A substation alarm-driven report generation method and system

Abstract

This invention discloses a method and system for generating alarm-driven reports in substations, relating to the fields of power system data processing and intelligent reporting. The method includes: acquiring substation alarm event information and constructing an alarm context object; generating candidate measurement point results and candidate time segments based on the alarm context object; combining alarm scenario knowledge constraint information to generate a target associated measurement point set and an adaptive collection time range, and forming a targeted collection strategy; acquiring an alarm-related dataset according to the targeted collection strategy; and selecting a report template and arranging content based on alarm event information, the alarm context object, the alarm-related dataset, template rules, and alarm scenario knowledge constraint information, generating and outputting a report instance. This invention enables automatic determination of associated measurement points and collection time ranges in substation alarm scenarios, improving the targeting of data collection, the completeness of abnormal process coverage, and the automation level of report generation.

Description

Technical Field

[0001] This invention relates to a method and system for generating alarm-driven reports for substations, belonging to the field of power system data processing and intelligent reporting technology. Background Technology

[0002] Substation operation monitoring, equipment alarm analysis, and data linkage processing surrounding alarm events have gradually become important directions in the digital application of power systems. Existing substation main and auxiliary monitoring systems can continuously collect and aggregate large amounts of operational and status data, and the industry is constantly promoting the development of capabilities such as multi-source data fusion, intelligent sensing, predictive early warning, and emergency response. However, existing technologies still have some shortcomings in the business linkage processing after alarm triggering: on the one hand, after an alarm occurs, further screening and analysis are usually required around the alarm-related measurement points and time ranges, resulting in low automation; on the other hand, existing report generation methods are mostly based on fixed cycles, fixed templates, or manual compilation, making it difficult to achieve dynamic triggering and scenario-based generation based on equipment status alarms, and also making it difficult to link alarm information, related measurement point information, and abnormal process information in reports, leading to low efficiency in anomaly analysis and insufficient relevance and timeliness of reports.

[0003] Therefore, how to automatically determine the set of associated measurement points and the time range of data collection around substation alarm events, and how to dynamically construct reports based on this, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method and system for generating substation alarm-driven reports. This is achieved by constructing a candidate measurement point generation mechanism and a candidate time segment generation mechanism for alarm scenarios. A decision-making intelligent agent then performs secondary correction on the candidate results based on alarm scenario knowledge constraints, generating a targeted data collection strategy. This enables targeted data collection and dynamic report construction for substation alarm scenarios. This invention at least solves the problems of existing technologies, such as reliance on manual selection of associated measurement points after alarm triggering, insufficient flexibility in determining the collection time range, and inadequate automation in report generation. It improves the targeting of data collection in alarm scenarios, the completeness of abnormal process coverage, and the timeliness and adaptability of report generation.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention provides a method for generating substation alarm-driven reports, comprising:

[0007] Obtain substation alarm event information and construct an alarm context object based on the alarm event information;

[0008] Based on the alarm context object, a multi-relationship fusion measurement point filtering mechanism is used to generate candidate measurement point results, and a segmented adaptive time range mechanism is used to generate candidate time segments.

[0009] Based on the candidate measurement point results, candidate time segments, and pre-constructed alarm scenario knowledge constraint information, a target-related measurement point set and an adaptive collection time range are generated, and a targeted collection strategy is formed accordingly.

[0010] Based on the aforementioned targeted data collection strategy, the corresponding data is retrieved and aggregated in a targeted manner to obtain an alarm-related dataset;

[0011] Based on the alarm event information, alarm context object, alarm association dataset, pre-built template rules, and alarm scenario knowledge constraint information, report templates are selected and content is arranged to generate and output a report instance corresponding to the current alarm scenario.

[0012] Furthermore, the alarm scenario knowledge constraint information is a pre-built structured knowledge set for substation alarm scenarios, used to provide scenario constraint basis for generating the targeted collection strategy; the template rules are a pre-built set of template rules for substation alarm scenarios, used to provide rule basis for report template selection and content arrangement; the alarm scenario knowledge constraint information and the template rules are stored in a rule base, knowledge base, or configuration base, and are called during alarm processing; the alarm scenario knowledge constraint information includes at least one or more of the following: the correspondence between alarm type and the object of interest, the correspondence between alarm type and the type of measurement point of interest, the correspondence between alarm type and the accompanying abnormal quantity, the correspondence between alarm type and the recommended time observation mode, and the correspondence between alarm type and the rule of mandatory measurement point; the template rules include at least one or more of the following: the correspondence between alarm type and report template type, the correspondence between alarm type and the focus of template display, template area structure definition, area field requirement definition, and area data mapping rules.

[0013] Furthermore, constructing the alarm context object includes: parsing the object identification information in the alarm event information; associating the device model relationship, interval affiliation relationship, primary and secondary device mapping relationship, and topology connection relationship corresponding to the object identification information; and structurally encapsulating the parsing results and association results to generate the alarm context object for use in generating the candidate measurement point results and the candidate time segment.

