Power plant intelligent fault monitoring and early warning system and method based on virtual reality enhancement

Abstract

The application discloses a power plant intelligent fault monitoring and early warning system and method based on virtual reality enhancement, relates to the technical field of power plant monitoring and virtual reality, and comprises the following: a data acquisition and edge computing layer, which is used for acquiring real-time operation data of physical equipment of the power plant; a digital twin modeling and mapping module, which is used for constructing a static digital twin model corresponding to the power plant and synchronously updating a state; an intelligent fault diagnosis module, which is used for analyzing the health state of the physical equipment of the power plant through a preset algorithm model and generating early warning information based on the fact that the health state exceeds a target health state; a VR enhanced interaction module, which is used for loading and rendering the digital twin model to form an immersive virtual scene and performing visual enhancement marking; and a collaborative management platform, which is used for generating maintenance instructions based on the early warning information and managing and collaboratively working on the virtual scene accessed by multiple users through the VR enhanced interaction module.

Description

Technical Field

[0001] This application relates to the technical fields of power plant monitoring and virtual reality, and in particular to a power plant intelligent fault monitoring and early warning system and method based on virtual reality enhancement. Background Technology

[0002] Power plants have dense equipment and complex pipelines, requiring maintenance personnel to monitor and troubleshoot a large number of devices in real time.

[0003] The existing technology has the following main drawbacks: 1. Lack of spatial information in two-dimensional monitoring interfaces: Traditional distributed control systems (DCS) use two-dimensional planar interfaces to display data. Equipment parameters are presented in numerical or simple chart form, resulting in scattered data displays and a lack of correspondence with physical space. In emergency situations, maintenance personnel find it difficult to quickly locate the actual physical location of the fault point within a complex equipment layout.

[0004] 2. Lack of immersive information presentation: Existing monitoring systems separate data from spatial location, requiring maintenance personnel to mentally map and convert "data-device-location," increasing cognitive load and affecting emergency response efficiency.

[0005] 3. Insufficient intelligence in 3D simulation systems: Most existing 3D simulation systems are for offline display or only for training purposes, lacking deep integration with real-time IoT data, and are unable to perform intelligent diagnosis and proactive early warning based on real-time data.

[0006] 4. Limited display of fault information: Traditional systems only use alarm lights or pop-up windows to indicate faults, lacking intuitive visual enhancement methods and making it difficult to quickly convey the severity, location, and cause of the fault.

[0007] 5. Weak collaborative troubleshooting capabilities: Traditional systems do not support collaborative consultations among multiple people in a virtual space, making it difficult for remote experts to provide intuitive guidance to on-site personnel in troubleshooting.

[0008] Therefore, there is an urgent need for a solution that can overcome the above-mentioned shortcomings. Summary of the Invention

[0009] This application proposes a power plant intelligent fault monitoring and early warning system and method based on virtual reality enhancement, which is used to solve the defects of the above-mentioned prior art.

[0010] According to a first aspect of the embodiments of this application, a power plant intelligent fault monitoring and early warning system based on virtual reality enhancement is provided, comprising: The data acquisition and edge computing layer is deployed on the physical equipment of the power plant to collect real-time operating data of the physical equipment and to preprocess and extract features from the real-time operating data. The digital twin modeling and mapping module is connected to the data acquisition and edge computing layer. It is used to construct a static digital twin model corresponding to the power plant and drive the virtual devices in the static digital twin model to update their status synchronously based on the real-time operation data. The intelligent fault diagnosis module is connected to the data acquisition and edge computing layer. It is used to analyze the health status of the power plant's physical equipment based on the real-time operating data and through a preset algorithm model, and to generate early warning information when the health status exceeds the target. The VR enhanced interaction module is connected to the digital twin modeling and mapping module and the intelligent fault diagnosis module, respectively, and is used to load and render the digital twin model to form an immersive virtual scene, and to perform visual enhancement marking on the corresponding devices in the immersive virtual scene according to the warning information. The collaborative management platform, connected to the intelligent fault diagnosis module and the VR enhanced interaction module, is used to generate maintenance instructions based on the early warning information and to manage and coordinate operations for multiple users accessing the virtual scene through the VR enhanced interaction module.

[0011] In some implementations, the VR enhanced interaction module performs visual enhancement marking on the corresponding device in the immersive virtual scene based on the warning information, including: The VR enhanced interaction module displays dynamically highlighted visual identifiers around power plant physical equipment with abnormal health status based on the immersive virtual scene, and overlays text labels containing fault probability, fault cause or suggested measures at the associated positions of the power plant physical equipment with abnormal health status.

