Steel structure lifting construction path intelligent planning method based on AI

Abstract

The invention relates to an AI-based intelligent planning method for a steel structure lifting construction path, which belongs to the technical field of steel structure building construction and comprises the following operation steps of: 1, establishing a net rack lifting database of a steel structure; 2, abstracting a steel structure weight map according to the grid coordinates and the weight; and 3, based on the factors of the lower structure, the shape of the net rack and the height difference of the net rack, combining a steel structure weight map, and carrying out total-factor steel structure intelligent partitioning. And 4, combining the AI model, the steel structure lifting database and an intelligent algorithm, fusing the existing construction data, and establishing an AI database to intelligently plan a steel structure lifting construction path. The AI model, the steel structure lifting database and the intelligent algorithm are combined, existing construction data are considered, and steel structure partitions and lifting construction path planning are intelligently determined. The problems that an existing steel structure partitioning and steel structure lifting construction path mainly depends on experience, scheme design is slow, and economical efficiency is not high are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of steel structure building construction, and in particular to an AI-based intelligent planning method for steel structure lifting construction paths. Background Art

[0002] The stress state of steel structures during the construction phase will be affected by many uncertain factors, and the situation is more complex and changeable. Improper setting of lifting points during lifting, sudden changes in internal forces of structural members during lifting, turning over and installing in place, and asynchronous lowering of jacks during temporary structure unloading will all lead to non-uniform deformation of the overall structure. The resulting redistribution of internal forces will cause sudden changes in internal forces of some members or collapse of temporary structures due to excessive loads. In addition, asynchrony during traction and sliding will also lead to uneven traction of the structure and deviation of the structure. An unsteady traction process will quickly cause the structure to vibrate under dynamic loads. Unsmooth tracks and mechanical problems will cause changes in the internal forces of the structural members, which may in turn cause damage to the structural components.

[0003] The invention patent "A method for analyzing the lifting and unloading construction of complex large-span spatial steel structures" has an application number of CN202211195652.6. It discloses a full-process simulation and determination of the lifting, unloading and monitoring construction plan for complex large-span spatial steel structures based on theoretical analysis and numerical simulation. It comprehensively considers the influence of component segmentation, lifting sequence and temperature stress during the lifting construction process, and also considers the influence of unloading sequence and unloading control method during the unloading construction process. At the same time, a reasonable and effective monitoring plan is formulated based on the complexity and uncertainty of actual construction. It can effectively improve the construction efficiency of subsequent complex large-span spatial steel structures, ensure construction safety, and solve the safety hazard problem caused by the large-span spatial steel structure due to the complexity of the structural form, construction technology and process. The stress and deformation of components in the construction process are often quite different from the design state.

[0004] Steel structure partitioning and construction path selection are crucial in steel structure construction. It is to divide the complex steel structure into multiple partitions according to factors such as steel structure weight, shape and size to facilitate construction. In the actual construction process, reasonable steel structure partitioning can effectively reduce the difficulty of construction and improve construction efficiency. Through the analysis of the steel structure, the boundaries and construction paths of each partition are determined to ensure that each partition can maintain the structural form and control costs during independent construction, thereby improving the overall project quality.

[0005] The existing steel structure zoning and construction path selection mainly rely on the experience of construction workers. The steel structure is preliminarily divided according to the shape symmetry and structural characteristics, and then the construction path is determined based on construction experience. Summary of the invention

[0006] The present invention mainly solves the deficiencies in the prior art and provides an AI-based intelligent planning method for steel structure lifting construction paths. It intelligently partitions the steel structure by abstracting the weight topographic map of the steel structure, and considers factors such as the lower structure, the shape of the steel structure, and the height difference of the steel structure. Combined with the weight topographic map of the steel structure, the steel structure partitioning method is corrected with all factors to solve the problem that the existing steel structure partitioning mainly relies on experience. By combining the AI ​​model, the steel structure lifting database and the intelligent algorithm, the existing construction data is considered to intelligently determine the steel structure partition and plan the lifting construction path. This solves the problem that the existing steel structure lifting construction path mainly relies on experience, the scheme design is slow, and the economy is not high.

