Intelligent operation and maintenance method and device of digital twinborn model based on traction substation

Abstract

The invention provides an intelligent operation and maintenance method and device for a digital twin model based on a traction substation, and relates to the technical field of operation and maintenance of traction substations. The method comprises the steps of detecting operation states of various devices in a digital twin model, and if a first target device with an abnormal operation state is detected, determining an associated device having a connection relationship with the first target device; if it is determined that the associated device is a first type device with preset internal structure features, splitting the associated device to obtain an internal structure displaying the associated device in a plane mode; and carrying out image feature extraction on the internal structure, and analyzing the extracted image features to obtain an operation state detection result of the associated equipment. The device executes the method. According to the method and the device provided by the embodiment of the invention, equipment with potential safety hazards in the traction substation can be found as soon as possible.

Description

Technical Field

[0001] The present invention relates to the technical field of traction substation operation and maintenance, and in particular to an intelligent operation and maintenance method and device based on a digital twin model of a traction substation. Background Art

[0002] During daily operation, traction substation power supply equipment faces problems such as corrosion and oxidation of the main circuit clamps, poor contact, excessive temperatures, pressure drops in sulfur hexafluoride circuit breakers, transformer oil leakage leading to low oil levels, poor heat dissipation during operation, and high oil temperatures affecting service life. The quality of daily inspections by on-duty personnel is poor, preventing them from promptly identifying hidden defects during equipment operation. Furthermore, due to the large size of the equipment and insufficient operational and maintenance personnel, repair and testing procedures are not prioritized, resulting in poor maintenance quality and causing equipment to operate with defects. Currently, with the exception of the integrated automation system, various online monitoring systems in traction substations have not formed an integrated whole due to interface protocols and other factors. Furthermore, the limited number of existing online monitoring systems prevents effective and comprehensive monitoring of traction substation equipment, hindering timely maintenance and causing serious losses. Summary of the Invention

[0003] In response to the problems in the prior art, an embodiment of the present invention provides an intelligent operation and maintenance method and device based on a digital twin model of a traction substation, which can at least partially solve the problems in the prior art.

[0004] On the one hand, the present invention proposes a smart operation and maintenance method based on a digital twin model of a traction substation, comprising:

[0005] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0006] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0007] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0008] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0009] The detection of the operating status of various devices in the digital twin model includes:

[0010] Performing fusion processing on multi-source data of the various devices;

[0011] Perform data cleaning on the multi-source data after fusion processing;

[0012] Extract data features from cleaned multi-source data;

[0013] Based on the preset defect recognition rules, the multi-source data after feature extraction is used to identify defects, and the operating status of various equipment is determined according to the defect recognition results;

[0014] Among them, the preset defect recognition rules include a threshold-based defect recognition method and / or an AI model-based fine screening defect method.

[0015] The smart operation and maintenance method based on the digital twin model of the traction substation also includes:

[0016] Acquire the first target device and the second target device whose operating status detection result is abnormal;

[0017] Performing cause analysis and diagnosis on the first target device and the second target device;

[0018] generating a maintenance processing strategy based on the cause analysis and diagnosis results, and verifying whether the first target device and the second target device can be restored to a normal state according to the maintenance processing strategy;

[0019] The maintenance processing strategy corresponding to the equipment that can be restored to normal state is used as a new fault case, and the new fault case is added to the training set to update the knowledge graph.

[0020] The dissecting of the associated device to obtain a planar representation of the internal structure of the associated device includes:

[0021] The geometric center points of the core components in the associated device are obtained, and explosions are simulated using the geometric center points as blasting source points to obtain a planar representation of the internal structure of the core components in the associated device.

[0022] The step of extracting image features from the internal structure includes:

[0023] Marking the internal structure according to abnormal local areas that appeared when historical faults of the associated device occurred, to obtain key areas presented on the surface of the internal structure;

[0024] Image features are extracted from the key area.

[0025] The step of extracting image features from the internal structure further includes:

[0026] Image features of the internal structure are extracted through a bimodal attention network.

[0027] On the one hand, the present invention proposes an intelligent operation and maintenance device based on a digital twin model of a traction substation, comprising:

[0028] A determination unit is configured to detect the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determine an associated device that has a connection relationship with the first target device;

[0029] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0030] a dissection unit, configured to dissect the associated device to obtain a planar representation of the internal structure of the associated device if it is determined that the associated device is a first type device having a preset internal structure feature;

[0031] The detection unit is used to extract image features of the internal structure and analyze the extracted image features to obtain a detection result of the operating status of the associated device.

