Structure disease automatic identification and early warning method and system based on BIM + Internet of Things technology

Abstract

The invention provides a structure disease automatic identification and early warning method and system based on a BIM + Internet of Things technology, and the system comprises a BIM model platform which creates and manages a digital three-dimensional model of a building structure or infrastructure based on a BIM model, and the BIM model comprises detailed component information; the Internet of Things sensor network comprises a plurality of sensors; the data acquisition module is used for acquiring physical data of a building structure or infrastructure in real time through an Internet of Things sensor network to obtain structure state data, and transmitting the structure state data through a wireless or wired network; the data processing and analyzing module is used for receiving the structure state data, performing association processing based on the component information and the structure state data, and analyzing and judging the type and degree of the structure disease through an algorithm; and the disease identification and early warning module performs disease identification by using big data analysis, machine learning and a physical model, automatically sends out early warning information after the disease is identified, and generates a report.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building information modeling and Internet of Things, and particularly relates to a structure disease automatic identification and early warning method and system based on BIM+Internet of Things technology. BACKGROUND

[0002] At present, in the use process of buildings or infrastructures, structure health monitoring (SHM) is paid more and more attention. The traditional structure health monitoring method relies on manual detection, such as periodic inspection or the use of single monitoring equipment, which often cannot provide real-time and continuous data, and it is difficult to find problems in time when the structure has a sudden failure. The building information model based on BIM, combined with Internet of Things technology, can realize the whole life cycle monitoring and disease automatic identification of building structures by deploying sensors at key positions and monitoring data in real time.

[0003] BIM technology can provide detailed three-dimensional digital models for buildings, and Internet of Things technology can obtain real-time structure state information by deploying a sensor network. Therefore, the combination of BIM and Internet of Things can effectively detect and identify diseases in buildings or infrastructures, and update them in real time to the BIM model, thereby providing managers with visual display of structure health status and maintenance decision support. SUMMARY

[0004] The present application is carried out to solve the above problems, and aims to provide a structure disease automatic identification and early warning method and system based on BIM+Internet of Things technology.

[0005] The present application provides a structure disease automatic identification and early warning system based on BIM+Internet of Things technology, which has the following characteristics: a BIM model platform, based on a BIM model, creates and manages a digital three-dimensional model of a building structure or infrastructure, and the BIM model contains detailed component information; an Internet of Things sensor network, including a plurality of sensors, arranged at key positions of the building structure; a data acquisition module, which acquires physical data of the building structure or infrastructure in real time through the Internet of Things sensor network, obtains structure state data, and transmits the structure state data through wireless or wired network; a data processing and analysis module, which receives the structure state data, performs associated processing based on the component information and the structure state data, and analyzes and judges the type and degree of structure disease through algorithm; a disease identification and early warning module, which uses big data analysis, machine learning and physical model, combines historical data and design parameters in the BIM model to identify diseases, and automatically sends early warning information and generates a report after identifying the diseases.

[0006] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the historical data analysis and trend prediction module can be used for storing structural state data to form a complete record of historical data, facilitating the review and trend prediction of historical data by using big data analysis methods, thereby early warning of diseases.

[0007] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the key parts of the building structure or infrastructure include the positions of beams, columns, pile foundations and load-bearing walls.

[0008] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the component information includes structural design parameters and material parameters, and the construction information is dynamically updated according to the structural state data.

[0009] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the sensor includes a strain sensor, a displacement sensor, a pressure sensor, a temperature and humidity sensor, the displacement sensor is used to collect the settlement, inclination and lateral displacement data of the building structure or infrastructure, the strain sensor is used to monitor the stress change inside the building structure or infrastructure in real time, so as to judge whether there is stress concentration or structural damage phenomenon in combination with the structural design parameters, the temperature and humidity sensor is used to monitor the change of environmental temperature and humidity, and the possibility of diseases such as thermal expansion and cold contraction caused by temperature difference or corrosion caused by humidity is judged in combination with the material parameters.

[0010] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the data acquisition module can acquire data according to a preset time interval or an event triggering mechanism.

