A distribution network primary and secondary fusion device adaptive fault detection system

Abstract

The present application relates to distribution network equipment fault detection technical field, especially distribution network primary and secondary fusion equipment adaptive fault detection system, the present application extracts performance characteristics and sorts according to the difference degree under normal and fault conditions by collecting primary side historical normal and fault working condition electrical data, constructs primary side performance characteristic sorting set;Through the collection secondary side closing action signal, the adjacent signal that meets the preset time difference condition is paired and sorted according to time difference;According to the sorting, the performance curve of the performance characteristic of the primary side in the corresponding period of each signal combination is collected and quantitatively extracted peak deviation and steady-state deviation;Abnormal risk performance characteristics are determined based on the quantification results, and the abnormal confidence is determined according to the sorting order, the comprehensive confidence is obtained by accumulating each abnormal confidence, and the line section is determined whether there is a fault and the fault type and position are output. Further, the reliability and accuracy of the fault detection result are improved.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology for distribution network equipment, and in particular to an adaptive fault detection system for integrated primary and secondary distribution network equipment. Background Technology

[0002] With the continuous advancement of intelligent transformation of power distribution networks, integrated primary and secondary circuit equipment has been widely used for sectional protection and fault isolation of distribution lines. The primary circuit is responsible for the transmission and disconnection of main power, while the secondary circuit is responsible for controlling, measuring, and protecting the operating status of the primary circuit. Existing fault detection technologies generally only judge the primary-side electrical data or secondary-side control signals of a single circuit breaker, without considering the cascading changes in the primary-side performance during the corresponding time periods of adjacent closing actions. This results in the reliability and accuracy of fault detection results failing to meet the requirements for high-reliability operation of power distribution networks.

[0003] For example, Chinese invention patent CN120742010A discloses a fault diagnosis method and system for a primary and secondary integrated ring main unit. By constructing electrical early warning packages, temperature early warning packages, and insulation early warning packages, it records the changes in physical state quantities before different alarms occur, including current change trends, temperature trends of each component, and insulation resistance change trends. When the real-time monitoring data matches the early warning package, an early warning can be triggered before the parameters reach the alarm threshold, thus buying time for maintenance personnel to intervene.

[0004] Existing technologies do not consider the electrical coupling and cascading effects between adjacent closing actions, and cannot comprehensively assess the confidence level of multiple anomalies within the closing action combination, resulting in insufficient reliability and accuracy of fault detection results. Summary of the Invention

[0005] To address this issue, the present invention provides an adaptive fault detection system for integrated primary and secondary distribution network equipment, which overcomes the problem that existing technologies do not consider the electrical coupling and cascading effects between adjacent closing actions, and cannot comprehensively assess the confidence level of multiple anomalies within the closing action combination, resulting in insufficient reliability and accuracy of fault detection results.

[0006] To achieve the above objectives, the present invention provides an adaptive fault detection system for integrated primary and secondary distribution network equipment, comprising: The primary side feature ranking construction module is used to collect electrical data under historical normal and fault conditions of the primary side, extract primary side performance features, and rank each performance feature according to the degree of difference between normal and fault conditions to construct a primary side performance feature ranking set. The secondary side action serialization module is used to collect the closing action signals of the secondary side equipment, pair the closing action signals that meet the preset time difference conditions according to the triggering order, and sort them according to the time difference of each combination of closing action signals. The feature extraction module is connected to the primary side feature sorting and construction module and the secondary side action serialization module respectively. It is used to collect the performance curves of the primary side performance characteristics within the corresponding time period of each closing action signal according to the order of the closing action signal combination and sorting, and to extract the features of the performance curves by feature quantization. An anomaly analysis module, which is connected to the feature extraction module, is used to determine the characteristics of anomaly risk based on the results of feature quantification extraction and to determine the anomaly confidence level. The fault detection module, which is connected to the anomaly analysis module, is used to accumulate the anomaly confidence of each anomaly risk performance feature in the closing action signal combination, determine whether there is a fault in the corresponding line section based on the result of the anomaly confidence accumulation, and output the fault type and the location of the corresponding line section.

[0007] Furthermore, the primary-side feature ranking construction module is used to construct a primary-side performance feature ranking set, including: The primary side feature sorting construction module is used to retrieve historical electrical data under normal and fault conditions of the primary side. The electrical data includes the peak value of the inrush current, the inrush current decay time, the voltage drop depth and the current distortion rate. Each electrical data is determined as a primary side performance feature. The primary side feature sorting construction module calculates the feature value difference of each primary side performance feature under historical normal operating conditions and fault operating conditions, and sorts each primary side performance feature in descending order of the relative difference rate of the feature value difference. The primary side feature sorting construction module determines the set of sorted primary side performance features as the primary side performance feature sorting set, and associates each primary side performance feature with a corresponding fault type label.

[0008] Furthermore, the secondary-side action serialization module is used to determine the closing action signal combination, including: The secondary-side action serialization module is used to collect the closing action signals of the secondary-side equipment within a preset monitoring period, record the trigger time of each closing action signal, and arrange the closing action signals in order from earliest to latest according to the trigger time. The secondary side action serialization module calculates the trigger time difference between two adjacent closing action signals, and pairs adjacent closing action signals whose time difference meets the preset time difference condition to form a closing action signal combination.

