Method for compensating influence of net rack deformation on inner molded surface precision

Abstract

The invention provides a method for compensating influence of net rack deformation on inner profile precision, and belongs to the technical field of wind tunnel body inner profile construction. A minimum dominating set algorithm is adopted to optimize sensor layout to construct a three-dimensional deformation monitoring network, and a pre-arching compensation optimization function is used to calculate a preset elevation compensation value; in the net rack splicing process, splicing errors are corrected in real time through a structural deformation self-adaptive compensation model, displacement synchronization in the lifting process is ensured through a synchronous hydraulic lifting system and a lifting deformation control algorithm, and after a net rack is in place, the actual geometrical shape of an inner molded surface is obtained through a three-dimensional laser scanning technology. According to the deviation value, comprehensive compensation measures of supporting jig frame elevation adjustment, temporary supporting structure arrangement, follow-up construction load distribution control and net rack node position fine adjustment are adopted for deformation compensation, and the technical problem that due to net rack deformation, the geometric accuracy of the inner profile of the lower hanging cast-in-place slab is difficult to control is solved.

Description

Technical Field

[0001] This invention belongs to the field of wind tunnel internal surface construction technology, specifically, it relates to a method for compensating for the impact of space frame deformation on the accuracy of the internal surface. Background Technology

[0002] In wind tunnel construction projects, traditional space frame deformation control mainly relies on theoretical calculations and empirical adjustments. During space frame assembly, pre-cambering is used to compensate for deflection caused by subsequent loads, and conventional measurement methods are employed for construction monitoring. However, traditional deformation control methods have significant limitations. Due to the complex stress on space frame structures, under the combined effects of self-weight, temperature changes, and wind loads, space frame deformation exhibits highly nonlinear characteristics, making it difficult for traditional linear prediction models to accurately predict actual deformation. Furthermore, existing monitoring methods are mostly discrete-point measurements, failing to achieve comprehensive monitoring of the entire space frame structure's deformation state, resulting in a lack of real-time and accurate deformation compensation. In existing technologies, the lack of effective multi-condition deformation prediction mechanisms and real-time adaptive compensation adjustment capabilities often leads to deviations in the geometric accuracy of the underlying cast-in-place slab exceeding design requirements due to space frame deformation, affecting the functionality and aesthetics of the building's interior space. In other words, existing technologies suffer from the technical problem of difficulty in controlling the geometric accuracy of the underlying cast-in-place slab due to space frame deformation. Summary of the Invention

[0003] In view of this, the present invention provides a method for compensating for the impact of space frame deformation on the accuracy of the inner surface, which can solve the technical problem in the prior art that the geometric accuracy of the inner surface of the lower cast-in-place slab is difficult to control due to space frame deformation.

[0004] This invention is implemented as follows: It provides a method for compensating for the impact of space frame deformation on the accuracy of the internal surface. A precise mathematical model of the space frame structure is established using finite element analysis software. The model is then corrected by combining measured data of deflection under the space frame's self-weight, thermal expansion and contraction caused by temperature changes, and deformation under wind load. A multi-condition deformation prediction database is established, including calculation results of the stress and deformation of the interlocking unloading structure, the roof concrete structure, the cast-in-place concrete structure under the space frame, and the deformation of the structure after formwork removal. This is achieved by modeling the space frame structure as an undirected graph and using minimum support... The optimal sensor placement algorithm is used to establish a three-dimensional deformation monitoring network and collect three-dimensional displacement data of key control points of the space frame structure in real time. Based on the multi-condition deformation prediction database and three-dimensional displacement data, the pre-arching compensation parameters for the space frame assembly stage are calculated. During the space frame assembly process, an adaptive structural deformation compensation model is used to correct assembly errors in real time. During the space frame lifting process, a lifting deformation control algorithm is used to dynamically adjust the lifting speed and displacement of each hydraulic lifter. The key control points of the inner surface of the cast-in-place slab under the space frame are precisely measured, and comprehensive compensation measures are used to compensate for deformation based on the deviation value of the inner surface of the cast-in-place slab under the space frame.

[0005] The multi-condition deformation prediction database is a data set containing the deformation laws of space frame under different load combinations, established through a large amount of finite element calculations and statistical analysis of measured data. It is used to predict the deformation response of space frame structures under various load combinations.

[0006] Specifically, the steps for establishing the three-dimensional deformation monitoring network involve deploying high-precision displacement sensors and laser rangefinders at the cast-in-place slab nodes, upper chord nodes, lower chord nodes, middle chord nodes, mid-span positions, and support points of the space frame to establish a three-dimensional deformation monitoring network covering the entire space frame structure.

[0007] The aforementioned space frame undercast slab node is a structural connection point where the welded ball at the bottom of the space frame structure is located at the original design location of the undercast concrete node, and the distance from the surface of the undercast slab template is designed to be 5cm.

[0008] The upper chord node of the space frame is a structural connection point where steel lattice columns are used as the assembly frame in the upper part of the space frame structure.

[0009] Specifically, the calculation steps for the pre-arching compensation parameters involve calculating the pre-arching compensation parameters during the space frame assembly stage using a pre-arching compensation optimization function, and determining the preset elevation compensation values ​​for the space frame lower slab nodes, upper chord nodes, lower chord nodes, and middle chord nodes.

[0010] The pre-camber compensation optimization function is used to calculate the optimal pre-camber compensation parameters for each node during the assembly stage of the space frame. The inputs include the self-weight distribution load of the space frame from the structural design, the design temperature variation range from the design specifications, the expected wind load value from the design specifications, the material elastic modulus from the material properties, and the maximum span of the space frame from the structural design. The output is the pre-camber compensation height value of each node used to guide the assembly construction.

[0011] Specifically, the step of real-time correction of assembly errors during the assembly of the space frame involves adjusting the elevations of the lower slab nodes, upper chord nodes, lower chord nodes, and middle chord nodes of the space frame according to the pre-arching compensation parameters during the assembly process. A laser total station is used for three-dimensional coordinate positioning and elevation control of the nodes. Staged detection is adopted, and the assembly error is corrected in real time through a structural deformation adaptive compensation model.

[0012] The adaptive compensation model for structural deformation adopts a variational autoencoder model with a multi-layer encoder-decoder architecture. The encoder part contains three convolutional layers and two fully connected layers to extract the deformation features of the grid structure. The decoder part adopts a symmetrical structure to output compensation parameters. The intermediate latent space uses a 16-dimensional vector to represent the deformation state of the grid structure.

[0013] The comprehensive compensation measures mentioned above are a systematic deformation compensation method that combines four compensation methods: adjustment of the support frame elevation, setting of temporary support structures, control of subsequent construction load distribution, and fine-tuning of the space frame node positions.

[0014] The phased testing refers to multi-level measurement and verification conducted during the assembly of the space frame according to the construction progress of node assembly, frame assembly, and segment installation, in order to promptly detect and correct assembly deviations.

[0015] Specifically, the step of dynamically adjusting the lifting speed and displacement of each hydraulic lifter during the lifting process of the space frame involves using a synchronous hydraulic lifting system and dynamically adjusting the lifting speed and displacement of each hydraulic lifter through a lifting deformation control algorithm to ensure that the displacement synchronization error of each lifting point is controlled within ±20mm.

[0016] The lifting deformation control algorithm is a control algorithm based on multi-objective optimization theory. It determines the optimal motion trajectory of each hydraulic lifter by solving the optimization problem under multiple constraints.

[0017] The step of precisely measuring the key control points of the inner surface of the cast-in-place slab under the space frame specifically involves precisely measuring the key control points of the inner surface of the cast-in-place slab under the space frame after the space frame is in place, using three-dimensional laser scanning technology to obtain the actual geometric shape of the inner surface of the cast-in-place slab under the space frame, and calculating the deviation value from the design reference surface.

[0018] Furthermore, following comprehensive compensation measures, the pre-arching compensation parameters for subsequent construction stages are adjusted through a deformation compensation feedback mechanism to form a closed-loop control system.

[0019] Furthermore, the structural deformation adaptive compensation model achieves a smooth transition between different deformation states through a latent spatial interpolation mechanism. The interpolation weight parameters are dynamically adjusted based on three key parameters: the complexity index of the space frame structure from the structural geometry parameters, the load variation coefficient from the load statistics, and the temperature sensitivity coefficient from the material's thermal expansion and contraction coefficient.

