Intelligent design optimization system for roof waterproof structure

Abstract

The invention relates to the technical field of design service, in particular to a roof waterproof structure intelligent design optimization system which comprises a node contour extraction module, a contact segment recognition module, a close boundary evaluation module, a disturbance area labeling module and a waterproof configuration reconstruction module. According to the invention, by extracting the direction vector change of the boundary point cloud sequence, the continuity and stability of the turning trend in the contour structure can be accurately obtained, the boundary of the roof structure can be identified, and the contact type of the boundary joint can be determined in combination with the direction change rate distribution of the contact surface, so that the boundary sealing state has quantitative index output; in addition, local positioning information of a fitting path and a structure defect can be provided, a node contour reconstruction path with an adjustment space is provided based on a pressing path continuity judgment and simulation edge shift strategy, and the adaptability of a node structure under the action of complex stress and the configuration optimization capability of the overall waterproof performance of a roof are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of design services, and in particular to an intelligent design optimization system for a roof waterproof structure. Background Art

[0002] The technical field of roof waterproofing structural design services primarily involves the engineering and technical aspects of designing, analyzing, and improving the waterproofing performance of roof structures within construction projects. These include aspects such as roof structure construction, waterproofing material selection, waterproofing joint treatment methods, construction process design, aging and leakage assessment, and maintenance strategy development. This technical area aims to ensure the long-term waterproofing performance of building rooftops. It optimizes the durability and adaptability of roof waterproofing systems through structural design and material combinations. This area represents a key application area at the intersection of architectural design and structural engineering. Specifically, intelligent roof waterproofing structural design optimization systems address technical considerations such as the diversity of structural combinations, material performance differences, joint connection complexity, and construction constraints that must be addressed during the development and evaluation of roof waterproofing structural design proposals. Traditional roof waterproofing structural design optimization utilizes empirical methods or standard-based rule-matching approaches for structural layout and material selection. Designers rely on engineering experience, consult standard drawings, and manually compare material parameters to configure structural proposals. These solutions are then supplemented with 2D CAD drawings or simple modeling tools to present and conduct preliminary analysis.

[0003] Existing technologies mainly rely on the designers' experience and judgment and reference to standard atlases, and mostly use two-dimensional graphic tools to construct solutions. When faced with complex nodes such as air outlets, it is difficult to refine the boundary joint status and contact segment type, resulting in the waterproofing solution being unable to accurately respond to deformation-sensitive areas at structural corners. In particular, when there are wind pressure disturbances or poor sealing on the roof, traditional methods have difficulty marking potential leakage paths and stress concentration areas. In areas where the node contours change frequently or the sealing strip fitting path is discontinuous, effective structural assessment and reconstruction suggestions cannot be achieved through two-dimensional rule matching, affecting the long-term durability and stability of the roof waterproofing. Summary of the Invention

[0004] In order to solve the technical problems that the existing technology is difficult to refine the boundary joint state and contact segment type when facing complex nodes such as air outlets, resulting in the waterproof solution being unable to make accurate responses to deformation-sensitive areas at structural corners, especially when there are wind pressure disturbances or poor sealing on the roof, traditional methods are difficult to mark potential leakage paths and stress concentration areas, and in areas where the node contour changes frequently or the sealing strip fitting path is discontinuous, it is impossible to achieve effective structural evaluation and reconstruction suggestions through two-dimensional rule matching, which affects the long-term durability and stability of roof waterproofing. An embodiment of the present invention provides a roof waterproof structure intelligent design optimization system. The technical solution is as follows: In one aspect, a roof waterproof structure intelligent design optimization system is provided, the system comprising: The node contour extraction module obtains the point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the turning point positions based on the change in the direction vector of the boundary point sequence, marks the continuous directional trend of the contour change, and generates the waterproof configuration node classification results; The contact segment identification module segments the corresponding contact segments on the roof outlet turning boundary and the roof joint surface based on the waterproof configuration node classification result, and generates a contact dense distribution layer; The close boundary assessment module uses the contact dense distribution layer to perform a structural comparison of the close boundary conditions in the contact segments, extract the sealing strip edge fitting path, the frequency of boundary folding, and the total length of the residual unpressed area on the edge of the contact zone, locate and annotate the fitting behavior in the node structure, and generate boundary close boundary status partition information; The disturbance area marking module calls the boundary tightness state partition information, analyzes the angle change trend between the wind pressure projection belt and the dominant direction of node deformation, and determines the location of the section where there is a sign of deformation concentration under the action of standard wind load. The node partition is marked according to the disturbance stress release path, and a wind pressure-induced waterproof weakening area map is generated.

[0005] As a further solution of the present invention, the waterproof configuration node classification results include directional trend categories, turning position indexes, and boundary segment type labels; the contact dense distribution layer includes continuous contact segments, discontinuous transition segments, and non-contact gap segments; the boundary tight state partition information includes fitting path morphology, boundary turning frequency index, and total length value of the unpressed area; the wind pressure-induced waterproof weakened area map includes disturbance-dominated segments, wind pressure projection belt areas, and concentrated deformation annotation belts.

[0006] As a further solution of the present invention, the node outline extraction module includes: The boundary junction identification submodule obtains the junction point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the three-dimensional coordinate values ​​of the boundary points in sequence, calculates the angle value between adjacent vectors based on the change of the direction vector of the coordinate point sequence, calls the angle value sequence to determine the vector turning mutation position, marks the coordinate position of the connection turning point of the vector segment, and obtains the spatial position data of the turning point; The directional trend classification submodule calculates a continuous directional vector sequence based on the turning point spatial position data, counts the angle change values ​​of adjacent vectors in multiple segments of the sequence, groups the boundary segments according to the continuous change characteristics of the angle change values, and divides them into directional consistent segments, directional change segments, and directional tilt segments, and obtains the boundary segment trend type data; The configuration type mapping submodule calls the boundary segment trend type data, detects the trend type tags of multiple paragraphs, matches the corresponding configuration type identifiers according to the differentiated trend types, calculates the trend matching feature values, summarizes them in sequence according to the segment number order, and generates the waterproof configuration node classification results.

[0007] As a further solution of the present invention, the trend matching characteristic value adopts the formula: ; in, Representative The trend matching feature value of the paragraph, Representative In the paragraph The numerical quantification results of trend type markers, Representative The average value in the paragraph, Representative In the paragraph The trend weight factor corresponding to each trend type mark, Representative In the paragraph The configuration identification interference coefficient corresponding to the trend type mark, Representative The difference in the number of configuration type identifiers between the paragraph and the previous paragraph, Representative The total number of trend type tags in the paragraph, Represents the number of the real-time segment, Represents the sequence number of the trend mark in the paragraph.