[0012] In some implementations, the VR enhanced interaction module is also used to provide a see-through function, including: The VR enhanced interaction module controls the outer shell of the marked power plant physical equipment to be transparent or semi-transparent when the user approaches the marked power plant physical equipment in the immersive virtual scene, in order to display the internal structure of the marked power plant physical equipment or simulate the diffusion path of a leaking medium.

[0013] In some implementations, the preset algorithm model in the intelligent fault diagnosis module is a hybrid neural network model that combines a convolutional neural network based on an attention mechanism with a long short-term memory network. This model is used to analyze the time-frequency domain characteristics of vibration signals and to perform fault diagnosis and prediction of remaining service life.

[0014] In some implementations, a voice alarm module is also included. The voice alarm module is connected to the intelligent fault diagnosis module and is used to trigger an alarm sound effect with three-dimensional spatial orientation attributes in the immersive virtual scene constructed by the VR enhanced interaction module when the intelligent fault diagnosis module determines that there is an emergency fault, and to locate the physical location of the fault.

[0015] In some implementations, the digital twin modeling and mapping module is used to construct a static geometric model of the power plant at a 1:1 scale based on building information model design drawings and / or 3D laser scanning point cloud data.

[0016] In some implementations, the collaborative management platform supports multiple users to log in to the immersive virtual scene simultaneously as virtual avatars, and supports voice calls, view sharing, annotation and instruction using virtual tools, and remote collaborative consultations within the immersive virtual scene.

[0017] According to a second aspect of this application, a method for intelligent fault monitoring and early warning in power plants based on virtual reality enhancement is provided, implemented based on the aforementioned intelligent fault monitoring and early warning system for power plants based on virtual reality enhancement, comprising: Real-time operating data of power plant physical equipment is collected and preprocessed through the data acquisition and edge computing layer; The preprocessed real-time running data is synchronized to the digital twin model constructed based on the digital twin modeling and mapping module, and the virtual devices of the digital twin model are driven to perform state synchronization updates. The intelligent fault diagnosis module analyzes the real-time operating data, generates equipment health status information, and outputs early warning information when an abnormality is detected. The VR enhanced interaction module receives the warning information in an immersive virtual scene and visually enhances the marking of the corresponding virtual devices. In response to user interactions in the virtual environment, historical data and records of associated power plant physical equipment are retrieved.

[0018] In some implementations, the step of visually enhancing the labeling of the corresponding virtual device includes: When the warning information indicates an emergency fault, an alarm sound effect with three-dimensional spatial orientation attributes is triggered in the immersive virtual scene to guide the operator to the location of the fault.

[0019] In some implementations, it also includes: The collaborative management platform generates maintenance work orders based on the early warning information and provides multi-user collaborative consultation functions when multiple users remotely access the immersive virtual scene.

[0020] The beneficial effects of the virtual reality-enhanced intelligent fault monitoring and early warning system and method for power plants in this application include at least the following: This application's embodiments address the shortcomings of traditional two-dimensional monitoring systems, such as missing spatial information, delayed fault location, and weak collaborative capabilities, by deploying a data acquisition and edge computing layer, a digital twin modeling and mapping module, an intelligent fault diagnosis module, a VR (virtual reality) enhanced interaction module, and a collaborative management platform. It achieves deep integration and dynamic mapping of the power plant's physical space and operational data, enabling maintenance personnel to intuitively and quickly locate faulty equipment and its physical location in an immersive, interactive virtual environment. Through early warning from the intelligent diagnosis module and enhanced visualization marking from the VR module, the difficulty of fault recognition and judgment is significantly reduced, and emergency response time is shortened. Simultaneously, the collaborative management platform supports remote intervention from multiple parties in the same virtual environment, allowing experts to provide intuitive guidance without being physically present on-site, thereby improving the efficiency of collaborative handling of complex faults and the effectiveness of maintenance training. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the structure of a power plant intelligent fault monitoring and early warning system based on virtual reality enhancement, according to an embodiment of this application. Figure 2 This is a schematic diagram of the CNN-LSTM attention network structure of an intelligent fault diagnosis module according to an embodiment of this application; Figure 3 This is a schematic diagram of a VR enhanced interaction model according to an embodiment of this application; Figure 4 This is a flowchart illustrating a power plant intelligent fault monitoring and early warning method based on virtual reality enhancement, according to an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the various embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this disclosure to facilitate a better understanding of the disclosure. However, the technical solutions claimed in this disclosure can be implemented even without these technical details and with various variations and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of this disclosure. The various embodiments can be combined with and referenced by each other without contradiction.

