A method for generating secondary beams based on main beams

Abstract

The present invention provides a method for generating secondary beams based on main beams, which solves problems such as poor construction effects of main and secondary beams in buildings. The method includes the following steps: generating an initial secondary beam scheme according to the original main beams and other elements; initially optimizing the wiring of the initial scheme; using the region growing algorithm; creating regional jitter; starting regional expansion; eliminating small regions; using the game tree algorithm to allocate all unallocated regions, and processing and correcting incorrect secondary beams again. The present invention has the advantages of high secondary beam design efficiency and convenient selection of the optimal solution, etc.

Description

Technical Field

[0001] The present invention belongs to the technical field of architectural design, and particularly relates to a secondary beam generation method based on a main beam. Background Art

[0002] The main beams and secondary beams in the construction industry are generally the overall support structures in the entire building, belonging to the structures that cannot be randomly damaged. Among them, the secondary beams are generally lapped on the main beams and mainly play the role of transferring loads. In the architectural design industry, when structural professional designers design the layout of secondary beams, they need to design a reasonable and beautiful layout of secondary beams according to the "environment" of the surrounding main beams and certain specifications in the construction industry. However, when designers design the layout of secondary beams, they will face a very common problem. Because the environment of the surrounding main beams is not static, there may be triangular main beam compartments, quadrilateral main beam compartments, pentagonal main beam compartments or more complex situations in the surrounding main beams. Moreover, according to the surrounding main beam compartment environment, several secondary beam layout schemes that can meet the requirements can be generated for each compartment. That is to say, there are several design schemes for each beam compartment area, and designers need to arrange secondary beams in all beam compartment areas, find an overall optimal scheme, and select the overall most beautiful and most in line with the architectural design requirements. This is a very time-consuming and laborious thing for designers. Moreover, the artificial selection of secondary beam components may also result in different effects for different people. Different designers are very likely to have inconsistent secondary beam layouts for the same drawing design.

[0003] In order to solve the deficiencies of the prior art, people have conducted long-term explorations and proposed various solutions. For example, a Chinese patent document discloses a secondary beam structure generation method and related equipment based on a BIM platform [202210757473.0], which includes: loading a BIM model containing a main beam structure; starting a secondary beam auxiliary design device and identifying the secondary beam layout direction and each structural partition of the BIM model; determining the corresponding secondary beam generation range in each structural partition according to a preset partition layout strategy; and sequentially generating secondary beam structures in the secondary beam generation ranges corresponding to each structural partition according to the secondary beam layout direction and a preset partition layout priority.

[0004] The above solution solves the problem of relatively low secondary beam design efficiency to a certain extent, but there are still many deficiencies in this solution. For example, problems such as the inability to select the optimal layout scheme by manual design. Summary of the Invention

[0005] The purpose of the present invention is to provide a secondary beam generation method based on a main beam with reasonable design and convenient selection of the optimal secondary beam design scheme for the above problems.

[0006] To achieve the above object, the present invention adopts the following technical solution: a method for generating a secondary beam based on a main beam, comprising the following steps:

[0007] S1: Generate the initial secondary beam solution based on the original main beam and other elements;

[0008] S2: Preliminary optimization of the initial wiring scheme to eliminate obvious abnormal areas;

[0009] S3: Using the region growing algorithm, starting with a set of seeds, adjacent pixels with similar properties to the seeds are added to the seeds in the growing region. The current layout result is divided into connected regions according to the direction of the secondary beam layout to establish the concept of region.

[0010] S4: manufacturing area jitter, shaking off irregular areas;

[0011] S5: Start region expansion, calculate the convex hull for each region, and then assign the unassigned segments in the convex hull to the region;

[0012] S6: Homologous regions merge and then re-expand, and heterologous regions divide the bordering regions and merge the bordering regions into two regions;

[0013] S7: Eliminate small areas and break them up and integrate them into undifferentiated areas;

[0014] S8: Use the game tree algorithm to allocate all unallocated areas and process and correct the erroneous secondary beams again. The overall idea is to unify the secondary beam generation logic, generate the optimal solution for each beam interval, and then gradually optimize the local optimal solution, so that the local optimal solution is gradually destroyed and the global optimal solution is gradually reached.

