, instantaneous overcurrent alarm records that do not conform to the preset topology are primarily those that have errors with the actual topology, and the preset topology is manually set. Instantaneous overcurrent alarm records commonly contain overcurrent alarm records caused by operators performing normal operation operations such as closing and switching certain equipment in the distribution network. For example, when a line closing and switching operation is performed, a current greater than 720A causes an overcurrent to be transmitted; when a DTU telemetry loop is debugged and increased, an overcurrent signal is transmitted; and when a line overload interval current exceeds the 480A set value corresponding to a 400 / 5CT for a long period of time, a signal is transmitted. Instantaneous overcurrent alarm records generated in these situations can be considered redundant data and removed. In this embodiment, the common feature of these redundant data is that no matter whether the closing and switching operation is performed manually on-site or by dispatching remote control, the closing time is at least more than 5 seconds. Similarly, the time for increasing the amount of the overcurrent signal for telemetering debugging is generally more than 5 seconds, that is, the reset time of the above-mentioned types of signals is greater than 5 seconds. Therefore, the instantaneous overcurrent alarm records with a reset time greater than 5 seconds can be regarded as redundant data. After the data is raised, the remaining instantaneous overcurrent alarm records are all marked as instantaneous overcurrent alarm records with flashover defects.

[0078] After screening out the transient overcurrent alarm records that may have flashover defects, in order to more accurately determine the location of the flashover defect, the flashover defect range will be determined based on the topological relationship of the distribution line and the generation of the transient overcurrent alarm records. Specifically, the following are included:

[0079] In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records;

[0080] The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range;

[0081] The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

[0082] After eliminating the instantaneous overcurrent alarm records with a reset time greater than the preset threshold, other fault records that may interfere with the flashover fault judgment still exist in the remaining instantaneous overcurrent alarm records. In this embodiment, these faults include actual line short-circuit faults that cause substation tripping, line equipment defect flashovers that cause line instantaneous overcurrent, such as the operation of the 10kV outgoing line switch of a 110kV substation, etc. These actual faults usually have a reset time of less than 1.5s under the action of the distribution network's own protection devices, and the overcurrent reset time corresponding to the equipment flashover is even shorter, which will affect the judgment of the flashover defect.

[0083] To address this issue and improve the accuracy of early warning results, this embodiment inputs the recorded waveform corresponding to the instantaneous overcurrent alarm record into a pre-trained neural network model to analyze whether the characteristics of the recorded waveform meet the characteristics of a flashover fault. By machine training the neural network model, waveform automatic judgment is achieved, accurately capturing the corresponding abnormal waveform in the massive data waveform. A neural network algorithm refers to a class of algorithms designed to simulate the learning process of the human brain. It is usually composed of an input layer, a hidden layer, and an output layer. Each layer contains multiple neurons, and all neurons are fully connected. It can simulate the complex nonlinear relationship between the input feature set (such as the characteristics of the recorded waveform) and the expected output value (such as whether the recorded waveform is a spike waveform unique to a flashover fault) and can further provide effective predictions for new input features. In this embodiment, the neural network model is pre-trained. During the training phase, the spike waveforms generated by each device during a flashover fault and the current waveform during normal operation are collected separately. The collected spike waveforms and current waveforms are then extracted through fast Fourier transform to obtain a waveform feature value sequence. The feature value sequence includes parameters such as the peak, trough, frequency, effective value, and duty cycle of the waveform. The feature value sequence can describe the shape of the waveform. These characteristic value sequences are input as training sets into the neural network model for training. During the training process, the weight parameters of each node in the neural network are continuously adjusted so that the neural network model can distinguish the sharp wave waveform among many waveforms. According to the analysis results, the waveform characteristics are determined to be consistent with the recorded waveform of the flashover fault. The alarm record corresponding to the recorded waveform is marked as an instantaneous overcurrent alarm record with a flashover defect, and whether the recorded waveform meets the flashover fault characteristics is used as one of the defect influencing factors.

