Partition fitting method for free hyperboloid curtain wall skin

Abstract

The invention discloses a partition fitting method for a free hyperboloid curtain wall skin, and aims to solve the problem that an optimal result of fitting sampling cannot be obtained by an existing hyperboloid curtain wall skin fitting method. The method comprises the following steps: setting arc edge sampling points on two arc edges II of an initial plate model, projecting a sampling line midpoint to the model to generate a sampling projection point, generating a first plane by the arc edge sampling points and the sampling projection point, projecting two arc edges I of the model to the first plane to generate an arc projection line and an arc edge projection line, obtaining a midpoint of a connecting line of corresponding nodes on the two projection lines and generating a middle curve, calculating a distance value from the nodes of the two projection lines to the middle curve, locking the positions of the arc edge sampling point and the middle curve when the distance value is minimum, forming a pressure arc line by the connecting line of two end points of the middle curve and the midpoint, and stretching to form a cylinder, and enabling the model to coincide with the normal of the side wall of the cylinder, projecting the arc edge of the model to the cylinder to form a single-curved-surface plate model, and outputting a pressure arc line, an angular point deviation value and an arc edge midpoint deviation value.

Description

Technical Field

[0001] This invention relates to the field of building curtain wall construction technology, and in particular to a method for partitioning and fitting the surface of a free hyperboloid curtain wall. Background Technology

[0002] With the increasing maturity of BIM technology for curtain walls both domestically and internationally, more and more free-form hyperboloid curtain wall skin designs are emerging. The entire free-form hyperboloid curtain wall skin is a hyperboloid. In actual construction, the entire free-form hyperboloid curtain wall skin is often divided into several panels by the curtain wall segmentation. However, the divided panels still have complex hyperboloid structures, making their manufacturing, processing, and installation quite difficult. Existing methods for fitting the divided panels of the hyperboloid curtain wall skin into a single-curved shape have the following problems:

[0003] 1. If the sampling range of the algorithm is too small, it is difficult to find the best result of the hyperboloid fitting sampling. If the random sampling algorithm is used, the sampling range is too large, the calculation time is long, and the optimization efficiency is low.

[0004] 2. Existing optimization fitting algorithms lack qualitative judgment and cannot determine whether the hyperbolic panel block is suitable to be fitted into a single surface shape or to retain the original hyperbolic surface shape based on the calculation results.

[0005] 3. Existing optimization fitting algorithms only consider the corner point deviations before and after fitting the hyperboloid panel, without considering the deviations of the overall hyperboloid curtain wall skin. Furthermore, the deviation data before and after fitting the hyperboloid and the subsequent processing data cannot be displayed intuitively. Summary of the Invention

[0006] Existing methods for fitting hyperboloid curtain wall skins into single-curved shapes lack qualitative judgment, failing to achieve both optimal fitting and sampling results and improved optimization efficiency. The purpose of this invention is to provide a zonal fitting method for free-form hyperboloid curtain wall skins.

[0007] The technical solution adopted by this invention to solve its technical problem is: a method for partitioning and fitting the surface of a free-form hyperboloid curtain wall, the steps of which are as follows:

[0008] S1: Divide the free-form hyperboloid curtain wall skin model into several panel models, extract a single initial panel model. The four arc edges of the initial panel model are two arc edges one and two arc edges two, which are set opposite to each other. Arc edge sampling points are set on the two arc edges two respectively. Connect the two arc edge sampling points to form a sampling line. Take the midpoint of the sampling line as the sampling midpoint. Project the sampling midpoint onto the initial panel model to generate a sampling projection point. Generate a first plane based on the two arc edge sampling points and one sampling projection point. Project the two arc edges one of the initial panel model onto the first plane to generate an arc projection line and an arc edge projection line.

[0009] S2: Divide the arc projection line and the arc edge projection line into several equal segments, with nodes between adjacent segments. Calculate the midpoint of the line connecting two corresponding nodes on the arc projection line and the arc edge projection line. Connect multiple midpoints to generate an intermediate curve. Simultaneously move two arc edge sampling points. The first plane moves along the two arc edges to sample as the two arc edge sampling points move. Calculate the distance L from the nodes of the arc projection line and the arc edge projection line to the intermediate curve. Extract the position of the arc edge sampling point when the distance L is the minimum, and lock the position of the arc edge sampling point and the intermediate curve.

