Boundary generation type design for 2.5-axis subtractive manufacturing constraint in computer-aided design and manufacturing

By using boundary-based generative design and level set topology optimization, the problem of CAD software struggling to handle precise surface boundary representations in generative design was solved, enabling an efficient 2.5-axis subtractive manufacturing process that simplifies the machining path and improves manufacturing efficiency.

CN118965601APending Publication Date: 2024-11-15AUTODESK INC
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410867836.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2018-11-09
Filing Date
2019-11-08
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Existing CAD software has difficulty directly manipulating the exact surface boundary representation in generative design, and there are problems with the complexity of machining tool paths and long machining time in the subtractive manufacturing process.

Method used

A boundary-based generative design process is adopted, using a level set topology optimization method to generate a smooth mathematical surface model, which is then converted into a boundary representation format to adapt to the 2.5-axis subtractive manufacturing process. The shape derivative is used to control the shape change of the model to simplify the machining path.

Benefits of technology

It improves the efficiency and accuracy of generative design, reduces machining time and toolpath complexity, ensures the manufacturability of the model, and supports fast and efficient 2.5-axis subtractive manufacturing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118965601A_ABST
    Figure CN118965601A_ABST
Patent Text Reader

Abstract

Methods, systems, and apparatus, including medium-encoded computer program products, for computer-aided design of physical structures using a generative design process in which a three-dimensional model of the physical structures is generated to facilitate manufacturing the physical structures using 2.5-axis subtractive manufacturing systems and techniques, the methods, systems, and apparatus, including medium-encoded computer program products, include obtaining a design space of an object to be manufactured and one or more design criteria including at least one manufacturability constraint of the object; performing a boundary-based generative design process to generate a generative model of the object using the one or more design criteria, the one or more design criteria including at least one manufacturability constraint corresponding to a 2.5-axis subtractive manufacturing process; and providing a three-dimensional model from the generative model for manufacturing a physical structure using one or more computer-controlled manufacturing systems employing the 2.5-axis subtractive manufacturing process.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This invention is a divisional application of a patent application with application number 201980073780.3, application date November 8, 2019, and invention name “Boundary-based generative design under 2.5-axis subtractive manufacturing constraints in computer-aided design and manufacturing”.

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims the benefit of U.S. Patent Application No. 62 / 758,454, filed on November 9, 2018, entitled “BOUNDARY BASED GENERATIVE DESIGN WITH 2.5-AXIS SUBTRACTIVE MANUFACTURING CONSTRAINT FOR COMPUTER AIDED DESIGN AND MANUFACTURING,” under 35 U.S.C. 119(e), which is incorporated herein by reference in its entirety. Technical Field

[0004] The present invention relates to the field of computer-aided design, and more particularly to computer-aided design of physical structures that can be manufactured using subtractive manufacturing systems and techniques. Background Art

[0005] Computer-aided design (CAD) software has been developed and used to generate three-dimensional (3D) representations of objects, and computer-aided manufacturing (CAM) software has been developed and used to manufacture the physical structure of these objects, for example using computer numerical control (CNC) manufacturing techniques. Typically, CAD software uses a boundary representation (B-Rep) format to store a 3D representation of the geometry of the modeled object. A B-Rep model is a set of connected surface elements that specify the boundaries between the solid and non-solid parts of the modeled 3D object. In a B-Rep model (commonly referred to as B-Rep), the geometry is stored in the computer using smooth and precise mathematical surfaces, which are contrasted with the discrete and approximate surfaces of a mesh model, which may be difficult to use in a CAD program.

[0006] In addition, CAD software has been designed to perform automatic generation of 3D geometry for one or more parts in a larger system of parts to be manufactured (generative design). This automated generation of 3D geometry is usually limited to the design space specified by the user of the CAD software, and the 3D geometry generation is usually dominated by design goals and constraints, which can be defined by the user of the CAD software or by another party and imported into the CAD software. Design goals (such as minimizing the waste of material or the weight of the designed part) can be used to drive the geometry generation process towards a better design direction. Design constraints may include both structural integrity constraints of individual parts (i.e., the requirement that the part must not fail under the expected structural load during the use of the part) and physical constraints imposed by the larger system (i.e., the requirement that one part will not interfere with another part in the system during use). In addition, the input of the generative design process may include a set of input entities (B-Rep inputs) that specify boundary conditions for the generative design process, but many modern generative design solvers do not directly operate on the exact surface boundary representation of their input entities. Instead, the B-Rep is sampled and replaced with a volume representation (such as level sets or tetrahedral or hexahedral meshes) that is significantly more convenient and efficient for physics simulation and material synthesis computed by the solver.

[0007] In addition, CAD programs have been used in conjunction with subtractive manufacturing systems and techniques. Subtractive manufacturing refers to any manufacturing process that creates a 3D part from a blank material by removing portions of the blank material (usually a "blank" or "workpiece" that is larger than the 3D part). Such manufacturing processes typically involve the use of multiple CNC machining cutting tools in a series of operations that begin with a roughing operation, an optional semi-finishing operation, and a finishing operation. During one or more roughing operations, one or more cutting tools of the CNC machining system are used to quickly (relative to the semi-finishing operation and the finishing operation) remove a large portion of the workpiece so as to approach the final shape of the manufactured part. Then, one or more optional semi-finishing operations and one or more finishing operations are used to complete the manufacture of the part. Summary of the invention

[0008] This specification describes techniques related to computer-aided design of physical structures using a generative design process in which a three-dimensional (3D) model of the physical structure is generated to facilitate manufacturing of the physical structure using 2.5-axis subtractive manufacturing systems and techniques.

[0009] Generally, one or more aspects of the subject matter described in this specification may be embodied as one or more methods (and also one or more non-transitory computer-readable media that tangibly encode a computer program that is operable to cause a data processing device to perform operations), the one or more methods comprising: obtaining, by a computer-aided design program, a design space of an object to be manufactured and one or more design criteria including at least one manufacturability constraint of the object; executing, by the computer-aided design program, a boundary-based generative design process to generate a generative model of the object using the one or more design criteria, wherein the at least one manufacturability constraint causes at least one shape derivative used during the boundary-based generative design process to direct a shape change of the generative model toward discrete height layers corresponding to a 2.5-axis subtractive manufacturing process; and providing, by the computer-aided design program, a three-dimensional model based on the generative model for use in manufacturing a physical structure corresponding to the object using one or more computer-controlled manufacturing systems that employ the 2.5-axis subtractive manufacturing process, wherein the three-dimensional model includes a flat area generated by the discrete height layers.

[0010] The generative model may include a level set representation of the three-dimensional topology of the object, and the boundary-based generative design process may employ a level set topology optimization method. The at least one shape derivative used during the boundary-based generative design process may direct the shape change of the generative model toward the sidewalls of the discrete height layers, the sidewalls being parallel to the milling direction of the 2.5-axis subtractive manufacturing process. The design space may include a bounding volume containing an initial description of one or more external shapes of the three-dimensional topology of the object, and executing the boundary-based generative design process may include: expanding the one or more external shapes of the three-dimensional topology in the milling direction to fill the bounding volume.

[0011] The milling direction may include two or more milling directions, and performing the boundary-based generative design process may include: iteratively modifying the one or more external shapes of the three-dimensional topology of the object. Each iteration may include: performing a physical simulation of the object to generate a physical evaluation of the current iteration of the one or more external shapes of the three-dimensional topology of the object; calculating a shape change rate of the one or more external shapes of the three-dimensional topology in the current iteration based on the physical evaluation and the at least one shape derivative; updating the one or more external shapes of the three-dimensional topology of the object using the shape change rate; extending the one or more external shapes of the three-dimensional topology in each of the two or more milling directions to fill the bounding volume, thereby forming two or more sets of one or more extended shapes; and changing the one or more external shapes to a Boolean intersection of the two or more sets of one or more extended shapes for the next iteration.

[0012] The at least one manufacturability constraint may include: (i) a number of discrete height layers, the number being determined by user input or by automatic detection, and (ii) two or more milling directions available for the 2.5-axis subtractive manufacturing process, the two or more milling directions being determined by user input or by automatic detection, and the two or more milling directions optionally including at least one off-axis milling direction. The at least one manufacturability constraint may include: an inner corner of the generative model having a radius equal to a tool corner radius of a milling tool used in the 2.5-axis subtractive manufacturing process.

[0013] The one or more design criteria may include: settings for physical simulation of the object, at least one design goal of the object (e.g., structural performance and / or mass minimization goal), and at least one design constraint of the object (e.g., volume and / or manufacturing time constraint). The performing may include: executing the boundary-based generative design process to generate two or more generative models according to the one or more design criteria and the changes in the discrete height layers, the milling direction, or both. The method (or the operation performed by the data processing device according to a computer program tangibly encoded in one or more non-transitory computer-readable media) may include: preparing and presenting a trade-off analysis between or among the two or more generative models based at least in part on the changes in the discrete height layers, the milling direction, or both. The 2.5-axis subtractive manufacturing process may be used to directly manufacture the physical structure, the three-dimensional model being a model of the object, or the 2.5-axis subtractive manufacturing process is used to manufacture a mold of the physical structure, the three-dimensional model being a model of the mold. The providing may include: saving the three-dimensional model to a permanent storage device for use in manufacturing the physical structure corresponding to the object using the one or more computer-controlled manufacturing systems.

[0014] The providing may include: converting (i) a polygonal mesh output by the boundary-based generative design process, or (ii) generative design data (e.g., level set distance field data) directly obtained from the boundary-based generative design process into a boundary representation model or a parameterized feature model to form the three-dimensional model, wherein the parameterized feature model can be edited as sketch geometry. The providing may include: using the three-dimensional model to generate a tool path description for a 2.5-axis subtractive manufacturing machine; and using the tool path description to manufacture the physical structure corresponding to the object or at least a portion of a mold of the physical structure by the 2.5-axis subtractive manufacturing machine.

[0015] One or more aspects of the subject matter described in this specification may also be embodied as one or more systems, the one or more systems comprising: a non-transitory storage medium having instructions of a computer-aided design program stored thereon; and one or more data processing devices, the one or more data processing devices being configured to run the instructions of the computer-aided design program to perform any of the one or more methods described above. The one or more systems may also include a 2.5-axis subtractive manufacturing machine.

[0016] The present specification also describes techniques related to computer-aided design of physical structures using data format conversion (e.g., one or more outputs from a generative design process), as well as user interface techniques that can be easily used with 2.5-axis subtractive manufacturing systems and techniques to facilitate generating 3D models of physical structures.