[0023] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed embodiments of the present application, but merely to illustrate selected embodiments of the present application. Other embodiments obtained by those skilled in the art based on the embodiments of the present application without inventive effort are all within the scope of protection of the embodiments of the present application.

[0024] This application discloses a power plant intelligent fault monitoring and early warning system and method based on virtual reality enhancement. The method is implemented based on the same system and aims to achieve immersive monitoring, intelligent early warning, and collaborative management of power plant equipment by deeply integrating real-time data acquisition, intelligent fault diagnosis, and virtual reality technology. This solves the problems of lack of spatial immersion, missing spatial information, delayed early warning information, and weak collaborative maintenance capabilities in existing power plant monitoring technologies.

[0025] See attached document Figure 1 The diagram illustrates a schematic representation of a power plant intelligent fault monitoring and early warning system based on virtual reality enhancement, according to an embodiment of this application. This system includes: a data acquisition and edge computing layer, a digital twin modeling and mapping module, an intelligent fault diagnosis module, a VR (virtual reality) enhanced interaction module, and a collaborative management platform.

[0026] In some implementations, the data acquisition and edge computing layer is deployed on the power plant physical equipment (hereinafter referred to as the equipment) to collect the real-time operating data of the power plant physical equipment and to preprocess and extract features from the real-time operating data.

[0027] For example, the data acquisition and edge computing layer includes IoT sensors. This layer is deployed on the physical equipment of the power plant, including IoT sensors, electrical parameter acquisition modules, and remote terminal units (RTUs). The IoT sensors include: vibration sensors deployed on rotating equipment (coal mills, fans, pumps, etc.) to collect vibration acceleration, velocity, and displacement signals; and temperature sensors deployed at key temperature measurement points such as bearings, motor windings, and pipe walls. Additionally, the electrical parameter acquisition modules are deployed in switchgear and motor control centers to collect parameters such as current, voltage, and power. The remote terminal units are deployed in the distributed control system (DCS) to collect existing process parameters, such as pressure, flow rate, and liquid level.

[0028] For example, preprocessing and feature extraction of the real-time operational data includes: preprocessing and feature extraction of the real-time operational data based on the deployment of edge computing nodes at the equipment site or in a nearby area. This preprocessing and feature extraction includes: protocol parsing, outlier removal, data cleaning, and feature extraction. Specifically, protocol parsing is used to parse various industrial communication protocols such as Modbus, OPC UA, and Profibus; outlier removal is used to identify and remove abnormal data caused by sensor failures or communication interruptions; data cleaning is used to interpolate missing values ​​and filter noisy data; and feature extraction is used to extract time-domain features (mean, variance, peak value) and frequency-domain features (FFT spectrum, power spectral density) to reduce data transmission volume.

[0029] In some implementations, the digital twin modeling and mapping module is connected to the data acquisition and edge computing layer to construct a static digital twin model corresponding to the power plant, and to drive the virtual devices in the static digital twin model to update their status synchronously based on the real-time operating data.

[0030] In some implementations, the digital twin modeling and mapping module is used to construct a static geometric model of the power plant at a 1:1 scale, based on Building Information Modeling (BIM) design drawings and / or 3D laser scanning point cloud data. The BIM design drawings are used to extract the 3D geometric dimensions and spatial location information of buildings, equipment, and pipelines; the 3D laser scanning point cloud data is used to perform laser scanning on the existing power plant to obtain high-precision spatial point cloud data and correct discrepancies between the drawings and the actual structure.

[0031] For example, the static geometric model is also built based on texture maps, which are real texture photos of the surface of power plant physical equipment. By mapping the texture maps onto the 3D model, the visual realism is enhanced.

[0032] For example, the virtual devices in the static digital twin model are synchronously updated based on the real-time operating data, including: accessing the real-time data from the data acquisition layer into the digital twin model, and performing device status refresh, parameter overlay display, and color status mapping. Specifically, device status refresh includes: displaying the running or stopped state of the virtual device in the model according to its start / stop status (e.g., fan blade rotation, valve opening change); parameter overlay display includes: displaying key parameters such as real-time temperature, vibration, and current floating above the virtual device; and color status mapping includes: presenting different colors (green for normal, yellow for caution, red for alarm) in the device model according to its health index.

[0033] In some implementations, the intelligent fault diagnosis module is connected to the data acquisition and edge computing layer to analyze the health status of the power plant's physical equipment based on the real-time operating data using a preset algorithm model and to generate early warning information when the health status exceeds the target.

