Construction method for arched large-span shell steel roof

Abstract

The invention provides a method for constructing an arch-shaped large-span shell steel roof, and belongs to the technical field of arch-shaped large-span shell steel roof construction. A temporary supporting system is built, and distributed optical fiber sensors are arranged to monitor the lifting process of a main arch; a cable force distribution scheme of a cable supporting system is calculated by adopting a double-layer game optimization model to realize cooperative control of stress uniformity and form precision, tensioning force is applied by using an intelligent tensioning system in stages, and a crane moving path is planned in combination with a path optimization algorithm based on a graph theory to realize graded unloading of a temporary supporting system. The technical problem that it is difficult to achieve structural stress uniform distribution and form accurate control of an arched large-span shell steel roof at the same time in the tensioning process of a cable supporting system is solved.

Description

Technical Field

[0001] This invention belongs to the field of construction technology for arched large-span shell steel roofs, and more specifically, it relates to a method for constructing arched large-span shell steel roofs. Background Technology

[0002] Arched large-span shell steel roofs, as an important structural form for large public buildings, are typically constructed using a combination of integral lifting and cable-stayed systems. Traditional construction techniques, after the main arch is lifted into place, adjust the structural stress state and geometry through tensioning the cable-stayed system. The conventional approach involves cable force distribution based on a single objective, either prioritizing stress control while neglecting morphological accuracy requirements, or focusing solely on morphological control leading to uneven stress distribution. In existing technologies, due to the complex mechanical coupling relationships between the cables in the cable-stayed system, adjusting the tension of a single cable simultaneously affects the stress distribution and geometry of the structure. Stress uniformity and morphological accuracy are often mutually restrictive and difficult to achieve simultaneously. When using traditional single-objective optimization methods for cable force distribution, the coupling relationship and constraint conflicts between the two control objectives cannot be effectively addressed, resulting in excessive local stress concentration or morphological deviations exceeding permissible limits after tensioning. This necessitates repeated cable force adjustments, increasing construction time and costs. In other words, existing technologies present a technical challenge in simultaneously achieving uniform stress distribution and precise morphological control during the tensioning process of the cable-stayed system for arched large-span shell steel roofs. Summary of the Invention

[0003] In view of this, the present invention provides a method for constructing an arched large-span shell steel roof, which can solve the technical problem in the prior art that it is difficult to simultaneously achieve uniform distribution of structural stress and precise shape control during the tensioning process of the cable support system for arched large-span shell steel roofs.

[0004] This invention is implemented as follows: A method for constructing an arched, large-span shell steel roof is provided. A temporary support system is built, and wind-resistant columns and a hydraulic synchronous lifting device are installed above the irregularly shaped concrete beams. The main arch structure is assembled on the ground, and distributed fiber optic strain sensors are deployed at key stress points. The hydraulic synchronous lifting device is activated for overall lifting, and strain data is collected in real time to adjust the lifting speed. After lifting, the cable support system is installed, and an influence matrix is ​​established. A two-layer game optimization model is used to calculate the cable force distribution scheme. The two-layer game optimization model outputs the upper-layer cable force adjustment coefficient with the goal of maximizing stress uniformity and transmits it to the lower-layer optimization model. The lower-layer optimization model outputs the lower-layer cable force adjustment coefficient with the goal of maximizing morphological accuracy and feeds it back to the upper-layer optimization model. Information exchange is achieved through coupling terms, and iterative solutions are used to achieve dynamic equilibrium between the two control objectives through mutual constraints. The target cable force value is calculated based on the cable force adjustment coefficient obtained from the iterative solution and the influence matrix. An intelligent tensioning system is used to apply tension force in stages, and a graded unloading scheme for the temporary support system is formulated while monitoring the structural displacement state.

[0005] In the step of building a temporary support system, a node network model is established at the construction site for crane path planning.

[0006] The node network model refers to discretizing the construction site into a 3m×3m grid, with each grid intersection serving as a node, and the lines connecting the nodes representing the travel paths of cranes and transportation equipment.

[0007] In the ground assembly of the main arch structure, the main arch adopts a box-section beam to complete the connection between the main arch and the single-layer steel grid shell.

[0008] The distributed optical fiber strain sensor adopts a wavelength modulation fiber optic grating sensor with a sensor spacing of ≤2m, a measurement accuracy of ±1 microstrain, and a sampling frequency of ≥10Hz.

[0009] During the overall lifting process, the stress deviation coefficient and displacement deviation coefficient of each monitoring point are calculated. When the stress deviation coefficient is greater than 0.15 or the displacement deviation coefficient is greater than 0.12, the lifting speed of the corresponding hydraulic lifting point is adjusted.

[0010] The stress deviation coefficient is calculated by dividing the absolute value of the difference between the measured stress value and the design stress value at each monitoring point by the design stress value, resulting in the maximum value.

[0011] The displacement deviation coefficient is calculated by dividing the absolute value of the difference between the measured displacement and the predicted displacement at each monitoring point by the predicted displacement, which is the maximum value.

[0012] When adjusting the jacking speed, reduce the jacking speed to 60-75% of the original speed until both the stress deviation coefficient and the displacement deviation coefficient are <0.10.

[0013] The influence matrix is ​​established by applying a unit tension increment to each cable, calculating the stress and displacement changes at key structural nodes, and arranging the influence coefficients of all cables in rows and columns to form a matrix.

[0014] The two-layer game optimization model is a master-slave game system consisting of an upper-layer optimization model and a lower-layer optimization model. The upper-layer optimization model acts as the dominant player, making priority decisions, while the lower-layer optimization model acts as the follower, responding based on the decisions made by the upper-layer model. During the iterative solution process, iteration stops when the changes in both the upper-layer and lower-layer force adjustment coefficients are less than 0.02.

[0015] The process involves using an intelligent tensioning system to apply tension to the cables. The tensioning process is divided into an initial tensioning stage and a fine tensioning stage. In the initial tensioning stage, 55% to 65% of the target cable force is applied, and in the fine tensioning stage, the remaining cable force is applied.

[0016] The tiered unloading scheme of the temporary support system divides the temporary support system into a main support area and a secondary support area. The unloading process is carried out in 4 to 6 batches, with each batch unloading 15% to 20% of the design load.

[0017] During the unloading process, a graph-based path optimization algorithm is used to plan the crane's movement path, and a spatiotemporal occupancy matrix is ​​established to avoid equipment collisions. After each batch of unloading is completed, the change in vertical displacement of the structure is monitored. When the change in vertical displacement exceeds 2.5% of the design elevation, unloading is paused and subsequent unloading plans are re-evaluated.

[0018] This invention establishes a two-layer game-theoretic optimization model to calculate cable force distribution schemes. Stress uniformity control and morphological accuracy control are set as two coupled optimization levels. The upper-layer optimization model aims to maximize stress uniformity and outputs upper-layer cable force adjustment coefficients, while the lower-layer optimization model aims to maximize morphological accuracy and outputs lower-layer cable force adjustment coefficients. The two models exchange information and iteratively solve the problem until convergence through coupling terms. This method transforms the traditional single-objective optimization problem into a master-slave game system. The upper-layer optimization model, as the dominant party, prioritizes the need for uniform structural stress distribution and outputs adjustment parameters. The lower-layer optimization model, as the follower party, responds with morphological control based on the upper-layer decision and feeds back adjustment parameters. Through bidirectional coupling and iterative mechanisms, the two control objectives achieve dynamic equilibrium through mutual constraints, overcoming the shortcomings of traditional single-objective optimization methods that cannot coordinate the contradiction between stress control and morphological control. In summary, this invention solves the technical problem mentioned in the background art of simultaneously achieving uniform structural stress distribution and precise morphological control during the tensioning process of an arched large-span shell steel roof system with cable support. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention.

[0020] Figure 2 This is a schematic diagram of an arched, large-span steel roof shell. Detailed Implementation

[0021] 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.

[0022] like Figure 1 The diagram shows a flowchart of a method for constructing an arched, large-span shell steel roof according to the present invention. This method includes the following steps:

[0023] S01. Construct a temporary support system, set up wind-resistant columns above the irregular concrete beams, install hydraulic synchronous lifting devices at the top of the wind-resistant columns, and establish a node network model at the construction site for crane path planning.

[0024] S02. Assemble the main arch structure on the ground. The main arch adopts a box-section beam. During the assembly process, distributed fiber optic strain sensors are installed at the key stress-bearing parts of the main arch. The sensor spacing does not exceed 2m. The connection between the main arch and the single-layer steel mesh shell is completed.