[0020] This invention solves the technical problem of difficulty in controlling the geometric accuracy of the inner surface of the cast-in-place slab caused by space frame deformation by establishing a multi-condition deformation prediction database and a three-dimensional deformation monitoring network, combined with the real-time adjustment mechanism of the structural deformation adaptive compensation model. This invention uses finite element analysis to establish an accurate mathematical model of the space frame structure, corrects model parameters through a large amount of measured data, and constructs a deformation prediction database containing multiple load conditions, providing a scientific basis for calculating pre-camber compensation parameters. Simultaneously, by optimizing sensor deployment through the minimum dominance set algorithm, a three-dimensional deformation monitoring network covering the entire space frame structure is established, achieving real-time and comprehensive monitoring of the space frame deformation state. This invention utilizes a variational autoencoder model to construct a structural deformation adaptive compensation model, which can dynamically adjust compensation parameters based on real-time monitoring data. Through the synergistic effect of staged detection and comprehensive compensation measures, precise control of space frame deformation is achieved. In summary, this invention solves the technical problem mentioned in the background art of difficulty in controlling the geometric accuracy of the inner surface of the cast-in-place slab caused by space frame deformation. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention.

[0022] Figure 2 This is a diagram of the adaptive compensation model architecture for structural deformation involved in the present invention.

[0023] Figure 3 This is a calculation diagram of the structural deformation during template removal in Example 2.

[0024] Figure 4 This is a cross-sectional view of the assembly site structure in Example 2.

[0025] Figure 5 This is a diagram showing the positioning of the tire frame in Example 2.

[0026] Figure 6 This is an isometric drawing of the tire frame elevation measurement in Example 2.

[0027] Figure 7 This is a diagram showing the setup of the space frame lifting points in Example 2.

[0028] Figure 8 This is a diagram showing the overall lifting deformation analysis of the space frame in Example 2.

[0029] Figure 9 This is a deformation analysis diagram of the segmented hoisting of the space frame in Example 2.

[0030] Figure 10 This is a distribution diagram of deformation monitoring data for the space frame nodes in Example 2.

[0031] Figure 11 This is a diagram illustrating the convergence process of optimizing the pre-arching compensation parameters in Example 2.

[0032] Figure 12 This is a diagram showing the effect of synchronization error control in the lifting process in Example 2. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0034] like Figure 1 The diagram shown is a flowchart of a method for compensating for the impact of space frame deformation on the accuracy of the inner surface, provided by the present invention. This method includes the following steps:

[0035] S01. A precise mathematical model of the space frame structure is established using finite element analysis software. The model is then corrected by combining measured data of deflection under the self-weight of the space frame, measured data of thermal expansion and contraction caused by temperature changes, and measured data of deformation under wind load. A multi-condition deformation prediction database is established, which includes the calculation results of stress and deformation of the interlocking unloading structure, the calculation results of stress and deformation of the roof concrete structure, the calculation results of stress and deformation of the cast-in-place concrete structure under the space frame, and the calculation results of deformation of the formwork removal structure.

[0036] S02. By modeling the space frame structure as an undirected graph and using the minimum dominating set algorithm to determine the optimal placement of sensors, high-precision displacement sensors and laser rangefinders are deployed at the cast-in-place slab nodes, upper chord nodes, lower chord nodes, middle chord nodes, mid-span positions, and support points of the space frame to establish a three-dimensional deformation monitoring network covering the entire space frame structure and collect three-dimensional displacement data of each key control point of the space frame structure in real time.

[0037] S03. Based on the multi-condition deformation prediction database and three-dimensional displacement data, the pre-arching compensation parameters of the space frame assembly stage are calculated by the pre-arching compensation optimization function, and the preset elevation compensation values ​​of the space frame undercast slab nodes, space frame upper chord nodes, space frame lower chord nodes, and space frame middle chord nodes are determined.

[0038] S04. During the assembly of the space frame, the elevations of the lower slab nodes, upper chord nodes, lower chord nodes, and middle chord nodes of the space frame are adjusted according to the pre-arching compensation parameters. A laser total station is used for three-dimensional coordinate positioning and elevation control of the nodes. The assembly is carried out in stages, with measurements taken after each node is assembled, after each frame is assembled, and after each segment is installed. The assembly error is corrected in real time through the structural deformation adaptive compensation model.

[0039] S05. During the lifting process of the space frame, a synchronous hydraulic lifting system is adopted. The lifting speed and displacement of each hydraulic lifting device are dynamically adjusted through the lifting deformation control algorithm to ensure that the displacement synchronization error of each lifting point is controlled within ±20mm.

[0040] S06. After the space frame is in place, the key control points of the inner surface of the cast-in-place slab under the space frame are precisely measured. The actual geometric shape of the inner surface of the cast-in-place slab under the space frame is obtained by using three-dimensional laser scanning technology, and the deviation value from the design reference surface is calculated.

[0041] S07. Establish a deformation compensation effect evaluation system based on the deviation value of the inner surface of the cast-in-place slab under the space frame. Deformation compensation is carried out through comprehensive compensation measures such as adjusting the elevation of the support frame, setting up temporary support structures, controlling the distribution of subsequent construction loads, and fine-tuning the position of the space frame nodes in real time. The pre-arching compensation parameters in the subsequent construction stage are adjusted through the deformation compensation feedback mechanism to form a closed-loop control system.

[0042] Among them, the multi-condition deformation prediction database is a data set containing the deformation law of the space frame under different load combinations, which is established by statistical analysis of a large number of finite element calculations and measured data. It is used to predict the deformation response of the space frame structure under various load combinations.

[0043] Among them, the three-dimensional deformation monitoring network is a spatial measurement system composed of multiple high-precision sensors, which is used to monitor the three-dimensional displacement changes of each key control point of the grid structure in real time.

[0044] Among them, the space frame under-cast slab node is a structural connection point where the welded ball at the bottom of the space frame structure adopts the original design of the under-cast concrete node position, and the distance from the surface of the space frame under-cast slab template is designed to be 5cm.

[0045] Among them, the upper chord node of the space frame is a structural connection point in the upper part of the space frame structure that uses adjustable steel lattice columns as assembly frames.

[0046] Among them, the lower chord node of the space frame is the structural connection point of the lower part of the space frame structure using the original design nodes and round tubes as the assembly frame.

[0047] Among them, the chord node in the space frame is the structural connection point located in the middle of the space frame structure that bears the main internal force transmission.

[0048] The mid-span position of the space frame is the center position of the space frame structure in the span direction, and it is the location where the maximum bending moment and maximum deflection occur.

[0049] Among them, the location of the support point of the space frame is the load-bearing transfer location where the space frame structure is connected to the lower support structure.

[0050] Among them, the key control points are important measurement locations that affect the geometric accuracy of the inner surface of the cast-in-place slab under the space frame. They are usually selected at the nodes of the cast-in-place slab under the space frame and at the mid-span of the space frame.

[0051] Among them, the pre-camber compensation parameter is a reverse camber value set in advance during the assembly stage of the space frame to offset the downward deflection caused by subsequent loads.

[0052] The pre-camber compensation optimization function is used to calculate the optimal pre-camber compensation parameters for each node during the assembly stage of the space frame. The inputs include the self-weight distribution load of the space frame from the structural design, the design temperature variation range from the design specifications, the expected wind load value from the design specifications, the material elastic modulus from the material properties, and the maximum span of the space frame from the structural design. The output is the pre-camber compensation height value of each node used to guide the assembly construction.

[0053] Among them, the synchronous hydraulic lifting system is a computer-controlled multi-point hydraulic lifting device that ensures the synchronous movement of each hydraulic lifter through stroke sensor feedback and control algorithms.

[0054] Among them, the lifting deformation control algorithm is a control algorithm based on multi-objective optimization theory. It determines the optimal motion trajectory of each hydraulic lifter by solving the optimization problem under multiple constraints.

[0055] Among them, the inner surface of the cast-in-place slab under the space frame is the lower surface of the cast-in-place slab concrete structure under the space frame, which serves as the top surface of the building's interior space.

[0056] Among them, phased testing is a multi-level measurement and verification carried out according to the construction progress of node assembly, frame assembly and segment installation during the assembly of the space frame, which is used to detect and correct assembly deviations in a timely manner.

[0057] Among them, the elevation adjustment of the support frame is to recalculate the installation height of the support frame based on the deviation value of the inner surface of the cast-in-place slab under the space frame, and to compensate for the deviation of the inner surface caused by the deformation of the space frame by adjusting the elevation of the adjustable steel lattice column.

[0058] The temporary support structure is set up by adding temporary support members between the cast-in-place slab formwork hanging under the space frame and the main steel beam hanging below, so as to reduce the impact of space frame deformation on the accuracy of the internal surface by dispersing the load transfer path.

[0059] Among them, the subsequent construction load distribution control is to evenly distribute the construction load during the construction of the upper chord concrete of the space frame and the construction of the roof project, so as to avoid local deformation caused by local concentrated load.

[0060] Among them, the fine-tuning of the space frame node position involves using equipment such as hydraulic jacks to make small displacement adjustments to the key control points of the space frame, directly correcting the geometric deviation of the inner surface of the cast-in-place slab hanging under the space frame.

