Electrical equipment intelligent fault early warning method and device based on edge calculation, equipment and medium

Abstract

The invention relates to the technical field of Internet of Things. The invention provides an electrical equipment intelligent fault early warning method and device based on edge computing, equipment and a medium. The method comprises the following steps: collecting equipment operation data and historical fault data through an edge computing node; based on the equipment operation data and the historical fault data, performing fault live-action reconstruction processing, and generating a three-dimensional fault evolution model; performing playback analysis processing on the three-dimensional fault evolution model through an interactive time axis to generate a fault root cause report; and on the basis of the three-dimensional fault evolution model, maintenance strategy simulation deduction processing is carried out, and a maintenance scheme evaluation result is generated, so that the technical effects of improving the fault response timeliness, enhancing the interpretability of equipment internal fault propagation and optimizing the maintenance scheme decision scientificity are achieved.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) technology, and in particular to a method, apparatus, device, and medium for intelligent fault early warning of electrical equipment based on edge computing. Background Technology

[0002] With the continuous advancement of intelligent upgrades to power systems, the safe and stable operation of electrical equipment has a significant impact on industrial production and social life. Fault early warning, as a core component in ensuring equipment reliability, directly determines equipment maintenance efficiency and accident prevention capabilities through its accuracy and timeliness.

[0003] However, the relevant fault early warning technologies have the following problems: the centralized cloud processing architecture leads to reduced response timeliness, which cannot meet the needs of rapid fault prevention and control of electrical equipment; the two-dimensional static model is difficult to reveal the spatial topological relationship between the deterioration of the internal structure of the equipment and the propagation of faults; and the human experience-driven mechanism cannot quantify and evaluate the chain effect of multiple maintenance schemes. Summary of the Invention

[0004] Therefore, it is necessary to provide methods, devices, equipment and media for intelligent fault early warning of electrical equipment based on edge computing to address the above-mentioned technical problems, so as to improve the timeliness of fault response, enhance the interpretability of fault propagation within the equipment and optimize the scientific nature of maintenance plan decisions.

[0005] In a first aspect, this application provides a method for intelligent fault early warning of electrical equipment based on edge computing, the method comprising:

[0006] Collect device operation data and historical fault data through edge computing nodes;

[0007] Based on equipment operation data and historical fault data, fault scene reconstruction processing is performed to generate a three-dimensional fault evolution model.

[0008] The three-dimensional fault evolution model is replayed and analyzed using an interactive timeline to generate a fault root cause report.

[0009] Based on a three-dimensional fault evolution model, maintenance strategies are simulated and deduced to generate maintenance plan evaluation results.

[0010] Furthermore, based on the three-dimensional fault evolution model, maintenance strategy simulation and deduction are performed to generate maintenance plan evaluation results, including:

[0011] A virtual maintenance scenario is constructed based on a three-dimensional fault evolution model;

[0012] Load multiple repair solutions into a virtual repair scenario;

[0013] Using digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram.

[0014] The recovery trend chart and the risk chain effect chart are compared and evaluated to generate the maintenance plan evaluation results.

[0015] Furthermore, through digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram, including:

[0016] For each maintenance plan, a virtual execution environment is constructed and processed to generate a unique digital twin of the plan;

[0017] Fault evolution parameters are injected into the solution-specific digital twin, and these parameters are derived from the three-dimensional fault evolution model.

[0018] By using a dynamic propagation algorithm, the cascading effect simulation of state changes in associated nodes of the device is performed to obtain the cascading effect simulation results;

[0019] Based on the simulation results of chain effects, multidimensional state trajectory reconstruction is performed to generate state recovery trend diagrams and risk chain effect diagrams.

[0020] Furthermore, the recovery trend chart and risk cascading effect chart are compared and evaluated to generate maintenance plan evaluation results, including:

[0021] The recovery efficiency feature is extracted from the state recovery trend map to generate a recovery efficiency feature vector.

[0022] The risk propagation feature vector is generated by performing deep feature extraction on the risk chain effect diagram.

[0023] Input the recovery efficiency feature vector and the risk propagation feature vector into the preset multidimensional evaluation model to generate the safety evaluation value, cost evaluation value and timeliness evaluation value of each maintenance plan;

[0024] Based on safety assessment values, cost assessment values, and timeliness assessment values, a comprehensive priority ranking process is performed to generate maintenance plan evaluation results.

[0025] Furthermore, the three-dimensional fault evolution model is replayed and analyzed using an interactive timeline to generate a fault root cause report, including:

[0026] Based on the three-dimensional fault evolution model, spatiotemporal decoupling is performed to generate the equipment state time series and spatial structure topology.

[0027] Based on the device status time series, a non-linear jumpable time axis interface is constructed and a dynamic playback control function is configured.

[0028] Based on spatial structure topology and dynamic playback control, a multi-scale fault propagation analysis algorithm is used to perform hierarchical marking of key fault nodes and generate hierarchical marking results.

