A fault analysis method and device based on line loss anomaly and a medium

Abstract

This invention discloses a fault analysis method, device, and medium based on line loss anomalies, belonging to the field of big data analytics. The method includes: assessing the causal explanatory power of behavioral explanation paths and identifying fault causes; generating a set of fault events; calculating the responsibility matching degree and responsibility contribution degree of the fault event set to form a fault responsibility structure set; mapping the fault responsibility structure set to the governance operation space for performance evolution evaluation and strategy response modeling, outputting a set of strategy response relationships; extracting strategy items from the strategy response relationship set and performing conditional trigger rule solidification processing to generate a set of strategy rule items. This invention, through in-depth big data analysis of management behavior records and simultaneous predictive residual analysis of line loss anomaly events, effectively matches management behavior with line loss anomaly events, achieving real-time and efficient fault early warning and analysis.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, and in particular to a fault analysis method, device, and medium based on line loss anomalies. Background Technology

[0002] In modern power networks, line loss analysis, as a key operational monitoring tool, has been widely applied. With the increasing complexity and scale of power equipment, effectively monitoring, analyzing, and reducing line losses has become crucial for improving power system efficiency and reducing operating costs. Traditional line loss monitoring methods mainly rely on periodic manual inspections, manual calculations, or simple equipment monitoring. While these methods can monitor line losses to some extent, they often suffer from problems such as data processing lag and inaccurate fault identification. Furthermore, with the growth of electricity demand and the expansion of network scale, traditional methods are unable to meet the challenges of the big data era, particularly in terms of real-time and intelligent analysis. In recent years, with the development of big data technology, combining big data analytics methods for line loss anomaly analysis has become an effective solution.

[0003] However, existing methods still have some shortcomings. While some methods can predict line loss using historical data, they lack effective integration and in-depth analysis of real-time dynamic data. Traditional methods' early warning mechanisms often rely on a single data source, making it difficult to detect potential faults in real time, leading to delays in fault response. Most existing line loss analysis methods only focus on a single indicator of line loss, failing to delve into the complex underlying causes that may lead to abnormal line loss, especially lacking sufficient analysis of fault causes under the combined influence of multiple factors. Therefore, existing technologies struggle to achieve comprehensive, accurate, and real-time detection and analysis of line loss anomalies. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a fault analysis method based on line loss anomalies to solve the problems of ineffective integration of real-time data, lack of multi-dimensional fault cause analysis, and delayed response.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: Firstly, this invention provides a fault analysis method based on line loss anomalies, comprising: collecting line loss measurement data and management behavior records; extracting the subject identifiers from the management behavior records and classifying the behaviors to generate management behavior fingerprints; statistically analyzing line loss anomalies in the line loss measurement data and performing predictive residual analysis to generate a set of line loss anomaly events; outputting behavior explanation paths by performing semantic event anchoring and causal semantic alignment on the management behavior fingerprints and the set of line loss anomaly events; evaluating the causal explanation degree of the behavior explanation paths and identifying fault causes to generate a set of fault events; calculating the responsibility matching degree and responsibility contribution degree of the fault event set to form a set of fault responsibility structures; mapping the set of fault responsibility structures to the governance operation space for effectiveness evolution evaluation and strategy response modeling to output a set of strategy response relationships; extracting strategy items from the set of strategy response relationships and performing condition trigger rule solidification processing to generate a set of strategy rule items.

[0007] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for generating management behavior fingerprints are as follows: By parsing the data fields of management behavior records, each management behavior record is broken down into the executing entity identifier field, behavior type field, behavior time field, and behavior target field to obtain the entity identity set; Based on the subject identifier field, the execution subjects in the subject identity set are mapped to business role entities and their behavior semantics are categorized to generate management behavior fingerprints.

[0008] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for generating the set of line loss anomaly events are as follows: The fluctuation range of the line loss measurement data is extracted and reconstructed at the time granularity to form a stable base value sequence; By extracting the baseline trend component and periodic component of the stationary base value sequence, the abnormal event interval of the stationary base value sequence is identified, and the set of abnormal line loss events is output.

[0009] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps of the output behavior interpretation path are as follows: Match the set of abnormal line loss events with the behavior types of management behavior fingerprints and perform semantic grouping to generate behavior semantic groups; By analyzing the temporal causal relationship between behaviors and events and the chronological order of events in behavioral semantic groups, a causal inference chain is constructed. Assess the internal consistency, temporal rationality, and logical order of the causal reasoning chain, and prioritize and output the behavioral explanation path.

[0010] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for generating the fault event set are as follows: Assess the causal credibility, temporal plausibility, and depth of impact of behavioral explanation paths, and generate causal strength indicators; Calculate the fault cause score of the causal intensity index and classify the cause patterns to generate a set of fault events.

[0011] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for forming the fault responsibility structure set are as follows: By analyzing the dynamic responsibility patterns of a set of fault events, potential sources of responsibility for the set of fault events are extracted, and a candidate set of potential sources of responsibility is generated. Based on the potential source of responsibility candidate set, calculate the responsibility matching degree of each failure event and classify it to generate a responsibility matching set; Cluster analysis is used to identify causal patterns, device correlations, and user impacts in the responsibility matching set, generating a responsibility cause classification set. By statistically analyzing the equipment damage degree, repair time, frequency of historical events, and user responsibility coefficient of each fault event in the fault cause classification set, the responsibility contribution of each fault event is calculated and decision reasoning is performed to form a fault responsibility structure set.

