Fabricated composite floor slab cast-in-place section span beam joint design method based on digital twinning

Abstract

The invention discloses an assembly type composite floor slab cast-in-place section span beam joint design method based on digital twinning, and relates to the technical field of intelligent construction, the method comprises the following steps: adopting a least square optimization algorithm to obtain a material performance correction parameter and a geometric compensation parameter; based on the historical material performance correction parameters and the geometric compensation parameters, training the structure model through a genetic algorithm to generate a digital twin reference model; inputting the material performance correction parameters and the geometric compensation parameters into the digital twin reference model, performing multi-scale finite element analysis, and outputting a multi-scale node performance evaluation report; identifying a multi-scale node performance influence factor and a sensitive area by adopting a response surface method, outputting a design parameter sensitivity analysis result, and converting the multi-scale node performance influence factor into a preliminary node design scheme through a design parameter conversion algorithm; according to the method, the real-time accuracy of node design is remarkably improved through least square dynamic parameter correction.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction technology, and in particular to a design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins. Background Technology

[0002] In recent years, prefabricated buildings have developed rapidly, with the design of beam-span joints in cast-in-place composite floor slabs being particularly crucial. Traditional methods rely on empirical formulas and simplified finite element analysis, making it difficult to accurately account for material property fluctuations and installation deviations during construction. Although structural modeling technology has achieved parametric modeling, it remains limited to static analysis and lacks real-time data feedback capabilities. While digital twin technology can achieve dynamic simulation, existing research in the field of prefabricated joints focuses primarily on macroscopic performance evaluation and has not yet established a method that integrates multi-source data, multi-scale analysis, and intelligent optimization.

[0003] Traditional methods rely on empirical formulas and static finite element analysis, making it difficult to respond in real time to construction deviations and material property fluctuations. While structural modeling technology enables parametric modeling, it lacks dynamic data feedback. Digital twin technology, although capable of dynamic simulation, still has limitations in multi-scale coupled analysis and intelligent optimization. Existing research mostly employs offline correction strategies, failing to achieve closed-loop optimization between construction and design, and machine learning methods are insufficient in multi-objective collaborative optimization. Therefore, there is an urgent need to establish a dynamic optimization method that integrates real-time monitoring, multi-scale simulation, and intelligent algorithms to improve the accuracy and adaptability of node design. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins to solve the problem of insufficient dynamic coordination between construction data and models.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a design method for cross-beam nodes of cast-in-place sections of prefabricated composite floor slabs based on digital twins, comprising,

[0008] Real-time data collection and pre-processing of material state data and precast slab installation deviation data for the cast-in-place beam-span joint construction process are performed to obtain joint construction data. A least-squares optimization algorithm is used to obtain material performance correction parameters and geometric compensation parameters. Based on historical material performance correction parameters and geometric compensation parameters, a genetic algorithm is used to train the structural model, generating a digital twin baseline model. The material performance correction parameters and geometric compensation parameters are input into the digital twin baseline model for multi-scale finite element analysis, outputting a multi-scale joint performance evaluation report. Response surface methodology is used to identify multi-scale joint performance influencing factors and sensitive areas, outputting design parameter sensitivity analysis results. A design parameter transformation algorithm is then used to convert the multi-scale joint performance influencing factors into preliminary joint design schemes. Finally, a multi-objective particle swarm optimization algorithm is used to optimize the preliminary joint design schemes to obtain the optimal joint design scheme.

[0009] As a preferred embodiment of the design method for cross-beam nodes of cast-in-place prefabricated composite floor slabs based on digital twins as described in this invention, the material state data of the cast-in-place section includes concrete temperature data, strain data, shrinkage rate data, and elastic modulus data.

[0010] The installation deviation data of the precast floor slab includes the axial offset, elevation deviation, torsion angle and joint misalignment data of the precast floor slab;

[0011] The preprocessing includes statistical thresholding to remove outliers, interpolation algorithms to fill in missing values, ICP algorithms to unify coordinates of multi-source data, principal component analysis to extract features, and Z-score standardization to unify the format.

[0012] As a preferred embodiment of the design method for the cross-beam joint of prefabricated composite floor slabs based on digital twins described in this invention, the following steps are taken: A least squares optimization algorithm is used to obtain material property correction parameters and geometric compensation parameters.

[0013] Based on the node construction data, the least squares optimization algorithm is used for iterative calculation. The node construction data and the calculation results are compared and analyzed to obtain the difference.

[0014] Based on the difference, iterative correction is performed using material parameterization calculation functions and geometric coordinate transformation functions to output material property correction parameters and geometric compensation parameters.

[0015] As a preferred embodiment of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins described in this invention, the step of training the structural model using a genetic algorithm based on historical material performance correction parameters and geometric compensation parameters to generate a digital twin baseline model is as follows:

[0016] The historical material performance correction parameters and geometric compensation parameters are divided into training set, validation set and test set;

[0017] Based on the training set, the structural model is optimized through multiple generations of selection, crossover, and mutation using a genetic algorithm. The validation set is used to monitor model performance and prevent overfitting. The test set is used to evaluate the generalization ability of the final model, generating a digital twin benchmark model.

