A power equipment digital twin model construction method, system, device and medium

Abstract

The application discloses a power equipment digital twin model construction method, system, device and medium, relates to the technical field of digital twin, and comprises the following steps: acquiring target multi-source features of power equipment in real time; performing pseudo-color coding based on the target multi-source features to obtain a dynamic color thermal map for the power equipment; constructing a static three-dimensional model of the power equipment; and performing model construction based on a preset abnormality identification algorithm, the dynamic color thermal map and the static three-dimensional model to obtain a target digital twin model of the power equipment. The method solves the problems that the prior art is prone to introducing errors due to manual operation, causes data input delay, reduces the timeliness of digital twin simulation, and is difficult to accurately reflect the running state of the power equipment in real time.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, and in particular to a method, system, equipment and medium for constructing a digital twin model of power equipment. Background Technology

[0002] Digital twins of power equipment are virtual mirror images of physical equipment (such as transformers, circuit breakers, generators, and transmission lines) throughout their entire lifecycle in digital space. Utilizing sensor data, physical models, operational history, and artificial intelligence algorithms, they enable real-time mapping, status monitoring, performance simulation, predictive analysis, and optimized control of physical equipment. It is a dynamic, interactive, and high-fidelity "digital clone" that allows maintenance personnel to see the real-time status of equipment in a virtual world, predict future changes, and test various decision-making schemes.

[0003] In existing technologies, digital twins of power equipment typically employ manual methods to input real-time collected data into a static 3D model. This method is not only prone to errors introduced by human operation but also leads to data input delays, thereby reducing the timeliness of the digital twin simulation and making it difficult to achieve a real-time and accurate reflection of the operating status of power equipment. Summary of the Invention

[0004] To overcome the problems of existing technologies, which are prone to errors due to human operation and data input delays, thereby reducing the timeliness of digital twin simulation and making it difficult to achieve real-time and accurate reflection of the operating status of power equipment, this application provides a method, system, equipment and medium for constructing a digital twin model of power equipment.

[0005] Firstly, in order to solve the above-mentioned technical problems, this application provides a method for constructing a digital twin model of power equipment, including: Real-time acquisition of target multi-source features of power equipment; Pseudo-color coding is performed based on target multi-source features to obtain dynamic color heatmaps for power equipment. Construct a static 3D model of the power equipment; Based on a pre-defined anomaly detection algorithm, dynamic color heatmap, and static 3D model, a target digital twin model of the power equipment is constructed.

[0006] Furthermore, real-time acquisition of target multi-source features of power equipment, including: Acquire historical multi-source monitoring data of power equipment and collect current multi-source monitoring data of power equipment in real time; Preprocessing historical and current multi-source monitoring data yields multi-source target data; Feature extraction is performed on multi-source target data to obtain multi-source target features of power equipment.

[0007] Furthermore, feature extraction is performed on the multi-source target data to obtain the target multi-source features of the power equipment, including: Temporal features are extracted from multi-source target data to obtain multi-source temporal features; Frequency domain features are extracted from multi-source target data to obtain multi-source frequency domain features; Multi-source wavelet features are obtained by performing wavelet transform based on multi-source time-domain features and multi-source frequency-domain features. Partial discharge features are extracted from multi-source target data to obtain multi-source partial discharge features; Thermal features are extracted from multi-source target data to obtain multi-source thermal features; Based on multi-source time-domain features, multi-source frequency-domain features, multi-source wavelet features, multi-source partial discharge features, and multi-source thermal features, target multi-source features of power equipment are formed.

[0008] Furthermore, pseudo-color encoding is performed based on the target's multi-source features to obtain a dynamic color heatmap for power equipment, including: Structural optimization is performed based on the target's multi-source features to obtain the target temperature matrix; The target temperature matrix is ​​pseudo-color encoded and mapped to obtain a dynamic color heat map for power equipment.

[0009] Furthermore, structural optimization is performed based on the target's multi-source features to obtain the target temperature matrix, including: Feature optimization is performed based on the target's multi-source features to obtain optimized target multi-source features; The target temperature matrix is ​​obtained by constructing a matrix based on the time attribute, type attribute, and feature attribute of each feature in the multi-source features optimized for the target.

[0010] Furthermore, based on a pre-defined anomaly detection algorithm, dynamic color heatmap, and static 3D model, a target digital twin model of the power equipment is constructed, including: By fusing dynamic color heatmaps and static 3D models, dynamic twin models and dynamic twin heatmaps are obtained. Using a pre-defined anomaly detection algorithm, anomalies are labeled based on a dynamic twin heatmap to obtain an anomaly twin heatmap; Based on the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model, a target digital twin model of the power equipment is formed, and the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model can be switched and displayed according to the display needs of the power equipment.

