Intelligent operation and maintenance management platform based on digital twinning

By performing data classification, preprocessing and feature splicing in the operation and maintenance management platform, the problems of poor data quality and difficulty in abnormal positioning are solved, efficient fault warning and processing are achieved, and the efficiency and stability of operation and maintenance management are improved.

CN118965230APending Publication Date: 2024-11-15ANHUI ANTAI TECH

Patent Information

Application Number
CN202411139278.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing operation and maintenance management platform did not perform further processing after data acquisition, resulting in poor data quality, inability to build an effective model, and inability to accurately locate abnormal areas and conduct targeted early warnings, resulting in low fault handling efficiency.

Method used

Through the intelligent operation and maintenance management platform based on digital twins, data classification and preprocessing are carried out, image and text data models are built, appropriate memory is selected for storage, and feature data is extracted for splicing to realize efficient data management and abnormal analysis, and according to faults Targeted warnings are made.

Benefits of technology

It improves the accuracy of data analysis and targeted warnings, ensures that operation and maintenance personnel can timely understand and handle faults, and improves the efficiency and stability of operation and maintenance management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118965230A_ABST
    Figure CN118965230A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent operation and maintenance management platform based on digital twinning, relates to the technical field of operation and maintenance management platforms, and aims to solve the problem that abnormal information cannot be acquired more intelligently during operation and maintenance data management. According to the method, the model storage space of the image data model and the text data model is confirmed, and the appropriate memory and the appropriate sub-storage space are selected for storage, so that efficient management of the data models is realized, timestamp labeling is performed on the stored image data model and the text data model, and the models have time attributes; the method enhances the traceability of data, can form a feature set containing rich information by respectively extracting feature data in an image and text data model and splicing feature vectors, and adopts corresponding early warning measures according to the severity of a fault, thereby improving the pertinence and effectiveness of early warning, and improving the accuracy of early warning. And early warning alarms of different degrees are transmitted to an abnormal area for alarm processing, so that operation and maintenance personnel can be ensured to know and process faults in time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance management platforms, and specifically to an intelligent operation and maintenance management platform based on digital twins. Background Art

[0002] The operation and maintenance management platform is an important tool designed to improve the efficiency and stability of enterprise operation and maintenance. It integrates various functions and tools to help the operation and maintenance team monitor, manage and optimize the company's IT infrastructure.

[0003] The patent application with publication number CN115643150A discloses an IDC intelligent operation and maintenance management system based on digital twins. It can timely reflect the operating status of the data center's software and hardware environment through digital twin technology, provide comprehensive data and information, and provide a reliable basis for decision makers, solving the problem of weak correlation between the current DC IM operation and maintenance platform and the real-time operation data of each device. The AI ​​algorithm module is used to identify and predict the hidden faults of each physical entity of the IDC and alarm, which improves the efficiency of fault identification and fault repair. Although the above patent solves the problem of operation and maintenance data management, the following problems still exist in actual operation:

[0004] 1. After the operation and maintenance management data is obtained, no further data processing is performed, resulting in poor data quality and the inability to build a model.

[0005] 2. The abnormal areas in the operation and maintenance management data model are not located more accurately, and targeted alarms are not issued based on the abnormal situations, resulting in the inability to warn of abnormalities in a timely manner.

[0006] 3. The acquired operation and maintenance management data is not used to build a targeted model according to the data type, which leads to limitations in the built model. Summary of the invention

[0007] The purpose of the present invention is to provide an intelligent operation and maintenance management platform based on digital twins. By confirming the model storage capacity of the image data model and the text data model, and selecting appropriate memory and sub-storage space for storage, efficient management of the data model is achieved, and the stored image data model and text data model are timestamped so that these models have time attributes, thereby enhancing the traceability of the data. By respectively extracting feature data from the image and text data models and splicing feature vectors, a feature set containing rich information can be formed. According to the severity of the fault, corresponding early warning measures are taken to improve the pertinence and effectiveness of the early warning, and early warning alarms of different degrees are transmitted to the abnormal area for alarm processing, so as to ensure that the operation and maintenance personnel can understand and handle the fault in time, and the problems in the prior art can be solved.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] The intelligent operation and maintenance management platform based on digital twins includes:

[0010] The data acquisition unit of the operation and maintenance management platform is used to:

[0011] Retrieve operation and maintenance management data from the database, classify the data according to the data type of the operation and maintenance management data, and perform data preprocessing after data classification to obtain target operation and maintenance management data;

[0012] The data model building unit of the operation and maintenance management platform is used to:

[0013] The target operation and maintenance management data is modeled using a digital twin model, where different data models are constructed according to different data types;

[0014] Build a model storage integration unit for:

[0015] Different data models are stored in different storages, and then the stored data models are integrated to obtain an integrated operation and maintenance management model;

[0016] Integrated model prediction analysis unit for:

[0017] Perform abnormal analysis on the integrated operation and maintenance management model, confirm the fault characteristics of the abnormal area according to the abnormal analysis results, and perform visual conversion on the fault characteristics of the abnormal area to obtain abnormal operation and maintenance management data after visual conversion;

[0018] Predictive analysis data early warning unit, used for:

[0019] The abnormal operation and maintenance management data is judged by fault level, different degrees of warning are issued according to different fault levels, and alarms are issued for abnormal areas based on the fault location.

[0020] Preferably, the operation and maintenance management platform data acquisition unit includes:

[0021] Data acquisition classification module for:

[0022] Operation and maintenance management data includes space data, equipment data, environment data and log data;

[0023] Classify spatial data, equipment data, environmental data and log data into text data and image data in turn;

[0024] After the classification is completed, the text data and image data of the space data, the text data and image data of the equipment data, the text data and image data of the environment data, and the text data and image data of the log data are obtained.

[0025] Preferably, the operation and maintenance management platform data acquisition unit further includes:

[0026] Categorical data preprocessing module for:

[0027] Preprocess the text data and image data of spatial data, equipment data, environmental data and log data respectively;

[0028] Among them, the preprocessing process of text data of spatial data, equipment data, environmental data and log data is data cleaning, data format unification and data standardization;

[0029] The preprocessing process of image data for spatial data, equipment data, environmental data and log data is image denoising, image size unification, image normalization, image quality enhancement and image feature extraction;

[0030] After the text data and image data preprocessing of space data, equipment data, environment data and log data is completed, the target operation and maintenance management data is obtained.

[0031] Preferably, the operation and maintenance management platform data model building unit is also used to:

[0032] Extracting text data and image data of spatial data, equipment data, environmental data and log data from target operation and maintenance management data respectively;

[0033] The text data of spatial data, equipment data, environmental data and log data are used to construct a text data model through a deep learning model;

[0034] The image data of spatial data, equipment data, environmental data and log data are classified using the ResNet model, target detection is performed using the YOLO model, and image segmentation is performed using the Mask R-CNN model. The image data model is obtained after image classification, target detection and image segmentation.

[0035] Preferably, the building model storage integrated unit includes:

[0036] Build a model storage module for:

[0037] Confirm the image data model and the text data model;

[0038] Confirming the model storage capacity of the image data model and the text data model, wherein the model storage capacity is the model size of the image data model and the text data model;

[0039] The memory includes a text memory and an image memory, wherein the text memory and the image memory respectively include a plurality of sub-storage spaces;

[0040] Confirm the remaining storage capacity of each sub-storage space;

[0041] Selecting a corresponding memory according to the image data model and the text data model;

[0042] After the memory selection is completed, the corresponding sub-storage space is selected according to the model storage capacity of the image data model and the text data model;

[0043] Wherein, the model storage amount of the image data model and the text data model is less than the remaining storage capacity of the sub-storage space;

[0044] After the image data model and the text data model are stored in the sub-storage space, the image data model and the text data model are timestamped according to the storage time;

[0045] Finally, we get the image data model and text data model stored at the same time.

