Multi-modal digital twin operation and maintenance management and control method, system and equipment for power monitoring system and storage medium

Abstract

The invention discloses a multi-mode digital twin operation and maintenance management and control method, system and device for a power monitoring system and a storage medium, and relates to the technical field of power monitoring systems, and the method comprises the steps: collecting various physical quantity data of power equipment in real time, and carrying out the preliminary format unification and timestamp marking; protocol analysis, space-time alignment and dynamic weighted fusion are carried out based on the collected multi-modal data, and a unified space-time data set is generated; based on the unified spatio-temporal data set, the state of the power equipment under the thermal-mechanical-electric multi-field coupling effect is simulated in real time through a multi-physics field simulation engine based on a finite element analysis model, fault evolution prediction is carried out, a specific operation and maintenance strategy and equipment health assessment are generated, and therefore an external execution unit is automatically triggered; according to the method, multi-modal data accurate fusion and fault dynamic deduction can be realized, the prediction accuracy is remarkably improved, an intelligent operation and maintenance closed loop is formed, and the operation intuition and the system efficiency can be enhanced.

Description

Technical Field

[0001] This invention relates to the field of power monitoring system technology, and in particular to a multimodal digital twin operation and maintenance management method, system, equipment and storage medium for power monitoring systems. Background Technology

[0002] Power monitoring systems are the core carriers for ensuring the safe operation of the power grid. With the accelerated construction of new power systems, the high proportion of new energy sources being integrated, and the surge in equipment complexity, traditional monitoring modes that rely on independent platforms such as SCADA (Supervisory Control and Data Acquisition) and EMS (Energy Management System) are no longer sufficient to meet the needs of real-time perception and dynamic control.

[0003] Digital twin technology, as a virtual mapping of the physical power grid, theoretically enables panoramic monitoring and predictive maintenance of equipment status and power grid operation by integrating IoT, AI, and multiphysics simulation. However, current technological maturity still faces significant bottlenecks: Inconsistent data formats and protocols across subsystems of the power monitoring system, such as SCADA, BIM (Building Information Model), and video inspection, create "information silos." Fault localization requires manual data retrieval across multiple platforms, averaging over 2 hours, and the false alarm rate is as high as 35% due to missing data fusion, failing to support cross-device collaborative analysis. Existing digital twin models often focus on a single dimension, such as geometric or electrical parameters, lacking dynamic coupling capabilities across thermal, mechanical, and electrical multiphysics fields. Mainstream simulation engines have refresh rates ≤10Hz, making it difficult to capture microsecond-level transient events, such as arc faults, and 90% of platforms cannot efficiently process unstructured data such as video streams and point clouds, resulting in fault prediction error rates exceeding 15%. Most platforms only achieve a one-way "monitoring-early warning" link, lacking a "decision-execution" intelligent closed loop. For example, after an equipment overheating alarm, manual repair is required, resulting in a response delay of over one hour. Furthermore, core components rely on foreign 3D engines such as Unity and databases like Oracle, posing risks of service outages and data security vulnerabilities. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention provides a multimodal digital twin operation and maintenance management method, system, equipment and storage medium for power monitoring systems.

[0005] Therefore, the technical problem solved by this invention is the problem of slow fault location, high prediction error and long response delay in power monitoring systems caused by fragmented multi-source data, insufficient model accuracy and non-closed-loop operation and maintenance.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a multimodal digital twin operation and maintenance management method for a power monitoring system, comprising: Based on distributed edge nodes, various IoT protocols are used to collect multiple physical quantity data of power equipment in real time, and preliminary format unification and timestamp marking are performed. Based on the collected multimodal data, protocol parsing, spatiotemporal alignment, and dynamic weighted fusion are performed to generate a unified spatiotemporal dataset; Based on a unified spatiotemporal dataset, a multiphysics simulation engine is used to simulate the state of power equipment under the coupling of multiple thermo-mechanical-electric fields in real time using a finite element analysis model, and to predict fault evolution. Based on the analysis of simulation and deduction results, specific operation and maintenance strategies and equipment health assessments are generated, thereby automatically triggering external execution units.