[0014] Furthermore, the multi-relationship fusion-based test point filtering mechanism for generating candidate test point results includes: generating an initial set of candidate test points based on the alarm object identifier, device model relationship, interval affiliation relationship, primary / secondary device mapping relationship, and topology connection relationship in the alarm context object, according to preset candidate expansion rules; and performing multi-relationship fusion correlation analysis on the candidate test points in the initial set of candidate test points based on the association, mapping, and semantic relationships between the candidate test points and the alarm objects. The calculation involves the following: the association relationship includes device affiliation relationship, interval affiliation relationship and topology connection relationship; the mapping relationship includes primary and secondary device mapping relationship; and the semantic relationship is the matching relationship between the semantics of candidate measurement points and the semantics of measurement points of interest in the current alarm scenario.

[0015] Furthermore, the initial candidate test point set is generated according to a preset candidate expansion rule, wherein the preset candidate expansion rule includes at least one or more of the following: including the test points corresponding to the device to which the alarm object belongs in the initial candidate test point set; including the test points within the interval to which the alarm object belongs in the initial candidate test point set; including the test points of devices that have a preset primary / secondary device mapping relationship with the device corresponding to the alarm object in the initial candidate test point set; including the test points of devices that meet a preset adjacency depth condition in the topology graph with the device corresponding to the alarm object in the initial candidate test point set; and including the test points corresponding to the key test point type predefined and associated with the current alarm type in the initial candidate test point set.

[0016] Furthermore, the candidate test points in the initial candidate test point set are subjected to multi-relation fusion correlation analysis. Calculation, where, for candidate measurement points Its multi-relationship fusion correlation Represented as:

[0017] ;

[0018] in, Indicates candidate measurement points The correlation score between the alarm object and the alarm object. Indicates candidate measurement points The mapping score between the alarm object and the alarm object Indicates candidate measurement points The matching score between the semantic tags and the semantics of the points of interest in the current alarm scenario; , , These are the weighting coefficients for the corresponding relationships;

[0019] Relationship score Represented as:

[0020] ;

[0021] in, , , For the corresponding weighting coefficients,

[0022] The score for equipment ownership is represented as follows:

[0023] ;

[0024] The interval attribution score is represented as follows:

[0025] ;

[0026] The score for topological connectivity is represented as follows:

[0027] ;

[0028] in, This represents the shortest path length between the candidate test point and the alarm device in the topology graph. The shorter the topology path length, the higher the degree of topological association between the candidate test point and the alarm object, and the higher its corresponding topology connection score.

[0029] Relationship score Represented as:

[0030] ;

[0031] Relationship score Represented as:

[0032] ;

[0033] in, Semantic labels representing candidate measurement points This indicates that the alarm scenario focuses on the semantic tags of the test points. This represents a semantic matching function.

[0034] Furthermore, the method for generating the target associated measurement point set includes: marking candidate measurement points with a comprehensive priority score higher than a preset threshold as priority candidate measurement points; identifying the device object of interest, the type of measurement point of interest, and the accompanying abnormal quantity corresponding to the current alarm scenario; retaining or eliminating the priority candidate measurement points; wherein, candidate measurement points that match the device object of interest and the type of measurement point of interest in the current alarm scenario are retained, and candidate measurement points with a comprehensive priority score higher than the preset threshold but not matching the current alarm scenario are eliminated; candidate measurement points that do not reach the preset threshold but are mandatory measurement points in the current alarm scenario or are related to the accompanying abnormal quantity are supplemented; and the set of candidate measurement points after retention, elimination, and supplementation is used as the target associated measurement point set.

[0035] Furthermore, the step of generating candidate time segments using a segmented adaptive time range mechanism includes: generating candidate pre-alarm time segments based on the alarm occurrence time; generating candidate duration segments based on the alarm occurrence time and alarm recovery time; and generating candidate post-alarm time segments based on the alarm recovery time. The candidate collection time range... Represented as:

[0036]

[0037] in, The candidate lead time before the alarm occurs. The duration from the time the alarm occurs to the time the alarm is recovered. The candidate post-recovery time length after alarm recovery;

[0038] The and Represented as:

[0039]

[0040]

[0041] in, Based on the length of the lead time. Based on the post-time length, Based on the sampling period, , For adjustment coefficients, The degree of abnormal fluctuation at the target measurement point. To restore the dispersion index;

[0042] The degree of abnormal fluctuation Represented as:

[0043]

[0044] in, The value of a measurement point within the alarm window at a specific moment. This is the normal reference value for the measuring point. It is a tiny positive number;

[0045] The recovery dispersion Represented as:

[0046]

[0047] Where n is the number of sampling points involved in the calculation surrounding the alarm recovery process. To focus on the j-th sampled value during the alarm recovery process, This is the average of all sampled values ​​during the recovery phase. The degree to which the j-th sampling point deviates from the average value.

[0048] Furthermore, the method for generating the adaptive acquisition time range includes: when the alarm event is a repeated alarm, performing a merging process on adjacent candidate time segments to form a complete abnormal process segment, specifically:

[0049] The candidate time interval set is defined as follows:

[0050]

[0051] When two adjacent candidate time intervals satisfy:

[0052] in, To set a preset merging threshold, the two candidate time segments are merged into:

[0053]

[0054] Repeat the merging operation of adjacent candidate time segments until there are no adjacent candidate time segments that meet the merging conditions; treat the final merging result as a complete abnormal process segment.

