Photovoltaic station virtual monitoring method based on digital twinborn technology

Abstract

The invention provides a photovoltaic station virtual monitoring method and system based on a digital twinning technology, and virtual monitoring and alarming of a photovoltaic station are realized by using the digital twinning technology. Through the virtual monitoring platform, operation and maintenance personnel can grasp the operation state of the power station in real time, discover and process abnormal conditions in time, and improve the operation and maintenance efficiency. Through virtual monitoring and alarming, the on-site inspection frequency and manual intervention can be reduced, and the operation and maintenance cost is reduced. Momo AI is used for monitoring alarm analysis management of the virtual model, comprehensive and real-time monitoring of various elements and behaviors in the virtual environment can be achieved, and an alarm is given out in time when abnormity or potential risks are detected. Momo AI is used for monitoring and alarming analysis management of the virtual model, efficient, accurate and real-time monitoring and alarming functions can be achieved, powerful guarantee is provided for safety and stability of the virtual environment, and intelligent, virtualized and unmanned monitoring of the photovoltaic power station is achieved.

Description

Technical Field

[0001] The present invention relates to the field of digital twin technology, and in particular to a photovoltaic station virtual monitoring method based on digital twin technology, a photovoltaic station virtual monitoring system based on digital twin technology, an electronic device and a computer-readable storage medium. Background Art

[0002] Photovoltaic power station monitoring is to connect the inverter, combiner box, irradiator, meteorological instrument, electric meter and other equipment of the photovoltaic power station through data cables, use the photovoltaic power station data collector to collect data from these devices, and upload them to the network server or local computer through GPRS, Ethernet, WIFI and other methods, so that users can view relevant data on the Internet or local computer, which is convenient for power station managers and users to view and manage the operation data of the photovoltaic power station.

[0003] Traditional methods of monitoring photovoltaic power plants mainly include manual inspections and setting up servers on site for monitoring. Although they can achieve certain effects of photovoltaic power plant monitoring, each has its own shortcomings: 1. The disadvantage of manual inspection is that it relies on a large amount of human resources. Especially for large photovoltaic power plants, inspection and maintenance work is heavy and time-consuming, which increases the operating costs of enterprises and limits the improvement of operation and maintenance efficiency. In addition, manual inspections are difficult to detect potential equipment failures in real time, which may lead to equipment performance degradation or even more serious consequences, thus affecting the overall power generation efficiency and economic benefits of the power plant.

[0004] 2. The disadvantage of setting up a server on site for monitoring is that the monitoring cost is high, and users need to remember the network address of each power station server and set up a user name and password, which is more troublesome to manage. At the same time, this method has certain limitations in data transmission, and it takes a certain amount of time to view data. In addition, the client needs to be continuously upgraded to improve its functions, and the settings will be more cumbersome.

[0005] 3. With the development of technology, photovoltaic power station monitoring should also develop in the direction of remote monitoring and intelligence to improve monitoring efficiency and accuracy and reduce operation and maintenance costs. However, the current monitoring mode of photovoltaic power stations is still mainly traditional, and AI monitoring is only applied in a few core places. It is impossible to achieve designated, rapid, comprehensive monitoring analysis and early warning, and lacks intelligent and broad development prospects. Summary of the invention

[0006] In order to solve the technical problems existing in the prior art, the present invention provides the following technical solutions: On the one hand, a photovoltaic station virtual monitoring method based on digital twin technology is provided, which is implemented based on a photovoltaic station digital twin platform. The photovoltaic station digital twin platform includes a data acquisition layer, a data processing layer, a digital twin model layer and a virtual monitoring platform layer, wherein: ‌Data Collection Layer‌: used to collect the status data of each element in the photovoltaic station in real time and transmit it to the data processing layer, where the elements include people, objects and behaviors in the photovoltaic station; ‌Data processing layer‌: used to receive the state data from the data acquisition layer, clean, verify and store it, and use big data analysis technology to process and analyze the collected data, extract the corresponding three-dimensional feature data from the state information of each element and upload it to the ‌digital twin model layer; ‌Digital twin model layer‌: used to construct a digital twin model of a photovoltaic power station through digital twin technology based on the three-dimensional feature data, obtain a photovoltaic station twin model, and classify and save the digital twin model data set of each element in the photovoltaic station twin model; ‌Virtual monitoring platform layer‌: used to visualize the digital twin model data set of the target management objects in the PV station twin model and its elements; and, using a multimodal AI model: Momo AI, to perform multimodal monitoring of each element of the PV station twin model. When an abnormal situation is monitored, the alarm mechanism is automatically triggered, and the corresponding alarm signal is generated and sent to the handheld inspection terminal of the operation and maintenance personnel; Management layer: used to select target management objects from various elements, and perform visual management and multi-modal monitoring on the target management objects; The method comprises: S1. Establish a virtual monitoring task for the photovoltaic station and select the target management object from each element; S2. Collecting the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; S3. Through Momo AI, traverse and identify the target detection data set, and determine whether abnormal data that triggers the alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; S4. Through Momo AI, the context data set is tracked for the abnormal data that appears, and the context data set obtained by tracking is sent together with the alarm signal to the inspection terminal.

[0007] Preferably, the S2, real-time acquisition of the digital twin model dataset of the target management object in the photovoltaic station twin model to obtain a target detection dataset, includes: Collecting the three-dimensional feature data of each element in real time; Based on the data set mapping relationship between the physical model and the virtual model, data mapping calculation is performed on the three-dimensional feature data of each element, and the digital twin model data set of each element in the photovoltaic station twin model is updated according to the calculation result; Traversing the digital twin model data set before and after each element is updated, and marking the elements of the digital twin model data set that have data changes, to obtain an element set S; Determine whether the target management object exists in the element set S: If so, extract the updated digital twin model dataset of the target management object to obtain the target detection dataset; If it does not exist, continue to update monitoring.