[0032] In another aspect, an embodiment of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following method is implemented:

[0033] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0034] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0035] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0036] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0037] An embodiment of the present invention provides a computer-readable storage medium, including:

[0038] The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following method is implemented:

[0039] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0040] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0041] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0042] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0043] An embodiment of the present invention further provides a computer program product, comprising a computer program. When the computer program is executed by a processor, the computer program implements the following method:

[0044] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0045] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0046] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0047] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0048] The embodiments of the present invention provide an intelligent operation and maintenance method and device based on a digital twin model of a traction substation, which detects the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device having a connection relationship with the first target device is determined; wherein the digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one to each physical device and includes the connection relationship between the physical devices; if the associated device is determined to be a first type of device with preset internal structural features, the associated device is dissected to obtain the internal structure of the associated device displayed in a planar manner; image features are extracted from the internal structure, and the extracted image features are analyzed to obtain the operating status detection results of the associated device, which can comprehensively, effectively and timely monitor the operating status of the equipment in the traction substation, so that equipment with safety hazards in the traction substation can be discovered early. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. 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 work. In the drawings:

[0050] Figure 1 It is a flow chart of an intelligent operation and maintenance method based on a digital twin model of a traction substation provided by one embodiment of the present invention.

[0051] Figure 2 It is a schematic diagram illustrating the dissection-related equipment provided by an embodiment of the present invention.

[0052] Figure 3 It is a modular structural diagram of the intelligent operation and maintenance method based on the digital twin model of the traction substation provided by an embodiment of the present invention.

[0053] Figure 4 It is a structural diagram of a model unit provided by an embodiment of the present invention.

[0054] Figure 5 It is a structural diagram of a patrol unit provided by an embodiment of the present invention.

[0055] Figure 6 It is a structural diagram of a device management unit provided by an embodiment of the present invention.

[0056] Figure 7 It is a structural diagram of the alarm management unit provided by an embodiment of the present invention.

[0057] Figure 8It is a structural diagram of a defect management unit provided by an embodiment of the present invention.

[0058] Figure 9 It is a structural diagram of a status evaluation unit provided by an embodiment of the present invention.

[0059] Figure 10 It is a structural schematic diagram of an intelligent operation and maintenance device based on a digital twin model of a traction substation provided by one embodiment of the present invention.

[0060] Figure 11 A schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of this application can be combined with each other in any manner.

[0062] Figure 1 This is a flow chart of a smart operation and maintenance method based on a digital twin model of a traction substation provided by an embodiment of the present invention. Figure 1 As shown, the intelligent operation and maintenance method based on the digital twin model of the traction substation provided by the embodiment of the present invention includes:

[0063] Step S1: Detecting the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0064] Among them, the digital twin model is a three-dimensional virtual model that is pre-established based on the physical equipment in the traction substation, corresponds one-to-one to each physical equipment, and includes the connection relationship between the physical equipment.

[0065] Step S2: If it is determined that the associated device is a first type device having preset internal structural features, the associated device is dissected to obtain a planar representation of the internal structure of the associated device.

[0066] Step S3: extracting image features of the internal structure and analyzing the extracted image features to obtain a detection result of the operating status of the associated device.

[0067] In the above step S1, the device detects the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device connected to the first target device is determined;

[0068] The digital twin model is a pre-built three-dimensional virtual model based on the physical devices in the traction substation, corresponding one-to-one with each physical device and including the connection relationships between the physical devices. The device can be a computer device that executes the method. The acquisition, storage, use, and processing of data in the technical solution of this application comply with relevant regulations. The physical devices in the traction substation may include high-voltage switches, circuit breakers, transformers, and busbars. Each physical device has a corresponding three-dimensional virtual device model. The connection relationships between the three-dimensional virtual device models can reflect the actual connection relationships between the physical devices. All three-dimensional virtual device models and their connection relationships constitute the three-dimensional virtual model, namely the digital twin model. It should be noted that this digital twin model is a refined digital twin model, which can simulate external changes of the physical devices in real time, such as tripping, indicator light flashing, and phenomena such as burning and smoke when a fault occurs. It can also simulate internal changes of the physical devices in real time, such as local short circuits and open circuits between turns in the transformer's internal coils and loose cores.