[0007] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The process of generating a unified spatiotemporal dataset by performing protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the collected multimodal data includes: Acquire multimodal monitoring data of power equipment collected by distributed edge nodes; The multimodal monitoring data is processed by protocol parsing, spatiotemporal alignment, and feature extraction. Based on a time-series database, dynamic weighted fusion is performed on processed heterogeneous data. Based on the results of dynamic weighted fusion, a multimodal dataset with a unified spatiotemporal benchmark and a mismatch rate below a preset threshold is generated.

[0008] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The dynamic weighted fusion of processed heterogeneous data includes: The weights for dynamic weighted fusion are determined as follows: the historical reliability indicators of the collected multimodal data sources are used as the main weighting coefficients, and the real-time indicators of the collected multimodal data itself are used as auxiliary weighting coefficients. The main weighting coefficients and auxiliary weighting coefficients are linearly combined to calculate the final fusion weights.

[0009] The beneficial effects of this preferred technical solution are as follows: by establishing a dual-weight evaluation mechanism based on historical reliability and real-time performance, it takes into account both the long-term stability of the data source and the timeliness value of the data, making the data fusion results more accurate and reliable, effectively reducing the mismatch rate, and providing a high-quality data foundation for subsequent simulation analysis.

[0010] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The method, based on a unified spatiotemporal dataset, uses a multiphysics simulation engine and a finite element analysis model to simulate the state of power equipment under the coupling of thermo-mechanical-electrical fields in real time, and performs fault evolution prediction, including: Based on a unified spatiotemporal dataset, a dynamic digital twin with coupled thermo-mechanical-electrical parameters is constructed using a multiphysics simulation engine. A numerical analysis model based on partial differential equations is used to solve the multiphysics coupling effect and simulate the state of the digital twin in real time. Based on real-time simulation results, the evolution of equipment failures is simulated at a frequency no lower than the predetermined refresh rate.

[0011] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The method of using a numerical analysis model based on partial differential equations to perform real-time solution and state simulation of multiphysics coupling effects on the digital twin includes: Establish the governing equations that include thermal energy storage, thermal conduction and diffusion, electromagnetic loss source, and force-thermal coupling terms; The governing equations enable bidirectional coupling calculations between thermal, mechanical, and electromagnetic fields.

[0012] The beneficial effects of this preferred technical solution are as follows: by establishing control equations for bidirectional coupling of multi-physics fields, accurate simulation of complex operating conditions of power equipment is achieved, which can truly reflect the changes in equipment state under the interaction of multiple fields of thermo-mechanical-electricity, and significantly improve the accuracy and reliability of fault prediction.

[0013] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The process of analyzing simulation and deduction results to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units, includes: Obtain device state simulation and fault inference results from the multiphysics simulation engine; Based on simulation and deduction results, an intelligent decision engine uses artificial intelligence models to perform causal reasoning and pattern recognition, generating comprehensive decision results that include operation and maintenance strategies, equipment health scores, and fault predictions.

[0014] As a preferred solution for a multimodal digital twin operation and maintenance management method for power monitoring systems, the following is provided: The process of analyzing simulation and deduction results to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units, also includes: The comprehensive decision-making results are converted into executable instructions, and at least one external execution unit is automatically triggered by the operation and maintenance execution linkage module to complete the inspection, material delivery or work order dispatch tasks. The visualization and interaction module integrates and renders equipment status, fault simulation, early warning information, and operation and maintenance progress, and outputs them to the interactive terminal.

[0015] The beneficial effects of this preferred technical solution are as follows: by establishing a complete closed loop from decision analysis to execution feedback, the automation and intelligence of the operation and maintenance process are realized, which greatly improves the efficiency of operation and maintenance; at the same time, through the integrated visualization of multi-dimensional information, it provides operation and maintenance personnel with intuitive and comprehensive situational awareness, and enhances the effectiveness of human-machine collaboration.