[0008] Preferably, the step S3, traversing and identifying the target detection data set through Momo AI, and determining whether abnormal data that triggers an alarm mechanism appears in the target detection data set, includes: Inputting the target detection data set of each target management object into the Momo AI anomaly detection model preset on the virtual monitoring platform layer; Through the Momo AI anomaly detection model, the target detection data set of each target management object is traversed and identified to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set; The method for generating the Momo AI anomaly detection model includes: Collect and pre-process abnormal motion data of each element in the photovoltaic station, including image, video or text data; Perform feature engineering on the abnormal action data of each element, use the multimodal algorithm in Momo AI to adaptively identify and extract abnormal action features in the abnormal action data of each element, and construct a feature set consisting of the abnormal action features of several elements; Dividing the feature set into a training set and a validation set according to a preset ratio; Input the training set into the preset Momo AI, perform feature training and learning, and construct a Momo AI anomaly detection model; The recognition performance of the Momo AI anomaly detection model is verified using the validation set: If the verification is successful, the Momo AI anomaly detection model is deployed on the virtual monitoring platform layer and put into use after adjusting the parameters; If the verification fails, repeat the above steps to rebuild the Momo AI anomaly detection model.

[0009] Preferably, when performing feature engineering, it also includes: Collect voice data / images describing abnormal motion data of each element; Parsing the voice data / image and obtaining corresponding abnormal action description text; Using the multimodal algorithm in Momo AI, adaptively identify and extract the description information of the abnormal actions of each element in the abnormal action description text; The description information is marked on the abnormal action feature.

[0010] On the other hand, a photovoltaic station virtual monitoring system based on digital twin technology is provided, and the photovoltaic station virtual monitoring system based on digital twin technology is used to implement the above-mentioned photovoltaic station virtual monitoring method based on digital twin technology, and the system includes: The management end is used to establish the virtual monitoring task of the photovoltaic station and select the target management object from each element; A virtual monitoring platform is used to collect the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; and, through Momo AI, traverse and identify the target detection data set to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; and, through Momo AI, perform context data set tracking on the abnormal data that appears, and send the context data set obtained by tracking together with the alarm signal to the inspection terminal; An inspection terminal, used for receiving and displaying the context data set and the alarm signal; The management terminal and the inspection terminal are respectively connected to the virtual monitoring platform for communication.

[0011] On the other hand, an electronic device is provided, comprising: a processor; and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the above-mentioned photovoltaic station virtual monitoring methods based on digital twin technology is implemented.

[0012] On the other hand, a computer-readable storage medium is provided, in which at least one instruction is stored. The at least one instruction is loaded and executed by a processor to implement any one of the above-mentioned photovoltaic station virtual monitoring methods based on digital twin technology.

[0013] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: The present invention uses digital twin technology to realize virtual monitoring and alarm of photovoltaic power stations, which has the following advantages: ‌Improve operation and maintenance efficiency‌: Through the virtual monitoring platform, operation and maintenance personnel can grasp the operating status of the power station in real time, promptly discover and handle abnormal situations, and improve operation and maintenance efficiency.

[0014] ‌Reduce operation and maintenance costs‌: Through virtual monitoring and alarm, the number of on-site inspections and manual intervention can be reduced, thereby reducing operation and maintenance costs.

[0015] Optimize operation strategy: Based on the analysis results of the digital twin model, the operation strategy of the power station can be optimized to improve power generation efficiency and equipment utilization.

[0016] This solution can be widely used in the monitoring and management of various types of photovoltaic power stations, providing strong support for the intelligent and efficient operation of power stations. ‌Using Momo AI for monitoring, alarm analysis and management of virtual models can achieve comprehensive and real-time monitoring of various elements and behaviors in the virtual environment, and issue alarms in time when abnormalities or potential risks are detected.‌ Using Momo AI for monitoring, alarm analysis and management of virtual models can achieve efficient, accurate and real-time monitoring and alarm functions, provide strong guarantees for the security and stability of the virtual environment, and realize intelligent, virtualized and unmanned monitoring of photovoltaic power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 It is a schematic diagram of the architecture of a photovoltaic station digital twin platform provided by an embodiment of the present invention; Figure 2 This is a flow chart of a photovoltaic station virtual monitoring method based on digital twin technology provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a construction process of a Momo AI anomaly detection model provided by an embodiment of the present invention; Figure 4 It is a block diagram of a photovoltaic station virtual monitoring system based on digital twin technology provided by an embodiment of the present invention; Figure 5 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0020] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0021] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0022] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are consistent.

[0023] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0024] An embodiment of the present invention provides a photovoltaic plant virtual monitoring method based on digital twin technology. The method can be implemented by an electronic device, which can be a terminal or a server.

[0025] Multimodal AI Model: Introduction to Momo AI Momo AI is a set of multimodal AI models. Its biggest feature is that it not only has image and text processing capabilities, but also can interact with the environment by pointing to specific objects. This pointing function gives Momo AI greater application potential in both the physical world and the virtual world.

[0026] In terms of architecture, Momo AI uses an efficient data training method, giving priority to training the model with more precise and detailed image descriptions. This strategy avoids the common AI "hallucination" problem, which is to mistakenly generate inaccurate information. Through the pointing function, Momo AI is able to identify objects and convey their information to users in a more intuitive way, such as pointing out objects in pictures and converting them into structured data in JSON format.