[0009] Furthermore, the secondary-side action serialization module is used to sort the signals according to the time difference of each adjacent closing action signal combination, including: The secondary side action serialization module is used to obtain the time difference between the two closing action signals in each closing action signal combination, and sort the closing action signal pairs in order of increasing time difference.

[0010] Furthermore, the feature extraction module is used to acquire performance curves according to the order of the closing action signal combinations, wherein: The feature extraction module processes each closing action signal combination in the order of the closing action signal combination sorting. This includes, for the currently processed closing action signal combination, collecting real-time electrical data corresponding to each performance feature in the primary side performance feature sorting set during the time period from the trigger time of the first closing action signal to the trigger time of the last closing action signal. The feature extraction module plots the performance curves of each performance feature over time based on the real-time electrical data.

[0011] Furthermore, the feature extraction module is used to perform feature quantization extraction on the performance curve, including: For each performance curve, the feature extraction module is used to identify the peak point in the performance curve, calculate the difference between the feature value of the peak point and the peak value of the corresponding performance feature under normal working conditions, and determine the difference as the peak deviation. The feature extraction module is used to identify the steady-state segment in the curve, calculate the difference between the average feature value of the steady-state segment and the steady-state value of the corresponding performance feature under normal operating conditions, and determine the difference as the steady-state offset.

[0012] Furthermore, the anomaly analysis module is also used to determine the characteristics of anomaly risk manifestations and to determine the anomaly confidence level, including: The anomaly analysis module is used to compare the peak deviation of each performance feature with a preset peak deviation threshold, and to compare the steady-state offset of each performance feature with a preset steady-state offset threshold. If the peak deviation exceeds the preset peak deviation threshold, or the steady-state deviation exceeds the preset steady-state deviation threshold, the corresponding performance characteristic will be judged as an abnormal risk performance characteristic. The anomaly analysis module is also used to obtain the ranking order of the abnormal risk performance characteristics in the primary side performance characteristic ranking set, and to determine the corresponding anomaly confidence level based on the ranking order; Among them, the earlier the sorting order, the higher the confidence level of the anomaly.

[0013] Furthermore, the fault detection module is used to accumulate the anomaly confidence level, including: The fault detection module identifies all abnormal risk characteristics corresponding to the combination of closing action signals, and sums up the abnormal confidence levels corresponding to all abnormal risk characteristics to obtain the comprehensive confidence level.

[0014] Furthermore, the fault detection module is used to compare the overall confidence level of the closing action signal combination with a preset confidence level reference value to determine whether a fault exists in the corresponding line section, including: If the overall confidence level is greater than or equal to the confidence level reference value, the fault detection module determines that there is a fault in the line section corresponding to the closing action signal combination.

[0015] Furthermore, the fault detection module is used to output the fault type and the corresponding line segment location, including: The fault detection module selects the abnormal risk performance feature with the highest abnormal confidence among all abnormal risk performance features corresponding to the closing action signal combination, and uses the fault type label associated with the abnormal risk performance feature with the highest abnormal confidence in the primary side performance feature sorting set as the fault type. The fault detection module determines the start and end positions of the faulty line section as the corresponding line section position based on the circuit breaker positions corresponding to the first and last closing action signals in the closing action signal combination.

[0016] The beneficial effects of the technical solution presented in this application include: constructing a baseline set by sorting primary side features according to their distinguishing ability, giving higher weight to features with strong distinguishing ability in fault determination and avoiding interference from weak distinguishing features; serializing adjacent closing actions by pairing and sorting them by time difference, and analyzing the sequential closing of multiple circuit breakers as a whole, capturing the electrical coupling and cascading effects between adjacent closing actions; and determining the anomaly confidence by quantifying the peak deviation and steady-state offset of the performance curve, combining the feature sorting order to obtain the comprehensive confidence, and achieving a comprehensive evaluation of multi-feature anomalies, thereby improving the fault section location accuracy, reducing the false alarm and missed alarm rate, and improving the reliability and accuracy of fault detection.

[0017] Furthermore, by calculating the difference in characteristic values ​​of each primary side performance feature under normal and fault conditions and sorting them from largest to smallest, this invention can objectively quantify the fault discrimination capability of each electrical feature, avoid interference from normal random fluctuations in features with weak discrimination capability, and thus improve the scientific nature of fault feature selection and the accuracy of fault judgment.

[0018] Furthermore, this invention sorts the closing action signal combinations according to the time difference from small to large, enabling the system to prioritize the processing of signal combinations with the strongest electrical coupling, the most significant superposition of transient processes, and the densest fault information, and to prioritize the allocation of computing resources to the analysis object with the largest amount of fault information, thereby improving the real-time performance of fault detection and the capture efficiency of critical faults.

[0019] Furthermore, this invention determines the anomaly confidence level based on the ranking order of the abnormal risk performance characteristics in the primary side performance characteristic ranking set, transforming the fault discrimination ability of the features into the credibility of the anomaly judgment. This allows features with strong discrimination ability to obtain higher confidence levels when they exhibit anomalies, and features with weak discrimination ability to obtain lower confidence levels when they exhibit anomalies, thereby effectively distinguishing between high-confidence anomalies and low-confidence anomalies and avoiding triggering fault alarms by a single weak discrimination feature anomaly.