[0055] Combining the temporal response characteristics of the candidate measurement points in the candidate measurement point results and the knowledge constraint information of the alarm scenario, the pre-time boundary and post-time boundary corresponding to the complete abnormal process segment are expanded or pruned to form the adaptive acquisition time range.

[0056] Specifically, when there are candidate test points that exhibit abnormal changes before the alarm occurs, the pre-time boundary is expanded; when there are candidate test points that continue to change after the alarm recovery time, the post-time boundary is expanded; when the abnormal changes of candidate test points are concentrated near the alarm occurrence time and the duration is shorter than a preset threshold, the pre-time boundary and post-time boundary are trimmed.

[0057] Furthermore, based on the targeted acquisition strategy, the corresponding data is targeted for retrieval and aggregation to obtain an alarm association dataset, including: acquiring data of the corresponding measurement points within the adaptive acquisition time range based on the target associated measurement point set; and forming the alarm association dataset based on the acquired data.

[0058] Furthermore, the step of selecting a report template and arranging content to generate a report instance corresponding to the current alarm scenario includes: selecting a target report template based on the alarm type, alarm object type, type of monitoring point, abnormal process characteristics, as well as the template rules and the knowledge constraint information of the alarm scenario; arranging the content of the data area, chart area, and abnormal annotation area in the target report template based on the key monitoring point data and abnormal segment data in the alarm association dataset; and generating a report instance including the abnormal annotation area and the chart area.

[0059] Secondly, the present invention provides a substation alarm-driven report generation system for implementing the substation alarm-driven report generation method described in any one of the preceding claims, comprising:

[0060] An alarm context construction module is used to obtain substation alarm event information and construct an alarm context object based on the alarm event information.

[0061] The candidate test point generation module is used to generate candidate test point results based on the alarm context object using a multi-relationship fusion test point filtering mechanism.

[0062] The candidate time segment generation module is used to generate candidate time segments based on the alarm context object using a segmented adaptive time range mechanism.

[0063] The decision-making intelligent agent module is used to generate a set of target-related measurement points and an adaptive collection time range based on candidate measurement point results, candidate time segments and alarm scenario knowledge constraints. Based on this, a targeted collection strategy is formed, and report templates are selected and content is arranged based on alarm event information, alarm context objects, alarm related datasets, template rules and alarm scenario knowledge constraints.

[0064] The data acquisition module is used to perform targeted retrieval and aggregation of relevant data according to the targeted acquisition strategy to obtain an alarm-related dataset;

[0065] The report output module is used to output report instances corresponding to the current alarm scenario.

[0066] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0067] Fourthly, the present invention provides an electronic device, comprising:

[0068] Memory, used to store computer programs / instructions;

[0069] A processor for executing the computer program / instructions to implement the steps of any of the methods described above.

[0070] Fifthly, the present invention provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of any of the methods described above.

[0071] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0072] This invention provides a method and system for generating alarm-driven reports in substations. By linking alarm events, measurement point model data, associated equipment information, and report templates, it solves the problems in existing technologies where manual selection of associated measurement points, manual determination of data collection time range, and manual compilation of reports are required after an alarm is triggered. It realizes targeted measurement point collection, associated data aggregation, and dynamic report generation around alarm events, thereby improving alarm analysis efficiency, data collection relevance, and the degree of automation in report generation. Attached Figure Description

[0073] Figure 1 This is a flowchart of a substation alarm-driven report generation method provided in an embodiment of the present invention;

[0074] Figure 2 This is a schematic diagram of a substation alarm-driven report generation system provided in an embodiment of the present invention. Detailed Implementation

[0075] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0076] Example 1: This example describes a method for generating alarm-driven reports for substations, including:

[0077] Obtain substation alarm event information and construct an alarm context object based on the alarm event information;

[0078] Based on the alarm context object, a multi-relationship fusion measurement point filtering mechanism is used to generate candidate measurement point results, and a segmented adaptive time range mechanism is used to generate candidate time segments.

[0079] Based on the candidate measurement point results, candidate time segments, and pre-constructed alarm scenario knowledge constraint information, a target-related measurement point set and an adaptive collection time range are generated, and a targeted collection strategy is formed accordingly.

[0080] Based on the aforementioned targeted data collection strategy, the corresponding data is retrieved and aggregated in a targeted manner to obtain an alarm-related dataset;

[0081] Based on the alarm event information, alarm context object, alarm association dataset, pre-built template rules, and alarm scenario knowledge constraint information, report templates are selected and content is arranged to generate and output a report instance corresponding to the current alarm scenario.

[0082] like Figure 1 As shown in the figure, the substation alarm-driven report generation method provided in this embodiment involves the following steps in its application process:

[0083] Step S100: Obtain substation alarm event information and construct an alarm context object based on the alarm event information.