[0027] Compared with other AI models, Momo AI's significant difference and innovation lies in its multimodal interaction capabilities. Many large AI models focus mainly on processing a single modality of images or text, while Momo AI is able to handle the complex relationship between images and text and excel in multimodal tasks. In visual-language evaluation tasks, Momo AI's performance is comparable to current top closed-source models such as GPT-4 and Gemini 1.5 Pro, and even surpasses them in some scenarios.

[0028] In addition, Momo AI's model size is relatively small, but its performance in multiple fields far exceeds that of its competitors that are ten times larger. This is due to its innovative data training strategy and focus on high-quality data sets. In contrast, other large multimodal language models are usually trained on huge data sets containing billions of images and text samples, and may contain trillions of parameters. Momo AI is trained on a smaller and more "carefully planned" data set, achieving better performance.

[0029] The advantages of Momo AI are also reflected in its practical applications. Its multimodal capabilities enable Momo AI to be widely used in multiple practical scenarios, such as virtual assistant systems, augmented reality, and robotics. In virtual assistant systems, Momo AI can help users complete various tasks, such as ordering drinks or identifying objects in pictures, through multimodal interaction capabilities. In augmented reality applications, Momo AI can be combined with AR devices to provide a vision-based enhanced interactive experience. In the field of robotics, Momo AI can help robots identify and manipulate objects more accurately through its powerful visual understanding capabilities.

[0030] Momo AI mainly integrates AI algorithms for image processing and text processing, and has the ability to interact with the environment through "pointing", which is actually a special interaction mode. Specifically, it includes: 1. Image processing algorithm: enables Momo AI to recognize and understand visual content, including objects, scenes, etc. This is a very critical part of multimodal AI because visual information is one of the most intuitive and richest sources of information in human-machine interaction.

[0031] 2. Text processing algorithm: enables Momo AI to understand and generate text information, thereby interacting with users in language. Text processing algorithms play an important role in multimodal AI, enabling machines to understand and respond to users' text instructions or queries.

[0032] 3. ‌Pointing Algorithm‌: This is an innovation of Momo AI, which is able to point to a specific object in the visual content. Through this pointing function, Momo AI can not only recognize objects, but also interact with the environment more intuitively. This algorithm actually combines image processing and interaction technology, providing Momo AI with a unique way of interaction. ‌ Different image processing algorithms or text processing algorithms can be customized by users or optimized and upgraded on the omoAI algorithm model architecture. Existing image or text processing algorithms include: Image processing algorithms mainly include the following: Filtering (smoothing, noise reduction): such as mean filtering, median filtering, Gaussian filtering, etc., used to remove noise from images or smooth images.

[0033] ‌Edge sharpening‌: such as Sobel operator, Laplace operator, Prewitt operator, etc., used to enhance the edge information in the image.

[0034] Image segmentation: such as threshold segmentation, boundary-based segmentation, Hough transform, region-based segmentation, etc., is used to divide an image into several regions with specific properties.

[0035] ‌Geometric transformation‌: such as image rotation, translation, mirroring, etc., used to change the spatial position or shape of an image.

[0036] Feature extraction: such as blob analysis, corner detection, contour extraction, etc., used to extract useful feature information from images.

[0037] Image enhancement: such as histogram equalization and contrast enhancement, used to improve the visual effect of the image.

[0038] ‌Image matching‌: such as template matching and search matching, used to find areas in an image that are similar to a specific template.

[0039] Text processing algorithms mainly include the following: Tokenization: Split text into words, phrases, or other meaningful elements.

[0040] ‌Remove stop words‌: Remove common, meaningless words like "the", "is", "in", etc.

[0041] ‌Stemming and Lemmatization‌: Restoring words to their base or dictionary form, such as restoring “running” to “run”.

[0042] ‌Part-of-speech tagging‌: Tag each word in the text with its part of speech, such as noun, verb, etc.

[0043] ‌Named Entity Recognition (NER): Identify entities in text, such as names of people, places, organizations, etc.

[0044] Sentiment Analysis: Determine the sentiment of a text, such as positive, negative, or neutral.

[0045] ‌Topic Modeling‌: Extracting topic information from text to understand and organize text data.

[0046] Momo AI integrates various modal algorithms and can adaptively identify data such as images and texts and extract corresponding image and text features.

[0047] The present invention utilizes digital twin virtual monitoring technology and a multimodal AI model, MomoAI, for photovoltaic stations to achieve multimodal intelligent supervision of abnormal behaviors of photovoltaic stations in virtual model scenarios. Momo AI can be used to perform monitoring, alarm analysis and management of virtual models, and comprehensive, real-time monitoring of various elements and behaviors in the virtual environment can be achieved, and timely alarms can be issued when abnormalities or potential risks are detected.

[0048] like Figure 1 As shown, this method needs to be implemented based on a photovoltaic station digital twin platform, which includes a data acquisition layer, a data processing layer, a digital twin model layer, and a virtual monitoring platform layer, wherein: ‌Data Collection Layer‌: used to collect the status data of each element in the photovoltaic station in real time and transmit it to the data processing layer, where the elements include people, objects and behaviors in the photovoltaic station; ‌Data processing layer‌: used to receive the state data from the data acquisition layer, clean, verify and store it, and use big data analysis technology to process and analyze the collected data, extract the corresponding three-dimensional feature data from the state information of each element and upload it to the ‌digital twin model layer; ‌Digital twin model layer‌: used to construct a digital twin model of a photovoltaic power station through digital twin technology based on the three-dimensional feature data, obtain a photovoltaic station twin model, and classify and save the digital twin model data set of each element in the photovoltaic station twin model; ‌Virtual monitoring platform layer‌: used to visualize the digital twin model data set of the target management objects in the PV station twin model and its elements; and, using a multimodal AI model: Momo AI, to perform multimodal monitoring of each element of the PV station twin model. When an abnormal situation is monitored, the alarm mechanism is automatically triggered, and the corresponding alarm signal is generated and sent to the handheld inspection terminal of the operation and maintenance personnel; Management layer: used to select target management objects from various elements, and perform visual management and multi-modal monitoring on the target management objects.