[0016] Secondly, this invention provides a multimodal digital twin operation and maintenance management system for power monitoring systems, comprising: The multi-source data acquisition module is used to collect various physical quantity data of power equipment in real time through various IoT protocols based on distributed edge nodes, and to perform preliminary format unification and timestamp marking. The multi-source data acquisition module is used to perform protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the acquired multimodal data to generate a unified spatiotemporal dataset. A multiphysics simulation engine is used to simulate the state of power equipment under the coupling of thermo-mechanical-electrical fields in real time based on a unified spatiotemporal dataset and finite element analysis model, and to predict fault evolution. The intelligent decision engine, operation and maintenance execution linkage module, and visualization interaction module are used to analyze simulation and deduction results, generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units.

[0017] Thirdly, the present invention provides a computer device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the multimodal digital twin operation and maintenance management method for power monitoring system are implemented.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of a multimodal digital twin operation and maintenance management method for a power monitoring system.

[0019] The beneficial effects of this invention are as follows: This invention collects multimodal data such as temperature, current, acoustic signature, and 4K video in real time through distributed edge nodes, performs protocol parsing, spatiotemporal alignment, and dynamic weighted fusion to generate a unified spatiotemporal dataset with a mismatch rate of <3%, completely eliminating the problem of data fragmentation across multiple systems and providing a precise data foundation for subsequent analysis; a multiphysics simulation engine employs a finite element analysis model to calculate the thermo-mechanical-electric coupling effect in real time, supporting dynamic fault simulation at refresh rates ≥60Hz; and an intelligent decision engine generates operation and maintenance strategies through an LSTM-GRU hybrid neural network, significantly improving fault prediction and equipment health assessment. The platform ensures accurate estimation; it automatically triggers drone inspections, unmanned vehicle spare parts delivery, and work order dispatch, and interfaces with the warehouse management system API to achieve automatic spare parts scheduling after a fault, forming an intelligent decision-making-execution closed loop; the security encryption module adopts national cryptographic algorithms and quantum key distribution units to ensure that the data transmission process is protected against man-in-the-middle attacks; it supports the overlay display of building information models and real-time video streams, and realizes 3D model operation through gesture recognition, enhancing the intuitiveness of operation and maintenance; the platform is deployed on a cloud-edge collaborative architecture, with the edge side completing 80% of real-time data processing, while the cloud focuses on historical data mining, realizing the rational allocation and efficient utilization of computing resources. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is an overall flowchart of a multimodal digital twin operation and maintenance management method for a power monitoring system provided by the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0023] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a multimodal digital twin operation and maintenance management method for a power monitoring system, including: S1: Based on distributed edge nodes, various IoT protocols are used to collect multiple physical quantity data of power equipment in real time, and preliminary format unification and timestamp marking are performed; S2: Based on the collected multimodal data, perform protocol parsing, spatiotemporal alignment, and dynamic weighted fusion to generate a unified spatiotemporal dataset; S3: Based on a unified spatiotemporal dataset, a multiphysics simulation engine is used to simulate the state of power equipment under the coupling of multiple fields of thermo-mechanical-electricity in real time, and to predict the evolution of faults. S4: Based on the simulation and deduction results, analyze the data to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units.

[0024] It should be noted that through steps S1-S4, a closed-loop system of "data acquisition-fusion analysis-simulation and deduction-decision execution" is constructed, which seamlessly connects the real-time status, dynamic evolution and intelligent response of physical equipment, thereby realizing accurate mapping, advanced prediction and autonomous control of the power system operation and maintenance process in the digital space.