[0025] S03. Start the hydraulic synchronous lifting device for overall lifting. During the lifting process, collect data from the distributed fiber optic strain sensor in real time, calculate the stress deviation coefficient and displacement deviation coefficient of each monitoring point. When the stress deviation coefficient exceeds 0.15 or the displacement deviation coefficient exceeds 0.12, adjust the lifting speed of the corresponding hydraulic lifting point and reduce the lifting speed to 60% to 75% of the original speed until the stress deviation coefficient and displacement deviation coefficient both fall back to below 0.10.

[0026] S04. After the main arch is lifted to the design elevation, the cable support system is installed, the initial length and installation angle of each cable are measured, an influence matrix is ​​established to describe the influence of each cable tension on the stress distribution and shape accuracy of the structure, and a two-layer game optimization model is used to calculate the cable force distribution scheme. The two-layer game optimization model includes an upper optimization model with the goal of maximizing the uniformity of structural stress and a lower optimization model with the goal of maximizing the shape control accuracy.

[0027] S05. In the upper-level optimization model, input the maximum stress monitoring value, the minimum stress monitoring value, and the average stress monitoring value of the structure, calculate the stress uniformity evaluation function, and the output value of the stress uniformity evaluation function is the upper-level cable force adjustment coefficient. The upper-level cable force adjustment coefficient is then passed to the lower-level optimization model as a coupling term.

[0028] S06. In the lower-level optimization model, the actual measured value of the structural sag, the allowable value of the morphological deviation, and the upper-level cable force adjustment coefficient are input. The morphological accuracy evaluation function is calculated. The output value of the morphological accuracy evaluation function is the lower-level cable force adjustment coefficient. The lower-level cable force adjustment coefficient is fed back to the upper-level optimization model as a coupling term. Iterative solutions are performed until the changes in both the upper-level cable force adjustment coefficient and the lower-level cable force adjustment coefficient are less than 0.02.

[0029] S07. Based on the upper and lower cable force adjustment coefficients obtained from the iterative solution, and combined with the influence matrix, calculate the target cable force value of each cable. Use an intelligent tensioning system to apply tension force to the cables. The tensioning process is divided into an initial tensioning stage and a fine tensioning stage. In the initial tensioning stage, apply 55% to 65% of the target cable force value. In the fine tensioning stage, apply the remaining cable force value. After each set of cables is tensioned, measure the three-dimensional coordinates and stress state of the key nodes of the structure.

[0030] S08. After the cable support system is tensioned, a graded unloading plan for the temporary support system is formulated. The temporary support system is divided into a main support area and a secondary support area. The unloading process is carried out in 4 to 6 batches. The unloading amount of each batch is 15% to 20% of the design load. Before unloading, the bearing capacity of the temporary support system is verified by a pre-compression test. During the unloading process, a path optimization algorithm based on graph theory is used to plan the crane's movement path and establish a spatiotemporal occupancy matrix to avoid equipment collisions.

[0031] S09. After each batch of unloading is completed, monitor the change in vertical displacement of the structure. When the change in vertical displacement exceeds 2.5% of the design elevation, suspend unloading and re-evaluate the subsequent unloading plan. When the change in vertical displacement is less than 2.5% of the design elevation, continue to execute the next batch of unloading until the temporary support system is completely unloaded.

[0032] The node network model discretizes the construction site into a 3m×3m grid, with each grid intersection serving as a node. The lines connecting the nodes represent the travel paths of cranes and transportation equipment. A path search algorithm is used to calculate the optimal path from the component storage area to the installation location.

[0033] The distributed optical fiber strain sensor is a wavelength-modulated fiber Bragg grating sensor. It reflects the strain state of the structure by measuring the center wavelength offset of the fiber Bragg grating. The sensor has a measurement accuracy of ±1 microstrain and a sampling frequency of not less than 10Hz.

[0034] The stress deviation coefficient is calculated by dividing the absolute value of the difference between the measured stress value and the design stress value at each monitoring point by the design stress value, and the displacement deviation coefficient is calculated by dividing the absolute value of the difference between the measured displacement and the predicted displacement at each monitoring point by the predicted displacement.

[0035] The influence matrix is ​​established by applying a unit tension increment to each cable, calculating the stress and displacement changes at key structural nodes, and arranging the influence coefficients of all cables in rows and columns to form a matrix. The influence matrix is ​​used to describe the linear relationship between cable force adjustment and structural response.

[0036] The two-layer game optimization model is a master-slave game system consisting of an upper-layer optimization model and a lower-layer optimization model. The upper-layer optimization model acts as the dominant party and makes priority decisions, while the lower-layer optimization model acts as the follower and responds based on the decisions made by the upper-layer model. The two-layer models achieve information exchange and collaborative optimization through coupling terms.

[0037] The objective function of the upper-level optimization model is used to maximize the uniformity of the structural stress distribution. The inputs include the maximum stress monitoring value, the minimum stress monitoring value, and the average stress monitoring value of the structure. The output is the upper-level cable force adjustment coefficient. The objective function is expressed as follows: ,in This represents the maximum stress monitoring value of the structure. This represents the minimum stress monitoring value for the structure. This represents the average stress monitoring value of the structure. The reference stress value is taken as 100 MPa. This is the upper cable tension adjustment coefficient. The value is a dimensionless stress uniformity evaluation value, and the constraint condition is: and .

[0038] The objective function of the lower-level optimization model is used to maximize the structural morphology control accuracy. The inputs include the actual measured structural elevation, the allowable morphological deviation, and the upper-level cable force adjustment coefficient. The output is the lower-level cable force adjustment coefficient. The objective function is described as follows: ,in This is the actual measured value of the structural elevation. The design elevation is set at 6m. The allowable value for morphological deviation is taken as 0.09m. This is the upper cable tension adjustment coefficient. This is the adjustment coefficient for the lower cable force. The value is a dimensionless morphological accuracy evaluation value, with the following constraints: and .

[0039] The upper cable tension adjustment coefficient is an adjustment parameter obtained by the upper optimization model through maximizing the stress uniformity evaluation value. It is used to control the distribution ratio of cable tension to achieve uniform stress distribution in the structure. The upper cable tension adjustment coefficient is transmitted to the lower optimization model as a coupling term to affect the morphological control decision.

[0040] The lower cable tension adjustment coefficient is an adjustment parameter obtained by the lower optimization model through maximizing the morphological accuracy evaluation value. It is used to control the distribution ratio of cable tension force to achieve precise control of structural morphology. The lower cable tension adjustment coefficient is fed back to the upper optimization model as a coupling term, affecting the stress uniformity decision.

[0041] The maximum stress monitoring value, minimum stress monitoring value, and average stress monitoring value of the structure are obtained by distributed fiber optic strain sensors at key stress-bearing parts of the main arch. The actual sag measurement value of the structure is obtained by measuring the elevation difference between the arch crown and the arch foot using a total station.

[0042] The intelligent tensioning system includes a tensioning jack, a pressure sensor, and a displacement sensor. It achieves precise application of tension force through closed-loop control, with tension force control accuracy reaching ±3% of the target cable force value and displacement control accuracy reaching ±2mm.

[0043] The preloading test involves applying 1.2 times the design load to the temporary support system for a duration of no less than 24 hours, monitoring the deformation and stress state of the temporary support system, and verifying whether the bearing capacity and stability of the temporary support system meet the construction requirements.

[0044] The graph-based path optimization algorithm uses the crane's starting and target positions as the start and end points of the graph, and uses a path search algorithm to calculate the shortest path after considering obstacles and restricted areas. The shortest path comprehensively considers path length, number of turns, and operation time.

[0045] The spatiotemporal occupancy matrix is ​​a three-dimensional matrix. The first two dimensions represent the spatial location of the construction site, and the third dimension represents time. The matrix elements record the equipment occupancy status at the corresponding spatiotemporal location. By detecting whether the spatiotemporal occupancy areas of different equipment overlap, it is determined whether there is a collision risk. When a collision risk is detected, the operating time window of the equipment is adjusted or the crane movement path is replanned.

[0046] The main support area is located at the mid-span and arch foot of the structure and bears the main construction load. The secondary support area is located in the area where the structure is under relatively less stress and plays an auxiliary support role. The unloading sequence is to unload the secondary support area first and then unload the main support area.