[0049] The steps of using digital twin technology to build a virtual model (digital twin model) of the corresponding entity will not be repeated here. The steps of using 3D feature data (such as extracting the corresponding 3D action posture features of the characters from the 3D point cloud data of the characters in the scene and building the corresponding 3D posture model; or collecting 3D data of the devices in the scene and extracting the corresponding 3D feature data) to build a virtual model can refer to the following scheme: 1. Data Collection and Preprocessing ‌Data Collection‌: Use drones, LiDAR, 3D scanners and other equipment to perform high-precision scanning and measurement of photovoltaic power stations to obtain original 3D point cloud data.

[0050] At the same time, the status information of each element in the photovoltaic power station (such as solar panels, inverters, brackets, etc.) is collected, including but not limited to temperature, voltage, current, power, etc.

[0051] ‌Data Preprocessing‌: The original 3D point cloud data is processed by denoising, registration, segmentation, etc. to extract clear and accurate 3D features.

[0052] Clean, calibrate and format status information to ensure data accuracy and consistency.

[0053] 2. 3D feature data extraction ‌Feature Identification‌: The three-dimensional features of each element of the photovoltaic power station, such as shape, size, position, etc., are extracted from the pre-processed three-dimensional point cloud data using three-dimensional image processing algorithms (such as edge detection, surface fitting, etc.).

[0054] ‌Feature digitization‌: The extracted three-dimensional features are converted into digital data formats, such as coordinate points, vectors, polygons, etc., to facilitate subsequent processing and analysis.

[0055] 3. Data upload and integration ‌Data upload‌: Upload the extracted three-dimensional feature data and status information to the data storage system of the digital twin model layer.

[0056] ‌Data Integration‌: In the digital twin model layer, the three-dimensional feature data and status information are integrated to form a three-dimensional feature dataset containing complete information.

[0057] 4. Digital twin model construction ‌Model initialization‌: Based on the 3D feature data set, a preliminary digital twin model of the PV power station is constructed using 3D modeling software (such as AutoCAD, SolidWorks, etc.) or a digital twin platform.

[0058] ‌Model refinement‌: According to the detailed information of the 3D feature data, the preliminary model is refined, including adding textures, adjusting colors, setting animations, etc., to more realistically reflect the actual situation of the photovoltaic power station.

[0059] ‌Model Validation‌: The accuracy and reliability of the digital twin model are ensured by comparison and verification with the actual photovoltaic power station.

[0060] 5. Classification, storage and data management ‌Save by Category‌: The various elements in the constructed photovoltaic station twin model (such as solar panel model, inverter model, administrator or technician model of each workstation (for posture recognition), etc.) are classified and saved to facilitate subsequent management and query.

[0061] ‌Data Management‌: Establish a complete data management mechanism, including data backup, data update, data access permission control, etc., to ensure the security and availability of the digital twin model data set.

[0062] 6. Continuous optimization and iteration ‌Real-time monitoring and feedback‌: Use digital twin models to conduct real-time monitoring and simulation analysis of photovoltaic power stations to promptly identify and solve problems.

[0063] ‌Model update and iteration‌: According to the actual changes in photovoltaic power stations and the emergence of new technologies, the digital twin model is updated and iterated regularly to maintain the advancement and accuracy of the model.

[0064] Through the above steps, big data analysis technology and digital twin technology can be used to build a digital twin model of the photovoltaic power station, and realize the extraction of three-dimensional feature data and model management. This will provide strong support for the operation, maintenance and management of photovoltaic power stations.

[0065] This solution uses digital twin technology, combined with the Internet of Things, big data analysis and visualization technology, to build a virtual monitoring platform corresponding to the real photovoltaic power station. The platform can collect the operating data of each device in the power station in real time, simulate and analyze it through the digital twin model, and realize comprehensive monitoring of the power station's operating status and abnormal alarm.

[0066] 1. The system architecture is as follows: 1. Data collection layer: Various sensors and intelligent instruments are installed, such as inverter output power sensor, combiner box current and voltage sensor, irradiator, meteorological instrument, electric meter, etc., to collect real-time operation data of various equipment in the power station.

[0067] The data acquisition equipment transmits the collected data to the data processing center via wired or wireless means.

[0068] 2. Data processing layer: The data processing center receives data from the data acquisition layer and performs cleaning, verification and storage.

[0069] Using big data analysis technology, the collected data is processed and analyzed to extract useful information and provide data support for the virtual monitoring platform.

[0070] 3. Digital Twin Model Layer: Build a digital twin model of the photovoltaic power station based on the integration of physical models, sensor updates, historical and real-time data.

[0071] The digital twin model can reflect the operating status of each equipment in the power station in real time, including the working status of the inverter, the current and voltage of the combiner box, the radiation intensity, meteorological conditions, power generation, etc.

[0072] 4. ‌Virtual monitoring platform layer‌: It provides a user-friendly interface to display the digital twin model of the photovoltaic power station (referred to as the photovoltaic station twin model), and displays the operating status and key parameters of each device in real time.

[0073] A multimodal AI model (Momo AI for short) is used to conduct multimodal monitoring of the photovoltaic station twin model and comprehensively monitor the operating status of the power station, including real-time monitoring of parameters such as power generation, equipment temperature, voltage, and current.