[0029] Based on the hierarchical labeling results, reverse reconstruction and forward verification of the fault causal chain are performed to generate an initial root cause analysis report.

[0030] Based on the initial root cause analysis report, the reliability verification process is carried out by integrating the physical mechanism constraints of the equipment, and a fault root cause report is generated.

[0031] Furthermore, based on the hierarchical labeling results, reverse reconstruction and forward verification of the fault causal chain are performed to generate an initial root cause analysis report, including:

[0032] Based on the hierarchical labeling results, a reverse traversal of the fault propagation tree is performed to generate an initial set of fault source points.

[0033] The initial set of fault source points is filtered by physical mechanism constraints to generate candidate fault source points;

[0034] Based on candidate fault sources, simulated fault signals are injected for forward propagation verification processing to generate fault reproduction data;

[0035] The fault reproduction data and the hierarchical labeling results are compared for spatiotemporal consistency to generate a verification matching degree.

[0036] When the verification match meets the preset threshold, an initial root cause analysis report is generated.

[0037] Furthermore, based on equipment operation data and historical fault data, fault scene reconstruction processing is performed to generate a three-dimensional fault evolution model, including:

[0038] Perform timestamp synchronization and spatial coordinate registration on equipment operation data and historical fault data to generate a spatiotemporal fusion dataset.

[0039] Based on the spatiotemporal fusion dataset, multidimensional feature parameter extraction is performed to generate a fault feature matrix;

[0040] The fault feature matrix is ​​loaded into a pre-set three-dimensional structural model through a physical field coupling mapping mechanism to generate an initial dynamic scene.

[0041] Based on the physical laws of electrical equipment, the evolution trajectory of the initial dynamic scene is iteratively corrected to generate a three-dimensional fault evolution model.

[0042] Secondly, this application also provides an intelligent fault early warning device for electrical equipment based on edge computing, the device comprising:

[0043] The data acquisition module is used to collect device operation data and historical fault data through edge computing nodes;

[0044] The scene reconstruction module is used to reconstruct the fault scene based on equipment operation data and historical fault data, and generate a three-dimensional fault evolution model.

[0045] The root cause analysis module is used to replay and analyze the 3D fault evolution model through an interactive timeline and generate a fault root cause report.

[0046] The strategy deduction module is used to simulate and deduce maintenance strategies based on a three-dimensional fault evolution model, and generate maintenance plan evaluation results.

[0047] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods in the first aspect of this application.

[0048] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods in the first aspect of this application.

[0049] This application provides a method, device, equipment, and medium for intelligent fault early warning of electrical equipment based on edge computing. The method includes: collecting equipment operation data and historical fault data through edge computing nodes; performing fault scene reconstruction processing based on equipment operation data and historical fault data to generate a three-dimensional fault evolution model; performing playback analysis processing of the three-dimensional fault evolution model through an interactive timeline to generate a fault root cause report; and performing maintenance strategy simulation and deduction processing based on the three-dimensional fault evolution model to generate maintenance plan evaluation results, so as to achieve the technical effects of improving the timeliness of fault response, enhancing the interpretability of fault propagation within the equipment, and optimizing the scientific nature of maintenance plan decision-making. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of an intelligent fault early warning method for electrical equipment based on edge computing, according to one embodiment of the present invention.

[0052] Figure 2This is a flowchart of generating a fault root cause report by replaying and analyzing the three-dimensional fault evolution model through an interactive timeline in one embodiment of the present invention.

[0053] Figure 3 This is a structural diagram of an intelligent fault early warning device for electrical equipment based on edge computing, according to one embodiment of the present invention. Detailed Implementation

[0054] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, the specific implementation methods of this application will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0055] First, the application scenarios of the embodiments of this application are described. In the embodiments of this application, a dynamic decision optimization method, system, device, and medium for multi-domain strategy collaboration applicable to, but not limited to, such scenarios are provided.

[0056] As an illustration, the dynamic decision optimization method, system, device and medium for multi-domain strategy collaboration provided in this application embodiment can also be applied to other application scenarios. This is only an example and does not limit the specific application scenarios.

[0057] like Figure 1 As shown, this application provides an intelligent fault early warning method for electrical equipment based on edge computing, the method comprising:

[0058] S101: Collects device operation data and historical fault data through edge computing nodes.

[0059] Specifically, based on the physical structure and monitoring requirements of the electrical equipment, edge computing nodes are deployed at key locations, configured with suitable sensor interfaces to receive real-time operating signals such as voltage, current, and temperature. Historical fault records, including fault type, occurrence time, and related component information, are retrieved from the equipment management system via data interfaces. The collected real-time data undergoes timestamp calibration and noise filtering, and the time dimension of historical fault data is synchronized with the equipment identifier to ensure consistency between the two types of data in the spatiotemporal dimensions. Through local preprocessing modules on the edge nodes, feature quantities in the operating data are extracted and correlated with feature labels in the historical fault data to generate a structured dataset.