[0084] Subsequently, considering the close correlation between weather factors and flashover faults, defect influencing factors such as weather factors were set. In addition, in order to further improve the accuracy of the early warning risk value, the total number of instantaneous overcurrent alarm records within the flashover defect range and the growth rate of the number of instantaneous overcurrent alarm records per unit time were set as defect influencing factors. The total number of instantaneous overcurrent alarm records and the growth rate mainly reflect the degree of influence of the flashover defect on each device within the flashover defect range. The more instantaneous overcurrent alarm records are generated, the more devices are affected in the flashover defect range, and the greater the risk of flashover fault. Combined with other preset defect influencing factors, including weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time, they are jointly used as defect influencing factors. Based on the preset influence weights, the defect influencing factors are weighted and summed, and the weighted calculation result is used as the risk value.

[0085] The order of partial discharge testing for distribution lines is then determined based on the risk values, from high to low. The higher the risk value, the earlier the test is ordered. The partial discharge detection instrument then performs partial discharge tests on each section of the distribution line within the flashover defect range, and the resulting partial discharge value is output as a warning end. Distribution network operators can then determine the severity of the flashover defect based on the partial discharge value and risk value. This allows them to determine whether to upgrade, repair, or replace the equipment within the flashover defect range, significantly reducing the risk of flashover faults affecting the normal operation of the distribution network.

[0086] Example 2

[0087] like Figure 2 As shown, this embodiment proposes an active early warning device 5 for equipment defects based on automated big data, comprising:

[0088] Generating unit 51: acquiring automation big data of equipment in the distribution network during operation, wherein the automation big data includes instantaneous overcurrent alarm records and recorded waveforms having corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records;

[0089] Screening unit 52: used to screen out instantaneous overcurrent alarm records with flashover defects;

[0090] Positioning unit 53: used to determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determine the flashover defect range based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record;

[0091] Calculation unit 54 is used to analyze whether the recorded waveform meets the characteristics of the flashover defect, calculate the risk value of the equipment within the flashover defect range based on the defect influencing factors, and perform partial discharge tests on the distribution line according to the risk value to obtain early warning results.

[0092] This embodiment does not require the accumulation of a large amount of experience data, but can directly use the real-time automated big data of the equipment to analyze fault defects, eliminating the tedious process of machine learning or experience accumulation for different distribution networks.

[0093] In this embodiment, the generating unit 51 is specifically configured to:

[0094] The three-phase current and zero-sequence current telemetry values ​​of the device are obtained through the three-remote DTU. The three-remote DTU refers to a data transfer unit (DTU) with three functions: telemetry, telesignaling, and remote control. Telemetry means remotely measuring the three-phase voltage, three-phase current, battery voltage, capacitor voltage, and zero-sequence current of the device. In this embodiment, telemetry data is used to monitor whether the device has faults in real time.

[0095] When the telemetry value reaches the preset 0-second instantaneous overcurrent condition, the fault recorder is triggered to collect the device's recorded waveform until the telemetry value returns to normal. In this embodiment, the 0-second instantaneous overcurrent condition is achieved when the telemetry value reaches 700A and persists for 0ms, triggering an alarm. In practical applications, this occurs when the telemetry value reaches 700A and persists for 5ms (5ms is considered the device's accuracy time). This demonstrates that this embodiment achieves indiscriminate maximum quantization to capture overcurrent signals, sending an alarm regardless of the duration of the overcurrent. Existing systems typically only send an alarm when the telemetry value persists for 100ms. Compared to existing systems, this generates more instantaneous overcurrent alarm records, improving the accuracy of subsequent risk analysis. Simultaneously, the three-remote DTU generates an instantaneous overcurrent alarm record, recording the duration of the recorded waveform as the reset time for the instantaneous overcurrent alarm record. The reset time and the recorded waveform are output together as the instantaneous overcurrent alarm record. In this embodiment, the uploaded instantaneous overcurrent alarm record is displayed in a table format, which respectively displays the action record and reset record of the instantaneous overcurrent alarm record, the name of the switch station where the equipment is located, the name of the transmission line where the equipment is located, and the weather. A hyperlink to the recorded waveform image is attached in the last column of the table to establish a correspondence between the instantaneous overcurrent alarm record and the recorded waveform. The operator can click on the hyperlink to view the recorded waveform during the overcurrent alarm. The recorded waveform will then be compared to see if it is consistent with the short-term discharge of the flashover fault, thereby increasing the credibility of the data analysis results.