[0047] The specific implementation methods of the above steps are described in detail below.

[0048] The specific implementation of step S01 is as follows: First, wind-resistant columns are installed on the top of the irregular concrete beams at the construction site. The stability of the wind-resistant columns is ensured by connecting them with embedded parts. The spacing of the wind-resistant columns is determined according to the structural span, generally 6 to 8 meters. A hydraulic synchronous lifting device is fixed on the top of the wind-resistant columns. This device uses a hydraulic pump station to centrally control multiple hydraulic jacks to achieve synchronous lifting. The system is equipped with displacement sensors and pressure sensors to monitor the synchronicity of the lifting process in real time. Then, the construction site is measured and discretized according to a 3m×3m grid size. Each grid intersection is defined as a node and assigned a unique number. The three-dimensional coordinates of the nodes are determined by measurement. Lines are established between nodes according to the actual passage conditions to represent feasible paths. For locations with obstacles or restricted areas, no lines are established. The shortest path algorithm in graph theory, such as Dijkstra's algorithm, is used to calculate the optimal path from the component storage area to the installation position. This path comprehensively considers the goals of minimizing the path length and the number of turns. By establishing a node network model, basic data support is provided for the subsequent path planning of cranes and transportation equipment, ensuring the efficiency and safety of equipment movement during construction.

[0049] The specific implementation of step S02 involves ground assembly of the main arch structure on an assembly platform at the construction site. The main arch adopts a box-section beam design to provide sufficient bending and torsional stiffness. The assembly process proceeds segment by segment from the arch foot to the arch crown. Adjacent beam segments are connected by high-strength bolts. Distributed fiber optic strain sensors are deployed in key stress-bearing areas of the main arch, including the arch foot region, the quarter-span region, and the arch crown region. The sensors employ wavelength-modulated fiber optic grating technology, reflecting the strain state of the structure by measuring the center wavelength offset of the fiber Bragg grating. The spacing between the sensors does not exceed 2m to ensure effective capture. To capture detailed changes in structural stress distribution, the sensor measurement accuracy reaches ±1 micro-strain, and the sampling frequency is set to no less than 10Hz to meet the needs of dynamic monitoring. After the main arch is assembled, a single-layer steel mesh shell is connected to the main arch. The steel mesh shell adopts a welded ball joint system and is bolted to the reserved connecting plate on the main arch through the rods. After the connection is completed, the geometric shape measurement and initial stress state detection of the overall structure are carried out to provide reference data for the monitoring of the subsequent lifting process. The purpose of deploying distributed fiber optic strain sensors is to realize the full-process monitoring of the structural stress state and provide real-time data support for active control during the lifting process.

[0050] The specific implementation of step S03 involves activating the hydraulic synchronous lifting device to lift the integral structure composed of the main arch and the steel mesh shell. The lifting process employs a staged lifting method, with each lifting height controlled between 0.3 and 0.5 meters and the lifting speed set at 2 to 4 meters per minute. During the lifting process, distributed fiber optic strain sensors continuously collect strain data from each monitoring point of the structure. The data acquisition system converts the strain values ​​into stress values ​​and compares them with the design stress values ​​to calculate the stress deviation coefficient for each monitoring point. This coefficient is defined as the absolute value of the difference between the measured stress value and the design stress value divided by the design stress value. The maximum value of the deviation coefficient among all monitoring points is taken as the judgment criterion. Simultaneously, displacement sensors installed on the hydraulic jacks measure the actual displacement at each lifting point. The measured displacement is compared with the displacement predicted based on the finite element model to calculate the displacement deviation. The coefficient is defined as the absolute value of the difference between the measured displacement and the predicted displacement divided by the predicted displacement. The maximum value of the deviation coefficient among all lifting points is taken as the judgment criterion. When the stress deviation coefficient exceeds 0.15 or the displacement deviation coefficient exceeds 0.12, the system determines that the structure has uneven stress or asynchronous lifting. At this time, it is necessary to adjust the lifting speed of the corresponding hydraulic lifting point, reducing the lifting speed of the point to 60% to 75% of the original speed. By reducing the lifting speed, the adaptive adjustment time of the structure is increased, making the stress redistribution and displacement coordination process more stable. The changes in the adjusted stress deviation coefficient and displacement deviation coefficient are continuously monitored. When both deviation coefficients fall back to below 0.10, the normal lifting speed is restored. This control strategy is based on the feedback control principle and achieves the safety and synchronization of the lifting process through real-time monitoring and dynamic adjustment.

[0051] The specific implementation of step S04 involves installing the cable support system after the main arch is raised to the design elevation and fixed. The cable support system consists of multiple cables, one end of which is connected to a pre-reserved cable lug on the main arch structure, and the other end is anchored to the ground or a support column. Before installation, a total station and laser rangefinder are used to measure the initial length and installation angle of each cable. The initial length measurement accuracy is required to be ±5mm, and the installation angle measurement accuracy is required to be ±0.5 degrees. An influence matrix is ​​established to describe the influence of each cable's tension on the structural stress distribution and morphological accuracy. The influence matrix is ​​established by sequentially applying unit tension increments to each cable, such as 1... Finite element method (FEM) software was used to calculate the stress and displacement changes at key structural nodes caused by a unit force increment. The influence coefficients of all cables on all key nodes were arranged into a matrix, where rows represent different cables and columns represent different monitoring nodes. The matrix element values ​​represent the degree of influence of the unit tension of the cable on the node. The influence matrix describes the linear relationship between cable force adjustment and structural response, providing basic data for subsequent cable force optimization calculations. A two-layer game optimization model was used to calculate the cable force distribution scheme. This model consists of an upper-layer optimization model and a lower-layer optimization model forming a master-slave game system. The upper-layer optimization model, as the dominant party, makes decisions with the goal of maximizing structural stress uniformity, while the lower-layer optimization model, as the follower, responds based on the upper-layer decision with the goal of maximizing shape control accuracy. The two models exchange information and perform collaborative optimization through coupling terms. This game model can simultaneously consider both stress control and shape control objectives, achieving a more balanced optimization result compared to traditional single-objective optimization methods.

[0052] The specific implementation of step S05 involves inputting the maximum stress monitoring value, minimum stress monitoring value, and average stress monitoring value of the structure into the upper-level optimization model. These monitoring values ​​are acquired by distributed fiber optic strain sensors at key stress-bearing locations in the main arch. The uniformity of stress distribution is calculated using a stress uniformity evaluation function. The input parameters of this evaluation function include the maximum stress monitoring value, minimum stress monitoring value, average stress monitoring value, and a reference stress value, where the reference stress value is set to 100. As a normalization benchmark, the evaluation function calculates the ratio of maximum stress to reference stress, multiplies it by the ratio of minimum stress to reference stress, divides it by the square of the ratio of average stress to reference stress, and finally multiplies it by the upper-level cable force adjustment coefficient to obtain a dimensionless stress uniformity evaluation value. The larger the evaluation value, the more uniform the stress distribution. The solution process of the upper-level optimization model is to search for the upper-level cable force adjustment coefficient that maximizes the evaluation value under constraints. The constraints include that the upper-level cable force adjustment coefficient ranges from 0.5 to 1.5 and that the ratio of the maximum stress to reference stress does not exceed 3.0. The optimal upper-level cable force adjustment coefficient is obtained by numerical optimization algorithms such as gradient descent or genetic algorithms. This coefficient is then passed as a coupling term to the lower-level optimization model, realizing the influence and constraint of the upper-level decision on the lower-level decision. This process reflects the characteristic of the dominant party's priority decision in the game model.

