Fabricated concrete building design method and system based on BIM

Abstract

The invention provides a BIM-based fabricated concrete building design method and system, and relates to the technical field of BIM, and the method comprises the steps: monitoring the internal micro-strain data of a concrete building corresponding to a target component in real time in a use stage of the target component; generating a dynamic strain distribution diagram by dynamically mapping the spatial distribution of the micro-strain data; based on historical strain data in a preset association database and real-time strain data in the dynamic strain distribution diagram, the deformation trend of the target component in the next N years is predicted in combination with a time sequence convolutional network, and N is larger than or equal to 3. And according to the deformation trend, adjusting attribute parameters between a target component and an adjacent component in a preset BIM model, and based on the adjusted attribute parameters, generating a reinforcement design scheme for the potential deformation risk in the BIM model. The early warning and prevention and control capability of the long-term deformation risk of the fabricated building node is improved.

Description

Technical Field

[0001] This application relates to the field of BIM technology, and in particular to a BIM-based design method and system for prefabricated concrete buildings. Background Technology

[0002] During long-term use, prefabricated concrete buildings are prone to hidden deformation due to factors such as load and environment. Traditional monitoring methods are difficult to detect potential risks in a timely manner. It is necessary to combine real-time data with historical trend analysis to achieve accurate deformation prediction and design optimization in order to ensure structural safety and durability.

[0003] Current research employs deformation prediction methods based on finite element analysis. These methods establish numerical models of components, combine them with real-time strain data collected by sensors, simulate the mechanical response of the components under different working conditions, and adjust design parameters based on the simulation results. This approach utilizes computer simulation technology to evaluate the long-term performance of components during the design phase.

[0004] Finite element analysis relies on precise model parameters and boundary conditions. However, uncertainties exist in material properties and connection node characteristics in actual engineering, leading to deviations between simulation results and actual deformation. Furthermore, this method involves a large amount of computation, making it difficult to achieve real-time dynamic prediction and hindering efficient support for rapid design optimization decisions. Summary of the Invention

[0005] This application provides a BIM-based design method and system for prefabricated concrete buildings to address the problems of poor early warning and low prevention and control capabilities for long-term deformation risks of prefabricated building nodes in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a BIM-based design method for prefabricated concrete buildings, comprising: During the use phase of the target component, real-time monitoring of micro-strain data inside the corresponding concrete structure is conducted. A dynamic strain distribution map is generated by dynamically mapping the spatial distribution of the micro-strain data; Based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network, the deformation trend of the target component in the next N years is predicted, where N is greater than or equal to 3. Based on the deformation trend, the attribute parameters between the target component and adjacent components in the preset BIM model are adjusted, and a reinforcement design scheme for potential deformation risks is generated in the BIM model based on the adjusted attribute parameters.

[0007] Optionally, the step of predicting the deformation trend of the target component over the next N years by combining historical strain data in a preset association database with real-time strain data in the dynamic strain distribution map and a temporal convolutional network includes: The historical strain data is divided into multiple first-period sample units according to a preset annual time window; The real-time strain data is divided into multiple second-period sample units according to a preset monthly time window; Each first-period sample unit and all corresponding second-period sample units are aligned along the time axis and then superimposed to form a mixed input sequence group. A multi-scale feature vector is generated based on the mixed input sequence group by a temporal convolutional network. The multi-scale feature vector is then input into a preset deformation prediction model to output the deformation change curve of the target component. The deformation change curve is used to characterize the deformation trend of the target component in the next N years.

[0008] Optionally, the step of generating multi-scale feature vectors based on the mixed input sequence group through a temporal convolutional network, inputting the multi-scale feature vectors into a preset deformation prediction model, and outputting the deformation change curve of the target component includes: In the hierarchical structure of the temporal convolutional network, time-dimensional convolution operations are performed layer by layer on each periodic sample unit in the mixed input sequence group. Specifically: in the first convolutional layer, the input data of the annual time window is processed, and the first feature vector reflecting the strain accumulation effect at the annual scale is output; in the second convolutional layer, the input data of the seasonal time window is processed, and the second feature vector reflecting the strain fluctuation period at the seasonal scale is output; in the third convolutional layer, the aligned input of historical strain data and real-time strain data is processed, and the third feature vector reflecting the strain evolution correlation between the two is output. The first feature vector, the second feature vector, and the third feature vector are concatenated and fused along the time axis to generate a multi-scale feature vector; The multi-scale feature vector is input into the deformation prediction model, and the deformation prediction model calculates the predicted value of the deformation variable step by step. Based on the predicted value of the deformation variable, a deformation variable change curve is formed.

[0009] Optionally, the step of inputting the multi-scale feature vector into the deformation prediction model, calculating the predicted deformation value step-by-step through the deformation prediction model, and forming the deformation change curve based on the predicted deformation value includes: In the reading module of the deformation prediction model, the multi-scale feature vectors corresponding to each prediction time point are read sequentially according to the time node order; The multi-scale feature vector is input into the parameterization function module of the deformation prediction model to calculate the predicted deformation value at each prediction time point; The deformation prediction model generation module connects the predicted deformation values ​​at all prediction time points in chronological order to generate a deformation time series. The deformation time series is input into the curve generation module of the deformation prediction model, and the deformation change curve is output.

[0010] Optionally, adjusting the attribute parameters between the target component and adjacent components in the preset BIM model according to the deformation trend, and generating a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters, includes: The deformation prediction value at each time point in the deformation change curve used to characterize the deformation trend is compared with the preset deformation threshold, and the deformation risk level at each time point is determined based on the comparison result. Based on the deformation risk level, calculate the parameter adjustment amount of the connection between the target component and adjacent components in the BIM model; Locate the connection nodes between the target component and adjacent components in the BIM model, and update the attribute parameters of the connection nodes based on the parameter adjustment amount; Based on the updated attribute parameters, a reinforcement design scheme is generated through the component connection rule base of the BIM model.

[0011] Optionally, generating a reinforcement design scheme based on the updated attribute parameters and the component connection rule base of the BIM model includes: Extract the steel reinforcement density value and grouting material mixing ratio value from the updated attribute parameters; When the density value of the steel reinforcement exceeds the preset density threshold, the reinforcement configuration rules of the component connection rule library are triggered to generate a geometric strengthening instruction set; When the mixing ratio of the grouting material exceeds the preset ratio range, the ratio optimization rule of the component connection rule library is triggered to generate a ratio optimization instruction set; The geometric reinforcement instruction set and the ratio optimization instruction set are combined to generate a deformation-resistant construction instruction set for the connection nodes; Based on the aforementioned anti-deformation construction instruction set, a reinforcement design scheme is generated.