[0034] In some implementations, the preset algorithm model in the intelligent fault diagnosis module is a hybrid neural network model combining a convolutional neural network (CNN) based on an attention mechanism and a long short-term memory network (LSTM), which is used to analyze the time-frequency domain characteristics of vibration signals and perform fault diagnosis and prediction of remaining service life.

[0035] For example, refer to the appendix. Figure 2 As shown, the preset algorithm model in this intelligent fault diagnosis module is a CNN-LSTM hybrid neural network based on an attention mechanism. This preset algorithm model includes: an input layer, a convolutional layer (with a convolutional neural network), a long short-term memory layer (with a long short-term memory network), an attention mechanism layer, a fully connected layer, and an output layer. The input layer receives multi-source sensor data, including vibration time-domain signals, temperature sequences, and current sequences. The convolutional layer extracts local feature patterns and captures short-term fluctuation features in the sensor data. The long short-term memory layer captures long-term dependencies in time-series data and learns fault evolution patterns. The attention mechanism layer automatically weights the importance of different time steps and different sensor channels, focusing on fault-sensitive features. The fully connected layer and output layer output the device's health index (0-100%) and the probability distribution of potential fault types, and can predict the remaining useful life (RUL).

[0036] In some implementations, before analyzing the health status of power plant physical equipment using a preset algorithm model and generating early warning information based on exceeding the target health status, the preset algorithm model is trained. This model training includes: data training, label annotation training, and loss function training. Specifically, data training includes training on historical normal operation data and historical fault case data (including fault type, fault time, and fault severity labels); label annotation training includes training on a health index composed of comprehensive labels from actual equipment operating time, historical maintenance records, and fault cases; and loss function training includes using mean squared error (MSE) for health index regression training and cross-entropy loss for fault classification training.

[0037] In some implementations, after analyzing the health status of the power plant's physical equipment using a preset algorithm model and generating early warning information based on exceeding the target health status, diagnostic output is included. This diagnostic output includes: an equipment health index, a 0-100% quantified value, with an early warning triggered when it falls below a threshold; potential fault types, such as "coal mill roller wear," "fan bearing failure," and "valve jamming"; fault probability: the probability of occurrence for each fault type; and remaining service life: the remaining usable time predicted based on the current degradation trend.

[0038] In some implementations, the VR enhanced interaction module is connected to the digital twin modeling and mapping module and the intelligent fault diagnosis module, respectively, to load and render the digital twin model to form an immersive virtual scene, and to perform visual enhancement marking on the corresponding devices in the immersive virtual scene according to the warning information.

[0039] In some implementations, the VR enhanced interaction module performs visual enhancement marking on the corresponding equipment in the immersive virtual scene based on the warning information, including: the VR enhanced interaction module displays dynamically highlighted visual marks on the periphery of power plant physical equipment with abnormal health status based on the immersive virtual scene, and overlays text labels containing fault probability, fault cause or suggested measures at the associated location of the power plant physical equipment with abnormal health status.

[0040] For example, the visualization enhancement markers are based on the output of the intelligent fault diagnosis module to visualize and enhance the marking of fault-risk devices, including: dynamic high-brightness halo markers, which display a dynamic high-brightness halo around devices with a health index below a threshold, and the color of the halo changes with the severity of the fault (yellow → orange → red); fault information label markers, which display real-time fault probability, fault cause, and suggested measures above the device; and trend indicator markers, which add warning arrows or flashing effects for devices with a rapidly declining health index.

[0041] In some implementations, the VR augmented interaction module is also used to provide a see-through function, including: the VR augmented interaction module controls the shell of the marked power plant physical equipment to be transparent or semi-transparent when the user approaches the marked power plant physical equipment in the immersive virtual scene, in order to show the internal structure of the marked power plant physical equipment or simulate the diffusion path of a media leak.

[0042] For example, the perspective function is automatically triggered when an operator approaches a high-risk equipment area in a virtual scene. This includes: the equipment casing automatically becoming semi-transparent or hidden to reveal the internal structure; for current leakage faults, using red particle flow to simulate the current leakage path; and for media leakage faults (such as steam, water, and oil), using dynamic particle effects to simulate the leakage diffusion direction to help determine the location of the leakage source.

[0043] In some implementations, the hardware configuration of the VR enhanced interaction module includes: a VR headset, using a high-resolution headset (such as HTC Vive, Pico 4 Enterprise, etc.), supporting 6DoF spatial positioning; interactive controllers, dual controllers supporting spatial positioning, used to interact with devices in the virtual scene; and positioning base stations, deployed in the power plant control room or inspection area to achieve centimeter-level spatial positioning.