[0011] In the BIM+ Internet of Things technology-based structural disease automatic identification and early warning system provided by the application, the data processing and analysis module can be accurately associated with the component code, spatial coordinates and attribute data,

[0012] The data processing and analysis module can automatically identify and classify the disease types inside the building structure or infrastructure by real-time processing of the input image through the deep learning target detection algorithm based on the YOLO series, obtain the identification result, reverse the actual physical size of the structural disease according to the disease area size, confidence and component attribute in the identification result, and set a multi-level judgment threshold, thereby realizing the quantitative analysis of the disease type and degree and obtaining the analysis result.

[0013] In the structural disease automatic identification and early warning system based on BIM+Internet of Things technology provided by the application, the disease identification and early warning module can further have the following features: the disease identification includes the following steps:

[0014] The target detection task is converted into a regression problem, the boundary box and the category probability of the disease area are directly predicted in the image, the input image is divided into SxS grids, each grid predicts B boundary boxes and the confidence score of each target, and each boundary box predicts a vector:

[0015] (x, y, w, h, c)

[0016] Wherein:

[0017] x, y: position coordinates of the center point of the boundary box relative to the grid unit;

[0018] w, h: width and height of the boundary box (normalized to the image size);

[0019] c: confidence score, calculated as follows:

[0020] c = P obj · IoU pred，truth

[0021] P is the probability of the existence of the target obj and the product of the intersection over union between the predicted box and the real box,

[0022] In the structural disease identification, the YOLO model uses a multi-layer convolutional neural network to extract image features, combines a feature pyramid network for multi-scale detection, and finally outputs the boundary box and the disease type, and its loss function can be divided into three parts:

[0023] L total = alpha1 x L conf + alpha2 x L cls + alpha3 x L bbox

[0024] Wherein L conf : confidence loss, used to determine whether there is a disease;

[0025] L cls : classification loss, used to identify the disease type (such as cracks, corrosion, leakage, etc.);

[0026] L bbox : boundary box positioning loss;

[0027] Alpha1, alpha2, alpha3: the weighted coefficients of the loss term, the position, category and confidence of the disease area are output quickly, combined with the image resolution and the component spatial information in the BIM model, the actual size of the disease can be further deduced, and three-dimensional spatial positioning and visual display can be realized, so as to build an intelligent identification and early warning mechanism of structural diseases under the digital twin platform.

[0028] In the structural disease automatic identification and early warning system based on BIM+Internet of Things technology provided by the application, the sending mode of the early warning information can include short message, email or platform notification to the management personnel, and the early warning information can include the specific disease type, position and severity in the building structure or infrastructure.

[0029] The application also provides a structural disease automatic identification and early warning method based on BIM+Internet of Things technology, which has the following characteristics: S1, arranging Internet of Things sensors at key positions of the building structure or infrastructure; S2, collecting the structural state data of the building structure or infrastructure in real time; S3, matching and updating the structural state data with the component information in the BIM model; S4, identifying the structural disease through data analysis algorithm and automatically generating early warning information.

[0030] Effects of the application

[0031] The structural disease automatic identification and early warning method and system based on BIM+Internet of Things technology according to the application have the following beneficial effects:

[0032] Real-time monitoring and data linkage: the structural state is monitored in real time through the sensor, and the data is linked with the BIM model, so that the state information of the structure is always updated.

[0033] Automatic disease identification: based on the Internet of Things sensor data, the machine learning algorithm is used to automatically identify the structural disease, reducing the error rate and time consumption of manual detection.

[0034] Early warning mechanism and rapid response: the system has flexible early warning configuration, and when the structural anomaly is detected, the early warning is automatically triggered to help the management personnel to discover and handle the problem in time.