[0015] In the aforementioned method for generating secondary beams based on primary beams, in step S1, the secondary beams are obtained using a local optimal solution for the entire primary beam region, where other elements include columns, walls, and other obstacles. Different optimal secondary beam solutions are generated for different primary beam regions.

[0016] In the above-mentioned method for generating secondary beams based on main beams, in step S2, if the connection direction of an area is up and down or left and right, and the directions of the upper, lower, left and right adjacent areas of the area are all left and right or up and down, then the direction of the area is adjusted to the left and right or up and down.

[0017] In the above-mentioned method for generating secondary beams based on primary beams, the seed property in step S3 is the arrangement direction.

[0018] In the above-mentioned method for generating a secondary beam based on a main beam, step S4 shakes off the edge portion and the sharp corner portion of the growth area.

[0019] In the above method for generating secondary beams based on main beams, step S5 includes the following steps:

[0020] S51: The growth area expands and collides with the surrounding area;

[0021] S52: Divide the collision area;

[0022] S53: Incorporate the divided collision part into the growth area.

[0023] In the above method for generating secondary beams based on main beams, in step S6, the homologous areas are areas with the same direction, and the heterologous areas are areas with different directions.

[0024] In the above method for generating secondary beams based on main beams, step S8 includes the following steps:

[0025] S81: Construct a game tree, regard each possible situation in the next step as a node on a tree, and regard each action as an edge on the tree;

[0026] S82: Apply the evaluation function to the leaf nodes to evaluate each step;

[0027] S83: Alpha-beta pruning, use the minimax algorithm to prune the entire tree, cutting off many unnecessary branches and impossible results;

[0028] S84: Select the optimal result, and select the optimal solution evaluated according to the pruning result.

[0029] In the above method for generating secondary beams based on main beams, step S81 includes the following steps:

[0030] S811: Regard the unallocated area as a game, and regard the adjacent allocated areas of this area as a player;

[0031] S812: After the players form teams, create a chessboard and preprocess the chessboard. Abstract the unallocated area into a chessboard, and regard each internal beam separation area as an empty position in the chessboard;

[0032] S813: Sort by group magnitude, start playing chess from the group with the largest magnitude, and the valuation is the overall score.

[0033] In the above method for generating secondary beams based on main beams, step S812 groups by the layout direction; step S813 does not allow dropping pieces across grids but allows connecting pieces.

[0034] Compared with the existing technologies, the advantages of the present invention are as follows: automatically generating a secondary beam layout scheme to replace manual design, improving the drawing efficiency, and quickly finding the optimal solution through a game tree; being able to modify and update the design structure according to its own needs, and independently choosing to manually or automatically generate the required secondary beam design results; automatically adapting to main beams of different shapes and being able to generate multiple design schemes at one time. Description of the Drawings

[0035] Figure 1 is the execution idea diagram of the secondary beam layout algorithm of the present invention;

[0036] Figure 2 is the flowchart of secondary beam generation of the present invention;

[0037] Figure 3 is the structural schematic diagram of the abnormal area of the present invention;

[0038] Figure 4 is the effect diagram of the region growing algorithm of the present invention.

[0039] Figure 5 is the input schematic diagram of the actual application case of the present invention.