[0096] Then, relying on the remote signaling function of the three remote DTUs, the instantaneous overcurrent alarm record and the recorded waveform are sent to the distribution automation master station. The distribution automation master station in this embodiment is set in the Supervisory Control And Data Acquisition (SCADA) system. The SCADA system is widely used in power systems and is used to collect big data from the equipment running on the distribution network. It usually consists of three parts: the master station, the communication system and the remote terminal unit.

[0097] The main purpose of this embodiment is to capture hidden flashover faults within the equipment. These faults will not produce external manifestations in the overall operation of the distribution network, but will gradually worsen as the flashover fault lasts longer. The power company analyzed the pre-alarm signals of multiple actual cable head, busbar, and cable intermediate joint burnout faults and found that they all matched abnormal spike waveforms. The earliest alarm was captured ranged from 10 days to 7 months before the fault, and only one alarm signal occurred 15 minutes before the accident. This shows that most of the time there is sufficient time to correct the fault after the signal is captured. Therefore, this embodiment uses the overcurrent alarm signal and its corresponding recorded waveform to realize the function of pre-determining equipment flashover faults.

[0098] In order to effectively monitor the flashover fault of the equipment, it is necessary to filter out the signals generated in the normal operation mode and the signals corresponding to the fault by the screening unit 52, specifically for:

[0099] Eliminate instantaneous overcurrent alarm records that do not conform to the preset topological relationship; obtain the action records and reset records in the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record and record it as the reset time; screen out instantaneous overcurrent alarm records with a reset time less than a preset threshold, and mark the instantaneous overcurrent alarm record as an instantaneous overcurrent alarm record with a flashover defect.

[0100] In this embodiment, instantaneous overcurrent alarm records that do not conform to the preset topology are primarily those that have errors with the actual topology, and the preset topology is manually set. Instantaneous overcurrent alarm records commonly contain overcurrent alarm records caused by operators performing normal operation operations such as closing and switching certain equipment in the distribution network. For example, when a line closing and switching operation is performed, a current greater than 720A causes an overcurrent to be transmitted; when a DTU telemetry loop is debugged and increased, an overcurrent signal is transmitted; and when a line overload interval current exceeds the 480A set value corresponding to a 400 / 5CT for a long period of time, a signal is transmitted. Instantaneous overcurrent alarm records generated in these situations can be considered redundant data and removed. In this embodiment, the common feature of these redundant data is that no matter whether the closing and switching operation is performed manually on-site or by dispatching remote control, the closing time is at least more than 5 seconds. Similarly, the time for increasing the amount of the overcurrent signal for telemetering debugging is generally more than 5 seconds, that is, the reset time of the above-mentioned types of signals is greater than 5 seconds. Therefore, the instantaneous overcurrent alarm records with a reset time greater than 5 seconds can be regarded as redundant data. After the data is raised, the remaining instantaneous overcurrent alarm records are all marked as instantaneous overcurrent alarm records with flashover defects.

[0101] After screening out the instantaneous overcurrent alarm records that may have flashover defects, in order to more accurately determine the location of the flashover defect, the flashover defect range is determined by the positioning unit 53, specifically including:

[0102] In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records;

[0103] The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range;

[0104] The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

[0105] After eliminating the instantaneous overcurrent alarm records with a reset time greater than the preset threshold, other fault records that may interfere with the flashover fault judgment still exist in the remaining instantaneous overcurrent alarm records. In this embodiment, these faults include actual line short-circuit faults that cause substation tripping, line equipment defect flashovers that cause line instantaneous overcurrent, such as the operation of the 10kV outgoing line switch of a 110kV substation, etc. These actual faults usually have a reset time of less than 1.5s under the action of the distribution network's own protection devices, and the overcurrent reset time corresponding to the equipment flashover is even shorter, which will affect the judgment of the flashover defect.