[0084] Specifically, the system receives alarm event information from the monitoring system. This alarm event information includes at least the alarm object identifier, equipment name, associated bay, alarm type, alarm level, alarm occurrence time, alarm recovery time, and associated operating conditions. An alarm context object is then constructed based on this information. For example, if an alarm is generated on the A-phase bushing of main transformer No. 1, the system first parses the original alarm event information to extract the object identifier, time information, and alarm attribute information. The equipment object is "A-phase bushing of main transformer No. 1," the associated bay is "220kV main transformer bay No. 1," the alarm type is "abnormal temperature rise," the alarm level is level two, and the adjacent objects in the equipment topology include the high-voltage side bushing current measuring point, the top-level oil temperature measuring point, the cooler A group status measuring point, and the winding hot spot temperature measuring point.

[0085] After completing the information parsing and relationship association, the system encapsulates the parsing and association results in a structured manner to generate an alarm context object. The alarm context object can be encapsulated in the form of a structured set of fields, object instances, or a set of key-value pairs.

[0086] Specifically, the generated alarm context object can be represented as follows: the alarm object is "A-phase bushing of No. 1 main transformer"; the equipment to which it belongs is "No. 1 main transformer"; the bay to which it belongs is "220kV No. 1 main transformer bay"; the alarm type is "abnormal temperature rise"; the alarm level is level two; the alarm occurrence time is 14:23:10; the alarm recovery time is 14:36:40; the operating condition is "high load operation"; the main and auxiliary equipment mapping objects include cooler group A and cooler outlet oil temperature measuring point; the topology association objects include high-voltage side bushing current measuring point, top layer oil temperature measuring point and winding hot spot temperature measuring point.

[0087] Step S200: Based on the alarm context object, a multi-relationship fusion measurement point filtering mechanism is used to generate candidate measurement point results, and a segmented adaptive time range mechanism is used to generate candidate time segments.

[0088] Specifically, the system first generates an initial candidate measurement point set based on the alarm object identifier, device model relationship, interval affiliation relationship, primary and secondary device mapping relationship, and topology connection relationship in the alarm context object, according to preset candidate expansion rules. In this embodiment, the preset candidate expansion rules specifically include: including measurement points of the device to which the alarm object belongs in the initial candidate measurement point set; including relevant measurement points within the interval to which the alarm object belongs in the initial candidate measurement point set; including measurement points of the cooler device that has a primary and secondary device mapping relationship with the alarm object in the initial candidate measurement point set; including current measurement points that meet the preset adjacency depth condition with the alarm object in the topology graph in the initial candidate measurement point set; and including measurement points corresponding to key measurement point types predefined in association with temperature rise alarms in the initial candidate measurement point set.

[0089] According to the above rules, the initial candidate measurement point set generated in this embodiment includes: A-phase bushing temperature, top layer oil temperature, winding hot spot temperature, cooler A group operating status, low-voltage side A-phase current, on-load tap changer position, No. 2 main transformer A-phase bushing temperature, A-phase bushing dielectric loss, and cooler outlet oil temperature. Subsequently, the system performs multi-relationship fusion correlation degree calculation on each candidate measurement point based on correlation, mapping, and semantic relationships.

[0090] In this embodiment, the formula for multi-relationship fusion correlation degree is as follows:

[0091]

[0092] Where α, β, and γ are taken as 0.5, 0.2, and 0.3 respectively; correlation score For the formula:

[0093]

[0094] in, , , Take values ​​of 0.4, 0.2, and 0.4 respectively.

[0095] Equipment ownership score for:

[0096]

[0097] Interval Attribution Score for:

[0098]

[0099] Topology connection score for:

[0100]

[0101] in, This represents the shortest path length between the candidate test point and the alarm device in the topology graph. The shorter the topology path length, the higher the degree of topological association between the candidate test point and the alarm object, and the higher its corresponding topology connection score.

[0102] Relationship score It can be represented as:

[0103]

[0104] Relationship score It can be represented as:

[0105]

[0106] in, Semantic labels representing candidate measurement points This indicates that the alarm scenario focuses on the semantic tags of the test points. This represents a semantic matching function. For a semantic matching function, an N×N matrix can be constructed to calculate the association between two semantic elements.

[0107] The calculation results for each measuring point are shown in Table 1 below:

[0108] Table 1 Calculation results for each measuring point

[0109]

[0110] In this example, the candidate collection time range It can be represented as:

[0111]

[0112] in, The candidate lead time before the alarm occurs. The duration from the time the alarm occurs to the time the alarm is recovered. The candidate post-alarm time length after alarm recovery.

[0113] and It can be represented as:

[0114]

[0115]

[0116] in, The base pre-load time is set to 10 minutes. The base post-processing time is set to 8 minutes. The basic sampling period is 1 minute. , The adjustment coefficients are 12 and 20, respectively; The degree of abnormal fluctuation at the target measurement point. To restore the dispersion index.

[0117] The degree of abnormal fluctuation It can be represented as:

[0118]

[0119] in, The value of a measurement point within the alarm window at a specific moment. This is the normal reference value for the measuring point. It is a tiny positive number, with a value of 0.001 (to prevent the denominator from being zero).