[0012] Optionally, generating a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data includes: Establish the correspondence between the spatial location of the target component and the micro-strain data; Based on the aforementioned correspondence, the strain change rate at each spatial location point is calculated; Based on the strain change rate at each spatial location, the strain distribution pattern of the target component is generated; The strain distribution pattern is transformed into visual graphic data, which is a dynamic strain distribution map.

[0013] Secondly, this application provides a BIM-based prefabricated concrete building design system, including: The monitoring module is used to monitor the micro-strain data inside the concrete structure corresponding to the target component in real time during the use phase of the target component. The mapping module is used to generate a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data; The prediction module is used to predict the deformation trend of the target component in the next N years, where N is greater than or equal to 3, based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network. The adjustment module is used to adjust the attribute parameters between the target component and adjacent components in the preset BIM model according to the deformation trend, and generate a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters.

[0014] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor, used to execute the computer program to implement the steps of the BIM-based prefabricated concrete building design method as described in the first aspect above.

[0015] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the BIM-based prefabricated concrete building design method described in the first aspect above.

[0016] This application provides a BIM-based prefabricated concrete building design method, which includes: real-time monitoring of micro-strain data inside the concrete building corresponding to the target component during the use stage; generating a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data; predicting the deformation trend of the target component in the next N years, where N is greater than or equal to 3, based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network; adjusting the attribute parameters between the target component and adjacent components in a preset BIM model according to the deformation trend; and generating a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters.

[0017] The technical solution provided in this application has the following beneficial effects: This application enables real-time perception of the internal stress state of concrete structures, providing an accurate data foundation for subsequent analysis. It visually displays the strain distribution of various parts of the components, facilitating rapid identification of high-stress areas. By integrating historical and real-time data, it accurately predicts the long-term deformation development patterns of the components, providing a basis for design optimization. Design parameters are dynamically adjusted based on the prediction results, ensuring consistency between the model and the actual structural state. It automatically outputs targeted reinforcement measures to improve structural safety and durability.

[0018] Furthermore, this application also forms a mixed input sequence group by dividing historical strain data into annual windows and real-time data into monthly windows and aligning them over time. After extracting multi-scale features using a temporal convolutional network, the input is fed into the deformation prediction model, and finally outputs a deformation change curve that characterizes the future deformation trend.

[0019] Furthermore, this prediction method, through multi-timescale data fusion and deep feature extraction, achieves a precise grasp of the long-term deformation law of prefabricated concrete components, providing a reliable basis for structural health assessment and optimized design.

[0020] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a BIM-based prefabricated concrete building design method provided for embodiments of this application; Figure 2 A schematic diagram illustrating a specific implementation of a BIM-based prefabricated concrete building design method provided in this application embodiment; Figure 3 This is a schematic diagram illustrating another specific implementation of a BIM-based prefabricated concrete building design method provided in this application. Figure 4 This is a structural schematic diagram of a BIM-based prefabricated concrete building design system provided in an embodiment of this application. Detailed Implementation

[0023] In the current field of prefabricated concrete building structural health monitoring, existing deformation prediction methods based on finite element analysis have limitations. On the one hand, these methods heavily rely on precise material parameters and boundary condition settings, while key parameters such as concrete creep characteristics and joint connection stiffness in actual engineering exhibit time-varying uncertainties, leading to systematic deviations between simulation results and actual deformation. On the other hand, finite element calculations consume significant computational resources, making it difficult to respond promptly to changes in real-time monitoring data and failing to meet the timeliness requirements for dynamic prediction in engineering sites. This contradiction between static modeling and dynamic monitoring severely restricts the accuracy of structural safety early warning and the real-time nature of design optimization.

[0024] To address the aforementioned issues, this application proposes a prefabricated concrete building design method based on Building Information Modeling (BIM). This method innovatively constructs a closed-loop system of "monitoring-prediction-optimization": by deeply mining the spatiotemporal correlation characteristics of historical strain data and real-time monitoring data through temporal convolutional networks, a dynamic prediction model considering the time-varying properties of materials is established; based on the prediction results, the node design parameters in the BIM model are automatically corrected to generate a reinforcement scheme adapted to long-term deformation patterns. This technical solution breaks through the dependence of traditional finite element analysis on idealized modeling conditions, achieving accurate capture of structural performance evolution through a data-driven approach. Simultaneously, by utilizing the real-time interactive capabilities of the BIM platform, the prediction cycle is shortened from hours to minutes. This technical path, which deeply integrates IoT monitoring, AI prediction, and BIM design optimization, not only solves the problem that the prediction accuracy of existing methods is limited by modeling assumptions, but also achieves the real-time response capability urgently needed in engineering practice through an embedded computing framework, providing reliable technical support for the full life-cycle safety management of prefabricated buildings.

[0025] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] The core of this application is to provide a BIM-based design method for prefabricated concrete buildings, and a flowchart of one specific implementation method is shown below. Figure 1 As shown, the method includes: Step 101: During the use phase of the target component, monitor the micro-strain data inside the concrete structure corresponding to the target component in real time.

[0027] In step 101, the target component refers to the precast component unit in the prefabricated concrete building that requires key monitoring. Micro-strain data are the measurements of minute deformations collected by sensors embedded inside the concrete.

[0028] In this embodiment of the application, after the prefabricated building is put into use, the micro-strain data inside the concrete is continuously collected by the resistance strain gauge sensor network embedded in the prefabricated components according to the set sampling frequency. The sensor data is sent to the central processing unit through wired transmission. The processing unit performs filtering and noise reduction processing on the original signal, removes environmental interference factors, and stores it as a time series data sequence. These data sequences constitute the original input for subsequent analysis.

[0029] For example, in a high-rise prefabricated residential project, the precast beam-column connection node is used as the target component. Twelve strain sensors are pre-embedded during concrete pouring. The sensors are arranged in a matrix in the key stress area of ​​the node. The system collects data once per minute. After the collected raw data is processed by a low-pass filtering algorithm, the interference signal caused by temperature changes is removed, and the effective strain data generated by the load is retained to form a continuous micro-strain monitoring dataset.

[0030] Step 102: Generate a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data.

[0031] In step 102, dynamic mapping is a computational process that converts discrete point measurement data into a continuous spatial distribution. The dynamic strain distribution map is a visual representation of the strain magnitude on the component surface using color gradients.

[0032] In this embodiment of the application, based on the micro-strain dataset obtained in step 101, a correspondence table between sensor spatial coordinates and strain values ​​is first established. Then, the strain value of the area on the component surface where no sensor is placed is calculated using an inverse distance weighted interpolation algorithm. Next, the strain data of the entire component is converted into a three-dimensional cloud map with color gradients through a three-dimensional rendering engine, where red represents high strain areas and blue represents low strain areas. Finally, a dynamically updatable strain distribution visualization graphic is generated.

