A 3D parametric design method for water conservancy projects based on deep learning

Abstract

The present invention discloses a three-dimensional parametric design method for water conservancy projects based on deep learning, which relates to the field of architectural design technology, including: obtaining a target building material strength model; obtaining a non-filled structure and a structure to be filled; obtaining at least one non-filled block set; numbering the non-filled block set; designing the construction structure of the non-filled blocks in the non-filled block set to obtain a non-filled block construction structure; verifying the non-filled block construction structure to obtain a non-filled block target structure; replacing the non-filled block target structure with the corresponding non-filled block in the water conservancy 3D model to obtain a target parameter design scheme. By designing the construction structure of the non-filled blocks in the non-filled block set, it is not necessary to fill all of them. As a result, the non-filled block construction structure can bear the forces acting on it, while using as little material as possible, thereby achieving the effect of cost saving.

Description

Technical Field

[0001] The present invention relates to the field of architectural design technology, and in particular to a three-dimensional parametric design method for water conservancy projects based on deep learning. Background Art

[0002] 3D parametric design of water conservancy projects is an emerging technology that utilizes modern computer-aided design (CAD) technology and parametric modeling methods for water conservancy project design. It combines traditional water conservancy project design methods with advanced 3D modeling technology, significantly improving design efficiency and quality.

[0003] The overall external outline of a water conservancy project is easy to determine based on the terrain, but its internal structure can be solid or non-solid. For the non-solid part, the thickness of its thin wall needs to be determined. Otherwise, using thicker or thinner thickness for structural design will either lead to excessive use of materials or insufficient structural bearing capacity. Summary of the Invention

[0004] In order to solve the above technical problems, a three-dimensional parametric design method for water conservancy projects based on deep learning is provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above objects, the technical solution adopted by the present invention is:

[0006] A deep learning-based 3D parametric design method for water conservancy projects, including:

[0007] Obtaining parameters of a water conservancy 3D model to be designed, and obtaining target building materials for constructing the water conservancy 3D model, wherein the water conservancy 3D model is a water conservancy contour model stored in a computer;

[0008] Conduct physical property analysis on target building materials to obtain the target building material strength model;

[0009] Perform structural recognition on the water conservancy 3D model to obtain the non-filled structure and the structure to be filled, where both the non-filled block and the structure to be filled are solid structures;

[0010] Splitting the non-filled structure into at least one non-filled block, and splitting the to-be-filled structure into at least one to-be-filled block, wherein the splitting is performed using at least one plane perpendicular to the x-axis, the y-axis, or the z-axis, wherein the distances between adjacent planes along the same direction are equal, the heights of the non-filled block and the to-be-filled block are consistent, the x-axis, the y-axis, and the z-axis are axes of a three-dimensional coordinate system, and the bottom surface of the water conservancy 3D model is perpendicular to the x-axis;

[0011] Pairing the non-filled blocks to obtain at least one non-filled block set;

[0012] Numbering the non-filling block set based on a positional relationship of the non-filling blocks in the non-filling block set;

[0013] According to the numbering of the non-filled block sets from small to large and based on the target building material strength model, the construction structure of the non-filled blocks in the non-filled block sets is designed to obtain the non-filled block construction structure;

[0014] Verify the non-filled block construction structure according to the numbers of the non-filled block sets from small to large, and based on the verification results, modify the non-filled block construction structure to obtain the non-filled block target structure;

[0015] The non-filled block target structure is replaced with the corresponding non-filled block in the water conservancy 3D model to obtain the target parameter design scheme.

[0016] Preferably, the physical property analysis of the target building material to obtain the target building material strength model comprises the following steps:

[0017] Get the height of the non-filled block as the feature height;

[0018] Obtaining an area value range of a cross section of the non-filled block, dividing the area value range of the cross section of the non-filled block at equal intervals to obtain at least one area point, wherein the cross section of the non-filled block is a plane parallel to a plane perpendicular to the x-axis;

[0019] A sample cube is formed, satisfying that the height of the sample cube is equal to the characteristic height and the bottom area of ​​the sample cube is equal to the value at the area point;

[0020] Apply uniform pressure to the top of the sample cube in a direction parallel to the height of the sample cube to obtain the maximum pressure that the sample cube can withstand as the bearing force;

[0021] The top of the sample cube is pulled uniformly, with the pulling direction parallel to the height of the sample cube, and the maximum pulling force that the sample cube can withstand is obtained as the tensile force;

[0022] The area point value and the load-bearing capacity are paired and fitted to obtain the load-bearing fitting function, where the area point value is the independent variable and the load-bearing capacity is the dependent variable;

[0023] The area point value and the tensile force are paired and fitted to obtain the tensile fitting function, where the area point value is the independent variable and the tensile force is the dependent variable;

[0024] The load-bearing fitting function and the tensile fitting function are used as the target building material strength model.