[0010] S3: Connect the two endpoints and the midpoint of the intermediate curve to form a smooth transition arc, forming the pressure arc line. Extract the pressure arc line and stretch it into a single-curved cylinder. Place the initial plate model close to the side wall of the cylinder so that the normals of the initial plate model and the side wall of the cylinder coincide. Project the four arc edges of the initial plate model onto the side wall of the cylinder to generate a single-curved plate block model. Output the pressure arc line to determine the bending direction of the single-curved plate block model. Output the corner deviation value D between the single-curved plate block model and the initial plate model, as well as the midpoint deviation value of the arc edge.

[0011] The method for partitioning and fitting the free-form hyperboloid curtain wall skin of the present invention first divides the free-form hyperboloid curtain wall skin into several initial panel models. Arc edge sampling points are set on the two arc edges of the initial panel models, and a sampling line is formed by connecting two arc edge sampling points. The midpoint of the sampling line is projected onto the initial panel model to generate a sampling projection point. A first plane is generated based on the two arc edge sampling points and one sampling projection point. The two arc edges of the initial panel model are projected onto the first plane to generate arc projection lines and arc edge projection lines. The corresponding sections on the arc projection lines and arc edge projection lines are calculated. The midpoints of the points are connected to form an intermediate curve. Then, the first plane samples along the two arc-edge sampling points as they move. The distance L from the nodes of the arc projection line and the arc-edge projection line to the intermediate curve is calculated. The location of the arc-edge sampling point when the distance L is minimum is extracted, and the positions of the arc-edge sampling point and the intermediate curve are locked. The two endpoints and the midpoint of the intermediate curve are connected to form a smooth transition arc line. The arc line is stretched into a single-curved cylinder, so that the normals of the initial plate model and the cylinder sidewall coincide. The four arc edges of the initial plate model are projected onto the cylinder sidewall to form a single-curved panel. The block model outputs the curved line to determine the bending direction of the single-curved panel block model, and outputs the corner deviation value and arc edge midpoint deviation value between the single-curved panel block model and the initial block model. The partition fitting method of this invention decomposes the initial block model of the free hyperboloid into lines, and then decomposes the lines into points. It adopts a coarse and fine combined sampling method to improve the computational efficiency. It locks the coarse range of the curved line through a geometric algorithm, and performs fine sampling on the determined sampling area. While ensuring the accuracy of the sampling results, it greatly improves the computational efficiency. The free hyperboloid curtain wall skin is optimized into a single-curved surface. After the curtain wall skin is applied, the geometry of the curtain wall becomes simpler, reducing the difficulty of processing and installation, simplifying the construction process, and making subsequent maintenance more convenient. It also saves construction costs by reducing the required construction equipment, manpower, and materials. This method can output deviation data between the single-curved panel block model and the initial panel model before and after surface fitting, and outputs the pressure curve to determine the bending direction and processing data of the single-curved panel block model, facilitating subsequent manual decision-making and verification. This ensures the construction quality of the free-form hyperboloid curtain wall skin while reducing the difficulty of processing and installation.

[0012] Furthermore, in step S2, the distance value L from each node of the arc projection line and the arc edge projection line to the middle curve is calculated respectively. When the distance value L is the smallest, the location of the arc edge sampling point and the middle curve is locked, and the pressure arc line used to generate the optimal single-curved panel block is generated and extracted; or, the distance value L from each node of the arc projection line and the arc edge projection line to the middle curve is calculated. When the sum of all distance values ​​L is the smallest, the location of the arc edge sampling point and the middle curve is locked, and the pressure arc line used to generate the optimal single-curved panel block is generated and extracted.

[0013] Furthermore, in step S2, the arc edge sampling point moves along the arc edge to sample, determines the initial sampling point of the arc line, obtains two extended sampling points adjacent to the initial sampling point on the sampling path, and further divides the arc segment between the two extended sampling points into equal parts for sampling to obtain the minimum value of the distance L.

[0014] Furthermore, in step S2, a distance standard value is set. When the distance value L is less than the distance standard value, the current initial state of the block model meets the conditions for optimization into a single-curved block model; otherwise, it is not suitable for optimization into a single-curved block model.

[0015] Furthermore, in step S3, the corner deviation value and the midpoint deviation value of the arc edge between the single-curved panel block model and the initial block model are output. The weighted average is calculated using the multiple output corner deviation values ​​D and multiple midpoint deviation values ​​E through the weighted average method.

[0016] Furthermore, in step S3, a standard value for corner deviation and a standard value for midpoint deviation are set. When the corner deviation value between the current initial plate model and the single-curved plate block model is less than the standard value for corner deviation and the midpoint deviation value is less than the standard value for midpoint deviation, it is determined that the current initial plate model is suitable to be fitted into a single-curved plate block model; otherwise, it is not suitable to be fitted into a single-curved plate block model. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a free-form hyperboloid curtain wall skin zoning fitting model according to an embodiment of the present invention;

[0018] Figures 2 to 8 This is a schematic diagram of the steps in a partitioning fitting method for a free-form hyperboloid curtain wall skin according to an embodiment of the present invention.