[0069] The establishment of the digital twin model is explained as follows:

[0070] Based on 3D modeling technology, real-time rendering is used to construct a digital twin substation that is consistent with the appearance, coordinates, and attributes of the substation and its internal facilities. It integrates and displays the substation's comprehensive indicator data, equipment operation data, alarm data, and risk management data, allowing macro-level control of the substation's operating status and providing a basis for decision-making for power users.

[0071] The detection of the operating status of various devices in the digital twin model includes:

[0072] The multi-source data of the various devices are fused and processed, including real-time sensor monitoring data (such as infrared thermal imagers capturing temperature anomalies), historical data (equipment ledgers, maintenance records, similar equipment fault database, etc.), and unstructured data (inspection reports, fault photos, voice recordings, etc.).

[0073] Perform data cleaning on the multi-source data after fusion processing, including removing outliers and filling missing values.

[0074] Perform data feature extraction on the cleaned multi-source data; including extracting time domain, frequency domain, and time-frequency domain features, such as wavelet packet energy entropy.

[0075] Based on the preset defect recognition rules, the multi-source data after feature extraction is used to identify defects, and the operating status of various equipment is determined according to the defect recognition results;

[0076] The preset defect recognition rules include threshold-based defect recognition methods and / or AI model-based fine-screening defect methods. The threshold-based defect recognition methods include terminal and transformer infrared overtemperature, transformer overheating, short circuit, etc.

[0077] The AI ​​model-based precise defect screening method includes insulator crack detection in drone inspection images based on image recognition and detection of current waveform distortion (sudden increase in harmonic content) using the LSTM model based on timing analysis.

[0078] If the defect identification result of the multi-source data for device S is that there is a defect, the operating status of device S is determined to be abnormal; if the defect identification result of the multi-source data for device S is that there is no defect, the operating status of device S is determined to be normal.

[0079] In step S2, if the device determines that the associated device is a first type device having predetermined internal structural characteristics, the device dissects the associated device to obtain a planar representation of the associated device's internal structure. If the operating status of device S is determined to be abnormal, device S is determined to be the first target device. If device S is connected to both device A and device B, both device A and device B are associated devices of device S.

[0080] The preset internal structure features can be determined according to the equipment type in the traction substation equipment. For example, the internal structure of the transformer is an iron core and a coil, etc. Correspondingly, the preset internal structure features corresponding to the transformer are the external features of the iron core, the external features of the coil, the external features of the oil tank, and the positional relationship features of the coil and the iron core, etc.

[0081] For example, a capacitor bank may be aging, but the associated switches and protective devices are well-designed, with a low probability of internal failure, so it can be assumed that internal failures are unlikely. Anomalies are typically triggered by changes in external parameters, rather than damage to the device itself. Therefore, it's unnecessary to designate a capacitor bank as a Type 1 device with predefined internal structural characteristics; instead, it can be designated as a Type 2 device, where these characteristics are not considered.

[0082] The associated device is dissected to obtain a planar display of the internal structure of the associated device, including:

[0083] The geometric center points of each core component in the associated equipment are obtained, and explosions are simulated using each geometric center point as the blasting source, resulting in a planar representation of the internal structure of each core component in the associated equipment. Core components are those that can cause internal failures in the associated equipment. For example, in a transformer, the core components of the transformer are the coil and tap changer. The fuel tank and iron core are generally not considered to cause internal failures in the associated equipment.

[0084] Referring to the above description, taking the coil as an example, the geometric center point of the coil is obtained, that is, the geometric center of the cylinder surrounded by the coil. The cross section of the simulated explosion can be set independently according to the actual situation, based on the ability to clearly and comprehensively display the internal features of the core component, such as Figure 2 As shown, for example, the cross section of the simulated explosion corresponding to the coil is Figure 2 As shown by the dotted line in the middle, the arrows point to the coil distribution after the simulated explosion, where each horizontal line represents a wound coil, so that the characteristics of each coil can be clearly displayed.

[0085] In the above step S3, the device extracts image features from the internal structure and analyzes the extracted image features to obtain the operating status detection result of the associated device. The image feature extraction of the internal structure includes:

[0086] The internal structure is marked according to the abnormal local area that appeared when the historical fault of the associated device occurred, and the key area presented on the surface of the internal structure is obtained; referring to the above example, if the 3rd and 4th coils of the transformer often fail, the locations of the 3rd and 4th coils are taken as the abnormal local area, such as Figure 2 As shown, the key area is an elliptical area.