[0007] The above technical problems of the present invention are mainly solved by the following technical solutions: An AI-based intelligent planning method for steel structure lifting construction path includes the following steps: Step 1: Establish a steel structure grid lifting database.

[0008] Step 2: Make an abstract steel structure weight topography map based on the grid coordinates and weight.

[0009] Step 3: Based on the factors of the lower structure, grid shape and grid height difference, combined with the steel structure weight topography, carry out intelligent zoning of the full-factor steel structure.

[0010] Step 4: Utilize the combination of AI models, steel structure improvement database and intelligent algorithms, and integrate existing construction data to establish an AI database to intelligently plan the steel structure improvement construction path.

[0011] Preferably, the grid lifting database is based on the existing steel structure engineering construction data to establish a steel structure lifting database, including data on the size of the lifted steel structure, the weight of the lifted steel structure, the number of steel structure lifting points, the lifting deformation, the stress ratio and the steel structure lifting construction cost.

[0012] Preferably, the abstract steel structure weight topographic map uses the steel structure cell as the basic unit to store the number and weight of each rod. The number and weight of each rod are stored by using cell 1 labeled D1, and the labels of the rods constituting cell 1 are D1-1, D1-2, D1-3, and the weights are M1-1, M1-2, and M1-3; the adjacent cell 2 is labeled D2, and the labels of the rods constituting cell 2 are D2-1, D2-2, and D2-3, and the weights are M2-1, M2-2, and M2-3; the same applies to the remaining cells; the X and Y coordinate points of the upper chord nodes of the steel structure are extracted to form a plane topographic map of the steel structure, and the closed figure surrounded by the upper chord rods is used as the topographic map unit, and the weight of the cell within the projection area of ​​the topographic map unit is assigned to the corresponding topographic map unit to form a steel structure weight topographic map.

[0013] As a preferred method, the influence of the lower structure on the weight map of the steel structure is corrected. If there are supporting columns or modeling columns at the bottom of the steel structure, the weight of the supporting columns or modeling columns is evenly superimposed on the corresponding topographic map range unit; the influence of the shape on the weight map of the steel structure is corrected. For steel structures with strange and irregular shapes, the topmost, bottommost, leftmost and rightmost protruding points of the ring steel structure are traversed, and each protruding point is used as a reference to form a closed rectangle containing the outer shape of the steel structure. The closed figure formed by the edge of the rectangle and the edge of the steel structure is divided into several topographic map units with a side length of 1 unit. Each topographic map unit is named E1, E2, and E3 in sequence, and is given a weight of 0, which is called a zero-weight topographic map cell.

[0014] As a preferred method, the center of the rectangular steel structure weight topographic map is used as the origin to form an outward-expanding quadrilateral similar to the rectangular steel structure weight topographic map. The outward-expanding quadrilateral stops expanding when there is a zero-weight topographic map cell in the outward-expanding quadrilateral, and detects whether the vertex of the outward-expanding quadrilateral is located in the zero-weight topographic map cell. If it is located in the zero-weight topographic map cell, the outward-expanding quadrilateral stops expanding; if it is located in a non-zero-weight topographic map cell, the short side of the outward-expanding quadrilateral stops expanding, and the long side continues to expand until the vertex of the outward-expanding quadrilateral is located in the zero-weight topographic map cell.

[0015] Preferably, the outer contour of the outward-expanding quadrilateral is used as the boundary, and each closed independent area outside the outward-expanding quadrilateral is named F zone; within the outward-expanding quadrilateral, according to the weight topographic map of the steel structure, the area with an average weight difference greater than δ is named S zone, and the single-layer grid structure, double-layer grid structure and other structural forms with weight differences in the complex steel structure are intelligently distinguished according to the size of δ, and then the areas are intelligently divided, and the other areas are named P zone.