[0035] Dynamic visualization and accurate positioning: the specific position and severity of the disease are visualized and displayed through the BIM model, so that the maintenance work is more intuitive and efficient. DETAILED DESCRIPTION

[0036] Figure 1 It is the main page diagram of the structural disease automatic identification and early warning system based on BIM+Internet of Things technology in the embodiment of the application;

[0037] Figure 2It is the access, processing, release flow chart of the digital three-dimensional model in the embodiment of the application;

[0038] Figure 3 It is a three-dimensional electronic sand table map of the prefabricated and assembled superimposed station of rail transit in the embodiment of the application;

[0039] Figure 4 It is the BIM model data principle and standard of the prefabricated and assembled superimposed structure in the embodiment of the application;

[0040] Figure 5 It is the hydration heat measuring point data detected by the Internet of Things sensor network in the embodiment of the application;

[0041] Figure 6 It is the shrinkage strain measuring point data detected by the Internet of Things sensor network in the embodiment of the application;

[0042] Figure 7 It is the flow chart of the data acquisition module in the embodiment of the application;

[0043] Figure 8 It is the identification process chart of the yolo model in the embodiment of the application;

[0044] Figure 9 It is the early warning rule configuration of the disease identification and early warning module in the embodiment of the application;

[0045] Figure 10 It is the SMS early warning push chart of the disease identification and early warning module in the embodiment of the application. DETAILED DESCRIPTION

[0046] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the following embodiments combine the drawings to specifically describe the structure disease automatic identification and early warning method and system based on BIM+Internet of Things technology of the application.

[0047] Figure 1 It is the main page chart of the structure disease automatic identification and early warning system based on BIM+Internet of Things technology in the embodiment of the application. Figure 2 It is the access, processing, release flow chart of the digital three-dimensional model in the embodiment of the application.

[0048] As Figures 1-2 shown, the structure disease automatic identification and early warning system based on BIM+Internet of Things technology in the embodiment includes a BIM model platform, an Internet of Things sensor network, a data acquisition module, a data processing and analysis module, a disease identification and early warning module, and a historical data analysis and trend prediction module.

[0049] Figure 3It is a three-dimensional electronic sand table map of the rail transit prefabricated and assembled superimposed station in the embodiment of the application.

[0050] As Figure 3 shown, the BIM model platform is based on a BIM model, which contains detailed component information and creates and manages a digital three-dimensional model of a building structure or infrastructure. The model not only contains detailed component information (such as the geometric size of the component, the material type, the mechanical property parameter, etc.), but also is constructed by using multi-source data fusion and intelligent modeling technology.

[0051] Figure 4 It is the BIM model data principle and standard of the prefabricated and assembled superimposed structure in the embodiment of the application.

[0052] As Figure 4 shown, in the embodiment, BIM model coding requirements and standard specifications are proposed in terms of file format, naming rules, coordinate system, etc., which provide a basis for subsequent model component and dynamic data model association. In terms of file format, the open IFC4.0 standard is uniformly adopted to ensure the cross-platform compatibility and scalability of the model; in terms of naming rules, the component naming follows the mode of "project code + floor number + component type + serial number" (such as "DX001-B2-COL-015" represents the 15th column in the B2 layer of the project DX001), and a unified letter coding is set for the component type (such as COL for column, BEAM for beam, and SLAB for slab); in terms of coordinate system, the model adopts the national geodetic coordinate system CGCS2000, and the origin, elevation reference and axis direction are unified to ensure the accurate alignment of all components in space position. Each component needs to be embedded with standardized attribute fields, including component ID, size, material, construction time, sensor number, etc., to realize the accurate correspondence between semantic data and monitoring data. These coding specifications not only improve the data consistency and engineering adaptability of the BIM model, but also lay a technical foundation for the realization of three-dimensional visualization, structure perception analysis and digital twin platform construction.

[0053] The key parts of the building structure or infrastructure include the positions of beams, columns, pile foundations and load-bearing walls. The component information includes structural design parameters and material parameters, and the construction information is dynamically updated according to the structural state data.

[0054] And the current health status of the structure is intuitively displayed in the three-dimensional model.

[0055] Figure 5 It is the hydration heat measurement point data detected by the Internet of Things sensor network in the embodiment of the application. Figure 6 It is the shrinkage strain measurement point data detected by the Internet of Things sensor network in the embodiment of the application.

[0056] As Figures 5-6As shown, the Internet of Things (IoT) sensor network includes several sensors deployed in key parts of the building structure.

[0057] The sensors include strain sensors, displacement sensors, pressure sensors, temperature and humidity sensors, ensuring comprehensive monitoring of the structure's health status.