[0087] Image feature extraction is performed on the key area. By performing image feature extraction on a part of the key area, the location features of the area where faults are prone to occur can be extracted more specifically.

[0088] The image feature extraction of the internal structure further includes:

[0089] The image features of the internal structure are extracted through the bimodal attention network. The following are the instructions:

[0090] The bimodal attention network is introduced to enhance the model's ability to focus on key areas (such as dynamic visual changes) and improve the accuracy of feature extraction.

[0091] 1. Independent enhancement of bimodal features:

[0092] The channel attention module (global average / maximum pooling combined with MLP) and the spatial attention module (spatial position weight distribution) are used for the features of each modality to screen important channels and key areas through dual attention.

[0093] The introduction of group receptive field blocks increases the scale diversity of local receptive fields and improves the ability to capture multi-scale features.

[0094] Cross-modal interaction mechanism:

[0095] The cross-attention mechanism is used to construct inter-modal feature associations, and the dynamic alignment and information complementarity of bimodal features are achieved through self-attention calculation.

[0096] Multimodal features are integrated into the Transformer encoder, and a multi-head attention mechanism is used to capture long-range dependencies.

[0097] 2. Multi-level integration strategy:

[0098] Feature-level fusion:

[0099] In the middle layer, the bimodal features are fused through adaptive weighting to automatically adjust the contribution of different modal features.

[0100] A dual-channel alignment mechanism is adopted to simultaneously calculate similarity attention and difference attention to enhance the contrastive learning ability of key areas.

[0101] Decision-level fusion:

[0102] The attention network is introduced into the independent detection results of the two modalities for weighted fusion, and the high-confidence detection areas are screened through the attention mechanism.

[0103] 3. Global feature optimization:

[0104] Global dependency modeling:

[0105] Second-order attention pooling is used to extract global features, and then the aggregated features are dynamically allocated to various spatial locations according to local needs to achieve collaborative optimization of global and local features4.

[0106] Build a multimodal graph structure, use graph neural networks to realize cross-modal message transmission, and enhance the semantic association between features.

[0107] 4. Dynamic calculation optimization:

[0108] Hybrid architecture design:

[0109] Combining the local feature extraction capabilities of CNN and the global modeling advantages of Transformer, a hybrid architecture is constructed to improve computing efficiency.

[0110] Redundant operations are reduced through parameter sharing and sparse attention calculation, which reduces computational complexity while ensuring accuracy.

[0111] Effectiveness: The above methods have been proven effective in tasks such as image recognition and object detection. For example, a bimodal network with adaptive weight fusion improved accuracy by 12% in object detection tasks, while a model incorporating a dual attention mechanism reduced the number of parameters by 40% while maintaining accuracy on the ImageNet dataset.

[0112] The intelligent operation and maintenance method based on the digital twin model of the traction substation also includes:

[0113] Acquire the first target device and the second target device whose operating status detection result is abnormal;

[0114] Perform cause analysis and diagnosis on the first target device and the second target device; this may include knowledge graph reasoning (it is necessary to build a device knowledge graph including: equipment, symptoms, faults and causal relationships, etc.); a deep learning model outputs fault probabilities (such as the probability of "casing discharge" is 85%, the probability of "loose core" is 10%) for multimodal data input (electrical parameters + vibration spectrum + oil chromatography data, etc.).

[0115] A maintenance processing strategy is generated based on the cause analysis and diagnosis results, and it is verified according to the maintenance processing strategy whether the first target device and the second target device can be restored to normal status; the maintenance processing strategy includes automated work order generation (generating maintenance work orders based on the diagnosis results, recommending spare parts and processing steps); maintenance verification includes the return of equipment data after maintenance to verify whether the indicators have returned to normal.

[0116] The repair and treatment strategies for devices that can be restored to normal status are treated as new fault cases and added to the training set to update the knowledge graph. Knowledge base iteration (adding new fault cases to the training set and updating the knowledge graph)

[0117] The intelligent operation and maintenance method based on the digital twin model of a traction substation provided in an embodiment of the present invention can be implemented based on modules, as described below:

[0118] like Figure 3 Shown, including:

[0119] The model unit uses 3D modeling technology to create a digital twin of the substation and its internal facilities, rendering it in real time to ensure that its appearance, coordinates, and attributes are consistent. This unit displays comprehensive indicators, equipment operation data, alarm data, and risk management data, enabling a macro-level understanding of the substation's operating status and providing a basis for decision-making for power users.