[0008] As a further solution of the present invention, the contact segment identification module includes: The contact zone segmentation module extracts the coordinates of the interaction area between the roof outlet turning boundary and the roof joint surface based on the classification results of the waterproof configuration nodes. The edge position of the point set is marked by the distance between adjacent points and the change in the boundary trend direction. The point set is divided into multiple segments according to the consistency of the boundary trend to obtain continuous contact zone coordinate data. The direction change judgment submodule calculates the distribution value of the direction vector change rate between the contact zone point series based on the continuous contact zone coordinate data, performs statistical judgment on the fluctuation amplitude and distribution pattern of the direction vector change rate, identifies the segment as direction continuous type, direction sudden change type or direction separation type, and marks the contact feature type as a continuous distribution segment, a discontinuous transition segment or a non-contact gap segment, thereby generating a contact dense distribution layer.

[0009] As a further solution of the present invention, the close boundary assessment module includes: The fitting path extraction submodule extracts the fitting trajectory data of the sealing strip edge on the contact zone in the node structure based on the contact segments in the contact dense distribution layer, performs position mapping on the edge node coordinate set and the segment curve contour, determines whether the node coordinate corresponds to the contour edge or internal location, and obtains the edge fitting trajectory path; The boundary behavior recognition submodule calls the fitted trajectory path, counts the path turning segments and the frequency of turning direction changes based on the position change value in the path node sequence, selects the length of the remaining unpressed segment based on the degree of curvature change on the node path segment, and uses the path turning frequency value and the total length of the remaining segment as boundary behavior determination factors to obtain the boundary behavior recognition degree; The state partitioning submodule extracts the turning frequency value and the total length value of the residual unpressed section on the multi-segment path according to the boundary behavior recognition degree, and makes interval judgments with the full coverage state threshold, gap existence state threshold, and structure eversion state threshold respectively. It classifies and labels the corresponding state types of the path nodes, and aggregates the state type nodes according to the structural area where the nodes are located to generate boundary tight state partition information.

[0010] As a further solution of the present invention, the disturbance area marking module includes: The wind pressure projection analysis submodule calls the boundary tightness state partition information, extracts the node set where gaps exist or the structure is in an outward-turned state, calculates the spatial distribution correction value of the node projection zone based on the node spatial distribution coordinate value and the wind load standard incident direction value, and generates a wind pressure projection direction angle group; The angle trend judgment submodule calls the wind pressure projection direction angle group, calculates the angle between the wind pressure projection direction vector and the node deformation dominant direction vector, obtains the change amplitude value in the angle sequence in the order of node numbers, and generates a disturbance angle concentrated section number set; The weakened zone map submodule calls the disturbance angle concentrated section number set, combines the connection relationship between nodes and the disturbance stress release path direction value, performs partition attribution matching on the nodes in the section number set, and performs regional identification on the contact surface structure node diagram based on the matching results. It evaluates the partition mapping relationship between wind pressure direction, disturbance trend and node attribution interval, and generates a wind pressure induced waterproof weakened zone map.

[0011] As a further solution of the present invention, the spatial distribution correction value adopts the formula: ; in, Representative node pair The spatial distribution correction value between Representative Node and The spatial coordinate value in the X-axis direction, Representative Node and The spatial coordinate value in the Y-axis direction, Representative Node and The spatial coordinate value in the Z-axis direction, Representative node pair The local center elevation value in the standard incident direction of wind load, Representative Node pairs The Euclidean distance between the nodes and the components that form the projection relationship, Representative Nodes and node pairs Forming the arithmetic mean of the projected distances, Representative The total number of associated nodes in the projection band corresponding to the node.

[0012] As a further solution of the present invention, the system further includes a waterproof configuration reconstruction module: The waterproof configuration reconstruction module screens disturbance-induced and insufficiently sealed corners or transition nodes based on the wind pressure-induced waterproof weakened area map, identifies boundary groups with discontinuous press paths in the structural contour, and determines whether boundary release line adjustment can be performed. If the release conditions are met, a rotational edge shift simulation is performed on the sealing strip boundary expansion path to generate a waterproof path node reconstruction map. The waterproof path node reconstruction graph includes releasing feasible boundary groups, simulating edge shift paths, and reconstructing node layout graphs.

[0013] As a further solution of the present invention, the waterproof configuration reconstruction module includes: The abnormal node screening submodule calls the wind pressure induced waterproof weakened area map, extracts the node set corresponding to the disturbance induced mark, combines the state labels marked as gap existence or structural eversion in the tight state partition, and screens the nodes with the characteristics of rotation angle value or curvature mutation value in the node attributes to generate the disturbance abnormal node index set; The boundary path judgment submodule calls the disturbance abnormal node index set, detects the continuity value and path curvature change rate of the fitting path between adjacent nodes according to the structural contour boundary sequence where the node index is located, screens whether the condition that the starting point and ending point of the release line coincide with the boundary endpoint of the sealing strip is met, and generates a set of releasable boundary path segment numbers; The waterproof path reconstruction submodule calls the releasable boundary path segment number set, extracts the sealing strip path node position coordinate values ​​on the corresponding boundary segment, combines the plane direction of the path and the expansion trend vector of the released boundary segment, and sequentially performs the node coordinate rotation angle offset operation around the release starting point and the path continuous segment edge shift translation operation to reconstruct the new sequence of fitting nodes of the pressing path and generate a waterproof path node reconstruction graph.