[0018] As a preferred embodiment of the digital twin-based design method for cast-in-place beam-span joints of prefabricated composite floor slabs according to the present invention, the specific steps for outputting a multi-scale joint performance evaluation report are as follows:

[0019] Input the material property correction parameters and geometric compensation parameters into the digital twin benchmark model to perform overall structural analysis and obtain the stress state of the nodes.

[0020] Based on the stress location of the node, analyze the concrete damage and steel reinforcement stress, and generate a multi-scale node performance evaluation report.

[0021] As a preferred embodiment of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins described in this invention, the step of using response surface methodology to identify multi-scale node performance influencing factors and sensitive areas, and outputting sensitivity analysis results for design parameters, includes the following specific steps.

[0022] Based on the multi-scale node performance evaluation report, principal component analysis was used to extract the comprehensive performance evaluation index of the nodes.

[0023] Using orthogonal experimental design, the performance response relationship between material property correction parameters, geometric compensation parameters and node comprehensive performance evaluation index is generated. Then, using variance decomposition method, the sensitivity index of material property correction parameters and geometric compensation parameters to node comprehensive performance evaluation index is calculated, and the sensitivity analysis results of design parameters are obtained.

[0024] As a preferred embodiment of the digital twin-based design method for cast-in-place beam-span joints of prefabricated composite floor slabs according to the present invention, the step of converting multi-scale joint performance influencing factors into preliminary joint design schemes through a design parameter conversion algorithm is as follows:

[0025] Based on the results of the sensitivity analysis of design parameters, multi-scale node performance influencing factors are screened, and the entropy weight method is used to quantify the design importance, forming multi-scale node performance influencing factors with priority.

[0026] Based on the priority-based multi-scale node performance impact factors, a preliminary node design scheme is transformed through a conversion rule base.

[0027] As a preferred embodiment of the digital twin-based design method for cast-in-place beam-span joints in prefabricated composite floor slabs according to the present invention, the preliminary joint design scheme is optimized using a multi-objective particle swarm optimization algorithm to obtain the optimal joint design scheme. The specific steps are as follows:

[0028] A high-dimensional parameter mapping encoding method is adopted to encode the adjustable parameters in the initial node design scheme into particle position vectors, and the particle swarm is randomly initialized to calculate the multi-target fitness value of each particle.

[0029] By using the particle swarm optimization algorithm, the particle positions and velocities are updated, the Pareto optimal solution is selected, and the optimal node design scheme is obtained.

[0030] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in the first aspect of the present invention.

[0031] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in the first aspect of the present invention.

[0032] The beneficial effects of this invention are as follows: Firstly, the real-time accuracy of node design is significantly improved through least-squares dynamic parameter correction, enabling the digital twin model to dynamically track changes in material properties and component installation deviations during construction, overcoming the shortcomings of traditional static models that cannot reflect actual working conditions. Secondly, multi-objective particle swarm optimization greatly enhances the overall economic efficiency of node solutions, intelligently balancing multiple objectives such as stress performance, material cost, and construction convenience while ensuring structural safety, breaking through the limitations of quantifying optimization indicators in manual experience-based design. The synergistic application of these two technologies improves the controllability of the construction process and enhances the adaptability of the design scheme, establishing an intelligent technical system for prefabricated building node design that combines precise control capabilities with multi-objective optimization capabilities. Attached Figure Description

[0033] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1This is a flowchart of a design method for cross-beam nodes in cast-in-place prefabricated composite floor slabs based on digital twins.

[0035] Figure 2 A flowchart for data preprocessing.

[0036] Figure 3 A flowchart for generating a digital twin baseline model.

[0037] Figure 4 A flowchart for multi-objective particle swarm optimization. Detailed Implementation

[0038] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0039] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0040] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0041] Reference Figure 1~Figure 4 This embodiment provides a design method for beam-span joints in cast-in-place prefabricated composite floor slabs based on digital twins, including the following steps:

[0042] S1. Real-time acquisition of material status data of cast-in-place section and installation deviation data of precast beam during the construction process of cross-beam joint of cast-in-place section, and preprocessing to obtain joint construction data.

[0043] S1.1 The material condition data for the cast-in-place section includes the temperature data, strain data, shrinkage rate data, and elastic modulus data of the concrete.