[0020] Furthermore, this invention sums up the anomaly confidence levels of each abnormal risk performance feature in the same closing action signal combination to obtain a comprehensive confidence level, thereby achieving weighted fusion of all abnormal evidence within the signal combination. The level of comprehensive confidence level directly reflects the sufficiency of abnormal evidence and the magnitude of the fault probability, thus improving the accuracy of comprehensive fault assessment. Attached Figure Description

[0021] Figure 1 This is a system block diagram of the adaptive fault detection system for the primary and secondary integrated distribution network equipment according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the logic of determining the characteristics of abnormal risks and establishing the confidence level of abnormality in an embodiment of the present invention. Figure 3 This is a flowchart illustrating the logic of determining whether a fault exists in a corresponding line section according to an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0025] Please see Figure 1 The diagram shown is a system block diagram of the adaptive fault detection system for primary and secondary distribution network integrated equipment according to an embodiment of the present invention. The adaptive fault detection system for primary and secondary distribution network integrated equipment of the present invention includes: The primary side feature ranking construction module is used to collect electrical data under historical normal and fault conditions of the primary side, extract primary side performance features, and rank each performance feature according to the degree of difference between normal and fault conditions to construct a primary side performance feature ranking set. The secondary side action serialization module is used to collect the closing action signals of the secondary side equipment, pair the closing action signals that meet the preset time difference conditions according to the triggering order, and sort them according to the time difference of each combination of closing action signals. The feature extraction module is connected to the primary side feature sorting and construction module and the secondary side action serialization module respectively. It is used to collect the performance curves of the primary side performance characteristics within the corresponding time period of each closing action signal according to the order of the closing action signal combination and sorting, and to extract the features of the performance curves by feature quantization. An anomaly analysis module, which is connected to the feature extraction module, is used to determine the characteristics of anomaly risk based on the results of feature quantification extraction and to determine the anomaly confidence level. The fault detection module, which is connected to the anomaly analysis module, is used to accumulate the anomaly confidence of each anomaly risk performance feature in the closing action signal combination, determine whether there is a fault in the corresponding line section based on the result of the anomaly confidence accumulation, and output the fault type and the location of the corresponding line section.

[0026] In this invention, the runtime sequence of the adaptive fault detection system for integrated primary and secondary distribution network equipment is as follows: The primary side feature sorting and construction module is executed once during system initialization or periodic updates. After completion, it sends the sorted set of primary side performance features to the feature extraction module and the anomaly analysis module. The secondary side action serialization module monitors the closing action signals in real time. Whenever a new combination of closing action signals is formed and sorted, the sorting result is sent to the feature extraction module. After receiving the sorting result, the feature extraction module collects and quantifies the performance curves in sequence, and sends the peak deviation and steady-state offset to the anomaly analysis module. After completing the anomaly judgment and confidence level determination, the anomaly risk performance features and their confidence levels are sent to the fault detection module. After completing the comprehensive confidence level calculation and fault judgment, the fault detection module outputs the final result.

[0027] This invention does not limit the specific structure of the primary side feature sorting construction module, the secondary side action serialization module, the feature extraction module, the anomaly analysis module, and the fault detection module. Each module can be constructed using logic components, such as field-programmable logic devices, microprocessors, and processors used in computers. Specifically, the module involving primary side electrical data acquisition can be constructed using sensing components such as current transformers and voltage sensors; the module involving secondary side closing action signal acquisition can be constructed using interface components such as optocouplers and communication interface chips; and the module involving detection result output can be constructed using output components such as wireless communication modules and display modules. Further details are omitted here.

[0028] Specifically, the primary-side feature ranking construction module is used to construct a primary-side performance feature ranking set, including: The primary side feature sorting construction module is used to retrieve historical electrical data under normal and fault conditions of the primary side. The electrical data includes the peak value of the inrush current, the inrush current decay time, the voltage drop depth and the current distortion rate. Each electrical data is determined as a primary side performance feature. The primary side feature sorting construction module calculates the feature value difference of each primary side performance feature under historical normal operating conditions and fault operating conditions, and sorts each primary side performance feature in descending order of the relative difference rate of the feature value difference. Specifically, the average characteristic value of each primary side performance feature under normal operating conditions for all historical closing events is calculated as the normal characteristic value; the average characteristic value of the same feature under fault operating conditions for all historical closing events is calculated as the fault characteristic value; the absolute value of the difference between the fault characteristic value and the normal characteristic value is determined as the characteristic value difference of the performance feature; the primary side feature sorting module calculates the characteristic value difference of each primary side performance feature under historical normal operating conditions and fault operating conditions, and divides the characteristic value difference by the characteristic value of the performance feature under normal operating conditions to obtain the relative difference rate of each performance feature; the primary side performance features are sorted in descending order of relative difference rate.

[0029] The primary side feature sorting construction module determines the set of sorted primary side performance features as the primary side performance feature sorting set, and associates each primary side performance feature with a corresponding fault type label.