[0120] The patrol unit gathers multi-dimensional monitoring data from robots, drones, sensors, cameras, etc., and combines image analysis algorithms and intelligent diagnosis technology to conduct visual monitoring of equipment and environmental fault points and safety hazard points, realize real-time notification / warning of abnormal events, and provide a basis for command decision-making;

[0121] A device management unit, which is used to assist power users in querying and analyzing data;

[0122] An alarm management unit is used to assist power users in promptly identifying and handling equipment anomalies, thereby achieving the purpose of predictive maintenance of power equipment;

[0123] A defect management unit, configured to register and manage defect issues;

[0124] A status evaluation unit, configured to evaluate the status of the device;

[0125] An operation and maintenance unit, wherein the operation and maintenance unit is used to generate an operation and maintenance strategy.

[0126] like Figure 4 As shown, the model unit includes: district overview, substation overview, building overview and equipment monitoring, wherein the district overview is based on digital twin technology, which models and simulates the real-world area to demarcate the monitoring area; the substation overview is based on digital twin technology, which models and simulates the real-world substation model to understand the real situation of the substation; the building overview is based on digital twin technology, which models and simulates the real-world building model to understand the real situation of the building; the equipment monitoring is based on digital twin technology, which models and simulates the real-world equipment model to achieve real-time monitoring of the equipment.

[0127] like Figure 5 As shown, the inspection unit includes: a robot inspection module, a video inspection module and a drone inspection module. The robot inspection module realizes intelligent inspection of power equipment based on the infrared and visible light video surveillance cameras carried by the robot. The video inspection module applies the precise mapping, visualization and other capabilities of the digital twin platform to provide real-time video image data, combined with the three-dimensional model virtual composition, to realize "pointing and viewing" of on-site equipment from different angles, and 360-degree panoramic restoration of equipment information. The drone inspection module is based on the drone in the power station, and is used to realize automated drone inspection of substations.

[0128] like Figure 6 As shown, the equipment management unit includes: a substation equipment management module and a power line equipment management module. Both the substation equipment management module and the power line equipment management module are implemented based on intelligent sensors. The operating status of the equipment is transmitted to the system in the form of data through the intelligent sensors. By controlling the collection frequency, full-time, high-density, and full-dimensional equipment status data collection is achieved.

[0129] like Figure 7As shown, the alarm management unit includes an alarm condition screening module, an alarm information statistics module, an alarm list display module, an alarm defect registration module and an alarm list export module, wherein the alarm condition screening module is used to screen conditions for alarms, and the alarm condition screening module filters and queries according to conditions such as the power station name, alarm type, alarm object, alarm level, alarm status (relieved or not relieved), alarm occurrence time period, and alarm name; the alarm information statistics module is used to count alarm information and view the number and proportion of alarm statistics under different screening conditions; the alarm list display module is used to display alarms The information is displayed in the form of a list, wherein the alarm information includes: alarm name, alarm type, alarm level, alarm object, power station to which it belongs, alarm occurrence time, alarm duration, and alarm status; the alarm defect registration module is used to register alarm defect information. For unprocessed alarms that occur, the user is supported to register defects. The equipment alarm data is automatically substituted during defect registration, and intelligent diagnosis of defects is supported. For processed alarms, defect registration and disposal information can be queried; the alarm list export module exports alarm list information under different screening conditions, which is convenient for analysis and reporting by power users.

[0130] like Figure 8 As shown, the defect management unit includes: a defect intelligent diagnosis module, a defect information registration module, a defect case management module and a knowledge graph management module;

[0131] Among them, the defect intelligent diagnosis module includes: 1. Data preprocessing function, multi-source data fusion processing, including real-time sensor monitoring data (such as infrared thermal imagers capturing temperature anomalies), historical data (equipment ledgers, maintenance records, similar equipment fault libraries, etc.), unstructured data (inspection reports, fault photos, voice records, etc.); 2. Data cleaning function (eliminating outliers and filling missing values); 3. Feature extraction function (extracting time domain, frequency domain, time-frequency domain features, such as wavelet packet energy entropy); 4. Defect recognition function: ① Threshold-based defects (such as terminals, transformer infrared overtemperature, transformer overheating, short circuit, etc.); ② Fine screening defects based on AI models (insulator crack detection in drone inspection pictures based on image recognition, based on time series analysis) ③ Cause analysis and diagnosis function, mainly divided into: knowledge graph reasoning (need to build equipment knowledge graph including: equipment, symptoms, faults, cause and effect, etc.); deep learning model, for multimodal data input (electrical parameters + vibration spectrum + oil chromatography data, etc.), output fault probability (such as "casing discharge" probability 85%, "core loosening" probability 10%); ④ Closed-loop processing and feedback function, including automatic work order generation (generating maintenance work orders based on diagnostic results, recommending spare parts and processing steps); maintenance verification (equipment data is transmitted back after maintenance to verify whether the indicators have returned to normal); knowledge base iteration (adding new fault cases to the training set, updating the model and knowledge graph).