[0033] For example, following the previous example, using the three-dimensional coordinate data and corresponding strain values ​​of 12 sensors, the strain value of each square centimeter area on the surface of the beam-column joint is calculated by spatial interpolation. These data are input into the graphics processing module to generate a three-dimensional cloud map with red, yellow, green and blue gradient colors to represent the strain magnitude. The map shows that the lower right corner of the joint is a clear red high strain concentration area, which matches the location of the fine cracks found in the field inspection.

[0034] Step 103: Based on the historical strain data in the preset associated database and the real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network, predict the deformation trend of the target component in the next N years, where N is greater than or equal to 3.

[0035] In step 103, the preset correlation database refers to the collection of strain monitoring data from prefabricated components of assembled concrete buildings in historical projects. Based on this strain monitoring data, a correlation database between strain and time is constructed. Prefabricated components refer to standardized concrete building parts prefabricated in a factory, while target components refer to specific prefabricated components that need to be monitored and optimized in the current project. The connection between the two lies in the fact that a target component is a specific instance of a prefabricated component; the difference is that "prefabricated component" is a general term, while "target component" specifically refers to the current research object. Historical data in the correlation database refers to recorded data on the strain changes of prefabricated components over time, collected from past projects. This data is accumulated through long-term monitoring. Strain monitoring data is a generalized strain measurement result, while historical data is a selected portion stored in the database; the two are related as a whole set and a subset. The real-time data of the dynamic strain distribution map is visualized distribution information generated by spatial mapping after real-time acquisition of micro-strain within the target component by sensors. "Monitoring micro-strain data inside concrete" refers to the original measured values, while real-time data is the derived data after processing; the two are related as raw materials and finished products. Temporal convolutional networks are deep learning models specifically designed for processing time-series data. Deformation trend is a function graph that describes the future deformation development law of a component.

[0036] In this embodiment, the real-time strain data generated in step 102 is aligned with the historical strain data of similar components stored in the database according to the time dimension. The historical data is segmented by year, and the real-time data is segmented by month. The data is input into a temporal convolutional network for multi-scale feature extraction. The network first captures long-term dependent features through dilated convolutional layers, then extracts short-term fluctuation features through ordinary convolutional layers, and finally inputs the feature vectors into a fully connected layer to predict future deformation trends, outputting a curve of deformation over time.

[0037] For example, by integrating strain data from similar nodes collected over the past five years with real-time monitoring data from the current six months, and using a temporal convolutional network for analysis, the network outputs a prediction formula. ,in Represents node shape variables, t The formula represents the number of months, and calculates the curve of the deformation at this node increasing from the current 152 microstrain to 182 microstrain over the next three years.

[0038] Step 104: Based on the deformation trend, adjust the attribute parameters between the target component and adjacent components in the preset BIM model, and generate a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters.

[0039] In step 104, the parameters refer to the design values ​​of the connection parameters between the target component and adjacent components (or other building parts) in prefabricated concrete buildings, specifically including the steel reinforcement distribution density and grouting material mixing ratio at the component connection. Hidden deformation risk refers to potential deformation problems in prefabricated components of prefabricated concrete buildings that are difficult to detect immediately through conventional testing methods under long-term loads. This deformation accumulates over time and eventually affects the structural safety performance. The reinforcement design scheme is a document automatically generated by the system that includes construction details and technical requirements.

[0040] In this embodiment of the application, based on the deformation trend predicted in step 103, the system first determines the type and magnitude of the design parameters that need to be adjusted, then locates the target component and the connection nodes of its adjacent components in the BIM model, modifies the corresponding reinforcement layout parameters and material ratio parameters, and finally calls the preset reinforcement rule library to automatically generate a complete set of design scheme documents containing the layout diagram of the newly added reinforcing bars, the grouting material ratio adjustment table, and construction precautions.

[0041] For example, in response to the deformation risk shown in the prediction results, the system automatically increases the density of steel bars in the node area from 8 to 9 per square meter and the cement ratio of the grout from 30% to 32%. After updating these parameters in the BIM model, it generates a layout drawing with 4 new diagonal reinforcing bars and outputs a construction technical requirement document that states "cement usage shall not be less than 32% and mixing time shall be extended by 30 seconds".

[0042] This method, through real-time monitoring, data fusion, and intelligent prediction, achieves closed-loop management of prefabricated concrete buildings from health monitoring to design optimization, improves the accuracy of structural safety early warning and the scientific nature of reinforcement design, and provides reliable technical support for the long-term performance assurance of prefabricated buildings.

[0043] To address the accuracy issue in predicting the long-term deformation of prefabricated concrete buildings, in some embodiments, step 103: predicting the deformation trend of the target component over the next N years based on historical strain data in a preset association database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network, includes: Step 201: Divide the historical strain data into multiple first-period sample units according to a preset annual time window.

[0044] In step 201, the annual time window refers to the time period in which data is divided into units of a complete year. The first periodic sample unit is a data block containing data for 12 consecutive months, formed after the annual window division.

[0045] In this embodiment of the application, historical strain monitoring data of similar products of the target component are extracted from the associated database, and the data is divided into multiple consecutive annual data segments according to the natural year. Each data segment contains complete monitoring data for 12 months of that year, forming several annual data units for subsequent analysis.

[0046] Step 202: Divide the real-time strain data into multiple second-period sample units according to a preset monthly time window.

[0047] In step 202, the monthly time window refers to a time period in which data is divided into calendar months. The second periodic sample unit is a data block containing approximately 30 days of data after being divided into months.

[0048] In this embodiment of the application, the real-time strain monitoring data currently being collected is segmented according to natural months. Each data segment contains complete data from all monitoring points in that month, forming several monthly data units for short-term feature analysis.

[0049] Step 203: Align each first-period sample unit with all corresponding second-period sample units along the time axis and then superimpose them to form a mixed input sequence group.

[0050] In step 203, timeline alignment refers to matching historical annual data with real-time monthly data according to calendar time. The mixed input sequence group is a composite dataset formed by superimposing aligned long-period annual data and short-period monthly data.

[0051] In this embodiment of the application, each annual data unit is precisely matched with the corresponding monthly data unit according to the actual calendar date to ensure that the seasonal change characteristics can correspond. Then, the matched long and short cycle data are superimposed and combined in the time dimension to form a hybrid dataset that contains both long-term trends and short-term fluctuations.

[0052] Step 204: Generate multi-scale feature vectors based on the mixed input sequence group through a temporal convolutional network, input the multi-scale feature vectors into a preset deformation prediction model, and output the deformation change curve of the target component. The deformation change curve is used to characterize the deformation trend of the target component in the next N years.