[0011] Furthermore, the dynamic color heat map includes local heat maps of multiple features of the power equipment, and the static three-dimensional model includes multiple spare parts models of the power equipment, with each spare parts model corresponding to a local heat map of at least one feature. By fusing dynamic color heatmaps and static 3D models, a dynamic twin model and a dynamic twin heatmap are obtained, including: Obtain the correspondence between spare parts features between dynamic color heatmaps and static 3D models; Based on the correspondence of spare parts features, the local heat map is integrated into the static three-dimensional model to form a dynamic twin model; Based on the correspondence of spare parts features, the spare parts model is integrated into the dynamic color heat map to form a dynamic twin heat map.

[0012] Secondly, this application also provides a system for constructing digital twin models of power equipment, including: The acquisition module is used to acquire target multi-source features of power equipment in real time; The encoding module is used to perform pseudo-color encoding based on the target's multi-source features to obtain a dynamic color heatmap of the power equipment. The first construction module is used to build static 3D models of power equipment; The second construction module is used to build a model based on a preset anomaly recognition algorithm, dynamic color heat map and static three-dimensional model, so as to obtain the target digital twin model of the power equipment.

[0013] Thirdly, this application also provides a computing device, including a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the above-described method for constructing a digital twin model of a power equipment.

[0014] Fourthly, this application also provides a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform the steps of a method for constructing a digital twin model of power equipment.

[0015] The beneficial effects of this application are as follows: First, pseudo-color encoding is performed based on the target multi-source features of the power equipment acquired in real time to obtain a dynamic color heatmap of the power equipment. Second, a static three-dimensional model of the power equipment is constructed, and a model is built based on a preset anomaly recognition algorithm, the dynamic color heatmap, and the static three-dimensional model to obtain a target digital twin model that can display the real-time status of the power equipment. In this way, dynamic twin simulation of the real-time status of the power equipment is achieved through this target digital twin model, avoiding errors introduced by human operation, thereby improving the timeliness of the digital twin simulation and achieving a real-time and accurate reflection of the operating status of the power equipment. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for constructing a digital twin model of power equipment, as shown in an exemplary embodiment of this application. Figure 2 This is a schematic diagram illustrating the structure of a digital twin model construction system for power equipment, which is an exemplary embodiment of this application. Detailed Implementation

[0017] The following embodiments are further explanations and supplements to this application and do not constitute any limitation on this application.

[0018] The following describes, with reference to the accompanying drawings, a method, system, device, and medium for constructing a digital twin model of power equipment according to an embodiment of this application.

[0019] The method for constructing a digital twin model of power equipment provided in this application can be specifically executed by a server. It should be noted that the server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. No limitation is imposed here.

[0020] Please see Figure 1 , Figure 1 An exemplary embodiment of this application illustrates a method for constructing a digital twin model of power equipment, such as... Figure 1 As shown, this application provides a method for constructing a digital twin model of power equipment, including: S11, real-time acquisition of target multi-source features of power equipment; S12, pseudo-color coding is performed based on the target multi-source features to obtain a dynamic color heat map for power equipment; S13, Construct a static 3D model of the power equipment; S14. Based on the preset anomaly recognition algorithm, dynamic color heat map and static three-dimensional model, the model is constructed to obtain the target digital twin model of the power equipment.

[0021] The method for constructing a digital twin model of power equipment provided in this application firstly performs pseudo-color encoding based on the target multi-source features of the power equipment acquired in real time, obtaining a dynamic color heatmap of the power equipment. Secondly, a static three-dimensional model of the power equipment is constructed, and the model is built based on a preset anomaly recognition algorithm, the dynamic color heatmap, and the static three-dimensional model, resulting in a target digital twin model capable of displaying the real-time status of the power equipment. In this way, the dynamic twin simulation of the real-time status of the power equipment is achieved through this target digital twin model, avoiding errors introduced by human operation, thereby improving the timeliness of the digital twin simulation and achieving a real-time and accurate reflection of the operating status of the power equipment.

[0022] In this embodiment, constructing a static three-dimensional model of the power equipment may include: obtaining three-dimensional design drawings of the power equipment, which include a key dimension parameter table, a bill of materials (BOM), a tolerance and fit table, and an assembly relationship description; and drawing the static three-dimensional model of the power equipment in three-dimensional software according to the three-dimensional design drawings.