[0044] In some implementations, immersive virtual scene rendering includes: importing a digital twin model into a VR engine (such as Unity3D or Unreal Engine) to construct an immersive virtual power plant scene; allowing users to freely roam and observe the equipment layout from any angle within the immersive virtual scene; and synchronously updating the equipment status and real-time data within the immersive virtual scene.

[0045] In some implementations, maintenance personnel interact with virtual devices using VR controllers, including: clicking on the device to access its detailed data panel, which includes real-time parameter curves, historical trends, maintenance records, and device files; gesture operations, allowing users to zoom in and rotate the device model to observe device details from multiple angles; and voice input, supporting voice queries for device information and recording of inspection findings.

[0046] In some implementations, a voice alarm module is also included. The voice alarm module is connected to the intelligent fault diagnosis module. When the intelligent fault diagnosis module determines that there is an emergency fault, it triggers an alarm sound effect with three-dimensional spatial orientation attributes in the immersive virtual scene constructed by the VR enhanced interaction module, and locates the physical location of the fault.

[0047] For example, the spatial stereo sound alarm is triggered when the system determines that an emergency fault exists. This includes: the alarm sound is spatially located based on the physical location of the faulty equipment; the operator can perceive the approximate direction of the fault through hearing and guide their vision to the location of the fault; different fault levels correspond to different alarm timbres and frequencies.

[0048] In some implementations, the collaborative management platform is connected to the intelligent fault diagnosis module and the VR enhanced interaction module to generate maintenance instructions based on the early warning information, and to manage and coordinate multiple users accessing the virtual scene through the VR enhanced interaction module.

[0049] In some implementations, the collaborative management platform supports multiple users to log in to the immersive virtual scene simultaneously as virtual avatars, and supports voice calls, view sharing, annotation and instruction using virtual tools, and remote collaborative consultations within the immersive virtual scene.

[0050] In some implementations, the maintenance work order generation process includes: when the intelligent fault diagnosis module determines that the equipment has a fault risk, an maintenance work order is automatically generated on the collaborative management platform, and the work order is automatically pushed to the mobile terminal of relevant maintenance personnel or the to-do list panel in the VR scene. The work order content includes at least: equipment name, location, fault type, health index, suggested handling measures, and urgency level.

[0051] In some implementations, the multi-user remote collaborative consultation function includes: multiple users simultaneously logging into the virtual power plant scenario, with each user appearing as a virtual avatar; remote specialists logging into the VR system from different locations to conduct consultations with on-site personnel in the same virtual scenario; enabling at least voice calls, screen sharing, and virtual laser pointer indication functions; experts marking fault points and drawing maintenance paths in the virtual scenario, which on-site personnel can view in real time.

[0052] In some implementations, the historical data tracing function includes: supporting the playback of historical scenarios by time to view the changes in equipment status before the failure occurred; and supporting the management of a failure case library to save typical failure VR scenarios as cases for training and new employee drills.

[0053] This application's embodiments address the shortcomings of traditional two-dimensional monitoring systems, such as missing spatial information, delayed fault location, and weak collaborative capabilities, by deploying a data acquisition and edge computing layer, a digital twin modeling and mapping module, an intelligent fault diagnosis module, a VR (virtual reality) enhanced interaction module, and a collaborative management platform. It achieves deep integration and dynamic mapping of the power plant's physical space and operational data, enabling maintenance personnel to intuitively and quickly locate faulty equipment and its physical location in an immersive, interactive virtual environment. Through early warning from the intelligent diagnosis module and enhanced visualization marking from the VR module, the difficulty of fault recognition and judgment is significantly reduced, and emergency response time is shortened. Simultaneously, the collaborative management platform supports remote intervention from multiple parties in the same virtual environment, allowing experts to provide intuitive guidance without being physically present on-site, thereby improving the efficiency of collaborative handling of complex faults and the effectiveness of maintenance training.

[0054] See attached document Figure 4 As shown, according to the second aspect of this application, a method for intelligent fault monitoring and early warning of power plants based on virtual reality enhancement is provided. The method is implemented based on the above-mentioned intelligent fault monitoring and early warning system for power plants based on virtual reality enhancement, and includes the following steps S1 to S5.

[0055] Step S1: Collect and preprocess real-time operating data of power plant physical equipment through data acquisition and edge computing layer.