[0060] S102: Based on equipment operation data and historical fault data, perform fault scene reconstruction processing to generate a three-dimensional fault evolution model.

[0061] Specifically, spatiotemporal calibration is performed on equipment operation data and historical fault data. Multidimensional data relationships are constructed through timestamp alignment and spatial coordinate mapping. Fault-sensitive parameters are extracted from the calibrated data using feature extraction algorithms, generating a high-dimensional feature vector space. These feature vectors are then mapped to a pre-defined 3D equipment structural model, and a physical field coupling algorithm is used to achieve spatial binding between the data and the model. Finally, based on the constraints of the equipment's physical laws, the initial mapping results are iteratively corrected, and the model parameters are dynamically adjusted until they conform to the fault evolution logic, generating a 3D fault evolution model that dynamically displays the occurrence and development of faults.

[0062] S103: Perform playback analysis on the 3D fault evolution model through an interactive timeline to generate a fault root cause report.

[0063] Specifically, the spatiotemporal dimensions of the 3D fault evolution model are decoupled, separating the equipment state time series from the spatial structural topology. A dynamically jumpable nonlinear time axis interface is constructed based on the time series data, and interactive functions such as rate control and breakpoint marking are configured. Combining the spatial topology, a multi-scale propagation analysis algorithm is used to hierarchically mark key fault nodes, identifying the root cause node, intermediate nodes, and affected nodes in the fault propagation path. Based on the marking results, the fault causal chain is traced backward to generate an initial set of root cause hypotheses. Then, the spatiotemporal consistency of the hypotheses is verified by forward injection of simulated fault signals, and candidate root causes that meet the physical mechanism constraints are selected, generating a fault root cause report containing the fault origin location, triggering conditions, propagation path, and responsibility determination.

[0064] S104: Based on the three-dimensional fault evolution model, perform maintenance strategy simulation and deduction to generate maintenance plan evaluation results.

[0065] Specifically, a virtual maintenance scenario consistent with the physical equipment characteristics is constructed based on a 3D fault evolution model, including equipment structure, connection relationships, and environmental parameters. Multiple preset maintenance schemes are loaded into the virtual scenario, including operation types such as component replacement, parameter adjustment, and system reconfiguration. A unique digital twin is created for each scheme, injecting real-time parameters from the fault evolution model. A dynamic propagation algorithm simulates the chain reaction of maintenance operations between related nodes in the equipment, generating multi-dimensional simulation results containing state recovery trajectories and risk propagation paths. Recovery efficiency features and risk propagation features are then extracted from the simulation results and input into a pre-set multi-dimensional evaluation model to generate quantitative evaluation values ​​and comprehensive priority rankings for each scheme across safety, cost, and timeliness dimensions, resulting in a maintenance scheme evaluation result.

[0066] One embodiment of this application also provides an intelligent fault early warning method for electrical equipment based on edge computing, including: collecting equipment operation data and historical fault data through edge computing nodes; performing fault scene reconstruction processing based on equipment operation data and historical fault data to generate a three-dimensional fault evolution model; performing playback analysis processing of the three-dimensional fault evolution model through an interactive timeline to generate a fault root cause report; and performing maintenance strategy simulation and deduction processing based on the three-dimensional fault evolution model to generate maintenance plan evaluation results, so as to achieve the technical effects of improving fault response timeliness, enhancing the interpretability of fault propagation within the equipment, and optimizing the scientific nature of maintenance plan decision-making.

[0067] Furthermore, based on the three-dimensional fault evolution model, maintenance strategy simulation and deduction are performed to generate maintenance plan evaluation results, including:

[0068] A virtual maintenance scenario is constructed based on a three-dimensional fault evolution model;

[0069] Load multiple repair solutions into a virtual repair scenario;

[0070] Using digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram.

[0071] The recovery trend chart and the risk chain effect chart are compared and evaluated to generate the maintenance plan evaluation results.

[0072] Specifically, based on the three-dimensional fault evolution model, the structural topology information, fault state parameters and physical field characteristics of the equipment are extracted, and the constraints of the equipment operation mechanism are integrated to construct a virtual maintenance scenario that includes interactive component models, dynamic fault environment and operation response rules. This scenario needs to accurately map the spatial relationship and fault evolution law of the physical equipment.

[0073] Multiple maintenance solutions are structurally decomposed and analyzed into specific operation units. These operation units are then transformed into execution instructions that can be recognized by the virtual scene. The interaction logic of the corresponding components in the scene is matched to complete the loading and adaptation of the solutions in the virtual environment.

[0074] For each type of maintenance scheme, a dedicated digital twin is constructed. This twin synchronizes the real-time fault parameters in the three-dimensional fault evolution model. The entire execution process of the scheme is simulated through digital twin simulation technology, dynamically capturing the recovery time trajectory of key equipment parameters, generating a state recovery trend chart, and simultaneously tracking the state anomalies of related nodes caused by the operation, marking the propagation path and impact boundary of secondary faults, and generating a risk chain effect chart.