[0025] Preferably, the structural recognition of the water conservancy 3D model to obtain the unfilled structure and the structure to be filled comprises the following steps:

[0026] Segment the water conservancy 3D model to obtain at least one 3D block;

[0027] Draw an inscribed circle inside the bottom surface of the 3D block. The inscribed circle is the largest circle tangent to the bottom edge of the 3D block.

[0028] The area of ​​the inscribed circle is divided by the bottom area of ​​the 3D block to get the inscribed circle ratio;

[0029] When the proportion of the inscribed circle is greater than a preset ratio, the 3D block is treated as a non-filled part; otherwise, the 3D block is treated as a to-be-filled part, wherein the preset ratio is set based on historical data;

[0030] At least one non-filled portion is aggregated into a non-filled structure, and at least one to-be-filled portion is aggregated into a to-be-filled structure.

[0031] Preferably, the pairing of the non-filled blocks to obtain at least one non-filled block set comprises the following steps:

[0032] The non-filled blocks with coplanar bottom surfaces are aggregated to form a non-filled block set, and the plane where the bottom surfaces of the non-filled blocks in the non-filled block set are located is used as the feature plane.

[0033] Preferably, numbering the non-filled block set based on the positional relationship of the non-filled blocks in the non-filled block set comprises the following steps:

[0034] Calculate the distance between the characteristic plane of the non-filled block set and the bottom surface of the water conservancy 3D model to obtain the characteristic distance;

[0035] The non-filled block sets are numbered from small to large according to the feature distance.

[0036] Preferably, the step of designing the building structure of the non-filled blocks in the non-filled block set to obtain the non-filled block building structure comprises the following steps:

[0037] At least one sampling point is evenly selected on the bottom surface of the non-filled block, and at least one sampling straight line is drawn through the sampling point, wherein the sampling straight line is perpendicular to the bottom surface of the non-filled block;

[0038] Obtain the non-filled block that intersects with the sampling point as the feature non-filled block;

[0039] Obtain the block to be filled that intersects with the sampling point as the feature block to be filled;

[0040] The characteristic non-filling block located above the non-filling block is used as the first non-filling block, and the characteristic non-filling block located below the non-filling block is used as the second non-filling block;

[0041] The characteristic block to be filled located above the non-filled block is used as the first block to be filled, and the characteristic block to be filled located below the non-filled block is used as the second block to be filled;

[0042] Calculating the upper weight limits of the first non-filled block and the first block to be filled, and summing them up to obtain a first characteristic weight;

[0043] Get the length of the bottom edge of the non-filled block as the characteristic length;

[0044] Substitute the characteristic weight into the load-bearing fitting function and inversely solve for the identification area;

[0045] The marked area is divided by the characteristic length to obtain the characteristic thickness;

[0046] A non-filled block building structure is formed, wherein the surface of the non-filled block building structure is consistent with the surface of the non-filled block, the non-filled block building structure is a hollow structure, and the thickness of the wall of the cavity of the non-filled block building structure is a characteristic thickness.

[0047] Preferably, the calculation of the upper weight limit of the first non-filled block and the first block to be filled comprises the following steps:

[0048] Fitting the surface of the first non-filled block to obtain a first non-filled fitting function;

[0049] Integrate within the area enclosed by the first non-filled fitting function to obtain the volume of the first non-filled block;

[0050] Multiplying the density and gravitational acceleration of the target building material by the volume of the first non-filled block to obtain the upper weight limit of the first non-filled block;

[0051] Fitting the surface of the first block to be filled to obtain a first fitting function to be filled;

[0052] Integrate the area enclosed by the first fitting function to be filled to obtain the volume of the first block to be filled;

[0053] The density and gravitational acceleration of the target building material are multiplied by the volume of the first block to be filled to obtain the upper limit of the weight of the first block to be filled.