[0023] Optionally, target multi-source features of power equipment can be acquired in real time, including: Acquire historical multi-source monitoring data of power equipment and collect current multi-source monitoring data of power equipment in real time; Preprocessing historical and current multi-source monitoring data yields multi-source target data; Feature extraction is performed on multi-source target data to obtain multi-source target features of power equipment.

[0024] In the embodiment provided in this application, by preprocessing the historical and current multi-source monitoring data of power equipment, interference data can be removed, data consistency can be improved, and standardized multi-source target data can be obtained. Furthermore, feature extraction is performed on the multi-source target data to obtain the target multi-source features of the power equipment. This facilitates the subsequent formation of a real-time dynamic color heatmap based on the target multi-source features, and the construction of a real-time target digital twin model. This target digital twin model enables dynamic twin simulation of the real-time state of the power equipment. The preprocessing includes statistical filtering (moving average filtering, wavelet denoising, Kalman filtering to remove impulse noise), anomaly detection (using Isolation Forest, or Local Outlier Factor (LOF), or the 3σ principle, or box plots to automatically identify abnormal sensor data), missing value imputation (including at least one method among MICE, time series interpolation, and adjacent device data imputation), and data alignment (unifying timestamps to handle time lag issues of different data components).

[0025] In this embodiment, historical multi-source monitoring data of power equipment is acquired, and current multi-source monitoring data of power equipment is collected in real time, including: ① determining the data sources of power equipment: SCADA system, online monitoring device (vibration, partial discharge, temperature), DCS, historical database; ② establishing data interfaces: OPC UA, Modbus, MQTT, direct database connection; ③ setting the data sampling frequency of power equipment: determined according to the equipment type of power equipment (vibration: 10kHz, temperature: 1Hz, partial discharge: 100MHz); ④ acquiring historical multi-source monitoring data of power equipment from data sources according to the data sampling frequency through the data interface, and collecting current multi-source monitoring data of power equipment in real time.

[0026] Optionally, feature extraction is performed on the multi-source target data to obtain the target multi-source features of the power equipment, including: Temporal features are extracted from multi-source target data to obtain multi-source temporal features; Frequency domain features are extracted from multi-source target data to obtain multi-source frequency domain features; Multi-source wavelet features are obtained by performing wavelet transform based on multi-source time-domain features and multi-source frequency-domain features. Partial discharge features are extracted from multi-source target data to obtain multi-source partial discharge features; Thermal features are extracted from multi-source target data to obtain multi-source thermal features; Based on multi-source time-domain features, multi-source frequency-domain features, multi-source wavelet features, multi-source partial discharge features, and multi-source thermal features, target multi-source features of power equipment are formed.

[0027] In the embodiment provided in this application, time-domain feature extraction, frequency-domain feature extraction, partial discharge feature extraction, and thermal feature extraction are performed on multi-source target data to fully extract the main key parameters under the real-time operating state of the power equipment, resulting in multi-source time-domain features, multi-source frequency-domain features, multi-source partial discharge features, and multi-source thermal features. Furthermore, wavelet transform is performed based on the multi-source time-domain and multi-source frequency-domain features to extract the energy characteristics during the operation of the power equipment, resulting in multi-source wavelet features, forming the target multi-source features of the power equipment. This allows for the full extraction of various key parameters under the real-time operating state of the power equipment, facilitating the subsequent intuitive simulation of these key parameters from the target digital twin model, thereby achieving a real-time and accurate reflection of the power equipment's operating state. The multi-source thermal features include temperature distribution parameters and thermal model parameters. Temperature distribution parameters include hot spot temperature, temperature gradient, temperature rise rate, and cooling characteristics; thermal model parameters include thermal resistance, heat capacity, and thermal time constant.

[0028] In this embodiment, time-domain features are extracted from multi-source target data to obtain multi-source time-domain features. This includes: extracting time-domain features from time-series signals such as vibration and current in the multi-source target data to obtain statistical features, shape features, and time-series features; and forming multi-source time-domain features based on the statistical features, shape features, and time-series features. The statistical features include: mean, variance, peak value, peak-to-peak value, and root mean square; the shape features include: skewness, kurtosis, waveform factor, peak factor, and impulse factor; and the time-series features include: autocorrelation coefficient, cross-correlation coefficient, and zero-crossing rate.

[0029] In this embodiment, frequency domain features are extracted from multi-source target data to obtain multi-source frequency domain features, including: performing FFT spectrum analysis on the multi-source target data to obtain spectral features, fault frequencies, and harmonic features; and forming multi-source frequency domain features based on the spectral features, fault frequencies, and harmonic features. The spectral features include: spectral centroid, spectral variance, spectral skewness, and spectral kurtosis; the fault frequencies include: bearing characteristic frequencies, gear meshing frequencies, and motor characteristic frequencies; and the harmonic features include: fundamental amplitude, harmonic content, and total harmonic distortion.