[0016] As a preferred method, traverse the Z coordinates of the steel structure nodes, obtain the height difference of the steel structure, and divide the steel structure according to the following rules; If the height difference of the steel structure is small, then: 1) divide the steel structure into an even number of equal parts along the long side of the quadrilateral; 2) divide the steel structure into an even number of equal parts along the short side of the quadrilateral; 3) divide the steel structure into equal parts along both the long and short sides at the same time.

[0017] If the height difference of the steel structure is large, determine whether there is more than one maximum height point. If not, take the top of the steel structure as the center: 1) divide the steel structure into odd equal parts along the long side of the quadrilateral; 2) divide the steel structure into odd equal parts along the short side of the quadrilateral; 3) divide the steel structure into equal parts along the long and short sides at the same time. If there is more than one maximum height point, first divide the larger area by the minimum height point, and then divide it according to the partition method when there is more than one maximum height point.

[0018] F ij , S ij , P ijName each steel structure partition in turn, among which F ij , S ij , P ij They are represented as the steel structure partition in the i-th row and j-th column of the F, S, and P areas respectively.

[0019] As a preference, the steel structure lifting database is used to select the characteristics that have a greater impact on the steel structure lifting construction cost, such as deformation, stress ratio, steel structure size, weight and number of lifting points, to train the preset AI model; the optimal steel structure lifting AI model is obtained by adjusting the hyperparameters of the AI ​​model; the main function of the steel structure lifting AI model is to evaluate and predict the construction costs of each sub-item of the steel structure partition according to the influencing characteristics of the steel structure size, weight and number of lifting points.

[0020] As a preferred method, based on the trained AI model for steel structure improvement, the deformation, stress ratio, total construction cost and construction cost of the partitioned steel structure measures are evaluated according to the rules; the genetic algorithm is combined with the AI ​​model for steel structure improvement, and the flow construction process is adopted. On the basis of meeting the standard deformation and stress ratio, the total construction cost is minimized as the control index to optimize the first three partition forms and construction paths with the minimum total construction cost; the single-step construction cost is the construction cost of this step minus the construction cost of the previous step that can be used for this step, and the block selected for the next construction step is at least one block away from the block selected for the previous construction step; the fitness function is as shown in Formula 1, where B k+1 is the construction cost of the k+1th construction step, C k is the cost of construction measures in the kth construction step that can be used in the k+1th construction step; Formula (1).

[0021] For steel structures with large height differences, the first construction step is to lift the highest block, and then lift the blocks cumulatively in order of height. For steel structures with small height differences, randomly select blocks with a number ≥ 1 to start lifting.

[0022] As a preferred method, when the genetic algorithm is combined with the steel structure improvement AI model to determine the partition size and number, each time the number of blocks is adjusted, the steel structure partitions are renamed in sequence according to the steel structure division method until the optimal partition size and number are confirmed; when optimizing the optimal construction path, the selected steel structure partition name is output for each construction step, and the construction path can be confirmed according to the construction step and partition name; based on the three selected partition forms and construction paths, a multi-objective genetic algorithm is selected, with the docking accuracy between the nodes of each partition bar as the control index, and the overall stress and deformation of the structure are considered with the minimum structural strain energy, and the steel structure partition hanging points and support frame positions are intelligently arranged for re-optimization; the partition form and construction path with high economic benefits and low docking accuracy are selected; the fitness function is as shown in Formula 2: Formula (2).

[0023] Where BIAS is the deformation of the partition docking point, Di,j is the deformation of the jth docking point in the i-th partition, and energy is the overall strain energy of the structure.

[0024] The present invention can achieve the following effects: The present invention provides an AI-based intelligent planning method for steel structure lifting construction paths. Compared with the prior art, the steel structure is intelligently partitioned by abstracting the weight topographic map of the steel structure, and the factors such as the lower structure, the shape of the steel structure, and the height difference of the steel structure are considered. Combined with the weight topographic map of the steel structure, the steel structure partitioning method is corrected with all factors to solve the problem that the existing steel structure partitioning mainly relies on experience. By combining the AI ​​model, the steel structure lifting database and the intelligent algorithm, the existing construction data is considered to intelligently determine the steel structure partition and plan the lifting construction path. The problem that the existing steel structure lifting construction path mainly relies on experience, the scheme design is slow, and the economy is not high is solved.