[0056] Step S2: The preprocessed real-time running data is synchronized to the digital twin model built based on the digital twin modeling and mapping module, and the virtual device of the digital twin model is driven to perform state synchronization update.

[0057] Step S3: Analyze the real-time operating data through the intelligent fault diagnosis module, generate equipment health status information, and output early warning information when an abnormality is detected.

[0058] Step S4: Receive the warning information in an immersive virtual scene through the VR enhanced interaction module, and visually enhance the marking of the corresponding virtual device.

[0059] Step S5: In response to the user's interactive operations in the virtual scene, retrieve the historical data and records of the associated power plant physical equipment.

[0060] In some implementations, the corresponding virtual device is marked with enhanced visualization, including: when the warning information indicates an emergency fault, triggering an alarm sound effect with three-dimensional spatial orientation attributes in the immersive virtual scene to guide the operator to the location of the fault.

[0061] In some implementations, the system also includes: generating maintenance work orders based on the early warning information through a collaborative management platform, and providing multi-user collaborative consultation functions when multiple users remotely access the immersive virtual scene.

[0062] In a specific implementation process, refer to the appendix Figure 3 As shown, this intelligent fault monitoring and early warning method for power plants based on virtual reality enhancement includes: real-time collection of operational data such as vibration, temperature, and electrical parameters of the equipment through IoT sensors, smart meters, and RTU acquisition terminals deployed on the power plant equipment; edge computing nodes performing protocol parsing, outlier removal, and data cleaning on the collected data, extracting key features, and uploading them to the central server; synchronizing the processed real-time data to the digital twin model to drive the virtual equipment in the virtual power plant to perform dynamic actions and status updates, achieving synchronized operation between the virtual and real systems; and an intelligent fault diagnosis module calculating the equipment health index based on the real-time data, which, when lower than a preset value, triggers an early warning system. When a threshold is reached or a fault characteristic is detected, an early warning message is generated, including the fault type, fault probability, and severity level. The VR enhanced interaction module receives the early warning message and visually enhances the marking of the corresponding faulty equipment in the virtual scene (dynamic halo, fault label, particle effects, etc.), and guides the operator to the fault location through 3D spatial stereo sound effects. Maintenance personnel can click on the virtual equipment with VR controllers to retrieve detailed historical trend curves and maintenance records. If necessary, remote collaborative consultations can be initiated to jointly develop maintenance plans with experts. The collaborative management platform automatically generates maintenance work orders. After maintenance is completed, maintenance personnel update the equipment status in the system to complete the work order closure.

[0063] This application's embodiments utilize VR immersive scenes and enhanced visualization markers, eliminating the need for maintenance personnel to mentally map data, equipment, and locations. They can directly locate faulty equipment in virtual space, reducing location time by over 60%, making spatial positioning intuitive and efficient. The combined use of dynamic high-brightness halos, particle effects, and 3D spatial sound effects makes fault information transmission more intuitive and rapid, significantly shortening emergency response time and improving emergency response speed. Based on a perspective function, this application's embodiments allow maintenance personnel to intuitively observe the internal structure and leakage paths of equipment, reducing uncertainty in fault diagnosis and improving accuracy. The multi-user remote collaborative consultation function allows experts to guide fault diagnosis without on-site visits, saving travel time and costs, improving fault handling efficiency, and significantly enhancing collaborative efficiency. Based on real data, VR scenes can be used for maintenance personnel training, supporting historical fault case playback and drills, enhancing the practicality and effectiveness of training, and improving training outcomes. This application's embodiments support historical data traceability and work order closed-loop management, providing data support for the entire equipment lifecycle management and achieving full lifecycle management.

[0064] To further demonstrate the effects of the embodiments of this application, the specific effects of the embodiments of this application are shown below based on Embodiments 1 to 3.