[0061] Among them, the comprehensive compensation measures are a systematic deformation compensation method that combines four compensation methods: adjustment of the support frame elevation, setting of temporary support structures, control of subsequent construction load distribution, and fine-tuning of the space frame node position.

[0062] Among them, the deformation compensation feedback mechanism is an adaptive control mechanism that adjusts the theoretical calculation parameters according to the actual construction effect, and is used to continuously optimize the compensation accuracy.

[0063] The adaptive compensation model for structural deformation employs a variational autoencoder model with a multi-layer encoder-decoder architecture. The encoder consists of three convolutional layers and two fully connected layers for extracting deformation features of the space frame. The decoder uses a symmetrical structure to output compensation parameters. The intermediate latent space uses a 16-dimensional vector to represent the deformation state of the space frame. The model achieves smooth transitions between different deformation states through a latent space interpolation mechanism. The interpolation weight parameters are dynamically adjusted based on three key parameters: the complexity index of the space frame structure from structural geometric parameters, the load variation coefficient from load statistics, and the temperature sensitivity coefficient from the material's thermal expansion and contraction coefficient. The training dataset establishment steps for the adaptive compensation model specifically include collecting design drawings and construction records of different types of space frame structures, extracting space frame geometric parameters, material properties, load conditions, and actual deformation data, establishing a deformation data sample library containing 10,000 space frame engineering cases, labeling each sample with ideal compensation parameters and actual construction effects, expanding the sample set to 50,000 training samples through data augmentation techniques, and establishing a standardized input and output data format. The training steps of the adaptive compensation model for structural deformation specifically include optimizing model parameters using the Adam optimizer and learning rate decay strategy, setting the batch size to 32, the number of training rounds to 200, using the mean squared error loss function to evaluate prediction accuracy, employing an early stopping mechanism to prevent overfitting during training, evaluating the model's generalization ability through cross-validation, and finally selecting the model parameters with the smallest validation set error as the optimal model.

[0064] The adaptive adjustment function for compensation parameters is used to adjust the interpolation weight parameters of the adaptive compensation model for structural deformation. This function is calculated based on four data points: the complexity index of the space frame structure from structural geometric parameters, the load variation coefficient from load statistics, the temperature sensitivity coefficient from the material's thermal expansion and contraction coefficient, and the construction accuracy requirement level from design requirements. This results in a comprehensive adjustment factor value for adjusting the interpolation weight parameters. When the comprehensive adjustment factor value is between 0 and 0.3, a conservative interpolation strategy is used to adjust the interpolation weight parameters of the adaptive compensation model for structural deformation. When the comprehensive adjustment factor value is between 0.3 and 0.7, a standard interpolation strategy is used to adjust the interpolation weight parameters of the adaptive compensation model for structural deformation. When the comprehensive adjustment factor value is between 0.7 and 1.0, an aggressive interpolation strategy is used to adjust the interpolation weight parameters of the adaptive compensation model for structural deformation.

[0065] The specific implementation methods of the above steps are described in detail below.

[0066] The specific implementation of step S01 involves first establishing a three-dimensional geometric model of the space frame structure. A complete geometric model containing all members, nodes, and connectors is constructed using computer-aided design software. Then, the geometric model is imported into finite element analysis software for mesh generation. Tetrahedral elements are used to discretize the members, with the element size controlled within one-tenth of the member length to ensure calculation accuracy. A material property database is established, inputting physical parameters such as the elastic modulus, Poisson's ratio, and coefficient of linear expansion of the steel. The elastic modulus of the steel is taken as 206000 MPa, the Poisson's ratio as 0.3, and the coefficient of linear expansion as 1.2 × 10⁻⁶. -5 / ℃. Define boundary conditions and constraints, setting the space frame support points as fixed or hinged constraints, determining the constraint type based on actual support conditions. Establish load case combinations, including the space frame's self-weight distributed load under dead load conditions, construction and service loads under live load conditions, thermal stress caused by temperature changes under temperature conditions, and wind pressure distributed loads under wind load conditions. Solve the structural response under each load case using nonlinear finite element analysis, employing the Newton-Raphson iterative method to solve the nonlinear equations, with the iterative convergence accuracy controlled within 0.001. Compare and analyze the finite element calculation results with measured data, using the least squares method to fit correction coefficients, adjusting material parameters and boundary conditions to control the deviation between the calculation results and measured data within 5%. Establish a multi-load case deformation prediction database, storing structural response data such as nodal displacements, member internal forces, and strain distributions under different load combinations, forming a basic dataset for subsequent predictive analysis.

[0067] The specific implementation of step S02 involves abstracting the grid structure as an undirected graph model, where grid nodes correspond to vertices and grid members correspond to edges. An adjacency matrix represents the topological connections of the grid. Based on the minimum dominating set theory in graph theory, the minimum number of sensor placement points needed to cover the entire grid structure is searched. A greedy algorithm is used to solve the minimum dominating set problem, adding the vertex that covers the most uncovered nodes to the dominating set each time, until all nodes are covered. High-precision displacement sensors are installed at the determined sensor placement locations, with the required sensor accuracy reaching [specific accuracy level].

[0068] ±0.1mm accuracy is achieved using resistance strain gauge or fiber optic grating sensors to ensure measurement stability under temperature variations and vibration. Laser rangefinders are deployed at key control points, with a measurement accuracy of ±0.05mm. Phase-based laser ranging is used, calculating distance changes by measuring the laser's round-trip time. A data acquisition system is established, employing multi-channel data acquisition units to synchronously acquire signals from each sensor at a frequency of 1Hz to capture the slow deformation process of the structure. A wireless transmission network is established to transmit the acquired data to the central processing system in real time, using LoRa wireless communication technology with a transmission distance of up to 5km, ensuring reliable communication on the construction site. A three-dimensional deformation monitoring network is established, unifying the measurement data from each sensor into the same coordinate system through a spatial coordinate transformation algorithm. Least squares method is used for coordinate system fitting, with a fitting accuracy controlled within ±0.02mm.

[0069] The specific implementation of step S03 is to establish a pre-arching compensation optimization model based on a multi-condition deformation prediction database. This model uses a multi-objective optimization algorithm to solve for the optimal pre-arching parameters. Input parameters include the distributed load of the space frame's self-weight, the design temperature range of -20℃ to 60℃, the expected wind load value determined based on the wind pressure height variation coefficient of the building's location, the material elastic modulus of 206000MPa, and the maximum span of the space frame, among other structural design parameters. A genetic algorithm is used to solve the multi-objective optimization problem, with a population size of 100 individuals, a crossover probability of 0.8, a mutation probability of 0.02, and 200 iterations. The optimization objective function includes minimizing the maximum deflection of the space frame, minimizing the internal surface deviation, and minimizing the pre-arching height, among other objectives, and is solved using the Pareto optimal solution set method. The multi-objective optimization problem is transformed into a single-objective optimization problem using a weighted coefficient method. The weighted coefficients are determined based on the project's importance and construction difficulty; the weighted coefficient for internal surface accuracy is set to 0.6, and the weighted coefficient for construction convenience is set to 0.4. Calculate the pre-camber compensation parameters for each node, including the preset elevation compensation values ​​for nodes with cast-in-place slabs hanging below the space frame, the upper chord nodes, the lower chord nodes, and the middle chord nodes. The compensation values ​​typically range from 5mm to 50mm. Establish a pre-camber compensation parameter database to store the optimal compensation parameters under different working conditions, forming a parameter lookup table to guide actual construction.

[0070] The specific implementation of step S04 involves formulating a space frame assembly construction plan based on the pre-arching compensation parameters, using a laser total station for high-precision measurement and positioning, with a measurement accuracy requirement of ±1mm, and the measurement range covering the entire space frame assembly area. A phased inspection method is adopted during the space frame assembly process. First, immediately after assembling a node, three-dimensional coordinate measurements are performed to check whether the node position meets the design requirements, with positional deviation controlled within ±2mm. After assembling a space frame, an overall geometric dimension check is performed, measuring the span, height, inclination, and other geometric parameters of the frame, with geometric deviations controlled within the design allowable range. After installing a segment, a comprehensive measurement verification is performed, including checking the three-dimensional coordinates of all nodes within the segment, member lengths, connection angles, and other parameters. A structural deformation adaptive compensation model is used to correct assembly errors in real time. This model, based on a deep learning algorithm, can predict the structural deformation trend based on real-time measurement data and automatically adjust subsequent assembly parameters. When an assembly error is detected to exceed the allowable range, an error correction program is immediately initiated, eliminating the error by adjusting node positions, changing member lengths, and recalibrating connection angles. Establish assembly quality control files to record assembly parameters, measurement data, error correction records, and other information for each node, forming a complete construction quality traceability system.