[0120] Restoring Discreteness It can be represented as:

[0121]

[0122] Where n is the number of sampling points involved in the calculation surrounding the alarm recovery process. To focus on the j-th sampled value during the alarm recovery process, This is the average of all sampled values ​​during the recovery phase. The degree to which the j-th sampling point deviates from the average value. It should be a small positive number (to prevent the denominator from being zero).

[0123] For the alarm object A phase bushing temperature, assuming the alarm occurred at 14:23:10 and recovered at 14:36:40; normal reference value. Let's take 68℃. The temperature sequence collected within the alarm window is 69℃, 72℃, 75℃, 79℃, 81℃, and 84℃. Therefore, the maximum temperature deviation is 84-68=16℃. Thus, the degree of abnormal fluctuation... Calculated according to the formula ≈0.2353.

[0124] During the alarm recovery process at the target measurement point, the number of sampling points n is 5, and the sampled values ​​are 80℃, 77℃, 74℃, 72℃, and 71℃ respectively. Their average value... =(80+77+74+72+71) / 5=74.8℃. Therefore, the dispersion is recovered. The calculated value is approximately 0.0396.

[0125] Calculations show that, It is 12.8236 minutes, which is rounded up to 13 minutes; The time interval is 9.792 minutes, which is rounded to 9 minutes. Therefore, the selected time interval before the time interval is [14:10:10, 14:23:10], the candidate duration interval is [14:23:10, 14:36:40], and the candidate time interval after the time interval is [14:36:40, 14:45:40].

[0126] In step S300, the decision-making agent generates a set of target-related measurement points and an adaptive collection time range based on the candidate measurement point results, candidate time segments, and alarm scenario knowledge constraints, and forms a targeted collection strategy accordingly.

[0127] In this embodiment, the decision-making agent marks candidate test points with a comprehensive priority score higher than a preset threshold of 0.55 as priority candidate test points, i.e. , , , , Subsequently, the decision-making agent invokes the alarm scenario knowledge constraint information to identify the relevant equipment object corresponding to the current alarm scenario as the main transformer bushing and its thermally related components, the relevant measurement point type as temperature-related and cooling status-related, and the accompanying abnormal quantity as load current. Based on this, the decision-making agent retains... , , , , and And although the threshold was not reached, it is a mandatory test point for the scenario. (Phase A current) is added to the target associated measurement point set; at the same time, p6 and p7, which are not relevant to the current alarm scenario, are removed. The final target associated measurement point set is determined as { , , , , , , }

[0128] Furthermore, considering that the same alarm object triggered a temperature rise anomaly alarm again at 14:47:20 and recovered at 14:49:10, and according to the same method, the candidate pre-alarm time segments for the second alarm are [14:34:20, 14:47:20], the candidate duration segments are [14:47:20, 14:49:10], and the candidate post-alarm time segments are [14:49:10, 14:58:30]. Since the intervals corresponding to the second set of candidate time segments are adjacent and overlap with the intervals of the first set in time, the calculated interval between the two is less than the preset merging threshold. The time interval is 5 minutes. Therefore, the decision-making agent performs merging processing on candidate time intervals that are adjacent in time to form a complete abnormal process interval.

[0129] The candidate time interval set is as follows:

[0130]

[0131] When two adjacent candidate time intervals satisfy:

[0132] in, To preset the merging threshold, it can be set according to the alarm type, device object type, or basic sampling period. In this example, Set to 5 minutes, the two candidate time segments are merged into:

[0133]

[0134] In this embodiment, after repeatedly performing the merging of adjacent segments, the complete abnormal process segment is obtained as [14:10:10, 14:58:30].

[0135] After merging the segments, the decision-making agent combines the temporal response characteristics of candidate measurement points and the knowledge constraint information of alarm scenarios to expand or prune the pre-time boundary and post-time boundary to form an adaptive acquisition time range.

[0136] In this embodiment, the decision-making agent detected that the current of phase A at candidate measurement point had been rising continuously at 14:08:40, earlier than the alarm occurrence time of 14:23:10; simultaneously, the top oil temperature at candidate measurement point continued to decline after 14:58:30, until it finally returned to the preset normal range at 15:02:20. Therefore, the decision-making agent extended the pre-time boundary of the complete abnormal process segment from 14:10:10 to 14:08:40, and extended the post-time boundary from 14:58:30 to 15:02:20.

[0137] If the abnormal changes of candidate measurement points are concentrated near the alarm occurrence time and the duration is shorter than a preset threshold, such as less than 3 minutes, the decision-making agent can also perform pruning on the preceding and following time boundaries to reduce data from normal operating time periods that are irrelevant to the current alarm scenario. In this embodiment, since both the precursor and lagging changes are quite obvious, the pruning condition was not triggered. Finally, the adaptive acquisition time range was determined to be [14:08:40, 15:02:20].

[0138] Step S400: Based on the targeted collection strategy, the corresponding data is targeted for retrieval and aggregation to obtain the alarm association dataset.

[0139] Specifically, the system acquires , , , , , and The time series data is within the interval [14:08:40, 15:02:20], where temperature measurement points are sampled once every 1 minute, and status measurement points are recorded according to the time of event change.