[0030] In this embodiment, wavelet transform is performed based on multi-source time-domain features and multi-source frequency-domain features to obtain multi-source wavelet features. This includes: wavelet packet decomposition based on multi-source time-domain features and multi-source frequency-domain features to obtain decomposed features; and continuous wavelet transform is performed on the decomposed features to obtain multi-source wavelet features. The decomposed features include: energy proportion of each frequency band, energy entropy, and wavelet coefficient variance; the multi-source wavelet features include: time-frequency energy distribution and feature scale.

[0031] In this embodiment, partial discharge (PD) features are extracted from multi-source target data to obtain multi-source PD features. This includes: performing PRPD (Phase-Resolved Partial Discharge) statistics on the multi-source target data to obtain charge characteristics; and extracting pulse features based on the charge characteristics to obtain multi-source PD features. The charge characteristics include: average discharge quantity, maximum discharge quantity, discharge repetition rate, and phase distribution; the multi-source PD features include: rise time, pulse width, pulse interval, and pulse shape.

[0032] Optionally, pseudo-color coding is performed based on the target multi-source features to obtain a dynamic color heatmap for power equipment, including: Structural optimization is performed based on the target's multi-source features to obtain the target temperature matrix; The target temperature matrix is ​​pseudo-color encoded and mapped to obtain a dynamic color heat map for power equipment.

[0033] In the embodiment provided in this application, structural optimization is performed based on the target multi-source features to obtain a target temperature matrix. The target temperature matrix is ​​then pseudo-color encoded and mapped to obtain a dynamic color heatmap for the power equipment. This achieves an intuitive mapping from the target multi-source features to the dynamic color heatmap, facilitating the subsequent construction of a target digital twin model based on the dynamic color heatmap to provide an intuitive and accurate display of the real-time operating status of the power equipment. The pseudo-color encoding includes: using color gradients to represent feature value magnitudes, adding a time axis and feature name annotations. The time axis and feature name annotations include: a time-series change graph for each key feature, a feature state graph (normal state is a regular circle, fault state is an irregular shape), and a state evolution trajectory graph.

[0034] Optionally, structural optimization is performed based on the target multi-source features to obtain the target temperature matrix, including: Feature optimization is performed based on the target's multi-source features to obtain optimized target multi-source features; The target temperature matrix is ​​obtained by constructing a matrix based on the time attribute, type attribute, and feature attribute of each feature in the multi-source features optimized for the target.

[0035] In the embodiment provided in this application, feature optimization is performed based on target multi-source features to obtain target optimized multi-source features. Then, a matrix is ​​constructed based on the time attribute, type attribute, and feature attribute of each feature in the target optimized multi-source features to obtain a target temperature matrix. This results in the target temperature matrix containing attribute information (time attribute, type attribute, and feature attribute, etc.) that can accurately describe the operating status of power equipment. This improves the matching degree between the dynamic color heat map obtained by subsequent encoding and the power equipment, and further enhances the timeliness and accuracy of the target digital twin model in dynamically simulating the operating status of power equipment.

[0036] In this embodiment, feature optimization is performed based on target multi-source features to obtain target optimized multi-source features, including: calculating the feature correlation between every two features in the target multi-source features; for each feature correlation, when the feature correlation is greater than a correlation threshold, selecting one feature from the two features corresponding to the feature correlation as a redundant feature; deleting redundant features in the target multi-source features to obtain initial optimized multi-source features; using a preset importance algorithm to score each feature in the initial optimized multi-source features to obtain an importance score; wherein the importance algorithm is the mutual information method or the chi-square test method; deleting features in the initial optimized multi-source features whose importance scores are less than the score threshold to obtain target optimized multi-source features.

[0037] Optionally, a target digital twin model of the power equipment is obtained by constructing a model based on a preset anomaly detection algorithm, a dynamic color heat map, and a static 3D model, including: By fusing dynamic color heatmaps and static 3D models, dynamic twin models and dynamic twin heatmaps are obtained. Using a pre-defined anomaly detection algorithm, anomalies are labeled based on a dynamic twin heatmap to obtain an anomaly twin heatmap; Based on the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model, a target digital twin model of the power equipment is formed, and the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model can be switched and displayed according to the display needs of the power equipment.