[0054] Preferably, the verification of the non-filled block construction structure comprises the following steps:

[0055] Obtaining a non-filled block building structure corresponding to the second non-filled block as a target non-filled block building structure;

[0056] Calculate the upper limit of the weight of the target non-filled block construction structure and the second to-be-filled block, and add them up to obtain a second characteristic weight;

[0057] Substitute the second characteristic weight into the tensile fitting function and inversely solve to obtain the target thickness;

[0058] Based on the target thickness, the characteristic thickness is checked. When the characteristic thickness is less than the target thickness, the non-filled block construction structure needs to be corrected. Otherwise, the non-filled block construction structure does not need to be corrected.

[0059] Preferably, the calculation of the upper limit of the weight of the target non-filled block building structure and the second block to be filled comprises the following steps:

[0060] Fitting the surface of the target non-filled block building structure to obtain a second non-filled fitting function;

[0061] Performing surface integral on the second non-filled fitting function to obtain the surface area of ​​the target non-filled block building structure;

[0062] The surface area of ​​the target non-filled block building structure is multiplied by the characteristic thickness to obtain the volume of the target non-filled block building structure;

[0063] Multiply the density and gravitational acceleration of the target building material by the volume of the target non-filled block building structure to obtain the upper weight limit of the target non-filled block building structure;

[0064] Fitting the surface of the second block to be filled to obtain a second fitting function to be filled;

[0065] Integrate the area enclosed by the second fitting function to be filled to obtain the volume of the second block to be filled;

[0066] The density and gravitational acceleration of the target building material are multiplied by the volume of the second block to be filled to obtain the upper weight limit of the second block to be filled.

[0067] Preferably, the step of modifying the non-filled block construction structure to obtain the non-filled block target structure comprises the following steps:

[0068] During correction, the characteristic thickness of the structure built in the non-filled block is updated using the target thickness.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] By obtaining a target building material strength model, obtaining a non-filled structure and a structure to be filled, designing the construction structure of the non-filled blocks in the non-filled block set, and modifying the non-filled block construction structure, the non-filled structure can be designed according to the strength properties of the target building material, so that the non-filled structure does not need to be fully filled during design. Therefore, when it is officially used, the non-filled block construction structure can bear the force acting on it while using as little material as possible, thereby achieving the effect of cost saving. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Schematic diagram of the process of the three-dimensional parametric design method of water conservancy projects based on deep learning of the present invention;

[0072] Figure 2 This is a flow chart of the present invention for analyzing the physical properties of a target building material to obtain a strength model of the target building material;

[0073] Figure 3 This is a schematic diagram of the process of performing structural recognition on a water conservancy 3D model to obtain a non-filled structure and a structure to be filled;

[0074] Figure 4 Schematic diagram of a process of numbering a non-filled block set based on the positional relationship of the non-filled blocks in the non-filled block set according to the present invention;

[0075] Figure 5 A schematic diagram of a process for designing a construction structure of a non-filled block in a non-filled block set according to the present invention to obtain the non-filled block construction structure;

[0076] Figure 6 Schematic diagram of a flow chart for calculating the upper weight limit of a first non-filled block and a first block to be filled according to the present invention;

[0077] Figure 7 A schematic diagram of a process for verifying a non-filled block construction structure according to the present invention;

[0078] Figure 8 Schematic diagram of a flow chart for calculating the weight upper limit of the target non-filled block construction structure and the second to-be-filled block according to the present invention. DETAILED DESCRIPTION

[0079] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are merely examples, and those skilled in the art may conceive of other obvious variations.

[0080] Reference Figure 1 As shown in FIG, a three-dimensional parametric design method for water conservancy projects based on deep learning includes:

[0081] Obtaining parameters of a water conservancy 3D model to be designed, and obtaining target building materials for constructing the water conservancy 3D model, wherein the water conservancy 3D model is a water conservancy contour model stored in a computer;

[0082] Conduct physical property analysis on target building materials to obtain the target building material strength model;

[0083] Perform structural recognition on the water conservancy 3D model to obtain the non-filled structure and the structure to be filled, where both the non-filled block and the structure to be filled are solid structures;

[0084] Splitting the non-filled structure into at least one non-filled block, and splitting the to-be-filled structure into at least one to-be-filled block, wherein the splitting is performed using at least one plane perpendicular to the x-axis, the y-axis, or the z-axis, wherein the distances between adjacent planes along the same direction are equal, the heights of the non-filled block and the to-be-filled block are consistent, the x-axis, the y-axis, and the z-axis are axes of a three-dimensional coordinate system, and the bottom surface of the water conservancy 3D model is perpendicular to the x-axis;

[0085] Pairing the non-filled blocks to obtain at least one non-filled block set;

[0086] Numbering the non-filling block set based on a positional relationship of the non-filling blocks in the non-filling block set;

[0087] According to the numbering of the non-filled block sets from small to large and based on the target building material strength model, the construction structure of the non-filled blocks in the non-filled block sets is designed to obtain the non-filled block construction structure;

[0088] Verify the non-filled block construction structure according to the numbers of the non-filled block sets from small to large, and based on the verification results, modify the non-filled block construction structure to obtain the non-filled block target structure;

[0089] The non-filled block target structure is replaced with the corresponding non-filled block in the water conservancy 3D model to obtain the target parameter design scheme.