[0017] One or more aspects of the subject matter described in this specification may be embodied in one or more methods, the one or more methods comprising: obtaining, by a computer-aided design (CAD) program, a first model of an object to be manufactured using a 2.5-axis subtractive manufacturing process, wherein the first model comprises smooth curves fitted to contours of discrete height layers representing the object to be manufactured using the 2.5-axis subtractive manufacturing process; modifying, by the CAD program, at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process, wherein the modifying comprises: detecting at least a portion of a first smooth curve of a first discrete height layer in the discrete height layers that is substantially coincident with at least a portion of a second smooth curve of a second discrete height layer in the discrete height layers; and replacing the at least a portion of the first smooth curve of the first discrete height layer in the discrete height layers with the at least a portion of the second smooth curve of the second discrete height layer in the discrete height layers; modifying, by the CAD program, at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process. The CAD program prepares an editable model of the object, the editable model comprising a series of construction steps represented by a parametric feature history, wherein the series of construction steps comprises: combining extruded versions of the smooth curves to form a solid three-dimensional (3D) model of the object in a boundary representation format, and the parametric feature history comprises sketch features associated with at least one of the discrete height layers; presenting, by the CAD program, a user interface element displaying the sketch features associated with the editable model; receiving, by the CAD program, user input through the user interface element; and reshaping, by the CAD program in response to the user input, a subset of the smooth curves in at least one of the discrete height layers to change the solid 3D model; and repeating, by the CAD program, the series of construction steps represented by the parametric feature history to construct, in the boundary representation format, the solid 3D model of the object as changed by the user input.

[0018] One or more aspects of the subject matter described in this specification may be embodied as one or more non-transitory computer-readable media encoding a computer-aided design program operable to cause one or more data processing devices to perform operations comprising: obtaining a first model of an object to be manufactured using a 2.5-axis subtractive manufacturing process, wherein the first model comprises smooth curves fitted to contours of discrete height layers representing the object to be manufactured using the 2.5-axis subtractive manufacturing process; modifying at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process, wherein the modification comprises: detecting at least a portion of a first smooth curve of a first discrete height layer in the discrete height layers that is substantially coincident with at least a portion of a second smooth curve of a second discrete height layer in the discrete height layers; and replacing the at least a portion of the first smooth curve of the first discrete height layer in the discrete height layers with the first smooth curve of the second discrete height layer in the discrete height layers. preparing an editable model of the object, the editable model comprising a series of construction steps represented by a parameterized feature history, wherein the series of construction steps comprises: combining extruded versions of the smooth curves to form a solid three-dimensional (3D) model of the object in a boundary representation format, and the parameterized feature history comprises a sketch feature associated with at least one of the discrete height layers; presenting a user interface element displaying the sketch feature associated with the editable model; receiving user input through the user interface element; and reshaping a subset of the smooth curves in at least one of the discrete height layers to change the solid 3D model in response to the user input; and repeating the series of construction steps represented by the parameterized feature history to construct the solid 3D model of the object as changed by the user input in the boundary representation format.

[0019] Both the method and non-transitory computer-readable medium aspects of the subject matter described in this specification may include the following features. The first model may include one or more modeling entities to be retained in the solid 3D model of the object; the modifying may include: detecting at least a portion of another smooth curve of the first discrete height layer in the discrete height layers that is nearly coincident with at least a portion of the one or more modeling entities, and aligning the at least a portion of the another smooth curve of the first discrete height layer in the discrete height layers with the at least a portion of the one or more modeling entities; wherein the shaping includes constraining changes to the at least a portion of the another smooth curve to maintain tangency and contact with the at least a portion of the one or more modeling entities; and wherein the series of construction steps includes combining the extruded version of the smooth curve with the one or more modeling entities to form the solid 3D model of the object in the boundary representation format.

[0020] The obtaining may include: receiving shape data corresponding to the object to be manufactured using the 2.5-axis subtractive manufacturing process; processing the shape data to generate a polyline that matches the contour of the discrete height layers representing the object; fitting the smooth curve to the polyline; and replacing at least one line segment of at least one of the smooth curves for at least one of the discrete height layers to reduce the curvature of the at least one line segment to be less than or equal to the curvature of the smallest milling tool that can be used in the 2.5-axis subtractive manufacturing process. Replacing the at least one line segment of the at least one of the smooth curves may include: for each set of smooth curves within one of the discrete height layers: offsetting the set of smooth curves outwardly by an amount based on the curvature of the smallest milling tool available; and offsetting the set of smooth curves inwardly by the amount, while using circular arcs to close any resulting gaps.

[0021] The shape data may include: a level set distance field output by a generative design process that employs a level set-based topology optimization method, and processing the shape data may include: for each of the discrete height layers: resampling the level set distance field to produce a two-dimensional (2D) level set grid at a slice plane corresponding to a current one of the discrete height layers, the slice plane being perpendicular to a milling direction to be used in the 2.5-axis subtractive manufacturing process (it should be noted that the milling direction may be arbitrarily aligned and is not necessarily parallel to any of the level set 3D coordinate system axes); and extracting a current layer polyline from the 2D level set grid, the current layer polyline matching the current layer contour of the object at the slice plane.

[0022] Receiving the shape data may include: receiving a mesh representation of the object, or receiving a level set representation of the object from a generative design process, and processing the shape data may include: converting the level set representation of the object into a mesh representation of the object; wherein processing the shape data to generate the polyline includes: slicing the mesh representation of the object along a milling direction by planes located within corresponding ones of the discrete height layers, each of the planes being perpendicular to the milling direction; and wherein obtaining may include: segmenting the polyline into a first set of line segments assigned to the one or more modeling entities and a second set of line segments not assigned to the one or more modeling entities, wherein the segmentation uses a tolerance value when assigning polyline vertices to line segments that are erroneous in assignment to the one or more modeling entities; and extending one or more line segments of the second set of line segments connected to one or more line segments of the first set of line segments until the one or more line segments of the second set of line segments meet at an intersection point or tangency point of at least one of the one or more modeling entities, while also checking to align intersection points of any extended line segments that intersect at least one same one of the one or more modeling entities.

[0023] The obtaining may include: receiving a level set representation of the object, or receiving a mesh representation of the object and converting the mesh representation into a level set representation of the object; modifying one or more level set values ​​in the level set representation of the object in each of one or more milling directions specified for the 2.5-axis subtractive manufacturing process to remove undercuts; modifying one or more additional level set values ​​in the level set representation of the object to move the faces of one or more planes of the object upward to a height level of a corresponding one of discrete height layers of the faces of the planes; and converting the modified level set representation of the object into an output mesh representation of the object; wherein the first model includes the output mesh representation of the object.

[0024] The method / operation may include: for each of the one or more modeling entities: for each layer in the discrete height layers, intersecting the layer with the modeling entity along the milling direction to produce a portion of the modeling entity, moving the portion of the modeling entity completely below the layer in the milling direction, sweeping the portion of the modeling entity upward through the layer opposite to the milling direction to produce a swept entity, and intersecting the layer with the swept entity to produce a replacement for the portion of the modeling entity; and sweeping one or more upper side surfaces of the modeling entity upward along the milling direction opposite to the milling direction to the topmost level of the discrete height layer to produce one or more upper side swept entities; sweeping any lower side surfaces of the modeling entity downward along the milling direction to the bottommost level of the discrete height layer to produce one or more lower side swept entities; intersecting the one or more upper side swept entities with the one or more lower side swept entities to produce one or more undercut fill entities; and combining the one or more undercut fill entities with the modeling entity.

[0025] One or more aspects of the subject matter described in this specification may be embodied as one or more systems, the one or more systems comprising: a display device; one or more data processing devices coupled to the display device; and a non-transitory storage medium coupled to the one or more data processing devices, the non-transitory storage medium encoding: program code that causes the one or more data processing devices to obtain a first model of an object to be manufactured using a 2.5-axis subtractive manufacturing process, wherein the first model includes smooth curves fitted to contours of discrete height layers representing the object to be manufactured using the 2.5-axis subtractive manufacturing process; program code that causes the one or more data processing devices to modify at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process, including detecting at least a portion of a first smooth curve of a first discrete height layer in the discrete height layers that is substantially coincident with at least a portion of a second smooth curve of a second discrete height layer in the discrete height layers, and replacing the at least a portion of the first smooth curve of the first discrete height layer in the discrete height layers with the at least a portion of the second smooth curve of the second discrete height layer in the discrete height layers. program code for causing the one or more data processing devices to prepare an editable model of the object, the editable model comprising a series of construction steps represented by a parameterized feature history, wherein the series of construction steps comprises: combining extruded versions of the smooth curves to form a solid three-dimensional (3D) model of the object in a boundary representation format, and the parameterized feature history comprises sketch features associated with at least one of the discrete height layers; program code for causing the one or more data processing devices to present a user interface element displaying the sketch features associated with the editable model to the display device; program code for causing the one or more data processing devices to receive user input through the user interface element; program code for causing the one or more data processing devices to reshape a subset of the smooth curves in the at least one of the discrete height layers in response to the user input to change the solid 3D model; and program code for causing the one or more data processing devices to replay the series of construction steps represented by the parameterized feature history to construct the solid 3D model of the object in the boundary representation format as changed by the user input.

[0026] The non-transitory storage medium of the one or more systems may encode program code, which causes the one or more data processing devices to perform any operation of the non-transitory computer-readable medium aspect of the subject matter described in this specification. For example, the non-transitory storage medium may encode program code, which causes the one or more data processing devices to perform a boundary-based generative design process to generate a generative model of the object using the one or more design criteria, wherein the at least one manufacturability constraint causes at least one shape derivative used during the boundary-based generative design process to direct the shape change of the generative model to discrete height layers corresponding to a 2.5-axis subtractive manufacturing process. In some implementations, the one or more systems include a subtractive manufacturing machine, and the non-transitory storage medium encodes program code, which causes the one or more data processing devices to generate one or more tool path instructions for the solid 3D model and use the one or more tool path instructions to manufacture at least a portion of the physical structure through the subtractive manufacturing machine.

[0027] Specific embodiments of the subject matter described in this specification can be implemented to achieve one or more of the following advantages. A three-dimensional model of a physical structure can be generated, the three-dimensional model including flat areas generated by discrete height layers imposed by at least one manufacturability constraint used during topology optimization in a generative design process, wherein the discrete height layers can help reduce CAM programming time, machining time, reduce the need for custom fixtures for clamping, or a combination of the foregoing during machining of the physical structure in a 2.5-axis subtractive manufacturing process. Trade-off analysis of design variants enables a sufficient number of design variants for the user to perform further analysis between designs. The described subject matter can provide the possibility of estimating the production cost of a 3D model and performing a final accuracy simulation of the production of the 3D model.

[0028] Specific embodiments of the subject matter described in this specification may also be implemented to achieve one or more of the following advantages. Geometries may be generated that are both easy to edit in a CAD application and easy and efficient to manufacture using a subtractive manufacturing method with 2.5 axis capabilities (e.g., by a 3-axis milling machine). The generated output geometry includes flat areas perpendicular to the 2.5D axis that can be quickly cleared using a flat milling cutter. Undercuts that cannot be reached by a 2.5 axis constrained tool path may be avoided by not creating undercuts in the first instance and / or eliminating undercuts during conversion to a boundary representation format. Thin shelves may be prevented from being generated in the final output geometry, as thin shelves may unnecessarily increase the milling time of a part. High curvature areas that cannot be reached and cut by a tool of a given radius may also be prevented from being generated in the final output geometry.