[0044] It should be noted that, in the data acquisition of material state for the cast-in-place section, the following key data were obtained through exemplary implementation: Temperature field distribution data was acquired using DS18B20 sensors with an accuracy of ±0.5℃, embedded in a 30cm×30cm grid, and collected in CSV format. For example, the internal temperature gradient of the floor slab 12 hours after pouring was measured as 25℃ at the surface → 65℃ at a depth of 20cm; Three-dimensional strain data was acquired using an array of FBG-OS3100 fiber optic sensors and HBM-LY41 strain gauges, collected in JSON format, exemplarily capturing the maximum tensile strain at the bottom of the beam during the initial setting stage as 156με (theoretical value 142με); Free shrinkage rate data was obtained by monitoring shrinkage deformation for 28 days using a Schaevitz HR-050 LVDT, collected in XML format, exemplarily showing a measured shrinkage rate of 0.032% with a laboratory error ≤0.004%; combined with the exemplary Proceq... The Pundit ultrasonic transducer with a wave velocity of 4120 m / s and the ZBL-S210 rebound hammer with an exemplary strength conversion value of 48.5 MPa in the test area yielded dynamic elastic modulus data in binary format. The exemplary result shows an elastic modulus of 35.2 GPa, with an error of <1.5% compared to the static load test.

[0045] S1.2 The precast floor slab installation deviation data includes the axial offset, elevation deviation, torsion angle and joint misalignment data of the precast floor slab.

[0046] It should be noted that in the data collection of precast floor slab installation deviations, key monitoring data were obtained through exemplary implementation: A Leica TS60 total station with an angle measurement accuracy of 0.5" and a distance measurement accuracy of 0.6mm + 1ppm was used to monitor axial offset. For example, the lateral deviation of the axial axis of precast floor slab #3 was measured to be +8.5mm, within the design allowable value of ±10mm. The data was collected in GEOJSON format and transmitted to MongoDB for storage via the OPC UA interface. A Trimble Dini03 electronic level was used to detect elevation deviation with a round-trip difference of 0.3mm per kilometer. For example, the elevation deviation at the end of beam section B2 was recorded as 6.2mm lower. This data was collected in CSV format and transmitted to InfluxDB for storage via the MQTT protocol. A SICK DT35 tilt sensor with a range of ±15° and a resolution of 0.001° was used to obtain the torsion angle. For example, the horizontal rotation deviation of beam C5 was captured to be 0.8°. This data was collected in JSON format and transmitted to TimescaleDB for storage via the Modbus TCP interface. Keyence was used... The LK-H020 laser displacement meter measures joint misalignment with a resolution of 0.01mm. For example, it displays a misalignment of 2.1mm / 2m length between adjacent precast component joints. The data is acquired in binary format and transmitted to PostgreSQL for storage via RS485 interface.

[0047] S1.3 Preprocessing includes statistical thresholding to remove outliers, interpolation algorithms to fill in missing values, ICP algorithm to unify coordinates of multi-source data, principal component analysis to extract features, and Z-score standardization to unify the format.

[0048] It should be noted that the statistical threshold for outlier removal is achieved by analyzing the statistical distribution characteristics of the material condition data of the cast-in-place section and the installation deviation data of the precast beam, and setting a dynamic threshold range based on the interquartile range. Abnormal material condition data of the cast-in-place section and installation deviation data of the precast beam that exceed the threshold range are removed. The interpolation algorithm for missing values ​​is used to complete the missing parts of the material condition data of the cast-in-place section and the installation deviation data of the precast floor slab. The cubic spline interpolation algorithm is used to reconstruct the missing parts of the material condition data of the cast-in-place section and the installation deviation data of the precast beam by curve fitting based on the adjacent valid material condition data of the cast-in-place section and the installation deviation data of the precast beam.

[0049] The ICP algorithm unifies the coordinates of multi-source data by uniformly registering the material state data of cast-in-place sections and the installation deviation data of precast beams acquired by different acquisition devices to the building information reference coordinate system through the iterative nearest point algorithm. The principal component analysis method extracts features by orthogonally transforming the high-dimensional material state data of cast-in-place sections and the installation deviation data of precast beams to extract the principal component dimensions that best reflect the variation characteristics of the material state data of cast-in-place sections and the installation deviation data of precast slabs. The Z-score standardization method unifies the format by calculating the mean and standard deviation of the material state data of cast-in-place sections and the installation deviation data of precast beams, and converting all the material state data of cast-in-place sections and the installation deviation data of precast beams into standardized values ​​that conform to the standard normal distribution to obtain the node construction data.

[0050] S2. The least squares optimization algorithm is used to obtain the material property correction parameters and geometric compensation parameters.

[0051] S2.1 Based on the node construction data, the least squares optimization algorithm is used for iterative calculation. The node construction data and the calculation results are compared and analyzed to obtain the difference.

[0052] It should be noted that, based on the preprocessed node construction data, a least squares optimization algorithm is used to establish the objective function, with the material state data of the cast-in-place section and the installation deviation data of the precast beam as input variables. In the initialization phase, initial estimates are assigned to the material performance correction parameters and geometric compensation parameters. The minimum value of the objective function is solved through an iterative calculation process. Each iteration includes three key calculation steps: calculating the theoretical prediction value under the current parameter combination, evaluating the residual between the predicted and measured values, and adjusting the parameters to minimize the sum of squared residuals. During each iteration, the currently calculated material performance correction parameters and geometric compensation parameters are substituted into the structural analysis calculation process, and the node performance prediction value is output based on the finite element analysis method. The node performance prediction value is compared item by item with the preprocessed node construction data. The comparison includes material state parameters such as concrete temperature, strain, shrinkage rate, and elastic modulus, as well as installation deviation parameters such as axis offset, elevation deviation, torsion angle, and joint misalignment. The difference is obtained by calculating the absolute difference between the measured and predicted values.