[0012] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for the output strategy response relationship set are as follows: The system analyzes the responsibility types, event levels, impact scope, historical data, and resource availability in the fault responsibility structure set, performs hierarchical analysis, and generates a set of governance operation strategies. Quantitatively assess the execution effectiveness, risks, resource consumption, and implementation difficulty of a set of governance operation strategies, and generate a set of effectiveness evolution predictions; The response process of the performance evolution prediction set is optimized by the response optimization algorithm to generate a governance strategy response set; Verify the repair effect, resource utilization efficiency, and time response speed of the governance strategy response set, and output the strategy response relationship set.

[0013] As a preferred embodiment of the fault analysis method based on line loss anomalies described in this invention, the specific steps for generating the strategy rule item set are as follows: Extract the responsibility type, fault level, and resource requirements of each fault event from the set of policy response relationships, and generate a governance operation entry for each fault event; Analyze the response conditions, execution order, and scope of impact for each governance operation item, and perform rule mapping to generate a set of triggering rules for each item; By solidifying the rules, the set of triggering rules for each entry is transformed into executable decision instructions, generating a set of policy rule entries.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the fault analysis method based on line loss anomalies as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the fault analysis method based on line loss anomalies as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by conducting in-depth big data analysis on management behavior records, management behavior fingerprints are generated, different management behaviors are accurately identified and classified, and management behaviors are effectively matched with abnormal line loss events, thereby improving the accuracy of fault cause analysis; at the same time, predictive residual analysis of abnormal line loss events is performed, which can accurately identify potential factors that may lead to abnormal line loss in the context of big data, and realize real-time and efficient fault early warning and analysis. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a fault analysis method based on line loss anomalies.

[0019] Figure 2 A flowchart for generating management behavior fingerprints.

[0020] Figure 3 A flowchart for generating a set of abnormal line loss events.

[0021] Figure 4 A flowchart for generating a set of policy rule entries. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a fault analysis method based on line loss anomalies, including the following steps: S1. Collect line loss measurement data and management behavior records, extract the subject identifier of the management behavior records and classify the behavior to generate management behavior fingerprints.

[0026] S1.1 By parsing the data fields of the management behavior records, each management behavior record is broken down into the execution subject identifier field, behavior type field, behavior time field, and behavior target field to obtain the subject identity set.

[0027] It should be noted that the line loss metering data includes a metering object identification field, a metering value field, and a metering time field. The metering object identification field is derived from the equipment code of the electricity meter or distribution transformer metering equipment and is generated by uploading it to the equipment sending end through the automatic meter reading acquisition channel. The metering value field represents the cumulative and instantaneous values ​​of the corresponding metering object, which are collected by the metering components inside the electricity meter and generated by uploading it through the automatic meter reading acquisition channel. The metering time field represents the acquisition time corresponding to the metering value field, which is automatically written by the automatic meter reading acquisition channel when the meter reading is successful and uploaded together with the metering value field.

[0028] The management behavior log includes a login account field, an operation category code field, an operation timestamp field, and a business object identifier field. The login account field is written by the business application when a user logs in successfully and is used in every behavior log. The operation category code field is written by the business application when generating the operation instruction. The operation timestamp field is written by the business application when recording the behavior event and contains the current time. The business object identifier field is written by the business application when recording the behavior event and contains the line identifier, transformer area identifier, or metering device identifier selected in the interface.

[0029] Read the login account field, operation category code field, operation timestamp field, and business object identifier field from the management behavior record. Use the login account field as the execution subject identifier field, the operation category code field as the behavior type field, the operation timestamp field as the behavior time field, and the business object identifier field as the behavior target field. Integrate all fields to form a subject identity set.

[0030] S1.2 Based on the subject identifier field, map the execution subject in the subject identity set to the business role entity and perform behavioral semantic classification processing to generate management behavior fingerprints.

[0031] It should be noted that the login account source corresponding to the execution entity identifier field of the subject identity set during the management behavior record collection stage is mapped to the business role entity, which is used to identify the responsibilities of the execution entity in the business operation process. The behavior type field in the management behavior record is read synchronously, and the business role entity and behavior type field are combined in a parallel aggregation manner to form a combined data structure of business role entity and behavior type field. The original values ​​of business role entity and behavior type field are retained in the combined data structure to form the management behavior fingerprint.

[0032] S2. Analyze the anomalies in the statistical line loss measurement data and perform predictive residual analysis to generate a set of line loss anomaly events. By performing semantic event anchoring and causal semantic alignment on the management behavior fingerprint and the set of line loss anomaly events, output the behavior explanation path.

[0033] S2.1 Extract the fluctuation range of the line loss measurement data and perform time-granular reconstruction processing to form a stable base value sequence.