[0029] The algorithm can use various forms of input, including 3D polygon meshes, polyline outlines of each 2.5D layer of the object, and levels if provided, or the levels can be detected from the 2.5 axis directions. The algorithm can be robust to imperfect inputs, such as geometry that is not 2.5D at all (the algorithm can force the shape to be completely 2.5D), geometry with rounded edges that should be sharp (the algorithm can create sharp edges), and geometry with undercuts either in preserved geometry or in organic (generatively designed) geometry (the algorithm can fill undercuts).

[0030] An easy-to-use editor interface can be created for users to work with 3D models of objects to be manufactured using the 2.5-axis subtractive manufacturing process. Sketch user interface elements associated with object outlines in parametric feature history can allow for rapid changes to the 3D model while also ensuring that the reconstructed 3D model after each edit will remain capable of being manufactured using the 2.5-axis subtractive manufacturing process. The resulting construction geometry can be assembled in a parametric CAD application as a feature recipe, which facilitates user editing while also maintaining 2.5-axis subtractive manufacturability.

[0031] The details of one or more embodiments of the subject matter described in this specification are set forth in the accompanying drawings and the description below.Other features, aspects, and advantages of the invention will become apparent from the description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1A An example of a system that can be used to design and manufacture physical structures under 2.5-axis subtractive manufacturing constraints is shown.

[0033] Figure 1B An example of a process for designing and manufacturing a physical structure under 2.5-axis subtractive manufacturing constraints is shown.

[0034] Figure 2A Graphical representation showing an example of 2.5-axis subtractive manufacturing.

[0035] Figure 2B Graphical representation showing an example of 3-axis subtractive manufacturing.

[0036] Figure 3A A graphical representation showing an example of a design under a single milling direction.

[0037] Figure 3B Graphical representation showing an example of a design in two opposite milling directions.

[0038] Figure 3C Graphical representation showing an example of a design in three milling directions.

[0039] Figure 3DA graphical representation showing an example of a design at a single height level.

[0040] Figure 3E A graphical representation showing an example of a design at two height levels.

[0041] Figure 3F A graphical representation showing an example of a design at three height levels.

[0042] Figure 3G A graphical representation showing an example of a design at three height levels that can be automatically extracted from the preserved geometry surfaces perpendicular to the milling direction.

[0043] Figure 4 An example of a process of using formal constraints to iteratively optimize the topology of one or more portions of a 3D model of an object generatively designed for 2.5-axis subtractive manufacturing is shown.

[0044] Figure 5 A graphical representation showing an example of the results of a trade-off analysis.

[0045] Fig. 6A Graphical representation showing an example of a milling head geometry definition.

[0046] Figure 6B Graphical representation showing an example of shape constraints for 2.5-axis subtractive manufacturing.

[0047] Fig. 7A , Figure 7B , Figure 7C , Fig.7D Graphical representation showing an example of the shape initialization process for 2.5-axis subtractive manufacturing performed in two milling directions.

[0048] Figure 8 An example of a process of iteratively optimizing the topology of one or more portions of a 3D model of an object generatively designed for 2.5-axis subtractive manufacturing using formal constraints and access fixes is shown.

[0049] Fig. 9 is a schematic diagram of a data processing system including a data processing device, which can be programmed as a client or a server.

[0050] Fig.10 An example of a process for converting received shape data (eg, mesh geometry) into a watertight boundary representation of an object to be manufactured using a 2.5-axis subtractive manufacturing process is shown.

[0051] Fig.11AShown is an example of a 3D mesh model from a generative design solver, contours of different height levels of this instance of the 3D mesh model, and the placement of cross sections used to generate the contours.

[0052] Fig. 11B An example is shown of resampling a level set representation of an object to produce polylines that match the current layer contour of the object at a slice plane that is perpendicular to the milling direction to be used in a 2.5-axis subtractive manufacturing process but is not necessarily aligned with the level set 3D coordinate system.

[0053] Fig. 11C is a schematic diagram showing a cross-sectional view of an example of an area of ​​a 2.5-axis millable level set entity.

[0054] Fig.11D An example showing how to impose concave curvature compliance.

[0055] Fig.11E A graphical representation showing an example of preparing an editable model of an object, receiving user input to a sketcher user interface element to reshape layers of a 3D model of the object, and replaying a parametric feature history to reconstruct the 3D model of the object.

[0056] Fig.12 Another example of a process for generating a watertight boundary representation for an object to be manufactured using a 2.5-axis subtractive manufacturing process is shown.

[0057] Fig.13A Shows an example of grid layering.

[0058] Fig. 13B Shows an example of modifying the level set values ​​to remove undercuts in a mesh.

[0059] Fig. 13C An example of modifying the level set value to move the face of a plane up to the layer level is shown.

[0060] Fig.13D An example of a process of homogenizing a layer of an input modeling entity preserving regions is shown.

[0061] Fig.13E An example of retention volume layering applied to a 3D retention volume is shown.

[0062] Fig.13F An example of a process for removing an undercut from a retention body that has been homogenized within a defining layer is shown.

[0063] Fig.14A An example of a polyline and initial segmentation of a polyline for a slice of an object is shown.

[0064] Fig. 14B and Fig. 14CShows the process of generating line segments of a polyline.

[0065] FIG. 15A to FIG. 15E A graphical representation showing the alignment of the curves is shown.

[0066] Fig.16 An example of a 2D sketch user interface element showing a parametric editable model.

[0067] Like reference numbers and designations in the various drawings indicate like elements. DETAILED DESCRIPTION

[0068] A method for imposing manufacturability constraints in a generated part so that the generated part can be manufactured using 2.5-axis subtractive manufacturing is described. Multi-axis CNC milling is enhanced by allowing material to be removed from the flank of the machining tool. Shape derivatives that impose a user-specified number of discrete height levels are added, which can produce large flat areas that can be easily machined. The discrete height levels and milling directions can be set manually by the user or can be automatically detected. A post-processing stage can be applied, which forces the surface to conform to the discrete height levels and optionally adds internal fillets that allow support for "bullnose" type machining tools. The final geometry can be exported as a mesh with additional metadata for conversion into CAD / CAM friendly geometry. In some implementations, a boundary-based generative design process (e.g., a level set method) is used, but it can be modified as described below. In some implementations, systems and techniques described in U.S. patent application Ser. No. 62 / 758,404, filed on Nov. 9, 2018, entitled “MACROSTRUCTURE TOPOLOGY GENERATION WITH DISPARATE PHYSICAL SIMULATION FOR COMPUTER AIDED DESIGN AND MANUFACTURING,” and PCT / US2019 / 060089, filed on Nov. 6, 2019 and published on May 14, 2020 as WO 2020 / 097216, are used, and the applications are hereby incorporated by reference.

[0069] Figure 1AAn example of a system 100 that can be used to design and manufacture a physical structure under 2.5-axis subtractive manufacturing constraints is shown. A computer 110 includes a processor 112 and a memory 114, and the computer 110 can be connected to a network 140, which can be a private network, a public network, a virtual private network, etc. The processor 112 can be one or more hardware processors that can each include multiple processor cores. The memory 114 can include both volatile memory and non-volatile memory, such as random access memory (RAM) and flash RAM. The computer 110 can include various types of computer storage media and devices, which can include a memory 114 for storing instructions for programs running on the processor 112, the programs including one or more computer-aided design (CAD) programs 116, the CAD programs 116 implementing three-dimensional (3D) modeling functions and including one or more generative design processes to achieve topology optimization (e.g., using at least one level set method as described) and physical simulation. The physical simulation performed by the systems and techniques described in this document can simulate one or more physical properties and can use one or more types of simulations. For example, finite element analysis (FEA) (including linear static FEA), one or more finite difference methods, and one or more material point methods may be used. In addition, simulations of physical properties may include computational fluid dynamics (CFD) simulations, acoustic / noise control simulations, heat conduction simulations, computational injection molding simulations, electrical or electromagnetic flux simulations, and / or material solidification (which may be used for phase changes during molding) simulations.

[0070] As used herein, CAD refers to any suitable program for designing a physical structure that meets design requirements, regardless of whether the program is capable of interfacing with and / or controlling manufacturing equipment. Thus, one or more CAD programs 116 may include one or more computer-aided engineering (CAE) programs, one or more computer-aided manufacturing (CAM) programs, and the like. One or more programs 116 may be run locally on the computer 110, remotely on a computer of one or more remote computer systems 150 (e.g., one or more server systems of one or more third-party providers accessible to the computer 110 via the network 140), or both locally and remotely. Thus, the CAD program 116 may be two or more programs that operate in concert on two or more separate computer processors, because the program 116 operating locally on the computer 110 may offload processing operations (e.g., generative design and / or physical simulation operations) to the “cloud,” thereby causing one or more programs 116 on one or more computers 150 to perform the offloaded processing operations.

[0071] The one or more CAD programs 116 present a user interface (UI) 122 on a display device 120 of the computer 110, which can be operated using one or more input devices 118 (e.g., a keyboard and a mouse) of the computer 110. Figure 1A 1 as separate devices, but the display device 120 and / or the input device 118 may also be integrated with each other and / or with the computer 110, such as in a tablet computer (e.g., a touch screen may be the input / output devices 118, 120). In addition, the computer 110 may include a virtual reality (VR) or augmented reality (AR) system, or may be part of such a system. For example, the input / output devices 118, 120 may include VR / AR input gloves 118a and / or a VR / AR head-mounted viewer 120a. In any case, the user 160 interacts with one or more CAD programs 116 to create and modify one or more 3D models, which may be stored in one or more 3D model documents 130.

[0072] The initial 3D model 132 may be an input to the generative design process. The user 160 may define a topology optimization problem for the generative design process to operate to generate the desired 3D model 133 from the starting 3D model 132, or the input may be a design space without a specific starting 3D model 132 (it should be noted that the generative design process itself may generate a starting geometry within the design space). In addition, the starting model 132 may be a retained geometry of the input, which may be an unconnected modeling entity, in which case the generative design process is used to generate a new 3D geometry connecting the input entity. As described herein, one or more CAD programs 116 implement at least one generative design process, which enables one or more CAD programs 116 to automatically generate one or more parts of one or more 3D models (or the entirety of the 3D model) based on one or more design goals and one or more constraints, wherein the geometric design is iteratively optimized based on simulation feedback. It should be noted that, as used herein, "optimization" (or "optimal") does not mean that the best of all possible designs is achieved in all cases, but rather that the best (or near-optimal) design is selected from a limited set of possible designs that can be generated within the allocated time, given the available processing resources.

[0073] Design constraints may be defined by the user 160 or by another party and imported into one or more CAD programs 116. For 2.5-axis subtractive manufacturing, the design constraints may include one or more manufacturability constraints in the generated 3D object such that the generated 3D object may be manufactured using 2.5-axis subtractive manufacturing. Additionally or in combination, the design constraints may include both structural integrity constraints of individual parts (e.g., a requirement that a part must not fail under expected structural loads during use of the part) and physical constraints imposed by a larger system (e.g., a requirement that a part be contained within a specified volume so as not to interfere with another or other parts in the system during use).