[0053] S2.2 Based on the difference, iterative correction is performed through material parameterization calculation function and geometric coordinate transformation function to output material property correction parameters and geometric compensation parameters.

[0054] It should be noted that, as one embodiment of the present invention, the expression of the material parameterization calculation function can be:

[0055] f(p)=[ E(1+ D E) α (1+ Da ) n (1+ Dn ) ] ;

[0056] in, It is a material parameterization calculation function. It is a vector of material property correction parameters. It is the material elastic modulus correction factor. It is the initial elastic modulus of the material. It is the correction amount for the coefficient of thermal expansion of the material. It is the initial thermal expansion coefficient of the material. It is the Poisson's ratio correction factor. It is the initial Poisson's ratio of the material;

[0057] It should be noted that, as one embodiment of the present invention, the expression for the geometric coordinate transformation function is:

[0058] ;

[0059] in, It is a geometric coordinate transformation function. It is a geometric compensation parameter vector. It is a rotation matrix. It is the corner compensation amount. They are geometric coordinates. It is the geometric coordinate compensation amount. It is the translation of the construction coordinate system;

[0060] It should be noted that, as one embodiment of the present invention, the expressions for the output modified material property correction parameters and geometric compensation parameters are as follows:

[0061] [ D p D g ]= H - 1 [ J m ⊤ W m ( y m - f(p)) A ⊤ W g ( y g - T(g)) ] ;

[0062] in, These are material property correction parameters. These are geometric compensation parameters. It is the coupling matrix of the inverse operation. It is the Jacobian matrix of the material constitutive model. It is the matrix transpose. It is a material data weight matrix. It is the measured material response vector. These are relevant measurement data of material properties. It is a material parameterization calculation function. It is a geometric constraint matrix. It is the geometric compensation parameter vector The weight matrix, It is the measured geometric compensation parameter vector coordinates It is a geometric coordinate transformation function;

[0063] It should be noted that the material parameterization calculation function typically uses polynomial regression or exponential functions to represent material properties such as the elastic modulus and shrinkage rate of concrete as functions of adjustable parameters. The geometric coordinate transformation function, on the other hand, is based on the principle of homogeneous coordinate transformation and constructs a composite transformation matrix that includes basic transformations such as translation, rotation, and scaling. In each iteration, the difference obtained in the previous step is first input into the parameterization calculation function and the geometric coordinate transformation function. For material performance parameters, the difference is decomposed into individual material parameter terms through sensitivity analysis, and the function coefficients are adjusted using the gradient descent method.

[0064] For geometric parameters, the optimal transformation parameters are solved using the least squares registration algorithm. An adaptive step-size control strategy is employed during the correction process. When the rate of change of the difference over three consecutive iterations is less than a threshold, the adjustment step size is automatically reduced to improve convergence accuracy. After each adjustment, the finite element analysis is re-executed, and the newly generated nodal performance predictions are compared and verified with the measured data, forming a closed-loop feedback mechanism of "calculation-comparison-correction". The entire iterative process continues until the dual convergence conditions are met: first, the root mean square error of the difference is less than 5% of the preset tolerance (exemplarily taken as the measured value); second, the change over five consecutive iterations is less than 1%. Finally, the material performance correction parameters and geometric compensation parameters are output. Example material performance correction parameters: The elastic modulus of concrete is corrected from 35.2 GPa to 38.5 GPa, the shrinkage rate is adjusted from 0.032% to 0.028%, and the coefficient of thermal expansion is adjusted from 10 × 10⁻⁶ to 10⁻⁶. -6 / ℃ corrected to 9.2×10 -6 / ℃; Exemplary geometric compensation parameters: X-axis offset compensation for precast floor slabs +3.3mm adjusted from +8.5mm to +5.2mm, elevation compensation +2.1mm adjusted from -6.2mm to -4.1mm, and torsion angle compensation 0.5° adjusted from 0.8° to 0.3°.

[0065] S3. Based on historical material performance correction parameters and geometric compensation parameters, a genetic algorithm is used to train the structural model and generate a digital twin benchmark model.

[0066] S3.1 Divide the historical material performance correction parameters and geometric compensation parameters into training set, validation set and test set.

