A Method and System for Predicting Building Deformation Risk Based on Multi-Source Data Fusion

Abstract

This invention discloses a method and system for predicting building deformation risks based on multi-source data fusion, relating to the field of building risk monitoring technology. The invention first constructs a building logical image model network, digitally integrating building components, monitoring equipment, and maintenance responsibilities. Once a risk trigger source is identified through multi-source data fusion, it is automatically mapped to this model. Based on risk characteristics, the effective impact domain is dynamically calculated, and the equipment requiring operation and the corresponding responsible parties within that domain are retrieved. Finally, through structured synthesis, an electronic work order integrating equipment control and personnel scheduling is automatically generated and issued, thereby achieving automated linkage from risk warning to response execution. This solves the problems of response delays and inappropriate resource allocation caused by the disconnect between risk warning and response execution in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of building risk monitoring technology, and in particular to a method and system for predicting building deformation risks based on multi-source data fusion. Background Technology

[0002] Currently, in the field of building safety monitoring, various sensors such as tiltmeters and strain gauges are typically installed to monitor building deformation. These sensors generate a large amount of data, and existing technologies can analyze this data and issue risk alerts when the data exceeds safety thresholds.

[0003] However, in reality, there is a serious disconnect between system alarms and manual intervention. After receiving an alarm, on-duty personnel often need to manually consult blueprints, contact relevant departments, and formulate a response plan. The entire process is time-consuming and labor-intensive, resulting in a significant delay in response and failing to meet the requirements of modern buildings for rapid and accurate safety responses. Summary of the Invention

[0004] This application provides a method and system for predicting building deformation risks based on multi-source data fusion, which solves the problems of response delay and improper resource allocation caused by the disconnect between risk warning and disposal execution in the prior art, and improves the response speed and disposal accuracy of building safety operation and maintenance.

[0005] This application provides a method for predicting building deformation risk based on multi-source data fusion. This method is applied to a building deformation risk prediction system based on multi-source data fusion, and includes:

[0006] Step S1: Construct the architectural logical image model network;

[0007] Step S2: Collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector.

[0008] Step S3: Map the risk triggering source to the building logic image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logic image model network with the risk triggering source as the center and the effective influence domain as the search radius;

[0009] Step S4: Obtain detailed information about the risk triggering source, control permissions for the associated device set, and pre-set handling strategies for the affected functional components, and perform structured synthesis to generate an electronic work order, which is then distributed to the corresponding IoT control terminal and responsible person terminal through the message routing interface.

[0010] Furthermore, the specific steps of the architectural logical image model network include:

[0011] Retrieve the original geometric data file of the building information model and convert it into the node and edge structure in the graph database;

[0012] Read the unique identification code, equipment model parameters, and local coordinates of the installation location of each sensor, and use the coordinate transformation matrix algorithm to convert the local coordinates of the sensor into global coordinate system coordinates consistent with the building information model;

[0013] And map the sensor nodes to their corresponding spatial locations in the graph database;

[0014] Read the organizational structure data of the operation and maintenance responsible entity, establish a mapping table between personnel and jurisdiction, and attach the responsible person information as attribute tags to the corresponding building component nodes;

[0015] The integrity of the completed graph structure is verified, and a building logical image model network is generated.

[0016] Furthermore, the specific steps for collecting multi-source sensor monitoring data of the building and performing spatiotemporal alignment and fusion include:

[0017] The analog signals are read in parallel and converted into digital signal streams according to the original sampling frequency set inside the multi-source sensor.

[0018] The received digital signal stream is stamped with a standard timestamp to form a raw data packet with time stamps;

[0019] Receive multiple raw data packets, select the data with the highest sampling frequency as the reference time axis, and use the cubic spline interpolation algorithm to calculate the corresponding values ​​at the reference time points for the remaining data to fill the data gaps in the time series;

[0020] The pre-stored spatial registration algorithm is invoked, and the scalar data collected by sensors at different locations are mapped to a unified three-dimensional spatial vector field based on the global coordinates of each sensor in the building's logical image model. The discrete monitoring point data are then fused into continuous field data describing the overall state of the building through spatial weighted calculation.

[0021] Complete spatiotemporal alignment and fusion.