[0034] It should be noted that the measurement value field and measurement time field of the line loss measurement data are read. The difference between two adjacent measurement value fields is calculated according to the time order of the measurement time field to form a difference sequence. The consecutive positions in the difference sequence where the difference is not zero are regarded as continuous fluctuation segments. The measurement time field within the continuous fluctuation segment is read and the time difference between adjacent measurement time fields is calculated. The minimum time difference is selected as the target time interval. The difference sequence within the continuous fluctuation segment is divided according to the target time interval. The difference of the difference sequence within each target time interval is accumulated to form the prediction base value of the corresponding target time interval. The prediction base values ​​corresponding to all target time intervals are recorded sequentially according to the time order of the measurement time field to form a stable base value sequence.

[0035] S2.2 By extracting the baseline trend component and periodic component of the stationary base value sequence, the abnormal event interval of the stationary base value sequence is identified, and the set of abnormal line loss events is output.

[0036] It should be noted that a fixed-length (e.g., 24-hour) sliding base value window is used to smooth the stationary base value sequence. Specifically, the average value within the sliding base value window before and after each time point is obtained to obtain the predicted baseline sequence. The difference between the stationary base value sequence and the predicted baseline sequence at each time point is used as the predicted residual value to form the predicted residual sequence. The difference between adjacent values ​​of the predicted residual in the predicted residual sequence is calculated to generate the predicted residual change sequence. Based on the absolute mean of the predicted residual change sequence, continuous intervals in the predicted residual change sequence with absolute values ​​greater than the absolute mean are identified as abnormal event intervals. All abnormal event intervals are recorded in chronological order as a set of line loss abnormal events.

[0037] S2.3 Match the set of abnormal line loss events with the behavior types of the management behavior fingerprint and perform semantic grouping to generate behavior semantic groups.

[0038] It should be noted that the start and end times of each abnormal event interval in the line loss abnormal event set are extracted, and the behavior time field and behavior type field in the management behavior fingerprint are read. By comparing the behavior time field with the time range of the abnormal event interval, if the behavior time field of the management behavior is within the abnormal event interval, the current management behavior is associated with the corresponding abnormal event. All matching management behaviors and abnormal events are grouped according to the behavior type field. Management behaviors with the same behavior type and related abnormal events are grouped together to form a behavior semantic group. Each behavior semantic group contains a behavior type field and multiple associated abnormal events.

[0039] S2.4. By analyzing the temporal causal relationship between behaviors and events and the chronological order of events in the behavioral semantic group, a causal reasoning chain is constructed.

[0040] It should be noted that the behavior time field in the behavior semantic group is read from the timestamp of the abnormal event. The behavior time field and the timestamp of the abnormal event are compared according to the time order to determine the causal relationship between the management behavior and the abnormal event. Specifically, if the behavior time field occurs within the time interval before and after the abnormal event, the current management behavior is considered to have a causal impact on the abnormal event. The management behaviors with causal relationship are associated with the abnormal events and arranged in chronological order to construct a causal inference chain. Each node in the causal inference chain represents a management behavior and a corresponding abnormal event, and the nodes are connected by time order and causal relationship.

[0041] S2.5. Evaluate the internal consistency, temporal rationality, and logical order of the causal reasoning chain, and prioritize and output the behavioral explanation path.

[0042] It should be noted that each management action and abnormal event in the causal inference chain is evaluated for consistency, temporal reasonableness, and logical order. In the consistency evaluation, the time field of each action and the corresponding abnormal event timestamp are checked to ensure they are arranged chronologically, ensuring the timestamp of the management action is earlier than the time of the abnormal event. If the timestamp of the management action is later than the time of the abnormal event, it is considered a contradiction and marked as inconsistent. In the temporal reasonableness evaluation, the difference between the time of the management action and the time of the abnormal event is obtained, and it is determined whether the difference is within a predetermined business time threshold. If the difference exceeds the business time threshold, it is marked as not meeting the reasonableness requirements. In the logical order evaluation, the order of management actions and abnormal events in the causal inference chain is checked to ensure that each management action occurs before the abnormal event. If the order of management actions and abnormal events does not conform to causal logic, it is considered not meeting the order requirements. Based on the results of the consistency, temporal reasonableness, and logical order evaluations, priorities are ranked, and the behavior explanation path is output.

[0043] It should also be noted that priority ranking includes causal relationship strength ranking and time rationality ranking. Causal relationship strength ranking refers to prioritizing management actions and abnormal events with the shortest time intervals. That is, the smaller the time difference between the management action and the abnormal event, the stronger the causal relationship and the higher the priority. Time rationality ranking refers to ranking based on the time difference between the management action and the abnormal event. The smaller the time difference, the stronger the rationality in line with the business time logic and the higher the priority.

[0044] The business time threshold is defined based on the historical time difference distribution between management actions and abnormal events; the value range is set to [0 minutes, 120 minutes], where the lower limit corresponds to the situation where management actions and abnormal events occur at almost the same time. When the device records an abnormal event, the management action will usually respond or execute immediately; the upper limit refers to the maximum allowable time interval between management actions and abnormal events.

[0045] S3. Evaluate the causal explanation level of the behavioral explanation path and identify the causes of failure, generate a set of failure events, calculate the responsibility matching degree and responsibility contribution degree of the set of failure events, and form a set of failure responsibility structures.