[0074] Various generative design processes that can optimize the shape and topology of at least a portion of a 3D model can be used. The optimization of the geometric design of one or more 3D models performed by one or more CAD programs 116 involves topology optimization, which is a lightweight method in which the optimal distribution of materials is determined by minimizing an objective function (e.g., structural compliance constrained by volume) subject to design constraints. There are two main types of topology optimization: density-based methods and boundary-based methods. Density-based methods discretize the volume of a part and assign density to each discrete unit. Then, the density is changed to solid and empty while supporting specified boundary conditions. In contrast, boundary-based methods track the shape of the external interface of a solid part and move the boundary so that the constraints are met. In order to impose one or more 2.5-axis manufacturability constraints, a boundary-based method can be selected because the boundary-based method has the advantage of providing an accurate knowledge of the boundary. The boundary-based method can track the boundary using an implicit method called a level set method. The advantage of this level set method is that the topology is allowed to change as the surface evolves without re-meshing.

[0075] Once the user 160 is satisfied with the generatively designed 3D model, the 3D model may be stored as a 3D model document 130 and / or another representation used to generate the model (e.g., a tool path description for 2.5-axis subtractive manufacturing). This may be done upon request by the user 160, or may be done upon request by the user for another action, such as sending the 3D model 132 to a subtractive manufacturing (SM) machine 170 or other manufacturing machinery that may be directly connected to the computer 110 as shown or connected via a network 140. This may involve post-processing performed on a local computer 110 or a cloud service to export the 3D model 132 as an electronic document upon which manufacturing is based. It should be noted that an electronic document (which will be referred to simply as a document for brevity) may be a file, but does not necessarily correspond to a file. A document may be stored in a portion of a file that holds other documents, in a single file dedicated to the document in question, or in multiple coordinated files.

[0076] The one or more CAD programs 116 may provide the document 135 (with tool path instructions in an appropriate format) to the SM machine 170 to create a complete structure 138 including an optimized topology and shape (e.g., a stepped design generated for 2.5-axis milling). The SM machine 170 may employ one or more subtractive manufacturing techniques, such as a computer numerical control (CNC) milling machine, such as a multi-axis multi-tool milling machine. Additionally, the user 160 may save or transfer the 3D model for later use. For example, the one or more CAD programs 116 may store the document 130 including the generated 3D model.

[0077] In various implementations, one or more CAD programs 116 of system 100 may implement one or more generative design processes as described in this document. A generative design process seeks optimal geometry, topology, or both. For example, a generative design process seeks optimal geometry among alternative designs by minimizing a performance-related objective function subject to constraints:

[0078] minimize

[0079] Make g i (s,u(s))=0i=1,…,n g (2)

[0080] where s is a vector of design variables related to the geometry of the domain, and u is a vector of state variables (e.g., displacements) that depend on s. Additional constraints (e.g., equilibrium) are imposed by the set g i For simplicity, the constraints are assumed to be equal. The mathematical programming methods used to minimize (1) can be gradient-based or non-gradient-based methods. Gradient-based methods (relative to non-gradient-based methods) usually use more information associated with the design sensitivity, such as:

[0081]

[0082] It is the derivative of the performance-related objective function with respect to the design variables. In level set-based topology optimization methods, s represents the boundary of the solid region.

[0083] Figure 1BAn example of a process for designing and manufacturing a physical structure under a 2.5-axis subtractive manufacturing constraint is shown. For example, a design space of an object and one or more design criteria are obtained 180 through one or more CAD programs 116 for generating a generative 3D model. The design criteria may include one or more manufacturability constraints of the object, such as one or more 2.5-axis subtractive manufacturability constraints. Different generative design processes can be formulated using different combinations of design variables (e.g., level sets). In some implementations, the design variables may include various types of inputs received through the UI 122, such as selecting from different generative design synthesis methods available in one or more CAD programs in the system 100. In some implementations, the available generative design synthesis methods may include a level set-based topology optimization that provides a basic level set method for topology optimization.

[0084] A design space is obtained for a subtractive manufacturing process, which typically begins with a solid block of stock material and then gradually removes material from the block. In some implementations, the obtained design space may include: (1) a design space for generative design geometry generation, e.g., a boundary representation (B-Rep) 3D model designed or loaded into one or more CAD programs 116 for use as a subspace of an optimization domain for the described generative design process, and / or (2) a set of input entities that specify boundary conditions for generative design geometry generation, e.g., a B-Rep selected using UI 122 to specify one or more subspaces that are reserved for use as one or more connection points with another or other components in a larger 3D model or one or more separate 3D models. For example, the obtained design space may include obstacles, e.g., B-Reps for removing volumes from the design space. In response to input from a user 160, for example, one or more CAD programs 116 may use different combinations of design variables. For example, the user 160 may select different generative design synthesis methods to use within corresponding different design spaces within a single 3D model.

[0085] One or more design criteria may include settings for physical simulation, such as element density in a FEA model to be used with an optimized 3D topology of a generatively designed part. One or more design criteria may include various design goals and manufacturability constraints such as those described in this document. In addition, for example, one or more CAD programs 116 may provide functions to assist a user in specifying one or more design criteria.

[0086] With the generative design space and design criteria specified, one or more CAD programs 116 use one or more generative design processes to generate 185 one or more 3D models, for example. For example, one or more CAD programs 116 perform a boundary-based generative design process to generate a generative model with discrete height layers corresponding to a 2.5-axis subtractive manufacturing process. The discrete height layers produce flat areas in the one or more 3D models, which facilitates the manufacture of one or more corresponding physical structures. In some implementations, one or more generative design processes using a level set topology optimization method are used to iteratively modify one or more external shapes of the 3D topology of the 3D model. In combination with Figure 4 and Figure 8 More details are described regarding iteratively optimizing the topology of a 3D model of an object for 2.5-axis subtractive manufacturing.

[0087] In some implementations, one or more generative design processes may use the described level set method, where s from equations (1), (2), and (3) represents the boundary of a solid region implicitly represented using one or more level sets, which are signed distance values ​​calculated on a Cartesian background grid. In a level set-based topology optimization method, the external shape of a structure is represented by a one-dimensional high-level level set function, and changes in shape and configuration are replaced by changes in the value of the level set function in order to obtain an optimal structure. The level set function refers to a function that indicates whether each part of a design domain in which an initial structure is set corresponds to a material domain (material phase) in which the structure is formed and occupied by a material, a void domain (void phase) in which voids are formed, or a boundary between these two domains, where a predetermined value between a value representing the material domain and a value representing the void domain represents a boundary between the material domain and the void domain.

[0088] like Figure 1B As shown, the generative design methods described in this document can be implemented in, for example, one or more CAD programs 116 to provide (1) substantial user control over the generative design process and post-processing of the generative design output to form a final acceptable 3D model of an object, and (2) control functions for providing the generatively designed 3D model for use in manufacturing a physical structure corresponding to the object.

[0089] In some implementations, at least one shape derivative is used in a boundary-based generative design process. The shape derivative can direct the shape change of the generative model to the sidewalls of the discrete height layers, which are parallel to the milling direction of the 2.5-axis subtractive manufacturing process. Figure 4 Describes more details about shape derivatives.

[0090] In some implementations, performing 185 the boundary-based generative design process may include shape initialization. The design space may include a bounding volume containing an initial description of one or more external shapes of the three-dimensional topology of the object. Performing 185 the boundary-based generative design process may include: extending the one or more external shapes of the three-dimensional topology to fill the bounding volume in directions that cannot be reached by the milling direction (there is no new fixture for the part during manufacturing). FIG. 7A to FIG. 7D Describes more details about shape initialization.

[0091] The results of the generative design process may be presented to the user, for example, in UI 122 on display device 120, along with an option 190 to accept or reject the design. For example, 3D models 133 generated by the generative design process may be presented to user 160 in UI 122. In some implementations, two or more 3D models generated by a generative design process (e.g., a process that iteratively modifies one or more external shapes of a 3D topology) may be presented to the user along with a trade-off analysis of design complexity versus manufacturing cost. The trade-off analysis may assist user 160 in accepting or rejecting one or more of the presented 3D models. Figure 5 Describes more details about the design changes used for the trade-off analysis.

[0092] If the design is rejected, Figure 1B The process may return to obtaining 180 a new design space and / or new design criteria for generating a new generative 3D model, for example, from one or more CAD programs 116. Once the design is not rejected 190, Figure 1B The process may be, for example, by one or more CAD programs 116 providing 195 a 3D model of an object based on one or more external shapes for 2.5-axis subtractive manufacturing of a physical structure. Providing 195 may involve sending or saving the 3D model to a permanent storage device for use in manufacturing a physical structure corresponding to the object using a SM manufacturing system. In some implementations, providing 195 involves, for example, one or more CAD programs 116 using the 3D model to generate 195A tool path instructions for one or more computer-controlled SM manufacturing systems, and, for example, one or more CAD programs 116 using the tool path instructions generated for a 2.5-axis SM machine to manufacture 195B at least a portion of the physical structure corresponding to the object through one or more computer-controlled SM manufacturing systems. In some implementations, providing 195 may include using the tool path instructions generated 195A to manufacture a mold for the physical structure through a 2.5-axis subtractive manufacturing machine.

[0093] The 3D model provided 195 may be a 3D model generated 185 by a generative design synthesis method or a post-processed version of a generative design output. In some implementations, a polygonal mesh that is an output of a boundary-based generative design process, or generative design data obtained directly from a boundary-based generative design process, may be converted into a boundary representation (B-Rep) model and / or a parametric feature model, for example, by one or more CAD programs 116. For example, the generative design data may be level set distance field data obtained directly from a boundary-based generative design process. The boundary representation model or the parametric feature model may be edited as sketch geometry. For example, in some implementations, a 3D mesh model generated by a generative design synthesis method may be converted into a watertight B-Rep 3D model before providing 195, as described in further detail below.

[0094] The following combination can also be achieved Figure 4 and Figure 8 The generative design processes described herein may be implemented using the post-processing, editing, and / or providing 195 systems and techniques described above. Although described in the context of a CAD program that provides multiple options for generative design, each of the generative design processes described in this document may be implemented as a stand-alone generative design process in a CAD program. Figure 4 and Figure 8 Not all of the generative design processes described need to be implemented together in any given implementation. In any case, the goal is to produce a 3D model of an object that facilitates 2.5-axis subtractive manufacturing of the object.

[0095] Subtractive manufacturing processes typically start with a solid block of stock material and then gradually remove material from the stock. One of the common subtractive methods is milling, which uses a rotary cutter (also called a "tool" or "drill") to remove material. The milling process can limit the types of shapes that can be manufactured because the milling machine must be able to hold the part securely and the rotating drill must be able to reach the surface of the material without interference. Other important considerations that should also be taken into account are part vibration during material removal and the stress of the milling process on the drill itself.