[0022] Furthermore, the specific steps for obtaining the deformation index of the monitoring point include:

[0023] The sensor time series data obtained after spatiotemporal alignment and fusion is received, and a smoothed measured value sequence is obtained by using a moving average filtering algorithm.

[0024] The initial reference value of the monitoring point is retrieved, and the difference between the measured value at the current time and the initial reference value is calculated to obtain the cumulative deformation of the monitoring point relative to the initial state.

[0025] At the same time, historical measurement data within the previous time window are extracted, and the data within the time window are linearly fitted using the least squares method. The slope of the fitted line is used as the current deformation rate of the monitoring point.

[0026] The deformation rate is subjected to a second difference operation to obtain the deformation acceleration index.

[0027] The cumulative deformation, deformation rate, and deformation acceleration are weighted and summed according to preset normalization weighting coefficients to generate a deformation index that comprehensively reflects the current state of the monitoring point.

[0028] Furthermore, the specific steps for mapping the risk triggering source to the building logical image model include:

[0029] Once the deformation index of a certain monitoring point is determined to be abnormal, the unique identification code and channel number carried in the data packet of that monitoring point are immediately read.

[0030] Using this unique identification code as the index key, a traversal query is initiated in the graph database of the building logical image model network to match nodes that are completely consistent with the unique identification code.

[0031] After locking onto the target node, read the global coordinate data of the node stored in the building logical image model network, the type identifier of the building component to which it belongs, and the hierarchical relationship of the node in the topology network.

[0032] Furthermore, the specific steps for dynamically calculating the effective influence domain of the risk triggering source based on the risk feature vector include:

[0033] First, analyze the risk feature vector of the risk triggering source and extract the deformation type component and deformation intensity component. The deformation type component indicates whether the risk is settlement, tilting or cracking, and the deformation intensity component indicates the magnitude of the value.

[0034] Starting from the risk triggering source node in the building logic mapping model network, a breadth-first search is performed along the connecting edges in the topology network to the surrounding adjacent nodes.

[0035] During the search process, the residual risk intensity when the risk intensity is transmitted to the adjacent node is calculated using the attenuation function based on the deformation type component. If the residual risk intensity is greater than the set minimum sensitivity threshold, the adjacent node is added to the effective influence domain set, and the calculation continues to propagate to the outer node from the adjacent node as a new starting point. Otherwise, the search for the path is stopped.

[0036] Repeat the above process until all propagation paths have been traversed or truncated. The final topological subgraph formed by all marked nodes is the effective influence domain of the risk triggering source.

[0037] Furthermore, the specific steps for generating the electronic work order include:

[0038] Create a blank electronic work order data object, and fill in the corresponding fields of the work order object with the retrieved detailed equipment parameters, the control instruction code segment that has passed the authorization verification, and the matched standard operation procedure text according to the predefined XML data format;

[0039] The timestamp, location coordinates, and risk level identifier of the risk occurrence are embedded in the electronic work order, and the entire data object is hashed and encrypted to generate a structured electronic work order file.

[0040] This application provides a building deformation risk prediction system based on multi-source data fusion, which is used to implement a building deformation risk prediction method based on multi-source data fusion, including: a network construction module, a vector extraction module, an information analysis module, and a work order generation module;

[0041] The network construction module is used to construct a building logical image model network.

[0042] The vector extraction module is used to collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector.

[0043] The information analysis module is used to map the risk triggering source to the building logical image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logical image model network with the risk triggering source as the center and the effective influence domain as the search radius.

[0044] The work order generation module is used to obtain detailed information about the risk triggering source, the control permissions of the associated device set, and the pre-set handling strategies for the affected functional components, and to perform structured synthesis to generate electronic work orders, which are then distributed to the corresponding IoT control terminal and the responsible person's terminal through the message routing interface.

[0045] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0046] The method and system for predicting building deformation risks based on multi-source data fusion provided in this application construct a building logical image model network; collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring points in real time, and lock the monitoring point as a risk trigger source when the deformation index exceeds a preset safety threshold; map the risk trigger source to the building logical image model, dynamically calculate the effective impact domain of the risk event based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the model network; obtain detailed information of the risk trigger source, control permissions of the related equipment sets and preset handling strategies for the affected functional components, perform structured synthesis, generate electronic work orders containing specific equipment control instructions and personnel scheduling tasks, and distribute them to the corresponding IoT control terminal and responsible person terminal through a message routing interface.