[0067] It should be noted that the historical material performance correction parameters and geometric compensation parameters are sorted according to the construction time sequence of the engineering projects to ensure the temporal continuity of these parameters. A stratified random sampling method is used to divide the complete dataset into three independent subsets: a training set, a validation set, and a test set. The training set contains the earliest 60% of the collected data samples for the historical material performance correction parameters and geometric compensation parameters; the validation set contains the subsequent 20%; and the test set contains the most recent 20%. During the partitioning process, special attention is paid to maintaining the original distribution ratio of each type of parameter in each subset. To ensure fairness in training, Z-score standardization is performed on the training set, validation set, and test set respectively, transforming all parameter values ​​to a standard normal distribution with a mean of 0 and a standard deviation of 1, resulting in standardized training, validation, and test sets.

[0068] S3.2. Based on the training set, the structural model is optimized through multiple generations of selection, crossover and mutation training using a genetic algorithm. The validation set is used to monitor the model performance and prevent overfitting. The test set is used to evaluate the generalization ability of the final model and generate a digital twin benchmark model.

[0069] It should be noted that when using a genetic algorithm to optimize the structural model for multiple generations, an initial population containing multiple combinations of structural model parameters is randomly generated during the initialization phase. The fitness of each structural model parameter combination is evaluated on the training set, and the mean square error between the predicted values ​​of the structural model and the measured values ​​of historical material performance correction parameters and geometric compensation parameters is used as the fitness index. During the selection phase, a roulette wheel selection mechanism is used, probabilistically selecting superior structural model parameter combinations to enter the next generation based on their fitness values; parameter combinations with higher fitness have a greater selection probability.

[0070] In the crossover phase, a single-point crossover operation is performed on the selected structural model parameter combinations. Crossover points are randomly selected, and some parameter values ​​from the parent combinations are swapped to generate offspring parameter combinations with new characteristics. In the mutation phase, a subset of structural model parameter combinations undergoes uniform mutation with a preset probability. Specific parameter values ​​are randomly adjusted within the allowed range to increase population diversity. After each generation of training, the generalization performance of the new generation of structural model parameter combinations is evaluated on the validation set, and the optimal fitness value is recorded. When the optimal fitness value on the validation set does not show a significant improvement over multiple generations, the algorithm is considered converged, and the training process is terminated. Finally, the optimized structural model parameter combinations are comprehensively evaluated on the test set, and the parameter combination with the best overall performance is selected as the final parameter configuration for the digital twin baseline model, generating the digital twin baseline model.

[0071] S4. Output a multi-scale node performance evaluation report.

[0072] S4.1 Input the material property correction parameters and geometric compensation parameters into the digital twin reference model to perform overall structural analysis and obtain the stress state of the nodes.

[0073] It should be noted that after inputting the material property correction parameters and geometric compensation parameters into the digital twin baseline model, the digital twin baseline model automatically executes the structural analysis process. Based on the material property correction parameters, the thermal expansion coefficient of concrete under temperature effects is adjusted, the stress-strain relationship curve is corrected, and the elastic modulus value is determined. Simultaneously, based on the geometric compensation parameters, the actual installation position of precast components is corrected, including adjusting beam-column node coordinates, correcting component elevations, compensating for torsional angle deviations, and simulating joint construction errors. A refined hexahedral mesh with a range of 0.5 times the beam height is established in the core area of ​​the nodes, while a gradually changing mesh size is used in the transition region. An improved Newton-Raphson iterative algorithm is used for nonlinear solution, with displacement and force convergence tolerances set, and a maximum of 20 iterations. During the analysis, the state of concrete elements, the stress development of reinforcing steel elements, and the interface bond-slip behavior are tracked in real time, and material parameters are dynamically adjusted. After each load step is completed, key indicators such as the principal stress of concrete, the stress of steel reinforcement, and the amount of interface slip in the joint area are recorded. Stress cloud diagrams and displacement field distributions are plotted to obtain the stress state of the joint. For example, the principal stress of concrete is reduced from 3.5MPa to 2.8MPa, the stress of steel reinforcement is optimized from 285MPa to 245MPa, the amount of interface slip is reduced from 0.15mm to 0.08mm, and the displacement at the beam end is adjusted from 12.3mm to 9.7mm.

[0074] S4.2. Based on the stress location of the node, analyze the concrete damage and steel reinforcement stress, and generate a multi-scale node performance evaluation report.