[0019] The numbers in the diagram are as follows:

[0020] Free hyperboloid curtain wall skin model 100; initial panel model 10; arc edge one 1; arc edge two 2; chord edge 3; arc edge sampling point 4; sampling midpoint 5; sampling projection point 6; arc projection line 8; arc edge projection line 9; first plane 7; intermediate curve 10; pressure arc line 12; cylinder 15; single-curved panel block model 16. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, and are only used to facilitate and clearly illustrate the embodiments of the present invention. For ease of description, the terms "upper" and "lower" used below are consistent with the upper and lower directions in the drawings, but this should not be construed as a limitation of the technical solution of the present invention.

[0022] Combination Figures 1 to 8The specific steps of the partition fitting method for the free-form hyperboloid curtain wall skin of the present invention are as follows:

[0023] S1: As Figure 1 As shown, the free-form hyperboloid curtain wall skin model 100 is divided into several panel models, such as... Figure 2 As shown, a single plate model is extracted. The initial plate model (referred to as initial plate model 10) has a hyperboloid shape with four curved edges. The four curved edges are two oppositely set curved edges 1 and two curved edges 2, and each curved edge corresponds to a chord edge 3; as shown Figure 3 As shown, sampling points 4 are set on the two arc edges 2 respectively. The two sampling points 4 are connected to form a sampling line. The midpoint of the sampling line is taken as the sampling midpoint 5. The sampling midpoint 5 is projected onto the initial plate model 10 to generate sampling projection point 6; as shown Figure 4 As shown, a longitudinal plane is determined based on two arc edge sampling points 4 and one sampling projection point 6, generating a first plane 7. The two arc edges 1 of the initial plate model 10 are projected onto the first plane 7 to generate arc projection lines 8 and arc edge projection lines 9.

[0024] S2: As Figure 5 As shown, using a parameterized program, the arc projection line 8 and the arc edge projection line 9 are divided into several equal segments, with nodes between adjacent segments. The midpoint of the line connecting two corresponding nodes on the arc projection line 8 and the arc edge projection line 9 is calculated. Multiple midpoints are connected to generate the intermediate curve 10. Two arc edge sampling points 4 are moved synchronously. The first plane 7 moves along the two arc edges 2 with the two arc edge sampling points 4 to sample. The distance L from the nodes of the arc projection line 8 and the arc edge projection line 9 to the intermediate curve 10 is calculated. The sampling data of the variable distance value L is recorded. The position of the arc edge sampling point 4 when the distance value L is the smallest is extracted from the sampling data, and the positions of the arc edge sampling point 4 and the intermediate curve 10 are locked.

[0025] S3: As Figure 5 and Figure 6 As shown, the two endpoints and the midpoint of the intermediate curve 10 are connected to form a smooth transition arc, forming the pressure arc line 12. The pressure arc line 12 is extracted and stretched into a single-curved cylinder 15; as shown... Figure 7 As shown, the initial plate model 10 is placed close to the side wall of the cylinder 15, so that the normals of the initial plate model 10 and the side wall of the cylinder 15 coincide. The four arc edges of the initial plate model 10 are projected onto the side wall of the cylinder 15 to generate a single-curved plate model 16; Figure 8 As shown, the output arc line 12 is used to determine the bending direction of the single-curved panel block model 16, and outputs the corner deviation value D between the single-curved panel block model 16 and the initial plate model 10, as well as the midpoint deviation value E of the arc edge.