[0065] Example 1: This system was deployed in a 600MW power plant to achieve early warning of coal mill bearing failures and VR interaction. The implementation process includes: data acquisition, edge computing, digital twin, intelligent diagnosis, VR enhanced interaction, collaborative consultation, and work order closed loop. Data acquisition includes: installing vibration and temperature sensors on the bearing housings at both the drive and non-drive ends of the coal mill to collect vibration acceleration, velocity, displacement, and bearing temperature signals at a sampling frequency of 10kHz. The system is connected to edge computing nodes via the Modbus TCP protocol. Edge computing includes: edge nodes calculating the effective value (RMS), peak value, and kurtosis of the vibration signals in real time, extracting the power spectral density characteristics of the 0-500Hz frequency band, and uploading the data to the central server. Digital twin includes: constructing a 1:1 3D model of the coal mill area based on the power plant's BIM drawings and laser point clouds, including the coal mill body, motor, reducer, lubrication station, and surrounding pipelines. The intelligent diagnostics include: real-time analysis of vibration and temperature data using a CNN-LSTM-attention network. When the effective value of the bearing vibration at the coal mill drive end continues to rise, and the frequency domain characteristics show sideband modulation in the 100-150Hz frequency band, the model output health index drops from 92% to 68%, determining an 85% probability of "bearing wear" and triggering a level three alarm. The VR-enhanced interaction includes: a dynamic orange high-brightness halo appearing around the bearing location at the coal mill drive end in the VR headset; a fault label displaying "Bearing Wear Probability 85%, Health Index 68%, Inspection Recommended" floating above the equipment; automatic triggering of the "see-through" function when the operator approaches the equipment, making the bearing housing semi-transparent, revealing the internal rolling elements and cage, with worn areas highlighted in red; and continuous warning sounds emanating from the direction of the coal mill drive end, guiding the operator to the fault location. The collaborative consultation includes: on-site operators clicking on the faulty equipment with VR controllers to retrieve the bearing vibration trend curve and maintenance history. Remote experts join the session via a VR system, marking key areas requiring inspection in a virtual environment. On-site operators simultaneously view and perform these inspections. The work order loop includes: the system automatically generating a maintenance work order and pushing it to the maintenance team. After maintenance is completed, the operator updates the equipment status in the VR environment; the health index recovers to 95%, completing the work order loop.

[0066] Example 2: This system was deployed in the boiler area of ​​a 1000MW ultra-supercritical unit to achieve steam pipeline leak detection and location. The implementation process includes: data acquisition, edge computing, intelligent diagnosis, VR-enhanced interaction, and collaborative handling. Data acquisition includes deploying acoustic and temperature sensor arrays in the boiler superheater and main steam pipeline areas to collect ultrasonic signals (20-100kHz) and temperature distribution changes generated by pipeline leaks. Edge computing includes edge nodes calculating the spectral characteristics of acoustic signals in real time and identifying energy changes in the characteristic frequency band of the leak. Intelligent diagnosis includes a deep learning model analyzing acoustic signals and temperature field data. When a sudden increase in energy in the 20-40kHz frequency band is detected, and a local abnormal drop in the temperature field is observed, a "steam leak" probability of 97% is output, triggering an emergency alarm. The VR-enhanced interactive features include: a dynamic, bright red halo and flashing warning light appearing at the location of the leak in the VR scene; an automatically triggered "see-through" function that makes the pipe casing semi-transparent, with red particle streams spraying out from the leak point, simulating the diffusion path and direction of a steam leak; 3D spatial stereo sound effects emanating from the direction of the leak, alerting operators to stay away from the danger zone; and operators can trace the leak source along the particle stream direction in the virtual scene to quickly locate the leak point. The collaborative response features include: remote safety experts joining the VR session to draw safety isolation zones and evacuation routes in the virtual scene, with on-site personnel following the instructions to carry out emergency response.

[0067] Example 3: This system was deployed during the commissioning phase of a newly built power plant to enable multi-user remote collaborative maintenance drills. The implementation process included scenario construction, training and drills, and remote collaboration. Scenario construction included building a 1:1 digital twin model of the entire plant based on BIM drawings, including all equipment such as boilers, turbines, generators, and auxiliary systems. Training and drills included newly hired maintenance personnel wearing VR headsets to conduct equipment inspections and troubleshooting drills in a virtual environment. The system replayed historical fault cases, with virtual equipment reproducing the fault occurrence process according to historical data. Remote collaboration included training experts logging into the VR system remotely and appearing as virtual avatars in the trainees' virtual environment. Experts could use laser pointers to point at equipment, explain fault characteristics via voice, and demonstrate standard operating procedures. Trainees observed and learned simultaneously, significantly improving training effectiveness.

[0068] The above description is merely a specific implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any simple modifications, changes, and equivalent changes made by those skilled in the art to the above embodiments based on the technical essence of the embodiments of this application within the technical scope disclosed in the embodiments of this application shall still fall within the protection scope of the technical solution of the embodiments of this application.