[0053] The specific implementation of step S06 involves inputting the actual structural elevation measurement, allowable morphological deviation, and upper cable force adjustment coefficient into the lower-level optimization model. The actual structural elevation measurement is obtained by measuring the elevation difference between the arch crown and arch foot positions using a total station. The design elevation is set to 6m, and the allowable morphological deviation is set to 0.09m. The accuracy of structural morphological control is calculated using a morphological accuracy evaluation function. This function first calculates the normalized deviation by dividing the absolute value of the difference between the actual and design elevation measurements by the allowable morphological deviation. Then, it calculates the negative exponential function value of this normalized deviation to reflect the relationship that a smaller deviation results in a higher evaluation value. Finally, it multiplies this by the ratio of the upper-level cable force adjustment coefficient to the lower-level cable force adjustment coefficient to obtain a dimensionless morphological accuracy evaluation value. A larger evaluation value indicates higher morphological control accuracy. The solution process of the lower-level optimization model involves searching for the value that maximizes the evaluation value under constraints. The lower-level cable force adjustment coefficient is constrained by conditions including a value range of 0.5 to 1.5 and the absolute value of the difference between the actual and designed sag heights not exceeding the allowable value of morphological deviation. The optimal lower-level cable force adjustment coefficient is obtained through a numerical optimization algorithm. This coefficient is then fed back to the upper-level optimization model as a coupling term, realizing the influence of lower-level decisions on upper-level decisions. Iterative solutions are then implemented. In each iteration, the upper-level model recalculates the upper-level cable force adjustment coefficient based on the coefficient fed back from the lower level, and the lower-level model recalculates the lower-level cable force adjustment coefficient based on the coefficient passed from the upper level. When the changes in both the upper and lower-level cable force adjustment coefficients obtained from two consecutive iterations are less than 0.02, the iteration is considered converged. The two adjustment coefficients obtained at this point are the optimal solution. This iterative process reflects the characteristic of the master and slave sides in a two-level game model reaching Nash equilibrium through information exchange.

[0054] The specific implementation of step S07 involves calculating the target cable force value for each cable based on the upper and lower cable force adjustment coefficients obtained through iterative solving, combined with a pre-established influence matrix. The calculation method involves multiplying the two adjustment coefficients as weighting coefficients with the corresponding influence coefficients in the influence matrix. Taking into account stress uniformity and morphological accuracy requirements, the required tension force for each cable is determined. An intelligent tensioning system is used to apply tension to the cables. This system includes a tensioning jack, a pressure sensor, and a displacement sensor. Closed-loop control enables precise application of tension force. The pressure sensor monitors the oil pressure of the tensioning jack in real time and converts it into a tension force value. The displacement sensor monitors the cable elongation. The control system automatically adjusts the output pressure of the hydraulic pump station based on the sensor feedback data, ensuring the actual tension force approaches the target cable force value. The tension force control accuracy reaches the target cable force. The displacement control accuracy reaches ±2mm, with a value of ±3%. The tensioning process is divided into two stages: the initial tensioning stage and the fine tensioning stage. In the initial tensioning stage, 55% to 65% of the target cable force is applied to all cables sequentially. The purpose of this stage is to eliminate cable slack and allow the structure to take initial shape, avoiding excessive tension in a single cable that could lead to local stress concentration. In the fine tensioning stage, the remaining cable force is applied to all cables sequentially to bring the total tension force to the target cable force value. The purpose of this stage is to finely adjust the stress distribution and morphological accuracy of the structure. After each set of cables is tensioned, the three-dimensional coordinates of the key nodes of the structure are measured using a total station, and the stress state of the key nodes is measured using distributed fiber optic strain sensors. The measurement results are compared with the design values. If the deviation exceeds the allowable range, the tensioned cables need to be fine-tuned. This process ensures that the installation quality of the cable support system meets the design requirements.

[0055] The specific implementation of step S08 involves formulating a graded unloading plan for the temporary support system after the cable support system is tensioned. First, based on the structural stress characteristics, the temporary support system is divided into a main support zone and a secondary support zone. The main support zone is located at the mid-span and arch foot of the structure, where the stress is greater, and bears the main construction load. The secondary support zone is located in areas where the structural stress is relatively less, serving as auxiliary support. The unloading sequence follows the principle of unloading the secondary support zone first, then the main support zone. The entire unloading process is divided into 4 to 6 batches, with each batch unloading 15% to 20% of the design load. Before unloading begins, a pre-loading test is conducted to verify the bearing capacity of the temporary support system. The pre-loading test method involves applying 1.2 times the design load to the temporary support system for at least 24 hours. During this period, displacement sensors and strain sensors are used to monitor the deformation and stress state of the temporary support system. If both deformation and stress are within the allowable range, the temporary support system is deemed to meet the construction requirements. During the unloading process, the base... A graph-based path optimization algorithm plans the movement path of a crane. This algorithm uses the crane's starting and target positions as the start and end points of the graph, respectively. It employs shortest path search algorithms such as Dijkstra's algorithm or A* algorithm to calculate the shortest path after considering obstacles and restricted areas. This path comprehensively considers multi-objective optimization, including minimizing path length, the number of turns, and the operation time. A spatiotemporal occupancy matrix is ​​established to avoid equipment collisions. This matrix is ​​three-dimensional; the first two dimensions represent the spatial location of the construction site using a grid, and the third dimension represents time using time slices. Matrix elements record whether the corresponding spatiotemporal location is occupied by a particular piece of equipment. When planning the movement path of a new piece of equipment, it checks whether the equipment's spatiotemporal occupancy area overlaps with that of existing equipment. If an overlap is detected, a collision risk is identified, requiring adjustment of the equipment's operation time window or replanning of the movement path. This method ensures the safety and efficiency of multiple pieces of equipment working collaboratively within a limited space.

[0056] The specific implementation of step S09 involves immediately monitoring the vertical displacement change of the structure after each batch of unloading is completed. A total station or level is used to measure the elevation changes of key structural nodes. The elevation after unloading is compared with the elevation before unloading to obtain the vertical displacement change. This change is then compared with the design elevation. If the vertical displacement change exceeds 2.5% of the design elevation, the structure is deemed to have excessive deformation and pose a safety hazard. Unloading must be paused, and subsequent unloading plans must be reassessed. The assessment includes checking the working status of the cable support system, verifying the structural stress distribution, analyzing the causes of deformation, and adjusting cable forces or adding temporary supports if necessary. If the vertical displacement change is less than 2.5% of the design elevation, the structural deformation is considered to be within a controllable range, and the next batch of unloading continues. This monitoring and judgment process is repeated until the temporary support system is completely unloaded. This control strategy is based on the safety judgment criteria of deformation monitoring. By setting a clear deformation threshold, the unloading process is actively controlled to ensure that the structure remains in a safe state throughout the unloading process. After complete unloading, a comprehensive geometric measurement and stress state test are performed on the structure to verify whether the final forming quality meets the design requirements.

[0057] It should be noted that the key technical ideas of this invention include real-time monitoring and active control technology based on distributed fiber optic strain sensors. This technology, by deploying high-density strain sensors at key stress points of the main arch, achieves full-process monitoring of the structural stress state. Combined with feedback control principles, it dynamically adjusts the jacking speed of the hydraulic lifting points during the lifting process. Compared with traditional empirical control methods, this effectively avoids uneven stress distribution and asynchronous lifting, significantly improving the safety and reliability of the lifting process for large-span structures. Another key technical idea is the cable force distribution method based on a two-layer game optimization model. This method incorporates stress uniformity control and shape accuracy control into a unified optimization framework. Through a master-slave game mechanism, it achieves synergistic optimization of the two objectives. Compared with traditional single-objective optimization or simple weighted multi-objective optimization methods, it can better balance the needs of stress control and shape control, obtaining a more reasonable cable force distribution scheme and avoiding the problem of sacrificing one for the other. The third key technological approach is equipment collision avoidance technology based on a spatiotemporal occupancy matrix. This technology transforms the equipment movement path planning problem into a spatiotemporal resource allocation problem. By establishing a three-dimensional spatiotemporal occupancy matrix, it achieves coordinated control of the operation process of multiple pieces of equipment. Compared with traditional manual coordination methods, it can significantly reduce the risk of equipment collisions and improve construction efficiency. The synergistic effect of the above three technological approaches lies in establishing a complete technical system from structural monitoring to active control and construction organization. Monitoring technology provides real-time data support for control decisions, optimization models provide a scientific basis for control execution, and construction organization technology provides a reliable guarantee for control implementation. The three work together to form a closed-loop control system, which can significantly improve the overall quality and safety level of large-span arched steel roof construction compared with traditional open-loop control methods, realizing intelligent and refined management of the construction process.