[0047] In this process, by constructing a building logical image model network, building components, monitoring equipment, and operation and maintenance responsibilities are digitally integrated, enabling risk warning information to be automatically associated with equipment and management entities within its impact range. Furthermore, when determining the scope of risk impact, by analyzing risk feature vectors and dynamically calculating the effective impact domain of risk events based on deformation type and deformation intensity components, the scope of equipment and components requiring attention and operation can be accurately defined according to the specific characteristics of the risk. This avoids over- or under-response measures due to fixed impact range settings, improving the accuracy of resource scheduling. Moreover, when generating disposal instructions, by structurally synthesizing detailed risk information, equipment control permissions, and pre-set disposal strategies, directly executable electronic work orders are automatically generated. These work orders simultaneously include control instructions from the equipment side and personnel scheduling tasks from the management side, achieving a seamless transition from risk assessment to specific operation and shortening the response time from risk identification to on-site intervention. Attached Figure Description

[0048] Figure 1 A flowchart of a building deformation risk prediction method based on multi-source data fusion provided in this application embodiment;

[0049] Figure 2 A schematic diagram of the structure of the building deformation risk prediction system based on multi-source data fusion provided in the embodiments of this application. Detailed Implementation

[0050] This application provides a method and system for predicting building deformation risks based on multi-source data fusion. This solves the problems of response delay and improper resource allocation caused by the disconnect between risk warning and disposal execution in the prior art. By constructing a building logical image model network, a unified risk analysis and disposal logical carrier is established. When a risk is triggered, the impact range is automatically calculated, related elements are retrieved, and a structured disposal work order is generated, thereby improving the response speed and disposal accuracy of building safety operation and maintenance.

[0051] In related technologies, the alarm information output by risk warning systems only includes basic data such as risk location and level, and cannot automatically associate specific response measures and responsible parties. When on-duty personnel receive an alarm, they need to manually consult building drawings to determine the scope of impact, contact relevant departments to determine liability, and formulate specific response plans. The entire process relies entirely on manual coordination and judgment, resulting in response delays and a high risk of errors. Existing risk prediction methods only complete the risk identification stage but fail to achieve automated integration with the response and execution stage.

[0052] Based on the aforementioned technical issues, this application pre-constructs a building logical image model network to digitally integrate building components, monitoring equipment, and operation and maintenance responsibilities. When a risk trigger source is identified through multi-source data fusion, it is automatically mapped to the model. Based on risk characteristics, the effective impact domain is dynamically calculated, and the equipment that needs to be operated and the corresponding responsible parties within the range are retrieved. Finally, through structured synthesis, an electronic work order integrating equipment control and personnel scheduling is automatically generated and issued, thereby realizing automated linkage from risk warning to disposal execution.

[0053] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0054] like Figure 1 As shown, this application provides a building deformation risk prediction method based on multi-source data fusion. This method is applied to a building deformation risk prediction system based on multi-source data fusion, and includes:

[0055] Step S1: Construct a building logical image model network. This model serves as a digital logical carrier corresponding to the physical building. At the data level, it predefines and stores the physical attributes of building components, the spatial distribution of monitoring equipment, and the jurisdiction of the operation and maintenance responsible entities, forming a multi-dimensional static association network.

[0056] Step S2: Collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector.

[0057] Step S3: Map the risk triggering source to the building logic image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logic image model network with the risk triggering source as the center and the effective influence domain as the search radius;

[0058] Step S4: Obtain detailed information about the risk trigger source, control permissions for the associated set of devices, and pre-set handling strategies for the affected functional components. Then, perform structured synthesis to generate an electronic work order containing specific device control instructions and personnel scheduling tasks. Distribute the work order to the corresponding IoT control terminal and the responsible person's terminal through the message routing interface.

[0059] Furthermore, the specific steps of the architectural logical image model network include:

[0060] The original geometric data file of the building information model is retrieved, and the three-dimensional coordinate vertex data, topological connection data and material property parameters of the building components in the original geometric data file are extracted by the parser. The above data is converted into node and edge structure in the graph database, where nodes represent independent building components or monitoring equipment, and edges represent the connection or spatial adjacency relationship between components.