[0090] In common 3D design, the designed object is usually fully filled, which inevitably takes more time than the case of non-full filling. However, the reason for using the full filling method for design is that after the hydraulic 3D model is designed and formed, every part of it will be subjected to the pressure of the weight of its top. At the same time, considering the seismic resistance, the part below the simulated vibration position has no support, and the weight of this part will produce a downward pulling force on the simulated vibration position. Therefore, if the non-filling method is used for design, if the building structure is not designed properly, it may not be able to withstand the two forces mentioned above. Therefore, in order to avoid this situation, the general design adopts the full filling method.

[0091] In this solution, a series of algorithms are used to design the structure so that the target parameter design solution can withstand the two forces mentioned above;

[0092] When segmenting unfilled structures and structures to be filled, the planes used are planes perpendicular to the x-axis, the y-axis, or the z-axis, and the spacing between adjacent planes in the same direction is equal, that is, the water conservancy 3D model is segmented using planes from three directions;

[0093] The three-dimensional coordinate system is used to coordinate model the water conservancy 3D model;

[0094] Here the contours of the unfilled structure and the structure to be filled are easy to determine, because the unfilled structure is a cavity used for some operations in water conservancy projects. The only thing that needs to be determined is the thickness of the thin wall of the cavity, which requires further calculation.

[0095] Reference Figure 2 As shown, performing physical property analysis on the target building material and obtaining the target building material strength model includes the following steps:

[0096] Get the height of the non-filled block as the feature height;

[0097] Obtaining an area value range of a cross section of the non-filled block, dividing the area value range of the cross section of the non-filled block at equal intervals to obtain at least one area point, wherein the cross section of the non-filled block is a plane parallel to a plane perpendicular to the x-axis;

[0098] A sample cube is formed, satisfying that the height of the sample cube is equal to the characteristic height and the bottom area of ​​the sample cube is equal to the value at the area point;

[0099] Apply uniform pressure to the top of the sample cube in a direction parallel to the height of the sample cube to obtain the maximum pressure that the sample cube can withstand as the bearing force;

[0100] The top of the sample cube is pulled uniformly, with the pulling direction parallel to the height of the sample cube, and the maximum pulling force that the sample cube can withstand is obtained as the tensile force;

[0101] The area point value and the load-bearing capacity are paired and fitted to obtain the load-bearing fitting function, where the area point value is the independent variable and the load-bearing capacity is the dependent variable;

[0102] The area point value and the tensile force are paired and fitted to obtain the tensile fitting function, where the area point value is the independent variable and the tensile force is the dependent variable;

[0103] The load-bearing fitting function and the tensile fitting function are used as the target building material strength model.

[0104] Here, the characteristic height is the height of the non-filled block, in order to reduce the factors that are different from the non-filled block in the target building material strength model. Therefore, the force of the non-filled block can be directly calculated through the target building material strength model, and then the construction structure design of the non-filled block can be carried out.

[0105] Reference Figure 3 As shown, performing structural recognition on a water conservancy 3D model to obtain a non-filled structure and a structure to be filled includes the following steps:

[0106] Segment the water conservancy 3D model to obtain at least one 3D block;

[0107] Draw an inscribed circle inside the bottom surface of the 3D block. The inscribed circle is the largest circle tangent to the bottom edge of the 3D block.

[0108] The area of ​​the inscribed circle is divided by the bottom area of ​​the 3D block to get the inscribed circle ratio;

[0109] When the proportion of the inscribed circle is greater than a preset ratio, the 3D block is treated as a non-filled part; otherwise, the 3D block is treated as a to-be-filled part, wherein the preset ratio is set based on historical data;

[0110] At least one non-filled portion is aggregated into a non-filled structure, and at least one to-be-filled portion is aggregated into a to-be-filled structure.