[0071] The specific implementation of step S05 involves a computer-controlled synchronous hydraulic lifting system. This system consists of multiple hydraulic lifters, stroke sensors, pressure sensors, and a central control system. The hydraulic lifters use double-acting hydraulic cylinders with a lifting capacity of 500kN to 2000kN, a stroke range of 0 to 2000mm, and an adjustable lifting speed range of 0.1mm / min to 10mm / min. High-precision stroke sensors are installed on each hydraulic lifter, with a measurement accuracy of ±0.1mm, to monitor the displacement changes of the lifter in real time. A lifting deformation control algorithm dynamically adjusts the operating parameters of each hydraulic lifter. This algorithm is based on multi-objective optimization theory, aiming to minimize the displacement deviation at each lifting point, minimize structural stress concentration, and maximize lifting efficiency. Through a real-time feedback control system, the control cycle of each hydraulic lifter is set to 100ms, adjusting the flow rate and pressure of the hydraulic cylinders based on the displacement data fed back by the stroke sensors. When a displacement deviation at a certain lifting point is detected to exceed ±10mm, the lifting speed of that lifter is automatically adjusted, either increasing the speed of the lagging lifter or decreasing the speed of the leading lifter, ensuring that the synchronization error of the displacement at each lifting point is controlled within ±20mm. Establish a lifting process monitoring system to display parameters such as displacement, speed, and pressure at each lifting point in real time. When an abnormal situation occurs, promptly sound an alarm and automatically stop the lifting operation.

[0072] The specific implementation of step S06 involves using 3D laser scanning technology to comprehensively measure the inner surface of the cast-in-place slab under the space frame after the space frame is in place and stabilized. The laser scanner uses a phase-based ranging principle, with a measurement accuracy of ±0.5mm and a scanning speed of 1 million points / second, covering the entire inner surface area. Before laser scanning, control points need to be set around the inner surface, and the 3D coordinates of the control points are measured using a total station to establish a unified coordinate system. During laser scanning, a multi-station scanning method is used to ensure that every part of the inner surface is scanned, and the point cloud density is controlled to be more than 1000 points per square meter. The point cloud data obtained from the scan is preprocessed using point cloud data processing software, including noise filtering, data stitching, and coordinate transformation. The actual geometric shape of the inner surface is fitted using the least squares method to establish a 3D digital model of the inner surface. The actual measured inner surface model is compared and analyzed with the design reference surface model, and the positional deviation of each key control point is calculated. The deviation values ​​include displacement components in the X, Y, and Z directions. Establish an internal surface accuracy assessment report, statistically analyze the distribution of deviation values, and calculate statistical parameters such as average deviation, maximum deviation, and standard deviation to provide data support for subsequent compensation measures.

[0073] The specific implementation of step S07 involves establishing a deformation compensation effect evaluation system based on the internal surface deviation value. This system is based on fuzzy comprehensive evaluation theory and establishes a comprehensive evaluation model that includes multiple evaluation indicators such as geometric accuracy, construction difficulty, and economic cost. The weight coefficients of each evaluation indicator are determined using the analytic hierarchy process (AHP): geometric accuracy has a weight coefficient of 0.5, construction difficulty has a weight coefficient of 0.3, and economic cost has a weight coefficient of 0.2. Deformation compensation is carried out using comprehensive compensation measures. First, the elevation of the support frame is adjusted. The installation height of each node of the support frame is recalculated based on the internal surface deviation value, with an adjustment range typically within ±30mm. Temporary support structures are set up, with temporary support members added between the cast-in-place slab formwork hanging under the space frame and the lower main steel beam. The support members use adjustable-length steel pipes, with an adjustment range of ±50mm. The distribution of subsequent construction loads is controlled. During the concrete construction of the upper chord of the space frame, a segmented pouring method is adopted, with each pouring area controlled within 200 square meters to avoid structural deformation caused by excessive local loads. Real-time fine-tuning of the space frame node positions is performed using hydraulic jacks to make minute displacement adjustments to key control points, with an adjustment accuracy of ±0.5mm and an adjustment range of ±5mm. A deformation compensation feedback mechanism is established, adjusting theoretical calculation parameters based on actual compensation effects, and continuously optimizing the prediction accuracy of the compensation model through machine learning algorithms. When the compensation effect does not meet the requirements, internal surface measurements and deviation analysis are repeated, compensation parameters are adjusted, and the compensation process is repeated until the internal surface accuracy meets the design requirements.

[0074] like Figure 2As shown, the adaptive compensation model for structural deformation employs a variational autoencoder model with a multi-layer encoder-decoder architecture. This model can learn the complex nonlinear mapping relationship of the space frame structure deformation and generate corresponding compensation parameters. The encoder part contains three convolutional layers: the first convolutional layer uses 32 3×3 convolutional kernels with the ReLU activation function; the second convolutional layer uses 64 3×3 convolutional kernels; and the third convolutional layer uses 128 3×3 convolutional kernels. Each convolutional layer is followed by a 2×2 max pooling layer for dimensionality reduction. The encoder also contains two fully connected layers: the first fully connected layer contains 256 neurons, and the second fully connected layer contains 128 neurons, used to further extract high-dimensional features of the space frame deformation. The decoder part adopts a symmetrical structure to the encoder, containing two fully connected layers and three deconvolutional layers, ultimately outputting a compensation parameter vector. The intermediate latent space uses a 16-dimensional vector to represent the space frame deformation state, which can capture the main deformation patterns and features of the space frame structure. The model achieves a smooth transition between different deformation states through a latent space interpolation mechanism. The interpolation employs a combination of linear and spherical interpolation to ensure that the generated compensation parameters have physical meaning. The interpolation weight parameters are dynamically adjusted based on three key parameters: the complexity index of the space frame structure, the load variation coefficient, and the temperature sensitivity coefficient. The complexity index is calculated using geometric parameters such as the number of space frame nodes, the number of members, and the span ratio. The load variation coefficient is obtained through load statistical analysis, and the temperature sensitivity coefficient is calculated using the material's thermal expansion and contraction coefficient.

[0075] The training dataset was established by first collecting design drawings and construction records for different types of space frame structures, including flat space frames, spherical space frames, cylindrical space frames, hyperboloid space frames, and other structural forms, with spans ranging from 20m to 200m and heights from 3m to 30m. Geometric parameters of the space frame were extracted from the design drawings, including node coordinates, member lengths, cross-sectional dimensions, and connection methods, and digitized using computer-aided design software. Material property parameters were extracted from material test reports, including physical parameters such as yield strength, tensile strength, elastic modulus, Poisson's ratio, and coefficient of linear expansion of steel. Load case data were extracted from the construction records, including the magnitude and distribution of different types of loads such as dead load, live load, temperature load, and wind load. Actual deformation data were extracted from measurement records, including measured data such as displacement of key nodes, strain of members, and deviation of internal surfaces. A deformation data sample library containing 10,000 space frame engineering cases was established. Each sample includes an input parameter vector and an output parameter vector. The input parameter vector contains information such as space frame geometric parameters, material properties, and load conditions, while the output parameter vector contains the deformation values ​​and compensation parameters of each node. Ideal compensation parameters and actual construction effects were labeled for each sample. The ideal compensation parameters were calculated using finite element analysis, and the actual construction effects were verified using field measurement data. The sample set was expanded using data augmentation techniques, employing methods such as noise addition, parameter perturbation, and geometric transformation to generate new training samples, ultimately expanding to 50,000 training samples. A standardized input and output data format was established, and all data were normalized to ensure that parameters with different dimensions have the same numerical range. Input parameters were normalized to between 0 and 1, and output parameters were normalized to between -1 and 1.

[0076] The model training process employed the Adam optimizer to optimize model parameters. The initial learning rate was set to 0.001, using an exponential decay strategy, decreasing to 0.9 times the original learning rate every 50 training epochs. The batch size was set to 32, meaning 32 samples were used for parameter updates per training cycle. The training epochs were set to 200, with each epoch encompassing a complete traversal of all training samples. The mean squared error loss function was used to evaluate the model's prediction accuracy. The loss function also included a KL divergence term to constrain the distribution of the latent space, with a KL divergence weight coefficient set to 0.01. An early stopping mechanism was used to prevent overfitting; training automatically stopped when the validation set loss showed no improvement for 20 consecutive training epochs. Five-fold cross-validation was used to evaluate the model's generalization ability. The training dataset was divided into five subsets; four subsets were selected for training each time, and the remaining subset was used for validation, repeated five times to obtain the average validation error. Dropout was used during training to prevent overfitting, with a dropout probability set to 0.2, randomly discarding 20% ​​of the neuron connections. Batch normalization is employed to accelerate model convergence by adding a batch normalization layer after each convolutional and fully connected layer. Finally, the model parameters with the smallest validation set error are selected as the optimal model, and these parameters are saved for practical applications.