[0096] Subtractive manufacturing techniques can include 2-axis milling, 2.5-axis milling, or 3-axis milling. 2-axis milling can cut into the stock without adjusting the height level of the milling head. 3-axis milling can cut into the stock while moving the milling tool in three separate dimensions. 2.5-axis milling uses a 3-axis milling machine because the milling tool (or a combination of a milling tool and a fixture support) can move in all three separate dimensions, but in most cutting operations, the milling tool only moves along 2 axes relative to the workpiece, which results in a more efficient manufacturing process. FIG. 2A to FIG. 2B Showing the difference between 2.5-axis subtractive manufacturing and 3-axis subtractive manufacturing. Figure 2A A graphical representation of an example of 2.5 axis subtractive manufacturing is shown. In the example of 2.5 axis milling, the subtractive process occurs with continuous movement in a plane perpendicular to the milling tool 202 but in discrete steps parallel to the milling tool 202 . Figure 2B A graphical representation of an example of 3-axis subtractive manufacturing is shown. An example of 3-axis milling may have continuous movement in a plane perpendicular to the milling tool 204, and may also have continuous movement parallel to the milling tool 204. In other words, three-axis milling may have continuous movement in all three directions. Compared to 3-axis subtractive manufacturing, the 2.5-axis subtractive manufacturing process can quickly remove layers of material in sequence and create a part that typically has a series of "pits" at different depths. Targeting 2.5-axis subtractive manufacturing means that the geometry should have a layered shape, which enables much faster programming of subtractive machine tool paths and much shorter machining times.

[0097] In some implementations, the 2.5-axis subtractive manufacturing process is used to directly manufacture the physical structure, in which case the 3D model can be a model of the object to be manufactured. In some implementations, the 2.5-axis subtractive manufacturing process is used to manufacture a mold for the physical structure, in which case the 3D model can be a model of the mold. For example, Figure 2A The 3D model 206 in may be a model of a mold for manufacturing a physical structure, wherein the mold is to be manufactured using a 2.5-axis subtractive manufacturing process.

[0098] In some implementations, manufacturability constraints in a generative design process may include: (i) a number of discrete height levels, and (ii) two or more available milling directions for a 2.5-axis subtractive manufacturing process. The number of discrete height levels may be determined by user input or by automatic detection. The two or more milling directions may be determined by user input or by automatic detection. In some implementations, the two or more milling directions may optionally include at least one off-axis milling direction.

[0099] FIG. 3A to FIG. 3C Three examples of designs under different milling directions in a 2.5-axis subtractive manufacturing process are shown. The milling direction is the direction in which the milling tool 202 moves toward the milling blank. The milling direction may also be referred to as the "mounting orientation." 2.5-axis parts typically limit the number of milling directions used to simplify 2.5-axis subtractive manufacturing. Figure 3A Graphical representation showing an example of design results in a single milling direction (+Z). Figure 3A Both a perspective view 310 and a perspective cross-sectional view 312 of the same 3D model to be manufactured are shown. A single milling direction (+Z) may produce a region 314 in the middle of the blank. Figure 3B Graphical representation showing an example of design results in two opposite milling directions (-Z, +Z). Figure 3BBoth a perspective view 320 and a perspective cross-sectional view 322 of the same 3D model to be manufactured are shown. In the blank, a milling direction +Z may produce an area 324 , while an opposite milling direction −Z may produce an area 326 . Figure 3C Graphical representation showing an example of design results in three milling directions (-Z, +Z, +Y). Figure 3B Both a perspective view 330 and a perspective cross-sectional view 332 of the same 3D model to be manufactured are shown. In the blank, a milling direction of +Z may produce an area 334, while the opposite milling direction of -Z may produce an area 336. Additionally, a milling direction of +Y may produce an area 338 in the blank. The milling direction is an important design constraint and design criteria when setting up a generative design process.

[0100] One or more milling directions may be specified in several different ways. In some implementations, one or more milling directions may be manually specified by a user through a user interface. A user may define a given number of orientations by defining a three-dimensional vector indicating one or more desired milling directions. In some implementations, one or more milling directions may be detected from user-provided retained geometry. The user-provided retained geometry may be added to the optimized geometry and may be an interface to other components in the assembly. In the case of a set of user-provided retained geometries, milling directions may be automatically extracted to infer the user's design intent. Geometric feature detection may be performed to extract one or more three-dimensional vectors from features that will require a given milling direction. For example, in the case where a hollow cylinder feature is used to represent a bolt hole interface, a vector along the axis of the cylinder may be extracted to be used as a milling direction. This approach may potentially reduce the number of required milling directions by utilizing existing required milling directions defined by retained geometry. In some implementations, one or more milling directions may be determined based on optimization. In this approach, topology optimization is followed by a clustering algorithm that finds the minimum number of milling directions for a set of performance criteria.

[0101] Figures 3D to 3F Three examples of designs at different numbers of height levels in a 2.5-axis subtractive manufacturing process are shown. A height level is a discrete step along the milling direction where material is removed in line with a fixed number of levels. A larger number of height levels may enable more complex geometries to be defined, but may affect manufacturing time. Figure 3D A graphical representation showing an example of design results at a single height level. Figure 3D Both a perspective view 340 and a perspective cutaway view 342 of the same 3D model to be manufactured are shown. A single height level can produce areas 344, 346 and three holes 348A, 348B and 348C in the blank. The SM machine can cut into the blank at this single height level. Figure 3E Graphical representation showing an example of design results at two height levels. Figure 3EBoth a perspective view 350 and a perspective cross-sectional view 352 of the same 3D model to be manufactured are shown. In the perspective view 350, the first height level can produce both regions 354A and 356 in the blank. The first height level can be more clearly seen in the perspective cross-sectional view 352, which shows a region 354B that is identical to region 354A. The second height level can produce regions 358A, 358B, and 358C, where the SM machine can cut into the blank at the second height level. Figure 3F Graphical representation showing an example of design results at three height levels. Figure 3F Both a perspective view 360 and a perspective cutaway view 362 of the same 3D model to be manufactured are shown. A first height level may produce region 364A, while a second height level may produce region 366A. In perspective cutaway view 362, region 364B (same as region 364A) and region 366B (same as region 366A) show two height levels. Similar to Figure 3E , a third height level is required to produce regions 368A, 368B, and 368C in the blank. The height level is an important design constraint and design criterion when setting up a generative design process.

[0102] The height levels may be specified in several different ways. In some implementations, one or more height levels may be manually specified by a user through a user interface. A user may specify a fixed number of height levels, or specify a ratio to derive the location and number of height levels, or explicitly set the number and location of height levels through a user interface or programmatically. In some implementations, one or more height levels may be automatically detected from retained geometry provided by a user. For each given milling direction, the height levels may be extracted from the height of the retained geometry surfaces perpendicular to the milling direction. Figure 3G An example of a set of height levels derived from user-provided retained geometry is shown. In this example, three height levels 304, 306, and 308 are extracted from surfaces at the top, middle, and bottom of the retained geometry 309 perpendicular to the milling direction 302. In some implementations, one or more height levels may be automatically obtained based on optimization. In this approach, topology optimization is followed by a clustering algorithm that finds the appropriate number and location of height levels for a set of performance criteria.

[0103] For example, if the set of height levels G required for a particular milling direction is not known a priori, the automatic detection algorithm may include determining the elements of G by a weighted k-means method. In this clustering, a set of surface points that satisfies -n(x)·m>0 is used, where m is the milling direction and n(x) refers to the outward unit normal vector field of the level set φ containing the part x. These surface points are clustered together at their heights close to the clipping plane, and the clusters are dynamically updated every n iterations. Since the 2.5-axis constraint is not strictly enforced in the early stages of the optimization, this allows the surface Ω to roughly form the final shape of the surface, thereby clustering useful height levels together when the constraint is strictly enforced.

[0104] Figure 4 A process for generating one or more portions of a 3D model of an object to be manufactured by 2.5-axis subtractive manufacturing (e.g., as described by Figure 1A An example of a 2.5 axis shape optimization method is provided in which the process includes iteratively optimizing the topology using one or more formalized constraints. This 2.5 axis shape optimization can impose discrete height layers through a shape derivative method, which can impose both vertical walls (parallel to the milling direction) and discrete height layers. The advantage of the shape derivative method is that it is not applied as a filter to the velocity during advection, but the shape derivatives define the velocity required to meet the constraints. The shape derivatives can be applied to the shape during level set based topology optimization. Once the constraints are met, it can help ensure that the resulting geometry can be manufactured from a given set of milling directions through a 2.5 axis milling process.

[0105] The generative design process to be used employs a level set topology optimization method. The design criteria, design space, and / or other design variables are previously obtained, for example, by one or more CAD programs 116. This may involve receiving user input, for example, via a UI 122 on a display device 120, importing information from another program or third-party source, and / or one or more of these inputs may be predefined in a given implementation.

[0106] An input shape is obtained as an initial 3D model or one or more parts of an initial 3D model to be used as a starting geometry. For example, the input shape 410 is a two-dimensional (2D) representation of an instance of the starting geometry. A physical simulation of the object is performed 402 to generate a physical evaluation of the current iteration of one or more external shapes. The physical evaluation is generated according to one or more design goals and one or more design constraints, and the one or more design constraints include one or more 2.5-axis manufacturability constraints. Various types of physical simulations can be used, and therefore various types of physical evaluations can be generated. In some implementations, performing 402 physical simulation involves performing a finite element analysis simulation using a body-fitting mesh-based solver, and changing the constitutive model by mapping geometric field data and finite element simulation data between voxel grid points of a level set representation of a generative model and elements of a solid mesh model used for the finite element analysis simulation. In some implementations, the solid mesh-based simulation can be non-body-fitting, for example, solved using a voxel solver. For example, after the physical simulation, a physical evaluation 412 of the input shape 410 is generated.

[0107] A shape change rate is calculated 404 based on the physical evaluation and at least one shape derivative. The shape change rate is calculated taking into account the effects of different objectives and constraints. In some implementations, calculating 404 the shape change rate involves calculating the advection rate, such as using an augmented Lagrangian method when there are multiple shape derivatives. For example, based on one or more formalized constraints 413 and the physical evaluation 412, the optimizer can calculate a shape change rate 414 for one or more external shapes of the current input geometry 410.

[0108] It should be noted that there may be several constraints on the shape of the generatively designed part so that the part can be produced by 2.5-axis milling technology. Figure 6B A graphical representation of an example of shape constraints for 2.5-axis subtractive manufacturing. The milling process involves moving the end mill in discrete increments along the milling direction. Therefore, the surface of the part should be divided into a series of height layers, where "walls" perpendicular to the height layers connect each layer. Figure 6B As shown, the 2.5-axis mold has a milling direction m and a marking g 0 , g 1 , g 2 and g 3 These layers may include recessed areas called “pits”. For example, Figure 6B Two recessed areas in the part are shown 606 and 608. The shape and size of these recessed areas are related to the shape of the tool itself.

[0109] Fig. 6AA graphical representation showing an example of a milling head geometry definition. An end mill (i.e., a tool bit) is defined using the following parameters: corner radius e 604 and drill radius r 602. The drill length defines the shape of the tool bit, known as a "bull nose," represented by a round cylinder oriented in the milling direction m. The tool head radius h defines another cylinder capped with a hemisphere, representing the tool holder and the tool head. The ends of the tool head cylinder are assumed to expand away from the curved surface.