[0014] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: By extracting the directional vector changes in the boundary point cloud sequence, the continuity and stability of the turning trend in the contour structure can be accurately obtained, the boundary of the roof structure can be identified, and the contact type at the boundary connection can be determined based on the distribution of the directional change rate of the contact surface. This allows the boundary tightness to not only have quantitative indicator outputs, but also provide local positioning information on the fitting path and structural defects. At the same time, the interaction between the tightness state and the wind pressure effect is integrated to construct a distribution map of the deformation concentration area, and to achieve the identification and labeling of the tightness weakening area caused by the disturbance stress. Based on the continuity judgment of the pressing path and the simulated edge shift strategy, a node contour reconstruction path with adjustment space is provided, which improves the adaptability of the node structure under complex stress and the configuration optimization capability of the overall waterproof performance of the roof. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 Schematic diagram of an intelligent design and optimization system for a roof waterproof structure provided by an embodiment of the present invention; Figure 2 Schematic diagram of the system framework of the present invention; Figure 3 This is a flow chart of the node outline extraction module in the present invention; Figure 4 This is a flow chart of the contact fragment identification module in the present invention; Figure 5 This is a flow chart of the close boundary assessment module in the present invention; Figure 6 This is a flow chart of the disturbance area marking module in the present invention; Figure 7 This is a flow chart of the waterproof configuration reconstruction module in the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0018] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0019] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0020] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0021] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0022] The embodiment of the present invention provides a roof waterproof structure intelligent design optimization system, such as Figure 1-2 The schematic diagram of the intelligent design optimization system for roof waterproofing structure shown in FIG. includes: The node contour extraction module obtains the point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the turning point positions based on the change in the direction vector of the boundary point sequence, marks the continuous directional trend of the contour change, and divides the boundary into three categories: uniform segment, turning segment, and bevel segment based on the trend stability, thus generating the waterproof configuration node classification results. The contact segment identification module, based on the waterproof configuration node classification results, segments the corresponding contact segments between the roof outlet turning boundary and the roof interface, extracts the continuous contact bands formed in the segments, and determines whether they are continuous distribution segments, discontinuous transition segments, or non-contact gap segments based on the distribution of the contact surface directional change rate, thus generating a contact dense distribution layer. The close boundary assessment module uses a densely distributed contact layer to perform a structural comparison of the close boundary conditions in the contact segments. It extracts the sealing strip edge fitting path, the frequency of boundary folding, and the total length of the residual unpressed area on the edge of the contact zone. It locates and annotates the fitting behavior in the node structure, determines whether the regional close boundary behavior is full coverage, gap existence, or structural eversion, and generates boundary close boundary status partition information. The disturbance area annotation module uses boundary tightness state partition information to analyze the angle change trend between the wind pressure projection zone on the surface of node blocks with gaps or eversion and the dominant direction of node deformation. It also determines the location of the segment with signs of deformation concentration under the action of standard wind loads, identifies the node partitions according to the disturbance stress release path, and generates a map of wind pressure-induced waterproof weakening areas. The waterproof configuration reconstruction module uses the wind pressure-induced waterproof weakening area map to screen disturbance-induced corners or transition nodes with insufficient sealing, identify boundary groups with discontinuous compression paths in the structural contour, and determine whether boundary release line adjustments can be made. If release conditions are met, a rotational edge shift simulation is performed on the expansion path of the sealing strip boundary to generate a waterproof path node reconstruction map. The classification results of waterproof configuration nodes include directional trend category, turning position index, and boundary segment type label. The contact dense distribution layer includes continuous contact segment, discontinuous transition segment, and non-contact gap segment. The boundary tight state zoning information includes the fitting path morphology, boundary turning frequency index, and total length value of the unpressed area. The wind pressure-induced waterproof weakening area map includes the disturbance-dominated section, wind pressure projection zone, and concentrated deformation annotation zone. The waterproof path node reconstruction diagram includes the released feasible boundary group, simulated edge displacement path, and reconstructed node layout diagram.

[0023] Specifically, if Figure 2 、 3 As shown, the node outline extraction module includes: The boundary junction identification submodule obtains the junction point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the three-dimensional coordinate values ​​of the boundary points in sequence, calculates the angle value between adjacent vectors based on the change of the direction vector of the coordinate point sequence, calls the angle value sequence to determine the vector turning mutation position, marks the coordinate position of the connection turning point of the vector segment, and obtains the spatial position data of the turning point; The 3D point cloud data generated by roof laser scanning or aerial photography is preprocessed. The process includes using statistical filtering to remove outliers whose local density is lower than two standard deviations of the global mean, and using radius filtering to remove noise points whose distance from the neighboring points is greater than 0.1 meters. The processed point cloud data will be limited to the boundary area containing the roof and the air outlet. The point cloud of the area of ​​interest is obtained by coordinate threshold screening. The X-direction coordinate is limited to 5 meters to 10 meters, and the Y-direction is limited to 2 meters to 6 meters. The main direction vector is obtained by performing principal component analysis on the local point set, and the curvature value is calculated. If the curvature of a point exceeds 1.5 times the median curvature of the local neighborhood, the point is determined to be a boundary point. The boundary points are selected and sorted according to the spatial distance of the adjacent points to make the point sequence direction consistent. The direction vector formed between each two points is calculated, and the angle between adjacent vectors is obtained in turn. The angle calculation is based on the projection angle of the 3D vector. For the angle of continuous vectors For points with significant value mutations, if the angle suddenly increases from about 10° in the previous section to more than 60°, for example, if the angle changes to 12°, 15°, and 65° in the point sequence, the point corresponding to 65° can be marked as a turning point; clustering processing is performed on such points to eliminate duplicate or misidentified points. A spatial clustering method is used with a point distance threshold of 0.1 meter and a minimum sample number of 3 as parameters to retain the key turning points of each effective boundary connection component. The points will serve as the basic data for subsequent directional trend analysis. In engineering practice, several ventilation tower openings are set on a flat roof, which form multiple geometric turning points with the boundary components at the connection with the roof. During the processing process, point sequences such as (6.2, 3.7, 2.8), (6.5, 3.9, 2.8), and (6.8, 4.2, 2.8) are extracted. The corresponding angles are measured to be 20°, 25°, and 75°, respectively. The third point is treated as a valid structural connection turning point to obtain the spatial position data of the turning point.

[0024] The directional trend classification submodule calculates the continuous directional vector sequence based on the spatial position data of the turning point, counts the angle change values ​​of adjacent vectors in multiple segments of the sequence, and groups the boundary segments according to the continuous change characteristics of the angle change values, dividing them into directional consistent segments, directional change segments, and directional tilt segments, and obtains the trend type data of the boundary segments; A direction line segment is established for every two adjacent points and its length and direction information are calculated. The direction information is determined by the three-dimensional coordinate difference between the start point and the end point. The turning points A (4.0, 2.0, 1.5) and B (5.0, 3.0, 1.5) are set, and the direction line segment is (1.0, 1.0, 0.0). After normalization, the unit direction vector is obtained for angle analysis. The angle calculation between consecutive segments will be measured in degrees. When judging the trend change, the angle sequence is traversed in a sliding window manner. The window size is set to 3 angles. The standard deviation is calculated within each window. If the standard deviation is less than 10°, it is judged to be consistent in direction. If it exceeds 30°, it is judged to be inconsistent. The angles between 10° and 30° are defined as direction changes, and the angles between 10° and 30° are defined as direction tilt segments. This rule is applicable to the classification of geometric turning trends of continuous boundary structures. The windows with angle values ​​of 8°, 10°, and 9° between the turning points are classified as consistent in direction, the windows with angle values ​​of 35°, 45°, and 50° are defined as direction changes, and the windows with angle values ​​of 15°, 20°, and 25° are defined as direction tilt. After traversing all segments, each segment is assigned a trend type. The classification results include segment 1 as a consistent segment, segment 2 as a changing segment, and segment 3 as an inclined segment. The trend type label will serve as the input basis for subsequent configuration mapping to obtain boundary segment trend type data.