[0061] The unique identification code, device model parameters, and local coordinates of the installation location of each sensor are read. A coordinate transformation matrix algorithm is then used to convert the sensor's local coordinates into global coordinates consistent with the Building Information Model (BIM). The formula for the coordinate transformation matrix algorithm is as follows:

[0062] [ X g Y g Z g 1 ] = T ⋅ [ X l Y l Z l 1 ] ;

[0063] in, Represents the coordinate values ​​in the global coordinate system. This represents the coordinate values ​​in the sensor's local coordinate system. It is a 4×4 homogeneous transformation matrix, including rotation matrix and translation vector, which is obtained through calibration data during sensor installation. The calibration data is calculated by measuring the relative position and orientation angle between the local coordinate system and the global coordinate system using a total station.

[0064] And map the sensor nodes to their corresponding spatial locations in the graph database;

[0065] Read the organizational structure data of the operation and maintenance responsible entity, establish a mapping table between personnel and jurisdiction, and attach the responsible person information as attribute tags to the corresponding building component nodes;

[0066] The completed graph structure is validated for integrity. Each building component node has at least one connecting edge, and each sensor node contains complete spatial location attributes and data transmission interface definitions. A building logical image model network is generated, which is a multi-dimensional topology network containing building entity category identifiers, sensing device names, and hierarchical relationships.

[0067] Furthermore, the specific steps for collecting multi-source sensor monitoring data of the building and performing spatiotemporal alignment and fusion include:

[0068] Inclinometers, strain gauges, crack gauges, and hydrostatic levels installed throughout the building are connected in parallel via fieldbus and wireless transmission protocols. The analog signals are read in parallel according to the original sampling frequency set inside the multi-source sensors and converted into digital signal streams.

[0069] A unified standard timestamp is sent to all data acquisition units via the Network Time Protocol, and the received digital signal stream is stamped with the standard timestamp to form a raw data packet with a time stamp.

[0070] Multiple raw data packets are received, and the data packet with the highest sampling frequency is selected as the reference time axis. For the remaining data packets, a cubic spline interpolation algorithm is used to calculate the corresponding values ​​at the reference time points. The cubic spline interpolation function is:

[0071] ;

[0072] in, Indicates time The interpolation results, and The timestamps of adjacent known data points, coefficients The coefficients are obtained by solving the three bending moment equations, ensuring the continuity of the function values, first derivative, and second derivative. These coefficients are calculated based on the known data points and boundary conditions.

[0073] Fill in the data gaps in the time series to achieve alignment along the time dimension;

[0074] A pre-stored spatial registration algorithm is invoked to map scalar data collected by sensors at different locations onto a unified three-dimensional spatial vector field based on the global coordinates of each sensor in the building's logical image model. Then, through spatial weighting calculations, the discrete monitoring point data is fused into continuous field data describing the overall state of the building. The spatial weighting formula is as follows:

[0075] ;

[0076] in, Representing a spatial point The fusion field value, Indicates the first Sensor data from each monitoring point Point With monitoring points The Euclidean distance between them is calculated using the global coordinates of each point, and the weights are... Inversely proportional to the square of the distance to emphasize the influence of nearest neighbor data;

[0077] Complete spatiotemporal alignment and fusion.

[0078] Furthermore, the specific steps for obtaining the deformation index of the monitoring point include:

[0079] The sensor time series data obtained after spatiotemporal alignment and fusion is received, and the moving average filtering algorithm is used to remove high-frequency random noise interference in the data to obtain a smoothed measured value sequence.

[0080] The initial reference value of the monitoring point is retrieved, and the difference between the current measured value and the initial reference value is calculated to obtain the cumulative deformation of the monitoring point relative to the initial state. ,in This represents the measured value at the current moment. Indicates the initial reference value;

[0081] Simultaneously, historical measurement data from the previous time window are extracted, and linear fitting of the data within that time window is performed using the least squares method. The slope of the fitted line is used as the current deformation rate of the monitoring point, as shown in the formula:

[0082] ;

[0083] in, Indicates the deformation rate. and These represent the times within the time window. The timestamps and measurements of each data point. and These represent the time and the average of all data points within the time window, respectively. The number of data points within the window is represented by these parameters, which are extracted directly from the historical data sequence.