[0046] Preferably, the building model storage integrated unit further includes:

[0047] Build model integration modules for:

[0048] Perform model integration of the image data model and the text data model in the memory;

[0049] The feature data in the image data model and the text data model are extracted respectively, and the image feature vector and the text feature vector are obtained after the feature data are extracted;

[0050] The extracted image feature vector and text feature vector are spliced ​​using vector splicing method;

[0051] The concatenated feature vectors are normalized. When the dimension of the concatenated feature vectors is higher than the preset dimension, PCA is used to reduce the dimension.

[0052] The normalized feature vectors are fused to obtain an integrated data model of the image data model and the text data model, and the integrated data model is labeled as an integrated operation and maintenance management model.

[0053] Preferably, the integrated model prediction and analysis unit is further used for:

[0054] Retrieve historical operation and maintenance data from the database, and match the data types of the historical operation and maintenance data with the data in the integrated operation and maintenance management model;

[0055] Compare the corresponding data types respectively, and determine whether there are abnormal areas in the integrated operation and maintenance management model based on the comparison results;

[0056] Among them, when the historical operation and maintenance data is compared with the data in the integrated operation and maintenance management model, if the comparison difference is not within the standard deviation range, the data in the integrated operation and maintenance management model is an abnormal area;

[0057] Confirm the fault feature according to the type of the abnormal area, wherein the fault feature is the abnormal value of the abnormal device in the abnormal area, and the abnormal value includes the frequency, amplitude, and duration of the device;

[0058] The fault characteristics of the abnormal area are visualized and converted through visualization tools;

[0059] After visualization conversion, abnormal operation and maintenance management data is obtained.

[0060] Preferably, the prediction and analysis data early warning unit is further used for:

[0061] Retrieve the comparison difference in the abnormal operation and maintenance management data obtained after the visualization conversion;

[0062] The fault level is determined based on the comparison difference, where the fault level is divided into level one fault, level two fault and level three fault;

[0063] Different fault levels correspond to different levels of warning, among which the warning levels are divided into level 1 warning, level 2 warning and level 3 warning;

[0064] Finally, early warning alarms of varying degrees are transmitted to the abnormal area for alarm processing.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] 1. The intelligent operation and maintenance management platform based on digital twins provided by the present invention can deeply explore the hidden information and correlation relationships in the data, improve the accuracy of data analysis and prediction, and realize the synergy of multi-dimensional data by processing the text data and image data separately and then fusing them, through the construction of deep learning models for the text data of spatial data, equipment data, environmental data and log data.

[0067] 2. The digital twin-based intelligent operation and maintenance management platform provided by the present invention confirms the model storage capacity of the image data model and the text data model, and selects appropriate memory and sub-storage space for storage, thereby realizing efficient management of the data model, and timestamping the stored image data model and text data model, so that these models have time attributes. This enhances the traceability of the data. By extracting the feature data in the image and text data models respectively and splicing these feature vectors, a feature set containing rich information can be formed, and information complementation and enhancement are realized through feature fusion, which improves the expression ability and generalization ability of the model.

[0068] 3. The digital twin-based intelligent operation and maintenance management platform provided by the present invention can accurately identify abnormal areas in the model by comparing historical operation and maintenance data with the data in the integrated operation and maintenance management model in detail, thereby improving the accuracy and reliability of anomaly detection. Once an abnormal area is found, the solution further refines and analyzes the fault characteristics, including specific parameters such as the frequency, amplitude, and duration of the abnormal equipment. According to the severity of the fault, corresponding early warning measures are taken to improve the pertinence and effectiveness of the early warning, and different degrees of early warning alarms are transmitted to the abnormal area for alarm processing, which can ensure that the operation and maintenance personnel can understand and handle the fault in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic diagram of the intelligent operation and maintenance management platform unit of the present invention;

[0070] Figure 2 It is a flow chart of the intelligent operation and maintenance management platform of the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] In order to solve the problem in the prior art that the operation and maintenance management data is not further processed after being acquired, resulting in poor data quality and the inability to build a model, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0073] The intelligent operation and maintenance management platform based on digital twins includes:

[0074] The data acquisition unit of the operation and maintenance management platform is used to:

[0075] Retrieve operation and maintenance management data from the database, classify the data according to the data type of the operation and maintenance management data, and perform data preprocessing after data classification to obtain target operation and maintenance management data;

[0076] The data model building unit of the operation and maintenance management platform is used to:

[0077] The target operation and maintenance management data is modeled using a digital twin model, where different data models are constructed according to different data types;

[0078] Build a model storage integration unit for:

[0079] Different data models are stored in different storages, and then the stored data models are integrated to obtain an integrated operation and maintenance management model;

[0080] Integrated model prediction analysis unit for:

[0081] Perform abnormal analysis on the integrated operation and maintenance management model, confirm the fault characteristics of the abnormal area according to the abnormal analysis results, and perform visual conversion on the fault characteristics of the abnormal area to obtain abnormal operation and maintenance management data after visual conversion;

[0082] Predictive analysis data early warning unit, used for:

[0083] The abnormal operation and maintenance management data is judged by fault level, different degrees of warning are issued according to different fault levels, and alarms are issued for abnormal areas based on the fault location.

[0084] Specifically, the data acquisition unit of the operation and maintenance management platform can ensure the data consistency between the digital twin model and the physical system, and improve the accuracy and credibility of the simulation. The data model construction unit of the operation and maintenance management platform can realize comprehensive and accurate analysis of image data and improve the visualization level of operation and maintenance management. The construction of the model storage integration unit helps operation and maintenance personnel to have a more comprehensive understanding of key information such as equipment status and environmental conditions, so as to make more scientific and reasonable decisions. The integrated model prediction and analysis unit improves the work efficiency of operation and maintenance personnel and reduces dependence on professional knowledge. The prediction and analysis data early warning unit transmits different degrees of early warning alarms to the abnormal area for alarm processing, which can ensure that operation and maintenance personnel can understand and handle faults in a timely manner.

[0085] The data acquisition unit of the operation and maintenance management platform includes:

[0086] Data acquisition classification module for:

[0087] Operation and maintenance management data includes space data, equipment data, environment data and log data;

[0088] Classify spatial data, equipment data, environmental data and log data into text data and image data in turn;

[0089] After the classification is completed, the text data and image data of the space data, the text data and image data of the equipment data, the text data and image data of the environment data, and the text data and image data of the log data are obtained.

[0090] Categorical data preprocessing module for:

[0091] Preprocess the text data and image data of spatial data, equipment data, environmental data and log data respectively;

[0092] Among them, the preprocessing process of text data of spatial data, equipment data, environmental data and log data is data cleaning, data format unification and data standardization;

[0093] The preprocessing process of image data for spatial data, equipment data, environmental data and log data is image denoising, image size unification, image normalization, image quality enhancement and image feature extraction;

[0094] After the text data and image data preprocessing of space data, equipment data, environment data and log data is completed, the target operation and maintenance management data is obtained.

[0095] Specifically, the image data in the data acquisition classification module (such as real-time photos of equipment, surveillance videos, space layout diagrams, etc.) can intuitively display the actual situation of the computer room, equipment or environment, and the operation and maintenance personnel can quickly obtain key information to improve decision-making efficiency and accuracy. The combination of text data (such as equipment model, parameters, operation logs, etc.) and image data (such as equipment appearance, operation status images, etc.) can provide more comprehensive and multi-dimensional information, which helps operation and maintenance personnel to have an in-depth understanding of the system status. By analyzing the image data, abnormal conditions of equipment or environment, such as equipment failure, environmental abnormalities, etc., can be identified. Combined with text data, the cause of the problem can be judged more accurately, and early warning can be given to avoid the occurrence or expansion of the failure. Artificial intelligence and big data analysis technology can be used to process text and image data to realize automated operation and maintenance, fault prediction and preventive maintenance, and improve the intelligent level of operation and maintenance management. The classification data preprocessing module can remove noise, correct errors, Steps such as processing missing values ​​ensure the quality and accuracy of input data, which is the basis for subsequent analysis and prediction. Unifying the formats of data from different sources facilitates data integration and comparison, and reduces errors and confusion caused by inconsistent formats. Standardization of text data (such as stem extraction and stop word removal) can reduce data complexity and improve processing speed. Image denoising, size unification, and normalization can optimize image data to make it more suitable for machine learning model processing and improve processing efficiency and accuracy. Image quality enhancement technology can improve image quality and make the details in the image clearer, thereby improving the accuracy and accuracy of image feature extraction. Extracting key features from the image helps the machine learning model better understand the image content, thereby improving the model's classification and recognition capabilities. It can process data from different sources (space, equipment, environment, log), support multiple data types (text, image), and provide a solid foundation for subsequent data fusion. Multi-source data fusion can provide more comprehensive information and help discover hidden associations and patterns. Through sophisticated data preprocessing, it can ensure data consistency between the digital twin model and the physical system, and improve the accuracy and credibility of the simulation.