[0046] S3.1 Evaluate the causal credibility, temporal rationality, and impact depth of the behavioral explanation path, and generate a causal strength index; It should be noted that for each management action and abnormal event in the causal inference chain, the causal credibility, temporal rationality, and impact depth are assessed. In the causal credibility assessment, the time difference between the time field of the management action and the timestamp of the abnormal event is statistically analyzed; the smaller the time difference, the stronger the causal relationship and the higher the credibility. In the temporal rationality assessment, the time difference between the management action and the abnormal event is statistically analyzed and its changes are recorded. If the time difference shows a continuous increasing or decreasing trend as the causal chain progresses, the temporal relationship is reasonable; if the trend of the time difference shows a jump or discontinuity, the temporal relationship is not reasonable. In the impact depth assessment, the number or scope of business objects involved in the management action and the number or scope of business objects involved in the abnormal event are statistically analyzed and recorded. If the degree of overlap between the number or scope of business objects involved in the management action and the abnormal event is higher, the impact depth is higher. The assessment results of causal credibility, temporal rationality, and impact depth are integrated to form a causal strength index.

[0047] S3.2 Calculate the fault cause score of the causal intensity index and perform cause pattern classification to generate a set of fault events.

[0048] It should be noted that all fault cause scores are sorted in ascending order. A fault difference sequence is formed by statistically analyzing the differences between adjacent fault cause scores. The locations where the magnitude of the difference changes abruptly in the fault difference sequence are identified as pattern segmentation boundaries. The sorted fault cause scores are divided into several cause score segments according to the pattern segmentation boundaries. The causal reasoning chain number, fault cause score interval, and associated management behavior and abnormal event information corresponding to each cause score segment are summarized to form a cause pattern. Each cause pattern is treated as an independent fault event, and these are summarized to form a fault event set.

[0049] The expression for calculating the fault cause score is: ; in, Represents a causal reasoning chain Fault cause score, Represents a causal reasoning chain The average value of the causal strength index Represents a causal reasoning chain The number of causal pairs of "management behavior - anomalous events" included. Represents a causal reasoning chain The variance of the causal strength index.

[0050] S3.3 By analyzing the dynamic responsibility pattern of the fault event set, the potential responsibility sources of the fault event set are extracted, and a candidate set of potential responsibility sources is generated.

[0051] It should be noted that the following steps are taken: First, the management behavior information, abnormal event information, and cause pattern identifier of each fault event in the fault event set are read. The fault event set is then grouped according to the cause pattern identifier. Within each cause pattern grouping result, the frequency of occurrence of the management behavior type field and the business object identifier field involved in the abnormal event is statistically analyzed to form a frequency distribution sequence. This sequence is then sorted from high to low frequency to form an ordered sequence. Within this ordered sequence, all elements from the beginning of the sequence up to the position before the first change in frequency are selected as potential liability feature fields and written into a potential liability feature set. From this potential liability feature set, the execution entity identifier field and business object identifier field corresponding to the sorted potential liability feature fields are extracted and combined to form a potential liability source candidate. Finally, all potential liability source candidates are summarized to generate a potential liability source candidate set.

[0052] S3.4. Based on the potential source of responsibility candidate set, calculate the responsibility matching degree of each fault event and classify it to generate a responsibility matching set.

[0053] It should be noted that the responsibility matching degree of each fault event and the corresponding potential source of responsibility candidate number are read. All fault events are sorted in descending order of responsibility matching degree, and the difference between adjacent responsibility matching degrees in the sorting results is counted to form a responsibility difference sequence. The position where the difference changes in the responsibility difference sequence is identified as the responsibility pattern segmentation boundary. The sorted responsibility matching degree sequence is divided according to the responsibility pattern segmentation boundary to form several responsibility pattern segments. Each responsibility pattern segment contains the responsibility matching degree, the fault event number, and the potential source of responsibility candidate number. Each responsibility pattern segment is recorded as a responsibility matching entry, and all responsibility matching entries are summarized to form a responsibility matching set.

[0054] The expression for calculating the responsibility matching degree is: ; in, Indicates a fault event With potential sources of liability The degree of matching of responsibilities between them Indicates a fault event With responsibility source The set of potential responsibility features that can be matched between them This represents the attribute matching judgment function, and its implementation logic is based on potential liability features. In the fault event It appears in related management behavior information or abnormal event information, and is related to potential sources of responsibility. The value is 1 if the execution entity identifier field or the business object identifier field is the same; otherwise, it is 0. Indicates fault events The associated set of fields, Indicates the source of potential liability The associated set of fields, Indicates potential liability characteristics The responsibility feature weights are calculated by counting the frequency of occurrence of the management behavior type field and the business object identifier field within each cause pattern group. The percentage of the historical occurrence frequency within the corresponding cause pattern group is used as the responsibility feature weight, with a value range of (0.2, 0.8). The lower limit of 0.2 is derived from the minimum effective statistical instance within the cause pattern group, and the upper limit of 0.8 is derived from the situation where potential responsibility features appear repeatedly within the current cause pattern group but do not appear in complete agreement. This is to prevent the weight value from degenerating to 1 and losing its distinguishability, thereby ensuring that the responsibility feature weights can stably reflect the relative representativeness and distinguishability of different responsibility features in the same type of fault causes in actual calculations.