[0058] Displacement sensors are used to collect data on settlement, tilt, and lateral displacement of building structures or infrastructure.

[0059] Strain sensors are used to monitor stress changes inside building structures or infrastructure in real time, thereby determining whether there is stress concentration or structural damage in combination with structural design parameters.

[0060] Temperature and humidity sensors are used to monitor changes in ambient temperature and humidity, and combined with material parameters, to determine the likelihood of thermal expansion and contraction caused by temperature differences or corrosion caused by humidity.

[0061] Figure 7 This is a flowchart of the data acquisition module in an embodiment of the present invention.

[0062] like Figure 7 As shown, the data acquisition module collects physical data of building structures or infrastructure in real time through an Internet of Things (IoT) sensor network to obtain structural status data, and transmits the structural status data through a wireless or wired network.

[0063] Enables the access and management of data from sensors, supporting the rapid and secure access of commonly used sensors in underground spaces and engineering, such as displacement, strain, pressure, stress, ambient temperature, humidity, and wind speed.

[0064] The data acquisition module collects data according to preset time intervals or event triggering mechanisms.

[0065] The data processing and analysis module receives the structure state data, performs correlation processing based on the component information and the structure state data, and analyzes and judges the type and degree of the structure disease through an algorithm. The data processing and analysis module is used to receive structure state data such as structure surface images, component strain, displacement, vibration and other monitoring information, and accurately associate the component code, spatial coordinates and attribute data in the BIM model. On this basis, the module processes the input image in real time through the deep learning target detection algorithm based on the YOLO series, automatically identifies and classifies the structure disease types such as cracks, corrosion, leakage and cement shedding. According to the disease area size, confidence and component properties in the identification result, the system reverses the actual physical size of the disease combined with the image resolution, and sets multiple judgment thresholds (such as cracks wider than 0.3mm and confidence higher than 0.85 are judged as moderate cracks), thereby realizing the quantitative analysis of the disease type and degree. The analysis result is finally bound with the component ID, which can be used for structure risk assessment, early warning triggering and subsequent maintenance decision support.

[0066] Figure 8 is the identification process diagram of the yolo model in the embodiment of the application.

[0067] As shown in Figure 8 , in the identification process of the yolo model, the input image on the left is sent to the backbone network after image preprocessing (such as scaling, normalization). The backbone network is composed of multiple convolutional layers (represented by blue cubes in the figure), which extract local features of the image layer by layer. This process uses C3 modules for feature extraction, and then the extracted multi-scale features are sent to the SPPF (Spatial Pyramid Pooling Fast) module, which is a lightweight feature enhancement module. The main function is to perform multiple maximum pooling operations (such as 5x5, 9x9, 13x13) on the input feature map, and then concatenate the pooling results with the original feature map, thereby fusing multi-scale spatial information, expanding the receptive field of the model, and enhancing the perception ability of different size targets. In the feature fusion stage, the network uses upsampling and feature concatenation operations to fuse the detailed features of the shallow layer with the semantic features of the deep layer. The result after feature fusion is processed again through C3 modules or C3TR modules, where C3TR is a module that introduces the Transformer mechanism, which is used to improve the model's understanding ability of long-distance dependencies and complex texture features (such as cracks, peeling, etc.). Then, each output path is connected to the feature extraction module and is finally used for disease target detection. Multi-scale outputs (upper, middle and lower three layers) correspond to the detection of different size targets (such as small cracks, large area corrosion, etc.), which improves the comprehensiveness and accuracy of identification.

[0068] Figure 9 is the early warning rule configuration of the disease identification and early warning module in the embodiment of the application. Figure 10It is a short message early warning pushing diagram of the disease identification and early warning module in the embodiment of the application.