[0140] After data acquisition, an alarm association dataset is formed. This dataset includes at least: key measurement point data, abnormal section data, and key field data. The key measurement point data includes time-series data of A-phase bushing temperature, top-layer oil temperature, winding hot spot temperature, and A-phase current; the abnormal section data includes the first abnormal temperature rise process section, the second repeated abnormal process section, and the merged complete abnormal process section; the key field data includes the alarm occurrence time, alarm recovery time, target associated measurement point set, complete abnormal process section, and adaptive acquisition time range.

[0141] In step S500, the decision-making agent selects candidate report templates based on the alarm event information, alarm context object, alarm associated dataset, and pre-built template rules and scenario knowledge constraints to confirm the target template. The target report template predefines data areas, chart areas, and anomaly annotation areas to form the template area structure definition. The decision-making agent generates a report instance corresponding to the current alarm scenario by arranging the content of the data areas, chart areas, and anomaly annotation areas in the target report template.

[0142] Specifically, since the alarm type is abnormal temperature rise of the main transformer bushing, and the template selection prior knowledge pre-sets the correspondence between "temperature rise alarm - main transformer equipment object - main transformer temperature rise alarm dynamic report template", the decision-making agent selects "main transformer temperature rise alarm dynamic report template" as the target report template.

[0143] The target report template structurally includes a data area, a chart area, and an anomaly annotation area. Based on the area field requirements and area data mapping rules defined in the template rules, the decision-making agent maps the basic information in the alarm-related dataset to the data area, maps the time-series data of A-phase bushing temperature, top-layer oil temperature, winding hotspot temperature, and A-phase current to the chart area, and identifies key moments such as 14:23:10, 14:36:40, 14:47:20, 14:49:10, and 15:02:20 in the anomaly annotation area.

[0144] Step S600: Display, output, or store the report instance.

[0145] Specifically, the final generated report instance includes: an anomaly annotation area to display key alarm times and complete anomaly process segments; a chart area to display the changing trends of key temperature measurement points and load current; and a set of data areas to display key information such as alarm objects, alarm types, target associated measurement point sets, and adaptive data acquisition time range. In this way, alarm information, associated measurement point information, and anomaly process information can be presented in a linked manner within the same report instance.

[0146] Example 2, as Figure 2 As shown, this embodiment provides a substation alarm-driven report generation system, including: an alarm context construction module, used to construct an alarm context object based on the alarm event information; a candidate measurement point generation module, used to generate candidate measurement point results based on the alarm context object using a multi-relationship fusion measurement point filtering mechanism; a candidate time segment generation module, used to generate candidate time segments based on the alarm context object using a segmented adaptive time range mechanism; a decision intelligence module, used to generate a target associated measurement point set and an adaptive collection time range based on the candidate measurement point results, candidate time segments, and alarm scenario knowledge constraint information, and to select report templates and arrange content based on alarm event information, alarm context object, alarm associated dataset, template rules, and alarm scenario knowledge constraint information; a data acquisition module, used to acquire data of corresponding measurement points within the adaptive collection time range according to the targeted collection strategy, and form an alarm associated dataset; and a report output module, used for displaying, outputting, or storing report instances.

[0147] The following section details the specific configuration of the decision-making agent module. This module, the core decision-making unit of this invention, includes a measurement point correction unit, a time correction unit, and a template selection and arrangement unit. It is used to generate a targeted data collection strategy based on candidate measurement point results, candidate time segments, and alarm scenario knowledge constraints, and to complete report template selection and content arrangement based on alarm event information, alarm context objects, alarm-related datasets, template rules, and alarm scenario knowledge constraints.

[0148] The measurement point correction unit is used to retain, eliminate, and supplement candidate measurement point results based on alarm scenario knowledge constraint information to generate a target-related measurement point set. Candidate measurement points with a comprehensive priority score higher than a preset threshold are marked as priority candidate measurement points. Furthermore, based on the current alarm scenario's corresponding attention device object, attention measurement point type, and accompanying anomaly quantity, scenario-based correction is performed on the candidate measurement points to generate the target-related measurement point set.

[0149] The time correction unit is used to merge, expand, and prune candidate time segments to generate an adaptive acquisition time range. By merging temporally adjacent candidate time segments to form a complete abnormal process segment, and combining the temporal response characteristics of candidate measurement points and alarm scenario knowledge constraint information, the preceding and following time boundaries corresponding to the complete abnormal process segment are expanded or pruned to form an adaptive acquisition time range.

[0150] The template selection and arrangement unit is used to select a target report template based on alarm event information, alarm context object, alarm associated dataset, template rules and alarm scenario knowledge constraint information, and to arrange the content of the data area, chart area and anomaly annotation area in the target report template to generate a report instance.

[0151] The specific functions of each module described above are explained in the relevant content of the method in Embodiment 1, and will not be repeated here.

[0152] Example 3: This example provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described in Example 1.

[0153] Example 4: This example provides an electronic device, including:

[0154] Memory, used to store computer programs / instructions;

[0155] A processor for executing the computer program / instructions to implement the steps of any of the methods described in Embodiment 1.