[0025] The configuration type mapping submodule calls the boundary segment trend type data, detects the trend type tags of multiple segments, matches the corresponding configuration type identifiers according to the differentiated trend types, calculates the trend matching eigenvalues, summarizes them in sequence according to the segment number, and generates the waterproof configuration node classification results; Trend matching characteristic value, using the formula: ; in, Representative The trend matching feature value of the paragraph, Representative In the paragraph The numerical quantification results of trend type markers, Representative The average value in the paragraph, Representative In the paragraph The trend weight factor corresponding to each trend type mark, Representative In the paragraph The configuration identification interference coefficient corresponding to the trend type mark, Representative The difference in the number of configuration type identifiers between the paragraph and the previous paragraph, Representative The total number of trend type tags in the paragraph, Represents the number of the real-time segment, Represents the sequence number of the trend mark in the paragraph; The calculation logic of the formula is as follows: based on the quantitative value of the trend type mark, by calculating the difference between each trend mark value and its average value within the paragraph, combined with the trend influence weight for weighted processing, and considering the fluctuations caused by interference factors in the configuration identification process, the uncertainty effect is weakened in the form of square root, the adjusted trend terms are summed, and the difference in the number of configuration identifiers between the current segment and the previous segment is subtracted. The absolute value of the difference is taken to form a quantifiable trend matching feature value, which is used to measure the correlation deviation between the trend distribution and the configuration identifier. The logical core of this value is to integrate the trend amplitude, trend weight and configuration disturbance effect to improve the numerical basis of configuration classification judgment; The trend matching feature value is used to measure the degree of matching between a certain segment of trend data and its configuration identifier. The larger the value, the more the trend distribution deviates from the configuration evolution characteristics. This parameter is an important numerical indicator for judging the effectiveness of the trend in paragraph classification. Parameter meaning and formula calculation derivation process: The paragraph number is paragraph 3, and the number is set to , this paragraph detects 5 types of trend type markers, set their number to According to the quantitative processing standard in the boundary segment trend data monitoring system, the non-numeric trend labels in the original text are mapped and transformed into numerical quantified values. , and its mapping rules are set according to the trend strength level as follows: Very strong trend mark: quantitative value is 9, strong trend mark: quantitative value is 7, medium trend mark: quantitative value is 5, weak trend mark: quantitative value is 3, very weak trend mark: quantitative value is 1; Corresponding to the 5 types of trend markers detected in the third section of this paragraph, each type of label after quantification is: , , , , ; Calculate the average of the trend quantitative value of the third segment: ; Configuration identification interference coefficient The data from the configuration detection sensor is normalized based on the degree of configuration boundary variation. The range is between 0 and 10, with higher values ​​representing greater interference. The data obtained from the test is as follows: , , , , ; Trend Weight Factor The importance assessment model for trend types is based on the mean change rate of the trend affecting the configuration recognition accuracy in the configuration classification. The weight coefficient is set based on the correlation between the change rate of the configuration recognition accuracy and the trend type. The test data is as follows: , , , , ; Configuration type identification difference value The difference in the number of configuration type identifiers between the third and second segments is obtained from the node classification record: The number of configuration identifiers in the third segment = 6, and the number of configuration identifiers in the second segment = 4, so: ; Item 1: ; Item 2: ; Item 3: ; Item 4: ; Item 5: ; Sum calculation: ; Substitute into the formula to calculate: ; The results show that the trend matching characteristic value of the third paragraph is 1.151. The higher the value, the higher the deviation between the strength of the trend mark and the difference in configuration identification. This value will be normalized with the remaining paragraph values ​​in the subsequent classification and sorting, and will serve as one of the criteria for the classification and summary of waterproof configuration nodes.

[0026] Specifically, if Figure 2 、 4 As shown, the contact fragment identification module includes: The contact zone segmentation module extracts the coordinates of the interaction area between the roof outlet turning boundary and the roof joint surface based on the waterproof configuration node classification results. The edge position of the point set is marked by the distance between adjacent points and the change in boundary trend direction. The point set is then segmented into multiple segments based on the consistency of the boundary trend to obtain continuous contact zone coordinate data. Extract the interactive area point set formed by the turning boundary of the roof outlet associated with the configuration node and the roof joint surface. In the point set extraction, the turning point corresponding to the configuration node is selected as the starting point, and the point cloud data is filtered within the range of 0.2 meters before and after it to form a three-dimensional coordinate set of the local interactive area. If the turning point coordinates are set to (5.0, 3.0, 2.8), then the points within the range of X = 4.8 to 5.2 meters and Y = 2.8 to 3.2 meters are retrieved as candidate sets; the Euclidean distance is calculated for adjacent points in the point set. If the distance change between a certain section of continuous points is within the range of ±15% of the average distance and the angle of change in the boundary direction is less than 10°, the boundary trend of this segment is considered consistent, and the point sequence of this segment is marked as a continuous segment. The boundary strike direction angle is calculated by constructing the direction vectors of adjacent points and then calculating the angle between the adjacent vectors. The direction angles are set to 5°, 6°, and 8°, respectively, all within the change threshold. When the angle difference of adjacent vectors in a segment suddenly increases, such as 8° in the previous segment and 35° in the next segment, this point is marked as a trend change point and is segmented as a segment boundary. According to this rule, all point sets are traversed, and the complete joint boundary is segmented into multiple segments according to the trend consistency principle. The coordinates of its continuous points are saved in each segment to obtain the coordinate data of the continuous contact zone.