[0074] When an abnormal situation is detected, the alarm mechanism is automatically triggered to alert the operation and maintenance personnel through sound, light, pop-up windows, etc.

[0075] (II) Function Implementation 1. Real-time data collection and transmission: The Internet of Things technology is used to realize real-time collection and transmission of data from various equipment in the power station.

[0076] Ensure the accuracy and real-time nature of data and provide a reliable data source for the virtual monitoring platform.

[0077] 2. Digital twin model construction: Construct a three-dimensional digital twin model based on the physical structure and equipment layout of the power station.

[0078] The collected data is mapped to the digital twin model in real time to achieve virtual reproduction of the power plant's operating status.

[0079] 3. Virtual operation monitoring: On the virtual monitoring platform, the operating status and key parameters of each device in the power station are displayed in real time.

[0080] Supports comprehensive monitoring of the power station's operating status, including real-time monitoring of power generation, equipment temperature, voltage, current and other parameters.

[0081] Provides a graphical interface to display data in the form of charts, curves, etc., so that operation and maintenance personnel can quickly understand the operation status of the power station.

[0082] 4. Abnormal alarm and fault diagnosis: When an abnormal situation is detected, the alarm mechanism is automatically triggered to alert the operation and maintenance personnel through sound, light, pop-up windows, etc.

[0083] Provide fault diagnosis suggestions to help operation and maintenance personnel quickly locate the cause of the problem and take appropriate measures to deal with it.

[0084] Therefore, the present invention utilizes Momo AI to monitor, alarm, analyze and manage virtual models, which can achieve comprehensive and real-time monitoring of various elements and behaviors in the virtual environment, and issue an alarm in a timely manner when an abnormality or potential risk is detected. Specifically: First of all, Momo AI has powerful multimodal processing capabilities and can simultaneously process multiple types of data such as images, videos, and text. In virtual model monitoring, Momo AI can fully perceive the state of the virtual world by analyzing visual information (such as character movements, object positions, etc.) and text information (such as system logs, user input, etc.) in the virtual environment.

[0085] Then, Momo AI uses advanced deep learning and machine learning algorithms to analyze and process the collected data in real time. It can identify abnormal behaviors or potential risks in the virtual environment, such as abnormal actions of characters, abnormal movement of objects, abnormal changes in system parameters, etc. These anomalies may be problems within the virtual world, or they may be caused by external attacks or failures.

[0086] Once Momo AI detects an anomaly, it will immediately trigger an alarm mechanism. The alarm mechanism can include various forms of notifications, such as sound prompts, pop-up warnings, email notifications, etc., to ensure that monitoring personnel can quickly learn about and respond to abnormal situations. At the same time, Momo AI can also provide detailed information and context of the abnormality to help monitoring personnel quickly locate the problem and take appropriate measures.

[0087] In addition, Momo AI also has the ability to self-learn and optimize. It can gradually improve the accuracy of identifying abnormal behaviors and the speed of alarm response by continuously analyzing and processing data in the virtual environment. This enables Momo AI to better adapt to the changes and development of the virtual environment and provide continuous support and improvement for monitoring alarm analysis management.

[0088] In summary, using Momo AI for monitoring, alarm analysis and management of virtual models can achieve efficient, accurate and real-time monitoring and alarm functions, providing strong guarantees for the security and stability of the virtual environment.

[0089] like Figure 2 As shown in FIG. 1 , a flow chart of a photovoltaic station virtual monitoring method based on digital twin technology of the present invention is shown. The processing flow of the method may include the following steps: S1. Establish a virtual monitoring task for the photovoltaic station and select the target management object from each element; S2. Collecting the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; S3. Through Momo AI, traverse and identify the target detection data set, and determine whether abnormal data that triggers the alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; S4. Through Momo AI, the context data set is tracked for the abnormal data that appears, and the context data set obtained by tracking is sent together with the alarm signal to the inspection terminal.

[0090] Because Momo AI is used to monitor, alarm, analyze and manage virtual models, it is possible to achieve comprehensive, real-time monitoring of various elements and behaviors in the virtual environment, and issue an alarm in a timely manner when anomalies or potential risks are detected. Therefore, the present invention allows administrators to construct virtual monitoring tasks for photovoltaic stations in advance, and the tasks may include: The elements to be monitored; The core elements of each monitored element are used as target management objects; Monitoring content and corresponding alarm mechanism; Monitoring signal routing addresses and patrol administrator allocation plans, etc.

[0091] The core elements can be specified by the administrator, such as abnormal supervision of the virtual model of the inverter and its administrator, including abnormal monitoring of human behavior and abnormal monitoring of the location status or movement behavior of objects (for example, the assembly / maintenance actions of the inverter accessories can be identified by Momo AI to determine whether their assembly / maintenance actions are qualified. Momo AI will receive training on abnormal action data of the assembly / maintenance actions of the inverter accessories in the early stage; the same applies to others).

[0092] Subsequently, when the platform is applied, data of each element (physical model) is collected in real time and processed based on data set mapping, and the digital twin model data set of each element in the photovoltaic station twin model is dynamically updated in real time.

[0093] The multimodal AI model, Momo AI, can virtually monitor each element in the virtual scene, monitor the abnormal behavior in the virtual scene based on the action feature images, voice or text data of the virtual model, and realize remote supervision and management. The multimodal monitoring of each element is carried out in the twin model of the photovoltaic station. When an abnormal situation is monitored, the alarm mechanism is automatically triggered, and the corresponding alarm signal is generated and sent to the handheld inspection terminal of the operation and maintenance personnel.