[0132] The intelligent defect diagnosis module can realize defect discovery, in-depth analysis and closed-loop processing; the defect information registration module is used to register defect problems discovered by power users during daily equipment inspections, and supports applications such as screening, previewing, editing, deletion, report generation and defect case sharing of registered defects; the defect case management module is based on defect case reports, technical guidelines, standard specifications and other document data imported by power users, and uses NLP artificial intelligence algorithms to identify user search intentions and realize intelligent knowledge retrieval applications; the knowledge graph management module is used to automatically locate the causes of equipment defects and provide users with defect elimination solutions or preventive measures.

[0133] The defect case management module includes applications such as case import, collection, review, query, preview, download, sharing, registration, and screening.

[0134] like Figure 9 As shown, the status evaluation unit includes: equipment status evaluation, risk assessment and operation and maintenance control, wherein the equipment status evaluation enhances the availability and accuracy of the equipment status evaluation function through the equipment status evaluation business operation process and according to the evaluation guideline elements; the risk assessment is used to perform equipment risk assessment business; the operation and maintenance control is used to review the operation and maintenance strategy and realize the final operation and maintenance strategy release.

[0135] The device status evaluation includes: device status evaluation overview, device status evaluation initiation and device status evaluation expert confirmation, wherein the device status evaluation overview is used to count the last evaluation results and perform performance optimization under large data volumes; the device status evaluation initiation is used to implement data permission control function, and users in various locations can only see their own data; the device status evaluation expert confirmation only modifies the results of the last evaluation.

[0136] The intelligent operation and maintenance method based on the digital twin model of a traction substation provided by the embodiment of the present invention has the following beneficial effects:

[0137] Based on the precise spatial location of 3D models of power stations, buildings, and equipment and their attribute information, combined with a high-performance, highly compatible 3D engine, real-time rendering is used to construct a digital twin power station that is consistent with the appearance, coordinates, and attributes of the smart substation and internal facilities. At the same time, multi-dimensional monitoring data from robots, drones, sensors, cameras, etc. is aggregated, and combined with image analysis algorithms and intelligent diagnosis technologies, visual monitoring of equipment and environmental fault points and safety hazard points is carried out to achieve real-time notification / warning of abnormal events, and provide a basis for command and decision-making, effectively improving the efficiency of power grid inspections, reducing equipment failure rates, ensuring the safe and stable operation of power equipment, realizing panoramic visual three-dimensional monitoring of power stations, and creating a panoramic monitoring center for smart substations.

[0138] An embodiment of the present invention provides an intelligent operation and maintenance method based on a digital twin model of a traction substation, which detects the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device having a connection relationship with the first target device is determined; wherein the digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one to each physical device and includes the connection relationship between the physical devices; if the associated device is determined to be a first type of device with preset internal structural features, the associated device is dissected to obtain the internal structure of the associated device displayed in a planar manner; image features are extracted from the internal structure, and the extracted image features are analyzed to obtain the operating status detection result of the associated device, which can comprehensively, effectively and timely monitor the operating status of the equipment in the traction substation, so that equipment with safety hazards in the traction substation can be discovered early.

[0139] Furthermore, the detection of the operating status of various devices in the digital twin model includes:

[0140] The multi-source data of the various devices are fused and processed; this can be described with reference to the above embodiment and will not be repeated here.

[0141] Perform data cleaning on the multi-source data after fusion processing; refer to the above embodiment for description and will not be repeated here.

[0142] Data feature extraction is performed on the cleaned multi-source data; this can be described with reference to the above embodiment and will not be repeated here.

[0143] Defect recognition is performed on the multi-source data after feature extraction based on preset defect recognition rules, and the operating status of various devices is determined according to the defect recognition results; please refer to the above embodiment for description and no further details will be given.

[0144] The preset defect recognition rules include a threshold-based defect recognition method and / or an AI model-based fine defect screening method. The above embodiments can be referred to for explanation and will not be repeated here.