[0069] As Figures 9-10 shown, the disease identification and early warning module utilizes big data analysis, machine learning and physical models, combines historical data and design parameters in the BIM model to identify diseases, determine disease types such as structural cracks, deformation, settlement, etc., and automatically issue early warning information after identifying diseases to generate reports. To achieve efficient identification of structural diseases, the target detection task is converted into a regression problem, and the bounding box and class probability of the disease area are directly predicted in the image. The main idea is to divide the input image into SxS grids, and each grid predicts B bounding boxes and the confidence score of each target class. Specifically, each bounding box prediction vector is:

[0070] (x, y, w, h, c)

[0071] Wherein:

[0072] x, y: position coordinates of the center point of the bounding box relative to the grid unit;

[0073] w, h: width and height of the bounding box (normalized to image size);

[0074] c: confidence score, calculated as follows:

[0075] c = P obj · IoU pred，truth

[0076] i.e. the product of the probability P obj of the target existing and the intersection over union between the predicted box and the real box.

[0077] In structural disease identification, the YOLO model uses a multi-layer convolutional neural network to extract image features, combines a feature pyramid network for multi-scale detection, and finally outputs the bounding box and disease class. Its loss function can be divided into three parts:

[0078] L total = α1×L conf + α2×L cls + α3×L bbox

[0079] Wherein L conf : confidence loss, used to determine whether there is a disease;

[0080] L cls : classification loss, used to identify disease types (such as cracks, corrosion, leakage, etc.);

[0081] L bbox : bounding box positioning loss;

[0082] α1, α2, α3: are the weighting coefficients of the loss term. The system can quickly output the location, category and confidence of the disease area. Combined with the image resolution and the component spatial information in the BIM model, the actual size of the disease can be further inferred, and three-dimensional spatial positioning and visualization display can be realized, thereby constructing an intelligent identification and early warning mechanism for structural diseases in the digital twin platform.

[0083] The system automatically identifies abnormal conditions by analyzing the monitored displacement, strain, stress and environmental data. When the set threshold is exceeded, the system sends an early warning to the management personnel through SMS, email or platform notification. The warning information will be displayed in conjunction with the BIM model, and the disease location will be directly marked in the BIM model, showing the specific disease type, location and severity.

[0084] The specific location of the disease can be visualized in the BIM model, and corresponding maintenance recommendations, including repair measures and monitoring schemes, can be provided to help management personnel make decisions.

[0085] The historical data analysis and trend prediction module is used to store structural state data to form a complete record of historical data, facilitating the review and trend prediction of historical data using big data analysis methods, thereby providing early warning of diseases.

[0086] The application also discloses a structural disease automatic identification and early warning method based on BIM+Internet of Things technology, which is carried out through the above-mentioned structural disease automatic identification and early warning system based on BIM+Internet of Things technology, and comprises the following steps:

[0087] S1, arranging Internet of Things sensors at key positions of the building structure or infrastructure.

[0088] S2, collecting structural state data of the building structure or infrastructure in real time.

[0089] S3, matching and updating the structural state data with the component information in the BIM model.

[0090] S4, identifying structural diseases through data analysis algorithms and automatically generating early warning information.

[0091] Those skilled in the art should understand that the application is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the application. Without departing from the spirit and scope of the application, various changes and improvements can be made to the application, and these changes and improvements all fall within the scope of the claimed application. The scope of protection of the application is defined by the appended claims and their equivalents.

Claims

1. A structural disease automatic identification and early warning system based on BIM+Internet of Things technology, characterized in that, Comprise: A BIM model platform, based on a BIM model, creates and manages a digital three-dimensional model of a building structure or infrastructure, which contains detailed component information; An Internet of Things sensor network, including a number of sensors, arranged at key positions of the building structure; A data acquisition module, which acquires physical data of the building structure or infrastructure in real time through the Internet of Things sensor network, obtains structure state data, and transmits the structure state data through a wireless or wired network; A data processing and analysis module that receives the structure state data, correlates the component information and the structure state data, and analyzes the type and degree of structural damage through an algorithm; A disease identification and early warning module that uses big data analysis, machine learning, and physical models, combined with historical data and design parameters in the BIM model, to identify diseases and automatically issue warning information and generate reports after identifying diseases.

2. The BIM+Internet of Things technology-based structural disease automatic identification and early warning system according to claim 1, characterized in that, Also include: A historical data analysis and trend prediction module for storing structure state data to form a complete record of historical data, facilitating the use of big data analysis methods to review and predict trends in the historical data, thereby providing early warning of diseases.