[0058] It should be noted that this invention also solves the following technical problems: During the overall lifting of an arched, large-span steel roof shell, due to the large span and complex stress of the main arch structure, the lifting speed of multiple hydraulic lifting points is difficult to synchronize precisely. This leads to deviations in the structural stress distribution and displacement state from the design expectations during the lifting process, which can easily cause local stress concentration or instability risks. This invention achieves high-density real-time monitoring by deploying distributed fiber optic strain sensors at key stress-bearing parts of the main arch. During the lifting process, the stress deviation coefficient and displacement deviation coefficient of each monitoring point are calculated. When the deviation coefficient exceeds the threshold, the lifting speed of the corresponding hydraulic lifting point is immediately adjusted to 60% to 75% of the original speed. Through a dynamic feedback adjustment mechanism, the stress deviation and displacement deviation are controlled within a safe range, effectively avoiding the structural stress anomalies caused by insufficient synchronization in traditional constant-speed lifting methods. During the unloading of the temporary support system, the unloading operation changes the stress path and load distribution of the structure. An unreasonable unloading sequence or unloading rate can lead to sudden changes in vertical displacement of the structure or even trigger a chain of instability. This invention divides the temporary support system into a primary support area and a secondary support area by formulating a graded unloading scheme and unloads the secondary support area in the order of primary support area first and secondary support area second. The unloading volume of each batch is strictly controlled between 50% and 20% of the design load. Before unloading, the bearing capacity is verified by applying a pre-loading test with twice the design load. During the unloading process, a spatiotemporal occupancy matrix is ​​established to plan the crane movement path to avoid equipment collisions. After each batch of unloading is completed, the vertical displacement change is monitored and the subsequent unloading scheme is dynamically adjusted according to the monitoring results. This method ensures a smooth transition of the structural stress state through a gradual unloading strategy and a real-time monitoring feedback mechanism, and solves the technical problem that traditional rapid unloading methods are prone to structural instability.

[0059] Specifically, the principle of this invention is as follows: The fundamental reason why this invention can solve this technical problem lies in the fact that the master-slave decision-making mechanism and bidirectional coupling characteristics of the two-layer game optimization model conform to the multi-objective control logic of the tensioning process of the cable-supported system. During the cable force distribution process, the structural stress state is the primary constraint to ensure structural safety, while morphological accuracy is a secondary control objective under the premise of meeting stress requirements. This master-slave relationship precisely corresponds to the hierarchical structure in the two-layer game model where the upper layer prioritizes the decision-making of the lower layer, and the lower layer responds accordingly. By transmitting the upper-layer cable force adjustment coefficient to the lower-layer model as a coupling term, the lower-layer optimization must consider the decision-making results of the upper-layer stress control when pursuing morphological accuracy. Simultaneously, by feeding back the lower-layer cable force adjustment coefficient to the upper-layer model as a coupling term, the upper-layer optimization can perceive the response state of morphological control when pursuing stress uniformity. This bidirectional information exchange mechanism ensures the coordination and consistency between the two control objectives, rather than their independence or conflict. The iterative solution process allows the adjustment coefficients of the upper and lower layers to gradually converge to the optimal equilibrium solution through multiple games. This equilibrium solution can ensure the uniformity of the stress distribution of the structure to meet safety requirements, and can also control the shape deviation within the allowable range to meet accuracy requirements. Therefore, the technical solution of this invention can effectively solve the technical problem of stress control and shape control being difficult to balance during the tensioning process of cable-supported systems, both in theoretical logic and engineering practice.

[0060] 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.

[0061] The specific implementation of step S01 is as follows: First, wind-resistant columns are installed on the top of the irregular concrete beams at the construction site. The stability of the wind-resistant columns is ensured by connecting them with embedded parts. The spacing of the wind-resistant columns is determined according to the structural span, generally 6 to 8 meters. A hydraulic synchronous lifting device is fixed on the top of the wind-resistant columns. This device uses a hydraulic pump station to centrally control multiple hydraulic jacks to achieve synchronous lifting. The system is equipped with displacement sensors and pressure sensors to monitor the synchronicity of the lifting process in real time. Then, the construction site is measured and discretized according to a 3m×3m grid size. Each grid intersection is defined as a node and assigned a unique number. The three-dimensional coordinates of the nodes are determined by measurement. Lines are established between nodes according to the actual passage conditions to represent feasible paths. For locations with obstacles or restricted areas, no lines are established. The shortest path algorithm in graph theory, such as Dijkstra's algorithm, is used to calculate the optimal path from the component storage area to the installation position. This path comprehensively considers the goals of minimizing the path length and the number of turns. By establishing a node network model, basic data support is provided for the subsequent path planning of cranes and transportation equipment, ensuring the efficiency and safety of equipment movement during construction.

[0062] The specific implementation of step S02 is similar to that of step S01. The main arch structure is assembled on the ground on the assembly platform at the construction site. The main arch adopts a box-section beam design to provide sufficient bending and torsional stiffness. The assembly process is carried out segment by segment from the arch foot to the arch crown. Adjacent beam segments are connected by high-strength bolts. Distributed fiber optic strain sensors are deployed in the key stress-bearing parts of the main arch, including the arch foot area, the quarter-span area, and the arch crown area. The sensors adopt wavelength modulation fiber optic grating technology and reflect the strain state of the structure by measuring the center wavelength offset of the fiber Bragg grating. The spacing between the sensors does not exceed 2m to ensure that detailed changes in the stress distribution of the structure can be captured. The sensor measurement accuracy reaches ±1 micro-strain, and the sampling frequency is set to not less than 10Hz to meet the needs of dynamic monitoring.

[0063] The specific implementation of step S03 involves activating the hydraulic synchronous lifting device to lift the overall structure composed of the main arch and the steel mesh shell. The lifting process adopts a staged lifting method, with each lifting height controlled between 0.3 and 0.5 meters and the lifting speed set at 2 to 4 meters per minute. During the lifting process, distributed fiber optic strain sensors continuously collect strain data from various monitoring points of the structure. The data acquisition system converts the strain values ​​into stress values ​​and compares them with the design stress values. The formula for calculating the stress deviation coefficient is as follows:

[0064] ;

[0065] In the formula, The stress deviation coefficient is dimensionless. For the first The measured stress values ​​at each stress monitoring point, in units of The values ​​were obtained by distributed fiber optic strain sensors. For the first The design stress value at each stress monitoring point, in units of , obtained by finite element analysis; This represents the total number of stress monitoring points. Number the stress monitoring points; This indicates that the maximum value of the deviation coefficient is taken from all monitoring points.

[0066] The formula for calculating the displacement deviation coefficient is as follows:

[0067] ;

[0068] In the formula, The displacement deviation coefficient is dimensionless. For the first The measured displacement of each hydraulic lifting point, in units of... The displacement is measured by a displacement sensor installed on a hydraulic jack; For the first The displacement of each hydraulic lifting point is predicted based on the finite element model, in units of... , obtained by finite element analysis; This represents the total number of hydraulic lifting points. Number the hydraulic lifting points; This indicates taking the maximum value of the deviation coefficient among all lifting points.

[0069] When the stress deviation coefficient Exceeding 0.15 or displacement deviation coefficient When the value exceeds 0.12, the system determines that the structure is experiencing uneven stress or asynchronous lifting. In this case, it is necessary to adjust the lifting speed of the corresponding hydraulic lifting point. The formula for calculating the adjusted lifting speed is as follows:

[0070] ;

[0071] In the formula, The adjusted jacking speed is expressed in units of... ; This is the speed adjustment coefficient, ranging from 0.60 to 0.75. It is dimensionless and is determined based on the degree of deviation; the greater the deviation, the smaller the coefficient value. The original lifting speed, in units of The empirical value is 2 to 4 meters per minute.

[0072] The specific implementation of step S04 involves installing the cable support system after the main arch is raised to the design elevation and fixed. The cable support system consists of multiple cables, one end of which is connected to a pre-reserved cable lug on the main arch structure, and the other end is anchored to the ground or a support column. Before installation, a total station and a laser rangefinder are used to measure the initial length and installation angle of each cable. The initial length measurement accuracy is required to be ±5mm, and the installation angle measurement accuracy is required to be ±0.5 degrees. An influence matrix is ​​established to describe the influence of the tension of each cable on the stress distribution and morphological accuracy of the structure. The method for establishing the influence matrix is ​​to apply a unit tension increment to each cable sequentially. The experience value is 1. The stress and displacement changes at key structural nodes caused by the unit force increment were calculated using finite element software. The formula for calculating the stress influence coefficient is as follows:

[0073] ;

[0074] In the formula, For the first The unit tension of the root cable relative to the first The stress influence coefficient of each structural monitoring node, in units of ; Number the cables; Number the structural monitoring nodes; For the first The stress change at each structural monitoring node, in units of , obtained by calculation using finite element software; The increment of unit tension force is 1, based on empirical values. .