[0075] It should be noted that, based on the analysis results of the nodal stress state, damage factors of concrete units and stress-strain data of steel reinforcement units in the core area of ​​the nodal are extracted. For concrete damage, the proportion of units with damage factors exceeding a set threshold is statistically analyzed, a damage area distribution map is plotted, and the maximum damage factor value and its location coordinates are recorded. For steel reinforcement stress, the number and distribution of yield steel reinforcement units are statistically analyzed, and the stress ratio is obtained by dividing the maximum Von Mises stress value of the steel reinforcement unit by the standard value of the steel reinforcement material, marking areas where stress exceeds the limit. Combining the concrete damage distribution and steel reinforcement stress state, nodal performance level regions are divided, and severely damaged, moderately damaged, and slightly damaged regions are marked. Integrating the concrete damage assessment results and steel reinforcement stress assessment results, a multi-scale nodal performance assessment report is generated, including a damage distribution map, stress cloud map, performance level classification map, and data statistics tables. The report covers the overall performance indicators of the node, the location of local weak points, and the quantitative analysis of key parameters. The final result is a multi-scale node performance evaluation report, including an exemplary concrete damage distribution map of the node core area, showing that the maximum damage factor of 0.42 is located at the beam-column junction; a reinforcement stress cloud map showing that the maximum stress ratio of 1.15 appears at the node stirrups; a performance grade zoning map showing that the severely damaged area accounts for 12% and the moderately damaged area accounts for 23%; and a key data statistics table showing that the concrete cracking area is 1.8 m² and the steel yield rate is 8.5%, forming a complete quantitative evaluation system for node performance.

[0076] S5. Use response surface methodology to identify multi-scale node performance influencing factors and sensitive areas, and output the sensitivity analysis results of design parameters.

[0077] S5.1 Based on the multi-scale node performance evaluation report, extract the comprehensive performance evaluation index of the nodes using principal component analysis.

[0078] S5.2. Using orthogonal experimental design, the performance response relationship between material property correction parameters, geometric compensation parameters and node comprehensive performance evaluation index is generated. Then, using variance decomposition method, the sensitivity index of material property correction parameters and geometric compensation parameters to node comprehensive performance evaluation index is calculated to obtain the design parameter sensitivity analysis results.

[0079] It should be noted that, based on the multi-scale node performance evaluation report, concrete damage factor distribution data, steel reinforcement stress ratio data, and node performance level regions were compiled to define input variables. Principal component analysis was then performed to standardize the input variables, calculate the covariance matrix, solve for eigenvalues ​​and eigenvectors, and select principal components with cumulative contribution rates exceeding a set threshold as the comprehensive node performance evaluation index.

[0080] Orthogonal experimental design was used to arrange different levels of material property correction parameters and geometric compensation parameters. For each parameter combination, the corresponding nodal comprehensive performance evaluation index value was obtained through structural analysis. Analysis of variance was used to calculate the influence of material property correction parameters and geometric compensation parameters on the nodal comprehensive performance evaluation index, establishing a quantitative relationship between parameters and performance indicators. Response surfaces of material property correction parameters, geometric compensation parameters, and nodal comprehensive performance evaluation index were plotted to determine the optimal parameter combination range and generate the performance response relationship between material property correction parameters, geometric compensation parameters, and nodal comprehensive performance evaluation index.

[0081] Based on the performance response relationship between material property correction parameters, geometric compensation parameters, and nodal comprehensive performance evaluation indicators, the sensitivity index is calculated using variance decomposition. In the preparation phase, material property correction parameters and geometric compensation parameters are used as input variables, and nodal comprehensive performance evaluation indicators are used as output variables. During calculation, all material property correction parameters and geometric compensation parameters except those currently being analyzed are fixed as baseline values. The values ​​of each material property correction parameter and geometric compensation parameter are adjusted one by one, and the range of change in the nodal comprehensive performance evaluation indicator is recorded. The variance components of the nodal comprehensive performance evaluation indicator caused by individual changes in each nodal performance parameter, as well as the variance components caused by the interaction between nodal performance parameters, are calculated.

[0082] The ratio of the variance component of each node's performance parameter to the total variance is used as a sensitivity index, ranging from 0 to 1. A larger value indicates a more significant impact of the parameter on the overall performance evaluation index of the node. Sensitivity indices are calculated separately for material property correction parameters and geometric compensation parameters. The calculation results are then compiled, sorted by sensitivity index, and used to generate the sensitivity analysis results for the design parameters.

[0083] S6. The multi-scale node performance influencing factors are converted into preliminary node design schemes through the design parameter conversion algorithm.

[0084] S6.1 Based on the results of the sensitivity analysis of design parameters, screen multi-scale node performance influencing factors, use the entropy weight method to quantify the design importance, and form multi-scale node performance influencing factors with priority.

[0085] It should be noted that, based on the sensitivity analysis results of the design parameters, material property correction parameters and geometric compensation parameters with sensitivity indices exceeding 0.1 were selected as the candidate set of multi-scale node performance influence factors. The entropy weight method was used to calculate the material property correction parameters and geometric compensation parameters: the node comprehensive performance evaluation index data corresponding to each parameter were standardized according to positive and negative indices respectively; the proportion distribution of the standardized data in each scheme was calculated; the information entropy value was calculated based on the proportion distribution to reflect the degree of data dispersion; the difference coefficient was calculated based on the information entropy value and normalized to obtain the objective weight of each parameter; the weights were weighted with the sensitivity index to calculate the comprehensive importance score; and the multi-scale node performance influence factors were formed by ranking the scores from highest to lowest.

[0086] S6.2 Based on the priority-based multi-scale node performance impact factor, a preliminary node design scheme is transformed through a conversion rule base.