[0094] Preferably, the S2, real-time acquisition of the digital twin model dataset of the target management object in the photovoltaic station twin model to obtain a target detection dataset, includes: Collecting the three-dimensional feature data of each element in real time; Based on the data set mapping relationship between the physical model and the virtual model, data mapping calculation is performed on the three-dimensional feature data of each element, and the digital twin model data set of each element in the photovoltaic station twin model is updated according to the calculation result; Traversing the digital twin model data set before and after each element is updated, and marking the elements of the digital twin model data set that have data changes, to obtain an element set S; Determine whether the target management object exists in the element set S: If so, extract the updated digital twin model dataset of the target management object to obtain the target detection dataset; If it does not exist, continue to update monitoring.

[0095] Because the management end needs to perform Momo AI designated management on the target management object. Therefore, when the virtual models of each subsequent element perform data updates, in order to avoid increasing the data operation and calculation pressure of Momo AI, the present invention first traverses and identifies the changes in the data set before and after the element update, determines and marks the elements that have generated data changes, and collects the elements in the data set that have generated data changes to obtain an element set S. Then, it is determined from the element set S whether there is an element in the element set where the target management object is located. If there is a target management object corresponding to the element with data changes, its updated data set is extracted as the target detection data set to be detected; if not, the model continues to be monitored for data changes. In the above manner, it is avoided that the Momo AI model performs data detection on all elements, thereby reducing its computing pressure and computing power cost.

[0096] Preferably, the step S3, traversing and identifying the target detection data set through Momo AI, and determining whether abnormal data that triggers an alarm mechanism appears in the target detection data set, includes: Inputting the target detection data set of each target management object into the Momo AI anomaly detection model preset on the virtual monitoring platform layer; Through the Momo AI anomaly detection model, the target detection data set of each target management object is traversed and identified to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set; Among them, Figure 3 As shown, the method for generating the Momo AI anomaly detection model includes: Collect and pre-process abnormal motion data of each element in the photovoltaic station, including image, video or text data; Perform feature engineering on the abnormal action data of each element, use the multimodal algorithm in Momo AI to adaptively identify and extract abnormal action features in the abnormal action data of each element, and construct a feature set consisting of the abnormal action features of several elements; Dividing the feature set into a training set and a validation set according to a preset ratio; Input the training set into the preset Momo AI, perform feature training and learning, and construct a Momo AI anomaly detection model; The recognition performance of the Momo AI anomaly detection model is verified using the validation set: If the verification is successful, the Momo AI anomaly detection model is deployed on the virtual monitoring platform layer and put into use after adjusting the parameters; If the verification fails, repeat the above steps to rebuild the Momo AI anomaly detection model.

[0097] Because Momo AI has a multimodal algorithm model, it can not only identify data features, but also train models based on the extracted features and build corresponding anomaly detection models.

[0098] Here, abnormal action data of people, equipment or corresponding behaviors in virtual scenes can be collected, such as the abnormal operation action image of a technician on an inverter accessory, equipment fault operation data, or abnormal transportation trajectory data of a photovoltaic device in a photovoltaic station, such as a photovoltaic panel.

[0099] Then use Momo AI to learn the features of abnormal action data of people, objects or behaviors. The steps to build the Momo AI anomaly detection model can refer to the following: ‌Data collection and preprocessing‌: Collect datasets containing normal and abnormal actions. This data can include images, videos, text, or other forms of data.

[0100] Preprocess the data, including data cleaning, denoising, and normalization, to ensure data quality and make it suitable for feature learning.

[0101] ‌Feature Extraction‌: Use Momo AI or its integrated feature extraction technology to extract useful features from the preprocessed data. These features should be able to reflect the essential attributes and regularities of the data.

[0102] Feature extraction may involve using deep learning algorithms such as convolutional neural networks (CNNs) to extract features from images or videos, or using text embedding models to extract features from text data.

[0103] Momo AI integrates multi-modal algorithm models, such as image algorithms, text algorithms, or machine learning algorithms such as CNN and RNN, so it can adaptively select algorithms suitable for processing each data type according to the data type to perform feature recognition and extraction. For example, for abnormal operation action images, CNN can be used to recognize and extract abnormal action posture features.

[0104] ‌Model Training‌: Choose an appropriate AI algorithm, such as Autoencoder, Isolation Forest, or a deep learning-based model, to train anomaly detection models.

[0105] The model is trained using the collected dataset. During the training process, the model will learn how to distinguish between normal and abnormal motion data.

[0106] The accuracy and generalization ability of the model can be improved by adjusting the model parameters and performing multiple iterative training.

[0107] ‌Model Evaluation and Optimization‌: Use the validation set to evaluate the performance of the model, including accuracy, recall, F1 score, etc. To be completed by the administrator.

[0108] Optimize the model based on the evaluation results, such as adjusting feature selection, improving model structure, or adjusting training parameters.

[0109] The process of training, evaluating, and optimizing is repeated until the model reaches a satisfactory level of performance.

[0110] ‌Deployment and Application‌: Deploy the trained Momo AI anomaly detection model to the virtual monitoring platform layer.

[0111] Use the model to perform real-time testing on new input data to identify anomalous actions.

[0112] Take appropriate actions based on the detection results, such as issuing an alarm, recording a log, or performing further analysis.

[0113] ‌Continuous Improvement‌: Collect new abnormal action data to continuously update and optimize the model.

[0114] Monitor the performance of your model and retrain or tune it if necessary.

[0115] Explore new feature extraction and model training methods to further improve model accuracy and robustness.