[0111] The preset ratio is based on existing experience to distinguish whether the 3D blocks in the historical data are suitable for filling design. According to the distinguished structure, the ratio of the inscribed circle is calculated. According to whether the 3D block corresponding to the inscribed circle ratio is suitable for non-filling design, a critical value of the inscribed circle ratio is formed and used as the preset ratio. In future designs, the preset ratio formed here will be used for judgment.

[0112] Because the 3D block is a very small block, even if there is a difference between its top and bottom, the difference is very small. Therefore, only the bottom surface can be considered. According to the above description, the structure to be filled is a structure in which the area of ​​the inscribed circle occupies a smaller proportion of the bottom surface of the structure to be filled, that is, a structure with a relatively flat bottom surface, such as a bottom surface that is almost a line. It is easy to know that if such a structure is designed to be hollow, the wall will be very thin and its supporting force will be insufficient. Therefore, the design adopted for the structure to be filled is a full filling design, while the non-filled structure is the opposite. Its bottom surface is similar to a shape that is relatively wide on all sides, such as a circle or a square. Therefore, its wall can be made thicker, so that it can support the gravity above itself and the falling force below itself. Therefore, the non-filled structure can be designed to be hollow.

[0113] Pairing the non-filled blocks to obtain at least one non-filled block set includes the following steps:

[0114] The non-filled blocks with coplanar bottom surfaces are aggregated to form a non-filled block set, and the plane where the bottom surfaces of the non-filled blocks in the non-filled block set are located is used as the feature plane.

[0115] The purpose of pairing is to synchronously consider the non-filled blocks at the same height in the water conservancy 3D model.

[0116] Reference Figure 4 As shown, numbering the non-filling block set based on the positional relationship of the non-filling blocks in the non-filling block set includes the following steps:

[0117] Calculate the distance between the characteristic plane of the non-filled block set and the bottom surface of the water conservancy 3D model to obtain the characteristic distance;

[0118] The non-filled block sets are numbered from small to large according to the feature distance.

[0119] The purpose of numbering is to enable the design of the building structure to be carried out in a certain order, thereby making the design process more organized.

[0120] Reference Figure 5 As shown, designing the building structure of the non-filled blocks in the non-filled block set to obtain the non-filled block building structure includes the following steps:

[0121] At least one sampling point is evenly selected on the bottom surface of the non-filled block, and at least one sampling straight line is drawn through the sampling point, wherein the sampling straight line is perpendicular to the bottom surface of the non-filled block;

[0122] Obtain the non-filled block that intersects with the sampling point as the feature non-filled block;

[0123] Obtain the block to be filled that intersects with the sampling point as the feature block to be filled;

[0124] The characteristic non-filling block located above the non-filling block is used as the first non-filling block, and the characteristic non-filling block located below the non-filling block is used as the second non-filling block;

[0125] The characteristic block to be filled located above the non-filled block is used as the first block to be filled, and the characteristic block to be filled located below the non-filled block is used as the second block to be filled;

[0126] Calculating the upper weight limits of the first non-filled block and the first block to be filled, and summing them up to obtain a first characteristic weight;

[0127] Get the length of the bottom edge of the non-filled block as the characteristic length;

[0128] Substitute the characteristic weight into the load-bearing fitting function and inversely solve for the identification area;

[0129] The marked area is divided by the characteristic length to obtain the characteristic thickness;

[0130] A non-filled block building structure is formed, wherein the surface of the non-filled block building structure is consistent with the surface of the non-filled block, the non-filled block building structure is a hollow structure, and the thickness of the wall of the cavity of the non-filled block building structure is a characteristic thickness.

[0131] Get the length of the bottom edge of the non-filled block as the characteristic length. The characteristic length can be easily obtained based on the method of calculating length in calculus. As long as the equation of the bottom edge of the non-filled block is known, the equation can be obtained by fitting, which will not be described in detail.

[0132] When designing the construction structure of the non-filled block, two issues need to be considered. The first is whether the non-filled block can support the gravity of the part above it, and the second is whether the non-filled block can withstand the downward force of the part below it when the hydraulic 3D model is picked up. Here, the structure is designed to meet the first point first, and then it will be modified to meet the second point during subsequent verification.

[0133] When designing the first point: consider the first characteristic weight of the part above the non-filled block. The part above the non-filled block refers to the part above the vertical area of ​​the non-filled block. The part above the non-filled block does not affect the non-filled block. The calculation of the first characteristic weight considers the part above the non-filled block as solid, which is greater than the actual pressure on the non-filled block. Therefore, such a design must meet the requirements. Therefore, according to the load-bearing fitting function, the identification area required to support the first characteristic weight is determined. Here, according to common sense, the characteristic thickness of the wall used for support is relatively small. Therefore, it can be approximately considered that the characteristic thickness multiplied by the characteristic length is equal to the identification area, and then using this relationship, the characteristic thickness can be obtained.