[0106] In order to solve this problem and improve the accuracy of the early warning results, this embodiment inputs the recorded waveform corresponding to the instantaneous overcurrent alarm record into a pre-trained neural network model through the calculation unit 54, and analyzes whether the characteristics of the recorded waveform meet the characteristics of the flashover fault. By machine training the neural network model, automatic waveform judgment is achieved, and the corresponding abnormal waveform is accurately captured in the massive big data waveform. A neural network algorithm refers to a type of algorithm that aims to simulate the learning process of the human brain. It is usually composed of an input layer, a hidden layer, and an output layer. Each layer contains multiple neurons, and all neurons are fully connected. It can simulate the complex nonlinear relationship between the input feature set (such as the characteristics of the recorded waveform) and the expected output value (such as whether the recorded waveform is a spike waveform unique to a flashover fault), and can further provide effective predictions for new input features. In this embodiment, the neural network model is pre-trained. During the training phase, spike waveforms generated by each device during a flashover fault and current waveforms during normal operation are collected. Fast Fourier transform is then used to extract characteristic value sequences of the waveforms. The characteristic value sequences include parameters such as the peaks, troughs, frequency, effective value, and duty cycle of the waveforms. The characteristic value sequences can describe the shape of the waveforms. These characteristic value sequences are input into the neural network model as training sets for training. During the training process, the weight parameters of each node in the neural network are continuously adjusted, enabling the neural network model to distinguish spike waveforms from numerous waveforms. Based on the analysis results, a waveform recording with waveform characteristics consistent with a flashover fault is determined. The alarm record corresponding to the waveform recording is marked as an instantaneous overcurrent alarm record indicating a flashover defect. Whether the waveform recording conforms to the flashover fault characteristics is considered as one of the defect influencing factors.

[0107] Subsequently, considering the close correlation between weather factors and flashover faults, defect influencing factors such as weather factors were set. In addition, in order to further improve the accuracy of the early warning risk value, the total number of instantaneous overcurrent alarm records within the flashover defect range and the growth rate of the number of instantaneous overcurrent alarm records per unit time were set as defect influencing factors. The total number of instantaneous overcurrent alarm records and the growth rate mainly reflect the degree of influence of the flashover defect on each device within the flashover defect range. The more instantaneous overcurrent alarm records are generated, the more devices are affected in the flashover defect range, and the greater the risk of flashover fault. Combined with other preset defect influencing factors, including weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time, they are jointly used as defect influencing factors. Based on the preset influence weights, the defect influencing factors are weighted and summed, and the weighted calculation result is used as the risk value.

[0108] The order of partial discharge testing for distribution lines is then determined based on the risk values, from high to low. The higher the risk value, the earlier the test is ordered. The partial discharge detection instrument then performs partial discharge tests on each section of the distribution line within the flashover defect range, and the resulting partial discharge value is output as a warning end. Distribution network operators can then determine the severity of the flashover defect based on the partial discharge value and risk value. This allows them to determine whether to upgrade, repair, or replace the equipment within the flashover defect range, significantly reducing the risk of flashover faults affecting the normal operation of the distribution network.

[0109] The serial numbers in the above embodiments are for description only and do not represent the order of assembly or use of the components.

[0110] The above descriptions are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An active early warning method for equipment defects based on automated big data, characterized in that: include: Obtaining automated big data during the operation of equipment in the distribution network, the automated big data including instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records; Filter out instantaneous overcurrent alarm records with flashover defects; Determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located. Based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record, determine the starting and ending points based on the generation order of the load switch alarm record, and thus determine the scope of the flashover defect; Input the recorded waveform corresponding to the instantaneous overcurrent alarm record into the pre-trained neural network model to analyze whether the characteristics of the recorded waveform meet the characteristics of the flashover fault; Whether the recorded waveform conforms to the flashover fault characteristics is combined with weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time as defect influencing factors; Based on the preset impact weights, the defect impact factors are weighted and summed, and the weighted calculation result is used as the risk value; According to the order of risk values ​​from high to low, partial discharge tests are carried out on each section of the distribution line within the flashover defect range in turn using a partial discharge detection instrument, and the obtained partial discharge value and risk value are output as the end of the early warning.