[0110] Return to Figure 6B ,like Fig. 6A The shape of the defined tool itself may affect the output shape of the part, because recessed areas (i.e., "pits") must be accessible to the tool. In other words, the tool should be able to contact every part of the surface without intersection between the tool and Ω. This also means that the inner corners of the part should have a radius equal to the tool corner radius e604. For example, in Figure 6B In the example, the sidewalls of the recessed area 606 are parallel to the milling direction m with an inner radius e604. The size of the recess is twice as wide as the tool radius r602.

[0111] A detailed example of formal constraints in an iterative topology optimization process is now described. As discussed above, the formal constraints may include: (1) height layer shape derivative constraints for 2.5-axis manufacturability, (2) tool head and / or tool drill accessibility constraints, and (3) multiple milling directions.

[0112] A penalty term is defined to penalize part shapes that do not conform to the 2.5 axis constraints. In the case of a set of discrete height levels G, the penalty function for each point on the surface is formed as a product of exponentials. This yields the value x when x is at level g. j ∈G and a penalty of 0 when far away from any g j The penalty is at most 1.

[0113]

[0114] where γ defines the exponentially decreasing tightness. γ is defined as follows:

[0115]

[0116] where t 0 and t 1 are the starting and ending ranges of the level set Ω in the milling direction m.

[0117] The next term in the penalty function ensures that surface points whose normals are perpendicular to m are not penalized:

[0118] f(x,m) 2 =(n(x)·m) 2 , (6)

[0119] And square it so that the sign does not affect the penalty. The final penalty function is formed as the product of (4) and (6) and integrated over the surface:

[0120] j(x,m)=f(x,m) 2 E(x,m),

[0121]

[0122] This penalty function (7) has a lower bound of 0 and an upper bound of 1, and therefore does not require further normalization. This penalty function can be used in the augmented Lagrangian framework to enforce equality constraints As part of the structural topology optimization problem. To be applied during shape updating, the shape gradient D of the penalty function is calculated by Ω J:

[0123] D Ω J=div || (2fm || )-f 2 H, (8)

[0124]

[0125] where div || and m || is the divergence and component of the milling direction parallel to the surface, and H is the mean curvature. A further regularization term of the form –vH can be added, which is equivalent to additionally penalizing small multiples v of the part surface area (e.g., v = 0.02)

[0126] In order to take into account the accessibility of the surface from the milling direction, D Ω J–vH is multiplied by a mask term A(x,m) that hides surface points that are not reachable from the milling direction. One type of reachability constraint is a line of sight from x to the bounding box in the –m direction. The minimum value φ is measured along a one-way ray. When Introducing a slightly different version of , which takes into account the unidirectionality of the ray function:

[0127]

[0128] To solve the interference of the tool head, an offset surface Ω can be used +A second ray intersection test is performed by multiplying the tool head radius h, and A(x,m):=1 only if no intersection is detected. To handle the reachability of the tool itself, two morphological operations are applied at the end of each shape update step. The first morphological operation ensures that cavities narrower than the tool radius r are eliminated. The velocity field v r (x) is created as follows:

[0129] vr(x)=r[n(x)-m(n(x)·m)], (11)

[0130] and then use Hamilton-Jacobi transport along v r (x) Advect the surface Ω. Then, reinitialize the level sets Ω and φ as signed distance fields. Finally, use -v r (x) to recover the surface, except when the cavity is already filled. The second morphological operator is similar to the first one, except that the velocity field is chosen to be perpendicular to the surface, as follows:

[0131] v e (x) = en(x). (12)

[0132] The effect of this morphological operator is to create corner rounding in the interior cavity as Figure 6B Tool corner radius e in .

[0133] An example of an algorithm for combining 2.5-axis constraints from multiple milling directions is formulated. In this example, the penalty becomes A The product of the penalties, where A(x,m)=1.

[0134]

[0135] And the shape derivative becomes:

[0136]

[0137] If desired, a main milling direction can be specified so that when the surface cannot be reached from the main direction, the height layer penalty is only applied in directions other than the main milling direction. This prevents excessive layering when multiple directions are imposed.

[0138] Return to Figure 4 , one or more external shapes are updated 406 using the shape change velocity. A level set representation of the object is updated 406 using the shape change velocity before the next iteration. For example, an updated shape 416 is generated according to the shape change velocity 414. In some implementations, the update 406 may use an advection velocity from an augmented Lagrangian method (ALM).

[0139] A check for convergence 407 is performed. In some implementations, the iterative topology optimization process may be determined to have converged if the update to one or more external shapes is less than a threshold value. In some implementations, convergence occurs when the shape updates have reached the accuracy limit supported by the computer or CAD program. In some implementations, the topology optimization converges when a fixed number of iterations have been performed. The physical simulation 402, the calculation of the shape change rate 404, the update of one or more external shapes 406, and the check for convergence 407 are then iterated until convergence.

[0140] In some implementations, after convergence, a trade-off analysis between or among the two or more generative design 3D models is prepared and presented 408. The trade-off analysis is based at least in part on changes in discrete height layers, milling directions, or both. The trade-off analysis can be presented to the user 160 in the UI 122. Figure 5 A graphical representation of an example of the results of a trade-off analysis is shown. Four generative design 3D models 502, 504, 506, 508 are presented. Design variations between the four models are visualized to compare cost and performance. In the example shown, the X-axis of the trade-off analysis is design complexity, and the Y-axis of the trade-off analysis is cost. Other axes and / or comparison metrics may be used.

[0141] Providing multiple design variations can be a key role for generative design. Trade-off analysis of design variations enables a sufficient number of design variations for the user to perform further analysis between designs. Milling direction and height level parameters are used to increase the possibility of design variations. For each parameter, the user is able to specify a single value, a range of values, or an automatic method. Based on these parameters, multiple designs can be synthesized covering a range of different performance, cost, and manufacturing outcomes.

[0142] In some implementations, for each 2.5-axis generative design job, assuming the user specifies "X" as the tool direction, three design variations are created as follows:

[0143] +X setup with 3 layers,

[0144] There are 3 layers of -X settings,

[0145] There are +X settings and -X settings with 5 levels.

[0146] This design process can create multiple design variations while requiring minimal user input. The design provides a cost / performance trade-off analysis that the user can evaluate (e.g., single setup vs. two setups). This design process creates different variants with the explicit goal of increasing the trade-off space. In some implementations, a special case of only one layer is provided as "2-axis cutting" where the only option for tool orientation is orientation (x,y,z). For example, the exploration options may be:

[0147] 2.5 axis with 2 layers,

[0148] 2.5 axis with 3 layers,

[0149] 2.5 axis with 5 layers,

[0150] 2-axis cutting.

[0151] In addition to design variants (e.g., milling direction and height level parameters), other parameters may include directions x, y, and z, tool diameter, tool length, and other parameters, which may change frequently. Typically, a given setup can be used to obtain generatively designed solutions for: 3-axis milling machines, 5-axis milling machines, unlimited additive casting, or die casting. Another design variant or parameter is material. For example, the same setup can be used to design generative models for different materials (such as aluminum or copper). Given the various possible design variants, including milling direction, height level, tool diameter, tool length, and material, a single setup can generate a large number of design results and possibly corresponding trade-off analyses.

[0152] Fixturing is one aspect of manufacturing cost, and the generative design process seeks to minimize the number of fixture steps during manufacturing. For example, in order to manufacture a part via 2.5-axis subtractive manufacturing, the part needs to be held fixed to the table of the machining tool. Using geometry that is easily fixed to the table greatly simplifies the machining process and machining costs. Creating parts with a flat base (like Figure 6B ) will make the task of clamping the part relatively simple. During the generative design process, the initial shape may have large unreachable areas. Therefore, an initialization process may be performed to automatically fill these areas, thereby creating a design with a flat base. This initialization may be called "reach area modification". Figure 4 Immediately at the end of the process or Figure 8 The touch area modification is performed in each iteration of the process. This touch area modification process can remove the burden from the shape update routine. The touch area modification takes as input a given shape Ω, a bounding box and the set of milling directions M. For each milling direction m i , calculate the new volume Ω by filling any volume that is inaccessible from the milling direction i The initialization volume is then generated by the Boolean intersection of all these volumes: After that, use Ω * To initialize the signed distance field φ.

[0153] For example, FIG. 7A to FIG. 7D A graphical representation showing an example of the shape initialization process for 2.5-axis subtractive manufacturing performed in two milling directions. Fig. 7AIn the initialization, the shape Ω, the bounding box and the set of milling directions m are given. 1 and m 2 As input. Figure 7B In the milling direction m, 1 The inaccessible volume is used to calculate the new volume Ω 1 .exist Figure 7C In the milling direction m, 2 The inaccessible volume is used to calculate the new volume Ω 2 .exist Fig.7D In , the initialization volume is generated by the Boolean intersection of these volumes: Ω * =Ω 1 ∩Ω 2 .

[0154] In some implementations, a reach zone modification process can be performed within a boundary-based generative design process and includes iteratively modifying one or more external shapes of an object's 3D topology. Figure 8 An example of a process of iteratively optimizing the topology of one or more portions of a 3D model of an object generatively designed for 2.5-axis subtractive manufacturing using formalized constraints and touch zone modifications is shown.

[0155] A physical simulation of the object is performed 802 to generate a physical evaluation of a current iteration of one or more external shapes of the object's 3D topology. Figure 8 804. The result of the physical simulation 818 of the current iteration of the input shape 816 is shown in FIG. A shape change speed of one or more external shapes of the 3D topology in the current iteration is calculated 804 based on the physical evaluation and at least one shape derivative. For example, the shape change speed 820 is calculated taking into account the physical evaluation result 818 and the constraint 819. The one or more external shapes are updated 806 using the shape change speed. For example, an updated shape 822 is obtained based on the shape change speed 820 after the update. It should be noted that the shape change speed corresponding to Figure 4 Operations 402, 404, and 406 in Figure 8 Operations 802, 804 and 806 in the above can be combined with Figure 4 Same process as described (i.e., same constraints).

[0156] One or more outer shapes are expanded 808 to fill the bounding volume in each of one or more corresponding directions that are not accessible to the one or more milling directions. This generates one or more sets of one or more expanded shapes. For example, a touch area modification 823 is performed on the updated shape 822 by expanding the outer shape of the shape 822. A new shape 824 (e.g., a dark area) is calculated by filling in the area 825 that is not accessible to the milling direction m. In some implementations (e.g., FIG. 7A to FIG. 7DIn the example of FIG. 1 , two or more sets of extended shapes are calculated in directions that are not accessible to the two or more milling directions. One or more external shapes are changed 810 to a Boolean intersection of the two or more sets of extended shapes. The final extended shape after the Boolean intersection (e.g., Fig.7D The shape of Ω * After the touch zone is modified, the corresponding Figure 4 Operations 407 and 408 Figure 8 Operations 812 and 814 in the above can be combined using Figure 4 Same process as described.