[0053] In step 204, the multi-scale feature vector is a set of values ​​reflecting features at different time scales. The deformation change curve is a prediction result of future deformation displayed in the form of a function graph. The deformation prediction model is constructed using a multilayer perceptron neural network structure. The model construction process specifically includes: constructing a fully connected network structure containing an input layer, three hidden layers, and an output layer. The number of nodes in the input layer corresponds to the dimension of the multi-scale feature vector. The three hidden layers are each set with several neurons and undergo nonlinear transformation using the ReLU activation function. The output layer uses a linear activation function to output continuous deformation prediction values. The model training process involves collecting a large amount of historical component monitoring data to form a training sample set. The mean squared error is used as the loss function, and the Adam optimization algorithm is used for backpropagation iterative calculation. The network weight parameters are updated through multiple iterations until the model prediction error converges to a predetermined range, finally obtaining a neural network model that can accurately predict the long-term deformation trend of components.

[0054] In this embodiment, a hybrid dataset is input into a temporal convolutional network. The network first extracts annual variation features using convolutional kernels with large time steps, then extracts monthly fluctuation features using convolutional kernels with small time steps, and finally combines features of different scales into a comprehensive feature vector, which is input into a prediction model to calculate a continuous curve of future deformation variables changing over time.

[0055] Here is a specific example: This embodiment follows the aforementioned high-rise prefabricated residential project case and specifically describes the deformation prediction process for prefabricated beam-column connection nodes: The system first extracts the historical strain monitoring data of this type of node from the database over the past five years. Dividing the data into five first-cycle sample units with a 365-day annual time window, the most recent year's sample unit includes complete data from January to December 2022, with the monthly average strain value gradually increasing from 148 microstrains in January to 156 microstrains in December. Simultaneously, the real-time monitoring data from January to June 2023 is divided into six second-cycle sample units, with monthly average strain values ​​of 158 microstrain, 157 microstrain, 159 microstrain, 160 microstrain, 158 microstrain, and 156 microstrain, respectively. The data from 2022... The annual sample units and the monthly sample units from the first half of 2023 are aligned and superimposed according to the actual calendar dates to form a mixed input sequence group containing long-term trends and short-term fluctuations. The alignment method is to match the data from January 2022 with the data from January 2023, and so on. This mixed sequence group is input into a temporal convolutional network. The network first extracts interannual variation features through a large time step convolutional kernel to obtain a feature component (0.38) reflecting the concrete creep effect. Then, it extracts monthly fluctuation features through a small step convolutional kernel to obtain a feature component (0.25) reflecting the temperature effect. Finally, it analyzes the correlation between historical and real-time data to obtain a feature component (0.40). These three feature components are combined into a multi-scale feature vector [0.38, 0.25, 0.40] and input into the deformation prediction model. The model is based on the formula... The future deformation was calculated, and the predicted curve of the deformation at this node from the current 156 microstrain to 186 microstrain was obtained over 36 months. The curve shows a decay trend of increasing by 8 microstrain in the first 12 months, 5 microstrain in the middle 12 months, and 3 microstrain in the last 12 months.

[0056] In the embodiments of this application, the method achieves accurate control over the long-term deformation law of prefabricated concrete components through the fusion analysis of multi-timescale data and deep learning prediction, providing a scientific basis for structural safety assessment and preventive maintenance, and effectively improving the safety management level of the entire building life cycle.

[0057] To improve the accuracy of deformation prediction for prefabricated concrete buildings, in some embodiments, step 204 involves generating a multi-scale feature vector based on the mixed input sequence group using a temporal convolutional network, inputting the multi-scale feature vector into a preset deformation prediction model, and outputting the deformation change curve of the target component, such as... Figure 2 As shown, it includes: Step 301: In the hierarchical structure of the temporal convolutional network, perform temporal dimension convolution operations layer by layer on each periodic sample unit in the mixed input sequence group, wherein: in the first convolutional layer, the input data of the annual time window is processed, and the first feature vector reflecting the strain accumulation effect at the annual scale is output; in the second convolutional layer, the input data of the seasonal time window is processed, and the second feature vector reflecting the strain fluctuation period at the seasonal scale is output; in the third convolutional layer, the aligned input of historical strain data and real-time strain data is processed, and the third feature vector reflecting the strain evolution correlation between the two is output.

[0058] In step 301, the time-dimensional convolution operation refers to feature extraction calculations performed along the time axis. The input data specifically refers to a mixed input sequence group, which is formed by superimposing long-period sample units of historical strain data and short-period sample units of real-time strain data aligned along the time axis, containing mixed time-series data at interannual and seasonal scales. Annual time window data contains complete annual monitoring values. Seasonal time window data contains quarterly monitoring values. Aligned input refers to the combined data after matching historical and real-time data at specific time points. The aligned input of historical and real-time strain data is a portion of the mixed input sequence group that has undergone time alignment processing; specifically, it refers to a combined data segment with a temporal correspondence formed by aligning long-period sample units and short-period sample units along the time axis, rather than the complete mixed input sequence group as a whole.

[0059] In this embodiment, the temporal convolutional network first uses a large-size convolutional kernel to scan and calculate the annual data to extract the long-term deformation accumulation features of the components. Then, it uses a medium-size convolutional kernel to process the quarterly data to capture the fluctuation features caused by seasonal temperature changes. Finally, it performs convolution operations on the small-size aligned data to analyze the correlation between historical data and real-time monitoring, and outputs three feature vectors respectively.

[0060] Step 302: The first feature vector, the second feature vector, and the third feature vector are spliced ​​and fused according to the time axis to generate a multi-scale feature vector.

[0061] In step 302, splicing and fusion refers to connecting and combining feature vectors from different time scales in the time dimension.

[0062] In this embodiment, the interannual feature vector output by the first convolutional layer, the seasonal feature vector output by the second convolutional layer, and the correlation feature vector output by the third convolutional layer are concatenated in chronological order to form a long feature vector that comprehensively reflects the deformation law of the component, which is used as the input of the prediction model.

[0063] Step 303: Input the multi-scale feature vector into the deformation prediction model, calculate the predicted deformation value step by step through the deformation prediction model, and form the deformation change curve based on the predicted deformation value.

[0064] In step 303, time-step calculation refers to predicting the deformation at each time point in month order.

[0065] In the embodiments of this application, after receiving the multi-scale feature vector, the deformation prediction model first decomposes the feature components of each time scale, then calculates the predicted deformation values ​​for each future month in turn, and finally connects these discrete prediction points with a smooth curve to form a complete deformation development trend map.