Claims

1. A power plant intelligent fault monitoring and early warning system based on virtual reality enhancement, characterized in that, include: The data acquisition and edge computing layer is deployed on the physical equipment of the power plant to collect real-time operating data of the physical equipment and to preprocess and extract features from the real-time operating data. The digital twin modeling and mapping module is connected to the data acquisition and edge computing layer. It is used to construct a static digital twin model corresponding to the power plant and drive the virtual devices in the static digital twin model to update their status synchronously based on the real-time operation data. The intelligent fault diagnosis module is connected to the data acquisition and edge computing layer. It is used to analyze the health status of the power plant's physical equipment based on the real-time operating data and through a preset algorithm model, and to generate early warning information when the health status exceeds the target. The VR enhanced interaction module is connected to the digital twin modeling and mapping module and the intelligent fault diagnosis module, respectively, and is used to load and render the digital twin model to form an immersive virtual scene, and to perform visual enhancement marking on the corresponding devices in the immersive virtual scene according to the warning information. The collaborative management platform, connected to the intelligent fault diagnosis module and the VR enhanced interaction module, is used to generate maintenance instructions based on the early warning information and to manage and coordinate operations for multiple users accessing the virtual scene through the VR enhanced interaction module.

2. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement as described in claim 1, characterized in that, The VR enhanced interaction module performs visual enhancement marking on the corresponding devices in the immersive virtual scene based on the warning information, including: The VR enhanced interaction module displays dynamically highlighted visual identifiers around power plant physical equipment with abnormal health status based on the immersive virtual scene, and overlays text labels containing fault probability, fault cause or suggested measures at the associated positions of the power plant physical equipment with abnormal health status.

3. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement according to claim 1, characterized in that, The VR enhanced interaction module is also used to provide perspective functionality, including: The VR enhanced interaction module controls the outer shell of the marked power plant physical equipment to be transparent or semi-transparent when the user approaches the marked power plant physical equipment in the immersive virtual scene, in order to display the internal structure of the marked power plant physical equipment or simulate the diffusion path of a leaking medium.

4. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement as described in claim 1, characterized in that, The preset algorithm model in the intelligent fault diagnosis module is a hybrid neural network model that combines a convolutional neural network based on an attention mechanism with a long short-term memory network. It is used to analyze the time-frequency domain characteristics of vibration signals and perform fault diagnosis and prediction of remaining service life.

5. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement according to claim 1, characterized in that, It also includes a voice alarm module. The voice alarm module is connected to the intelligent fault diagnosis module and is used to trigger an alarm sound effect with three-dimensional spatial orientation attributes in the immersive virtual scene constructed by the VR enhanced interaction module when the intelligent fault diagnosis module determines that there is an emergency fault, and to locate the physical location of the fault.

6. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement according to claim 1, characterized in that, The digital twin modeling and mapping module is used to construct a static geometric model of the power plant at a 1:1 scale based on building information model design drawings and / or 3D laser scanning point cloud data.

7. The power plant intelligent fault monitoring and early warning system based on virtual reality enhancement according to claim 1, characterized in that, The collaborative management platform supports multiple users to log in to the immersive virtual scene simultaneously as virtual avatars, and supports voice calls, view sharing, annotation and instruction using virtual tools, and remote collaborative consultations within the immersive virtual scene.

8. A method for intelligent fault monitoring and early warning in power plants based on virtual reality enhancement, implemented based on the intelligent fault monitoring and early warning system for power plants based on virtual reality enhancement as described in any one of claims 1 to 7, characterized in that, include: Real-time operating data of power plant physical equipment is collected and preprocessed through the data acquisition and edge computing layer; The preprocessed real-time running data is synchronized to the digital twin model constructed based on the digital twin modeling and mapping module, and the virtual devices of the digital twin model are driven to perform state synchronization updates. The intelligent fault diagnosis module analyzes the real-time operating data, generates equipment health status information, and outputs early warning information when an abnormality is detected. The VR enhanced interaction module receives the warning information in an immersive virtual scene and visually enhances the marking of the corresponding virtual devices. In response to user interactions in the virtual environment, historical data and records of associated power plant physical equipment are retrieved.

9. The intelligent fault monitoring and early warning method for power plants based on virtual reality enhancement as described in claim 8, characterized in that, The step of visually enhancing and marking the corresponding virtual device includes: When the warning information indicates an emergency fault, an alarm sound effect with three-dimensional spatial orientation attributes is triggered in the immersive virtual scene to guide the operator to the location of the fault.

10. The intelligent fault monitoring and early warning method for power plants based on virtual reality enhancement according to claim 8, characterized in that, Also includes: The collaborative management platform generates maintenance work orders based on the early warning information and provides multi-user collaborative consultation functions when multiple users remotely access the immersive virtual scene.

Images