[0096] In order to solve the problem in the prior art that the acquired operation and maintenance management data is not used to build a targeted model according to the data type, which leads to the limitation of the built model, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0097] The data model building unit of the operation and maintenance management platform is also used for:

[0098] Extracting text data and image data of spatial data, equipment data, environmental data and log data from target operation and maintenance management data respectively;

[0099] The text data of spatial data, equipment data, environmental data and log data are used to construct a text data model through a deep learning model;

[0100] The image data of spatial data, equipment data, environmental data and log data are classified using the ResNet model, target detection is performed using the YOLO model, and image segmentation is performed using the Mask R-CNN model. The image data model is obtained after image classification, target detection and image segmentation.

[0101] Specifically, by building deep learning models for the text data of spatial data, equipment data, environmental data, and log data, we can deeply explore the hidden information and associations in the data and improve the accuracy of data analysis and prediction. This provides strong support for operation and maintenance management decisions. Using the ResNet model for image classification, we can quickly identify key elements and scenes in the image, such as equipment type, environmental conditions, etc. The target detection function of the YOLO model can track and locate the target object in the image in real time, while the image segmentation capability of the Mask R-CNN model can further provide refined image analysis results. The combination of these technologies can achieve comprehensive and accurate analysis of image data and improve the visualization level of operation and maintenance management. After processing text data and image data separately and then fusing them, the synergy of multi-dimensional data can be achieved. Through real-time processing and analysis of image data, problems such as equipment failures and environmental anomalies can be discovered in a timely manner, and the source of the problem can be quickly located.

[0102] Build a model storage integration unit, including:

[0103] Build a model storage module for:

[0104] Confirm the image data model and the text data model;

[0105] Confirming the model storage capacity of the image data model and the text data model, wherein the model storage capacity is the model size of the image data model and the text data model;

[0106] The memory includes a text memory and an image memory, wherein the text memory and the image memory respectively include a plurality of sub-storage spaces;

[0107] Confirm the remaining storage capacity of each sub-storage space;

[0108] Selecting a corresponding memory according to the image data model and the text data model;

[0109] After the memory selection is completed, the corresponding sub-storage space is selected according to the model storage capacity of the image data model and the text data model;

[0110] Wherein, the model storage amount of the image data model and the text data model is less than the remaining storage capacity of the sub-storage space;

[0111] After the image data model and the text data model are stored in the sub-storage space, the image data model and the text data model are timestamped according to the storage time;

[0112] Finally, we get the image data model and text data model stored at the same time.

[0113] Build model integration modules for:

[0114] Perform model integration of the image data model and the text data model in the memory;

[0115] The feature data in the image data model and the text data model are extracted respectively, and the image feature vector and the text feature vector are obtained after the feature data are extracted;

[0116] The extracted image feature vector and text feature vector are spliced ​​using vector splicing method;

[0117] The concatenated feature vectors are normalized. When the dimension of the concatenated feature vectors is higher than the preset dimension, PCA is used to reduce the dimension.

[0118] The normalized feature vectors are fused to obtain an integrated data model of the image data model and the text data model, and the integrated data model is labeled as an integrated operation and maintenance management model.