[0066] Here is a specific example: This embodiment continues the deformation prediction analysis of precast beam-column connection nodes in the aforementioned high-rise prefabricated residential project. After the time-series convolutional network obtains a multi-scale feature vector composed of an interannual feature component (0.38), a seasonal feature component (0.25), and a correlation feature component (0.40), this vector is first input into the feature decomposition module of the deformation prediction model to separate and extract the three feature components. The interannual component (0.38) is used to calculate the long-term creep effect of concrete, using the formula 0.38 × ln( t ) Calculate the variable Xu, where t This represents the number of months from the current point in time. A seasonal component of 0.25 is used to calculate the periodic fluctuations caused by temperature changes, using the formula 0.25 × sin(2π). t / 12) Calculate the seasonal fluctuation amount. The correlation component 0.40 is used to calculate the cumulative correlation effect between historical data and real-time monitoring, using the formula 0.40 × t / 36 Calculate the associated impact; the model calculates the predicted deformation values ​​for each future month sequentially, starting with the first month. t Taking 1, we get an increment of 0.38×ln(1)+0.25×sin(2π×1 / 12)+0.40×1 / 36≈0.15 microstrain. Adding this to the current value of 156 microstrain, we get the first month's predicted value of 156.15 microstrain. In the second month, taking 2, we calculate an increment of 0.21 microstrain and a predicted value of 156.36 microstrain. This process continues until the 36th month, with a cumulative increment of 30 microstrain and a final predicted value of 186 microstrain. These discrete monthly predicted values ​​are connected into a smooth curve using a cubic spline interpolation algorithm to form a complete curve of deformation change. This curve shows that the deformation increases by 8 microstrain in the first 12 months, 5 microstrain in the middle 12 months, and 3 microstrain in the last 12 months, showing a typical decaying growth trend.

[0067] In the embodiments of this application, the method achieves accurate understanding of the long-term deformation law of concrete components through feature extraction and fusion prediction at multiple time scales, providing a reliable basis for structural safety assessment and effectively guiding the formulation of subsequent reinforcement design schemes.

[0068] To improve the accuracy and visualization of deformation prediction for prefabricated concrete buildings, in some embodiments, step 303 involves inputting the multi-scale feature vector into the deformation prediction model, calculating the predicted deformation value step-by-step through the deformation prediction model, and forming a deformation change curve based on the predicted deformation value, such as... Figure 3 As shown, it includes: Step 401: In the reading module of the deformation prediction model, read the multi-scale feature vectors corresponding to each prediction time point in sequence according to the time node order.

[0069] In step 401, a time node refers to a discrete time point on the future prediction timeline.

[0070] In this embodiment of the application, the deformation prediction model first divides the future time period into several equal division points according to a preset time interval, and then extracts the feature components corresponding to each time point from the input multi-scale feature vector to prepare for subsequent calculations.

[0071] Step 402: Input the multi-scale feature vector into the parameterization function module of the deformation prediction model to calculate the predicted deformation value at each prediction time point.

[0072] In step 402, the parameterized function module refers to the computational unit containing a preset calculation formula. The deformation prediction value is the deformation prediction amount at a single time point. The step-by-step process is represented in the flowchart by the following terms: sequentially reading the multi-scale feature vectors corresponding to each prediction time point according to the time node order, calculating the deformation prediction value at each prediction time point, and connecting the deformation prediction values ​​of all prediction time points according to the time axis order. The terms "in sequence," "each prediction time point," "time axis order," and "all prediction time points" together constitute a complete description of the specific implementation method of the "step-by-step" operation.

[0073] In this embodiment of the application, the characteristic components corresponding to each time point are input into a preset mathematical formula, and the long-term creep component, seasonal fluctuation component and related influence component are calculated respectively. Then, the components are added together to obtain the deformation prediction value at that time point.

[0074] Step 403: Using the generation module of the deformation prediction model, connect the predicted deformation values ​​of all prediction time points in chronological order to generate a deformation time series.

[0075] In step 403, the deformable time series is a set of discrete predicted values ​​arranged in chronological order.

[0076] In this embodiment of the application, the predicted values ​​of all time points are arranged in chronological order to form a set of discrete data points that reflect the change of deformation over time.

[0077] Step 404: Input the deformation time series into the curve generation module of the deformation prediction model and output the deformation change curve.

[0078] In step 404, the curve generation module refers to the processing unit that converts discrete data into continuous curves.

[0079] In this embodiment, spline interpolation algorithm is used to smooth discrete time series data and connect the data points to form a continuous deformation curve.

[0080] Here is a specific example: This embodiment continues to use the precast beam-column connection node in a high-rise prefabricated residential project as the analysis object. After obtaining the multi-scale feature vector composed of interannual feature component 0.38, seasonal feature component 0.25, and correlation feature component 0.40, the deformation prediction model first sets a prediction period of 36 months and divides the time into 36 time nodes. The model reading module processes each monthly node in turn. At the first monthly node, the feature vector is input into the parameterization function module, and the deformation increment of 0.15 micro-strain is calculated by the formula 0.38×ln(1)+0.25×sin(2π×1 / 12)+0.40×1 / 36. Among them, the ln(1) term calculates the long-term creep as 0, the sin term calculates the seasonal fluctuation as 0.125, and the correlation term calculates the cumulative influence as 0.011. The three are added together to obtain the total increment of 0.15 micro-strain, which is added to the initial value of 156 micro-strain to obtain the predicted value of 156.15 micro-strain for the first month. At the second monthly node, t Taking a value of 2, the calculated increment is 0.21 microstrain, with a predicted value of 156.36 microstrain. This process continues until the 12th month, when the cumulative increment is 8 microstrain, with a predicted value of 164 microstrain; at the 24th month, the cumulative increment is 13 microstrain, with a predicted value of 169 microstrain; and at the 36th month, the cumulative increment is 30 microstrain, with a final predicted value of 186 microstrain. The generation module arranges these 36 months of predicted values ​​in chronological order to form a discrete time series. Then, the cubic spline interpolation algorithm of the curve generation module connects these discrete points into a smooth curve. The final output deformation curve shows that the deformation increases rapidly by about 8 microstrains in the first 12 months, slows down by about 5 microstrains in the middle 12 months, and slows down the most by about 3 microstrains in the last 12 months.

[0081] In the embodiments of this application, the method realizes the visualization of deformation prediction results through step-by-step calculation and smoothing, providing engineers with an intuitive structural performance evolution trend diagram, which helps to identify potential risks in a timely manner and take preventive measures.

[0082] To ensure the safety of prefabricated concrete building structures and optimize reinforcement design schemes, in some embodiments, step 104 involves adjusting the attribute parameters between the target component and adjacent components in a preset BIM model based on the deformation trend, and generating a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters, including: Step 501: Compare the predicted deformation value at each time point in the deformation change curve used to characterize the deformation trend with the preset deformation threshold, and determine the deformation risk level at each time point based on the comparison results.

[0083] In step 501, the deformation threshold is a safety limit determined according to specifications. The deformation risk level represents a classification identifier of the component's safety status.

[0084] In this embodiment, the system compares the deformation of each time point in the prediction curve with a preset safety threshold, classifies the risk into three levels (low, medium, and high) based on the degree to which the threshold is exceeded, and records the risk status corresponding to each time point.

[0085] Step 502: Based on the deformation risk level, calculate the parameter adjustment amount of the connection between the target component and adjacent components in the BIM model.