[0040] Figure 6 is the output schematic diagram of the actual application case of the present invention. Detailed Embodiment

[0041] The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0042] As Figure 1-2 shown, a method for generating secondary beams based on main beams includes the following steps:

[0043] S1: Generating an initial secondary beam scheme according to the original main beam and other elements, wherein the layout of the secondary beams is basically divided into three layout structures, double secondary beams, frame large slabs, and single secondary beams; the main beam partition areas are also mainly divided into three areas, triangular areas, quadrilateral areas, and pentagonal areas;

[0044] S2: Conducting preliminary optimization on the wiring of the initial scheme to eliminate obvious abnormal areas; different shapes may layout different "optimal solution" secondary beam schemes, but at the same time there will be 3-4 alternative sub-optimal solutions, and it is necessary to automatically find one or more sub-optimal solutions and the optimal solution at one time according to the surrounding environment of this beam partition interval;

[0045] S3: Adopting a region growing algorithm, starting with a group of seeds, attaching adjacent pixels with similar properties to the seeds of the growing region, dividing the current layout result according to the wiring area in the direction of secondary beam layout, and establishing the concept of a region;

[0046] S4: Creating regional jitter to shake off irregular regions;

[0047] S5: Start the regional expansion. Calculate the convex hull for each region, and then assign the unassigned compartments within the convex hull to that region;

[0048] S6: Fuse the homologous regions and then re-expand. Divide the adjacent regions for the heterologous regions and merge the adjacent regions into the two regions;

[0049] S7: Eliminate the small regions and break them up and integrate them among the undifferentiated regions;

[0050] S8: Use the game tree algorithm to assign all unassigned regions and process and correct the error secondary beams again. The overall core idea is to first uniformly generate the optimal solution for each beam compartment interval according to the secondary beam generation logic, and then gradually optimize this local optimal solution to gradually destroy the local optimum but gradually reach the global optimum.

[0051] As Figure 2 shown, in step S1, the secondary beams in the main beam region are all obtained using the local optimal solution, where other elements include columns, walls, and other obstacles. The layout algorithm for the secondary beams includes the following steps: construct the connection relationship; generate the initial secondary beams for all beam compartment regions according to the designer's requirements; and preliminarily adjust the layout of the secondary beams.

[0052] As Figure 3 shown, in step S2, if the connection direction of a region is the up-down or left-right direction, and the directions of the adjacent regions above, below, left, and right of this region are all the left-right or up-down direction, then adjust the direction of this region to the left-right or up-down direction.

[0053] As Figure 4 shown in the boxed area, the seed property in step S3 is the layout direction, and the adjacent pixels with the same layout direction are attached to the seeds of the growing region.

[0054] Furthermore, in step S4, shake off the edge parts and corner parts of the growing region. Because the initial growing region is just a simple set of compartments in the same direction and adjacent to each other, the shape of the region will not be regular, and there may be linear regions and some other irregular regions generated. At this time, shake off these irregular regions through shaking to make the region "slim down" but become more "healthy", and the region no longer has scattered tentacles.

[0055] In addition, although the regional shaking will shake off the scattered tentacles, it may still cause additional regional damage or misjudgment. At this time, calculate the convex hull for each region, and then assign the unassigned compartments (scattered points not within the region) within the convex hull to that region. This step can cause the expansion of the region and will not destroy the "health" of the region, nor create new "tentacles" for the region. Step S5 includes the following steps:

[0056] S51: The growth area expands and collides with the surrounding area;

[0057] S52: Divide the collision area;

[0058] S53: Incorporate the divided collision part into the growth area.

[0059] Meanwhile, after step S6, the area range will become larger after area expansion, which may cause the areas to border each other. The bordering areas may be homologous (in the same direction) or heterologous (in different directions). Homologous areas can be merged and then expand again (refer to the fifth step). Heterologous areas can divide the bordering area and incorporate the bordering area into the two areas.

[0060] The purpose of step S7 is to eliminate small areas. Since small areas are relatively small themselves and their own gravitational force is relatively small, the simplest example is a 2*2 small area. Although it is "healthy" itself (a cube without extra tentacles), due to its own scale factor, it may cause damage to the final area. Therefore, eliminate small areas and break them up and incorporate them into the undivided areas.