[0026] The method for partitioning and fitting the free-form hyperboloid curtain wall skin of the present invention first divides the free-form hyperboloid curtain wall skin into several initial panel models 10. Arc edge sampling points 4 are set on the two arc edges 2 of the initial panel models 10. A sampling line is formed by connecting two arc edge sampling points 4. The midpoint of the sampling line is projected onto the initial panel model 10 to generate a sampling projection point 6. A first plane 7 is generated based on the two arc edge sampling points 4 and one sampling projection point 6. The two arc edges 1 of the initial panel model 10 are projected onto the first plane 7 to generate arc projection lines 8 and arc edge projection lines 9. The relative positions on the arc projection lines 8 and arc edge projection lines 9 are calculated. The midpoints of the corresponding nodes are connected to generate the intermediate curve 10. Then, the first plane 7 moves along the two arc edge sampling points 4 and the two arc edges 2 to sample. The distance L from the nodes of the arc projection lines 8 and 9 to the intermediate curve 10 is calculated. The position of the arc edge sampling point 4 when the distance L is the minimum is extracted, and the positions of the arc edge sampling point 4 and the intermediate curve 10 are locked. The two endpoints and the midpoint of the intermediate curve 10 are connected to form a smooth transition arc line 12. The arc line 12 is stretched into a single-curved cylinder 15, so that the normals of the sidewalls of the initial plate model 10 and the cylinder 15 coincide. The four arc edges of the initial plate model 10 are projected onto the cylinder 15. The image is projected onto the side wall of cylinder 15 to form a single-curved panel block model 16. An arc line 12 is output to determine the bending direction of the single-curved panel block model 16, and the corner deviation value and arc edge midpoint deviation value between the single-curved panel block model 16 and the initial plate model 10 are also output. The partitioning fitting method of this invention decomposes the initial plate model 10 of the free hyperboloid into lines, and then further decomposes the lines into points. A coarse and fine combined sampling method is used to improve computational efficiency. A geometric algorithm is used to lock the approximate range of the arc line 12, and fine sampling is performed on the determined sampling area. This greatly improves computational efficiency while ensuring the accuracy of the sampling results. The free hyperboloid... After optimizing the curtain wall skin into a single-curved curtain wall skin, the geometry of the curtain wall becomes simpler, reducing the difficulty of processing and installation, simplifying the construction process, and making subsequent maintenance more convenient. This method also saves construction costs by reducing the required construction equipment, manpower, and materials. Furthermore, it can output deviation data between the single-curved panel block model 16 and the initial panel model 10 before and after surface fitting, and output the pressure arc line 12 to determine the bending direction and processing data of the single-curved panel block model 16, facilitating subsequent manual decision-making and verification. This ensures the construction quality of the free-form hyperboloid curtain wall skin while reducing the difficulty of processing and installation.

[0027] like Figure 5As shown, in step S2, the distance L from each node of the arc projection line 8 and the arc edge projection line 9 to the middle curve 10 is calculated respectively. When the distance L is the smallest, the positions of the arc edge sampling point 4 and the middle curve 10 are locked, and the pressure arc line 12 used to generate the optimal single-curved panel block is generated and extracted; or, the distance L from each node of the arc projection line 8 and the arc edge projection line 9 to the middle curve 10 is calculated. When the sum of all distance values ​​L is the smallest, the positions of the arc edge sampling point 4 and the middle curve 10 are locked, and the pressure arc line 12 used to generate the optimal single-curved panel block is generated and extracted.

[0028] In step S2, the arc edge sampling point 4 moves along the arc edge 2 to sample, determining the initial sampling point of the pressing arc line 12. Two extended sampling points adjacent to the initial sampling point are obtained on the sampling path, one before and one after. The arc segment between the two extended sampling points is further subdivided into more refined equal parts to increase sampling accuracy and perform fine sampling to find the minimum value of distance L. For example, the arc edge 2 is first divided into 100 sampling points, and the initial sampling point is determined to be at the 48th sampling point. Then, within the range of sampling points 47, 48, and 49, another 100 sampling points are subdivided into refined equal parts, and then precise sampling is performed. The advantage of this step is that only 200 calculations are needed to achieve the precise effect of dividing the arc edge 2 into 10,000 equal parts, significantly improving computational efficiency.

[0029] In step S2, a distance standard value is set according to the construction requirements. When the distance value L is less than the distance standard value, the current initial state of the block model meets the conditions for optimization into a single curved block model 16. Otherwise, it is not suitable to be optimized into a single curved block model 16.

[0030] like Figure 8 As shown, in step S3, the corner deviation value D and the arc edge midpoint deviation value E between the single curved panel block model 16 and the initial block model 10 are output. The weighted average is calculated using the multiple output corner deviation values ​​D and multiple midpoint deviation values ​​E, thereby improving the surface fitting accuracy of the single curved panel block model 16.

[0031] like Figure 8As shown, in step S3, based on the characteristics of the project and the installation position of the free hyperboloid curtain wall skin, the standard values ​​for corner deviation and midpoint deviation are set. Based on the standard values ​​for corner deviation and midpoint deviation, it is determined whether the current initial panel model 10 is suitable to be fitted into a single-curved panel model 16. When the corner deviation value D between the current initial panel model 10 and the single-curved panel model 16 is less than the standard value for corner deviation, and the midpoint deviation value E is less than the standard value for midpoint deviation, it is determined that the current initial panel model 10 is suitable to be fitted into a single-curved panel model 16; otherwise, it is not suitable to be fitted into a single-curved panel model 16. This ensures that the optimized hyperboloid curtain wall can meet the architectural appearance design requirements. At the same time, the corner deviation value D and midpoint deviation value E after surface fitting, as well as the subsequent processing data, are output to facilitate subsequent manual decision-making and verification.