2. The active early warning method for equipment defects based on automated big data according to claim 1 is characterized in that: The step of obtaining automated big data during operation of equipment in the distribution network, wherein the automated big data includes instantaneous overcurrent alarm records and recorded waveforms having corresponding relationships, includes: The three-phase current and zero-sequence current telemetry values ​​of the equipment are obtained through the three-remote DTU. When the telemetry value reaches the preset 0-second time limit instantaneous overcurrent condition, the fault recorder is triggered to collect the recorded waveform of the equipment until the telemetry value returns to normal; Generate instantaneous overcurrent alarm records and send them together with the recorded waveform to the distribution automation master station.

3. The active early warning method for equipment defects based on automated big data according to claim 1 is characterized in that: The screening out of instantaneous overcurrent alarm records with flashover defects includes: Eliminate instantaneous overcurrent alarm records that do not conform to the preset topology relationship; Obtain the action records and reset records from the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record, and record it as the reset time; The instantaneous overcurrent alarm records with a reset time less than a preset threshold are screened out, and the instantaneous overcurrent alarm records are marked as instantaneous overcurrent alarm records with flashover defects.

4. The active early warning method for equipment defects based on automated big data according to claim 1 is characterized in that: The determining of the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determining the flashover defect range based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record, includes: In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records; The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range; The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

5. The active early warning device for equipment defects based on automated big data is characterized by: include: Generating unit: used to obtain automation big data of equipment in the distribution network during operation, wherein the automation big data includes instantaneous overcurrent alarm records and recorded waveforms with corresponding relationships, wherein the instantaneous overcurrent alarm records include zero-sequence overcurrent alarm records and phase overcurrent alarm records; Screening unit: used to screen out instantaneous overcurrent alarm records with flashover defects; Positioning unit: used to determine the distribution line where the equipment involved in the instantaneous overcurrent alarm record is located, and determine the flashover defect range based on the topological relationship of the distribution line and the generation of the instantaneous overcurrent alarm record; Calculation unit: used to input the recorded waveform corresponding to the instantaneous overcurrent alarm record into the pre-trained neural network model, and analyze whether the characteristics of the recorded waveform meet the characteristics of the flashover fault; Whether the recorded waveform conforms to the flashover fault characteristics is combined with weather factors, the total number of instantaneous overcurrent alarm records within the flashover defect range, and the growth rate of the number of instantaneous overcurrent alarm records per unit time as defect influencing factors; Based on the preset impact weights, the defect impact factors are weighted and summed, and the weighted calculation result is used as the risk value; According to the order of risk values ​​from high to low, partial discharge tests are carried out on each section of the distribution line within the flashover defect range in turn using a partial discharge detection instrument, and the obtained partial discharge value is output as the end of the early warning.

6. The active early warning device for equipment defects based on automated big data according to claim 5 is characterized in that: The generating unit is specifically configured to: The three-phase current and zero-sequence current telemetry values ​​of the equipment are obtained through the three-remote DTU. When the telemetry value reaches the preset 0-second time limit instantaneous overcurrent condition, the fault recorder is triggered to collect the recorded waveform of the equipment until the telemetry value returns to normal; Generate instantaneous overcurrent alarm records and send them together with the recorded waveform to the distribution automation master station.

7. The active early warning device for equipment defects based on automated big data according to claim 5 is characterized in that: The screening unit is specifically used for: Obtain the action records and reset records from the remaining instantaneous overcurrent alarm records, calculate the generation time difference between the action record and the reset record, and record it as the reset time; The instantaneous overcurrent alarm records with a reset time less than a preset threshold are screened out, and the instantaneous overcurrent alarm records are marked as instantaneous overcurrent alarm records with flashover defects.

8. The active early warning device for equipment defects based on automated big data according to claim 5 is characterized in that: The positioning unit is specifically used for: In the order from the power supply side to the load side of the distribution line, determine in turn whether the load switches passing through generate instantaneous overcurrent alarm records; Determine the first and last load switches that generate instantaneous overcurrent alarm records respectively; The last load switch that generates an instantaneous overcurrent alarm record is regarded as the starting point of the flashover defect range; The first load switch that does not generate an instantaneous overcurrent alarm record is regarded as the end point of the flashover defect range.

Images