[0030] Specifically, fault type labels are associated with historical fault records. Specifically, when a certain characteristic in a historical fault record shows an anomaly, and the corresponding fault type has the highest statistical proportion, then the fault type label is associated with that characteristic. For example, if the inrush current peak is abnormal and the short-circuit fault has the highest proportion in the historical record, then the label "short-circuit fault" is associated; if the voltage drop depth is abnormal and the ground fault has the highest proportion, then the label "ground fault" is associated; if the inrush current decay time is abnormal and the insulation aging fault has the highest proportion, then the label "insulation aging fault" is associated; if the current distortion rate is abnormal and the contact failure fault has the highest proportion, then the label "contact failure" is associated.

[0031] Those skilled in the art will understand that in the actual operation of integrated primary and secondary distribution network equipment, the sensitivity of different electrical characteristics to faults varies significantly. For example, the inrush current peak value is extremely sensitive to short-circuit faults; when a short circuit occurs, the inrush current peak value can reach several to tens of times that of normal closing, while the current distortion rate is relatively less sensitive to short-circuit faults. If all characteristics are assigned the same weight, the random fluctuations of characteristics with weak distinguishing ability under normal operating conditions may have the same judgment weight as the fault signals of characteristics with strong distinguishing ability, thereby interfering with fault judgment. This invention quantifies and ranks the fault distinguishing ability of each characteristic by calculating the difference in characteristic values ​​under normal and fault operating conditions, thus avoiding noise interference from weak distinguishing characteristics.

[0032] This invention does not specifically limit the peak inrush current, inrush current decay time, voltage sag depth, and current distortion rate, as these are all fundamental concepts in power systems. The peak inrush current refers to the maximum instantaneous surge current, which is much larger than the normal operating current, generated when the circuit breaker closes due to factors such as line distributed capacitance charging and distribution transformer no-load excitation. The inrush current decay time is the time required for the inrush current to decay exponentially from its peak value to the steady-state operating current, typically measured in milliseconds. The voltage sag depth is the magnitude of the system voltage drop from its rated value when the circuit breaker closes or a fault occurs. For example, if the rated voltage is 10kV and the voltage drops to 8.2kV during a fault, the sag depth is 1.8kV. The current distortion rate is the degree to which the current waveform deviates from an ideal sine wave; in engineering, it usually refers to the total harmonic distortion (THD), which is the ratio of the square root of the sum of the squares of the effective values ​​of each harmonic to the effective value of the fundamental wave, expressed as a percentage. All of the above are well-known in the field and will not be elaborated further here.

[0033] Specifically, the secondary-side action serialization module is used to determine the closing action signal combination, including: The secondary-side action serialization module is used to collect the closing action signals of the secondary-side equipment within a preset monitoring period, record the trigger time of each closing action signal, and arrange the closing action signals in order from earliest to latest according to the trigger time. The secondary side action serialization module calculates the trigger time difference between two adjacent closing action signals, and pairs adjacent closing action signals whose time difference meets the preset time difference condition to form a closing action signal combination.

[0034] In this invention, the preset time difference condition is: the time difference between the triggering times of two adjacent closing action signals is greater than or equal to 0.5 seconds and less than or equal to 30 seconds. Adjacent signals with a time difference of less than 0.5 seconds are considered to be linked operations triggered by the same control command and have no independent analytical significance; adjacent signals with a time difference greater than 30 seconds have already had their transient process of the previous closing completely attenuated, and there is no electrical coupling between the two closing operations, therefore they are not included in the pairing scope.

[0035] Those skilled in the art will understand that in power distribution line fault recovery scenarios, maintenance personnel or automatic devices typically perform segmented testing by sequentially closing multiple circuit breakers from the power source side to the load side. There is a natural electrical coupling between the closing operations of two adjacent circuit breakers: after the previous circuit breaker closes, its downstream line has already established voltage and carries distributed capacitance and load; the inrush current characteristics and voltage drop when the subsequent circuit breaker closes are directly affected by the line state after the previous circuit breaker closes. If the time difference between the two closing operations is too large, the transient process of the previous closing operation has completely decayed, and the electrical coupling between the two closing operations has disappeared. Therefore, a preset time difference condition is set to pair adjacent closing operations with a time difference within a reasonable range, ensuring that there is indeed electrical coupling between the paired operations.

[0036] Specifically, the secondary-side action serialization module is used to sort the signals according to the time difference of each adjacent closing action signal combination, including: The secondary side action serialization module is used to obtain the time difference between the two closing action signals in each closing action signal combination, and sort the closing action signal pairs in order of increasing time difference.

[0037] The purpose of sorting by time difference from smallest to largest in this invention is as follows: the smaller the time difference, the less complete the transient process of the previous closing operation has been, resulting in stronger electrical interference to the subsequent closing operation, more concentrated fault information, and a greater need for priority analysis; the larger the time difference, the weaker the coupling effect, and the lower the analysis priority. This sorting rule allows the system to prioritize the allocation of computing resources to the signal combination with the strongest electrical coupling and the largest amount of fault information.