[0025] It has the following advantages: 1. Improve the scientificity and accuracy of steel structure zoning and construction path selection, and avoid the limitations of artificial experience methods.

[0026] 2. Improve the economy of steel structure partitioning and construction path selection. Based on construction factors and taking economy as the control indicator, determine the size of steel structure partitions and plan the partition construction path.

[0027] 3. The efficiency of steel structure lifting construction path planning has been improved. The method combining intelligent algorithm, steel structure lifting database and AI can quickly plan the steel structure lifting construction path.

[0028] 4. It is not only applicable to regular shaped steel structures, but also to special shaped steel structures.

[0029] 5. It is not only applicable to steel structures with a single structural form, but also to steel structures with complex structural forms composed of multiple structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0031] The technical solution of the invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0032] Example 1: Figure 1 As shown, an AI-based intelligent planning method for steel structure lifting construction path includes the following steps: The first step: Establish a steel structure grid lifting database. The grid lifting database is based on the existing steel structure engineering construction data, and establishes a steel structure lifting database, including the lifting steel structure size, lifting steel structure weight, steel structure lifting point number, lifting deformation, stress ratio and steel structure lifting construction cost data.

[0033] Step 2: Make an abstract steel structure weight topography map based on the grid coordinates and weight.

[0034] The abstract steel structure weight topographic map uses the steel structure cell as the basic unit to store the number and weight of each member. The cell 1 is labeled D1 to store the number and weight of each member. The numbers of the members that make up cell 1 are D1-1, D1-2, and D1-3, and the weights are M1-1, M1-2, and M1-3 respectively; the adjacent cell 2 is labeled D2, and the numbers of the members that make up cell 2 are D2-1, D2-2, and D2-3, and the weights are M2-1, M2-2, and M2-3 respectively; the same applies to the remaining cells. The X and Y coordinate points of the upper chord nodes of the steel structure are extracted to form a plane topographic map of the steel structure. The closed figure surrounded by the upper chord members is taken as the topographic map unit. The weight of the cell within the projection area of ​​the topographic map unit is assigned to the corresponding topographic map unit to form a steel structure weight topographic map.

[0035] Step 3: Based on the factors of the lower structure, grid shape and grid height difference, combined with the steel structure weight topography, carry out intelligent zoning of the full-factor steel structure.

[0036] Correct the influence of the lower structure on the weight map of the steel structure. This involves the existence of supporting columns or modeling columns at the bottom of the steel structure. The weight of the supporting columns or modeling columns is evenly superimposed on the corresponding topographic map range unit. Correct the influence of the shape on the weight map of the steel structure. For steel structures with strange and irregular shapes, traverse the top, bottom, leftmost and rightmost protruding points of the ring steel structure, and use each protruding point as a reference to form a closed rectangle containing the outer shape of the steel structure. The closed figure formed by the edge of the rectangle and the edge of the steel structure is divided into several topographic map units with a side length of 1 unit. Each topographic map unit is named E1, E2, and E3 in turn, and is given a weight of 0, which is called a zero-weight topographic map cell.

[0037] Taking the center of the rectangular steel structure weight terrain figure as the origin, an outward expansion quadrilateral similar to the rectangular steel structure weight terrain figure is formed. Through the traversal algorithm, it is continuously expanded. When there is a zero-weight terrain map cell in the outward expansion quadrilateral, the outward expansion quadrilateral stops expanding and detects whether the vertex of the outward expansion quadrilateral is located in the zero-weight terrain map cell. If it is located in the zero-weight terrain map cell, the outward expansion quadrilateral stops expanding; if it is located in a non-zero-weight terrain map cell, the short side of the outward expansion quadrilateral stops expanding, and the long side continues to expand until the vertex of the outward expansion quadrilateral is located in the zero-weight terrain map cell.