[0157] Fig. 9 9 is a schematic diagram of a data processing system including a data processing device 900, which can be programmed as a client or a server. The data processing device 900 is connected to one or more computers 990 via a network 980. Fig. 9 Only one computer is shown as the data processing device 900, but multiple computers may be used. The data processing device 900 includes various software modules that may be distributed between the application layer and the operating system. These software modules may include executable and / or interpretable software programs or libraries, including tools and services of one or more 3D modeling programs 904 that implement the above-mentioned systems and techniques. Therefore, the one or more 3D modeling programs 904 may be one or more CAD programs 904, and may implement (1) one or more generative design processes (e.g., using one or more level set-based methods for generative design) to perform topology optimization and physical simulation operations (finite element analysis (FEA) or other), and / or (2) converting polygonal meshes (e.g., output from a generative design process or from a scanner) and / or generative design data obtained directly from a boundary-based generative design process into a boundary representation model and / or a parametric feature model. In addition, one or more programs 904 may potentially implement manufacturing control operations (e.g., generating and / or applying tool path instructions to affect the manufacturing of the designed object). The number of software modules used may vary depending on the implementation. Furthermore, software modules may be distributed across one or more data processing devices connected via one or more computer networks or other suitable communications networks.

[0158] The data processing device 900 also includes hardware or firmware devices, which include one or more processors 912, one or more additional devices 914, computer-readable media 916, communication interfaces 918, and one or more user interface devices 920. Each processor 912 is capable of processing instructions for execution within the data processing device 900. In some implementations, the processor 912 is a single-threaded or multi-threaded processor. Each processor 912 is capable of processing instructions stored on a computer-readable medium 916 or on a storage device such as one of the additional devices 914. The data processing device 900 uses a communication interface 918 to communicate with one or more computers 990, for example, via a network 980. Examples of user interface devices 920 include displays, cameras, speakers, microphones, tactile feedback devices, keyboards, mice, and VR and / or AR equipment. The data processing device 900 may store instructions for implementing operations associated with the above-mentioned one or more programs, for example, on a computer-readable medium 916 or one or more additional devices 914 (e.g., one or more of a hard disk device, an optical disk device, a tape device, and a solid-state memory device).

[0159] Fig.10 An example of a process for converting received shape data (e.g., mesh geometry) into a watertight boundary representation of an object to be manufactured using a 2.5-axis subtractive manufacturing process is shown. A first model of the object is obtained 1010, for example, by one or more programs 116, 904, wherein the first model includes smooth curves fitted to contours of discrete height layers representing the object to be manufactured using the 2.5-axis subtractive manufacturing process. The curves are smooth in the computer modeling sense because each curve is stored in the computer using a precise mathematical definition (e.g., a spline curve) that has derivatives of all orders everywhere in the function domain defining the curve. In addition, the first model may include a 3D model, such as one or more modeled entities associated with the smooth curves.

[0160] The received shape data may be output data from a generative design process, such as one or more of the generative design processes described above, and the received shape data may be a mesh representation of the object or another representation of the object from the generative design process (e.g., a level set representation). In some implementations, the received shape data includes a polygonal mesh, which may be a native or converted triangular mesh. Various types of mesh formats may be used, including tetrahedral or hexahedral meshes. Additionally, the one or more modeled entities are in a boundary representation format (e.g., B-Rep), and may be regular geometric shapes that have been fit to the scanned mesh or one or more modeled entities used as input to a generative design process that employs a volume mesh approach or a level set approach to create a generative design for the object, such as described above.

[0161] Thus, the polygonal mesh may be an output from one or more generative design processes that employ one or more modeling entities (i.e., retained volumes) as one or more subspaces of an optimization domain of the generative design process. For example, an initial 3D model of an object may be a triangular mesh (or level set geometry, such as level set voxel data) generated by a generative design solver that has been constrained to produce a 2.5-axis millable shape, where such output from the generative design solver needs to be converted into an editable B-Rep volume. Such an initial 3D model of the object may be processed to extract 1012 a polyline profile for each height layer of the object that describes the shape of the layer. Smooth curves may then be fitted 1014 to the polyline profiles, for example while respecting the maximum curvature allowed by the tool radius, so as to produce smooth curves for the first model.

[0162] Obtaining 1010 may involve receiving or generating an initial 3D model, for example, any polygonal mesh geometry or level set geometry that can be plane-sliced ​​to produce closed polyline loops. Obtaining 1010 may also include receiving or determining other information associated with the object to be manufactured and therefore the first 3D model, such as one or more milling directions for a 2.5-axis subtractive manufacturing process, one or more retained regions and potentially one or more obstacle regions (subspaces to be excluded from the design space or optimization domain of a generative design process), a B-Rep volume, a milling height layer location, and at least one milling tool radius (e.g., a minimum milling tool radius or the radius of the smallest milling tool available when cutting a part to form the object during manufacturing). It should be noted that one or more larger tools are typically used to remove a large portion of the blank material, while one or more smaller tools are typically used to enter corners that cannot be reached by the larger tools. The received milling tool radius information may be provided per layer, but the algorithm may be designed to run without receiving any milling tool radius information, in which case a default minimum tool radius (or no minimum tool radius) may be used during the conversion process.

[0163] Furthermore, the milling direction is also referred to as the attack direction, since this is the direction in which the milling tool approaches the stock material, i.e., the vector from the tool base to the tool tip. The algorithm can be designed to operate without receiving 2.5D levels in the input shape relative to a selected origin. However, the algorithm should receive some form of 2.5D extrusion direction, such as a 2.5D axis that is bidirectionally parallel to the normal of each level or planar region of the input geometry. Therefore, in some implementations, the process automatically determines the levels based on the received 2.5D axis information, and then divides the input shape into layers (material between two adjacent levels) as determined by these levels.

[0164] Fig.11AAn example of a 3D mesh model 1100 from a generative design solver is shown, where the 3D mesh model 1100 has height levels L0, L1, L2 perpendicular to an attack direction 1102 (which is the milling direction for 2.5-axis subtractive manufacturing). In some implementations, the algorithm assumes that all geometries generated by the generative design solver can be 2.5-axis milled from one main direction or two bidirectional parallel directions, as shown, and that the retention bodies themselves can be 2.5-axis milled from the same directions. Obstacle areas can have any shape. In some implementations, the input shape data can be checked to confirm that the input geometry and the retention and obstacle B-Rep bodies have these properties. In some implementations, the algorithm processes any retention and obstacle B-Rep bodies based on one or more milling directions to ensure that the input shape data has the correct 2.5-axis millable properties after conversion to B-Rep format.

[0165] In the case of such a mesh 1100, to extract 1012 the polyline contours, a planar cross section of the mesh 1100 may be taken to create polyline rings. Standard procedures for reverse engineering a mesh into a B-Rep may be used, as may variations of these standard procedures, as described herein. It should be noted that since rounding of sharp edges is a natural consequence of the marching cubes algorithm used to fit the mesh to the level set geometry, placing the cross section at the midpoint between the layers may be more reliable. Fig.11A As shown, cross sections 1104 (perpendicular to the direction of attack 1102) are placed at midpoints between layers of the mesh 1100. For example, in the case of a triangular mesh input, the polyline is the union of all intersections of a given infinite plane with the triangular surfaces of the mesh (with special treatment for triangles that completely coincide with the plane). By positioning the two cross-sectional planes 1104 between the height layers to avoid rounded areas, the contours 1106 of the different height levels of the 3D mesh model 1100 can be improved. It should be noted that the contours 1106 are rings that describe the shape of each height layer, and each contour 1106 can have a generated contour segment 1106A adjacent to a retained volume 1108 (or more generally, a segment separated from any retained volume 1108) and a keep-in contour segment 1106B.

[0166] This is just one method of extracting 1012 contours from the received shape data, which can potentially introduce shape errors into the final B-Rep conversion due to the fact that the mesh is a noisy approximation of the original horizontal group voxel data that the generative solver naturally uses, or in the case of a scanned mesh, the mesh can be a noisy approximation of the scanned physical object. Various processing operations can be employed to help avoid introducing shape errors when using a mesh, such as described in further detail below. Off-axis features (such as side holes) may make cross-section extraction methods on a mesh less than optimal because the side holes may cause the contour to move into an undercut region or be topologically split into multiple loops. To handle these off-axis features, such features can be eliminated from the model before extracting the contour. Doing this on a curved triangle mesh is challenging, but it is much simpler to do so on a horizontal group voxel representation.

[0167] Thus, in some implementations, the mesh itself is not used to extract 1012 contours 1106. In some implementations, the algorithm selects a technique for extracting 1012 polylines representing planar slices of the input geometry based on the type of input geometry. In some implementations, contours are extracted 1012 from level set data, where the level set input is defined by a set of values ​​on a 3D voxel grid, where each value represents a distance to an implicit level set surface. Polyline approximations of planar slices of the level set geometry may be created, where the planes are orthogonal to the milling direction 1102. Thus, each such plane defines a slice of the 3D level set, and even where the planes are not aligned with the voxel grid, the level set distance field may be resampled to produce a 2D level set grid for each slice plane.

[0168] The shape data may be a level set distance field output by a generative design process employing a level set-based topology optimization method, and processing the shape data to produce polylines that match contours of discrete height layers representing an object may include: resampling the level set distance field to produce a two-dimensional (2D) level set grid at a slice plane corresponding to a current discrete height layer in the discrete height layers, and extracting a current layer polyline from the 2D level set grid, wherein the current layer polyline matches the current layer contour of the object at the slice plane. Input level set voxel data with a voxel grid at an arbitrary orientation relative to the slice plane may be processed to create polyline approximations of planar slices of the level set geometry, as described below. Each such plane defines a slice of the 3D level set, and even in the case where the plane is not aligned with the voxel grid, the level set distance field may be resampled to produce a 2D level set grid for each slice plane.

[0169] Fig. 11B1 shows an example of resampling a level set representation of an object to produce a polyline that matches the current layer contour of the object at a slice plane that is perpendicular to the milling direction to be used in a 2.5-axis subtractive manufacturing process; the milling direction is not necessarily aligned with the level set 3D coordinate system. The first graphical representation 1110 shows a general step level plane π i Seam frame aligned with grid axes The second graphical representation 1112 shows a 2D level set grid Γ covering the minimum area junction rectangle P. i The points in the graph 1112 represent the grid voxels γ st Consider the grid that is aligned by assigning uniform axes to The value of the vertex Represents the level set. Each axis (x, y, z) in the selected coordinate system is scaled using an independent factor rescaled, so it can be assumed without any general restriction that is an integer grid. That is, each cell is defined as:

[0170]

[0171] Accordingly, let and yes The minimum and maximum grid points (voxels) of , that is, we get:

[0172]

[0173] Furthermore, to describe the CNC machine tool subtractive operation along the axis, it is assumed that the unit ray in is the source point, and is an arbitrary milling direction per unit length that is not necessarily aligned with the coordinate system. Finally, assume that in the subtractive operation of the CNC machine tool, a set of layer levels representing "height layers" is obtained

[0174] make is the volume corresponding to each subtractive layer step along r. Extracting the plane π represents the “intermediate level” i (p):D r (p)=(l i +l i+1 ) / 2 on each T i 2D slices of Slices are covered by π i2D level set grid Γ on i Composition. In the common case of being collinear with the coordinate axis, Γ i is trivially set to Orthogonal 2D grid tiles Otherwise, by Resample to construct Γ i .