[0145] Furthermore, the intelligent operation and maintenance method based on the digital twin model of the traction substation also includes:

[0146] The first target device and the second target device whose operating status detection result is abnormal are obtained; reference can be made to the above embodiment for description, which will not be repeated here.

[0147] Perform cause analysis and diagnosis on the first target device and the second target device; refer to the above embodiment for description and will not be repeated here.

[0148] A maintenance processing strategy is generated based on the cause analysis and diagnosis results, and it is verified according to the maintenance processing strategy whether the first target device and the second target device can be restored to a normal state; the above embodiment can be referred to for description and will not be repeated here.

[0149] The maintenance and treatment strategies corresponding to the equipment that can be restored to normal status are used as new fault cases, and the new fault cases are added to the training set to update the knowledge graph.

[0150] Furthermore, the associated device is dissected to obtain a planar representation of the internal structure of the associated device, including:

[0151] The geometric center points of the core components of the associated device are obtained, and explosions are simulated using the geometric center points as blasting source points to obtain a planar representation of the internal structure of the core components of the associated device.

[0152] Furthermore, the extracting image features of the internal structure includes:

[0153] The internal structure is marked according to the abnormal local area appearing when the historical fault of the associated device occurs, and the key area presented on the surface of the internal structure is obtained; the description can be made with reference to the above embodiment and will not be repeated here.

[0154] The image feature extraction of the key area can be referred to the above embodiment and will not be described in detail.

[0155] Furthermore, the image feature extraction of the internal structure further includes:

[0156] The image features of the internal structure are extracted by the bimodal attention network. Please refer to the above embodiment for details, which will not be repeated here.

[0157] Figure 10 This is a structural diagram of an intelligent operation and maintenance device based on a digital twin model of a traction substation provided by an embodiment of the present invention. Figure 10 As shown, the smart operation and maintenance device based on the digital twin model of the traction substation provided by the embodiment of the present invention includes a determination unit 1001, a dissection unit 1002 and a detection unit 1003, wherein:

[0158] The determination unit 1001 is used to detect the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device that has a connection relationship with the first target device is determined; wherein the digital twin model is a three-dimensional virtual model that is pre-established based on the physical equipment in the traction substation, corresponds to each physical device one-to-one and includes the connection relationship between the physical devices; the dissection unit 1002 is used to dissect the associated device if it is determined that the associated device is a first type of device with preset internal structural features, so as to obtain the internal structure of the associated device displayed in a planar manner; the detection unit 1003 is used to extract image features of the internal structure, and analyze the extracted image features to obtain the operating status detection result of the associated device.

[0159] Specifically, the determination unit 1001 in the device is used to detect the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device that has a connection relationship with the first target device is determined; wherein the digital twin model is a three-dimensional virtual model that is pre-established based on the physical equipment in the traction substation, corresponds to each physical device one-to-one and includes the connection relationship between the physical devices; the dissection unit 1002 is used to dissect the associated device if it is determined that the associated device is a first type of device with preset internal structural features, so as to obtain the internal structure of the associated device displayed in a planar manner; the detection unit 1003 is used to extract image features of the internal structure, and analyze the extracted image features to obtain the operating status detection result of the associated device.

[0160] An embodiment of the present invention provides an intelligent operation and maintenance device based on a digital twin model of a traction substation, which detects the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device having a connection relationship with the first target device is determined; wherein the digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one to each physical device and includes the connection relationship between the physical devices; if the associated device is determined to be a first type of device with preset internal structural features, the associated device is dissected to obtain the internal structure of the associated device displayed in a planar manner; image features are extracted from the internal structure, and the extracted image features are analyzed to obtain the operating status detection result of the associated device, which can comprehensively, effectively and timely monitor the operating status of the equipment in the traction substation, so that equipment with safety hazards in the traction substation can be discovered early.

[0161] The embodiment of the intelligent operation and maintenance device based on the digital twin model of the traction substation provided by the embodiment of the present invention can be specifically used to execute the processing flow of the above-mentioned method embodiments. Its functions will not be repeated here, and you can refer to the detailed description of the above-mentioned method embodiments.