[0075] Recovery efficiency features (such as parameter compliance period and stable fluctuation range) are extracted from the condition recovery trend diagram, and risk level features (such as secondary failure rate and impact range level) are extracted from the risk chain effect diagram. The two types of features are input into a pre-set multi-dimensional evaluation framework, and feature weight allocation and comprehensive comparison are performed through a multi-criteria decision algorithm to generate quantitative scores and priority rankings for each scheme, and finally generate maintenance scheme evaluation results.

[0076] Furthermore, through digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram, including:

[0077] For each maintenance plan, a virtual execution environment is constructed and processed to generate a unique digital twin of the plan;

[0078] Fault evolution parameters are injected into the solution-specific digital twin, and these parameters are derived from the three-dimensional fault evolution model.

[0079] By using a dynamic propagation algorithm, the cascading effect simulation of state changes in associated nodes of the device is performed to obtain the cascading effect simulation results;

[0080] Based on the simulation results of chain effects, multidimensional state trajectory reconstruction is performed to generate state recovery trend diagrams and risk chain effect diagrams.

[0081] Specifically, for each maintenance solution, the operational steps, equipment components, and execution constraints are analyzed. Based on the physical structure model of the equipment and the maintenance process logic, a virtual execution environment is constructed that includes interactive operation interfaces and dynamic response mechanisms for components. This generates a solution-specific digital twin that can accurately map the execution scenario of the solution.

[0082] Key fault evolution parameters such as fault location, degradation degree, propagation rate, and impact range are extracted from the three-dimensional fault evolution model. Based on the spatial correlation and signal transmission relationship of equipment components, these parameters are mapped and injected into the corresponding modules of the solution-specific digital twin, so that the initial state of the twin is consistent with the actual fault state.

[0083] The dynamic propagation algorithm is used. Based on the association rules of electrical connection, mechanical coupling, heat conduction and other relationships between nodes of the equipment, the algorithm simulates the parameter fluctuation, state transition and anomaly transmission process of the associated nodes during the maintenance operation and the natural evolution of the fault. It tracks the state change data of each node from the initial state to each subsequent moment and generates chain effect simulation results.

[0084] Multidimensional state trajectory reconstruction is performed on the simulation results of chain effect. The change trajectory of the core performance parameters of the equipment over time is extracted from the results. A state recovery trend diagram is generated after time series processing. At the same time, the occurrence time sequence, propagation path and impact level information of abnormal state nodes are extracted. A risk chain effect diagram is generated through spatial correlation analysis.

[0085] Furthermore, the recovery trend chart and risk cascading effect chart are compared and evaluated to generate maintenance plan evaluation results, including:

[0086] The recovery efficiency feature is extracted from the state recovery trend map to generate a recovery efficiency feature vector.

[0087] The risk propagation feature vector is generated by performing deep feature extraction on the risk chain effect diagram.

[0088] Input the recovery efficiency feature vector and the risk propagation feature vector into the preset multidimensional evaluation model to generate the safety evaluation value, cost evaluation value and timeliness evaluation value of each maintenance plan;

[0089] Based on safety assessment values, cost assessment values, and timeliness assessment values, a comprehensive priority ranking process is performed to generate maintenance plan evaluation results.

[0090] Specifically, for the state recovery trend chart, the key indicators it contains, such as the regression rate, fluctuation amplitude, stabilization time and target achievement, are analyzed by the feature extraction algorithm. The above indicators are then structured and integrated according to preset dimensions to generate a recovery efficiency feature vector that can quantitatively characterize the ability of equipment performance to recover after maintenance.

[0091] Simultaneously, deep feature mining is performed on the risk chain effect diagram to extract core elements such as the scope of impact nodes, propagation speed, impact level, and probability of occurrence. These elements are then reorganized according to the dimensions of the risk assessment system to generate a risk propagation feature vector that can quantify the degree of risk.

[0092] The two types of feature vectors mentioned above are input into a pre-set multidimensional evaluation model, which includes a safety evaluation unit, a cost evaluation unit, and a timeliness evaluation unit. The safety evaluation unit calculates a safety evaluation value that reflects the risk control capability of the scheme based on the risk propagation feature vector. The cost evaluation unit generates a cost evaluation value by combining the resource consumption information implicit in the recovery efficiency feature. The timeliness evaluation unit outputs a timeliness evaluation value based on the time dimension parameter in the recovery efficiency feature.

[0093] A pre-defined weighting system is introduced to weight and integrate the safety assessment value, cost assessment value, and timeliness assessment value. A multi-attribute decision algorithm is used for comprehensive calculation to determine the comprehensive score of each maintenance plan and prioritize them. The result is a maintenance plan evaluation that includes a ranking of the plan’s merits, a comprehensive performance score, and an analysis of key influencing factors.