[0027] The direction change judgment submodule calculates the distribution value of the direction vector change rate between the contact zone point series based on the continuous contact zone coordinate data, performs statistical judgment on the fluctuation amplitude and distribution pattern of the direction vector change rate, identifies the segment as direction continuous, direction sudden or direction separation, and marks the contact feature type as continuous distribution segment, discontinuous transition segment or non-contact gap segment, thereby generating a contact dense distribution layer; Process the point sequence in each segment in turn, construct a point sequence direction vector sequence and calculate the direction change rate for each vector, that is, the angle between two consecutive direction vectors divided by the length of the corresponding line segment to obtain the direction change rate distribution value. Set the angles of three consecutive direction vectors in a certain segment to 10°, 15°, and 12°, and the corresponding lengths to 0.2 meters, 0.2 meters, and 0.2 meters, then the direction change rates are 50° / m, 75° / m, and 60° / m respectively. Statistically calculate the average value and standard deviation of the direction change rate of this segment. If the standard deviation is less than 10° / m and the difference between the maximum and minimum change rates is less than 20° / m, the segment is judged to be directionally continuous. If the standard deviation is greater than 30° / m and there is a mutation point (the change rate exceeds The contact zone is divided into two groups (twice the average value), which is a directional mutation type. If there is a section with less than 3 points in the middle and the direction difference between adjacent points exceeds 50°, it is marked as a directional separation type. In the above judgment classification, each segment is assigned a contact feature label according to the judgment type. The directional continuous segment is marked as a continuous distribution segment, the directional mutation type is marked as a discontinuous transition segment, and the directional separation type is a non-contact gap segment. The direction change rate in the contact zone No. 1 segment is 45° / m, 48° / m, 44° / m, and the standard deviation is 2° / m, which is a directional continuity type. The direction change rate in the No. 2 segment is 45° / m, 110° / m, 60° / m, and the standard deviation is 30° / m, which is a mutation type. A contact dense distribution layer is generated.

[0028] Specifically, if Figure 2 、 5 As shown in the figure, the close boundary assessment module includes: The fitting path extraction submodule extracts the fitting trajectory data of the sealing strip edge on the contact zone in the node structure based on the contact segments in the contact dense distribution layer. It maps the edge node coordinate set to the segment curve contour, determines whether the node coordinates correspond to the contour edge or internal location, and obtains the edge fitting trajectory path. The contact segments classified as "continuous distribution segments" and "intermittent transition segments" are processed to screen out the areas where the sealing strip is fitted. This step extracts the edge point set distributed along the edge of the curve from the contact segment point sequence, uses the contour extraction algorithm to identify the segment boundary morphology, and sets the α-shape shell boundary of the point cloud contour as the segment curve contour. The edge point coordinate extraction uses an α value of 0.15 meters to ensure that the contour curve does not deviate from the actual contact boundary. The node data is read in sequence from the edge node coordinate set in the waterproof configuration node and compared with the proposed curve contour in three-dimensional space. The row position mapping comparison is performed to determine whether the node is within the bounding box of the curve. If the node is within the bounding box, the shortest distance from the node to the contour boundary is determined. If the shortest distance is less than 0.05 meters, it is considered to be aligned with the contour edge. The edge nodes are set to (4.3, 2.9, 2.7), and the points on the curve contour boundary are set to (4.2, 2.9, 2.7). The Euclidean distance between them is 0.1 meters, which exceeds the alignment threshold and is determined to be unaligned. On the contrary, if the distance is 0.03 meters, it is a aligned point. According to this rule, the edge nodes are judged point by point, and the node sequence that meets the alignment conditions is retained to obtain the edge alignment trajectory path.

[0029] The boundary behavior recognition submodule calls the fitted trajectory path, counts the path turning segments and the frequency of turning direction changes based on the position change values ​​in the path node sequence, and selects the length of the remaining unpressed segment based on the degree of curvature change on the node path segment. The path turning frequency value and the total length of the remaining segment are used as boundary behavior determination factors to obtain the boundary behavior recognition degree. Construct continuous line segments in sequence according to the path point sequence and calculate the turning angle of each segment. The turning angle is calculated in units of three points. The current point is set as the midpoint, and the two points before and after it form an angle. If the angle changes by more than 20°, it is recorded as a turning point. Set the path points as (4.0, 2.8, 2.7), (4.1, 2.9, 2.7), (4.2, 3.2, 2.7), then the angle is 28°, which is determined to be a turning point; count the number of turning points in the entire path as the turning frequency value, and calculate the degree of curvature change between the two points of each turning segment. , using a simplified curvature estimation method, the turning angle of each section is divided by the length of the section, the angle is set to 30° and the length is 0.15 meters, then the curvature estimate is 200° / m, if the curvature is greater than 150° / m and the distance between points is less than 0.05 meters, then the section is marked as a residual unpressed section, and its length is recorded. The accumulated residual section length is used as the total unpressed length value; the overall characteristics of the path are determined by the turning frequency and the residual unpressed length. The path turning frequency is set to 5 times, and the total residual unpressed length is 0.35 meters, and the boundary behavior recognition degree is obtained.

[0030] The state partitioning submodule extracts the turning frequency value and the total length of the residual unpressed section on the multi-segment path based on the boundary behavior recognition degree, and makes interval judgments with the full coverage state threshold, gap existence state threshold, and structural eversion state threshold. It classifies and labels the corresponding state types of the path nodes, and aggregates the state type nodes according to the structural area where the nodes are located to generate boundary tight state partition information. Classification is performed according to preset judgment intervals, and three status thresholds are set: the full coverage status threshold is a turning frequency of no more than 3 times and a residual uncompressed length of less than 0.1 meter; the gap existence status threshold is a turning frequency of 3 times and a residual uncompressed length of 0.3 meter; and the structural inversion status threshold is a turning frequency of more than 6 times or a residual uncompressed length of more than 0.3 meter. Conditional judgment is performed on each path. Path A is set with a turning frequency of 4 and a total residual length of 0.15 meters, meeting the gap existence condition. Path B is set with a turning frequency of 7 and a total residual length of 0.4 meters, indicating a structural inversion state. Path status types are aggregated according to the structural region to which the nodes belong. That is, the node states of paths within the same configuration region are merged to form a boundary fit status label corresponding to the structural unit. All node states are output in a partitioned manner. For example, if structural region 1 contains 10 paths, 8 of which are fully covered and 2 have gaps, structural region 1 is labeled "partially gapped," and boundary fit status partition information is generated.