[0156] Example 5: This example provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the method described in any one of Examples 1.

[0157] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0158] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0160] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this disclosure and not to limit its protection scope. Although this disclosure has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading this disclosure, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the pending claims.

Claims

1. A method for generating alarm-driven reports in substations, characterized in that, include: Obtain substation alarm event information and construct an alarm context object based on the alarm event information; Based on the alarm context object, a multi-relationship fusion measurement point filtering mechanism is used to generate candidate measurement point results, and a segmented adaptive time range mechanism is used to generate candidate time segments. Based on the candidate measurement point results, candidate time segments, and pre-constructed alarm scenario knowledge constraint information, a target-related measurement point set and an adaptive collection time range are generated, and a targeted collection strategy is formed accordingly. Based on the aforementioned targeted data collection strategy, the corresponding data is retrieved and aggregated in a targeted manner to obtain an alarm-related dataset; Based on the alarm event information, alarm context object, alarm association dataset, pre-built template rules, and alarm scenario knowledge constraint information, report templates are selected and content is arranged to generate and output a report instance corresponding to the current alarm scenario.

2. The substation alarm drive report generation method according to claim 1, characterized in that, The alarm scenario knowledge constraint information is a pre-built structured knowledge set for substation alarm scenarios, used to provide scenario constraint basis for generating the targeted collection strategy; the template rules are a pre-built template rule set for substation alarm scenarios, used to provide rule basis for report template selection and content arrangement; the alarm scenario knowledge constraint information and the template rules are stored in a rule base, knowledge base, or configuration base, and are called during alarm processing; the alarm scenario knowledge constraint information includes at least one or more of the following: the correspondence between alarm type and the object of interest, the correspondence between alarm type and the type of measurement point of interest, the correspondence between alarm type and the accompanying abnormal quantity, the correspondence between alarm type and the recommended time observation mode, and the correspondence between alarm type and the rule of mandatory measurement point; the template rules include at least one or more of the following: the correspondence between alarm type and report template type, the correspondence between alarm type and the focus of template display, template area structure definition, area field requirement definition, and area data mapping rules.

3. The substation alarm drive report generation method according to claim 1, characterized in that, The construction of the alarm context object includes: parsing the object identification information in the alarm event information; associating the device model relationship, interval affiliation relationship, primary and secondary device mapping relationship and topology connection relationship corresponding to the object identification information; and encapsulating the parsing results and association results in a structured manner to generate the alarm context object for the generation and invocation of the candidate measurement point results and the candidate time segment.

4. The substation alarm drive report generation method according to claim 1, characterized in that, The multi-relationship fusion-based test point filtering mechanism for generating candidate test point results includes: generating an initial set of candidate test points based on the alarm object identifier, device model relationship, interval attribution relationship, primary / secondary device mapping relationship, and topology connection relationship in the alarm context object, according to preset candidate expansion rules; and performing multi-relationship fusion correlation analysis on the candidate test points in the initial set of candidate test points based on the association, mapping, and semantic relationships between the candidate test points and the alarm objects. The calculation involves the following: the association relationship includes device affiliation relationship, interval affiliation relationship and topology connection relationship; the mapping relationship includes primary and secondary device mapping relationship; and the semantic relationship is the matching relationship between the semantics of candidate measurement points and the semantics of measurement points of interest in the current alarm scenario.

5. The substation alarm drive report generation method according to claim 4, characterized in that, The initial candidate test point set is generated according to a preset candidate expansion rule, wherein the preset candidate expansion rule includes at least one or more of the following: including the test points corresponding to the device to which the alarm object belongs in the initial candidate test point set; including the test points within the interval to which the alarm object belongs in the initial candidate test point set; including the test points of devices that have a preset primary / secondary device mapping relationship with the device corresponding to the alarm object in the initial candidate test point set; including the test points of devices that meet a preset adjacency depth condition with the device corresponding to the alarm object in the topology graph in the initial candidate test point set; and including the test points corresponding to the key test point type predefined and associated with the current alarm type in the initial candidate test point set.

6. The substation alarm drive report generation method according to claim 4, characterized in that, Perform multi-relation fusion correlation analysis on the candidate test points in the initial candidate test point set. Calculation, where, for candidate measurement points Its multi-relationship fusion correlation Represented as: ； in, Indicates candidate measurement points The correlation score between the alarm object and the alarm object. Indicates candidate measurement points The mapping score between the alarm object and the alarm object Indicates candidate measurement points The matching score between the semantic tags and the semantics of the points of interest in the current alarm scenario; , , These are the weighting coefficients for the corresponding relationships; Relationship score Represented as: ； in, , , For the corresponding weighting coefficients, The score for equipment ownership is represented as follows: ； The interval affiliation score is represented as: ； The score for topological connectivity is represented as follows: ； in, This represents the shortest path length between the candidate test point and the alarm device in the topology graph. The shorter the topology path length, the higher the degree of topological association between the candidate test point and the alarm object, and the higher its corresponding topology connection score. Relationship score Represented as: ； Relationship score Represented as: ； in, Semantic labels representing candidate measurement points This indicates that the alarm scenario focuses on the semantic tags of the test points. This represents a semantic matching function.