[0025] Example 2, refer to Figure 1 As an embodiment of the present invention, based on the previous embodiment, a multimodal digital twin operation and maintenance management method for a power monitoring system is provided, comprising: In this embodiment, step S1 above, which involves collecting various physical quantity data of power equipment in real time based on distributed edge nodes and various IoT protocols, and performing preliminary format unification and timestamp marking, includes: Based on distributed edge nodes, various physical data of power equipment are collected in real time through IoT protocols, including: operation data, infrared thermal images, vibration sensor data, video inspection streams, and environmental parameters; among them, the distributed edge nodes support simultaneous access to temperature sensors, current transformers, acoustic fingerprint detectors, and 4K video streaming devices, with a data acquisition latency of no more than 50 milliseconds.

[0026] In this embodiment, each node is preferably configured with an Intel Atom C3758R processor and connects to a temperature sensor (PT100), a current transformer (0.2S class), a voiceprint detector (frequency response 20Hz-20kHz), and a 4K video streaming device (H.265 encoding) via the OPCUA / Modbus protocol.

[0027] In another possible implementation, real-time data acquisition can also be carried out by deploying intelligent sensing terminals with edge computing capabilities at key nodes of power equipment. These terminals integrate multiple data acquisition interfaces, are capable of performing local AD conversion and filtering on analog signals, and directly upload data packets to the edge gateway through 5G URLLC network slicing technology, ensuring that end-to-end transmission latency requirements as low as 10 milliseconds can still be met even in complex electromagnetic environments.

[0028] In another possible implementation, real-time acquisition can also be carried out by building an edge data acquisition bus based on a time-sensitive network. This bus allocates dedicated transmission channels for high-bandwidth data such as video streams and vibration waveforms, and sets the highest transmission priority for key state parameters such as current and temperature. Microsecond-level time synchronization between acquisition nodes is achieved through the IEEE 802.1AS protocol, thereby ensuring that all data have a unified and accurate time scale.

[0029] In this embodiment, step S2 above, which involves protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the collected multimodal data to generate a unified spatiotemporal dataset, includes: The data fusion processing module performs protocol parsing, spatiotemporal alignment, and feature extraction on the collected multimodal data to generate a multimodal dataset with a unified spatiotemporal benchmark.

[0030] The data fusion processing module has a built-in time-series database that performs dynamic weighted fusion of unstructured point cloud data and structured operational data, with a mismatch rate of less than three percent.

[0031] In this embodiment, the data fusion processing module preferably uses the time-series database InfluxDBv2.0 to store the data, and the dynamic weighted fusion formula is as follows: in, The weights are assigned to the i-th data type (e.g., temperature weight 0.6, video stream weight 0.3). For the historical error standard deviation of the data source, =e^(-0.02t) (t is the delay in milliseconds), coefficients α=0.7, β=0.3 (optimized through training with historical data).

[0032] In another possible implementation, dynamic weighted fusion can also be implemented by introducing an adaptive Kalman filter algorithm, which adjusts the observation noise coefficient in real time based on the residual covariance matrix of the sensor's historical data, assigns higher confidence weights to data sources with smaller recent fluctuations, and automatically reduces their weight ratio when abnormal data jumps are detected, thereby achieving dynamic suppression of noise and outliers.

[0033] In another possible implementation, dynamic weighted fusion can also be carried out by constructing a multi-source information confidence assessment model based on DS evidence theory. This model treats the state information provided by different sensors as independent evidence, calculates the basic probability assignment function and uses Dempster's combination rule to synthesize evidence. It is particularly suitable for solving the confidence fusion problem when there is a conflict between unstructured detection data such as infrared thermal imaging and ultraviolet imaging.