[0086] In step 502, the parameter adjustment amount refers to the change value of the design parameters that need to be modified, including the increase or decrease in steel reinforcement density and the change in material ratio.

[0087] In this embodiment of the application, for each risk level, a preset parameter adjustment rule table is queried to determine the amount of steel reinforcement that needs to be added or the proportion of materials that needs to be adjusted, and the specific modification value is calculated.

[0088] Step 503: Locate the connection nodes between the target component and adjacent components in the BIM model, and update the attribute parameters of the connection nodes based on the parameter adjustment amount.

[0089] In step 503, the connection node refers to the three-dimensional model of the connection part between components.

[0090] In this embodiment of the application, the connection parts that need to be modified are accurately located in the BIM software, and the corresponding steel reinforcement layout parameters and material ratio parameters in the model are directly modified according to the calculated adjustment amount.

[0091] Step 504: Based on the updated attribute parameters, generate a reinforcement design scheme through the component connection rule base of the BIM model.

[0092] In step 504, the component connection rule base is a database that stores various reinforcement measures, and the reinforcement design scheme includes construction drawings and technical requirements.

[0093] In this embodiment of the application, the BIM software automatically matches the reinforcement measures in the rule base according to the updated parameters, and generates a complete set of design scheme documents including the newly added steel reinforcement layout drawing, material proportion table and other contents.

[0094] Here is a specific example: This embodiment continues the deformation prediction and reinforcement design process of precast beam-column connection nodes in the aforementioned high-rise prefabricated residential project. After obtaining the change curve of deformation from 152 microstrain to 182 microstrain in the next 36 months, the system first compares the deformation at each time point of the curve with the preset safety threshold of 170 microstrain. It finds that the deformation reaches 165 microstrain in the 18th month, exceeding the threshold for the first time, reaches 172 microstrain in the 24th month, and reaches 182 microstrain in the 36th month. Based on the magnitude of the exceedance, the risk is divided into three levels: the 18th-23rd month is a level 1 risk that needs to be observed, the 24th-35th month is a level 2 risk that needs to be warned, and the 36th month and above is a level 3 risk that needs to be reinforced immediately. Based on the risk level calculation parameters, the reinforcement density is adjusted by 5% for Level 1 risk (from 8 to 8.4 bars per square meter), 10% for Level 2 risk (to 8.8 bars per square meter), and 15% for Level 3 risk (to 9.2 bars per square meter). The cement ratio of the grouting material is adjusted accordingly: from 30% to 31% for Level 1 risk, to 32% for Level 2 risk, and to 33% for Level 3 risk. After accurately locating the beam-column connection nodes in the BIM model, the reinforcement density parameter is modified to 9.2 bars per square meter, and the cement ratio is adjusted to 33%. The system automatically calls the rule base to generate a reinforcement scheme: adding 6 diagonal reinforcing bars with a diameter of 16 mm in the node area, arranged at a 45-degree angle with a spacing of 200 mm. Simultaneously, the grouting material mix is ​​updated in the material parameter table to 33% cement, 40% sand, 25% gravel, and 2% additives, and a construction instruction document marked "mixing time not less than 5 minutes" is generated. The calculation process for the increase in steel reinforcement is as follows: original design 8 bars / square meter × node area 1.5 square meters × 115% ≈ 13.8 bars, rounded to 14 bars, after deducting the original 8 bars, 6 bars are added.

[0095] In the embodiments of this application, the method realizes an automated process from deformation prediction to reinforcement design through intelligent risk level determination and parameter adjustment, thereby improving the accuracy of structural safety assessment and the efficiency of design optimization.

[0096] To optimize the automation level of reinforcement design for prefabricated concrete buildings, in some embodiments, step 504: generating a reinforcement design scheme based on the updated attribute parameters and the component connection rule base of the BIM model includes: Step 601: Extract the steel reinforcement distribution density value and grouting material mixing ratio value from the updated attribute parameters.

[0097] In step 601, the steel reinforcement distribution density value refers to the number of steel reinforcements per unit area. The grouting material mixing ratio value refers to the proportion of each component material.

[0098] In this embodiment, the system reads two key design parameters, the steel reinforcement density and the grout mix ratio, from the modified BIM model parameters, providing a data basis for subsequent rule judgments.

[0099] Step 602: When the density value of the steel reinforcement exceeds the preset density threshold, the reinforcement configuration rules of the component connection rule library are triggered to generate a geometric reinforcement instruction set.

[0100] In step 602, the reinforcement configuration rules are a logical set that determines the arrangement of supplementary reinforcement based on changes in reinforcement density. The geometric reinforcement instruction set contains information such as the quantity, location, and size of the newly added reinforcement.

[0101] In this embodiment, when the read rebar density value exceeds a preset standard, the system queries the corresponding reinforcement measures in the rule base, calculates the number, arrangement position, and angle of the rebars to be added, and forms a specific rebar supplementation scheme instruction. The specific implementation process is as follows: the system compares the current mixing ratio with a preset range. When it exceeds the range, it calculates the component ratios that need to be adjusted based on the direction (too high or too low) and magnitude of the excess, including the changes in the proportions of cement, aggregate, and additives. Then, the adjusted proportion parameters are packaged into an instruction set. Example: If the cement ratio of 35% is detected, exceeding the upper limit of 34%, it is calculated that the aggregate needs to be reduced by 1% to balance the ratio, generating a proportion optimization instruction set containing 34% cement, 65% aggregate, and 1% additives.

[0102] Step 603: When the mixing ratio of the grouting material exceeds the preset ratio range, the ratio optimization rule of the component connection rule library is triggered to generate a ratio optimization instruction set.

[0103] In step 603, the proportioning optimization rules are a logical set of rules for adjusting the mix proportions based on changes in material ratios. The proportioning optimization instruction set contains new proportion data for each material component.

[0104] In this embodiment, when the cement ratio is detected to exceed the standard range, the system automatically calculates the aggregate and additive ratios that need to be adjusted, generates material mix modification instructions, and ensures that the grout performance meets the requirements. The specific implementation process is as follows: The system first compares the current rebar density value with a threshold. When it exceeds the threshold, it calculates the number and arrangement parameters of the reinforcing bars that need to be added based on the excess ratio, including the number of bars, diameter, spacing, and arrangement angle. These parameters are then packaged into an instruction set containing the geometric characteristics of the reinforcing bars. Example: If the current rebar density of 9.2 bars / m² exceeds the threshold of 9 bars / m², it calculates that two reinforcing bars with a diameter of 16mm need to be added, arranged at a spacing of 200mm and a 45-degree angle, generating a geometric strengthening instruction set containing these parameters.

[0105] Step 604: Combine the geometric strengthening instruction set with the ratio optimization instruction set to generate a deformation-resistant construction instruction set for the connection nodes.

[0106] In step 604, the anti-deformation construction instruction set is a comprehensive instruction document that integrates the adjustment requirements for both reinforcement and materials.