[0038] Specifically, the feature extraction module is used to acquire performance curves according to the order of the closing action signal combinations, wherein: The feature extraction module processes each closing action signal combination in the order of the closing action signal combination sorting. This includes, for the currently processed closing action signal combination, collecting real-time electrical data corresponding to each performance feature in the primary side performance feature sorting set during the time period from the trigger time of the first closing action signal to the trigger time of the last closing action signal. The feature extraction module plots the performance curves of each performance feature over time based on the real-time electrical data.

[0039] It is understood that the present invention selects the time from the first triggering moment to the last triggering moment as the data acquisition window. This window completely covers the entire process from the voltage establishment when the first circuit breaker closes to the transient impact generated when the second circuit breaker closes. The primary side electrical performance within the window directly reflects the insulation status, load condition and fault characteristics of the line segment between the two circuit breakers.

[0040] Specifically, the feature extraction module is used to perform feature quantization extraction on the performance curve, including: For each performance curve, the feature extraction module is used to identify the peak point in the performance curve, calculate the difference between the feature value of the peak point and the peak value of the corresponding performance feature under normal working conditions, and determine the difference as the peak deviation. In the implementation of this invention, the specific process of identifying the peak point in the performance curve includes: scanning the performance curve with a sliding window at a fixed sampling interval of 1ms, with the window length set to 11 data points; if the value of the center data point in the window is greater than the values ​​of the other 10 data points in the window, and the value of the center data point is greater than 1.2 times the average value of all 11 data points in the window, then the center data point is identified as the peak point.

[0041] The feature extraction module is used to identify the steady-state segment in the curve, calculate the difference between the average feature value of the steady-state segment and the steady-state value of the corresponding performance feature under normal operating conditions, and determine the difference as the steady-state offset.

[0042] In the implementation of this invention, the specific process of identifying the steady-state segment in the curve includes: when there are 10 consecutive data points in the curve whose numerical fluctuation amplitude is less than 5% of the average value of the consecutive data points, the segment where the consecutive data points are located is identified as a steady-state segment; if there are multiple steady-state segments on the curve, the first steady-state segment with a duration of more than 20ms after the end of the closing transient process is selected for calculating the steady-state offset.

[0043] In the implementation of this invention, the peak value and steady-state value under normal operating conditions are taken from the historical statistics of normal operating conditions in the process of constructing the primary side performance feature sorting set. Specifically, the average value of the peak point of the performance feature in all historical closing events under normal operating conditions is taken as the peak value under normal operating conditions, and the average value of the steady-state segment of the performance feature in all historical closing events under normal operating conditions is taken as the steady-state value under normal operating conditions.

[0044] In this invention, based on the physical characteristics of power distribution faults, two indicators—peak deviation and steady-state offset—are selected in the curve quantification process. Peak deviation reflects the transient impact at the moment of closing, corresponding to sudden faults such as short circuits and grounding faults. These faults cause a sharp increase in the peak inrush current and an abnormal drop in the peak voltage. Steady-state offset reflects the continuous operating state after closing, corresponding to progressive faults such as overloads, poor contact, and insulation aging. These faults do not produce severe transient impacts but cause the steady-state current and voltage to continuously deviate from normal values. These two indicators cover both transient and steady-state dimensions, ensuring that both sudden and progressive faults can be quantified and captured.

[0045] Specifically, please refer to Figure 2As shown, this is a flowchart illustrating the logic of determining abnormal risk characteristics and anomaly confidence levels according to an embodiment of the present invention. The anomaly analysis module is further used to determine abnormal risk characteristics and anomaly confidence levels, including: The anomaly analysis module is used to compare the peak deviation of each performance feature with a preset peak deviation threshold, and to compare the steady-state offset of each performance feature with a preset steady-state offset threshold. If the peak deviation exceeds the preset peak deviation threshold, or the steady-state deviation exceeds the preset steady-state deviation threshold, the corresponding performance characteristic will be judged as an abnormal risk performance characteristic. In this invention, the preset peak deviation threshold is set to 20% of the peak value under normal operating conditions. That is, when the peak deviation is greater than 20% of the peak value under normal operating conditions, the corresponding performance characteristic is judged as an abnormal risk performance characteristic. The preset steady-state offset threshold is set to 15% of the steady-state value under normal operating conditions. That is, when the steady-state offset is greater than 15% of the steady-state value under normal operating conditions, the corresponding performance characteristic is judged as an abnormal risk performance characteristic.

[0046] The anomaly analysis module is also used to obtain the ranking order of the abnormal risk performance characteristics in the primary side performance characteristic ranking set, and to determine the corresponding anomaly confidence level based on the ranking order; Among them, the earlier the sorting order, the higher the confidence level of the anomaly.

[0047] This document provides an example of determining the anomaly confidence level based on the ranking order, specifically calculated using the following formula: Anomaly Confidence Level = (Total number of performance features in the primary side performance feature ranking set - Ranking order + 1) / Total number of performance features. For example, if there are 4 performance features in the primary side performance feature ranking set, the anomaly confidence level of the feature ranked 1st is (4-1+1) / 4 = 1, the anomaly confidence level of the feature ranked 2nd is (4-2+1) / 4 = 0.75, the anomaly confidence level of the feature ranked 3rd is (4-3+1) / 4 = 0.5, and the anomaly confidence level of the feature ranked 4th is (4-4+1) / 4 = 0.25.