[0031] Specifically, if Figure 2 、 6 As shown, the disturbance area annotation module includes: The wind pressure projection analysis submodule uses boundary tightness state partition information to extract node sets where gaps exist or the structure is in an outward-turned state. Based on the spatial distribution coordinate values ​​of the nodes and the standard incident direction values ​​of the wind load, it calculates the spatial distribution correction values ​​of the node projection zone and generates a wind pressure projection direction angle group. Spatial distribution correction value, using the formula: ; in, Representative node pair The spatial distribution correction value between Representative Node and The spatial coordinate value in the X-axis direction, Representative Node and The spatial coordinate value in the Y-axis direction, Representative Node and The spatial coordinate value in the Z-axis direction, Representative node pair The local center elevation value in the standard incident direction of wind load, Representative Node pairs The Euclidean distance between the nodes and the components that form the projection relationship, Representative Nodes and node pairs Forming the arithmetic mean of the projected distances, Representative The total number of associated nodes in the projection band corresponding to the node; The calculation logic of the formula: Based on the coordinate information of the nodes in three-dimensional space, the difference in the center elevation of the wind load incident direction, and the node dispersion within the wind pressure projection zone, the geometric correlation of the node pair in the wind load projection direction is comprehensively evaluated. The first term describes the horizontal distribution of the node foundation through the Euclidean distance between the two nodes on the XY plane. The second term calculates the difference between the average elevation of the two nodes and the center elevation in the wind incident direction to reflect the inconsistency of the vertical structure. The third term introduces the spatial dispersion deviation of multiple associated nodes that form a wind pressure projection relationship with the node pair to measure the representativeness and stability of the node pair in the projection zone. The three terms together constitute the spatial distribution correction value on the wind pressure action line. The spatial distribution correction value is an important quantitative indicator that reflects the geometric stability of the node pair in the wind pressure projection direction. The larger the value, the higher the degree of dispersion of the node pair in the wind load projection zone and the lower the geometric representativeness. It is used to adjust the influence intensity and spatial weight distribution of the wind pressure direction angle. Parameter meaning and formula calculation derivation process: parameter Represents the spatial coordinate values ​​of node a and node b in the X, Y, and Z directions respectively. After sampling the point cloud of the building structure through laser scanning modeling, it is calculated by the point cloud coordinate extraction module. Based on the spatial distribution point cloud data of the edge nodes of the known structure, the three-dimensional coordinates of the node numbered a are selected in the point cloud. m, m, m, the three-dimensional coordinates of node b are m, m, m; It represents the local center elevation value of node pair ab in the standard incident direction of wind load. After nodes a and b are projected onto the structural contact surface in the direction of wind load, the median value of the local structural surface wind pressure projection elevation is taken as the representative value. According to the laser scanning elevation data of the structural surface and the wind pressure direction calibration, it is obtained. m; The Euclidean distance between the kth associated node that forms a wind pressure projection relationship with node ab and ab is calculated using the three-dimensional coordinate Euclidean distance formula: ; in, is the three-dimensional coordinate of the kth associated node, which is obtained by the intersection detection between the structure surface and the wind direction projection path. The total number of valid nodes in the wind pressure projection zone is , and its three sets of coordinates are: Node 1: m, m, m; Node 2: m, m, m; Node 3: m, m, m; The geometric center coordinates of nodes a and b are: ; ; ; Calculate each value: m; m; m; The average distance is: ; The sum of absolute deviations is: ; The mean deviation term is: ; Compute the first Euclidean distance: ; Second elevation offset value: ; Substituting into the formula: ; The results show that the spatial distribution correction value of the wind pressure projection zone formed by nodes a and b in the direction of wind load incidence is 1.295 meters. This value reflects the geometric quantity of the effective wind pressure projection association path between the node pairs under wind pressure incidence conditions, and provides the basic basis for the subsequent correction of the wind pressure direction angle group distribution weight.

[0032] The angle trend judgment submodule calls the wind pressure projection direction angle group, calculates the angle between the wind pressure projection direction vector and the node deformation dominant direction vector, obtains the change amplitude value in the angle sequence according to the node number order, and generates a disturbance angle concentrated section number set; Each angle in the wind pressure projection direction angle group needs to be converted into a direction vector in two-dimensional space and marked in sequence. The direction vector can be based on the typical angles defined in the wind pressure test or simulation, such as the projection direction divided into 15-degree intervals around the building structure. The dominant deformation direction of each structural node under wind load needs to be determined. From the node displacement output results generated by the structural simulation software, the maximum principal deformation direction of each node under multiple loading conditions is extracted as the dominant direction vector of the node. The dominant deformation of a certain node is set to extend to the northeast direction under the wind pressure loading direction of 30 degrees. The above-mentioned wind pressure projection direction vector is paired with the node dominant deformation direction vector to obtain the angle value. The closer the angle is to 90 degrees, the closer the wind load and node deformation are. The smaller the degree of coupling and the closer the angle is to 0 or 180 degrees, the stronger the coupling is. This step requires processing the nodes in sequence and sorting the results from small to large by node number to obtain a set of angle sequences; the angle change amplitude is calculated by comparing the absolute difference between the angle values ​​of two adjacent nodes in sequence to form an angle change sequence, and then setting the judgment standard value of the angle mutation to 10 degrees. If the angle changes between adjacent nodes in a certain section exceed this standard value, it means that there is a disturbance trend in this section. The node numbers that meet the disturbance standard are aggregated, and the node groups with prominent angle changes in the continuous numbers are marked as concentrated disturbance sections. Continuous angle mutations are set between node numbers 5 and 10 to generate a set of disturbance angle concentrated section numbers.

[0033] The weakened zone map submodule calls the disturbance angle concentrated segment number set, combines the connection relationship between nodes and the disturbance stress release path direction value, and performs partition attribution matching on the nodes in the segment number set. Based on the matching results, regional identification is performed on the contact surface structure node diagram, and the partition mapping relationship between wind pressure direction, disturbance trend and node attribution interval is evaluated to generate a wind pressure-induced waterproof weakened zone map. The graph is traversed based on the node number set and the node connection relationship matrix to construct the disturbance propagation path. For example, the BFS algorithm is used to mark the nodes adjacent to the disturbance number to form a disturbance diffusion map. Then, combined with the disturbance stress release path direction value, which can be obtained from the equivalent stress release path in the structural mechanics simulation, the force line density path formed by the normal direction of the node contact surface and the external force direction is set to calculate the main conduction direction of the disturbance. The corresponding node area affiliation relationship is marked in combination with the angle concentration number to form a node-segment matching table. Then, the disturbance identification area is drawn on the contact structure node diagram (such as the node layout in the structural model built based on CAD) according to the matching relationship. The color annotation of the regional distribution can be set. The nodes with matching numbers are set to be highlighted in red and the remaining areas are gray. The regional identification map is output according to the image coordinate mapping method to generate a wind pressure induced waterproof weakening area map.