[0084] Subsequently, a second difference calculation was performed on the deformation rate to obtain the deformation acceleration index. ,in and These represent the deformation rates at the current and previous moments, respectively.

[0085] Finally, the cumulative deformation, deformation rate, and deformation acceleration are weighted and summed according to preset normalization weighting coefficients to generate a dimensionless deformation index value that comprehensively reflects the current state of the monitoring point. The weighting coefficient Obtained from historical data through principal component analysis, and satisfying the following conditions: ;

[0086] This value will be used as the sole quantitative criterion for determining whether a risk has been triggered and stored in a real-time database.

[0087] Furthermore, the specific steps for mapping the risk triggering source to the building logical image model include:

[0088] Once the deformation index of a certain monitoring point is determined to be abnormal, the unique identification code and channel number carried in the data packet of that monitoring point are immediately read.

[0089] Using this unique identification code as the index key, a traversal query is initiated in the graph database of the building logical image model network to match nodes that are completely consistent with the unique identification code.

[0090] After locking the target node, read the global coordinate data of the node stored in the building logical image model network, the type identifier of the building component to which it belongs, and the hierarchical relationship of the node in the topology network;

[0091] In the visualization rendering layer of the building logic image model network, the highlight attribute of the 3D model primitive corresponding to the node is set to the on state, and the color attribute of the node is modified to a specific color value representing the warning state.

[0092] At the same time, the status bit of the node in the graph database is activated, marking it as the starting point of risk diffusion, and the historical data snapshot associated with the node is frozen, linking the snapshot to the risk trigger source record table.

[0093] Furthermore, the specific steps for dynamically calculating the effective influence domain of the risk triggering source based on the risk feature vector include:

[0094] First, analyze the risk feature vector of the risk triggering source and extract the deformation type component and deformation intensity component. The deformation type component indicates whether the risk is settlement, tilting or cracking, and the deformation intensity component indicates the magnitude of the value.

[0095] The remaining risk intensity is obtained by using the attenuation function model based on the deformation type component. The attenuation function model is used to reflect the loss rate of risk energy when it is transferred in the building structure.

[0096] The attenuation function model is as follows:

[0097]

[0098] in, Indicates the transmission distance The remaining strength of the risk afterward The initial intensity of the risk trigger source is represented by the deformation intensity component of the risk feature vector. This represents the attenuation coefficient, which is retrieved from the rule base based on the deformation type component. The rule base is pre-built based on historical risk trigger source data and structural mechanics simulations. The topological distance along the risk propagation path is calculated using the number of connecting edges between nodes in the building logic mapping model.

[0099] Starting from the risk triggering source node in the building logic mapping model network, a breadth-first search is performed along the connecting edges in the topology network to the surrounding adjacent nodes.

[0100] During the search process, the residual risk intensity when the risk intensity is transmitted to the adjacent node is calculated using the attenuation function based on the deformation type component. If the residual risk intensity is greater than the set minimum sensitivity threshold, the adjacent node is added to the effective influence domain set, and the calculation continues to propagate to the outer node from the adjacent node as a new starting point. Otherwise, the search for the path is stopped.

[0101] Repeat the above process until all possible propagation paths have been traversed or cut off. The final topological subgraph formed by all marked nodes is the effective influence domain of the risk triggering source.

[0102] Furthermore, the specific steps for generating the electronic work order include:

[0103] Access the equipment asset database, and based on the risk triggering source and the unique identifier of the equipment in the effective impact domain, retrieve the manufacturer's technical manual, the maximum rated operating parameters, and the current operating status code for each equipment. At the same time, query the access control list to confirm the current remote write permission level of these equipment.

[0104] Based on the type code of the affected functional components, the corresponding emergency response plan text is matched in the expert knowledge base. This text contains the standard operating procedures for that type of component at a specific risk level.

[0105] Create a blank electronic work order data object, and fill in the corresponding fields of the work order object with the retrieved detailed equipment parameters, the control instruction code segment that has passed the authorization verification, and the matched standard operation procedure text according to the predefined XML data format;

[0106] The timestamp, location coordinates, and risk level identifier of the risk occurrence are embedded in the electronic work order, and the entire data object is hashed and encrypted to generate a structured electronic work order file with tamper-proof properties.