[0075] The formula for calculating the displacement influence coefficient is as follows:

[0076] ;

[0077] In the formula, For the first The unit tension of the root cable relative to the first Displacement influence coefficient of each structural monitoring node, in units of ; For the first The displacement change of each structural monitoring node, in units of The result was obtained by calculation using finite element software.

[0078] The formula for the stress influence matrix is ​​as follows:

[0079] ;

[0080] In the formula, This is the stress influence matrix; The total number of cables; This represents the total number of structural monitoring nodes.

[0081] The formula for the displacement influence matrix is ​​as follows:

[0082] ;

[0083] In the formula, This is the displacement influence matrix.

[0084] The specific implementation of step S05 involves inputting the maximum stress monitoring value, minimum stress monitoring value, and average stress monitoring value of the structure into the upper-level optimization model. These monitoring values ​​are acquired by distributed fiber optic strain sensors at key stress-bearing locations in the main arch. The uniformity of stress distribution is calculated using a stress uniformity evaluation function, the formula of which is expressed as follows:

[0085] ;

[0086] In the formula, This is a dimensionless evaluation value for stress uniformity. This represents the maximum stress monitoring value of the structure, in units of... It is determined by the maximum value among all the stress values ​​collected by the distributed fiber optic strain sensors at all monitoring points; This represents the minimum stress monitoring value for the structure, in units of... It is determined by the minimum stress value among all monitoring points collected by the distributed fiber optic strain sensor; The average stress monitoring value of the structure is expressed in units of 1. It is determined by the arithmetic mean of the stress values ​​at all monitoring points collected by the distributed fiber optic strain sensor; For reference stress value, take 100. , used for dimensionless normalization; This is the adjustment coefficient for the upper cable force, which is dimensionless.

[0087] The constraints of the upper-level optimization model are:

[0088] and ;

[0089] The upper-level optimization model maximizes The optimal upper cable force adjustment coefficient is obtained by solving the problem. This coefficient is passed to the lower-level optimization model as a coupling term.

[0090] The specific implementation of step S06 involves inputting the actual structural rise measurement value, the allowable value of morphological deviation, and the upper cable force adjustment coefficient into the lower-level optimization model. The actual structural rise measurement value is obtained by measuring the elevation difference between the arch crown and the arch foot positions using a total station. The design rise is... Take 6m, allowable value for morphological deviation Taking a value of 0.09m, the accuracy of structural morphology control is calculated using a morphological accuracy evaluation function, the formula of which is as follows:

[0091] ;

[0092] In the formula, This is a dimensionless morphological accuracy evaluation value; This is the actual measured elevation of the structure, in units of... The elevation difference between the position of the arch crown and the position of the arch foot is obtained by measuring the elevation difference using a total station. The design elevation is set to 6m; The allowable value for morphological deviation is taken as 0.09m; This is the adjustment coefficient for the upper cable force, which is dimensionless. This is the adjustment coefficient for the lower cable force, which is dimensionless. This represents the natural exponential function.

[0093] The constraints of the lower-level optimization model are:

[0094] and ;

[0095] The lower-level optimization model maximizes The optimal lower cable force adjustment coefficient is obtained by solving the problem. This coefficient is fed back as a coupling term to the upper-level optimization model for iterative solution. The convergence condition for the iteration is:

[0096] and ;

[0097] In the formula, For the first The upper cable force adjustment coefficient for the next iteration; For the first The upper cable force adjustment coefficient for the next iteration; For the first The lower layer cable force adjustment coefficient for the next iteration; For the first The lower layer cable force adjustment coefficient for the next iteration; This represents the number of iterations.

[0098] The specific implementation of step S07 is to calculate the target cable force value of each cable based on the upper and lower cable force adjustment coefficients obtained by iterative solution, combined with the pre-established influence matrix. First, the reference cable force value is calculated. This value was obtained through finite element analysis, and the calculation formula is as follows:

[0099] ;

[0100] In the formula, For the first The reference cable force value of the root cable, in units of ; For the first Genlaso to the first The stress influence coefficient of each structural monitoring node, in units of ; For the first The target stress value of each structural monitoring node, in units of As specified in the design documents; As the reference value for stress normalization, it is set to 100. ; For the first Genlaso to the first Displacement influence coefficient of each structural monitoring node, in units of ; For the first The target displacement value of each structural monitoring node, in units of As specified in the design documents; As the reference value for displacement normalization, it is set to 100. ; The reference value for normalized cable force is set to 100. ; This represents the total number of structural monitoring nodes.

[0101] The formula for calculating the target cable force value is expressed as follows:

[0102] ;

[0103] In the formula, For the first The target cable force value of the root cable, in units of ; This is the stress uniformity weighting coefficient, usually taken as 0.6, dimensionless. This weighting coefficient is determined by engineering experience and reflects the importance of stress uniformity in the optimization objective. This is the morphological accuracy weighting coefficient, typically set to 0.4. It is dimensionless and determined by engineering experience, reflecting the importance of morphological accuracy in the optimization objective. ; This is the adjustment coefficient for the upper cable force; This is the adjustment coefficient for the lower layer cable force.

[0104] The tensioning process is divided into the initial tensioning stage and the fine tensioning stage. The formula for calculating the cable force applied in the initial tensioning stage is as follows:

[0105] ;

[0106] In the formula, For the first The cable force applied during the initial tensioning stage of the root cable, in units of ; The initial tension factor ranges from 0.55 to 0.65 and is dimensionless. This factor is determined by construction experience to avoid excessive tension in a single operation, which could lead to localized stress concentration in the structure.

[0107] The formula for calculating the cable force applied during the fine tensioning stage is as follows:

[0108] ;

[0109] In the formula, For the first The cable force applied during the fine tensioning stage of the root cable, in units of .

[0110] The specific implementation of steps S08 and S09 involves formulating a graded unloading plan for the temporary support system after the cable support system is tensioned. First, based on the structural stress characteristics, the temporary support system is divided into a main support zone and a secondary support zone. The main support zone is located at the mid-span and arch foot of the structure, where the stress is greater, and bears the main construction load. The secondary support zone is located in areas where the structural stress is relatively smaller, serving as auxiliary support. The unloading sequence follows the principle of unloading the secondary support zone first, then the main support zone. The entire unloading process is divided into 4 to 6 batches, with each batch unloading 15% to 20% of the design load. Before unloading begins, a pre-loading test is conducted to verify the bearing capacity of the temporary support system. The pre-loading test method involves applying 1.2 times the design load to the temporary support system for at least 24 hours. During this period, displacement sensors and strain sensors are used to monitor the deformation and stress state of the temporary support system. Immediately after each batch of unloading is completed, the vertical displacement change of the structure is monitored. A total station or level is used to measure the elevation changes of key structural nodes. The formula for calculating the vertical displacement change is as follows:

[0111] ;

[0112] In the formula, For the first The vertical displacement change of each key structural node, in units of ; Number the key nodes of the structure; For the first Elevation of each key structural node after unloading, in units of , obtained by total station or level; For the first Elevation of each key structural node before unloading, in units of , obtained by total station or level.

[0113] The formula for calculating the ratio of vertical displacement change to design elevation is as follows:

[0114] ;

[0115] In the formula, For the first The vertical displacement rate of change of each key structural node, dimensionless; The design elevation is set to 6m.

[0116] When the rate of change of vertical displacement of any critical structural node exceeds a threshold, the structural deformation is deemed excessive. The judgment condition is expressed by the following formula:

[0117] ;

[0118] In the formula, The maximum value of the vertical displacement rate of change among all critical structural nodes, dimensionless; 0.025 represents the total number of critical structural nodes; 0.025 represents the vertical displacement rate threshold, which is 2.5% of the design elevation.

[0119] when When this happens, the uninstallation process is paused and subsequent uninstallation plans are reassessed; when If necessary, continue with the next batch of uninstallation until the temporary support system is completely uninstalled.