[0038] In the embodiment provided in this application, a dynamic twin model and a dynamic twin heatmap are obtained by interlinking and fusing a dynamic color heatmap and a static 3D model. A preset anomaly identification algorithm is then used to annotate the dynamic twin heatmap, resulting in an anomaly twin heatmap. This forms the target digital twin model of the power equipment, achieving a qualitative leap from a static 3D model to a multi-state fusion, intelligent, and interactive dynamic twin. By accurately fusing real-time monitoring data with the 3D model in the form of a dynamic color heatmap, the system generates a dynamic twin model and heatmap that intuitively reflects the distribution of multiple physical fields such as temperature, stress, and partial discharge in the equipment. In addition, based on preset algorithms, the system intelligently analyzes dynamic heatmaps, automatically identifies and marks abnormal areas, generates anomaly twin heatmaps with superimposed risk information, and ultimately forms a target digital twin model. This model supports on-demand switching between panoramic dynamic views, anomaly focused views, and pure 3D structural views, enabling maintenance personnel to not only have a comprehensive view of the overall real-time status of the equipment but also to locate risk points with a single click. This achieves a leap from "passively viewing data" to "actively perceiving risks," providing intuitive and intelligent visualization support for accurate equipment status assessment, early fault warning, and efficient maintenance decision-making.

[0039] In this embodiment, the algorithm flow of the anomaly detection algorithm is as follows: ① Based on the dynamic twin heatmap, local blocks are divided to obtain local heatmaps, and the temperature matrix corresponding to the local heatmaps is used. Temporal feature extraction is performed to obtain the temperature time gradient. and the accumulation of temperature changes ; Temperature time gradient The calculation formula is as follows: in, Indicates the frame time interval; Accumulated temperature change The calculation formula is as follows: ② Perform spatial feature extraction on the temperature matrix to obtain the temperature spatial gradient. and local contrast ; Temperature spatial gradient The calculation formula is as follows: Local contrast The calculation formula is as follows: in, and For Statistics of the local heatmap centered on the image (local heatmap size: typically 7×7 or 15×15 pixels). =10 -6 To prevent division by zero; ③ Based on the temperature matrix, temperature time gradient, temperature spatial gradient, and local contrast, feature cubes are obtained by combining features. : Calculate reconstruction error value based on feature cube : ④ Perform multi-layer 3D convolutional feature extraction based on feature cubes to obtain high-dimensional features. Calculate feature difference values ​​based on high-dimensional features : ; ⑤ Based on the reconstruction error value and feature difference value, the following is calculated: Anomaly score of local heatmap centered on The calculation formula is as follows: in, , For Historical normal fractional statistics of the local heatmap centered on the subject; ⑥ When the abnormal score is greater than or equal to the abnormal threshold, it is identified as an abnormal heatmap. The local heatmap in the dynamic twin heatmap is marked as abnormal, and multiple abnormal heatmaps are arranged in descending order of abnormal score to obtain an abnormal twin heatmap.

[0040] In this way, by identifying abnormal features (precursors of failure) in power equipment based on dynamic twin heatmaps, and updating the dynamic twin heatmap, abnormal twin heatmap, and dynamic twin model in real time based on these abnormal features, it is possible to subsequently locate the abnormal location in the dynamic twin model by associating the abnormal heatmap marked with anomalies in the dynamic twin heatmap. This allows maintenance personnel to quickly locate and repair abnormalities in the power equipment based on the anomaly locations identified in the dynamic twin model. Simultaneously, through continuous interaction between the digital twin model and actual operating data, accurate capture and spatial location of precursors of failure are achieved, improving the timeliness and reliability of operation and maintenance decisions. Furthermore, based on changes in the dynamic twin heatmap and abnormal twin heatmap, the operating status and health index of various components of the power equipment can be mapped in real time. Color gradient changes reflect the dynamic evolution trends of key parameters such as temperature, vibration, and current. Combined with multi-dimensional synchronous analysis over a time axis, abnormal features are identified, associated with corresponding locations in the dynamic twin model, and an early warning mechanism is triggered.

[0041] The target digital twin model in this embodiment can achieve the following functions: fault diagnosis (outputting fault type probability), health assessment (outputting health score 0-1), life prediction (outputting remaining life hours), and anomaly detection (outputting reconstruction error score).