[0077] It should be noted that the key technical ideas of this invention mainly include core technologies such as sensor optimization deployment based on graph theory, establishment of a multi-condition deformation prediction database, and construction of an adaptive compensation model for structural deformation.

[0078] Graph theory-based sensor optimization deployment technology abstracts the network structure into an undirected graph model and uses a minimum dominating set algorithm to determine the optimal sensor placement locations. Compared to traditional empirical deployment methods, this technology can minimize the number of sensors while ensuring monitoring accuracy, thus reducing monitoring costs and improving monitoring efficiency. Traditional methods typically rely on engineering experience for sensor deployment, which can easily lead to monitoring blind spots or sensor redundancy. In contrast, the graph theory-based optimization deployment method can mathematically guarantee the optimality and integrity of the monitoring network.

[0079] The multi-condition deformation prediction database establishment technology, through extensive finite element calculations and statistical analysis of measured data, has established a dataset containing the deformation patterns of space frame structures under different load combinations. This provides a reliable data foundation for predicting the deformation response of space frame structures under various working conditions. Compared to traditional single-condition analysis methods, this technology considers the comprehensive influence of multiple factors such as temperature changes, wind loads, and construction loads, enabling more accurate prediction of the actual deformation of space frame structures and providing a scientific basis for developing reasonable pre-camber compensation schemes.

[0080] The structural deformation adaptive compensation model construction technology employs deep learning algorithms to establish an intelligent model capable of automatically adjusting compensation parameters based on real-time monitoring data. This model learns deformation patterns and compensation effects from a large amount of historical data, enabling accurate prediction and adaptive adjustment for unknown working conditions. Compared to traditional static compensation methods, this technology is dynamic and adaptive, allowing for timely adjustments to the compensation strategy based on changes during actual construction, significantly improving compensation accuracy and construction efficiency.

[0081] The synergistic effect of these key technological approaches forms a complete closed-loop control system for space frame deformation compensation. From the optimized layout of the monitoring network to the establishment of the deformation prediction database and the application of the adaptive compensation model, each link works closely together and supports each other. Compared with traditional open-loop control methods, this system has significant technical advantages, enabling a shift from passive response to proactive prevention, from experience-driven to data-driven approaches, and from static compensation to dynamic adaptive compensation. Ultimately, it achieves high-precision control of the internal surface accuracy of the space frame structure and a significant improvement in construction quality.

[0082] It should be noted that the present invention also solves the following two technical problems: (1) The technical problem of difficulty in controlling the accumulation of deformation error during the assembly of the space frame. In the traditional space frame assembly process, due to the lack of real-time deformation monitoring and error correction mechanism, the assembly error often accumulates gradually with the progress of construction, resulting in the final geometric accuracy deviation exceeding the allowable range. The present invention establishes a structural deformation adaptive compensation model and adopts a variational autoencoder model with a multi-layer encoder-decoder architecture, which can extract the deformation characteristics of the space frame in real time and output the corresponding compensation parameters. A phased detection strategy is implemented in the assembly process, that is, multi-level verification is performed after assembling a node, after assembling a frame, and after installing a segment. The assembly error is corrected in real time by the structural deformation adaptive compensation model, which effectively prevents the occurrence of error accumulation. (2) The technical problem of difficulty in accurately predicting the coupling effect of multiple load conditions during the construction of the space frame. During the construction process, the space frame structure will be subjected to the combined action of multiple loads such as self-weight, temperature change, and wind load. The coupling effect of these loads presents complex nonlinear characteristics, and the traditional single-condition analysis method is difficult to accurately predict the actual deformation. The multi-condition deformation prediction database established by this invention is corrected by combining finite element analysis software with measured data. It includes deformation calculation results for various construction conditions such as patching and unloading, roof concrete structure, cast-in-place concrete structure under space frame, and formwork removal. It can accurately predict the deformation response of space frame under various load combinations, providing a scientific theoretical basis for deformation compensation.

[0083] Specifically, the principle of this invention is as follows: The fundamental reason why this invention can solve the problem of difficulty in controlling the geometric accuracy of the inner surface of the cast-in-place slab caused by the deformation of the space frame lies in its establishment of a complete deformation prediction-monitoring-compensation closed-loop control system. First, the multi-condition deformation prediction database constructed by combining a precise mathematical model established through finite element analysis software with measured data correction can accurately predict the deformation response law of the space frame under different load combinations, providing a reliable theoretical basis for the calculation of compensation parameters. Second, the sensor deployment scheme optimized by the minimum dominance set algorithm ensures the full coverage and economy of the monitoring network, and the real-time collected three-dimensional displacement data provides accurate data support for deformation state assessment. Third, the pre-camber compensation optimization function, based on the pre-camber compensation parameters calculated from the multi-condition deformation prediction database, can effectively offset the downward deflection deformation caused by subsequent loads, controlling the impact of space frame deformation on the accuracy of the inner surface from the source. Fourth, the structural deformation adaptive compensation model adopts a variational autoencoder architecture, extracts the deformation features of the space frame through a multi-layer encoder-decoder structure, and the 16-dimensional latent space representation can capture complex deformation patterns. The interpolation mechanism realizes a smooth transition between different deformation states, ensuring the continuity and stability of compensation parameter adjustment. Finally, the synergistic effect of phased detection and comprehensive compensation measures forms a multi-level deformation control system. The deformation compensation feedback mechanism adjusts theoretical parameters based on feedback from actual construction results, achieving continuous optimization of compensation accuracy. This prediction-monitoring-compensation closed-loop control system can effectively intervene at each stage of space frame deformation, ensuring that the geometric accuracy of the inner surface of the lower cast-in-place slab always meets design requirements.

[0084] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0085] The specific implementation of step S01 involves establishing a finite element analysis model of the space frame structure and predicting the deformation response of the space frame structure under different load combinations by constructing a multi-load condition deformation prediction database. The deformation response of the space frame structure can be expressed as:

[0086] {u}=[K] -1 {F};

[0087] In the formula, {u} is the nodal displacement vector in mm, which includes the displacement components of all nodes in the x, y, and z directions; [K] is the overall structural stiffness matrix in N / mm, which is obtained by assembling the stiffness matrices of each element; {F} is the nodal load vector in N, which includes the concentrated forces and moments acting on each node.

[0088] The establishment of the multi-condition deformation prediction database is based on the superposition principle of temperature deformation, self-weight deformation, and wind load deformation, as specifically expressed below:

[0089] Δtotal =Δ temp +Δ self +Δ wind +ε model ;

[0090] In the formula, Δ total Δ represents the total deformation. temp Δ represents the temperature deformation. self Δ is the deformation due to its own weight. wind ε represents wind load deformation. model This is the model error term.

[0091] The formula for calculating temperature deformation is as follows:

[0092] Δ temp =α·L·ΔT;

[0093] In the formula, α is the coefficient of linear expansion of steel, which is taken as 1.2 × 10⁻⁶. -5 / ℃; L is the length of the rod, in meters; ΔT is the temperature change, in degrees Celsius.

[0094] The deformation due to self-weight was calculated using finite element analysis:

[0095]

[0096] In the formula, N e The number of units; [N] e [K] is the shape function matrix of the e-th element, dimensionless; e ] represents the stiffness matrix of the e-th element, in N / mm; {F self，e} represents the self-weight load vector of the e-th element, in units of N, calculated based on the element volume and material density.

[0097] The formula for calculating wind load deformation is:

[0098]

[0099] In the formula, q wind This is the wind pressure value, in N / m. 2 A projected The projected area of ​​the space frame in the wind direction, in meters (m²). 2 ;k structure C represents the overall stiffness characteristic value of the space frame, expressed in N / m, and is calculated through eigenvalue analysis. wind This is the wind load shape factor, dimensionless, determined based on the geometry of the grid structure and the wind direction.

[0100] The parameter is obtained as follows: Δ tempThe temperature-displacement relationship was obtained experimentally, including step 1: measuring the displacement of key nodes of the space frame under different temperature conditions; and step 2: determining the temperature-displacement relationship coefficients through linear fitting. self Static analysis was performed by inputting the geometric parameters and material properties of the space frame through finite element method calculations. Δ wind Obtained through wind tunnel testing or numerical simulation, including step 1: establishing a wind load model of the space frame structure; step 2: calculating the structural response under different wind directions and speeds. ε model The range is 0.05 to 0.15.

[0101] The specific implementation of step S02 is the same as described above, and will not be repeated in detail here.

[0102] The specific implementation of step S03 is to calculate the pre-camber compensation parameters during the space frame assembly stage using a pre-camber compensation optimization function. The pre-camber compensation optimization function is based on multi-objective optimization theory and is specifically expressed as follows:

[0103] h comp =f(W self T range P wind E material L max );

[0104] In the formula, h comp This is the pre-camber compensation height value, in mm; W self This is the distributed load of the space frame's self-weight, in N / m. 2 ;T range The design temperature range is given in °C; P wind The expected wind load value is expressed in N / m. 2 E material The elastic modulus of the material is expressed in Pa; L max This represents the maximum span of the space frame, measured in meters (m).