[0055] S3.5 Cluster analysis is performed on the causal patterns, device correlation, and user impact of the responsibility matching set to generate a responsibility cause classification set.

[0056] It should be noted that the fault event number, potential responsibility source candidate number, and responsibility matching degree of each responsibility matching entry in the responsibility matching set are read. Simultaneously, the cause pattern identifier, behavior target field, and execution subject identifier field corresponding to the fault event number in the fault event set are read. The cause pattern identifier field is mapped to causal pattern features, the behavior target field is mapped to equipment association features, and the execution subject identifier field is mapped to user impact features. The causal pattern features, equipment association features, and user impact features are combined into a multi-dimensional feature vector set in a fixed order. The differences between feature dimensions are obtained within the multi-dimensional feature vector set, and the responsibility pattern grouping results are formed based on the grouping structure of the differences. Each responsibility pattern grouping result is recorded as a responsibility cause classification entry, which includes the fault event number, potential responsibility source candidate number, and responsibility matching degree information. All responsibility cause classification entries are summarized to form a responsibility cause classification set.

[0057] It should also be noted that the difference between feature dimensions refers to the difference in whether the feature values ​​at the same position in the multidimensional feature vector set are consistent. If the feature values ​​at the same position are consistent, the difference is zero; if the feature values ​​at the same position are inconsistent, the difference is one.

[0058] S3.6. By statistically analyzing the equipment damage degree, repair time, frequency of historical events, and user responsibility coefficient of each fault event in the fault cause classification set, calculate the responsibility contribution of each fault event and perform decision reasoning to form a fault responsibility structure set.

[0059] It should be noted that by reading the relevant information of each fault event in the responsibility cause classification set, the equipment damage degree, repair time, historical event frequency, and user responsibility coefficient of each fault event are obtained. Among them, the ratio of repair cost to the maximum repair cost of the equipment is used as the equipment damage degree; the time from the occurrence of the fault to the completion of the repair is calculated to obtain the repair time; the historical event frequency is obtained by statistically analyzing the number of occurrences of the same type of fault event in history; and the ratio of the number of user operation errors to the total number of equipment operations is used as the user responsibility coefficient.

[0060] The repair time and frequency of historical events are normalized. Combined with the degree of equipment damage and the user responsibility coefficient, the responsibility contribution of each fault event is calculated. Fault events are sorted according to their responsibility contribution and divided into different responsibility levels. The higher responsibility contribution (e.g., the top 30%) corresponds to primary responsibility, the middle range (e.g., 30%-70%) corresponds to secondary responsibility, and the lower range (e.g., 70%-100%) is classified as minor responsibility. Each fault event is summarized with its corresponding responsibility level, responsibility contribution, candidate responsibility source number, and responsibility allocation ratio to form a fault responsibility structure set.

[0061] The expression for calculating the contribution of responsibility is: ; ; ; in, Indicates a fault event The degree of equipment damage, Indicates a fault event Repair time after normalization Indicates a fault event Normalized frequency of historical events Indicates a fault event User responsibility coefficient, Indicates a fault event The intensity index of the responsibility effect, Indicates a fault event Responsibility dispersion index Indicates a fault event The degree of responsibility and contribution.

[0062] S4. Map the set of fault responsibility structures to the governance operation space for performance evolution assessment and strategy response modeling, and output the set of strategy response relationships.

[0063] S4.1 Analyze the responsibility types, event levels, scope of impact, historical data, and resource availability in the statistical fault responsibility structure set, and perform hierarchical analysis to generate a set of governance operation strategies.

[0064] It should be noted that the following steps are taken: First, the responsibility contribution, responsibility level, and fault event number are read from the fault responsibility structure set. The responsibility level corresponding to the fault event is obtained from the fault event set using the fault event number, and this responsibility level is mapped to a responsibility type field. Second, the historical event frequency corresponding to the fault event is obtained from the fault event set using the fault event number, and this historical event frequency is mapped to an event level field. Third, the behavioral target field of the fault event is obtained from the fault event set using the fault event number, and the scope of impact field is determined based on the equipment identifier type pointed to by the behavioral target field. Fourth, the historical event frequency is obtained from the fault event set using the fault event number. Fifth, equipment maintenance resources, personnel dispatch resources, and material inventory resources are obtained from resource management data using the fault event number. The quantity and availability of resources available for current fault recovery are recorded as resource availability fields. Historical statistical information such as the historical event frequency, corresponding repair time, and equipment damage degree during the fault event's historical operation are uniformly used as historical data fields. Finally, the responsibility type field, event level field, scope of impact field, historical data field, and resource availability field are combined to form a strategy analysis record. Hierarchical analysis is then performed on all strategy analysis records to form a governance operation strategy set.