[0061] Visibly, step S8 is the most important step. Steps S1 to S7 are just preprocessing to create a better overall environment for the final eighth step. At this time, the environment consists of relatively "healthy" areas, the remaining discrete undistributed partitions, and the undistributed partitions gathered together. Then, for each relatively independent undistributed partition area, only known allocated areas may exist around it. Then, these known allocated areas will compete through game theory. Here, the algorithm idea of the game tree is adopted. The idea of the game tree is a two-player game. The core idea is to establish an evaluation system. Two players, A and B, play the game, taking one step each. The purpose of A's turn is to increase A's score, and the purpose of B's turn is to decrease A's score. Based on this, the game is carried out. However, the actual application is more complex than the ordinary game tree (it may not necessarily be a two-player game and may involve multi-player games), the evaluation system is more troublesome (the evaluation of the game tree is to calculate the score of a certain player, while the evaluation of the secondary beam layout game is the overall aesthetic score), and the gameplay is more complex (the game tree is one step per person, and the secondary beam layout game is not one step per person). The specific steps are as follows:

[0062] S81: Construct a game tree, regard each possible situation in the next step as a node on a tree, regard each action as an edge on the tree, and the development of each step can be regarded as the expansion from one node to the next-level node;

[0063] S82: Apply the evaluation function to the leaf nodes and evaluate each step;

[0064] S83: Alpha-beta pruning is used to prune the entire tree using the minimax algorithm, cutting off many unnecessary branches and impossible outcomes, thus optimizing the overall tree structure.

[0065] S84: Select the optimal result by choosing the best solution evaluated based on the pruning result.

[0066] Obviously, this algorithm only adopts the idea of the game tree, and its internal logic is not exactly the same. It should be more appropriately called a "cooperation tree". The algorithmic idea of the secondary beam layout cooperation tree includes the following steps:

[0067] S811: Consider the unallocated area as a game, and then find players to participate, that is, find the adjacent allocated areas of this area and regard the adjacent allocated areas as a player.

[0068] S812: After the players form teams, create a chessboard and preprocess the chessboard. Abstract the unallocated area as a chessboard, and the internal beam separation areas as empty positions on the chessboard, waiting for the players to play and compete.

[0069] S813: Sort by group magnitude and start playing chess from the group with the largest magnitude. The steps of playing chess are also different. It is no longer one step per person because the core idea of one step per person is for two players to play against each other to determine the winner, while the purpose of this algorithm is to maximize the total score. Therefore, the steps of playing chess are sorted by group magnitude and start "playing chess" from the group with the largest magnitude. Its score is not to maximize the score of that group but the overall score.

[0070] Preferably, in step S812, group by the layout direction and initially assign the same original direction teams to the players. The advantage of forming pairs into groups is that it reduces the number of players, and all players cooperate in the form of groups. In step S813, it is not allowed to place a piece across grids but allowed to connect pieces. For example, if grid 2 is connected to grid 1 and grid 1 is connected to player 1, then the player can only place a piece on grid 1 first and then on grid 2. The rule for connecting pieces is that if a player places a piece somewhere, it is equivalent to having a short tentacle in this area, and then for the surrounding areas of this tentacle, the connecting piece operation is performed, reducing the number of algorithm iterations.

[0071] If the above player places a piece at position 3, then the "connecting piece" operation will be triggered, and pieces will be placed at all positions from 1 to 7. Because the remaining positions are allocated to other players and it is considered that it will definitely cause a bad impact, at this time, all seven pieces are placed at once.

[0072] At this time, all areas have been fully allocated, and there will be no unallocated areas. Subsequently, process and correct the possible incorrect secondary beams at the junction, referring to step S2, and the final result can be completed.

[0073] Such asFigure 5 As shown, for a residential community project, there are 8 above-ground individual buildings, and the basement area is approximately 27,000 square meters. Before generation, the layout lines of basement walls, columns, and main beams start from Figure 5 As can be seen from Figure 6 shown.