[0034] In this embodiment, step S3 above, based on a unified spatiotemporal dataset, uses a multiphysics simulation engine to simulate the state of power equipment under the coupling of thermo-mechanical-electrical multiple fields in real time using a finite element analysis model, and performs fault evolution prediction, including: Based on a unified spatiotemporal dataset, a dynamic digital twin coupled with thermodynamics, mechanical stress, and electrical parameters is constructed through a multiphysics simulation engine, supporting fault simulation at a refresh rate of no less than 60 frames per second. The multiphysics simulation engine uses a finite element analysis model to calculate the coupling effect between the mechanical stress of the transformer winding and the oil temperature diffusion in real time, with an error rate of less than 5%. The core equations of the finite element analysis model of the multiphysics simulation engine are as follows (taking transformer oil temperature diffusion as an example): in, This is the cumulative term for heat capacity per unit volume, and the dominant process is thermal energy storage; This is the heat conduction diffusion term, and the dominant process is Fourier heat conduction. This is an electromagnetic loss source term, and the dominant process is the Joule heating effect; It is a force-heat coupling term, and the dominant process is the conversion of mechanical work into heat.

[0035] In another possible implementation, the multiphysics simulation engine can also be implemented by using a fluid-structure interaction solver based on the material point method. This solver can accurately simulate the fluid dynamics behavior of transformer oil under winding vibration and its impact on heat dissipation efficiency. By coupling Eulerian grids and Lagrange point sets, it can calculate the deformation and aging process of insulating paperboard under the action of multiple electric-thermal-fluid fields in real time.

[0036] In another possible implementation, the multiphysics simulation engine can also be implemented by integrating a real-time computing framework based on a proxy model. This framework pre-calculates offline under different working conditions using high-fidelity finite element analysis and generates a database of reduced-order models. When running online, it queries the reduced-order model that is closest to the current state and interpolates to solve the problem, thereby reducing the computation time of complex coupled fields to the millisecond level while ensuring accuracy.

[0037] In another possible implementation, fault evolution prediction can also be carried out by constructing a prediction model based on a physical information neural network. This model embeds the physical laws describing heat conduction and material fatigue as constraints into the neural network training process. By inputting the current state parameters of the device, it can directly deduce the thermal aging trajectory and remaining life probability distribution of the insulation material over a future period of time.

[0038] In another possible implementation, fault evolution prediction can also be carried out by using a dynamic Bayesian network for probabilistic reasoning. This network transforms the equipment fault mechanism into a conditional probability relationship between nodes. By continuously inputting the latest monitoring evidence, the network node status is updated online, thereby calculating and outputting a spatiotemporal probability map of the development from local overheating to cascading faults such as insulation breakdown in real time.

[0039] In this embodiment, step S4 above involves analyzing simulation and deduction results to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units, including: Based on the output of the multiphysics simulation engine, the intelligent decision engine generates operation and maintenance strategies through the artificial intelligence causal reasoning model, and outputs equipment health scores and fault prediction results.

[0040] The intelligent decision engine integrates a hybrid neural network with long short-term memory gated recurrent units, and generates an optimized spare parts scheduling path based on a historical fault database, with a response latency of less than 10 seconds.

[0041] In this embodiment, the intelligent decision engine preferably uses an LSTM-GRU hybrid neural network structure: input layer (128 dimensions) → LSTM layer (64 units) → GRU layer (32 units) → fully connected layer (Softmax); training data: a fault database from the past 5 years, including 12 types of transformer faults and 8 types of line faults.

[0042] The operation and maintenance execution linkage module receives instructions from the intelligent decision engine and automatically triggers the drone inspection, unmanned vehicle spare parts delivery and work order dispatch system. The operation and maintenance execution linkage module interfaces with the warehouse management system application interface to automatically retrieve spare parts inventory and plan unmanned vehicle delivery routes after fault location.

[0043] It integrates a 3D rendering engine to map the status, early warning information, and operation and maintenance progress of the digital twin to augmented reality, virtual reality terminals, and large screens; it supports the overlay display of building information models and real-time video streams, and enables 3D model scaling and fault point marking through gesture recognition.

[0044] It should be noted that this invention also includes a security encryption module, which uses national cryptographic algorithms to encrypt cross-module data transmission and implements access control based on a domestic operating system. A quantum key distribution unit is deployed at the data bus layer to prevent man-in-the-middle attacks. The security encryption module is present throughout all data transmission links.