[0162] Figure 11 A schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention is shown in FIG. Figure 11 As shown, the computer device includes: a memory 1101, a processor 1102, and a computer program stored in the memory 1101 and executable on the processor 1102. When the processor 1102 executes the computer program, the following method is implemented:

[0163] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0164] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0165] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0166] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0167] This embodiment discloses a computer program product, which includes a computer program. When the computer program is executed by a processor, the following method is implemented:

[0168] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0169] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0170] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0171] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0172] This embodiment provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the following method is implemented:

[0173] Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device;

[0174] The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices;

[0175] If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device;

[0176] Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

[0177] Compared with the technical solutions in the prior art, the embodiments of the present invention provide an intelligent operation and maintenance method based on the digital twin model of the traction substation, which detects the operating status of various devices in the digital twin model. If a first target device with an abnormal operating status is detected, the associated device with a connection relationship with the first target device is determined; wherein, the digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one to each physical device and includes the connection relationship between the physical devices; if it is determined that the associated device is a first type of device with preset internal structural characteristics, the associated device is dissected to obtain the internal structure of the associated device displayed in a planar manner; image features are extracted from the internal structure, and the extracted image features are analyzed to obtain the operating status detection results of the associated device, which can comprehensively, effectively and timely monitor the operating status of the equipment in the traction substation, so that equipment with safety hazards in the traction substation can be discovered early.

[0178] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0179] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0180] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0181] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0182] Throughout this specification, reference to terms such as "one embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations 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 any one or more embodiments or examples.

[0183] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A smart operation and maintenance method based on the digital twin model of a traction substation, characterized by: include: Detecting the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determining associated devices that have a connection relationship with the first target device; The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices; If it is determined that the associated device is a first type device having a preset internal structural feature, dissecting the associated device to obtain a planar representation of the internal structure of the associated device; Image features of the internal structure are extracted, and the extracted image features are analyzed to obtain a detection result of the operating status of the associated device.

2. The intelligent operation and maintenance method based on the digital twin model of the traction substation according to claim 1 is characterized in that: The detection of the operating status of various devices in the digital twin model includes: Performing fusion processing on multi-source data of the various devices; Perform data cleaning on the multi-source data after fusion processing; Extract data features from cleaned multi-source data; Based on the preset defect recognition rules, the multi-source data after feature extraction is used to identify defects, and the operating status of various equipment is determined according to the defect recognition results; Among them, the preset defect recognition rules include a threshold-based defect recognition method and / or an AI model-based fine screening defect method.

3. The intelligent operation and maintenance method based on the digital twin model of the traction substation according to claim 1 is characterized in that: The intelligent operation and maintenance method based on the digital twin model of the traction substation also includes: Acquire the first target device and the second target device whose operating status detection result is abnormal; Performing cause analysis and diagnosis on the first target device and the second target device; generating a maintenance processing strategy based on the cause analysis and diagnosis results, and verifying whether the first target device and the second target device can be restored to a normal state according to the maintenance processing strategy; The maintenance processing strategy corresponding to the equipment that can be restored to normal state is used as a new fault case, and the new fault case is added to the training set to update the knowledge graph.

4. The intelligent operation and maintenance method based on the digital twin model of the traction substation according to claim 1 is characterized in that: The associated device is dissected to obtain a planar display of the internal structure of the associated device, including: The geometric center points of the core components in the associated device are obtained, and explosions are simulated using the geometric center points as blasting source points to obtain a planar representation of the internal structure of the core components in the associated device.

5. The intelligent operation and maintenance method based on the digital twin model of the traction substation according to claim 1 is characterized in that: The extracting image features of the internal structure includes: Marking the internal structure according to abnormal local areas that appeared when historical faults of the associated device occurred, to obtain key areas presented on the surface of the internal structure; Image features are extracted from the key area.

6. The intelligent operation and maintenance method based on the digital twin model of the traction substation according to claim 5 is characterized in that: The image feature extraction of the internal structure further includes: Image features of the internal structure are extracted through a bimodal attention network.

7. An intelligent operation and maintenance device based on the digital twin model of a traction substation, characterized in that: include: A determination unit is configured to detect the operating status of various devices in the digital twin model, and if a first target device with an abnormal operating status is detected, determine an associated device that has a connection relationship with the first target device; The digital twin model is a three-dimensional virtual model pre-established based on the physical devices in the traction substation, which corresponds one-to-one with each physical device and includes the connection relationship between the physical devices; a dissection unit, configured to dissect the associated device to obtain a planar representation of the internal structure of the associated device if it is determined that the associated device is a first type device having a preset internal structure feature; The detection unit is used to extract image features of the internal structure and analyze the extracted image features to obtain a detection result of the operating status of the associated device.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images