[0065] It should also be noted that the specific steps of the hierarchical analysis process are as follows: First layer, the responsibility type layer: The strategy analysis record set is grouped according to the responsibility type field, dividing the records into multiple responsibility type groups corresponding to primary responsibility, secondary responsibility, and minor responsibility, and arranging each group according to the priority of the responsibility type. Second layer, the event level layer: Within each responsibility type group obtained in the first layer, the strategy analysis records are sorted within the group according to the event level field. Within the same responsibility type group, the strategy analysis records with higher event level fields are ranked first, forming the result of the second layer. Third layer, the impact scope layer: In the result of the second layer, for strategy analysis records with the same event level field, priority adjustment is performed according to the impact scope field. When the event levels are the same, the strategy analysis records with larger impact scopes are ranked first. Fourth layer, the resource availability layer: Based on the result of the third layer, the strategy analysis records are differentiated according to the resource availability field. When the resource availability field... When the required resources are available, the corresponding strategy analysis record remains in its position within the existing sorting results. When the resource availability field indicates that the required resources are unavailable, the corresponding strategy analysis record is moved to the end of the current sorting results. In the fifth layer of processing, the historical data layer, after completing each layer of processing, if strategy analysis records still exist with the same sorting position in the responsibility type, event level, impact scope, and resource availability processing results, disambiguation processing is performed on the strategy analysis records based on historical data fields. Specifically, for strategy analysis records with the same sorting position, the historical occurrence frequency field, historical repair duration field, and equipment damage degree field are compared sequentially. When the historical occurrence frequency field is different and distinguishable, the strategy analysis record with the higher historical occurrence frequency is retained first. When the historical occurrence frequency field is the same and indistinguishable, the strategy analysis record with the longer historical repair duration is retained first. When the historical repair duration field is still the same and indistinguishable, the strategy analysis record with the higher equipment damage degree is retained first. This completes the disambiguation processing of the sorting position and generates a set of governance operation strategies.

[0066] S4.2 Quantitatively assess the execution effectiveness, risks, resource consumption, and implementation difficulty of the set of governance operation strategies, and generate a set of effectiveness evolution predictions.

[0067] It should be noted that the responsibility type field, event level field, scope of impact field, historical data field, and resource availability field of the governance operation strategy set are read. Based on the correspondence between the strategy items recorded in the governance operation strategy set and the equipment maintenance resources, personnel dispatch resources, and material inventory resources, the resource quantity required to execute the strategy is extracted from the resource management data and recorded as the resource quantity required by the strategy. The absolute value of the level difference between the responsibility type field and the event level field is used as the execution effectiveness value. The sum of the levels between the event level field and the scope of impact field is used as the risk value. The difference between the resource quantity required by the strategy and the available resource quantity represented by the resource availability field is used as the resource consumption value. The historical frequency represented by the historical data field is used as the implementation difficulty value. The execution effectiveness value, risk value, resource consumption value, and implementation difficulty value are recorded as effectiveness analysis records. Time series analysis is performed on the historical data fields to form a trend series. The trend series and effectiveness analysis records are combined to form an effectiveness evolution prediction set.

[0068] S4.3 Optimize the response process of the performance evolution prediction set through the response optimization algorithm to generate the governance strategy response set.

[0069] It should be noted that the execution performance value, risk value, resource consumption value, implementation difficulty value, and resource availability fields in the performance evolution prediction set are read. The execution performance value and risk value constitute strategy screening factors. The execution performance value is judged to be at a high level (e.g., higher than the median of the entire execution performance value set) by comparing its position with the statistical distribution of the entire risk value set. Similarly, the risk value is judged to be at a low level (e.g., lower than the median of the entire risk value set) by comparing its position with the statistical distribution of the entire risk value set. Strategy entries that simultaneously satisfy both a high execution performance value and a low risk value are retained to form the initial screening set. The resource consumption value and resource availability fields in the initial screening set are then compared. The resource consumption value is used for matching and validation. When the resource consumption value is less than or equal to the resource availability value, it is considered that the resource can meet the policy execution requirements. When the resource consumption value is greater than the resource availability value, it is considered that the resource cannot meet the policy execution requirements and the policy entry is removed from the initial filtering set, forming a candidate set of executable policies. In the candidate set of executable policies, the implementation difficulty value is prioritized. When the implementation difficulty value is in a lower position in the set of all implementation difficulty values ​​(for example, it is considered lower if it is lower than the median of the set of all implementation difficulty values), it is considered simple to implement and has the priority execution attribute. Policy entries with the priority execution attribute are recorded as governance policy response records in the order of execution. All governance policy response records are summarized to form a governance policy response set.

[0070] S4.4 Verify the repair effect, resource utilization efficiency and time response speed of the governance strategy response set, and output the strategy response relationship set.

[0071] It should be noted that, through big data analysis, the execution efficiency, resource consumption, implementation difficulty, and repair time of each policy response record in the governance policy response set are read, and the success rate and quality of equipment repair after the corresponding policy execution are statistically analyzed. Policy response records with successful and high-quality repairs are assigned higher repair effectiveness scores, while policy responses with short repair times and success are assigned higher time response speed scores. By matching the resource consumption and resource availability fields of each policy response record, resource utilization efficiency is evaluated, and policy response records with higher resource utilization efficiency receive higher scores. The repair effectiveness score, resource utilization efficiency score, and time response speed score of each policy response record are weighted and summed to obtain the comprehensive score of each policy response record. By sorting the comprehensive scores of all policy response records, policy response records with higher comprehensive scores are executed first, and all policy response records are aggregated into a policy response relationship set.