[0042] Optionally, the dynamic color heat map includes local heat maps of multiple features of the power equipment, and the static three-dimensional model includes multiple spare part models of the power equipment, with each spare part model corresponding to a local heat map of at least one feature. By fusing dynamic color heatmaps and static 3D models, a dynamic twin model and a dynamic twin heatmap are obtained, including: Obtain the correspondence between spare parts features between dynamic color heatmaps and static 3D models; Based on the correspondence of spare parts features, the local heat map is integrated into the static three-dimensional model to form a dynamic twin model; Based on the correspondence of spare parts features, the spare parts model is integrated into the dynamic color heat map to form a dynamic twin heat map.

[0043] In the embodiment provided in this application, based on the correspondence between spare parts features between the dynamic color heat map and the static three-dimensional model, the local heat map is fused into the static three-dimensional model to form a dynamic twin model, and the spare parts model is fused into the dynamic color heat map to form a dynamic twin heat map. This facilitates the dynamic switching between the dynamic twin model and the dynamic twin heat map in the target digital twin model, as well as the location of spare parts or the location of spare parts features of the power equipment, thereby achieving a real-time and accurate reflection of the operating status of the power equipment.

[0044] Please see Figure 2 , Figure 2An exemplary embodiment of this application illustrates a system for constructing a digital twin model of power equipment, such as... Figure 2 As shown, this application provides a digital twin model construction system 200 for power equipment, including: The acquisition module 201 is used to acquire the target multi-source features of power equipment in real time; Encoding module 202 is used to perform pseudo-color encoding based on target multi-source features to obtain a dynamic color heat map for power equipment; The first construction module 203 is used to construct a static three-dimensional model of the power equipment; The second construction module 204 is used to construct a model based on a preset anomaly recognition algorithm, dynamic color heat map and static three-dimensional model to obtain a target digital twin model of the power equipment.

[0045] The power equipment digital twin model construction system 200 provided in this application firstly uses an encoding module 202 to perform pseudo-color encoding based on the target multi-source features of the power equipment acquired in real time by an acquisition module 201, obtaining a dynamic color heatmap of the power equipment. Secondly, a static three-dimensional model of the power equipment is constructed using a first construction module 203, and a second construction module 204 is used to construct a model based on a preset anomaly recognition algorithm, the dynamic color heatmap, and the static three-dimensional model, obtaining a target digital twin model capable of displaying the real-time status of the power equipment. In this way, the dynamic twin simulation of the real-time status of the power equipment is achieved through this target digital twin model, avoiding errors introduced by human operation, thereby improving the timeliness of the digital twin simulation and realizing a real-time and accurate reflection of the operating status of the power equipment.

[0046] Optionally, module 201 is used specifically for: Acquire historical multi-source monitoring data of power equipment and collect current multi-source monitoring data of power equipment in real time; Preprocessing historical and current multi-source monitoring data yields multi-source target data; Feature extraction is performed on multi-source target data to obtain multi-source target features of power equipment.

[0047] Optionally, module 201 is used specifically for: Temporal features are extracted from multi-source target data to obtain multi-source temporal features; Frequency domain features are extracted from multi-source target data to obtain multi-source frequency domain features; Multi-source wavelet features are obtained by performing wavelet transform based on multi-source time-domain features and multi-source frequency-domain features. Partial discharge features are extracted from multi-source target data to obtain multi-source partial discharge features; Thermal features are extracted from multi-source target data to obtain multi-source thermal features; Based on multi-source time-domain features, multi-source frequency-domain features, multi-source wavelet features, multi-source partial discharge features, and multi-source thermal features, target multi-source features of power equipment are formed.

[0048] Optionally, the encoding module 202 is specifically used for: Structural optimization is performed based on the target's multi-source features to obtain the target temperature matrix; The target temperature matrix is ​​pseudo-color encoded and mapped to obtain a dynamic color heat map for power equipment.

[0049] Optionally, the encoding module 202 is specifically used for: Feature optimization is performed based on the target's multi-source features to obtain optimized target multi-source features; The target temperature matrix is ​​obtained by constructing a matrix based on the time attribute, type attribute, and feature attribute of each feature in the multi-source features optimized for the target.

[0050] Optionally, the second building module 204 is specifically used for: By fusing dynamic color heatmaps and static 3D models, dynamic twin models and dynamic twin heatmaps are obtained. Using a pre-defined anomaly detection algorithm, anomalies are labeled based on a dynamic twin heatmap to obtain an anomaly twin heatmap; Based on the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model, a target digital twin model of the power equipment is formed, and the abnormal twin heatmap, dynamic twin heatmap, and dynamic twin model can be switched and displayed according to the display needs of the power equipment.