[0116] Because Momo AI is able to use optimized data sets to build model training data, including: Detailed image description: Describe the abnormal action through images (that is, abnormal action annotations and explanation information on the abnormal action map and on it); Voice description: Describe the image in detail through voice. This type of voice input usually contains more details, helping AI to better understand and learn the content in the image. For example, for an abnormal operation of a certain inverter accessory, the administrator can prepare the corresponding abnormal operation explanation voice data. The voice needs to explain which points are prone to operation errors, what are the action requirements for each error position, what is the correction strategy for operation errors, etc.

[0117] Through enhanced training data, Momo AI can deeply understand the characteristic data of abnormal action data of each element, thereby enhancing the accuracy of model training and learning.

[0118] Preferably, when performing feature engineering, it also includes: Collect voice data / images describing abnormal motion data of each element; Parsing the voice data / image and obtaining corresponding abnormal action description text; Using the multimodal algorithm in Momo AI, adaptively identify and extract the description information of the abnormal actions of each element in the abnormal action description text; The description information is marked on the abnormal action feature.

[0119] The collection, analysis, identification and annotation of the description data of abnormal action data of each element can refer to the following steps: ‌First, collect abnormal action data‌: For voice data, voice information describing abnormal actions of each element can be collected through management-side recording or related application device programs (such as the voice recording function of a smartphone).

[0120] For image data, you can use a camera or screenshot tool to capture images of each element in an abnormal motion state.

[0121] ‌Next, parse the voice / image data‌: Voice data: Use voice recognition technology to convert the collected voice data into text. This step can be achieved with the help of existing voice recognition software or APIs to convert the abnormal action descriptions in the voice into processable text information.

[0122] Image data: Use image recognition technology to identify the elements in the image and extract the features of abnormal movements. This usually involves steps such as image preprocessing, feature extraction and classification, which can be done using computer vision libraries or deep learning models.

[0123] ‌Then, using the multimodal algorithm in Momo AI‌: The parsed abnormal action description text is used as input, and the multimodal algorithm in Momo AI is used for adaptive recognition. The multimodal algorithm can comprehensively consider multiple information such as text, voice, and image to improve the accuracy and robustness of recognition.

[0124] Through algorithm processing, specific descriptive information about the abnormal actions of each element in the text is extracted, and this information will be used for subsequent labeling work.

[0125] ‌Finally, the description information is annotated on the abnormal action features‌: For image data, the extracted abnormal action description information can be annotated in the form of a text label on the abnormal action feature of the corresponding element (the annotation position can be the upper right corner by default). This can be achieved through image annotation tools or programming to ensure that the description information corresponds to the abnormal action feature in the image.

[0126] For voice data, the parsed text description can also be annotated on the abnormal action feature of the corresponding element. In this way, in subsequent use, the voice description of a specific abnormal action can be found by querying the database.

[0127] Through the above steps, the goal of collecting, parsing, identifying and annotating abnormal action data of each element can be achieved. Momo AI has a multimodal algorithm that can extract descriptive information about abnormal actions from the parsed text, such as abnormal action correction strategies.

[0128] Therefore, the present invention uses Momo AI to perform monitoring, alarm analysis and management of virtual models, which can achieve efficient, accurate and real-time monitoring and alarm functions, and provide strong protection for the security and stability of the virtual environment.

[0129] Figure 4 The block diagram of a photovoltaic station virtual monitoring system based on digital twin technology is shown according to an exemplary embodiment. The system is used for a photovoltaic station virtual monitoring method based on digital twin technology. Figure 4 On the other hand, a photovoltaic station virtual monitoring system based on digital twin technology is provided, and the photovoltaic station virtual monitoring system based on digital twin technology is used to implement the above-mentioned photovoltaic station virtual monitoring method based on digital twin technology, and the system includes: The management end is used to establish the virtual monitoring task of the photovoltaic station and select the target management object from each element; A virtual monitoring platform is used to collect the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; and, through Momo AI, traverse and identify the target detection data set to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; and, through Momo AI, perform context data set tracking on the abnormal data that appears, and send the context data set obtained by tracking together with the alarm signal to the inspection terminal; An inspection terminal, used for receiving and displaying the context data set and the alarm signal; The management terminal and the inspection terminal are respectively connected to the virtual monitoring platform for communication.

[0130] The interactive management functions between the management terminal and the inspection terminal and the virtual monitoring platform respectively should be understood in conjunction with the above-mentioned platform and method, and will not be elaborated here.

[0131] Figure 5 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 5 As shown, optionally, the electronic device 410 may include a first processor 2001 .

[0132] Optionally, the electronic device 410 may further include a memory 2002 and a transceiver 2003 .

[0133] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0134] Combine the following Figure 5 The components of the electronic device 410 are described in detail: The first processor 2001 is the control center of the electronic device 410, and may be a processor or a general term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or may be application specific integrated circuits (ASICs), or may be one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0135] Optionally, the first processor 2001 can execute various functions of the electronic device 410 by running or executing a software program stored in the memory 2002 and calling data stored in the memory 2002 .

[0136] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 5 CPU0 and CPU1 are shown in FIG.

[0137] In a specific implementation, as an embodiment, the electronic device 410 may also include multiple processors, such as Figure 5The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0138] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled to be executed by the first processor 2001. The specific implementation method can refer to the above method embodiment, which will not be repeated here.

[0139] Optionally, the memory 2002 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001, or may exist independently and access the first processor 2001 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0140] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0141] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 5 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0142] Optionally, the transceiver 2003 may be integrated with the first processor 2001, or may exist independently and communicate with the first processor 2001 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0143] It should be noted that Figure 5 The structure of the electronic device 410 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure recognition device may include more or fewer components than those shown in the figure, or combine certain components, or arrange the components differently.

[0144] In addition, the technical effects of the electronic device 410 can refer to the technical effects of the photovoltaic station virtual monitoring method based on digital twin technology described in the above method embodiment, and will not be repeated here.