[0072] It should also be noted that the weighting coefficient corresponding to the repair effectiveness score is obtained by statistically analyzing the number of times the same policy response record was successfully repaired in the historical execution records, and comparing the number of successful repairs with the total number of executions of the current policy response record; the weighting coefficient corresponding to the time response speed score is determined by sorting the repair duration of each policy response record in the governance policy response set from smallest to largest. Policy response records with the same repair duration are considered parallel and assigned the same sorting position, and the basic weighting value of the time response speed score is generated based on the sorting position, so that the earlier the sorting position, the larger the corresponding basic weighting value; the weighting coefficient corresponding to the resource utilization efficiency score is determined by comparing the resource consumption value of the policy response record with the number of available resources recorded in the resource availability field, using the proportion of resource consumption to available resources as the comparison result, the lower the proportion, the larger the corresponding basic weighting value of the resource utilization efficiency score; the three basic weighting values ​​are normalized so that the sum of the weighting coefficients of the repair effectiveness score, time response speed score, and resource utilization efficiency score is 1.

[0073] S5. Extract strategy entries from the strategy response relationship set and perform condition trigger rule solidification processing to generate a strategy rule entry set.

[0074] S5.1 Extract the responsibility type, fault level, and resource requirements of each fault event from the policy response relationship set, and generate the corresponding governance operation entry for each fault event.

[0075] It should be noted that the fault event number, responsibility type field, impact scope field, event level field, and resource consumption field of each policy response record are read from the policy response relationship set. Based on the responsibility type field, the responsibility type to which the fault event belongs (including primary responsibility, secondary responsibility, and minor responsibility) is extracted and marked as the responsibility type field in the governance operation entry. The basic severity level of the fault event is determined based on the event level field, and this basic severity level is used as the initial reference level for fault level determination, limiting the minimum level of urgency for fault event handling. The extent of the impact of the fault event is assessed based on the impact scope field. When the impact scope field indicates that the affected object has expanded from a single device to multiple devices, or from a local business scope to a cross-business scope, ... When the impact of a fault event expands from an individual user to a group of users, it is determined that the current fault event has an expanded scope. If there is no expansion of the scope, the base level corresponding to the event level field is maintained as the fault level. If there is an expansion of the scope, the base level corresponding to the event level field is raised by one level to represent the urgency of handling the fault event in the current operating state. The resource requirement field is extracted based on the difference between the resource consumption field and the resource availability field, and the types and quantities of resources required by the fault event are counted and marked as the resource requirement field in the governance operation item. The responsibility type, fault level, and resource requirement fields in each governance operation item are associated with the corresponding fault event number to generate a set of governance operation items corresponding to each fault event.

[0076] S5.2 Analyze the response conditions, execution order and scope of impact of each governance operation item and perform rule mapping to generate a set of triggering rules for the item.

[0077] It should be noted that the responsibility type field, fault level field, resource requirement field, and event impact scope field of each governance operation item are read from the governance operation item set. Based on the responsibility type, fault level, resource requirement, and impact scope fields, the response conditions, execution order, and impact scope of each governance operation item are analyzed.

[0078] The response condition analysis specifically determines the response conditions corresponding to the governance operation entries based on the responsibility type field. Among them, the governance operation entries for primary responsibilities are executed first when all conditions are met, while the governance operation entries for secondary and minor responsibilities can be executed even when the conditions are not fully met.

[0079] The execution order analysis specifically assesses the execution order of each governance operation based on the matching of the resource requirement field and the resource availability field. Operations with higher resource availability can be executed earlier, while operations with higher resource consumption need to be executed later.

[0080] Impact scope analysis specifically assesses the impact of governance operation items on equipment, users, regions, etc., based on the impact scope field. Items with a wider impact scope are given priority, while the time span and urgency of the impact are also considered. Items with a smaller impact scope are arranged in a later stage of the strategy implementation process.

[0081] The response conditions, execution order, and scope of impact of each governance operation item are mapped to corresponding rule items, generating an item trigger rule set. Each item trigger rule contains the corresponding trigger conditions, execution order, and scope of impact. All item trigger rule sets are summarized to form a trigger rule set.

[0082] S5.3. Through rule solidification processing, the set of triggering rules for entries is transformed into executable decision instructions, generating a set of policy rule entries.

[0083] It should be noted that the response conditions, execution order, and scope of impact of each triggering rule are read from the set of triggering rules. Based on the response conditions in each triggering rule, specific triggering conditions are set to determine when to trigger governance operations. This includes defining fixed logical rules for data such as resource requirements, responsibility levels, and fault levels to ensure that corresponding operations are automatically triggered when specific conditions are met. The execution priority is determined based on the execution order in each triggering rule, prioritizing strategies with a larger scope of impact and higher resource availability. The scope of impact field is converted into executable scope control rules to ensure that strategies with a larger scope of impact are executed first when resources allow, while ensuring that covered areas, devices, or users are processed promptly. The response conditions, execution order, and scope of impact of each triggering rule are converted into decision instructions, including specific operational steps, required resources, and timelines, ensuring accurate execution during actual implementation. All decision instructions are summarized to generate a set of policy rule entries, each containing complete execution rules, conditions, and order, ensuring consistency and accuracy in decision execution.