[0087] It should be noted that, based on the priority-based multi-scale node performance influencing factors, by collecting and organizing the actual application data of material performance correction parameters and geometric compensation parameters from historical engineering cases, and combining them with specification requirements, the correspondence between various parameters and engineering measures is determined. For material performance parameters such as concrete temperature correction parameters and strain correction parameters, their allowable adjustment range and corresponding construction control methods are analyzed. For geometric compensation parameters such as axis offset compensation parameters and elevation deviation compensation parameters, treatment schemes for typical deviation cases are studied. The matching effect of parameters and measures is verified through field tests, ultimately forming a conversion rule library containing specific technical measures, applicable scope, and implementation conditions. Through this conversion rule library, concrete temperature correction parameters are mapped to temperature control measures, strain correction parameters to reinforcement adjustment measures, shrinkage rate correction parameters to curing scheme optimization, and elastic modulus correction parameters to material proportion adjustment. Axis offset compensation parameters are mapped to installation positioning measures, elevation deviation compensation parameters to support system optimization, torsion angle compensation parameters to connection node reinforcement, and joint misalignment compensation parameters to joint treatment schemes. A preliminary node design scheme is obtained.

[0088] S7. The initial node design scheme is optimized using a multi-objective particle swarm optimization algorithm to obtain the optimal node design scheme.

[0089] S7.1. Using a high-dimensional parameter mapping encoding method, the adjustable parameters in the initial node design scheme are encoded into particle position vectors, and the particle swarm is randomly initialized to calculate the multi-target fitness value of each particle.

[0090] It should be noted that when using the high-dimensional parameter mapping coding method to process the preliminary node design scheme, adjustable parameters such as concrete mix proportion parameters, reinforcement configuration parameters, node construction parameters, and construction control parameters are extracted as coding objects. Each adjustable parameter is divided into multiple discrete levels according to its allowable adjustment range, concrete strength grades are divided into multiple strength ranges, reinforcement diameters are divided into multiple specification ranges, and node connection methods are divided into multiple types. An independent coding dimension is assigned to each parameter, establishing a mapping relationship between parameter levels and numerical coordinates. An initial particle swarm is randomly generated, and the position vector of each particle consists of random coordinate values ​​of each parameter dimension. When calculating the particle fitness value, the position vector is decoded and restored to the specific node design scheme parameter combination. Through structural analysis calculation, three target performance indicators—node bearing capacity, stiffness, and ductility coefficient—are obtained as fitness values, resulting in a multi-objective fitness value.

[0091] S7.2. Using the particle swarm optimization algorithm, update the particle positions and velocities, select the Pareto optimal solution, and obtain the optimal node design scheme.

[0092] It should be noted that the particle swarm optimization algorithm is executed based on multi-objective fitness values ​​to update particle position and velocity vectors. The current particle fitness value is compared with the individual's historical best fitness value to update the individual's optimal position; the fitness values ​​of all particles in the swarm are compared with the global historical best fitness value to update the global optimal position. A non-dominated sorting method is used to handle the three objectives of bearing capacity, stiffness, and ductility coefficient, dividing the particle swarm into multiple Pareto front levels. Solutions in the first Pareto front level are selected as the candidate optimal solution set, and the crowding distance index of the candidate solutions is calculated. Elite particles are selected based on the Pareto front level and crowding distance, retaining solutions with good diversity. The process of position and velocity update, fitness calculation, Pareto sorting, and elite selection is repeated until the preset number of iterations is reached. From the final Pareto optimal solution set, the solution that satisfies the engineering constraints is selected, and decoded and restored into a specific combination of concrete mix proportion parameters, reinforcement configuration parameters, node construction parameters and construction control parameters to form the optimal node design scheme. For example, the concrete mix proportion parameters are optimized from the initial C30 to C35, the reinforcement diameter is adjusted from 6mm to 8mm, the node stirrup spacing is increased from 150mm to 100mm, and the curing time is extended from 7 days to 10 days. Finally, the bearing capacity is increased by 12%, the stiffness is increased by 8%, and the ductility coefficient reaches 5.2.

[0093] This embodiment also provides a computer device applicable to the design method of cross-beam nodes for cast-in-place prefabricated composite floor slabs based on digital twins, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the design method of cross-beam nodes for cast-in-place prefabricated composite floor slabs based on digital twins as proposed in the above embodiment.