[0045] In this embodiment, the SM4 national cryptographic algorithm is preferred for encryption, and the quantum key distribution unit (QKD) updates the 256-bit key every 10 minutes.

[0046] Example 3: The above is an illustrative scheme of a multimodal digital twin operation and maintenance management method for a power monitoring system according to this embodiment. It should be noted that the technical solution of a multimodal digital twin operation and maintenance management system for a power monitoring system and the technical solution of the above-described multimodal digital twin operation and maintenance management method for a power monitoring system belong to the same concept. Details not described in detail in the technical solution of the multimodal digital twin operation and maintenance management system for a power monitoring system in this embodiment can be found in the description of the technical solution of the above-described multimodal digital twin operation and maintenance management method for a power monitoring system.

[0047] This embodiment also provides a multimodal digital twin operation and maintenance management system for power monitoring systems, including: The multi-source data acquisition module is used to collect various physical quantity data of power equipment in real time through various IoT protocols based on distributed edge nodes, and to perform preliminary format unification and timestamp marking. The multi-source data acquisition module is used to perform protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the acquired multimodal data to generate a unified spatiotemporal dataset. A multiphysics simulation engine is used to simulate the state of power equipment under the coupling of thermo-mechanical-electrical fields in real time based on a unified spatiotemporal dataset and finite element analysis model, and to predict fault evolution. The intelligent decision engine, operation and maintenance execution linkage module, and visualization interaction module are used to analyze simulation and deduction results, generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units.

[0048] This embodiment also provides an electronic device applicable to a multimodal digital twin operation and maintenance management method for a power monitoring system, including: The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement a multimodal digital twin operation and maintenance management method for a power monitoring system as proposed in the above embodiments.

[0049] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements a multimodal digital twin operation and maintenance management method for a power monitoring system as proposed in the above embodiments.

[0050] The storage medium proposed in this embodiment belongs to the same inventive concept as the multimodal digital twin operation and maintenance management method for power monitoring system proposed in the above embodiment. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0051] Example 4, referring to Tables 1-2, is an embodiment of the present invention, providing a multimodal digital twin operation and maintenance management method for a power monitoring system. To verify the beneficial effects of the present invention, a simulation experiment is conducted for scientific demonstration.

[0052] Scenario 1: Troubleshooting Transformer Overheating Step 1: Data Acquisition and Fusion Edge node data collected: winding temperature 98℃ (exceeding standard), B-phase vibration acceleration 0.8g (normal <0.3g), infrared thermal imaging showed local hot spots on the bushing (142℃); The data fusion module performs weighted calculations (temperature weight 0.7, vibration weight 0.2, thermal imaging weight 0.1) to generate a unified spatiotemporal dataset, with a mismatch rate of 2.1%. Step 2: Simulation and Decision Making Multiphysics engine simulation of oil temperature diffusion (error rate 3.8%) shows that hot spots will cause insulation failure after 30 seconds. Intelligent decision engine output: Health score: 32 points (warning threshold <60); Operation and maintenance strategy: 1. Immediately reduce the load to 50%; 2. Dispatch a drone for inspection (coordinates: X=35.7, Y=128.6); 3. Retrieve the spare part "Bushing Insulator - Model CZ-203".

[0053] Step 3: Operation and Maintenance Execution Automatically generate work orders and trigger: The drone flew to the location of the malfunction. The inventory status returned by the warehousing system is shown in Table 1: Table 1. Inventory Status Returned by the Warehousing System

[0054] Unmanned vehicle delivery route: Area A → TR-2024-085 (distance 342m, time 2.1 minutes).

[0055] Step 4: Visualization and Security Monitoring AR glasses display: BIM model overlaid with real-time video, gesture recognition zoom and mark fault points; The quantum key distribution unit intercepted a man-in-the-middle attack log digest once.