[0094] like Figure 2 As shown, the three-dimensional fault evolution model is replayed and analyzed using an interactive timeline to generate a fault root cause report, including:

[0095] S201: Based on the three-dimensional fault evolution model, perform spatiotemporal decoupling processing to generate the equipment state time series and spatial structure topology;

[0096] S202: Based on the device status time series, construct a non-linear jumpable time axis interface and configure dynamic playback control function;

[0097] S203: Based on spatial structure topology and dynamic playback control functions, key fault node hierarchical marking is performed through multi-scale fault propagation analysis algorithm to generate hierarchical marking results;

[0098] S204: Based on the hierarchical labeling results, perform reverse reconstruction and forward verification of the fault causal chain to generate an initial root cause analysis report;

[0099] S205: Based on the initial root cause analysis report, the reliability verification process is carried out by integrating the physical mechanism constraints of the equipment, and a fault root cause report is generated.

[0100] Specifically, spatiotemporal decoupling is performed on the three-dimensional fault evolution model. By separating the time and space dimensions in the model, the operating parameters and status indicators of each equipment component at different times are extracted and integrated to generate a time series of equipment status. At the same time, the physical connection relationships, location distribution, and signal transmission paths of the components in the model are analyzed to construct a spatial structure topology, so as to clearly present the spatial relationship logic of each node of the equipment.

[0101] Based on the device status time series, key time nodes are identified, and non-linear jumpable time axis interfaces are constructed using these as anchor points. The interface supports displaying status changes at different time granularities and is configured with dynamic playback control functions, including playback speed adjustment, fixed-point viewing at specific times, and segmented playback of fault stages, so as to achieve flexible tracing of the fault evolution process.

[0102] By combining the correlation strength of components in the spatial structure topology with the state change trajectory captured by the dynamic playback control function, a multi-scale fault propagation analysis algorithm is used to analyze the fault propagation path layer by layer from three scales: equipment system level, component level, and element level. Key fault nodes are marked according to the hierarchy of "root node - first-level propagation node - second-level propagation node - affected node", generating a hierarchical marking result containing node identifier, hierarchical attributes, and correlation weight.

[0103] Based on the hierarchical labeling results, the fault causal chain is reconstructed in reverse. Starting from the final fault state node, the upstream related nodes are traced back along the propagation path to screen out potential initial fault sources. Then, simulated fault signals are injected into the potential sources, and forward propagation simulation is used to verify whether they can reproduce the fault evolution trajectory in the hierarchical labeling results. The verified source information and propagation path are integrated to generate an initial root cause analysis report.

[0104] Based on the physical constraints of the equipment (such as Ohm's law of circuits, heat conduction equations, mechanical coupling principles, etc.), the fault source and propagation logic in the initial root cause analysis report are verified, assumptions that do not conform to physical laws are eliminated, explanations at the mechanism level are added, and a fault root cause report containing the fault origin location, triggering conditions, complete propagation path and mechanism verification conclusions is generated.

[0105] Furthermore, based on the hierarchical labeling results, reverse reconstruction and forward verification of the fault causal chain are performed to generate an initial root cause analysis report, including:

[0106] Based on the hierarchical labeling results, a reverse traversal of the fault propagation tree is performed to generate an initial set of fault source points.

[0107] The initial set of fault source points is filtered by physical mechanism constraints to generate candidate fault source points;

[0108] Based on candidate fault sources, simulated fault signals are injected for forward propagation verification processing to generate fault reproduction data;

[0109] The fault reproduction data and the hierarchical labeling results are compared for spatiotemporal consistency to generate a verification matching degree.

[0110] When the verification match meets the preset threshold, an initial root cause analysis report is generated.

[0111] Specifically, based on the hierarchical attributes and association weights of each node in the hierarchical labeling results, a fault propagation tree structure is constructed. Starting from the affected node at the highest level, the direct upstream associated nodes are traced back layer by layer along the propagation path. The triggering sequence and association strength of each node are recorded until the initial node with no upstream node is traced back. The above nodes are then integrated to generate an initial fault source point set.

[0112] Based on the physical constraints of the equipment, the initial set of fault source points is filtered, eliminating nodes that cannot logically trigger subsequent fault propagation and retaining nodes that conform to the mechanism as candidate fault source points. For each candidate fault source point, its typical fault characteristic signals (such as abnormal voltage, temperature jumps, vibration frequency shifts, etc.) are simulated and injected into the fault propagation model. The forward propagation algorithm is used to simulate the transmission process of signals between associated nodes of the equipment, recording the state change sequence of each node, the fault propagation path, and the scope of influence, generating fault reproduction data.

[0113] The fault reproduction data and hierarchical labeling results are compared for spatiotemporal consistency. The temporal dimension verifies the matching of fault occurrence sequence and duration, while the spatial dimension verifies the consistency of propagation path and node hierarchy. The overlap ratio and logical consistency of the two are comprehensively calculated to generate a verification matching score. When the verification matching score reaches a preset threshold, candidate fault source information, forward verification trajectories, and matching score analysis results are integrated to generate an initial root cause analysis report containing the initial fault source, core propagation link, and verification basis.