[0094] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0095] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins, as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0096] In summary, this invention significantly improves the real-time accuracy of node design through least-squares dynamic parameter correction, enabling the digital twin model to dynamically track changes in material properties and component installation deviations during construction, overcoming the shortcomings of traditional static models that cannot reflect actual working conditions. Simultaneously, multi-objective particle swarm optimization greatly enhances the overall economic efficiency of node solutions, intelligently balancing multiple objectives such as seismic performance, material cost, and construction convenience while ensuring structural safety, overcoming the limitations of quantifying optimization indicators in manual experience-based design. The synergistic application of these two technologies improves the controllability of the construction process and enhances the adaptability of the design scheme, establishing an intelligent technical system for prefabricated building node design that combines precise control capabilities with multi-objective optimization capabilities.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. The embodiments of the present invention may be deleted or otherwise adjusted without affecting the implementation of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A design method for beam-span joints in cast-in-place prefabricated composite floor slabs based on digital twins, characterized in that: include, Real-time data collection of material status of cast-in-place sections and installation deviation data of precast floor slabs during the construction of cross-beam joints in cast-in-place sections is performed and preprocessed to obtain joint construction data. The least squares optimization algorithm was used to obtain the material property correction parameters and geometric compensation parameters; Based on historical material performance correction parameters and geometric compensation parameters, a genetic algorithm is used to train the structural model and generate a digital twin benchmark model. Input the material property correction parameters and geometric compensation parameters into the digital twin benchmark model, perform multi-scale finite element analysis, and output a multi-scale node performance evaluation report. The response surface methodology is used to identify the performance influencing factors and sensitive areas of multi-scale nodes, output the sensitivity analysis results of design parameters, and convert the performance influencing factors of multi-scale nodes into preliminary node design schemes through a design parameter transformation algorithm. The initial node design scheme is optimized using a multi-objective particle swarm optimization algorithm to obtain the optimal node design scheme.

2. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 1, characterized in that: The material condition data of the cast-in-place section includes the temperature data, strain data, shrinkage rate data, and elastic modulus data of the concrete; The installation deviation data of the precast floor slab includes the axial offset, elevation deviation, torsion angle and joint misalignment data of the precast floor slab; The preprocessing includes, Statistical thresholding removes outliers, interpolation algorithms fill in missing values, ICP algorithms unify coordinates of multi-source data, principal component analysis extracts features, and Z-score standardization unifies formats.

3. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 2, characterized in that: The least squares optimization algorithm is used to obtain the material property correction parameters and geometric compensation parameters. The specific steps are as follows. Based on the node construction data, the least squares optimization algorithm is used for iterative calculation. The node construction data and the calculation results are compared and analyzed to obtain the difference. Based on the difference, iterative correction is performed using material parameterization calculation functions and geometric coordinate transformation functions to output material property correction parameters and geometric compensation parameters.

4. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 3, characterized in that: The process involves training the structural model using a genetic algorithm based on historical material property correction parameters and geometric compensation parameters to generate a digital twin baseline model. The specific steps are as follows: The historical material performance correction parameters and geometric compensation parameters are divided into training set, validation set and test set; Based on the training set, the structural model is optimized through multiple generations of selection, crossover, and mutation using a genetic algorithm. The validation set is used to monitor model performance and prevent overfitting. The test set is used to evaluate the generalization ability of the final model, generating a digital twin benchmark model.

5. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 4, characterized in that: The specific steps for outputting the multi-scale node performance evaluation report are as follows. Input the material property correction parameters and geometric compensation parameters into the digital twin benchmark model to perform overall structural analysis and obtain the stress state of the nodes. Based on the stress location of the node, analyze the concrete damage and steel reinforcement stress, and generate a multi-scale node performance evaluation report.

6. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 5, characterized in that: The method employs response surface methodology to identify multi-scale node performance influencing factors and sensitive regions, and outputs sensitivity analysis results for design parameters. The specific steps are as follows: Based on the multi-scale node performance evaluation report, principal component analysis was used to extract the comprehensive performance evaluation index of the nodes. Using orthogonal experimental design, the performance response relationship between material property correction parameters, geometric compensation parameters and node comprehensive performance evaluation index is generated. Then, using variance decomposition method, the sensitivity index of material property correction parameters and geometric compensation parameters to node comprehensive performance evaluation index is calculated, and the sensitivity analysis results of design parameters are obtained.

7. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 6, characterized in that: The process of converting multi-scale node performance influencing factors into preliminary node design schemes using a design parameter transformation algorithm is as follows: Based on the results of the sensitivity analysis of design parameters, multi-scale node performance influencing factors are screened, and the entropy weight method is used to quantify the design importance, forming multi-scale node performance influencing factors with priority. Based on the priority-based multi-scale node performance impact factors, a preliminary node design scheme is transformed through a conversion rule base.

8. The design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in claim 7, characterized in that: The preliminary node design scheme is optimized using a multi-objective particle swarm optimization algorithm to obtain the optimal node design scheme. The specific steps are as follows. A high-dimensional parameter mapping encoding method is adopted to encode the adjustable parameters in the initial node design scheme into particle position vectors, and the particle swarm is randomly initialized to calculate the multi-target fitness value of each particle. By using the particle swarm optimization algorithm, the particle positions and velocities are updated, the Pareto optimal solution is selected, and the optimal node design scheme is obtained.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the design method for cross-beam nodes of prefabricated composite floor slabs based on digital twins as described in any one of claims 1 to 8.

Images