[0119] Specifically, by constructing a model storage module to confirm the model storage capacity of the image data model and the text data model, and selecting appropriate storage and sub-storage space for storage, efficient management of the data model is achieved. This management method ensures that the data model can be stored correctly and is easy to retrieve and call later. When selecting a sub-storage space, the remaining storage capacity of each sub-storage space is taken into account to ensure the rational use of storage resources. This avoids the waste of storage space, and can flexibly adjust the storage strategy as the amount of data grows. The stored image data model and text data model are timestamped so that these models have time attributes. This enhances the traceability of the data, helps to quickly locate the data model at a specific time point when needed, helps to improve the intelligence level of the platform, and enables the platform to more accurately reflect the actual situation of operation and maintenance management. By constructing a model integration module, cross-modal fusion of image data and text data is achieved, breaking the boundaries of data types, so that the platform can comprehensively utilize information from multiple data sources. This fusion method can provide a more comprehensive and accurate operation and maintenance management perspective, and help discover problems and patterns that are difficult to reveal with a single data source. By extracting feature data from the image and text data models respectively and splicing these feature vectors, a feature set containing rich information can be formed. This feature splicing method can retain the key information of the original data while avoiding information loss. Normalizing the spliced ​​feature vectors can eliminate the dimensional differences between different features, making the feature vectors more fair and effective in subsequent processing. When the feature vector dimension is too high, using dimensionality reduction techniques such as PCA to process it can reduce the computational complexity and improve processing efficiency while retaining key information. The feature vectors after normalization and dimensionality reduction are fused to obtain an integrated data model of the image data model and the text data model. This integrated model not only contains information from the two data types of image and text, but also achieves information complementarity and enhancement through feature fusion, improves the model's expression ability and generalization ability, and helps operation and maintenance personnel to have a more comprehensive understanding of key information such as equipment status and environmental conditions, so as to make more scientific and reasonable decisions.

[0120] In order to solve the problem that the existing technology does not accurately locate the abnormal area in the operation and maintenance management data model, and does not issue targeted alarms based on abnormal situations, resulting in the inability to issue early warnings for abnormalities, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions:

[0121] The integrated model prediction analysis unit is also used for:

[0122] Retrieve historical operation and maintenance data from the database, and match the data types of the historical operation and maintenance data with the data in the integrated operation and maintenance management model;

[0123] Compare the corresponding data types respectively, and determine whether there are abnormal areas in the integrated operation and maintenance management model based on the comparison results;

[0124] Among them, when the historical operation and maintenance data is compared with the data in the integrated operation and maintenance management model, if the comparison difference is not within the standard deviation range, the data in the integrated operation and maintenance management model is an abnormal area;

[0125] Confirm the fault feature according to the type of the abnormal area, wherein the fault feature is the abnormal value of the abnormal device in the abnormal area, and the abnormal value includes the frequency, amplitude, and duration of the device;

[0126] The fault characteristics of the abnormal area are visualized and converted through visualization tools;

[0127] After visualization conversion, abnormal operation and maintenance management data is obtained.

[0128] Specifically, by making a detailed comparison between the historical operation and maintenance data and the data in the integrated operation and maintenance management model, the abnormal areas in the model can be accurately identified. In particular, when the comparison difference exceeds the preset standard deviation range, it is judged as abnormal. This method improves the accuracy and reliability of abnormality detection. Once an abnormal area is found, the solution further refines and analyzes the fault characteristics, including specific parameters such as the frequency, amplitude, and duration of the abnormal equipment. This detailed analysis helps operation and maintenance personnel to quickly locate the root cause of the problem, improve the efficiency and accuracy of fault handling, and convert the fault characteristics of the abnormal area through visualization tools, so that the originally complex data information can be presented in an intuitive and easy-to-understand way. This not only improves the work efficiency of operation and maintenance personnel, but also reduces the dependence on professional knowledge, allowing more people to participate in operation and maintenance management work, helping operation and maintenance personnel to make scientific and reasonable decisions in a timely manner and ensure the smooth progress of operation and maintenance work.