[0051] Optionally, the dynamic color heat map includes local heat maps of multiple features of the power equipment, and the static three-dimensional model includes multiple spare part models of the power equipment, with each spare part model corresponding to a local heat map of at least one feature. The second building module 204 is specifically used for: Obtain the correspondence between spare parts features between dynamic color heatmaps and static 3D models; Based on the correspondence of spare parts features, the local heat map is integrated into the static three-dimensional model to form a dynamic twin model; Based on the correspondence of spare parts features, the spare parts model is integrated into the dynamic color heat map to form a dynamic twin heat map.

[0052] It should be noted that the power equipment digital twin model construction system and the power equipment digital twin model construction method provided in the above embodiments belong to the same concept. The specific methods of operation of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the power equipment digital twin model construction system provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0053] A computing device according to an embodiment of this application includes a memory, a processor, and a program stored in the memory and running on the processor. When the processor executes the program, it implements some or all of the steps of the above-described method for constructing a digital twin model of power equipment.

[0054] The computing device can be a computer, and the corresponding program is computer software. The parameters and steps of the computing device described above can be referred to the parameters and steps in the embodiment of the method for constructing a digital twin model of power equipment in the above text, and will not be repeated here.

[0055] This application embodiment provides a computer-readable storage medium storing instructions that, when executed, perform the steps of the above-described method for constructing a digital twin model of power equipment.

[0056] The computer-readable storage medium may be a transient computer-readable storage medium or a non-transitory computer-readable storage medium.

[0057] The technical solutions of this disclosure can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of this disclosure. The aforementioned computer-readable storage medium can be a non-transitory computer-readable storage medium, including: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, and other media capable of storing program code; it can also be a transient computer-readable storage medium.

[0058] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0059] Those skilled in the art will recognize that this application can be implemented as a system, method, or computer program product. Therefore, this disclosure can be implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "module" or "system." Furthermore, in some embodiments, this application can also be implemented as a computer program product contained in one or more computer-readable media, which contains computer-readable program code. Computer-readable storage media can be, for example, but not limited to—electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof.

[0060] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0061] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for constructing a digital twin model of power equipment, characterized in that, include: Real-time acquisition of target multi-source features of power equipment; Based on the target multi-source features, pseudo-color encoding is performed to obtain a dynamic color heatmap for the power equipment. Construct a static three-dimensional model of the power equipment; Based on the dynamic color heatmap and the static three-dimensional model, a dynamic twin model and a dynamic twin heatmap are obtained by cross-referencing and fusing them. Using a preset anomaly detection algorithm, local blocks are segmented based on the dynamic twin heatmap to obtain a local heatmap, and a temperature matrix corresponding to the local heatmap is generated. Temporal feature extraction is performed to obtain the temperature time gradient. ; The temperature time gradient The calculation formula is as follows: in, Indicates the frame time interval; Spatial feature extraction is performed on the temperature matrix to obtain the temperature spatial gradient. and local contrast ; The temperature spatial gradient The calculation formula is as follows: Local contrast The calculation formula is as follows: in, and For Local heatmap statistics centered on the center =10 -6 To prevent division by zero; Based on the temperature matrix, the temperature time gradient, the temperature spatial gradient, and the local contrast, a feature cube is obtained by feature combination. : Calculate reconstruction error value based on feature cube : ; Based on the feature cube, multi-layer 3D convolutional feature extraction is performed to obtain high-dimensional features. Calculate the feature difference value based on the high-dimensional features. : ; Based on the reconstruction error value and the feature difference value, the following is calculated: Anomaly score of local heatmap centered on The calculation formula is as follows: in, , For Historical normal fractional statistics of the local heatmap centered on the subject; When the abnormal score is greater than or equal to the abnormal threshold, it is determined to be an abnormal heatmap, and the local heatmaps in the dynamic twin heatmap are marked as abnormal. Multiple abnormal heatmaps are arranged in descending order of abnormal score to obtain an abnormal twin heatmap. Based on the abnormal twin heatmap, the dynamic twin heatmap, and the dynamic twin model, a target digital twin model of the power equipment is formed, so that the abnormal twin heatmap, the dynamic twin heatmap, and the dynamic twin model can be switched and displayed according to the display needs of the power equipment.

2. The method according to claim 1, characterized in that, The real-time acquisition of target multi-source features of power equipment includes: Acquire historical multi-source monitoring data of the power equipment, and collect current multi-source monitoring data of the power equipment in real time; The historical multi-source monitoring data and the current multi-source monitoring data are preprocessed to obtain multi-source target data; Feature extraction is performed on the multi-source target data to obtain the target multi-source features of the power equipment.