[0038] Taking the outer contour of the expanded quadrilateral as the boundary, the non-zero weight closed independent areas outside the expanded quadrilateral are named F areas; inside the expanded quadrilateral, according to the steel structure weight topography, the area with an average weight difference greater than δ is named S area, and the single-layer grid structure, double-layer grid structure and other structural forms with weight differences in complex steel structures are intelligently distinguished according to the size of δ, and then the areas are intelligently divided, and the other areas are named P areas.

[0039] Traverse the Z coordinates of the steel structure nodes, obtain the height difference of the steel structure, and divide the steel structure according to the following rules: If the height difference of the steel structure is small, then: 1) divide the steel structure into an even number of equal parts along the long side of the quadrilateral; 2) divide the steel structure into an even number of equal parts along the short side of the quadrilateral; 3) divide the steel structure into equal parts along both the long and short sides at the same time.

[0040] If the height difference of the steel structure is large, determine whether there is more than one maximum height point. If not, take the top of the steel structure as the center: 1) divide the steel structure into odd equal parts along the long side of the quadrilateral; 2) divide the steel structure into odd equal parts along the short side of the quadrilateral; 3) divide the steel structure into equal parts along the long and short sides at the same time. If there is more than one maximum height point, first divide the larger area by the minimum height point, and then divide it according to the partition method when there is more than one maximum height point.

[0041] F ij , S ij , P ij Name each steel structure partition in turn, among which F ij , S ij , P ij They are represented as the steel structure partition in the i-th row and j-th column of the F, S, and P areas respectively.

[0042] Step 4: Utilize the combination of AI models, steel structure improvement database and intelligent algorithms, and integrate existing construction data to establish an AI database to intelligently plan the steel structure improvement construction path.

[0043] By using the steel structure lifting database, we selected the features that have a greater impact on the steel structure lifting construction cost, such as deformation, stress ratio, steel structure size, weight and number of lifting points, and trained the preset AI model; by adjusting the hyperparameters of the AI ​​model, we obtained the optimal steel structure lifting AI model; the main function of the steel structure lifting AI model is to evaluate and predict the construction costs of each sub-item of the steel structure partition according to the influencing characteristics of the steel structure size, weight and number of lifting points.

[0044] Based on the trained AI model for steel structure improvement, the deformation, stress ratio, total construction cost and construction cost of measures that can be used for other zoning measures of the steel structure partition are evaluated according to the rules; the genetic algorithm is combined with the AI ​​model for steel structure improvement, and the flow construction process is adopted. On the basis of meeting the deformation and stress ratio of the specifications, the total construction cost is minimized as the control index to optimize the first three zoning forms and construction paths with the minimum total construction cost; the single-step construction cost is the construction cost of this step minus the construction cost of the previous step that can be used for this step, and the block selected for the next construction step is at least one block away from the block selected for the previous construction step; the fitness function is as shown in Formula 1, where B k+1 is the construction cost of the k+1th construction step, C k is the cost of construction measures in the kth construction step that can be used in the k+1th construction step; Formula (1); For steel structures with large height differences, the first construction step is to lift the highest block, and then lift the blocks cumulatively in order of height. For steel structures with small height differences, randomly select blocks with a number ≥ 1 to start lifting.

[0045] When the genetic algorithm is combined with the steel structure improvement AI model to determine the partition size and number, each time the number of blocks is adjusted, the steel structure partitions are renamed in sequence according to the steel structure division method until the optimal partition size and number are confirmed; when optimizing the optimal construction path, the selected steel structure partition name is output for each construction step, and the construction path can be confirmed according to the construction step and partition name; based on the three selected partition forms and construction paths, a multi-objective genetic algorithm is selected, with the docking accuracy between the nodes of each partition bar as the control indicator, and the overall stress and deformation of the structure are considered with the minimum structural strain energy, and the steel structure partition hanging points and support frame positions are intelligently arranged for re-optimization; select the partition form and construction path with high economic benefits and low docking accuracy; the fitness function is as shown in Formula 2: Formula (2); Where BIAS is the deformation of the partition docking point, Di,j is the deformation of the jth docking point in the i-th partition, and energy is the overall strain energy of the structure.