3. The structural disease automatic identification and early warning system based on BIM+Internet of Things technology according to claim 1, wherein: wherein The key positions of the building structure or infrastructure include the positions of beams, columns, pile foundations, and load-bearing walls.

4. The structural disease automatic identification and early warning system based on BIM+Internet of Things technology according to claim 1, wherein: wherein The component information includes structural design parameters, material parameters, component codes, spatial coordinates, and attribute data, and the component information is dynamically updated according to the structure state data.

5. The structural disease automatic identification and early warning system based on BIM+Internet of Things technology according to claim 4, wherein: wherein The sensors include strain sensors, displacement sensors, pressure sensors, temperature and humidity sensors, The displacement sensors are used to collect settlement, inclination, and lateral displacement data of the building structure or infrastructure, The strain sensors are used to monitor the stress changes inside the building structure or infrastructure in real time, thereby combining the structural design parameters to determine whether there is stress concentration or structural damage, The temperature and humidity sensors are used to monitor changes in environmental temperature and humidity, and in combination with the material parameters, to determine the possibility of thermal expansion and contraction caused by temperature differences or corrosion caused by humidity.

6. The structural disease automatic identification and early warning system based on BIM+Internet of Things technology according to claim 1, wherein: wherein The data acquisition module collects data according to a preset time interval or event trigger mechanism.

7. The structural disease automatic identification and early warning system based on BIM+Internet of Things technology according to claim 4, wherein: wherein, The data processing and analysis module is accurately associated with the component codes, spatial coordinates, and attribute data, The data processing and analysis module automatically identifies and classifies the disease types inside the building structure or infrastructure by real-time processing of the input image based on the YOLO series deep learning target detection algorithm, obtains an identification result, reverses the actual physical size of the structure disease according to the disease area size, confidence and component attribute in the identification result, and sets a multi-level judgment threshold, thereby realizing quantitative analysis of the disease type and degree and obtaining an analysis result.

8. The BIM+Internet of Things technology-based structure disease automatic identification and early warning system according to claim 1, characterized in that: wherein In the disease identification and early warning module, the disease identification comprises the following steps: The target detection task is converted into a regression problem, and the boundary box and class probability of the disease area are directly predicted in the image. The input image is divided into SxS grids, each grid predicts B boundary boxes and confidence scores of each target class, and each boundary box prediction vector is: (x, y, w, h, c) wherein: x, y: position coordinates of the center point of the boundary box relative to the grid unit; w, h: width and height of the boundary box (normalized to the image size); c: confidence score, calculated as follows: c = P obj • IoU pred，truth i.e. the probability P that the target exists obj the product of the intersection over union between the predicted and the real box, In structure disease identification, the YOLO model uses a multi-layer convolutional neural network to extract image features, combines a feature pyramid network for multi-scale detection, and finally outputs the boundary box and disease type. Its loss function can be divided into three parts: L total = a1 x L conf + a2 x L cls + a3 x L bbox wherein L conf : confidence loss for determining whether a disease exists; L cls : classification loss for identifying disease types (e.g., crack, exfoliation, leakage, etc.); L bbox : bounding box localization loss; α1, α2, α3: weighting coefficients of the loss term, quickly output the position, class and confidence of the disease area, combine the image resolution and the component space information in the BIM model, further reverse the actual size of the disease, and realize three-dimensional spatial positioning and visual display, thereby building an intelligent identification and early warning mechanism for structure diseases under the digital twin platform.

9. The BIM+Internet of Things technology-based structure disease automatic identification and early warning system according to claim 1, characterized in that: wherein The sending mode of the early warning information includes short message, email or platform notification to the management personnel, The early warning information includes the specific disease type, location and severity in the building structure or infrastructure.

10. A structural disease automatic identification and early warning method based on BIM+Internet of Things technology, characterized in that, Comprising the following steps: S1, arranging Internet of Things sensors at key positions of the building structure or infrastructure; S2, real-time collection of structure state data of the building structure or infrastructure; S3, matching and updating the structure state data with the component information in the BIM model; S4, identifying structure diseases through data analysis algorithms and automatically generating early warning information.

Images