3. The method according to claim 2, characterized in that, The step of extracting features from the multi-source target data to obtain the target multi-source features of the power equipment includes: Temporal features are extracted from the multi-source target data to obtain multi-source temporal features; Frequency domain features are extracted from the multi-source target data to obtain multi-source frequency domain features; Based on the multi-source time-domain features and the multi-source frequency-domain features, wavelet transform is performed to obtain multi-source wavelet features; The partial discharge features of the multi-source target data are extracted to obtain multi-source partial discharge features; Thermal features are extracted from the multi-source target data to obtain multi-source thermal features; Based on the multi-source time-domain features, the multi-source frequency-domain features, the multi-source wavelet features, the multi-source partial discharge features, and the multi-source thermal features, the target multi-source features of the power equipment are formed.

4. The method according to claim 1, characterized in that, The step of performing pseudo-color encoding based on the target multi-source features to obtain a dynamic color heatmap for the power equipment includes: Structural optimization is performed based on the target multi-source features to obtain the target temperature matrix; The target temperature matrix is ​​pseudo-color encoded and mapped to obtain a dynamic color heat map for the power equipment.

5. The method according to claim 4, characterized in that, The structural optimization based on the target multi-source features to obtain the target temperature matrix includes: Feature optimization is performed based on the target multi-source features to obtain the target optimized multi-source features; Based on the time attribute, type attribute, and feature attribute of each feature in the multi-source features optimized for the target, a matrix is ​​constructed to obtain the target temperature matrix.

6. The method according to claim 1, characterized in that, The dynamic color heat map includes local heat maps of multiple features of the power equipment, and the static three-dimensional model includes multiple spare part models of the power equipment, with each spare part model corresponding to a local heat map of at least one feature. The process of fusing the dynamic color heatmap and the static 3D model to obtain a dynamic twin model and a dynamic twin heatmap includes: Obtain the correspondence between spare parts features between the dynamic color heat map and the static three-dimensional model; Based on the corresponding relationship of the spare parts features, the local heat map is fused into the static three-dimensional model to form a dynamic twin model; Based on the corresponding relationship of the spare parts features, the spare parts model is fused into the dynamic color heat map to form a dynamic twin heat map.

7. A system for constructing digital twin models of power equipment, characterized in that, include: The acquisition module is used to acquire target multi-source features of power equipment in real time; The encoding module is used to perform pseudo-color encoding based on the target multi-source features to obtain a dynamic color heat map for the power equipment. The first construction module is used to construct a static three-dimensional model of the power equipment; The second construction module is used to perform mutual correlation and fusion based on the dynamic color heat map and the static three-dimensional model to obtain a dynamic twin model and a dynamic twin heat map; Using a preset anomaly detection algorithm, local blocks are segmented based on the dynamic twin heatmap to obtain a local heatmap, and a temperature matrix corresponding to the local heatmap is generated. Temporal feature extraction is performed to obtain the temperature time gradient. ; The temperature time gradient The calculation formula is as follows: in, Indicates the frame time interval; Spatial feature extraction is performed on the temperature matrix to obtain the temperature spatial gradient. and local contrast ; The temperature spatial gradient The calculation formula is as follows: Local contrast The calculation formula is as follows: in, and For Local heatmap statistics centered on the center =10 -6 To prevent division by zero; Based on the temperature matrix, the temperature time gradient, the temperature spatial gradient, and the local contrast, a feature cube is obtained by feature combination. : Calculate reconstruction error value based on feature cube : ; Based on the feature cube, multi-layer 3D convolutional feature extraction is performed to obtain high-dimensional features. Calculate the feature difference value based on the high-dimensional features. : ; Based on the reconstruction error value and the feature difference value, the following is calculated: Anomaly score of local heatmap centered on The calculation formula is as follows: in, , For Historical normal fractional statistics of the local heatmap centered on the subject; When the abnormal score is greater than or equal to the abnormal threshold, it is determined to be an abnormal heatmap, and the local heatmaps in the dynamic twin heatmap are marked as abnormal. Multiple abnormal heatmaps are arranged in descending order of abnormal score to obtain an abnormal twin heatmap. Based on the abnormal twin heatmap, the dynamic twin heatmap, and the dynamic twin model, a target digital twin model of the power equipment is formed, so that the abnormal twin heatmap, the dynamic twin heatmap, and the dynamic twin model can be switched and displayed according to the display needs of the power equipment.

8. A computing device comprising a memory, a processor, and a program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the steps of the method for constructing a digital twin model of power equipment as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a terminal device, cause the terminal device to perform the steps of a method for constructing a digital twin model of power equipment as described in any one of claims 1 to 6.

Images