[0046] In summary, this AI-based intelligent planning method for steel structure lifting construction paths intelligently partitions steel structures by abstracting the weight topographic map of steel structures, and considers factors such as the lower structure, steel structure shape, and height difference of steel structures. Combined with the weight topographic map of steel structures, the steel structure partitioning method is fully corrected to solve the problem that existing steel structure partitioning mainly relies on experience. By combining AI models, steel structure lifting databases, and intelligent algorithms, and considering existing construction data, the steel structure partitions and lifting construction paths are intelligently determined. This solves the problem that existing steel structure lifting construction paths mainly rely on experience, slow scheme design, and low economy.

[0047] The above description is only a specific embodiment of the present invention, but the structural features of the present invention are not limited thereto. Any changes or modifications made by any technician in the field of the present invention are included in the patent scope of the present invention.

Claims

1. An AI-based intelligent planning method for steel structure lifting construction path, characterized in that The steps are as follows: Step 1: Establish a steel structure grid lifting database; Step 2: Make an abstract steel structure weight topographic map based on the grid coordinates and weight; Step 3: Based on the factors of the lower structure, grid shape and grid height difference, combined with the steel structure weight topography, all-element steel structure intelligent zoning is carried out; Step 4: Utilize the combination of AI models, steel structure improvement database and intelligent algorithms, and integrate existing construction data to establish an AI database to intelligently plan the steel structure improvement construction path.

2. The AI-based intelligent planning method for steel structure lifting construction path according to claim 1 is characterized by: The grid lifting database is based on the existing steel structure engineering construction data to establish a steel structure lifting database, including data on the size of the lifted steel structure, the weight of the lifted steel structure, the number of steel structure lifting points, lifting deformation, stress ratio and steel structure lifting construction costs.

3. The AI-based intelligent planning method for steel structure lifting construction path according to claim 1 is characterized by: The abstract steel structure weight topographic map takes the steel structure cell as the basic unit and stores the number and corresponding weight of each member. The X and Y coordinate points of the upper chord nodes of the steel structure are extracted to form a plane topographic map of the steel structure. The closed figure surrounded by the upper chord members is taken as the topographic map unit. The weight of the cell within the projection area of ​​the topographic map unit is assigned to the corresponding topographic map unit to form a steel structure weight topographic map.

4. The AI-based intelligent planning method for steel structure lifting construction path according to claim 3 is characterized by: Correct the influence of the lower structure on the weight topographic map of the steel structure. If there are supporting columns or modeling columns under the steel structure, the weight of the supporting columns or modeling columns will be evenly superimposed on the corresponding topographic map range unit; Correct the influence of shape on the weight map of steel structure. For steel structures with strange and irregular shapes, traverse the protruding points at the top, bottom, leftmost and rightmost parts of the ring steel structure, and use each protruding point as a reference to form a closed rectangle containing the outer shape of the steel structure. The closed figure formed by the edge of the rectangle and the edge of the steel structure is divided into several map units with a side length of unit length. The weight of each map unit is assigned to be 0, which is called a zero-weight map cell.

5. The AI-based intelligent planning method for steel structure lifting construction path according to claim 4 is characterized in that: Taking the center of the rectangular steel structure weight topography as the origin, an outward-expanding quadrilateral similar to the rectangular steel structure weight topography is formed, and the outward expansion is continuously carried out through the traversal algorithm. When there is a zero-weight topography cell in the outward-expanding quadrilateral, the outward-expanding quadrilateral stops expanding, and it is detected whether the vertex of the outward-expanding quadrilateral is located in the zero-weight topography cell; if it is located in the zero-weight topography cell, the outward-expanding quadrilateral stops expanding; if it is located in a non-zero-weight topography cell, the short side of the outward-expanding quadrilateral stops expanding, and the long side continues to expand until the vertex of the outward-expanding quadrilateral is located in the zero-weight topography cell.