[0105] The calculation of the pre-camber compensation parameters adopts the principle of influence line superposition, as shown below:

[0106]

[0107] In the formula, h comp，i η represents the pre-camber compensation height for the i-th node, in mm. ij Δ is the dimensionless coefficient representing the influence of the j-th load on the i-th node; j σ represents the deformation caused by the j-th load, in mm; n represents the number of load cases; β is the safety factor, ranging from 1.2 to 1.5, dimensionless; deviation This represents the standard deviation of the prediction bias, in mm.

[0108] The parameter is obtained as follows: η ij The influence coefficient matrix of each load case on each node was calculated using finite element analysis. j The results were obtained experimentally, including step 1: measuring the structural response under different loads in an actual engineering project; and step 2: statistically analyzing the deformation patterns under each working condition. σ deviation The range is 2–8 mm.

[0109] The specific implementation of step S04 involves using a structural deformation adaptive compensation model to correct assembly errors in real time. The adjustment of the compensation parameters of the structural deformation adaptive compensation model is based on the feedback control principle, as specifically expressed below:

[0110]

[0111] In the formula, Δh adjust The compensation parameter adjustment amount is in mm; K p K is a proportionality constant, dimensionless. i The integral coefficient is expressed in units of s. -1 ;K d is the differential coefficient, in units of s; e is the measurement error, in units of mm.

[0112] The cumulative error control equation for staged detection is:

[0113]

[0114] In the formula, E cumulative The cumulative error is expressed in mm; m is the number of measurement points; x measured，k y measured，k , z measured，k x represents the measured coordinates of the k-th measurement point, in mm. design，k y design，k , z design，k Here are the design coordinates for the k-th measurement point, in mm.

[0115] The parameter acquisition method is: K p K i K d The control parameters, K, were obtained through a system identification method and determined using a step response experiment. p Typical values ​​are 0.5–2.0, K i Typical values ​​are 0.01–0.1 s. -1 K d Typical values ​​are 0.1–0.5 s. e is obtained in real-time using a laser total station with a measurement accuracy of ±1 mm. The calculation formula is e = h. measured -h target , where h measuredThe measured height is in mm or h. target The target height value is in mm.

[0116] The specific implementation of step S05 involves using a lifting deformation control algorithm to dynamically adjust the lifting speed and displacement of each hydraulic lifter. The lifting deformation control algorithm is based on multi-objective optimization theory and is specifically expressed as follows:

[0117] v i (t)=v ref +Δv i (t);

[0118] In the formula, v i (t) represents the lifting speed of the i-th hydraulic lifter at time t, in mm / min; v ref For reference lifting speed, the unit is mm / min; Δv i (t) represents the speed adjustment amount, in mm / min.

[0119] The calculation of the speed adjustment is based on the principle of minimizing displacement synchronization error:

[0120]

[0121] In the formula, K sync Synchronization control gain, unit is min -1 ;K damp s is the damping coefficient, dimensionless; i (t) represents the displacement of the i-th lifting point, in mm; s avg (t) represents the average displacement of all lifting points, in mm, and is calculated using the following formula: Where N lift This represents the total number of hydraulic lifters.

[0122] The parameter acquisition method is: K sync The typical value was determined through control system debugging and is 0.1–0.3 min. -1 K damp Determined through system stability analysis, the value is 0.05–0.15, dimensionless. i (t) is obtained in real time through a stroke sensor with an accuracy of ±0.1mm. lift The number of lifting points is determined based on the arrangement of the lifting points of the space frame structure, and is usually 4 to 20.

[0123] The specific implementation method of step S06 is the same as described above, and will not be repeated in detail here.

[0124] The specific implementation of step S07 involves establishing a deformation compensation effect evaluation system based on the surface deviation values ​​of the cast-in-place slab hanging under the space frame. The adaptive adjustment function of the compensation parameters is calculated based on a comprehensive adjustment factor, and is specifically expressed as follows:

[0125] γ=w1·C complexity +w2·C load +w3·C temp +w4·P precision ;

[0126] In the formula, γ is the comprehensive adjustment factor value, which is dimensionless and ranges from 0 to 1; C complexity C is a dimensionless index representing the complexity of a space frame structure. load C is the load variation coefficient, dimensionless; temp This is the temperature sensitivity coefficient, measured in °C. -1 ;P precision The required construction accuracy level is dimensionless; w1, w2, w3, and w4 are weighting coefficients, dimensionless, satisfying w1+w2+w3+w4=1.

[0127] The formula for calculating the complexity index of a space frame structure is:

[0128]

[0129] In the formula, N node N represents the number of network nodes. member L represents the number of space frame members. span λ represents the span of the space frame, in meters (m). topology is a topological complexity correction coefficient, dimensionless, with a value range of 0.8 to 1.2.

[0130] The formula for calculating the load variation coefficient is:

[0131]

[0132] In the formula, σ load This represents the standard deviation of the load, in N / m. 2 It is obtained through statistical calculation of historical load data; μ load This is the average load, in N / m. 2 , which is the arithmetic mean of all load samples.

[0133] The formula for calculating the temperature sensitivity coefficient is:

[0134]

[0135] In the formula, α is the coefficient of linear expansion of the material, with units of °C. -1 ;ΔT design The design temperature range is given in °C; h allowableThe allowable deformation value is expressed in mm and is determined according to building codes and usage requirements.

[0136] The adjustment strategy for the interpolation weight parameters is as follows: when γ∈[0, 0.3], a conservative interpolation strategy is adopted; when γ∈[0.3, 0.7], a standard interpolation strategy is adopted; and when γ∈[0.7, 1.0], an aggressive interpolation strategy is adopted.

[0137] The parameter acquisition method is: C complexity C is obtained through calculation of the geometric parameters of the space frame. load The data was obtained using statistical analysis methods, including step 1: collecting historical load data; and step 2: calculating the coefficient of variation of the loads. temp P is obtained by calculation based on the material's coefficient of thermal expansion and contraction and the range of temperature change. precision The accuracy level is determined according to the engineering accuracy requirements. Level 1 accuracy is 0.9–1.0, Level 2 accuracy is 0.7–0.9, and Level 3 accuracy is 0.5–0.7. Typical values ​​for w1, w2, w3, and w4 are 0.3, 0.25, 0.25, and 0.2, respectively.

[0138] It should be noted that the deformation response equation for a space frame structure is {u}=[K]. -1 {F} is based on linear elasticity theory and solves for nodal displacements through the inverse operation of the stiffness matrix. This equation can accurately describe the deformation law of the space frame structure under external loads. Compared with the traditional simplified calculation method, it provides more accurate deformation prediction results and provides a reliable theoretical basis for subsequent compensation parameter calculation.

[0139] Multi-condition deformation superposition equation Δ total =Δ temp +Δ self +Δ wind +ε model Based on the superposition principle, the deformation effects of different load conditions are linearly superimposed. This equation considers the comprehensive influence of multiple factors such as temperature, self-weight, and wind load. Compared with single-condition analysis methods, it can more comprehensively reflect the actual deformation state of the space frame structure, significantly improving the accuracy and reliability of deformation prediction. The self-weight deformation is calculated using the finite element analysis equation. Based on finite element theory, the overall structural response is calculated by assembling the element stiffness matrix and load vector. This equation can accurately consider the geometric characteristics and material nonlinearities of the space frame structure. Compared with traditional beam theory calculation methods, it provides more accurate prediction results of self-weight deformation, laying the foundation for multi-condition superposition analysis. Wind load deformation calculation equation. Based on structural dynamics theory, wind load deformation is calculated by the relationship between wind pressure, projected area and structural stiffness. This equation takes into account the pulsating characteristics of wind load and the dynamic response of the structure. Compared with static wind load calculation methods, it can more realistically reflect the impact of wind load on space frame structures.

[0140] Pre-camber compensation parameter calculation equation Based on influence line theory and probabilistic statistical methods, the contribution of each load to the nodal displacement is described by the influence coefficient matrix. By introducing a safety factor and statistical deviation term, the equation can comprehensively consider a variety of uncertainties. Compared with traditional empirical estimation methods, it provides more scientific and reasonable calculation results of compensation parameters and effectively reduces the cumulative error in the construction process.

[0141] PID control equations for the adaptive compensation model of structural deformation Based on classical control theory, the assembly error is corrected in real time through three stages: proportional, integral, and derivative. This equation has the characteristics of fast response speed, high steady-state accuracy, and strong anti-interference ability. Compared with the traditional open-loop control method, it can realize closed-loop feedback control, which significantly improves the assembly accuracy and construction quality.