[0056] Scenario 2: Handling Partial Discharge Faults in Substation GIS Equipment Step 1: Multi-source data acquisition and edge processing Real-time data acquisition from edge nodes (≤50ms latency): Table 2 Multimodal Monitoring Data Acquisition Table

[0057] Edge side: Complete data compression, time stamp alignment, and feature extraction (accounting for 85% of real-time data processing), and generate structured messages to upload to the cloud.

[0058] Step 2: Multiphysics Simulation and Decision Generation Finite element method for solving equations involving coupled electric and thermal fields: in For relative permittivity, For potential distribution, For space charge density, It is the vacuum permittivity; Discharge energy conversion equation: Tensor expansion: Where D is the electric displacement vector and E is the electric field intensity vector. The discharge energy density is given by Q = 3.8 mJ.

[0059] Intelligent decision engine output: { Fault type: "Insulator surface discharge" Health Score: 41 "Operation and Maintenance Instructions":[ {"Action":"Drone close-range scan","Coordinates":"GIS-B-Phase"}, {"Action":"Retrieve Spare Parts","Part Number":"GIM-70-40-00"}, {"Action":"Isolate Phase B gas chamber"} ] } Step 3: Closed-loop operation and maintenance execution Warehouse system interaction code: SELECT stock_locFROMinventory WHEREpart_no='GIM-70-40-00'ANDqty>=1.

[0060] Step 4: Security and Visual Monitoring AR overlay display: Highlight faulty air chambers in the BIM model; Gesture recognition marks the discharge area; Quantum Encrypted Log (weight 8): 2024-06-11 09:17:22 | Data packet encryption: SM4-CTR mode | Key update: QKD serial number.

[0061] In summary, this invention utilizes distributed edge nodes to collect multimodal data such as temperature, current, acoustic signature, and 4K video in real time. A data fusion processing module performs protocol parsing, spatiotemporal alignment, and dynamic weighted fusion to generate a unified spatiotemporal dataset with a mismatch rate of <3%, completely eliminating the problem of data fragmentation across multiple systems and providing a precise data foundation for subsequent analysis. A multiphysics simulation engine employs a finite element analysis model to calculate thermo-mechanical-electrical coupling effects in real time, supporting dynamic fault simulation at refresh rates ≥60Hz. An intelligent decision engine generates maintenance strategies using an LSTM-GRU hybrid neural network, significantly improving the accuracy of fault prediction and equipment health assessment. This is further enhanced by a maintenance execution linkage module. The system automatically triggers drone inspections, unmanned vehicle spare parts delivery, and work order dispatch. It also interfaces with the warehouse management system's API to enable automatic spare parts scheduling after a failure, forming an intelligent decision-making-execution closed loop. The security encryption module uses national cryptographic algorithms and a quantum key distribution unit to ensure that data transmission is protected against man-in-the-middle attacks. Access control is implemented through a domestic operating system. The visual interaction module supports the overlay display of building information models and real-time video streams, and gesture recognition enables 3D model operation, enhancing the intuitiveness of operation and maintenance. The platform is deployed in a cloud-edge collaborative architecture, with 80% of real-time data processing completed at the edge, while the cloud focuses on historical data mining, achieving reasonable allocation and efficient utilization of computing resources.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multimodal digital twin operation and maintenance management method for a power monitoring system, characterized in that, include: Based on distributed edge nodes, various IoT protocols are used to collect multiple physical quantity data of power equipment in real time, and preliminary format unification and timestamp marking are performed. Based on the collected multimodal data, protocol parsing, spatiotemporal alignment, and dynamic weighted fusion are performed to generate a unified spatiotemporal dataset; Based on a unified spatiotemporal dataset, a multiphysics simulation engine is used to simulate the state of power equipment under the coupling of multiple thermo-mechanical-electric fields in real time using a finite element analysis model, and to predict fault evolution. Based on the analysis of simulation and deduction results, specific operation and maintenance strategies and equipment health assessments are generated, thereby automatically triggering external execution units.