[0120] The proposed two-layer game-theoretic optimization model achieves a multi-objective balance in cable force distribution by synergistically optimizing the upper-layer model's focus on stress uniformity and the lower-layer model's focus on morphological accuracy. The stress uniformity evaluation function... Dividing the product of the maximum and minimum stresses by the square of the average stress can effectively punish extreme non-uniform stress distribution. When the stress distribution is uniform... , and When the stress distribution is uneven, the difference between the maximum and minimum stresses increases, and the evaluation function value decreases. This affects the morphological accuracy evaluation function. The exponential function approach makes the influence of shape deviation on the evaluation value non-linear. When the absolute value of the difference between the actual sag and the design sag approaches zero, the exponential term... A value close to 1 indicates good control performance; as the deviation increases, the exponential term rapidly decays, indicating a deterioration in control performance. The two-layer model adjusts the coefficients... and Information exchange is achieved through coupling and transmission; the upper-level model will... The impact on morphological control decisions is transmitted to the lower-level model, and the lower-level model will... Feedback to the upper-level model influences stress uniformity decisions. The iterative solution process enables the two objectives to reach Nash equilibrium through multiple games. The final cable force distribution scheme not only ensures the uniformity of structural stress distribution but also meets the accuracy requirements of morphological control. Compared with traditional single-objective optimization methods, it significantly improves the overall quality and safety of large-span shell steel roof construction.

[0121] It should be noted that the variables involved in this invention are explained in detail in Table 1.

[0122] Table 1. Variable Explanation Table

[0123]

[0124] To better understand and implement this invention, the following is a specific application scenario of the invention, Example 2: The main arch of the arched steel roof of a large airport terminal building adopts a box-section beam with specifications of B1200×700×34, made of Q460GJ high-strength steel. The single-layer steel grid shell main beam has specifications of H400×250×8×16, and the diagonal brace is P159×10. The overall structure is supported by 26 wind-resistant columns with dimensions of B450×300×20. The lower foundation is an irregularly shaped concrete beam. The total weight of the roof is approximately 3120t. Figure 2 The schematic diagram of the arched, large-span shell steel roof is shown. This project faced technical challenges such as construction restrictions within the terminal's operational area, complex transportation routes for large components, stringent requirements for overall lifting deformation control, and significant difficulties in optimizing cable tensioning. The technical team decided to adopt the construction method of this invention for construction organization.

[0125] In the early stages of construction, the technical team installed 26 wind-resistant columns above the irregularly shaped concrete beams. Each column was topped with a hydraulic synchronous lifting device, with a rated lifting capacity of 145t, a stroke of 8m, and an adjustable lifting speed ranging from 0.5 to 12m / min. The construction site was discretized using a 3m×3m grid, creating a network model with 3168 nodes for subsequent crane path planning and equipment scheduling management. The team deployed two QTZ160 tower cranes and three 30t truck cranes at the construction site. The optimal transportation route from the component storage area to the assembly area was calculated using the node network model. The route planning comprehensively considered three factors: path length, number of turns, and operation time, avoiding the terminal's operational areas and passenger passageways.

[0126] During the ground assembly phase, the technical team divided the main arch structure into 11 assembly units, each 12m in length. The main arch was assembled using a combination of factory prefabrication and on-site assembly. During assembly, 168 distributed fiber optic strain sensors were deployed at key stress points on the main arch. These sensors, employing wavelength-modulated fiber optic gratings, achieved a measurement accuracy of ±1 micro-strain, with a sampling frequency set at 12Hz and sensor spacing controlled within the range of 1.8–2.0m. After the main arch was assembled, a single-layer steel grid shell structure was installed. The grid shell nodes used welded spherical joints, totaling 512 nodes, completing the rigid connection between the main arch and the grid shell, forming a complete roof structure system.

[0127] During the overall lifting phase, the technical team activated 26 sets of hydraulic synchronous lifting devices, setting the initial lifting speed to 5.5 m / min. Data was collected in real-time from 168 fiber optic sensors at 0.083 s intervals. When the lifting reached a height of 2.8 m, the monitoring system showed that the stress deviation coefficient at the mid-span of the main arch reached 0.168, exceeding the control threshold of 0.15. Simultaneously, the displacement deviation coefficient at the north arch foot reached 0.142, exceeding the control threshold of 0.12. The technical team immediately adjusted the lifting speed of the five hydraulic lifting points in the corresponding areas, reducing the speed from 5.5 m / min to 3.85 m / min, a reduction of 70%. After 9 minutes of adjustment, the stress deviation coefficient decreased to 0.088, and the displacement deviation coefficient decreased to 0.082, both meeting the control requirement of below 0.10. During the continued lifting process, the technical team triggered another warning at a height of 4.2m, indicating that the displacement deviation coefficient in the southern area of ​​the main arch reached 0.135. The lifting speed at the four lifting points in this area was adjusted to 3.58m / min, a reduction of 65%. After 7 minutes of adjustment, the deviation coefficient returned to 0.091. The entire lifting process lasted 8.2 hours, and the main arch was successfully lifted to the design elevation of 6m. A total of 19 adjustments were made to the hydraulic lifting speed during the process, effectively controlling structural deformation. The measurement data after the lifting was completed is shown in Table 2.

[0128] Table 2. Measurement data of key nodes after the main arch lifting is completed.

[0129]

[0130] During the cable-stayed system installation phase, the technical team deployed 36 cables on both sides of the main arch. The cables are 7×139 steel strands, with a single cable designed tension of 850kN. Cable lengths range from 16 to 36 meters, and installation angles range from 32° to 62°. The team used a total station to measure the initial length and installation angle of each cable, establishing a 36×36 influence matrix to describe the relationship between cable tension and structural response. During the influence matrix establishment process, a unit tension increment of 50kN was applied to each cable, and the stress and displacement changes at key nodes of the main arch were measured. The complete influence coefficient matrix was then calculated using finite element analysis.

[0131] During the cable stress optimization calculation phase, the technical team employed a two-layer game theory optimization model to calculate the cable stress distribution scheme. The upper-layer optimization model aims to maximize the uniformity of structural stress, inputting the maximum stress monitoring value of the structure. =252.3MPa, minimum structural stress monitoring value =142.8MPa and structural average stress monitoring value =189.6MPa, reference stress value =100MPa, calculate the stress uniformity evaluation function. The initial upper cable force adjustment coefficient is obtained through optimization. =1.18. The lower-level optimization model aims to maximize the accuracy of morphological control, and the input is the actual measured value of the structural sag. =6.014m, Design elevation =6m and allowable morphological deviation =0.09m, calculation of morphological accuracy evaluation function The initial lower cable force adjustment coefficient is obtained by solving the problem. =1.06.

[0132] The technical team performed iterative optimization calculations, and after the first iteration... Adjusted to 1.14, Adjusting to 1.10, with changes of 0.04 and 0.04 respectively, did not meet the convergence condition. After the second iteration... =1.12, =1.12, with changes of 0.02 and 0.02 respectively, failing to meet the convergence condition. After the 3rd iteration =1.11, =1.13, with changes of 0.01 and 0.01 respectively, satisfying the convergence condition of less than 0.02, and the iteration terminates. Based on the final solution... =1.11 and =1.13, and the target cable force value of each cable was calculated based on the influence matrix. The target cable force value of the 36 cables ranges from 768 to 936 kN. The data of the iterative calculation process are shown in Table 3.

[0133] Table 3 Iterative Calculation Process of the Two-Level Game Optimization Model

[0134]

[0135] During the cable tensioning phase, the technical team employed an intelligent tensioning system to tension the 36 cables. The system included tensioning jacks, pressure sensors, and displacement sensors, using closed-loop control to precisely apply the tension force. The tensioning process was divided into two stages: initial tensioning and fine tensioning. In the initial tensioning stage, 60% of the target cable force was applied, i.e., 461–562 kN of tension per cable. After the initial tensioning was completed, the three-dimensional coordinates and stress state of key nodes in the main arch were measured. In the fine tensioning stage, the remaining 40% of the cable force was applied using a grouped symmetrical tensioning method, with four cables in each group tensioned simultaneously. The tensioning sequence progressed from the mid-span towards the arch foot, and the fine tensioning time for each cable was controlled within 18 minutes. After completing the tensioning of one group of cables, a total station was used to measure the coordinates of key structural nodes, and distributed fiber optic sensors were used to monitor the stress state, ensuring that structural deformation and stress remained within controllable ranges during the tensioning process. After all 36 cables were tensioned, the maximum stress of the structure was 225.8 MPa, the minimum stress was 158.4 MPa, and the average stress was 192.3 MPa. The actual rise of the arch was 5.996 m, with a deviation of only 4 mm from the design rise, meeting the shape control requirements. Some cable tensioning data are shown in Table 4.