[0034] Specifically, if Figure 2 、 7As shown, the waterproof configuration reconstruction module includes: The abnormal node screening submodule calls the wind pressure-induced waterproof weakening zone atlas to extract the node set corresponding to the disturbance-induced marker. Combined with the state labels marked as gap existence or structural eversion in the tight state partition, it screens the nodes with rotation angle value or curvature mutation value characteristics in the node attributes to generate the disturbance abnormal node index set. The wind pressure induced waterproof weakened area map is processed. The map comes from the wind pressure test data on the building structure surface and the simulation results of the calculation model. The setting can be generated with the help of pressure sensing array or CFD modeling. Each node in the map has spatial coordinate attributes and wind pressure response characteristic values. The system first classifies the nodes in the map according to the disturbance induced identification, and selects the regional nodes where the wind pressure change rate is higher than the preset disturbance intensity threshold as the candidate disturbance node set. The system calls the tight state partition map, which is formed by thermal imaging scanning or optical measurement. The nodes are classified according to the states of complete sealing, gap existence, and structural eversion. Among the candidate nodes, the system compares the partition icons The results are annotated and the nodes located in the "gap existence" and "structural overturning" areas are further extracted. This process uses a unified spatial reference coordinate system to reproject and align the two maps to make the node mapping accurate. The system searches for points with structural corner changes or sudden changes in geometric contour curvature from the node attributes. Common turning structural areas such as window corners and the lower edge of door frames have obvious contour turning characteristics. The nodes appear as sudden changes in path direction in the structural boundary line. The system marks the nodes that meet the requirements of strong wind pressure disturbance, abnormal tightness, and sudden changes in geometric characteristics, and records their path index numbers in combination with their connection positions in the structural contour to generate an index set of disturbance abnormality nodes.

[0035] The boundary path judgment submodule calls the disturbance abnormal node index set and detects the continuity value and path curvature change rate of the fitting path between adjacent nodes based on the structural contour boundary sequence where the node index is located. It then screens whether the condition that the starting and ending points of the release line coincide with the boundary endpoint of the sealing strip is met, and generates a set of releasable boundary path segment numbers. It is necessary to judge the context structure of each node in the structural outline boundary path. The structural outline path comes from the architectural design drawings or 3D modeling data. The boundary direction is defined in the order of node numbers. The system extracts the path segment information before and after each abnormal node according to the node index position, and checks the boundary path continuity between it and the adjacent nodes to determine whether there is a drastic change in the path connection direction. In the path structure, the reduction of continuity means that the fitting surface is displaced, broken or mutated. When a window frame boundary node suddenly reverses from clockwise to counterclockwise, it means that there is an abnormal point of compression stress in the structure. The system also needs to Compare the degree of curvature change between the nodes in the path segment. If the direction change rate between the nodes is drastic, it indicates that the path segment has non-negligible geometric disturbance characteristics. The system also needs to retrieve the boundary endpoint information from the sealing strip design drawing to compare whether the abnormal path segment is consistent with the sealing strip endpoint position. If the starting node of the abnormal path segment matches the starting node number of the sealing strip, and the ending node is also near the tail number of the sealing strip, then it can be confirmed that the path segment is a releasable path segment. The path numbers that meet the requirements of drastic fitting path mutation, concentrated curvature change, and corresponding start and end nodes to the sealing strip endpoint are marked and output as a set of releasable boundary path segment numbers.

[0036] The waterproof path reconstruction submodule calls the set of releasable boundary path segment numbers, extracts the coordinate values ​​of the sealing strip path nodes on the corresponding boundary segments, combines the plane direction of the path with the expansion trend vector of the released boundary segment, and sequentially performs node coordinate rotation angle offset operations around the release starting point and path continuous segment edge shifting operations to reconstruct a new sequence of fitting nodes of the pressed path and generate a waterproof path node reconstruction graph. The node coordinate information on each marked path segment is extracted and stored as a spatial point cloud or CAD node with three-dimensional coordinate information. The system uses the starting point of each path segment as a rotation reference to determine the plane direction of the path segment and the expansion direction of the released boundary. The system then performs rotation and offset operations on the path nodes in sequence by calculating the spatial deflection direction of the node around the starting point. In actual operation, for example, for a window corner structure, the original path will cause the local contour of the entire path to expand in a certain direction after being deflected 15 degrees around the point. The system sets a unified shift direction and offset based on the structural shift requirements of the building joint release, and performs node translation operations on the path to ensure that the path is released from the original pressed state to a relatively symmetrical or neutral position of the structure. The system recombines the processed position of each node into a continuous path segment, forming a new fitted path point set. For areas such as window corners and curtain wall joints, their pressed path structure can be restored in this way. After the path segment reconstruction is completed, a waterproof path node reconstruction map is generated.

[0037] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An intelligent design and optimization system for roof waterproofing structure, characterized in that: The system comprises: The node contour extraction module obtains the point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the turning point positions based on the change in the direction vector of the boundary point sequence, marks the continuous directional trend of the contour change, and generates the waterproof configuration node classification results; The contact segment identification module segments the corresponding contact segments on the roof outlet turning boundary and the roof joint surface based on the waterproof configuration node classification result, and generates a contact dense distribution layer; The close boundary assessment module uses the contact dense distribution layer to perform a structural comparison of the close boundary conditions in the contact segments, extract the sealing strip edge fitting path, the frequency of boundary folding, and the total length of the residual unpressed area on the edge of the contact zone, locate and annotate the fitting behavior in the node structure, and generate boundary close boundary status partition information; The disturbance area marking module calls the boundary tightness state partition information, analyzes the angle change trend between the wind pressure projection belt and the dominant direction of node deformation, and determines the location of the section where there is a sign of deformation concentration under the action of standard wind load. The node partition is marked according to the disturbance stress release path, and a wind pressure-induced waterproof weakening area map is generated.

2. The intelligent design and optimization system for roof waterproof structure according to claim 1 is characterized in that: The waterproof configuration node classification results include directional trend categories, turning position indexes, and boundary segment type labels; the contact dense distribution layer includes continuous contact segments, discontinuous transition segments, and non-contact gap segments; the boundary tight state partition information includes the fitting path morphology, boundary turning frequency index, and the total length value of the unpressed area; the wind pressure-induced waterproof weakening area map includes the disturbance-dominated segment, the wind pressure projection zone, and the concentrated deformation annotation zone.

3. The intelligent design and optimization system for roof waterproof structure according to claim 1, characterized in that: The node outline extraction module includes: The boundary junction identification submodule obtains the junction point cloud coordinates of the boundary connection components between the roof outlet and the main roof body, extracts the three-dimensional coordinate values ​​of the boundary points in sequence, calculates the angle value between adjacent vectors based on the change of the direction vector of the coordinate point sequence, calls the angle value sequence to determine the vector turning mutation position, marks the coordinate position of the connection turning point of the vector segment, and obtains the spatial position data of the turning point; The directional trend classification submodule calculates a continuous directional vector sequence based on the turning point spatial position data, counts the angle change values ​​of adjacent vectors in multiple segments of the sequence, groups the boundary segments according to the continuous change characteristics of the angle change values, and divides them into directional consistent segments, directional change segments, and directional tilt segments, and obtains the boundary segment trend type data; The configuration type mapping submodule calls the boundary segment trend type data, detects the trend type tags of multiple paragraphs, matches the corresponding configuration type identifiers according to the differentiated trend types, calculates the trend matching feature values, summarizes them in sequence according to the segment number order, and generates the waterproof configuration node classification results.