[0107] In this embodiment, the reinforcement supplementation scheme instruction and the material ratio modification instruction are merged into a complete structural optimization instruction file, covering all technical requirements for the reinforcement of connection nodes.

[0108] Step 605: Generate a reinforcement design scheme based on the aforementioned anti-deformation construction instruction set.

[0109] In this embodiment, the system automatically generates a new reinforcement layout drawing, an updated material ratio table, and corresponding construction process instructions based on the comprehensive instruction file, forming a complete solution that can be directly used for construction.

[0110] Here is a specific example: This embodiment continues the reinforcement design process for beam-column connection nodes in the aforementioned high-rise prefabricated residential project. After updating the parameters in the BIM model to adjust the rebar density to 9.2 bars / m² and the cement ratio of the grouting material to 33%, the system first extracts these two key values ​​from the updated parameters. When it detects that the rebar density of 9.2 bars / m² exceeds the preset threshold of 9 bars / m², it triggers the reinforcement configuration rule. Based on the excess value of 0.2 bars / m² and the node area of ​​1.5 m², it calculates that 3 more rebars are needed. Considering the stress requirements, it is finally determined to add 4 diagonal reinforcement bars with a diameter of 16 mm, generating a geometric reinforcement instruction set containing the number, diameter, angle, and arrangement of the rebars. At the same time, it detects that the cement ratio of 33% exceeds the standard range of 30-32%, triggering the mix optimization rule. The sand ratio is calculated from 40% to 39% using the formula: New Sand Ratio = Original Sand Ratio - (New Cement Ratio - Original Cement Ratio), generating a set of instructions containing... The BIM system optimizes the mix proportions of cement (33%), sand (39%), gravel (26%), and additives (2%). After merging these two sets of instructions into a deformation-resistant structural instruction set, the BIM system automatically performs the following operations: Adding one 45-degree diagonal reinforcing bar at each corner of the beam-column joint in the 3D model, with a total length of 1.2 meters and an anchorage length of 0.3 meters at both ends; updating the grout mix proportion parameters in the material database; generating construction drawings that indicate the accurate location of the newly added reinforcing bars and the grout mixing requirements: "Cement content 33%, mixing time 5 minutes"; finally, outputting a reinforcement design document containing all the above content. The calculation process for the number of newly added reinforcing bars is 9.2 bars / square meter × 1.5 square meters = 13.8 bars, while the original design had 8 bars / square meter × 1.5 square meters = 12 bars. The difference of 1.8 bars is rounded down to 2 bars, but considering symmetrical arrangement, the final determination is 4 bars. This precise parametric design method satisfies structural safety requirements while avoiding material waste.

[0111] In this embodiment of the application, the method realizes the intelligence and standardization of the reinforcement design process through parameter-driven automated rule judgment and scheme generation, which not only ensures structural safety but also improves design efficiency.

[0112] To more accurately visualize the stress state of prefabricated concrete components, in some embodiments, step 102: generating a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data, includes: Step 701: Establish the correspondence between the spatial position of the target component and the micro-strain data.

[0113] In step 701, spatial location refers to the three-dimensional coordinates of monitoring points on or inside the component surface. The correspondence is a matching table of coordinate points and strain values.

[0114] In this embodiment of the application, the system reads the three-dimensional coordinate information of the sensors embedded in the component, establishes a data table corresponding one-to-one with the position coordinates of each sensor and the strain value it collects, and forms a mapping relationship between spatial position and strain data.

[0115] Step 702: Based on the correspondence, calculate the strain change rate at each spatial location point.

[0116] In step 702, spatial location points refer to key monitoring points arranged on or inside the precast concrete component. These points collect micro-strain data in real time through sensors. Each point has X, Y, and Z coordinate values ​​determined in the component's three-dimensional spatial coordinate system, and the spacing between adjacent points is set to 0.3-0.5 meters according to the component size and monitoring accuracy requirements, collectively forming a monitoring network node covering the entire stress area of ​​the component. The strain change rate refers to how quickly the strain value between adjacent monitoring points changes with distance.

[0117] In this embodiment of the application, based on the established coordinate-strain correspondence table, the strain difference between adjacent sensors within a unit distance is calculated, and the strain gradient change of each region of the component is determined by the spatial difference algorithm.

[0118] Step 703: Generate the strain distribution pattern of the target component based on the strain change rate at each spatial location point.

[0119] In step 703, the strain distribution pattern is a continuous spatial model that reflects the overall strain condition of the component.

[0120] In this embodiment, the calculated strain change rate data at each point is used to construct a continuous strain distribution model covering the entire surface of the component using the radial basis function interpolation method, thus forming a complete expression of the stress state.

[0121] Step 704: Convert the strain distribution pattern into visual graphic data, wherein the visual graphic data is a dynamic strain distribution map.

[0122] In step 704, the visualized graphic data is a data format that converts numerical information into intuitive images.

[0123] In this embodiment of the application, the strain distribution model is input into the graphics rendering engine, and a gradient color system from blue to red is used to represent the strain magnitude, with red representing the high strain area and blue representing the low strain area, generating a dynamically updatable three-dimensional color cloud map.

[0124] Here is a specific example: This embodiment continues the monitoring process of precast beam-column connection nodes in the aforementioned high-rise prefabricated residential project. Twelve strain sensors are arranged in a matrix on the component surface in the node area. The system first establishes the correspondence between the three-dimensional coordinates of each sensor and the real-time strain value. The strain value corresponding to sensor number 5 (coordinates 1.2, 0.8, 2.5) located in the lower right corner of the node reaches 158 microstrains, while the adjacent sensor number 6 (coordinates 1.5, 0.8, 2.5) has a strain value of 152 microstrains. The two sensors are 0.3 meters apart, and the strain change rate in this area is calculated to be (158-152) / 0.3 = 20 microstrains / meter. Based on the strain change rate data of all adjacent sensor pairs, an inverse distance weighted interpolation algorithm is used to calculate the strain value per square centimeter of the node surface. The lower right corner shows the highest change rate of 25 microstrains / meter, while the upper left corner shows the lowest at 5 microstrains / meter. These data are input into the graphics processing module to generate a three-dimensional cloud map representing the strain magnitude using a gradient of red, yellow, green, and blue colors, with the red area concentrated in the lower right corner of the node.

[0125] In this embodiment of the application, the method realizes an intuitive display of the stress state of concrete components through spatial data mapping and visualization processing, which helps engineers quickly identify high-risk areas and provides an important basis for structural safety assessment.

[0126] Figure 4 This is a structural schematic diagram of a specific implementation of a BIM-based prefabricated concrete building design system provided in this application, with reference to... Figure 4 The system may include: The monitoring module 41 is used to monitor the micro-strain data inside the concrete structure corresponding to the target component in real time during the use phase of the target component.