[0129] The predictive analysis data early warning unit is also used to:

[0130] Retrieve the comparison difference in the abnormal operation and maintenance management data obtained after the visualization conversion;

[0131] The fault level is determined based on the comparison difference, where the fault level is divided into level one fault, level two fault and level three fault;

[0132] Different fault levels correspond to different levels of warning, among which the warning levels are divided into level 1 warning, level 2 warning and level 3 warning;

[0133] Finally, early warning alarms of varying degrees are transmitted to the abnormal area for alarm processing.

[0134] Specifically, by retrieving the comparison difference in the abnormal operation and maintenance management data obtained after the visual conversion, the problems or faults in the operation and maintenance process can be accurately identified. This data-based fault identification method is more accurate and reliable than the traditional method that relies on manual experience. The fault level is divided into primary fault, secondary fault and tertiary fault, which helps to scientifically classify and evaluate different types of faults. This division method is not only easy for operation and maintenance personnel to understand, but also can provide clear guidance for subsequent early warning and processing. Different fault levels correspond to different degrees of early warning, including primary warning, secondary warning and tertiary warning. This differentiated early warning strategy can take corresponding early warning measures according to the severity of the fault, improve the pertinence and effectiveness of the early warning, and transmit different degrees of early warning alarms to the abnormal area for alarm processing, which can ensure that the operation and maintenance personnel can understand and handle the fault in time. This timely alarm processing mechanism helps to reduce the impact of faults on operation and maintenance work, improve operation and maintenance efficiency and service quality. Through the implementation of the above scheme, the operation and maintenance efficiency can be significantly improved. On the one hand, the automated fault identification and early warning mechanism can reduce manual intervention and reduce operation and maintenance costs; on the other hand, timely alarm processing can shorten the fault processing time and improve the stability and reliability of the system.

[0135] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0136] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. The intelligent operation and maintenance management platform based on digital twins is characterized by: include: The data acquisition unit of the operation and maintenance management platform is used to: Retrieve operation and maintenance management data from the database, classify the data according to the data type of the operation and maintenance management data, and perform data preprocessing after data classification to obtain target operation and maintenance management data; The data model building unit of the operation and maintenance management platform is used to: The target operation and maintenance management data is modeled using a digital twin model, where different data models are constructed according to different data types; Build a model storage integration unit for: Different data models are stored in different storages, and then the stored data models are integrated to obtain an integrated operation and maintenance management model; Integrated model prediction analysis unit for: Perform abnormal analysis on the integrated operation and maintenance management model, confirm the fault characteristics of the abnormal area according to the abnormal analysis results, and perform visual conversion on the fault characteristics of the abnormal area to obtain abnormal operation and maintenance management data after visual conversion; Predictive analysis data early warning unit, used for: The abnormal operation and maintenance management data is judged by fault level, different degrees of warning are issued according to different fault levels, and alarms are issued for abnormal areas based on the fault location.

2. The intelligent operation and maintenance management platform based on digital twin according to claim 1 is characterized by: The operation and maintenance management platform data acquisition unit includes: Data acquisition classification module for: Operation and maintenance management data includes space data, equipment data, environment data and log data; Classify spatial data, equipment data, environmental data and log data into text data and image data in turn; After the classification is completed, the text data and image data of the space data, the text data and image data of the equipment data, the text data and image data of the environment data, and the text data and image data of the log data are obtained.

3. The intelligent operation and maintenance management platform based on digital twin according to claim 2 is characterized in that: The operation and maintenance management platform data acquisition unit also includes: Categorical data preprocessing module for: Preprocess the text data and image data of spatial data, equipment data, environmental data and log data respectively; Among them, the preprocessing process of text data of spatial data, equipment data, environmental data and log data is data cleaning, data format unification and data standardization; The preprocessing process of image data for spatial data, equipment data, environmental data and log data is image denoising, image size unification, image normalization, image quality enhancement and image feature extraction; After the text data and image data preprocessing of space data, equipment data, environment data and log data is completed, the target operation and maintenance management data is obtained.