[0145] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0146] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0147] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable system. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0148] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0149] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0150] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0151] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0152] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0153] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of the system or unit, which can be electrical, mechanical or other forms.

[0154] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0155] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0156] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0157] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A photovoltaic station virtual monitoring method based on digital twin technology, implemented based on a photovoltaic station digital twin platform, characterized in that: The photovoltaic station digital twin platform includes a data acquisition layer, a data processing layer, a digital twin model layer and a virtual monitoring platform layer, wherein: ‌Data Collection Layer‌: used to collect the status data of each element in the photovoltaic station in real time and transmit it to the data processing layer, where the elements include people, objects and behaviors in the photovoltaic station; ‌Data processing layer‌: used to receive the state data from the data acquisition layer, clean, verify and store it, and use big data analysis technology to process and analyze the collected data, extract the corresponding three-dimensional feature data from the state information of each element and upload it to the ‌digital twin model layer; ‌Digital twin model layer‌: used to construct a digital twin model of a photovoltaic power station through digital twin technology based on the three-dimensional feature data, obtain a photovoltaic station twin model, and classify and save the digital twin model data set of each element in the photovoltaic station twin model; ‌Virtual monitoring platform layer‌: used to visualize the digital twin model data set of the target management objects in the PV station twin model and its elements; and, using a multimodal AI model: Momo AI, to perform multimodal monitoring of each element of the PV station twin model. When an abnormal situation is monitored, the alarm mechanism is automatically triggered, and the corresponding alarm signal is generated and sent to the handheld inspection terminal of the operation and maintenance personnel; Management layer: used to select target management objects from various elements, and perform visual management and multi-modal monitoring on the target management objects; The method comprises: S1. Establish a virtual monitoring task for the photovoltaic station and select the target management object from each element; S2. Collecting the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; S3. Through Momo AI, traverse and identify the target detection data set, and determine whether abnormal data that triggers the alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; S4. Through Momo AI, the context data set is tracked for the abnormal data that appears, and the context data set obtained by tracking is sent together with the alarm signal to the inspection terminal.

2. The photovoltaic station virtual monitoring method based on digital twin technology according to claim 1 is characterized in that: The step S2, collecting the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set, includes: Collecting the three-dimensional feature data of each element in real time; Based on the data set mapping relationship between the physical model and the virtual model, data mapping calculation is performed on the three-dimensional feature data of each element, and the digital twin model data set of each element in the photovoltaic station twin model is updated according to the calculation result; Traversing the digital twin model data set before and after each element is updated, and marking the elements of the digital twin model data set that have data changes, to obtain an element set S; Determine whether the target management object exists in the element set S: If so, extract the updated digital twin model dataset of the target management object to obtain the target detection dataset; If it does not exist, continue to update monitoring.

3. The photovoltaic station virtual monitoring method based on digital twin technology according to claim 1 is characterized in that: S3, through Momo AI, traversing and identifying the target detection data set, and determining whether abnormal data that triggers an alarm mechanism appears in the target detection data set, includes: Inputting the target detection data set of each target management object into the Momo AI anomaly detection model preset on the virtual monitoring platform layer; Through the Momo AI anomaly detection model, the target detection data set of each target management object is traversed and identified to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set; The method for generating the Momo AI anomaly detection model includes: Collect and pre-process abnormal motion data of each element in the photovoltaic station, including image, video or text data; Perform feature engineering on the abnormal action data of each element, use the multimodal algorithm in Momo AI to adaptively identify and extract abnormal action features in the abnormal action data of each element, and construct a feature set consisting of the abnormal action features of several elements; Dividing the feature set into a training set and a validation set according to a preset ratio; Input the training set into the preset Momo AI, perform feature training and learning, and construct a Momo AI anomaly detection model; The recognition performance of the Momo AI anomaly detection model is verified using the validation set: If the verification is successful, the Momo AI anomaly detection model is deployed on the virtual monitoring platform layer and put into use after adjusting the parameters; If the verification fails, repeat the above steps to rebuild the Momo AI anomaly detection model.

4. The photovoltaic station virtual monitoring method based on digital twin technology according to claim 3 is characterized in that: When performing feature engineering, it also includes: Collect voice data / images describing abnormal motion data of each element; Parsing the voice data / image and obtaining corresponding abnormal action description text; Using the multimodal algorithm in Momo AI, adaptively identify and extract the description information of the abnormal actions of each element in the abnormal action description text; The description information is marked on the abnormal action feature.

5. A photovoltaic station virtual monitoring system based on digital twin technology, wherein the photovoltaic station virtual monitoring system based on digital twin technology is used to implement the photovoltaic station virtual monitoring method based on digital twin technology as claimed in any one of claims 1 to 4, characterized in that: The system comprises: The management end is used to establish the virtual monitoring task of the photovoltaic station and select the target management object from each element; A virtual monitoring platform is used to collect the digital twin model data set of the target management object in the photovoltaic station twin model in real time to obtain a target detection data set; and, through Momo AI, traverse and identify the target detection data set to determine whether abnormal data that triggers an alarm mechanism appears in the target detection data set: if so, generate an alarm signal corresponding to the target management object; otherwise, continue monitoring; and, through Momo AI, perform context data set tracking on the abnormal data that appears, and send the context data set obtained by tracking together with the alarm signal to the inspection terminal; An inspection terminal, used for receiving and displaying the context data set and the alarm signal; The management terminal and the inspection terminal are respectively connected to the virtual monitoring platform for communication.

6. An electronic device, characterized in that: The electronic device comprises: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, which can be called by a processor to execute the method according to any one of claims 1 to 4.

Images