[0127] The mapping module 42 is used to generate a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data.

[0128] The prediction module 43 is used to predict the deformation trend of the target component in the next N years, where N is greater than or equal to 3, based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network.

[0129] The adjustment module 44 is used to adjust the attribute parameters between the target component and adjacent components in the preset BIM model according to the deformation trend, and generate a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters.

[0130] The BIM-based prefabricated concrete building design system of this application embodiment is used to implement the aforementioned BIM-based prefabricated concrete building design method. Therefore, the specific implementation of the BIM-based prefabricated concrete building design system can be found in the embodiment section of the BIM-based prefabricated concrete building design method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0131] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described BIM-based prefabricated concrete building design methods.

[0132] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the BIM-based prefabricated concrete building design methods described above.

[0133] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0134] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the BIM-based prefabricated concrete building design method.

[0135] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. 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 implementations should not be considered beyond the scope of this invention.

[0136] The above provides a detailed description of a BIM-based prefabricated concrete building design method and system provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A BIM-based design method for prefabricated concrete buildings, characterized in that, include: During the use phase of the target component, real-time monitoring of micro-strain data inside the corresponding concrete structure is conducted. A dynamic strain distribution map is generated by dynamically mapping the spatial distribution of the micro-strain data; Based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network, the deformation trend of the target component in the next N years is predicted, where N is greater than or equal to 3. Based on the deformation trend, the attribute parameters between the target component and adjacent components in the preset BIM model are adjusted, and a reinforcement design scheme for potential deformation risks is generated in the BIM model based on the adjusted attribute parameters.

2. The method according to claim 1, characterized in that, The method of predicting the deformation trend of the target component over the next N years by combining historical strain data from a preset associated database with real-time strain data from the dynamic strain distribution map and a temporal convolutional network includes: The historical strain data is divided into multiple first-period sample units according to a preset annual time window; The real-time strain data is divided into multiple second-period sample units according to a preset monthly time window; Each first-period sample unit and all corresponding second-period sample units are aligned along the time axis and then superimposed to form a mixed input sequence group. A multi-scale feature vector is generated based on the mixed input sequence group by a temporal convolutional network. The multi-scale feature vector is then input into a preset deformation prediction model to output the deformation change curve of the target component. The deformation change curve is used to characterize the deformation trend of the target component in the next N years.

3. The method according to claim 2, characterized in that, The step of generating multi-scale feature vectors based on the mixed input sequence group through a temporal convolutional network, inputting the multi-scale feature vectors into a preset deformation prediction model, and outputting the deformation change curve of the target component includes: In the hierarchical structure of the temporal convolutional network, time-dimensional convolution operations are performed layer by layer on each periodic sample unit in the mixed input sequence group. Specifically: in the first convolutional layer, the input data of the annual time window is processed, and the first feature vector reflecting the strain accumulation effect at the annual scale is output; in the second convolutional layer, the input data of the seasonal time window is processed, and the second feature vector reflecting the strain fluctuation period at the seasonal scale is output; in the third convolutional layer, the aligned input of historical strain data and real-time strain data is processed, and the third feature vector reflecting the strain evolution correlation between the two is output. The first feature vector, the second feature vector, and the third feature vector are concatenated and fused along the time axis to generate a multi-scale feature vector; The multi-scale feature vector is input into the deformation prediction model, and the deformation prediction model calculates the predicted value of the deformation variable step by step. Based on the predicted value of the deformation variable, a deformation variable change curve is formed.

4. The method according to claim 3, characterized in that, The step of inputting the multi-scale feature vector into the deformation prediction model, calculating the predicted deformation value step by step through the deformation prediction model, and forming the deformation change curve based on the predicted deformation value includes: In the reading module of the deformation prediction model, the multi-scale feature vectors corresponding to each prediction time point are read sequentially according to the time node order; The multi-scale feature vector is input into the parameterization function module of the deformation prediction model to calculate the predicted deformation value at each prediction time point; The deformation prediction model generation module connects the predicted deformation values ​​at all prediction time points in chronological order to generate a deformation time series. The deformation time series is input into the curve generation module of the deformation prediction model, and the deformation change curve is output.

5. The method according to claim 1, characterized in that, The step involves adjusting the attribute parameters between the target component and adjacent components in the preset BIM model according to the deformation trend, and generating a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters, including: The deformation prediction value at each time point in the deformation change curve used to characterize the deformation trend is compared with the preset deformation threshold, and the deformation risk level at each time point is determined based on the comparison result. Based on the deformation risk level, calculate the parameter adjustment amount of the connection between the target component and adjacent components in the BIM model; Locate the connection nodes between the target component and adjacent components in the BIM model, and update the attribute parameters of the connection nodes based on the parameter adjustment amount; Based on the updated attribute parameters, a reinforcement design scheme is generated through the component connection rule base of the BIM model.

6. The method according to claim 5, characterized in that, The process of generating a reinforcement design scheme based on the updated attribute parameters and the component connection rule base of the BIM model includes: Extract the steel reinforcement density value and grouting material mixing ratio value from the updated attribute parameters; When the density value of the steel reinforcement exceeds the preset density threshold, the reinforcement configuration rules of the component connection rule library are triggered to generate a geometric strengthening instruction set; When the mixing ratio of the grouting material exceeds the preset ratio range, the ratio optimization rule of the component connection rule library is triggered to generate a ratio optimization instruction set; The geometric reinforcement instruction set and the ratio optimization instruction set are combined to generate a deformation-resistant construction instruction set for the connection nodes; Based on the aforementioned anti-deformation construction instruction set, a reinforcement design scheme is generated.

7. The method according to claim 1, characterized in that, The step of generating a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data includes: Establish the correspondence between the spatial location of the target component and the micro-strain data; Based on the aforementioned correspondence, the strain change rate at each spatial location point is calculated; Based on the strain change rate at each spatial location, the strain distribution pattern of the target component is generated; The strain distribution pattern is transformed into visual graphic data, which is a dynamic strain distribution map.

8. A BIM-based prefabricated concrete building design system, characterized in that, include: The monitoring module is used to monitor the micro-strain data inside the concrete structure corresponding to the target component in real time during the use phase of the target component. The mapping module is used to generate a dynamic strain distribution map by dynamically mapping the spatial distribution of the micro-strain data; The prediction module is used to predict the deformation trend of the target component in the next N years, where N is greater than or equal to 3, based on historical strain data in a preset associated database and real-time strain data in the dynamic strain distribution map, combined with a temporal convolutional network. The adjustment module is used to adjust the attribute parameters between the target component and adjacent components in the preset BIM model according to the deformation trend, and generate a reinforcement design scheme for potential deformation risks in the BIM model based on the adjusted attribute parameters.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the BIM-based prefabricated concrete building design method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the BIM-based prefabricated concrete building design method as described in any one of claims 1 to 7.

Images