4. The intelligent design and optimization system for roof waterproof structure according to claim 3 is characterized in that: The trend matching characteristic value adopts the formula: ； in, Representative The trend matching feature value of the paragraph, Representative In the paragraph The numerical quantification results of trend type markers, Representative The average value in the paragraph, Representative In the paragraph The trend weight factor corresponding to each trend type mark, Representative In the paragraph The configuration identification interference coefficient corresponding to the trend type mark, Representative The difference in the number of configuration type identifiers between the paragraph and the previous paragraph, Representative The total number of trend type tags in the paragraph, Represents the number of the real-time segment, Represents the sequence number of the trend mark in the paragraph.

5. The intelligent design and optimization system for roof waterproof structure according to claim 3, characterized in that: The contact segment identification module includes: The contact zone segmentation module extracts the coordinates of the interaction area between the roof outlet turning boundary and the roof joint surface based on the classification results of the waterproof configuration nodes. The edge position of the point set is marked by the distance between adjacent points and the change in the boundary trend direction. The point set is divided into multiple segments according to the consistency of the boundary trend to obtain continuous contact zone coordinate data. The direction change judgment submodule calculates the distribution value of the direction vector change rate between the contact zone point series based on the continuous contact zone coordinate data, performs statistical judgment on the fluctuation amplitude and distribution pattern of the direction vector change rate, identifies the segment as direction continuous type, direction sudden change type or direction separation type, and marks the contact feature type as a continuous distribution segment, a discontinuous transition segment or a non-contact gap segment, thereby generating a contact dense distribution layer.

6. The intelligent design and optimization system for roof waterproof structure according to claim 5, characterized in that: The close boundary assessment module includes: The fitting path extraction submodule extracts the fitting trajectory data of the sealing strip edge on the contact zone in the node structure based on the contact segments in the contact dense distribution layer, performs position mapping on the edge node coordinate set and the segment curve contour, determines whether the node coordinate corresponds to the contour edge or internal location, and obtains the edge fitting trajectory path; The boundary behavior recognition submodule calls the fitted trajectory path, counts the path turning segments and the frequency of turning direction changes based on the position change value in the path node sequence, selects the length of the remaining unpressed segment based on the degree of curvature change on the node path segment, and uses the path turning frequency value and the total length of the remaining segment as boundary behavior determination factors to obtain the boundary behavior recognition degree; The state partitioning submodule extracts the turning frequency value and the total length value of the residual unpressed section on the multi-segment path according to the boundary behavior recognition degree, and makes interval judgments with the full coverage state threshold, gap existence state threshold, and structure eversion state threshold respectively. It classifies and labels the corresponding state types of the path nodes, and aggregates the state type nodes according to the structural area where the nodes are located to generate boundary tight state partition information.

7. The intelligent design and optimization system for roof waterproof structure according to claim 6, characterized in that: The disturbance area marking module includes: The wind pressure projection analysis submodule calls the boundary tightness state partition information, extracts the node set where gaps exist or the structure is in an outward-turned state, calculates the spatial distribution correction value of the node projection zone based on the node spatial distribution coordinate value and the wind load standard incident direction value, and generates a wind pressure projection direction angle group; The angle trend judgment submodule calls the wind pressure projection direction angle group, calculates the angle between the wind pressure projection direction vector and the node deformation dominant direction vector, obtains the change amplitude value in the angle sequence in the order of node numbers, and generates a disturbance angle concentrated section number set; The weakened zone map submodule calls the disturbance angle concentrated section number set, combines the connection relationship between nodes and the disturbance stress release path direction value, performs partition attribution matching on the nodes in the section number set, and performs regional identification on the contact surface structure node diagram based on the matching results. It evaluates the partition mapping relationship between wind pressure direction, disturbance trend and node attribution interval, and generates a wind pressure induced waterproof weakened zone map.

8. The intelligent design and optimization system for roof waterproof structure according to claim 7, characterized in that: The spatial distribution correction value adopts the formula: ； in, Representative node pair The spatial distribution correction value between Representative Node and The spatial coordinate value in the X-axis direction, Representative Node and The spatial coordinate value in the Y-axis direction, Representative Node and The spatial coordinate value in the Z-axis direction, Representative node pair The local center elevation value in the standard incident direction of wind load, Representative Node pairs The Euclidean distance between the nodes and the components that form the projection relationship, Representative Nodes and node pairs Forming the arithmetic mean of the projected distances, Representative The total number of associated nodes in the projection band corresponding to the node.

9. The intelligent design and optimization system for roof waterproof structure according to claim 1, characterized in that: The system further includes a waterproof configuration reconstruction module: The waterproof configuration reconstruction module screens disturbance-induced and insufficiently sealed corners or transition nodes based on the wind pressure-induced waterproof weakened area map, identifies boundary groups with discontinuous press paths in the structural contour, and determines whether boundary release line adjustment can be performed. If the release conditions are met, a rotational edge shift simulation is performed on the sealing strip boundary expansion path to generate a waterproof path node reconstruction map. The waterproof path node reconstruction graph includes releasing feasible boundary groups, simulating edge shift paths, and reconstructing node layout graphs.

10. The intelligent design and optimization system for roof waterproof structure according to claim 9, characterized in that: The waterproof configuration reconstruction module includes: The abnormal node screening submodule calls the wind pressure induced waterproof weakened area map, extracts the node set corresponding to the disturbance induced mark, combines the state labels marked as gap existence or structural eversion in the tight state partition, and screens the nodes with the characteristics of rotation angle value or curvature mutation value in the node attributes to generate the disturbance abnormal node index set; The boundary path judgment submodule calls the disturbance abnormal node index set, detects the continuity value and path curvature change rate of the fitting path between adjacent nodes according to the structural contour boundary sequence where the node index is located, screens whether the condition that the starting point and ending point of the release line coincide with the boundary endpoint of the sealing strip is met, and generates a set of releasable boundary path segment numbers; The waterproof path reconstruction submodule calls the releasable boundary path segment number set, extracts the sealing strip path node position coordinate values ​​on the corresponding boundary segment, combines the plane direction of the path and the expansion trend vector of the released boundary segment, and sequentially performs the node coordinate rotation angle offset operation around the release starting point and the path continuous segment edge shift translation operation to reconstruct the new sequence of fitting nodes of the pressing path and generate a waterproof path node reconstruction graph.

Images