[0084] This embodiment also provides a computer device applicable to the fault analysis method based on line loss anomalies, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the fault analysis method based on line loss anomalies proposed in the above embodiment.

[0085] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0086] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the fault analysis method based on line loss anomalies as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0087] In summary, this invention improves the accuracy of fault cause analysis by: conducting in-depth big data analysis on management behavior records to generate management behavior fingerprints, accurately identifying and classifying different management behaviors, and effectively matching management behaviors with abnormal line loss events; and simultaneously performing predictive residual analysis on abnormal line loss events, which can accurately identify potential factors that may cause abnormal line loss in the context of big data, thus achieving real-time and efficient fault early warning and analysis.

[0088] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fault analysis method based on line loss anomalies, characterized in that: include, Collect line loss measurement data and management behavior records, extract the subject identifier of the management behavior records and classify the behavior to generate management behavior fingerprints; The system analyzes line loss anomalies in statistical line loss measurement data and performs predictive residual analysis to generate a set of line loss anomaly events. By performing semantic event anchoring and causal semantic alignment between the management behavior fingerprint and the set of line loss anomaly events, the system outputs the behavior explanation path. Assess the causal explanatory power of the behavioral explanation path and identify the causes of failure, generate a set of failure events, calculate the responsibility matching degree and responsibility contribution degree of the set of failure events, and form a set of failure responsibility structures; The set of fault responsibility structures is mapped to the governance operation space for effectiveness evolution assessment and strategy response modeling, and the set of strategy response relationships is output. Extract strategy entries from the strategy response relationship set and perform condition trigger rule solidification processing to generate a strategy rule entry set.

2. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for generating the management behavior fingerprint are as follows: By parsing the data fields of management behavior records, each management behavior record is broken down into the executing entity identifier field, behavior type field, behavior time field, and behavior target field to obtain the entity identity set; Based on the subject identifier field, the execution subjects in the subject identity set are mapped to business role entities and their behavior semantics are categorized to generate management behavior fingerprints.

3. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for generating the set of abnormal line loss events are as follows. The fluctuation range of the line loss measurement data is extracted and reconstructed at the time granularity to form a stable base value sequence; By extracting the baseline trend component and periodic component of the stationary base value sequence, the abnormal event interval of the stationary base value sequence is identified, and the set of abnormal line loss events is output.

4. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for interpreting the output behavior are as follows: Match the set of abnormal line loss events with the behavior types of management behavior fingerprints and perform semantic grouping to generate behavior semantic groups; By analyzing the temporal causal relationship between behaviors and events and the chronological order of events in behavioral semantic groups, a causal inference chain is constructed. Assess the internal consistency, temporal rationality, and logical order of the causal reasoning chain, and prioritize and output the behavioral explanation path.

5. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for generating the set of fault events are as follows: Assess the causal credibility, temporal plausibility, and depth of impact of behavioral explanation paths, and generate causal strength indicators; Calculate the fault cause score of the causal intensity index and classify the cause patterns to generate a set of fault events.

6. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for forming the fault responsibility structure set are as follows: By analyzing the dynamic responsibility patterns of a set of fault events, potential sources of responsibility for the set of fault events are extracted, and a candidate set of potential sources of responsibility is generated. Based on the potential source of responsibility candidate set, calculate the responsibility matching degree of each failure event and classify it to generate a responsibility matching set; Cluster analysis is used to identify causal patterns, device correlations, and user impacts in the responsibility matching set, generating a responsibility cause classification set. By statistically analyzing the equipment damage degree, repair time, frequency of historical events, and user responsibility coefficient of each fault event in the fault cause classification set, the responsibility contribution of each fault event is calculated and decision reasoning is performed to form a fault responsibility structure set.

7. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for defining the output strategy response relationship set are as follows. The system analyzes the responsibility types, event levels, impact scope, historical data, and resource availability in the fault responsibility structure set, performs hierarchical analysis, and generates a set of governance operation strategies. Quantitatively assess the execution effectiveness, risks, resource consumption, and implementation difficulty of a set of governance operation strategies, and generate a set of effectiveness evolution predictions; The response process of the performance evolution prediction set is optimized by the response optimization algorithm to generate a governance strategy response set; Verify the repair effect, resource utilization efficiency, and time response speed of the governance strategy response set, and output the strategy response relationship set.

8. The fault analysis method based on line loss anomalies as described in claim 1, characterized in that: The specific steps for generating the set of strategy rule entries are as follows: Extract the responsibility type, fault level, and resource requirements of each fault event from the set of policy response relationships, and generate a governance operation entry for each fault event; Analyze the response conditions, execution order, and scope of impact for each governance operation item, and perform rule mapping to generate a set of triggering rules for each item; By solidifying the rules, the set of triggering rules for each entry is transformed into executable decision instructions, generating a set of policy rule entries.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the fault analysis method based on line loss anomalies as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the fault analysis method based on line loss anomalies as described in any one of claims 1 to 8.

Images