[0114] Furthermore, based on equipment operation data and historical fault data, fault scene reconstruction processing is performed to generate a three-dimensional fault evolution model, including:

[0115] Perform timestamp synchronization and spatial coordinate registration on equipment operation data and historical fault data to generate a spatiotemporal fusion dataset.

[0116] Based on the spatiotemporal fusion dataset, multidimensional feature parameter extraction is performed to generate a fault feature matrix;

[0117] The fault feature matrix is ​​loaded into a pre-set three-dimensional structural model through a physical field coupling mapping mechanism to generate an initial dynamic scene.

[0118] Based on the physical laws of electrical equipment, the evolution trajectory of the initial dynamic scene is iteratively corrected to generate a three-dimensional fault evolution model.

[0119] Specifically, timestamp synchronization is performed on equipment operation data and historical fault data. Time deviations from different data sources are calibrated using a unified time base to eliminate timing misalignments. At the same time, spatial coordinate registration is performed to map each data item to the three-dimensional coordinates of equipment components, ensuring a one-to-one correspondence between data and physical location, and integrating them into a spatiotemporal fusion dataset containing the three dimensions of "time-space-state".

[0120] Based on the above dataset, multidimensional feature parameters are extracted, fault-sensitive indicators are parsed from the operational data, fault feature labels are extracted from historical fault data, and the above parameters are arranged in a structured manner according to the time series and spatial partition dimensions to generate a fault feature matrix with rows corresponding to time nodes, columns corresponding to spatial locations and feature types.

[0121] Based on the physical field coupling mapping mechanism, the eigenvalues ​​in the fault feature matrix are transformed into corresponding physical field parameters and loaded into the preset three-dimensional structural model of the equipment. This allows each component of the model to change dynamically according to the physical field parameters, generating an initial dynamic scene that can intuitively display the initial evolution state of the fault.

[0122] Based on the constraints of the physical laws of electrical equipment, the evolution trajectory of the initial dynamic scenario is iteratively corrected: by verifying the state changes at each moment in the scenario, identifying trajectory deviations that conflict with physical laws, adjusting the feature parameter mapping relationship and evolution rate, and iterating through multiple rounds until the scenario fully conforms to the equipment operation mechanism, a three-dimensional fault evolution model that can dynamically reproduce the entire process of fault occurrence and propagation is generated.

[0123] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0124] In one embodiment, such as Figure 3 As shown, this application also provides an intelligent fault early warning device 300 for electrical equipment based on edge computing, the device 300 comprising:

[0125] Data acquisition module 301 is used to collect device operation data and historical fault data through edge computing nodes;

[0126] The scene reconstruction module 302 is used to perform fault scene reconstruction processing based on equipment operation data and historical fault data, and generate a three-dimensional fault evolution model.

[0127] The root cause analysis module 303 is used to perform playback analysis on the three-dimensional fault evolution model through an interactive timeline and generate a fault root cause report.

[0128] The strategy deduction module 304 is used to perform maintenance strategy simulation and deduction based on the three-dimensional fault evolution model, and generate maintenance plan evaluation results.

[0129] Specifically, the data acquisition module 301, based on the monitoring needs of electrical equipment, deploys edge computing nodes on key components and configures suitable sensor interfaces to access real-time operating signals such as voltage, temperature, and vibration. Simultaneously, it retrieves historical fault records from the equipment management system via a data interface, including fault type, occurrence time, and associated components. The acquired real-time data undergoes timestamp calibration and noise filtering, and the time dimension of historical fault data is synchronized with the equipment identifier to ensure spatiotemporal consistency between the two types of data. Finally, it outputs equipment operating data and historical fault data.

[0130] The scene reconstruction module 302, based on equipment operation data and historical fault data, performs timestamp synchronization and spatial coordinate registration, mapping the data to the equipment's three-dimensional coordinates to generate a spatiotemporal fusion dataset. It extracts multi-dimensional fault feature parameters such as temperature gradient and vibration frequency to generate a fault feature matrix. Through a physical field coupling mapping mechanism, the feature matrix is ​​loaded onto a pre-set three-dimensional structural model to generate an initial dynamic scene. Based on the constraints of the physical laws governing electrical equipment, the evolution trajectory is iteratively corrected, outputting a three-dimensional fault evolution model.

[0131] The root cause analysis module 303, based on a three-dimensional fault evolution model, performs spatiotemporal decoupling to generate a time series of equipment states and a spatial structural topology. A nonlinear, jumpable time axis interface is constructed based on the time series, and dynamic playback control functionality is configured. Combining the spatial structural topology and playback function, a multi-scale fault propagation analysis algorithm is used to hierarchically mark key fault nodes, generating hierarchical marking results. Based on the marking results, the fault causal chain is reconstructed in reverse and forward verified, generating an initial root cause analysis report. The reliability is verified by integrating equipment physical mechanism constraints, and a fault root cause report is output.