[0032] The above description is merely a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the scope of the claims.

Claims

1. A partition fitting method for a free hyperbolic curtain wall skin, characterized in that: Here are the steps: S1: Divide the free hyperbolic curtain wall skin model into a plurality of plate models, extract a single initial plate model, wherein the four arc edges of the initial plate model are two arc edges one and two arc edges two that are arranged opposite to each other, arc edge sampling points are respectively arranged on the two arc edges two, the two arc edge sampling points are connected to form a sampling line, the midpoint of the sampling line is taken as the sampling midpoint, and the sampling midpoint is projected onto the initial plate model to generate a sampling projection point; a first plane is generated according to the two arc edge sampling points and one sampling projection point, and the two arc edges one of the initial plate model are projected onto the first plane to generate an arc projection line and an arc edge projection line; S2: Divide the arc projection line and the arc edge projection line into a number of equal segments, with nodes between two adjacent segments, calculate the midpoint of the line connecting two nodes corresponding to the positions on the arc projection line and the arc edge projection line, connect multiple midpoints to generate an intermediate curve, move two arc edge sampling points synchronously, sample the first plane along the two arc edges along the two arc edge sampling points, calculate the distance value L from the nodes of the arc projection line and the arc edge projection line to the intermediate curve, extract the position of the arc edge sampling point when the distance value L is the smallest, and lock the position of the arc edge sampling point and the intermediate curve; S3: Connect the two endpoints and the midpoint of the middle curve to form a smooth transition arc, form an arc pressure line, extract the arc pressure line and stretch it into a cylinder with a single-curved surface shape; move the initial plate model close to the side wall of the cylinder so that the normal lines of the initial plate model and the side wall of the cylinder coincide, project the four arc edges of the initial plate model to the side wall of the cylinder respectively and generate a single-curved panel block model; output the arc pressure line to determine the bending processing direction of the single-curved panel block model, output the corner point deviation value between the single-curved panel block model and the initial plate model, and the arc edge midpoint deviation value.

2. The partition fitting method of the free hyperbolic curtain wall skin according to claim 1 is characterized in that: In the step S2, the distance values ​​L from each node of the arc projection line and the arc edge projection line to the middle curve are calculated respectively. When the distance value L is the smallest, the arc edge sampling point and the position of the middle curve are locked, and the arc pressure line for generating the optimal single-curved panel block is generated and extracted; Alternatively, the distance value L from each node of the arc projection line and the arc edge projection line to the middle curve is calculated. When the sum of all distance values ​​L is the smallest, the arc edge sampling point and the position of the middle curve are locked to generate and extract the arc pressure line for generating the optimal single-curved panel block.

3. The partition fitting method of the free hyperbolic curtain wall skin according to claim 1 is characterized in that: In step S2, the arc edge sampling point is sampled along the arc edge, a preliminary sampling point of the arc pressure line is determined, two extended sampling points adjacent to the initial sampling point before and after the sampling path are obtained, and the arc segment between the two extended sampling points is further divided into equal parts for sampling to obtain the minimum value of the distance value L.

4. The partition fitting method of the free hyperbolic curtain wall skin according to claim 1 is characterized in that: In step S2, a distance standard value is set. When the distance value L is less than the distance standard value, the plate model in the current initial state meets the conditions for being optimized into a single-curved plate block model. Otherwise, it is not suitable for being optimized into a single-curved plate block model.

5. The partition fitting method of the free hyperbolic curtain wall skin according to claim 1 is characterized in that: In step S3, the corner point deviation values ​​between the single curved panel block model and the initial plate model, as well as the arc edge midpoint deviation values ​​are output, and the weighted average is calculated using the output multiple corner point deviation values ​​D and multiple midpoint deviation values ​​E by weighted average method.

6. The partition fitting method of the free hyperbolic curtain wall skin according to claim 1 is characterized in that: In step S3, a corner point deviation standard value and a midpoint deviation standard value are set. When the corner point deviation value between the current initial plate model and the single curved panel block model is less than the corner point deviation standard value, and the midpoint deviation value is less than the midpoint deviation standard value, it is judged that the current initial plate model is suitable for being fitted into the single curved panel block model, otherwise it is not suitable for being fitted into the single curved panel block model.

Images