[0136] Table 4. Partial Cable Tensioning Data

[0137]

[0138] During the unloading phase of the temporary support system, the technical team developed a tiered unloading plan, dividing the 26 wind-resistant columns into a main support zone and a secondary support zone. The main support zone consisted of 14 wind-resistant columns: 4 at the mid-span and 5 at each of the two arch foot sections. The secondary support zone comprised the remaining 12 wind-resistant columns. Before unloading, a pre-loading test was conducted on the temporary support system with a load 1.2 times the design load. The design load was 3120 kN, and the pre-loading load was 3744 kN, lasting for 28 hours. Monitoring data showed that the maximum deformation of the temporary support system was 9.3 mm, and the maximum stress was 183.5 MPa, both meeting the load-bearing capacity requirements. The unloading process was carried out in 5 batches, with each batch unloading 18% of the design load, or 562 kN. The first batch unloaded the 4 wind-resistant columns on the outer side of the secondary support zone. After unloading, the monitored vertical displacement change was 10.5 mm, accounting for 0.18% of the design rise of 6 m, far less than the control threshold of 2.5%. The second batch unloaded the four middle wind-resistant columns in the secondary support area, with a vertical displacement change of 13.2 mm, accounting for 0.22% of the design elevation. The third batch unloaded the remaining four wind-resistant columns in the secondary support area, with a vertical displacement change of 15.8 mm, accounting for 0.26% of the design elevation. The fourth batch unloaded six non-critical wind-resistant columns in the main support area, with a vertical displacement change of 20.4 mm, accounting for 0.34% of the design elevation. The fifth batch unloaded the last eight wind-resistant columns in the main support area, with a vertical displacement change of 24.6 mm, accounting for 0.41% of the design elevation. The vertical displacement changes in all batches met the control requirements.

[0139] During the unloading process, the technical team employed a graph-based path optimization algorithm to plan the crane movement paths. The starting and target positions of the two tower cranes and three truck cranes were used as nodes in the graph. Dijkstra's algorithm was used to calculate the shortest path after considering obstacles, restricted areas, and terminal restrictions. A spatiotemporal occupancy matrix for the construction site was established, with dimensions of 102×88×24. The first two dimensions represent the spatial location of the construction site after being divided into 3m×3m grids, and the third dimension represents time, with a time step of 1 hour. Matrix elements record the equipment occupancy status at the corresponding spatiotemporal location. During the fourth batch of unloading, the system detected a spatiotemporal overlap between the operating area of ​​tower crane QTZ160-01 at time T=14h and the movement path of the truck cranes, posing a collision risk. The technical team adjusted the truck crane's operating time window, extending the originally scheduled T=14h operation time to T=15h, thus eliminating the collision risk. The entire unloading process took 15 days, involving 9 adjustments to equipment operation windows and 6 replanning of crane movement paths, ensuring the unloading process was safe and efficient without affecting the normal operation of the terminal. Monitoring data for each batch of unloading is shown in Table 5.

[0140] Table 5. Monitoring data on the graded unloading of temporary support systems

[0141]

[0142] After the temporary support system was completely unloaded, the technical team conducted a comprehensive inspection of the structure. The final elevation of the arch was 5.994m, a deviation of 6mm from the design elevation of 6m. The maximum stress was 229.6MPa, the minimum stress was 155.2MPa, and the average stress was 193.8MPa, with good stress distribution uniformity. The actual tension of the 36 cables deviated from the design tension within ±2%, indicating that the cable support system was functioning well. The main arch weld quality inspection pass rate reached 100%, and the grid shell node connection quality inspection pass rate reached 98.4%. The overall structure met the design and specification requirements, and the construction of the terminal roof was successfully completed.

[0143] This invention represents a significant advancement over traditional methods for constructing large-span steel structures. Regarding deformation control, traditional methods, relying on limited monitoring points, struggle to fully grasp the structural stress state, easily leading to localized stress concentrations and exceeding deformation limits. This invention, however, employs distributed fiber optic strain sensors for full-process stress monitoring, combined with dual control standards of stress deviation coefficient and displacement deviation coefficient. This ensures vertical displacement deviation during the lifting process is controlled within 0.23% of the design sagitta, an improvement of approximately 12% compared to the traditional method's 0.48% deviation control accuracy. In cable tension optimization, traditional methods often rely on single-objective optimization or empirical adjustments, failing to simultaneously address stress uniformity and morphological accuracy. This invention utilizes a two-layer game-theoretic optimization model to achieve synergistic optimization of stress and morphological control. After cable tensioning, the structural stress distribution uniformity evaluation value reaches 1.068, an improvement of approximately 13% compared to the traditional method's 0.945, and the morphological accuracy evaluation value reaches 0.912, an improvement of approximately 14% compared to the traditional method's 0.798. In terms of construction safety management, traditional methods rely on manual experience to plan equipment paths, which easily leads to equipment collisions and operational conflicts. This invention, however, employs a graph-based path optimization algorithm and a spatiotemporal occupancy matrix to achieve intelligent equipment scheduling. The equipment collision warning response time during construction is reduced to 32% of that of traditional methods, and equipment operating efficiency is improved by approximately 17%, effectively ensuring construction safety and progress. Overall, this invention, through the comprehensive application of distributed monitoring, two-layer game theory optimization, and intelligent scheduling technologies, significantly improves the accuracy, efficiency, and safety of arched large-span shell steel roof construction, providing an effective technical solution for the construction of large-span spatial structures such as airport terminals.

[0144] 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 constructing an arched, large-span shell steel roof, characterized in that, A temporary support system was constructed, and wind-resistant columns and a hydraulic synchronous lifting device were installed above the irregularly shaped concrete beam. The main arch structure was assembled on the ground, and distributed fiber optic strain sensors were deployed at key stress points. The hydraulic synchronous lifting device was activated for overall lifting, and strain data was collected in real time to adjust the lifting speed. After lifting, the cable support system was installed, and an influence matrix was established. A two-layer game optimization model was used to calculate the cable force distribution scheme. The two-layer game optimization model outputs the upper-layer cable force adjustment coefficient with the goal of maximizing stress uniformity and transmits it to the lower-layer optimization model. The lower-layer optimization model outputs the lower-layer cable force adjustment coefficient with the goal of maximizing morphological accuracy and feeds it back to the upper-layer optimization model. Information exchange was achieved through coupling terms, and iterative solutions were used to achieve dynamic equilibrium between the two control objectives under mutual constraints. The target cable force value was calculated based on the cable force adjustment coefficient obtained from the iterative solution and the influence matrix. A smart tensioning system was used to apply tension force in stages, and a graded unloading scheme for the temporary support system was formulated to monitor the structural displacement status.

2. The method according to claim 1, characterized in that, In the process of setting up a temporary support system, a node network model is established at the construction site for crane path planning.

3. The method according to claim 2, characterized in that, The node network model refers to discretizing the construction site into a 3m×3m grid, with each grid intersection serving as a node, and the lines connecting the nodes representing the travel paths of cranes and transportation equipment.

4. The method according to claim 3, characterized in that, In the ground assembly of the main arch structure, the main arch adopts a box-section beam to complete the connection between the main arch and the single-layer steel grid shell.

5. The method according to claim 4, characterized in that, The distributed optical fiber strain sensor adopts a wavelength-modulated fiber optic grating sensor with a sensor spacing of ≤2m, a measurement accuracy of ±1 microstrain, and a sampling frequency of ≥10Hz.

6. The method according to claim 5, characterized in that, During the overall lifting process, the stress deviation coefficient and displacement deviation coefficient of each monitoring point are calculated. When the stress deviation coefficient is greater than 0.15 or the displacement deviation coefficient is greater than 0.12, the lifting speed of the corresponding hydraulic lifting point is adjusted.

7. The method according to claim 6, characterized in that, The stress deviation coefficient is calculated as the maximum value obtained by dividing the absolute value of the difference between the measured stress value and the design stress value at each monitoring point by the design stress value.

8. The method according to claim 7, characterized in that, The displacement deviation coefficient is calculated by dividing the absolute value of the difference between the measured displacement and the predicted displacement at each monitoring point by the predicted displacement, resulting in the maximum value.

9. The method according to claim 8, characterized in that, When adjusting the jacking speed, reduce the jacking speed to 60-75% of the original speed until both the stress deviation coefficient and the displacement deviation coefficient are <0.

10.

10. The method according to claim 9, characterized in that, The influence matrix is ​​established by applying a unit tension increment to each cable, calculating the stress and displacement changes at key structural nodes, and arranging the influence coefficients of all cables in rows and columns to form a matrix.

Images