[0132] The strategy simulation module 304, based on a three-dimensional fault evolution model, constructs a virtual maintenance scenario that includes equipment structure and connection relationships. It loads multiple maintenance schemes broken down into specific operational units, adapting to the scenario's interaction logic. Through digital twin simulation technology, it constructs a dedicated twin for each scheme and injects fault evolution parameters, simulating and generating state recovery trend diagrams and risk chain effect diagrams. It extracts the recovery efficiency and risk propagation feature vectors from these two types of diagrams, inputs them into a multi-dimensional evaluation model to generate safety, cost, and timeliness evaluation values, and outputs the maintenance scheme evaluation results after comprehensive ranking.

[0133] Strategy deduction module 304 is also used for:

[0134] A virtual maintenance scenario is constructed based on a three-dimensional fault evolution model;

[0135] Load multiple repair solutions into a virtual repair scenario;

[0136] Using digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram.

[0137] The recovery trend chart and the risk chain effect chart are compared and evaluated to generate the maintenance plan evaluation results.

[0138] Strategy deduction module 304 is also used for:

[0139] For each maintenance plan, a virtual execution environment is constructed and processed to generate a unique digital twin of the plan;

[0140] Fault evolution parameters are injected into the solution-specific digital twin, and these parameters are derived from the three-dimensional fault evolution model.

[0141] By using a dynamic propagation algorithm, the cascading effect simulation of state changes in associated nodes of the device is performed to obtain the cascading effect simulation results;

[0142] Based on the simulation results of chain effects, multidimensional state trajectory reconstruction is performed to generate state recovery trend diagrams and risk chain effect diagrams.

[0143] Strategy deduction module 304 is also used for:

[0144] The recovery efficiency feature is extracted from the state recovery trend map to generate a recovery efficiency feature vector.

[0145] The risk propagation feature vector is generated by performing deep feature extraction on the risk chain effect diagram.

[0146] Input the recovery efficiency feature vector and the risk propagation feature vector into the preset multidimensional evaluation model to generate the safety evaluation value, cost evaluation value and timeliness evaluation value of each maintenance plan;

[0147] Based on safety assessment values, cost assessment values, and timeliness assessment values, a comprehensive priority ranking process is performed to generate maintenance plan evaluation results.

[0148] Root cause analysis module 303 is also used for:

[0149] Based on the three-dimensional fault evolution model, spatiotemporal decoupling is performed to generate the equipment state time series and spatial structure topology.

[0150] Based on the device status time series, a non-linear jumpable time axis interface is constructed and a dynamic playback control function is configured.

[0151] Based on spatial structure topology and dynamic playback control, a multi-scale fault propagation analysis algorithm is used to perform hierarchical marking of key fault nodes and generate hierarchical marking results.

[0152] Based on the hierarchical labeling results, reverse reconstruction and forward verification of the fault causal chain are performed to generate an initial root cause analysis report.

[0153] Based on the initial root cause analysis report, the reliability verification process is carried out by integrating the physical mechanism constraints of the equipment, and a fault root cause report is generated.

[0154] Root cause analysis module 303 is also used for:

[0155] Based on the hierarchical labeling results, a reverse traversal of the fault propagation tree is performed to generate an initial set of fault source points.

[0156] The initial set of fault source points is filtered by physical mechanism constraints to generate candidate fault source points;

[0157] Based on candidate fault sources, simulated fault signals are injected for forward propagation verification processing to generate fault reproduction data;

[0158] The fault reproduction data and the hierarchical labeling results are compared for spatiotemporal consistency to generate a verification matching degree.

[0159] When the verification match meets the preset threshold, an initial root cause analysis report is generated.

[0160] Scene reconstruction module 302 is also used for:

[0161] Perform timestamp synchronization and spatial coordinate registration on equipment operation data and historical fault data to generate a spatiotemporal fusion dataset.

[0162] Based on the spatiotemporal fusion dataset, multidimensional feature parameter extraction is performed to generate a fault feature matrix;

[0163] The fault feature matrix is ​​loaded into a pre-set three-dimensional structural model through a physical field coupling mapping mechanism to generate an initial dynamic scene.

[0164] Based on the physical laws of electrical equipment, the evolution trajectory of the initial dynamic scene is iteratively corrected to generate a three-dimensional fault evolution model.

[0165] In one embodiment, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0166] In one embodiment, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0167] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0168] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for intelligent fault early warning of electrical equipment based on edge computing, characterized in that, The method includes: Collect device operation data and historical fault data through edge computing nodes; Based on the equipment operation data and historical fault data, fault scene reconstruction processing is performed to generate a three-dimensional fault evolution model. The three-dimensional fault evolution model is replayed and analyzed using an interactive timeline to generate a fault root cause report. Based on the three-dimensional fault evolution model, maintenance strategy simulation and deduction are performed to generate maintenance plan evaluation results.

2. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 1, characterized in that, The process of simulating and extrapolating maintenance strategies based on the three-dimensional fault evolution model to generate maintenance plan evaluation results includes: Based on the aforementioned three-dimensional fault evolution model, a virtual maintenance scenario is constructed; Multiple repair solutions are loaded into the virtual repair scenario; Using digital twin simulation technology, the effects of each maintenance plan are predicted, generating a state recovery trend diagram and a risk chain effect diagram. The recovery trend chart and risk chain effect chart are compared and evaluated to generate the maintenance plan evaluation results.

3. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 2, characterized in that, The process of using digital twin simulation technology to predict the effects of each maintenance plan and generate a state recovery trend diagram and a risk chain effect diagram includes: For each maintenance plan, a virtual execution environment is constructed and processed to generate a unique digital twin of the plan; Fault evolution parameters are injected into the digital twin of the solution, and the fault evolution parameters are derived from the three-dimensional fault evolution model; By using a dynamic propagation algorithm, the cascading effect simulation of state changes in associated nodes of the device is performed to obtain the cascading effect simulation results; Based on the simulation results of the chain effect, multidimensional state trajectory reconstruction processing is performed to generate the state recovery trend map and the risk chain effect map.

4. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 2, characterized in that, The process of comparing and evaluating the state recovery trend diagram and the risk chain effect diagram to generate the maintenance plan evaluation result includes: The recovery efficiency feature is extracted from the state recovery trend graph to generate a recovery efficiency feature vector; The risk propagation deep feature extraction process is performed on the risk chain effect diagram to generate a risk propagation feature vector; The recovery efficiency feature vector and risk propagation feature vector are input into a preset multidimensional evaluation model to generate a safety evaluation value, a cost evaluation value, and a timeliness evaluation value for each maintenance plan. Based on the safety assessment value, cost assessment value, and timeliness assessment value, a comprehensive priority ranking process is performed to generate the maintenance plan assessment result.

5. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 1, characterized in that, The step of replaying and analyzing the three-dimensional fault evolution model through an interactive timeline to generate a fault root cause report includes: Based on the aforementioned three-dimensional fault evolution model, spatiotemporal decoupling processing is performed to generate the equipment state time series and spatial structure topology. Based on the device status time series, a nonlinear jumpable time axis interface is constructed, and a dynamic playback control function is configured. Based on the spatial structure topology and the dynamic playback control function, a multi-scale fault propagation analysis algorithm is used to perform hierarchical marking of key fault nodes and generate hierarchical marking results. Based on the hierarchical labeling results, reverse reconstruction and forward verification of the fault causal chain are performed to generate an initial root cause analysis report. Based on the initial root cause analysis report, the reliability verification process is performed by integrating the physical mechanism constraints of the equipment to generate the fault root cause report.

6. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 5, characterized in that, Based on the hierarchical labeling results, the process involves reverse reconstruction and forward verification of the fault causal chain to generate an initial root cause analysis report, including: Based on the hierarchical labeling results, a reverse traversal of the fault propagation tree is performed to generate an initial set of fault source points. The initial set of fault sources is subjected to physical mechanism constraint filtering to generate candidate fault sources; Based on the candidate fault sources, simulated fault signals are injected for forward propagation verification processing to generate fault reproduction data. The fault reproduction data and the hierarchical labeling results are compared for spatiotemporal consistency to generate a verification matching degree. When the verification matching degree meets the preset threshold, the initial root cause analysis report is generated.

7. The intelligent fault early warning method for electrical equipment based on edge computing according to claim 1, characterized in that, The process of reconstructing a fault scene based on the equipment's operating data and historical fault data to generate a three-dimensional fault evolution model includes: The device operation data and historical fault data are time-stamped and spatially registered to generate a spatiotemporal fusion dataset. Based on the spatiotemporal fusion dataset, a multidimensional feature parameter extraction operation is performed to generate a fault feature matrix; The fault feature matrix is ​​loaded into a pre-set three-dimensional structural model through a physical field coupling mapping mechanism to generate an initial dynamic scene. Based on the physical constraints of electrical equipment, the initial dynamic scene is subjected to iterative correction of the evolution trajectory to generate the three-dimensional fault evolution model.

8. An intelligent fault early warning device for electrical equipment based on edge computing, characterized in that, The device includes: The data acquisition module is used to collect device operation data and historical fault data through edge computing nodes; The scene reconstruction module is used to perform fault scene reconstruction processing based on the equipment operation data and historical fault data, and generate a three-dimensional fault evolution model. The root cause analysis module is used to perform playback analysis on the three-dimensional fault evolution model through an interactive timeline and generate a fault root cause report. The strategy deduction module is used to perform maintenance strategy simulation and deduction based on the three-dimensional fault evolution model, and generate maintenance plan evaluation results.

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 intelligent fault early warning method for electrical equipment based on edge computing as described in any one of claims 1 to 7.

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 intelligent fault early warning method for electrical equipment based on edge computing as described in any one of claims 1 to 7.

Images