[0134] Reference Figure 6 As shown, calculating the upper weight limit of the first non-filled block and the first block to be filled includes the following steps:

[0135] Fitting the surface of the first non-filled block to obtain a first non-filled fitting function;

[0136] Integrate within the area enclosed by the first non-filled fitting function to obtain the volume of the first non-filled block;

[0137] Multiplying the density and gravitational acceleration of the target building material by the volume of the first non-filled block to obtain the upper weight limit of the first non-filled block;

[0138] Fitting the surface of the first block to be filled to obtain a first fitting function to be filled;

[0139] Integrate the area enclosed by the first fitting function to be filled to obtain the volume of the first block to be filled;

[0140] The density and gravitational acceleration of the target building material are multiplied by the volume of the first block to be filled to obtain the upper limit of the weight of the first block to be filled.

[0141] The volume calculation and surface area calculation here are both performed through the method of calculus. As long as the equation of the surface is known, the corresponding calculation can be performed. The equation of the surface is obtained by taking marked points on the surface and fitting the marked points.

[0142] Reference Figure 7 As shown, the verification of the non-filled block construction structure includes the following steps:

[0143] Obtaining a non-filled block building structure corresponding to the second non-filled block as a target non-filled block building structure;

[0144] Calculate the upper limit of the weight of the target non-filled block construction structure and the second to-be-filled block, and add them up to obtain a second characteristic weight;

[0145] Substitute the second characteristic weight into the tensile fitting function and inversely solve to obtain the target thickness;

[0146] Based on the target thickness, the characteristic thickness is checked. When the characteristic thickness is less than the target thickness, the non-filled block construction structure needs to be corrected. Otherwise, the non-filled block construction structure does not need to be corrected.

[0147] During verification, it is necessary to determine whether the non-filled block can withstand the downward force of the part below it when the hydraulic 3D model is picked up. In other words, it is necessary to determine the part below the non-filled block in the vertical direction, namely the second non-filled block and the second block to be filled. Since the second non-filled block is a hollow design, the overall weight of the corresponding target non-filled block building structure is calculated using an approximate calculation. Since the characteristic thickness is very small, the volume of the target non-filled block building structure can be approximately regarded as the surface area of ​​the target non-filled block building structure multiplied by the characteristic thickness, and then the density is used to complete the weight calculation;

[0148] When making corrections, keep the feature thickness name unchanged, but use the target thickness to update the feature thickness value. Therefore, when updating from bottom to top according to the number, the updated non-filled block building structure must be able to withstand the falling force of the part below it.

[0149] The increase in gravity caused by the thickening of the wall will not affect the existing design, that is, the characteristic thickness must be sufficient to withstand the generated gravity. Because when designing whether the non-filled block can support the gravity of the part above itself, the solid design of the part above the non-filled block is taken into consideration. Therefore, the increased gravity of the wall will not exceed the solid case, and therefore, there will be no impact.

[0150] Reference Figure 8 As shown, calculating the upper limit of the weight of the target non-filled block building structure and the second block to be filled includes the following steps:

[0151] Fitting the surface of the target non-filled block building structure to obtain a second non-filled fitting function;

[0152] Performing surface integral on the second non-filled fitting function to obtain the surface area of ​​the target non-filled block building structure;

[0153] The surface area of ​​the target non-filled block building structure is multiplied by the characteristic thickness to obtain the volume of the target non-filled block building structure;

[0154] Multiply the density and gravitational acceleration of the target building material by the volume of the target non-filled block building structure to obtain the upper weight limit of the target non-filled block building structure;

[0155] Fitting the surface of the second block to be filled to obtain a second fitting function to be filled;

[0156] Integrate the area enclosed by the second fitting function to be filled to obtain the volume of the second block to be filled;

[0157] The density and gravitational acceleration of the target building material are multiplied by the volume of the second block to be filled to obtain the upper weight limit of the second block to be filled.

[0158] Correcting the non-filled block construction structure to obtain the non-filled block target structure includes the following steps:

[0159] During correction, the characteristic thickness of the structure built in the non-filled block is updated using the target thickness.

[0160] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned three-dimensional parametric design method for water conservancy projects based on deep learning is executed.