4. The intelligent operation and maintenance management platform based on digital twin according to claim 3 is characterized by: The operation and maintenance management platform data model building unit is also used to: Extracting text data and image data of spatial data, equipment data, environmental data and log data from target operation and maintenance management data respectively; The text data of spatial data, equipment data, environmental data and log data are used to construct a text data model through a deep learning model; The image data of spatial data, equipment data, environmental data and log data are classified using the ResNet model, target detection is performed using the YOLO model, and image segmentation is performed using the Mask R-CNN model. The image data model is obtained after image classification, target detection and image segmentation.

5. The intelligent operation and maintenance management platform based on digital twin according to claim 4 is characterized in that: The building model storage integrated unit comprises: Build a model storage module for: Confirm the image data model and the text data model; Confirming the model storage capacity of the image data model and the text data model, wherein the model storage capacity is the model size of the image data model and the text data model; The memory includes a text memory and an image memory, wherein the text memory and the image memory respectively include a plurality of sub-storage spaces; Confirm the remaining storage capacity of each sub-storage space; Selecting a corresponding memory according to the image data model and the text data model; After the memory selection is completed, the corresponding sub-storage space is selected according to the model storage capacity of the image data model and the text data model; Wherein, the model storage amount of the image data model and the text data model is less than the remaining storage capacity of the sub-storage space; After the image data model and the text data model are stored in the sub-storage space, the image data model and the text data model are timestamped according to the storage time; Finally, we get the image data model and text data model stored at the same time.

6. The intelligent operation and maintenance management platform based on digital twin according to claim 5 is characterized by: The building model storage integrated unit also includes: Build model integration modules for: Perform model integration of the image data model and the text data model in the memory; The feature data in the image data model and the text data model are extracted respectively, and the image feature vector and the text feature vector are obtained after the feature data are extracted; The extracted image feature vector and text feature vector are spliced ​​using vector splicing method; The concatenated feature vectors are normalized. When the dimension of the concatenated feature vectors is higher than the preset dimension, PCA is used to reduce the dimension. The normalized feature vectors are fused to obtain an integrated data model of the image data model and the text data model, and the integrated data model is labeled as an integrated operation and maintenance management model.

7. The intelligent operation and maintenance management platform based on digital twin according to claim 6 is characterized by: The integrated model prediction and analysis unit is further used for: Retrieve historical operation and maintenance data from the database, and match the data types of the historical operation and maintenance data with the data in the integrated operation and maintenance management model; Compare the corresponding data types respectively, and determine whether there are abnormal areas in the integrated operation and maintenance management model based on the comparison results; Among them, when the historical operation and maintenance data is compared with the data in the integrated operation and maintenance management model, if the comparison difference is not within the standard deviation range, the data in the integrated operation and maintenance management model is an abnormal area; Confirm the fault feature according to the type of the abnormal area, wherein the fault feature is the abnormal value of the abnormal device in the abnormal area, and the abnormal value includes the frequency, amplitude, and duration of the device; The fault characteristics of the abnormal area are visualized and converted through visualization tools; After visualization conversion, abnormal operation and maintenance management data is obtained.

8. The digital twin-based intelligent operation and maintenance management platform according to claim 7 is characterized by: The prediction and analysis data early warning unit is also used for: Retrieve the comparison difference in the abnormal operation and maintenance management data obtained after the visualization conversion; The fault level is determined based on the comparison difference, where the fault level is divided into level one fault, level two fault and level three fault; Different fault levels correspond to different levels of warning, among which the warning levels are divided into level 1 warning, level 2 warning and level 3 warning; Finally, early warning alarms of varying degrees are transmitted to the abnormal area for alarm processing.

Citation Information

Patent Citations

  • Article identification method for efficiently labeling samples

    CN111523610A

  • IDC intelligent operation and maintenance management system based on digital twinning

    CN115643150A

  • Multi-modal data fusion method based on image-text interaction

    CN115659279A

  • Complex industrial system operation monitoring method and system based on digital twinning

    CN115857447A

  • Grading alarm method based on universal index adaptive threshold

    CN116206427A

Cited By

  • Digital twinning-based legal compliance full life cycle management platform

    CN121413971A

  • A legal compliance full life cycle management platform based on digital twinning

    CN121413971B