[0142] Improve the velocity adjustment equation of the deformation control algorithm Based on synchronous control theory, a proportional-derivative control strategy is used to achieve displacement synchronization at each lifting point. This equation effectively suppresses displacement deviations and vibrations during the lifting process. Compared to the traditional constant-speed lifting method, it significantly improves the synchronization and stability of the lifting process, ensuring the geometric accuracy of the space frame structure during lifting. The formula for calculating the average displacement is as follows: The reference displacement of all lift points is determined by the arithmetic mean method, which provides a unified benchmark for synchronous control and ensures the stability and convergence of the control algorithm.

[0143] Compensation parameter adaptive adjustment function γ=w1·C complexity +w2·C load +w3·C temp +w4·P precision Based on fuzzy comprehensive evaluation theory, a comprehensive adjustment factor is calculated through multi-factor weighted summation. This function can automatically select an appropriate interpolation strategy according to the specific characteristics of the project. Compared with compensation methods with fixed parameters, it has stronger adaptability and flexibility, and can achieve optimal compensation results under different engineering conditions, realizing intelligent and personalized adjustment of the compensation strategy. The formula for calculating the complexity index of the space frame structure is also included. The geometric complexity of a space frame is reflected by the ratio of the number of nodes and members to the square of the span. The topological complexity correction coefficient further considers the influence of the space frame connection method, providing a quantitative complexity evaluation index for adaptive adjustment. (Load variation coefficient calculation formula) The ratio of standard deviation to mean quantifies the dispersion of the load, reflecting the level of uncertainty and providing a basis for the safe adjustment of compensation parameters. Temperature sensitivity coefficient calculation formula. Taking into account the influence of material properties, temperature variation range and structural dimensions on temperature deformation, the temperature sensitivity level is determined by comparing it with the allowable deformation value, providing a scientific basis for the formulation of temperature compensation strategies.

[0144] To better understand and implement this invention, a specific application scenario of the invention is provided below as Example 2: A major technical challenge faced by a wind tunnel project is that the designed distance between the lower end of the main steel beam suspended from the space frame and the surface of the cast-in-place slab template is 5cm. Furthermore, bidirectional bottom reinforcement bars for the cast-in-place slab are installed between the lower end of the main steel beam and the template, with the bars having a diameter of 20mm and occupying a height of 40mm. Given the high precision requirements of the cast-in-place slab, there is no room for adjustment. Therefore, controlling the deformation of the space frame is a key challenge in controlling the internal surface accuracy of the cast-in-place slab. The technical team decided to use the space frame deformation compensation method of this invention to address this technical problem.

[0145] The technical team first established an accurate finite element analysis model of the space frame structure, using ANSYS software to construct a three-dimensional geometric model containing 8760 nodes and 15420 members. During the model calibration phase, model parameters were adjusted based on measured data of deflection under the space frame's self-weight, thermal expansion and contraction caused by temperature changes, and deformation under wind load. Through comparative analysis, the elastic modulus of the steel was corrected to 208000 MPa, Poisson's ratio to 0.28, and the coefficient of linear expansion was set to 1.2 × 10⁻⁶. -5 / ℃. The technical team established a multi-condition deformation prediction database containing calculation results for the stress and deformation of the infill unloading structure, the roof concrete structure, the cast-in-place concrete structure under the space frame, and the deformation of the formwork removal structure. This database covers all stages of the construction process, providing complete data support for subsequent deformation prediction and compensation parameter calculation. The deformation calculation results for the formwork removal structure are as follows: Figure 3 As shown.

[0146] During the sensor deployment phase, the technical team abstracted the space frame structure as an undirected graph model and used the minimum dominating set algorithm to determine the optimal placement of 128 sensors. Based on the algorithm's calculations, 64 high-precision displacement sensors were deployed at the nodes where the cast-in-place slabs are suspended below the space frame, 32 sensors at the upper chord nodes, 24 sensors at the lower chord nodes, and 8 sensors at the middle chord nodes. All sensors achieved a measurement accuracy of ±0.1mm, with a sampling frequency set to 1Hz. Simultaneously, 16 laser rangefinders were deployed at key control points, achieving a measurement accuracy of ±0.05mm, establishing a three-dimensional deformation monitoring network covering the entire space frame structure. Real-time data transmission was achieved using LoRa wireless communication technology, covering the entire 5km construction site.

[0147] Based on a multi-condition deformation prediction database, the technical team calculated the pre-camber compensation parameters for the space frame assembly stage. A multi-objective optimization problem was solved using a genetic algorithm, with a population size of 100 individuals, a crossover probability of 0.8, and a mutation probability of 0.02. After 200 generations of iterative calculations, the pre-camber compensation parameters for each node were determined. The pre-camber compensation values ​​for the nodes of the cast-in-place slabs hanging below the space frame range from 15mm to 35mm, for the upper chord nodes from 8mm to 25mm, for the lower chord nodes from 12mm to 28mm, and for the middle chord nodes from 10mm to 22mm. The technical team created a lookup table of these compensation parameters to guide on-site assembly construction. Deformation control measures during the space frame assembly stage are shown in Table 1.

[0148] Table 1 Deformation control during the space frame assembly stage

[0149]

[0150] During the assembly of the space frame, the technical team used a laser total station for high-precision positioning, achieving a measurement accuracy of ±1mm. A phased inspection method was strictly implemented during assembly; three-dimensional coordinate measurements were performed immediately after each node was assembled, ensuring that node position deviations were controlled within ±2mm. After assembling a space frame, an overall geometric check was conducted, measuring parameters such as span, height, and inclination. After installing a segment, a comprehensive verification was performed, including checking the coordinates of all nodes, member lengths, and connection angles within the segment. An adaptive compensation model for structural deformation analyzed the measurement data in real time, such as... Figure 2 The model shown employs a variational autoencoder architecture. The encoder contains three convolutional layers and two fully connected layers, while the decoder uses a symmetrical structure. The intermediate latent space is represented by a 16-dimensional vector to indicate the deformation state of the space frame. When an assembly error exceeds the allowable range, the model automatically calculates error correction parameters to guide on-site adjustments to node positions, replacement of member lengths, and recalibration of connection angles.

[0151] The technical team fabricated a steel space frame test specimen and conducted a simulation analysis of the construction process, taking the first diffusion section A-3 / A-5 axis as an example. The specimen fabrication process is shown in Table 2. By comparing the deformation values ​​before and after the pre-camber setting with the finite element simulation analysis results, the effectiveness of the compensation method was verified. The test results show that the pre-camber compensation can effectively offset the deflection deformation caused by the self-weight of the space frame and the construction load, and the internal surface deviation is controlled within ±3mm.

[0152] Table 2. Fabrication of Steel Space Frame Test Specimens

[0153]

[0154] Regarding the assembly site setup, the technical team addressed the assembly area's needs. After compacting the subgrade, a 50mm thick layer of crushed stone was laid, followed by a 200mm thick layer of plain concrete, and then steel plates were laid to form a stable assembly platform. Figure 4 As shown in Table 3, the assembly jig uses the original design nodes and circular tubes of the lower chord as the jig, while the upper chord uses adjustable steel lattice columns to precisely control the position of each node and member. The jig setup ensures the high precision required during the assembly process.

[0155] Table 3 On-site assembly jig setup

[0156]

[0157] Based on the space frame structure drawings, ground survey lines are laid out on the ground. The process of setting up the ground survey lines is shown in Table 4. Precise positioning using the ground survey lines ensures the accurate placement of the assembly jig, providing a benchmark for subsequent member installation and overall measurement. The jig positioning and elevation measurement process is as follows: Figure 5 , Figure 6 As shown, a total station was used for precise measurement to ensure that the position and elevation of the jig met the design requirements.

[0158] Table 4. Plot Line Setting

[0159] Serial Number Installation steps Key Control Points 1 Laying out the layout and positioning the assembly frame Ensure the accuracy of the tire frame position 2 Measurement and positioning during pole installation Real-time monitoring of assembly accuracy 3 Measurement and positioning during pole installation Detect and correct deviations in a timely manner 4 Measurements were taken after the entire space frame was installed. Verify overall accuracy

[0160] After the space frame assembly was completed, the lifting phase began. The technical team employed a synchronous hydraulic lifting system, equipped with 16 hydraulic lifters, each with a lifting capacity of 1500kN and a stroke range of 2000mm. The lifting process was computer-controlled, with stroke sensors monitoring the displacement of each lifting point in real time, and the control cycle set to 100ms. When a displacement deviation exceeding ±10mm was detected, the system automatically adjusted the lifting speed to ensure that the synchronization error of the displacement at each lifting point was controlled within ±20mm. The lifting process control measures are shown in Table 5, ensuring the geometric accuracy of the space frame during the lifting process through precise synchronous control.