7. The substation alarm drive report generation method according to claim 1, characterized in that, The method for generating the target associated measurement point set includes: marking candidate measurement points with a comprehensive priority score higher than a preset threshold as priority candidate measurement points; identifying the device object of interest, the type of measurement point of interest, and the accompanying abnormal quantity corresponding to the current alarm scenario; retaining or eliminating the priority candidate measurement points; wherein, retaining candidate measurement points that match the device object of interest and the type of measurement point of interest in the current alarm scenario, and eliminating candidate measurement points with a comprehensive priority score higher than the preset threshold but not matching the current alarm scenario; supplementing candidate measurement points that do not reach the preset threshold but are mandatory measurement points in the current alarm scenario or are related to the accompanying abnormal quantity; and using the set of candidate measurement points after retention, elimination, and supplementation as the target associated measurement point set.

8. The substation alarm drive report generation method according to claim 1, characterized in that, The segmented adaptive time range mechanism for generating candidate time segments includes: generating candidate pre-alarm time segments based on the alarm occurrence time; generating candidate duration segments based on the alarm occurrence time and alarm recovery time; and generating candidate post-alarm time segments based on the alarm recovery time. The candidate collection time range... Represented as: in, The candidate lead time before the alarm occurs. The duration from the time the alarm occurs to the time the alarm is recovered. The candidate post-recovery time length after alarm recovery; The and Represented as: in, Based on the length of the lead time. Based on the post-time length, Based on the sampling period, , For adjustment coefficients, The degree of abnormal fluctuation at the target measurement point. To restore the dispersion index; The degree of abnormal fluctuation Represented as: in, The value of a measurement point within the alarm window at a specific moment. This is the normal reference value for the measuring point. It is a tiny positive number; The recovery dispersion Represented as: Where n is the number of sampling points involved in the calculation surrounding the alarm recovery process. To focus on the j-th sampled value during the alarm recovery process, This is the average of all sampled values ​​during the recovery phase. The degree to which the j-th sampling point deviates from the average value.

9. The substation alarm drive report generation method according to claim 1, characterized in that, The method for generating the adaptive acquisition time range includes: when the alarm event is a repeated alarm, performing a merging process on adjacent candidate time segments to form a complete abnormal process segment, specifically: The candidate time interval set is defined as follows: When two adjacent candidate time intervals satisfy: in, To set a preset merging threshold, the two candidate time segments are merged into: Repeat the merging operation of adjacent candidate time segments until there are no adjacent candidate time segments that meet the merging conditions; treat the final merging result as a complete abnormal process segment. Combining the temporal response characteristics of the candidate measurement points in the candidate measurement point results and the knowledge constraint information of the alarm scenario, the pre-time boundary and post-time boundary corresponding to the complete abnormal process segment are expanded or pruned to form the adaptive acquisition time range. Specifically, when there are candidate test points that exhibit abnormal changes before the alarm occurs, the pre-time boundary is expanded; when there are candidate test points that continue to change after the alarm recovery time, the post-time boundary is expanded; when the abnormal changes of candidate test points are concentrated near the alarm occurrence time and the duration is shorter than a preset threshold, the pre-time boundary and post-time boundary are trimmed.

10. The substation alarm drive report generation method according to claim 1, characterized in that, Based on the targeted acquisition strategy, the corresponding data is targeted for retrieval and aggregation to obtain an alarm association dataset, including: acquiring data of the corresponding measurement points within the adaptive acquisition time range based on the target associated measurement point set; and forming the alarm association dataset based on the acquired data.

11. The substation alarm drive report generation method according to claim 1, characterized in that, The step of selecting a report template and arranging content to generate a report instance corresponding to the current alarm scenario includes: selecting a target report template based on the alarm type, alarm object type, type of monitoring point, abnormal process characteristics, as well as the template rules and the knowledge constraint information of the alarm scenario; arranging the content of the data area, chart area, and abnormal annotation area in the target report template based on the key monitoring point data and abnormal segment data in the alarm association dataset; and generating a report instance including the abnormal annotation area and the chart area.

12. A substation alarm-driven report generation system, used to implement the substation alarm-driven report generation method according to any one of claims 1-11, characterized in that, include: An alarm context construction module is used to obtain substation alarm event information and construct an alarm context object based on the alarm event information. The candidate test point generation module is used to generate candidate test point results based on the alarm context object using a multi-relationship fusion test point filtering mechanism. The candidate time segment generation module is used to generate candidate time segments based on the alarm context object using a segmented adaptive time range mechanism. The decision-making intelligent agent module is used to generate a set of target-related measurement points and an adaptive collection time range based on candidate measurement point results, candidate time segments and alarm scenario knowledge constraints. Based on this, a targeted collection strategy is formed, and report templates are selected and content is arranged based on alarm event information, alarm context objects, alarm related datasets, template rules and alarm scenario knowledge constraints. The data acquisition module is used to perform targeted retrieval and aggregation of relevant data according to the targeted acquisition strategy to obtain an alarm-related dataset; The report output module is used to output report instances corresponding to the current alarm scenario.

Images