[0107] The specific methods for obtaining the preset security threshold include:

[0108] First, read the material mechanical property parameters from the building structure design specifications, including the standard values ​​of the compressive strength of concrete and the yield strength of steel reinforcement. Then, calculate the maximum allowable deformation of each component under the ultimate bearing capacity state according to the building structure design code, and define this value as the theoretical limit threshold. ;

[0109] Next, historical monitoring data for the building over the past full year was retrieved, and the average value of the data at each monitoring point was calculated. with standard deviation Using the principle of normal distribution in statistics, the statistical fluctuation threshold is set as the sum of the mean and three standard deviations. ,in and Calculated directly from historical data sequences;

[0110] Subsequently, using finite element simulation analysis software, the actual load distribution of the current building is input, and the critical deformation value at which the structure exhibits plastic damage is simulated and calculated, which is defined as the simulation early warning threshold. ;

[0111] Finally, the minimum value among the theoretical limit threshold, statistical fluctuation threshold, and simulation early warning threshold is selected, and a safety factor less than one is introduced. The minimum value is scaled, and the resulting value is used as the final preset safety threshold. Among them, the safety factor The values ​​are obtained from a preset table based on the building's service life and environmental factors, with a range of 0.7 to 0.9, and are written into the parameter table of the system's configuration file.

[0112] like Figure 2 As shown, this application provides a building deformation risk prediction system based on multi-source data fusion to implement the building deformation risk prediction method based on multi-source data fusion, including: a network construction module, a vector extraction module, an information analysis module, and a work order generation module;

[0113] The network construction module is used to construct a building logical image model network.

[0114] The vector extraction module is used to collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector.

[0115] The information analysis module is used to map the risk triggering source to the building logical image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logical image model network with the risk triggering source as the center and the effective influence domain as the search radius.

[0116] The work order generation module is used to obtain detailed information about the risk triggering source, the control permissions of the associated device set, and the pre-set handling strategies for the affected functional components, and to perform structured synthesis to generate electronic work orders, which are then distributed to the corresponding IoT control terminal and the responsible person's terminal through the message routing interface.

[0117] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0118] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0119] Those skilled in the art will recognize that the modules and algorithm steps of the various examples 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 implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0120] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0121] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0122] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting building deformation risk based on multi-source data fusion, characterized in that, Includes the following steps: Step S1: Construct the architectural logical image model network; Step S2: Collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector. Step S3: Map the risk triggering source to the building logic image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logic image model network with the risk triggering source as the center and the effective influence domain as the search radius; Step S4: Obtain detailed information about the risk triggering source, control permissions for the associated device set, and pre-set handling strategies for the affected functional components, and perform structured synthesis to generate an electronic work order, which is then distributed to the corresponding IoT control terminal and responsible person terminal through the message routing interface.

2. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps of the building logical image model network include: Retrieve the original geometric data file of the building information model and convert it into the node and edge structure in the graph database; Read the unique identification code, equipment model parameters, and local coordinates of the installation location of each sensor, and use the coordinate transformation matrix algorithm to convert the local coordinates of the sensor into global coordinate system coordinates consistent with the building information model; And map the sensor nodes to their corresponding spatial locations in the graph database; Read the organizational structure data of the operation and maintenance responsible entity, establish a mapping table between personnel and jurisdiction, and attach the responsible person information as attribute tags to the corresponding building component nodes; The integrity of the completed graph structure is verified, and a building logical image model network is generated.

3. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps for collecting multi-source sensor monitoring data of the building and performing spatiotemporal alignment and fusion include: The analog signals are read in parallel and converted into digital signal streams according to the original sampling frequency set inside the multi-source sensor. The received digital signal stream is stamped with a standard timestamp to form a raw data packet with time stamps; Receive multiple raw data packets, select the data with the highest sampling frequency as the reference time axis, and use the cubic spline interpolation algorithm to calculate the corresponding values ​​at the reference time points for the remaining data to fill the data gaps in the time series; The pre-stored spatial registration algorithm is invoked, and the scalar data collected by sensors at different locations are mapped to a unified three-dimensional spatial vector field based on the global coordinates of each sensor in the building's logical image model. The discrete monitoring point data are then fused into continuous field data describing the overall state of the building through spatial weighted calculation. Complete spatiotemporal alignment and fusion.

4. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps for obtaining the deformation index of the monitoring point include: The sensor time series data obtained after spatiotemporal alignment and fusion is received, and a smoothed measured value sequence is obtained by using a moving average filtering algorithm. The initial reference value of the monitoring point is retrieved, and the difference between the measured value at the current time and the initial reference value is calculated to obtain the cumulative deformation of the monitoring point relative to the initial state. At the same time, historical measurement data within the previous time window are extracted, and the data within the time window are linearly fitted using the least squares method. The slope of the fitted line is used as the current deformation rate of the monitoring point. The deformation rate is subjected to a second difference operation to obtain the deformation acceleration index. The cumulative deformation, deformation rate, and deformation acceleration are weighted and summed according to preset normalization weighting coefficients to generate a deformation index that comprehensively reflects the current state of the monitoring point.

5. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps for mapping the risk triggering source to the building logical image model include: Once the deformation index of a certain monitoring point is determined to be abnormal, the unique identification code and channel number carried in the data packet of that monitoring point are immediately read. Using this unique identification code as the index key, a traversal query is initiated in the graph database of the building logical image model network to match nodes that are completely consistent with the unique identification code. After locking onto the target node, read the global coordinate data of the node stored in the building logical image model network, the type identifier of the building component to which it belongs, and the hierarchical relationship of the node in the topology network.

6. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps for dynamically calculating the effective influence domain of the risk triggering source based on the risk feature vector include: First, analyze the risk feature vector of the risk triggering source and extract the deformation type component and deformation intensity component. The deformation type component indicates whether the risk is settlement, tilting or cracking, and the deformation intensity component indicates the magnitude of the value. Starting from the risk triggering source node in the building logic mapping model network, a breadth-first search is performed along the connecting edges in the topology network to the surrounding adjacent nodes. During the search process, the residual risk intensity when the risk intensity is transmitted to the adjacent node is calculated using the attenuation function based on the deformation type component. If the residual risk intensity is greater than the set minimum sensitivity threshold, the adjacent node is added to the effective influence domain set, and the calculation continues to propagate to the outer node from the adjacent node as a new starting point. Otherwise, the search for the path is stopped. Repeat the above process until all propagation paths have been traversed or truncated. The final topological subgraph formed by all marked nodes is the effective influence domain of the risk triggering source.

7. The building deformation risk prediction method based on multi-source data fusion as described in claim 1, characterized in that, The specific steps for generating an electronic work order include: Create a blank electronic work order data object, and fill in the corresponding fields of the work order object with the retrieved detailed equipment parameters, the control instruction code segment that has passed the authorization verification, and the matched standard operation procedure text according to the predefined XML data format; The timestamp, location coordinates, and risk level identifier of the risk occurrence are embedded in the electronic work order, and the entire data object is hashed and encrypted to generate a structured electronic work order file.

8. A building deformation risk prediction system based on multi-source data fusion, used to implement the building deformation risk prediction method based on multi-source data fusion as described in any one of claims 1-7, characterized in that, include: Network construction module, vector extraction module, information analysis module, and work order generation module; The network construction module is used to construct a building logical image model network. The vector extraction module is used to collect multi-source sensor monitoring data of the building and perform spatiotemporal alignment and fusion, calculate the deformation index of the monitoring point in real time, and lock the monitoring point as a risk trigger source when the deformation index of the monitoring point exceeds the preset safety threshold, and extract its risk feature vector. The information analysis module is used to map the risk triggering source to the building logical image model, dynamically calculate the effective influence domain of the risk triggering source based on the risk feature vector, and traverse and retrieve all affected related equipment sets, functional component sets and corresponding responsible entities in the building logical image model network with the risk triggering source as the center and the effective influence domain as the search radius. The work order generation module is used to obtain detailed information about the risk triggering source, the control permissions of the associated device set, and the pre-set handling strategies for the affected functional components, and to perform structured synthesis to generate electronic work orders, which are then distributed to the corresponding IoT control terminal and the responsible person's terminal through the message routing interface.

Images