[0048] It's important to understand that the principle behind this invention's method of determining anomaly confidence based on ranking order is as follows: features ranked higher have larger differences in feature values ​​between normal and fault conditions, meaning they have stronger fault discrimination capabilities. When a feature with strong discrimination capability exhibits an anomaly, the probability that the anomaly is caused by a genuine fault is much higher than the probability of it being caused by random fluctuations; therefore, it should be assigned a higher confidence level. Conversely, when a feature with weak discrimination capability exhibits an anomaly, the probability that it is caused by noise is higher, and therefore, the confidence level should be correspondingly lower. This invention transforms the fault discrimination capability of a feature into the credibility of anomaly determination, avoiding the triggering of fault alarms by a single weakly discriminative feature anomaly.

[0049] Specifically, the fault detection module is used to accumulate the anomaly confidence level, including: The fault detection module identifies all abnormal risk characteristics corresponding to the combination of closing action signals, and sums up the abnormal confidence levels corresponding to all abnormal risk characteristics to obtain the comprehensive confidence level.

[0050] This invention employs anomaly confidence accumulation because different anomaly features have varying degrees of credibility. The accumulation operation essentially summarizes all anomalous evidence within the signal combination; a higher overall confidence level indicates more sufficient anomalous evidence and a greater likelihood of failure.

[0051] Please see Figure 3 As shown, this is a logic flowchart for determining whether a fault exists in a corresponding line section according to an embodiment of the present invention. The fault detection module is used to compare the comprehensive confidence level of the closing action signal combination with a preset confidence level reference value to determine whether a fault exists in the corresponding line section, including: If the overall confidence level is less than the confidence level reference value, the fault detection module determines that there is no fault in the line section corresponding to the closing action signal combination; If the overall confidence level is greater than or equal to the confidence level reference value, the fault detection module determines that there is a fault in the line section corresponding to the closing action signal combination.

[0052] In this invention, the preset confidence reference value is set to 1. The basis for this setting is: the anomaly confidence of the most discriminative feature in the primary side performance feature ranking set is 1. When the overall confidence is greater than or equal to 1, it means that either the ranked first strong discriminative feature is abnormal, or at least two or more weaker discriminative features are abnormal at the same time. Both situations indicate that the evidence of anomaly is sufficient and the possibility of failure is high.

[0053] Specifically, the fault detection module is used to output the fault type and the corresponding line segment location, including: The fault detection module selects the abnormal risk performance feature with the highest abnormal confidence among all abnormal risk performance features corresponding to the closing action signal combination, and uses the fault type label associated with the abnormal risk performance feature with the highest abnormal confidence in the primary side performance feature sorting set as the fault type. The fault detection module determines the start and end positions of the faulty line section as the corresponding line section position based on the circuit breaker positions corresponding to the first and last closing action signals in the closing action signal combination.

[0054] In this invention, determining the fault section based on the positions of the first and last circuit breakers is based on the physical logic of segment-by-segment testing: a closing action signal combination corresponds to a line segment between two adjacent circuit breakers, and the primary side performance curve collected during this period reflects the electrical state of this line segment. Therefore, the positions of the first and last circuit breakers are the start and end boundaries of the fault section. This allows for precise fault location down to a specific line segment.

[0055] Here is an example of the operation of an adaptive fault detection system for primary and secondary integrated equipment in a distribution network. Suppose that there are 3 primary and secondary integrated pole-mounted circuit breakers on a certain distribution line, numbered A, B, and C. After a fault occurs on the line, the automatic device closes the circuit breaker in sequence from the power supply side to the load side to perform segmented test. The primary side feature ranking construction module retrieves 12 months of historical electrical data for the line, calculates the average value of four features under normal operating conditions and the average value under fault operating conditions, and calculates the relative difference rate of each feature: Peak inrush current at closing: 800A under normal conditions, 1650A under fault conditions, difference 850A, relative difference rate = 850 / 800 = 106.25%; Voltage drop depth: Normal 0.2kV, fault 2.0kV, difference 1.8kV, relative difference rate = 1.8 / 0.2 = 900%; Inrush current decay time: 15ms under normal conditions, 27ms under fault conditions, difference 12ms, relative difference rate = 12 / 15 = 80%; Current distortion rate: 2% for normal, 5% for fault, 3% for difference, relative difference rate = 3 / 2 = 150%; The relative difference rates are sorted from largest to smallest as follows: voltage drop depth (associated with the "grounding fault" label), current distortion rate (associated with the "poor contact fault" label), peak inrush current (associated with the "short circuit fault" label), and inrush current decay time (associated with the "insulation aging fault" label). The peak values ​​of each characteristic under normal operating conditions are as follows: peak inrush current of 800A, voltage drop depth of 0.2kV, inrush current decay time of 15ms, and current distortion rate of 2%; the steady-state values ​​of each characteristic under normal operating conditions are as follows: peak inrush current of 350A, voltage drop depth of 0.2kV, inrush current decay time of 15ms, and current distortion rate of 2%.