[0161] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state disk (SSD).

[0162] In summary, the advantages of the present invention are: by obtaining the target building material strength model, obtaining the non-filled structure and the structure to be filled, designing the construction structure of the non-filled blocks in the non-filled block set and correcting the non-filled block construction structure, the non-filled structure can be designed according to the strength properties of the target building material, so that the non-filled structure does not need to be fully filled during design. Therefore, when it is officially used, the non-filled block construction structure can bear the force acting on it, and at the same time use as little material as possible, thereby achieving the effect of cost saving.

[0163] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions merely illustrate the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A three-dimensional parametric design method for water conservancy projects based on deep learning, characterized in that: include: Obtaining parameters of a water conservancy 3D model to be designed, and obtaining target building materials for constructing the water conservancy 3D model, wherein the water conservancy 3D model is a water conservancy contour model stored in a computer; Conduct physical property analysis on target building materials to obtain the target building material strength model; Perform structural recognition on the water conservancy 3D model to obtain the non-filled structure and the structure to be filled, where both the non-filled block and the structure to be filled are solid structures; Splitting the non-filled structure into at least one non-filled block, and splitting the to-be-filled structure into at least one to-be-filled block, wherein the splitting is performed using at least one plane perpendicular to the x-axis, the y-axis, or the z-axis, wherein the distances between adjacent planes along the same direction are equal, the heights of the non-filled block and the to-be-filled block are consistent, the x-axis, the y-axis, and the z-axis are axes of a three-dimensional coordinate system, and the bottom surface of the water conservancy 3D model is perpendicular to the x-axis; Pairing the non-filled blocks to obtain at least one non-filled block set; Numbering the non-filling block set based on a positional relationship of the non-filling blocks in the non-filling block set; According to the numbering of the non-filled block sets from small to large and based on the target building material strength model, the construction structure of the non-filled blocks in the non-filled block sets is designed to obtain the non-filled block construction structure; Verify the non-filled block construction structure according to the numbers of the non-filled block sets from small to large, and based on the verification results, modify the non-filled block construction structure to obtain the non-filled block target structure; The non-filled block target structure is replaced with the corresponding non-filled block in the water conservancy 3D model to obtain the target parameter design scheme.

2. A three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 1, characterized in that: The physical property analysis of the target building material to obtain the target building material strength model comprises the following steps: Get the height of the non-filled block as the feature height; Obtaining an area value range of a cross section of the non-filled block, dividing the area value range of the cross section of the non-filled block at equal intervals to obtain at least one area point, wherein the cross section of the non-filled block is a plane parallel to a plane perpendicular to the x-axis; A sample cube is formed, satisfying that the height of the sample cube is equal to the characteristic height and the bottom area of ​​the sample cube is equal to the value at the area point; Apply uniform pressure to the top of the sample cube in a direction parallel to the height of the sample cube to obtain the maximum pressure that the sample cube can withstand as the bearing force; The top of the sample cube is pulled uniformly, with the pulling direction parallel to the height of the sample cube, and the maximum pulling force that the sample cube can withstand is obtained as the tensile force; The area point value and the load-bearing capacity are paired and fitted to obtain the load-bearing fitting function, where the area point value is the independent variable and the load-bearing capacity is the dependent variable; The area point value and the tensile force are paired and fitted to obtain the tensile fitting function, where the area point value is the independent variable and the tensile force is the dependent variable; The load-bearing fitting function and the tensile fitting function are used as the target building material strength model.

3. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 2 is characterized in that: The structural recognition of the water conservancy 3D model to obtain the unfilled structure and the structure to be filled comprises the following steps: Segment the water conservancy 3D model to obtain at least one 3D block; Draw an inscribed circle inside the bottom surface of the 3D block. The inscribed circle is the largest circle tangent to the bottom edge of the 3D block. The area of ​​the inscribed circle is divided by the bottom area of ​​the 3D block to get the inscribed circle ratio; When the proportion of the inscribed circle is greater than a preset ratio, the 3D block is treated as a non-filled part; otherwise, the 3D block is treated as a to-be-filled part, wherein the preset ratio is set based on historical data; At least one non-filled portion is aggregated into a non-filled structure, and at least one to-be-filled portion is aggregated into a to-be-filled structure.

4. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 3 is characterized in that: Pairing the non-filled blocks to obtain at least one non-filled block set includes the following steps: The non-filled blocks with coplanar bottom surfaces are aggregated to form a non-filled block set, and the plane where the bottom surfaces of the non-filled blocks in the non-filled block set are located is used as the feature plane.

5. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 4 is characterized in that: The step of numbering the non-filling block set based on the positional relationship of the non-filling blocks in the non-filling block set comprises the following steps: Calculate the distance between the characteristic plane of the non-filled block set and the bottom surface of the water conservancy 3D model to obtain the characteristic distance; The non-filled block sets are numbered from small to large according to the feature distance.

6. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 5 is characterized in that: The designing of the building structure of the non-filled blocks in the non-filled block set to obtain the non-filled block building structure comprises the following steps: At least one sampling point is evenly selected on the bottom surface of the non-filled block, and at least one sampling straight line is drawn through the sampling point, wherein the sampling straight line is perpendicular to the bottom surface of the non-filled block; Obtain the non-filled block that intersects with the sampling point as the feature non-filled block; Obtain the block to be filled that intersects with the sampling point as the feature block to be filled; The characteristic non-filling block located above the non-filling block is used as the first non-filling block, and the characteristic non-filling block located below the non-filling block is used as the second non-filling block; The characteristic block to be filled located above the non-filled block is used as the first block to be filled, and the characteristic block to be filled located below the non-filled block is used as the second block to be filled; Calculating the upper weight limits of the first non-filled block and the first block to be filled, and summing them up to obtain a first characteristic weight; Get the length of the bottom edge of the non-filled block as the characteristic length; Substitute the characteristic weight into the load-bearing fitting function and inversely solve for the identification area; The marked area is divided by the characteristic length to obtain the characteristic thickness; A non-filled block building structure is formed, wherein the surface of the non-filled block building structure is consistent with the surface of the non-filled block, the non-filled block building structure is a hollow structure, and the thickness of the wall of the cavity of the non-filled block building structure is a characteristic thickness.

7. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 6 is characterized in that: Calculating the upper weight limit of the first non-filled block and the first to-be-filled block includes the following steps: Fitting the surface of the first non-filled block to obtain a first non-filled fitting function; Integrate within the area enclosed by the first non-filled fitting function to obtain the volume of the first non-filled block; Multiplying the density and gravitational acceleration of the target building material by the volume of the first non-filled block to obtain the upper weight limit of the first non-filled block; Fitting the surface of the first block to be filled to obtain a first fitting function to be filled; Integrate the area enclosed by the first fitting function to be filled to obtain the volume of the first block to be filled; The density and gravitational acceleration of the target building material are multiplied by the volume of the first block to be filled to obtain the upper limit of the weight of the first block to be filled.

8. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 7 is characterized in that: The verification of the non-filled block construction structure includes the following steps: Obtaining a non-filled block building structure corresponding to the second non-filled block as a target non-filled block building structure; Calculate the upper limit of the weight of the target non-filled block construction structure and the second to-be-filled block, and add them up to obtain a second characteristic weight; Substitute the second characteristic weight into the tensile fitting function and inversely solve to obtain the target thickness; Based on the target thickness, the characteristic thickness is checked. When the characteristic thickness is less than the target thickness, the non-filled block construction structure needs to be corrected. Otherwise, the non-filled block construction structure does not need to be corrected.

9. The three-dimensional parametric design method for water conservancy projects based on deep learning according to claim 8 is characterized in that: The calculation of the upper limit of the weight of the target non-filled block building structure and the second to-be-filled block comprises the following steps: Fitting the surface of the target non-filled block building structure to obtain a second non-filled fitting function; Performing surface integral on the second non-filled fitting function to obtain the surface area of ​​the target non-filled block building structure; The surface area of ​​the target non-filled block building structure is multiplied by the characteristic thickness to obtain the volume of the target non-filled block building structure; Multiply the density and gravitational acceleration of the target building material by the volume of the target non-filled block building structure to obtain the upper weight limit of the target non-filled block building structure; Fitting the surface of the second block to be filled to obtain a second fitting function to be filled; Integrate the area enclosed by the second fitting function to be filled to obtain the volume of the second block to be filled; The density and gravitational acceleration of the target building material are multiplied by the volume of the second block to be filled to obtain the upper weight limit of the second block to be filled.

10. The deep learning-based three-dimensional parametric design method for water conservancy projects according to claim 9, characterized in that: The method of modifying the non-filled block construction structure to obtain the non-filled block target structure includes the following steps: During correction, the characteristic thickness of the structure built in the non-filled block is updated using the target thickness.

Images