[0161] Table 5. Improvement of Process Control Measures

[0162] Serial Number content Control requirements 1 Lifting point measurement Real-time monitoring of displacement synchronization at each lifting point 2 Measurements were taken after the space frame was lifted into place. Verify the elevation and position accuracy after the upgrade. 3 Post-component measurement Ensure connection accuracy meets requirements

[0163] The technical team optimized the lifting point placement scheme by comparing multiple options. The space frame underwent overall hydraulic lifting and segmented hoisting simulation analysis. By comparing the deformation of the space frame under various lifting schemes, the 16-point lifting scheme with the least deformation was ultimately selected. The space frame lifting point placement and deformation analysis are as follows: Figure 7 , Figure 9 , Figure 9 As shown, ensure that the deformation of the space frame meets the design requirements. After being lifted into place, the maximum deflection of the space frame is controlled within 25mm, and the displacement deviation of each key control point is within ±3mm.

[0164] After the space frame was in place, the technical team used 3D laser scanning technology to comprehensively measure the internal surface of the cast-in-place slab hanging under the space frame. The laser scanner had a measurement accuracy of ±0.5mm, a scanning speed of 1 million points / second, and a point cloud density controlled at 1000 points per square meter. The deviation between the internal surface and the design reference surface was calculated using point cloud data processing software, establishing a deformation compensation effect evaluation system. Based on the deviation analysis results, the technical team adopted comprehensive compensation measures for precise adjustments. The elevation of the support frame was adjusted within ±25mm, temporary support structures were set up to distribute the load transfer path, and the subsequent construction load was controlled to be evenly distributed. Hydraulic jacks were used to make fine adjustments within a range of ±3mm to key control points.

[0165] After systematic deformation compensation processing, the final surface accuracy of the cast-in-place slab under the space frame reached ±3.5mm, meeting the design requirement of ±5mm. In this embodiment, the distribution diagram of space frame node deformation monitoring data, the optimization and convergence process of pre-arching compensation parameters, and the effect of synchronous error control during the lifting process are shown below. Figure 10-12 As shown, the technical team established a deformation compensation feedback mechanism, adjusting theoretical calculation parameters based on actual compensation effects and continuously optimizing the prediction accuracy of the compensation model through machine learning algorithms. During subsequent construction, the deformation of the space frame was continuously monitored, and timely measures were taken to address abnormal deformations, ensuring that the internal surface accuracy always met requirements throughout the entire construction process.

[0166] The technological advancements brought about by this invention compared to traditional methods are mainly reflected in the following aspects: Traditional methods for controlling the deformation of space frames typically employ empirical estimation and static compensation strategies, lacking systematic theoretical support and accurate predictive capabilities. This invention, by establishing a multi-condition deformation prediction database and employing finite element analysis, can accurately predict the deformation response of the space frame under various load combinations, providing a scientific basis for calculating compensation parameters. Traditional methods often rely on engineering experience in sensor deployment, easily leading to monitoring blind spots or redundant deployment. This invention uses a minimum dominator set algorithm based on graph theory, mathematically ensuring the optimality and integrity of the monitoring network, improving monitoring accuracy while reducing the number of sensors. Traditional compensation strategies are mostly open-loop control, unable to dynamically adjust according to actual construction conditions. The adaptive compensation model for structural deformation constructed in this invention has closed-loop feedback control capabilities, automatically adjusting compensation parameters based on real-time monitoring data, realizing a shift from passive response to proactive prevention. Traditional methods for lifting space frames often employ constant-speed lifting, which makes it difficult to ensure synchronization at each lifting point. The synchronous hydraulic lifting system employed in this invention significantly improves the synchronization and stability of the lifting process through real-time feedback control and dynamic speed adjustment, ensuring the geometric accuracy of the space frame during lifting. Traditional deformation compensation methods lack systematicity and adaptability. This invention, through an adaptive adjustment function for compensation parameters, can automatically select the optimal compensation strategy based on the specific characteristics of the project. Compared to methods with fixed parameters, it has greater adaptability and flexibility, achieving the best compensation effect under different engineering conditions.

[0167] It should be noted that the variables involved in this invention are explained in detail in Tables 6 and 7 below.

[0168] Table 6. Variable Explanation Table (Part 1)

[0169]

[0170]

[0171] Table 7. Variable Explanation Table (Part Two)

[0172]

[0173] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for compensating for the influence of grid deformation on the accuracy of the inner profile, characterized in that, The precise mathematical model of the grid structure is established by using a finite element analysis software, model correction is performed in combination with measured data of deflection deformation under the self-weight of the grid structure, measured data of thermal expansion and contraction caused by temperature change, and measured data of deformation under wind load, and a multi-working condition deformation prediction database is established, which contains stress deformation calculation results of the embedded unloading structure, stress deformation calculation results of the roof concrete structure, stress deformation calculation results of the grid structure underhung cast-in-place slab concrete structure, and deformation calculation results of the formwork removal structure; A three-dimensional deformation monitoring network is established by modeling the grid structure as an undirected graph and determining the optimal sensor layout position by using a minimum dominating set algorithm, and three-dimensional displacement data of each key control point of the grid structure are collected in real time; Pre-camber compensation parameters in the grid assembly stage are calculated based on the multi-working condition deformation prediction database and the three-dimensional displacement data; the assembly error is corrected in real time in the grid assembly process by using a structure deformation adaptive compensation model; the lifting speed and displacement of each hydraulic lifter are dynamically adjusted in the grid lifting process by using a lifting deformation control algorithm; the key control points of the inner surface of the grid underhung cast-in-place slab are precisely measured, and comprehensive compensation measures are adopted for deformation compensation according to the deviation value of the inner surface of the grid underhung cast-in-place slab.

2. The method of claim 1, wherein, The multi-working condition deformation prediction database is a data set containing the deformation law of the grid structure under different load combinations, which is established by a large number of finite element calculations and statistical analysis of measured data, and is used to predict the deformation response of the grid structure under various load combinations.

3. The method of claim 2, wherein the method further comprises: determining a deformation of the grid structure; and determining a compensation value for the inner profile based on the deformation of the grid structure. The three-dimensional deformation monitoring network is established by arranging high-precision displacement sensors and laser range finders at the grid underhung cast-in-place slab nodes, the grid top chord nodes, the grid bottom chord nodes, the grid middle chord nodes, the grid mid-span positions, and the grid support positions, and establishing a three-dimensional deformation monitoring network covering the entire grid structure.

4. The method of claim 3, wherein the method further comprises: determining a compensation value based on the determined deformation of the grid structure; and applying the compensation value to the inner profile of the grid structure. The grid underhung cast-in-place slab node is a structure connection point with a design distance of 5 cm from the surface of the grid underhung cast-in-place slab formwork.

5. The method of claim 4, wherein, The grid top chord node is a structure connection point using a steel lattice column as an assembly jig frame at the upper part of the grid structure.

6. The method of claim 5, wherein, The calculation steps of the pre-camber compensation parameters are specifically calculating the pre-camber compensation parameters in the grid assembly stage by using a pre-camber compensation optimization function, and determining the preset elevation compensation values of the grid underhung cast-in-place slab nodes, the grid top chord nodes, the grid bottom chord nodes, and the grid middle chord nodes.

7. The method of claim 6, wherein the method further comprises: The pre-camber compensation optimization function is used to calculate the optimal pre-camber compensation parameters of each node in the grid assembly stage, and the input includes the grid self-weight distributed load from the structure design, the design temperature change range from the design specification, the expected wind load value from the design specification, the material elastic modulus from the material performance, and the maximum span of the grid structure from the structure design, and the output is the pre-camber compensation height value of each node for guiding the assembly construction.

8. The method of claim 7, wherein, The step of correcting the assembly error in real time during the assembly of the net rack, specifically, adjusting the elevation of the net rack lower hanging cast-in-place slab node, the net rack upper chord node, the net rack lower chord node and the net rack middle chord node according to the pre-arch compensation parameters during the assembly of the net rack, using a laser total station to position and control the node three-dimensional coordinates and elevation, using phased detection, and correcting the assembly error in real time through a structure deformation adaptive compensation model.

9. The method of claim 8, wherein, The structure deformation adaptive compensation model adopts a variational autoencoder model with a multi-layer encoder-decoder architecture, the encoder part includes three convolutional layers and two fully connected layers for extracting net rack deformation features, the decoder part adopts a symmetric structure to output compensation parameters, and the intermediate latent space adopts a 16-dimensional vector to represent the net rack deformation state.

10. The method of claim 9, wherein, The comprehensive compensation measure is a systematic deformation compensation method that combines the use of four compensation methods, namely, support cradle elevation adjustment, temporary support structure setting, subsequent construction load distribution control and net rack node position fine tuning.

Images