2. The multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 1, characterized in that, The process of generating a unified spatiotemporal dataset by performing protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the collected multimodal data includes: Acquire multimodal monitoring data of power equipment collected by distributed edge nodes; The multimodal monitoring data is processed by protocol parsing, spatiotemporal alignment, and feature extraction. Based on a time-series database, dynamic weighted fusion is performed on processed heterogeneous data. Based on the results of dynamic weighted fusion, a multimodal dataset with a unified spatiotemporal benchmark and a mismatch rate below a preset threshold is generated.

3. The multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 2, characterized in that, The dynamic weighted fusion of processed heterogeneous data includes: The weights for dynamic weighted fusion are determined as follows: the historical reliability indicators of the collected multimodal data sources are used as the main weighting coefficients, and the real-time indicators of the collected multimodal data itself are used as auxiliary weighting coefficients. The main weighting coefficients and auxiliary weighting coefficients are linearly combined to calculate the final fusion weights.

4. The multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 3, characterized in that, The method, based on a unified spatiotemporal dataset, uses a multiphysics simulation engine and a finite element analysis model to simulate the state of power equipment under the coupling of thermo-mechanical-electrical fields in real time, and performs fault evolution prediction, including: Based on a unified spatiotemporal dataset, a dynamic digital twin with coupled thermo-mechanical-electrical parameters is constructed using a multiphysics simulation engine. A numerical analysis model based on partial differential equations is used to solve the multiphysics coupling effect and simulate the state of the digital twin in real time. Based on real-time simulation results, the evolution of equipment failures is simulated at a frequency no lower than the predetermined refresh rate.

5. The multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 4, characterized in that, The method of using a numerical analysis model based on partial differential equations to perform real-time solution and state simulation of multiphysics coupling effects on the digital twin includes: Establish the governing equations that include thermal energy storage, thermal conduction and diffusion, electromagnetic loss source, and force-thermal coupling terms; The governing equations enable bidirectional coupling calculations between thermal, mechanical, and electromagnetic fields.

6. The multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 5, characterized in that, The process of analyzing simulation and deduction results to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units, includes: Obtain device state simulation and fault inference results from the multiphysics simulation engine; Based on simulation and deduction results, an intelligent decision engine uses artificial intelligence models to perform causal reasoning and pattern recognition, generating comprehensive decision results that include operation and maintenance strategies, equipment health scores, and fault predictions.

7. A multimodal digital twin operation and maintenance management method for a power monitoring system as described in claim 6, characterized in that, The process of analyzing simulation and deduction results to generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units, also includes: The comprehensive decision-making results are converted into executable instructions, and at least one external execution unit is automatically triggered by the operation and maintenance execution linkage module to complete the inspection, material delivery or work order dispatch tasks. The visualization and interaction module integrates and renders equipment status, fault simulation, early warning information, and operation and maintenance progress, and outputs them to the interactive terminal.

8. A multimodal digital twin operation and maintenance management system for power monitoring systems, using the method described in any one of claims 1 to 7, characterized in that, include: The multi-source data acquisition module is used to collect various physical quantity data of power equipment in real time through various IoT protocols based on distributed edge nodes, and to perform preliminary format unification and timestamp marking. The multi-source data acquisition module is used to perform protocol parsing, spatiotemporal alignment, and dynamic weighted fusion based on the acquired multimodal data to generate a unified spatiotemporal dataset. A multiphysics simulation engine is used to simulate the state of power equipment under the coupling of thermo-mechanical-electrical fields in real time based on a unified spatiotemporal dataset and finite element analysis model, and to predict fault evolution. The intelligent decision engine, operation and maintenance execution linkage module, and visualization interaction module are used to analyze simulation and deduction results, generate specific operation and maintenance strategies and equipment health assessments, thereby automatically triggering external execution units.

9. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores computer-executable instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images