6. The AI-based intelligent planning method for steel structure lifting construction path according to claim 5 is characterized by: The outer contour of the expanded quadrilateral is used as the boundary to enclose the non-zero weight areas outside the expanded quadrilateral; inside the expanded quadrilateral, the average weight difference of each area is compared with δ according to the steel structure weight topography map, and the single-layer grid structure, double-layer grid structure and other structural forms with weight differences in the complex steel structure are intelligently distinguished according to the size of δ, and then the areas are intelligently divided and named.

7. The AI-based intelligent planning method for steel structure lifting construction path according to claim 6 is characterized by: Traverse the Z coordinates of the steel structure nodes, obtain the height difference of the steel structure, and divide the steel structure according to the following rules; If the height difference of the steel structure is small, then: 1) divide the steel structure into even numbers along the long side of the quadrilateral; 2) divide the steel structure into even numbers along the short side of the quadrilateral; 3) divide the steel structure into equal parts along both the long and short sides; If the height difference of the steel structure is large, determine whether there is more than one maximum height point. If not, take the top of the steel structure as the center: 1) divide the steel structure into odd equal parts along the long side of the quadrilateral; 2) divide the steel structure into odd equal parts along the short side of the quadrilateral; 3) divide the steel structure into equal parts along the long and short sides at the same time; if there is more than one maximum height point, first divide the larger area by the minimum height point, and then divide it according to the partition method when there is more than one maximum height point.

8. The AI-based intelligent planning method for steel structure lifting construction path according to claim 1 is characterized by: By using the steel structure lifting database, we selected the features that have a greater impact on the steel structure lifting construction cost, such as deformation, stress ratio, steel structure size, weight and number of lifting points, and trained the preset AI model; by adjusting the hyperparameters of the AI ​​model, we obtained the optimal steel structure lifting AI model; the main function of the steel structure lifting AI model is to evaluate and predict the construction costs of each sub-item of the steel structure partition according to the influencing characteristics of the steel structure size, weight and number of lifting points.

9. The AI-based intelligent planning method for steel structure lifting construction path according to claim 8 is characterized by: Based on the trained AI model for steel structure improvement, the deformation, stress ratio, total construction cost and construction cost of the partitioned steel structure measures are evaluated according to the rules; the genetic algorithm is combined with the AI ​​model for steel structure improvement, and the flow construction process is adopted. On the basis of meeting the deformation and stress ratio of the specifications, the total construction cost is minimized as the control index to optimize the first three partition forms and construction paths with the minimum total construction cost; the single-step construction cost is the construction cost of this step minus the construction cost of the previous step that can be used for this step, and the block selected for the next construction step is at least one block away from the block selected for the previous construction step; the fitness function is as shown in Formula 1, where B k+1 is the construction cost of the k+1th construction step, C k is the cost of construction measures in the kth construction step that can be used in the k+1th construction step; Formula (1); For steel structures with large height differences, the first construction step is to lift the block with the highest height, and then lift the blocks cumulatively in order of height. For steel structures with small height differences, randomly select blocks with a number ≥1 to start lifting.

10. The AI-based intelligent planning method for steel structure lifting construction path according to claim 9 is characterized in that: When the genetic algorithm is combined with the steel structure improvement AI model to determine the partition size and number, each time the number of blocks is adjusted, the steel structure partitions are renamed in sequence according to the steel structure division method until the optimal partition size and number are confirmed; when optimizing the optimal construction path, the selected steel structure partition name is output for each construction step, and the construction path can be confirmed according to the construction step and partition name; based on the three selected partition forms and construction paths, a multi-objective genetic algorithm is selected, with the docking accuracy between the nodes of each partition bar as the control indicator, and the overall stress and deformation of the structure are considered with the minimum structural strain energy, and the steel structure partition hanging points and support frame positions are intelligently arranged for re-optimization; select the partition form and construction path with high economic benefits and low docking accuracy; the fitness function is as shown in Formula 2: Formula (2); Where BIAS is the deformation of the partition docking point, Di,j is the deformation of the jth docking point in the i-th partition, and energy is the overall strain energy of the structure.

Images