[0056] The secondary-side action serialization module detected the closing trigger times of circuit breakers A, B, and C as t=0s, t=2s, and t=5s, respectively. The adjacent time differences are 2s and 3s, both satisfying the condition of 0.5s to 30s, forming two signal combinations: (A,B) and (B,C). Sorted from smallest to largest time difference: (A,B) is ranked 1st, and (B,C) is ranked 2nd.

[0057] The feature extraction module processes the (A, B) combination: Data is collected from t=0s to t=2s, and four performance curves are plotted. Peak inrush current curve: The characteristic value of the peak point is 920A, the peak deviation is 920-800=120A, the average characteristic value of the steady-state section is 350A, and the steady-state offset is 0A. Voltage sag depth curve: peak value is 0.22kV, peak deviation is 0.02kV; steady-state average value is 0.2kV, steady-state offset is 0kV.

[0058] Inrush decay time curve: peak value 16ms, peak deviation 1ms; steady-state average value 15ms, steady-state offset 0; Current distortion rate curve: peak point characteristic value 2.2%, peak deviation 0.2%; steady-state average characteristic value 2%, steady-state offset 0.

[0059] Processing combination (B, C): Data was collected from t=2s to t=5s. Peak inrush current curve: characteristic value of the peak point is 1500A, peak deviation = 1500 - 800 = 700A; average characteristic value of the steady-state section is 380A, steady-state offset is 30A; Voltage sag depth curve: peak value 2.5kV, peak deviation 2.3kV, steady-state average characteristic value 2kV, steady-state offset 1.8kV; Inrush decay time curve: peak value 16ms, peak deviation 1ms, steady-state average value 15ms, steady-state offset 0; Current distortion rate curve: peak point characteristic value 2.3%, peak deviation 0.3%, steady-state average characteristic value 2%, steady-state offset 0.

[0060] The anomaly analysis module calculates the peak deviation threshold and steady-state offset threshold for each feature based on 20% of the peak value under normal operating conditions and 15% of the steady-state value under normal operating conditions. Peak inrush current at closing: Peak deviation threshold = 800 × 20% = 160A, steady-state deviation threshold = 350 × 15% = 52.5A; Voltage sag depth: Peak deviation threshold = 0.2 × 20% = 0.04 kV, steady-state deviation threshold = 0.2 × 15% = 0.03 kV; Inrush decay time: Peak deviation threshold = 15 × 20% = 3ms, steady-state deviation threshold = 15 × 15% = 2.25ms; Current distortion rate: peak deviation threshold = 2 × 20% = 0.4%, steady-state deviation threshold = 2 × 15% = 0.3%.

[0061] The anomaly analysis module determines the combination (A, B): The peak deviation of the closing inrush current is 120A < 160A, and the steady-state deviation is 0A < 52.5A; The peak deviation of the voltage drop depth is 0.02kV < 0.04kV, and the steady-state deviation is 0kV < 0.03kV; The peak deviation of the inrush decay time is 1ms < 3ms, and the steady-state deviation is 0ms < 2.25ms; The peak deviation of the current distortion rate is 0.2% < 0.4%, and the steady-state deviation is 0% < 0.3%. If the deviation of each performance characteristic does not exceed the threshold, then there are no abnormal risk performance characteristics.

[0062] Determine the combination (B, C): The peak deviation of the closing inrush current is 700A > 160A, and the steady-state deviation is 30A < 52.5A, ranking 3rd. Confidence level = (4-3+1) / 4 = 0.5.

[0063] The peak deviation of the voltage drop depth is 2.3kV > 0.04kV, and the steady-state deviation is 1.8kV > 0.03kV, ranking first. Confidence level = (4-1+1) / 4 = 1.

[0064] The peak deviation of the inrush decay time is 1ms < 3ms, and the steady-state deviation is 0ms < 2.25ms; The peak deviation of the current distortion rate was 0.3% < 0.4%, and the steady-state deviation was 0% < 0.3%, both of which did not exceed the threshold.

[0065] The fault detection module analyzes the (A, B) combination: no abnormal features are found, indicating that section AB is normal. Analyzing the (B, C) combination: abnormal features are voltage dip depth and inrush current peak value. The overall confidence level is 1 + 0.5 = 1.5, which is greater than or equal to the confidence level reference value of 1, indicating a fault in section BC. The voltage dip depth feature with the highest confidence level is selected, and the associated fault type is "ground fault." The fault section is the line segment between circuit breakers B and C.

[0066] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0067] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive fault detection system for integrated primary and secondary distribution network equipment, characterized in that, include: The primary side feature ranking construction module is used to collect electrical data under historical normal and fault conditions of the primary side, extract primary side performance features, and rank each performance feature according to the degree of difference between normal and fault conditions to construct a primary side performance feature ranking set. The secondary side action serialization module is used to collect the closing action signals of the secondary side equipment, pair the closing action signals that meet the preset time difference conditions according to the triggering order, and sort them according to the time difference of each combination of closing action signals. The feature extraction module is connected to the primary side feature sorting and construction module and the secondary side action serialization module respectively. It is used to collect the performance curves of the primary side performance characteristics within the corresponding time period of each closing action signal according to the order of the closing action signal combination and sorting, and to extract the features of the performance curves by feature quantization. An anomaly analysis module, which is connected to the feature extraction module, is used to determine the characteristics of anomaly risk based on the results of feature quantification extraction and to determine the anomaly confidence level. The fault detection module, which is connected to the anomaly analysis module, is used to accumulate the anomaly confidence of each anomaly risk performance feature in the combination of closing action signals, determine whether there is a fault in the corresponding line section based on the result of the anomaly confidence accumulation, and output the fault type and the location of the corresponding line section.

2. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 1, characterized in that, The primary-side feature ranking construction module is used to construct a primary-side performance feature ranking set, including: The primary side feature sorting construction module is used to retrieve historical electrical data under normal and fault conditions of the primary side. The electrical data includes the peak value of the inrush current, the inrush current decay time, the voltage drop depth and the current distortion rate. Each electrical data is determined as a primary side performance feature. The primary side feature sorting construction module calculates the feature value difference of each primary side performance feature under historical normal operating conditions and fault operating conditions, and sorts each primary side performance feature in descending order of the relative difference rate of the feature value difference. The primary side feature sorting construction module determines the set of sorted primary side performance features as the primary side performance feature sorting set, and associates each primary side performance feature with a corresponding fault type label.

3. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 1, characterized in that, The secondary-side action serialization module is used to determine the closing action signal combination, including: The secondary-side action serialization module is used to collect the closing action signals of the secondary-side equipment within a preset monitoring period, record the trigger time of each closing action signal, and arrange the closing action signals in order from earliest to latest according to the trigger time. The secondary side action serialization module calculates the trigger time difference between two adjacent closing action signals, and pairs adjacent closing action signals whose time difference meets the preset time difference condition to form a closing action signal combination.

4. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 3, characterized in that, The secondary-side action serialization module is used to sort the signals according to the time difference of each adjacent closing action signal combination, including: The secondary side action serialization module is used to obtain the time difference between the two closing action signals in each closing action signal combination, and sort the closing action signal pairs in order of increasing time difference.

5. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 4, characterized in that, The feature extraction module is used to acquire performance curves according to the order of the closing action signal combinations, wherein: The feature extraction module processes each closing action signal combination in the order of the closing action signal combination sorting. This includes, for the currently processed closing action signal combination, collecting real-time electrical data corresponding to each performance feature in the primary side performance feature sorting set during the time period from the trigger time of the first closing action signal to the trigger time of the last closing action signal. The feature extraction module plots the performance curves of each performance feature over time based on the real-time electrical data.

6. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 5, characterized in that, The feature extraction module is used to perform feature quantization extraction on the performance curve, including: For each performance curve, the feature extraction module is used to identify the peak point in the performance curve, calculate the difference between the feature value of the peak point and the peak value of the corresponding performance feature under normal working conditions, and determine the difference as the peak deviation. The feature extraction module is used to identify the steady-state segment in the curve, calculate the difference between the average feature value of the steady-state segment and the steady-state value of the corresponding performance feature under normal operating conditions, and determine the difference as the steady-state offset.

7. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 6, characterized in that, The anomaly analysis module is also used to determine the characteristics of anomaly risk and to determine the confidence level of anomalies, including: The anomaly analysis module is used to compare the peak deviation of each performance feature with a preset peak deviation threshold, and to compare the steady-state offset of each performance feature with a preset steady-state offset threshold. If the peak deviation exceeds the preset peak deviation threshold, or the steady-state deviation exceeds the preset steady-state deviation threshold, the corresponding performance characteristic will be judged as an abnormal risk performance characteristic. The anomaly analysis module is also used to obtain the ranking order of the anomaly risk performance characteristics in the primary side performance characteristic ranking set, and to determine the corresponding anomaly confidence level based on the ranking order; Among them, the earlier the sorting order, the higher the confidence level of the anomaly.

8. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 7, characterized in that, The fault detection module is used to accumulate the anomaly confidence level, including: The fault detection module identifies all abnormal risk characteristics corresponding to the combination of closing action signals, and sums up the abnormal confidence levels corresponding to all abnormal risk characteristics to obtain the comprehensive confidence level.

9. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 8, characterized in that, The fault detection module is used to compare the overall confidence level of the closing action signal combination with a preset confidence level reference value to determine whether there is a fault in the corresponding line section, including: If the overall confidence level is greater than or equal to the confidence level reference value, the fault detection module determines that there is a fault in the line section corresponding to the closing action signal combination.

10. The adaptive fault detection system for integrated primary and secondary distribution network equipment according to claim 9, characterized in that, The fault detection module is used to output the fault type and the corresponding line section location, including: The fault detection module selects the abnormal risk performance feature with the highest abnormal confidence among all abnormal risk performance features corresponding to the closing action signal combination, and uses the fault type label associated with the abnormal risk performance feature with the highest abnormal confidence in the primary side performance feature sorting set as the fault type. The fault detection module determines the start and end positions of the faulty line section as the corresponding line section position based on the circuit breaker positions corresponding to the first and last closing action signals in the closing action signal combination.

Images