[0074] In summary, the principle of this embodiment is as follows: The manual operations of designers are processed through algorithms to become procedural. Starting from the original requirements of designers, combined with different drawings and different businesses, automatic layout is achieved. A large number of optimizations are used to shorten the time for secondary beam layout, and a brand-new concept of cooperation tree is used to abstract the overall structure of the main and secondary beams into a chessboard, automatically obtaining the optimal solution.

[0075] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Those skilled in the art to which the present invention pertains can make various modifications or supplements to the described specific embodiments or use similar ways of substitution, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

[0076] Although terms such as region growing algorithm and game tree algorithm are used more frequently in this article, the possibility of using other terms is not excluded. Using these terms is only for more conveniently describing and explaining the essence of the present invention; interpreting them as any additional limitation is contrary to the spirit of the present invention.

Claims

1. A method for generating secondary beams based on main beams, characterized in that The steps include: S1: Generate the initial secondary beam solution based on the original main beam and other elements; S2: Preliminary optimization of the initial wiring scheme to eliminate obvious abnormal areas; S3: Using the region growing algorithm, starting with a set of seeds, adjacent pixels with similar properties to the seeds are added to the seeds in the growing region. The current layout result is divided into connected regions according to the direction of the secondary beam layout to establish the concept of region. S4: manufacturing area jitter, shaking off irregular areas; S5: Start region expansion, calculate the convex hull for each region, and then assign the unassigned segments in the convex hull to the region; S6: Homologous regions merge and then re-expand, and heterologous regions divide the bordering regions and merge the bordering regions into two regions; S7: Eliminate small areas and break them up and integrate them into undifferentiated areas; S8: Use the game tree algorithm to allocate all unallocated areas, and process and correct the erroneous secondary beams again.

2. The method for generating secondary beams based on a main beam according to claim 1, wherein In step S1, the secondary beams are obtained by using the local optimal solution in the main beam area, wherein other elements include columns, walls and other obstacles.

3. The method for generating secondary beams based on main beams according to claim 1, wherein In step S2, if the connection direction of an area is up and down or left and right, and the directions of the upper, lower, left and right adjacent areas of the area are all left and right or up and down, then the direction of the area is adjusted to left and right or up and down.

4. A secondary beam generation method based on a main beam according to claim 1, characterized in that The seed property in step S3 is the arrangement direction.

5. A method for generating secondary beams based on main beams according to claim 1, characterized in that, The step S4 shakes off the edge portion and the sharp corner portion of the growth area.

6. The method for generating secondary beams based on main beams according to claim 1, characterized in that, The step S5 includes the following steps: S51: The growing area expands and collides with the surrounding area; S52: Divide the collision area; S53: The collision part is divided and integrated into the growth area.

7. A method for generating secondary beams based on main beams according to claim 1, characterized in that, In step S6, the homologous regions are regions with the same direction, and the heterologous regions are regions with different directions.

8. A secondary beam generation method based on a main beam according to claim 1, characterized in that, The step S8 includes the following steps: S81: Construct a game tree, treating each possible situation in the next step as a node on a tree, and each action as an edge on the tree; S82: Apply the evaluation function to the leaf nodes and perform evaluation for each step; S83: α-β pruning, using the minimax algorithm to prune the entire tree, removing many unnecessary branches and impossible results; S84: Select the optimal result, and select the evaluated optimal solution based on the pruning result.

9. A method for generating secondary beams based on main beams according to claim 8, characterized in that The step S81 includes the following steps: S811: Treat the unassigned area as a game and the assigned area adjacent to the area as a player; S812: After the players have formed a team, they create a chessboard and pre-process the chessboard. The unallocated area is abstracted into the chessboard, and the internal beam-separated areas are used as empty spaces on the chessboard. S813: Sort by group size, start playing chess from the group with the largest size, and estimate the overall score.

10. A method for generating secondary beams based on main beams according to claim 9, characterized in that, The step S812 groups the pieces according to the arrangement direction; the step S813 does not allow pieces to be placed across squares but allows pieces to be connected.

Images