[0175] First, calculate π i Align box with (integer) axis The intersecting polygons Then, calculate the minimum area coplanar rectangle Since P is convex, B can be efficiently computed by applying the rotating caliper algorithm (see Toussaint, G., 1983, Solving Geometric Problems with the Rotating Calipers, In Proceedings IEEE MELECON, 10–17). Then, by choosing an arbitrary corner point B as the origin O and replacing the axis unit vector and Aligned with the incident side B to define π i 2D Cartesian coordinate system on

[0176] Let a and b be and Align the length of side B. Set along and The grid cell length and Since max(h u ,h v )≤1 / 2(i.e., the grid size This ensures that there are enough grid points to mitigate the resampling accuracy loss. In order to cover P, Γ i A×B units are required, where and The unit is composed of A+1×B+1 uniformly placed voxel points {γ st |s∈[0,A],t∈[0,B]}, at last, Positioning containing unit And use the corner G ijk The value at {d ijk} to estimate the 2D level set value {d st =G ijk (γst )}。

[0177] In some implementations, a projection method can be used. If the grid is aligned with one of the milling directions (e.g., along the z-axis), for a given volume representing the region between any two planes that will align with the height layers, the algorithm can iterate over each grid point, and for each z-column of grid points, the algorithm can project the minimum voxel value (including negative values) onto the xy-plane. In practice, only the voxel grid points near the level set surface have values (positive if outside the solid and negative if inside the solid). Additionally, if the direction is not aligned with one of the grid directions, the voxel grid can be rotated, and the grid values can be resampled relative to the original grid.

[0178] The algorithm can create a 2D level set voxel grid with a set of pairs (2D points, values), and the 2D level set voxel grid is a "shadow" of the 3D shape. By performing a 2D marching cubes algorithm on the 2D level set, a polyline contour of the height layer can be extracted. An important property of this contour is that it ensures that all the material is enclosed within it, and any undercuts or internal voids in the 3D shape are naturally eliminated by always projecting the most negative level set distance value. These undercuts can cause mesh intersection problems and should therefore be removed, such as by using the projection method described here or other methods that help extract the contour for each layer (such as the layering method described further below).

[0179] An example of the projection method for the 3-level case is now provided. Fig. 11C FIG. is a schematic cross-sectional view showing an example of a region of a 2.5-axis millable level set solid 1120A. There is a side hole 1124 whose radius is large enough to intersect both the L1 plane and the L2 plane. This side hole is to demonstrate that it is better to analyze multiple sets of layers at once rather than one layer. The "blank" contour 1120B, which is the outermost contour of the entire solid, can be extracted by applying the projection method to the entire solid relative to the milling direction 1122. Then the deepest layer (L2 in this example) can be determined, and the solid volume 1120C above L2 relative to the milling direction 1122 can be identified. Then the level set projection method can be performed to extract a contour for this <L2 region 1120C. Then this process can be repeated for the <L1 region, and generally, the layers can be continued to be moved upward until a polyline contour is extracted for each layer.

[0180] Although processing the level set representation as described above is an option for extracting 1012 contours, this is not required. In some implementations, the mesh itself is used to extract 1012 contours, and a layering operation (such as described in further detail below) is performed to eliminate off-axis features from the mesh and improve the extracted contours. As described above, in some implementations, the received shape data is not from the output of a generative design process, but from the output of another process. For example, the shape data can be the output from a 3D scanner, which provides a representation of a physical object that the 3D scanner has scanned, which can use contact-based and / or non-contact-based (active or passive) scanning technology, such as laser triangulation, structured light technology, modulated light technology, laser pulse-based 3D scanning, laser phase shift 3D scanning, coordinate measuring machine (CMM) technology, arm-based 3D scanning and detection systems, optical tracking 3D scanners and detection systems, conoscopic holography systems, computed tomography (CT) imaging, magnetic resonance imaging (MRI), etc. Such implementations may also use one or more modeling entities that should be retained in the solid 3D model of the object, such as one or more retained regions associated with the polygonal mesh from the 3D scanner and a B-Rep volume for one or more obstacle regions.

[0181] In the case where an object has been scanned for manufacturing using 2.5 axis milling and this object needs to be connected to other objects in an assembly, reservations and obstacles for the 3D scanned mesh may be generated. For example, screw holes may be needed to connect the scanned object to another object. These screw holes may already exist in the scanned mesh, but will not have the precise circular radius required for a particular screw or bolt specification. Therefore, one or more programs 116, 904 may enable the user to import or directly create reservation models representing these precisely defined screw / bolt holes (usually cylindrical holes), and these reservation models (including obstacle areas) can then be used as input to the algorithms described in this application. These additional modeling objects (e.g., modeling objects for screw / bolt holes) can therefore affect the conversion of the mesh to B-Rep, and will also be included in the final 3D model of the object to provide connection points of the correct size for the larger assembly.

[0182] Return to Fig.10Regardless of the source of the initial 3D model, and regardless of the technique used to extract 1012 the polyline contour, fitting 1014 the curve to the polyline contour can generate a smooth curve of the first model for subsequent processing. In some implementations, obtaining 1010 includes: processing received shape data (from a generative design process or output data from a scanner) to produce a polyline that matches the contour of discrete height layers representing an object to be manufactured using a 2.5-axis subtractive manufacturing process, fitting 1014 the smooth curve to the polyline, and replacing at least one line segment of at least one of the smooth curves for at least one of the discrete height layers to reduce the curvature of the at least one line segment to be less than or equal to the curvature of a smallest milling tool that can be used in the 2.5-axis subtractive manufacturing process.

[0183] In some implementations, fitting 1014 includes fitting a spline curve to the broken line points using a least squares minimization technique for computational geometry. However, the problem becomes more complicated when a manufacturability constraint associated with a minimum tool radius is applied. If the profile is fitted to a curve having a region of high curvature in a concave direction that exceeds the curvature of the tool, i.e., the minimum tool radius for milling the part is greater than the radius of the curve to be milled, then the tool will not be able to mill the region. The problem is further complicated by the need to ensure that profiles from adjacent layers are aligned when viewed along the direction to avoid thin shelves between layers, as described in further detail below.

[0184] It should be noted that in addition to the condition that the radius of curvature should be greater than the minimum tool radius, some implementations also use a condition that the distance from any point on the profile to the median axis of the profile should be greater than the minimum tool radius. In this case, coin fitting can close the narrow necks in the profile. But in some cases, due to milling constraints, such geometries with these narrow necks are unlikely to be generated by 2.5D generative algorithms. However, in some implementations, the voronoi diagram algorithm used in coin fitting can be used to check to detect such narrow necks in the profile, and then a similar modification strategy can be used, as described below. Then, some points can be pushed outward as needed to widen the narrow necks.

[0185] Fig.11D An example showing how to impose concave curvature compliance. Known optimization methods for fitting a curve to a point that minimizes curvature can be used, but a requirement of 2.5-axis millable designs is to fit curves with strict upper curvature limits in concave directions (i.e., nooks and crannies). In some implementations, this requirement is addressed by detecting non-compliant regions and replacing them with compliant curve segments.

[0186] For example, each polyline may be segmented so as to produce polyline subsegments of the fitted curve with known endpoints. Splines 1130A may be fitted (via standard linear methods) using arbitrary but appropriate relative tolerances based on the ratio of the bounding box diagonals of the entire model. The curve may then be sampled at a higher density than the polyline vertices. The curvature at each sampled point may be estimated, and non-compliant sections 1132 of the curve may be detected. Having found the closest polyline vertices (points A and B in spline 1130A) where the curve is compliant, the spline curve may be trimmed back to be a little further away from the midpoint 1134 of the polyline segment where the curve is compliant for 2.5 axis manufacturing, thereby forming a trimmed curve 1130B.

[0187] There are several options for making the non-compliant section 1132 compliant with the minimum tool radius. The gap can be refitted using a spline with a larger fitting tolerance to bring it below the high curvature limit, or an arc with a compliant curvature can be fitted into the gap. Fitting the arc in an optimal way while keeping the tangent line connected to the existing curve is a known problem with the described solution. Regardless of which method is used, at least one line segment of at least one of the smooth curves is replaced with a new line segment 1136 to reduce the curvature of the at least one line segment to less than or equal to the curvature of the smallest milling tool available, thereby forming a new spline 1130C that meets the curvature requirements of milling.

[0188] Return to Fig.10 Regardless of whether obtaining 1010 includes generating an initial model or generating a first model from an initial model, one or more of the smooth curves fitted 1014 to the extracted 1012 broken line profiles are modified 1020, for example, by one or more programs 116, 904. The modification 1020 is performed to facilitate the 2.5-axis subtractive manufacturing process, and the modification 1020 may include detecting 1022 at least a portion of a first smooth curve of a first layer in the discrete height layers that is almost coincident with at least a portion of a second smooth curve of a second layer in the discrete height layers. To achieve almost coincidence, the lateral distance (perpendicular to the milling direction) between the portion of the first smooth curve and the portion of the second smooth curve should be less than a threshold value determined according to the 2.5-axis subtractive manufacturing process. For example, the threshold value can be set to the minimum tool radius available in the 2.5-axis subtractive manufacturing process. In some implementations, a tolerance of 1 / 1000 of the grid box size is used. It should be noted that it is desirable to maintain the physical properties of the solution, and by selecting this tolerance, the volume of the solution can be expected to vary within a given ratio. In some implementations, the base tolerance for a grid with size (dx, dy, dz) and z attack direction may be set to a value of sqrt(dx^2+dy^2) / 1000. Additionally, it should be noted that thin shelves may be detected by calculating the perimeter and area of ​​a planar face.

[0189] Each such portion of the first smooth curve in one height layer that is nearly coincident with a portion of the second smooth curve in another height layer can be replaced 1024 with a portion from the second smooth curve, thereby removing the thin shelf that would otherwise be produced in the output 3D model. Thus, the contour curves are aligned with each other to at least prevent the thin shelf from appearing between layers, thereby facilitating the 2.5-axis subtractive manufacturing process. Additional detection and alignment techniques for contours that can be performed as part of the modification 1020 are described in further detail below.

[0190] For example, the first model may include one or more modeled entities that are to be retained in the solid 3D model of the object. As described above, these one or more modeled entities may be retention volumes with one or more obstruction areas that are input into the generative design process or created by a user for a scanned object (a mesh or level set representation). In either case, at least one extrusion direction (corresponding to at least one milling direction) is obtained (automatically received or determined). Modifying 1020 may include detecting 1022 at least a portion of another smooth curve of a first layer in the discrete height layers that is nearly coincident with at least a portion of the one or more modeled entities, and modifying 1020 may include aligning a portion of another smooth curve of a first layer in the discrete height layers with at least a portion of the one or more modeled entities (e.g., by replacing 1024 a portion of the curve that is not aligned with its adjacent retention volume).

[0191] After modification 1020, an editable model of the object is prepared 1030, for example, by one or more programs 116, 904. The editable model may include a series of construction steps represented by a parametric feature history, wherein the series of construction steps includes combining extruded versions of smooth curves to form a solid three-dimensional (3D) model of the object in a boundary representation format. The smooth curves from each height layer form planar rings, which can be extruded 1032 into solid B-Rep bodies. The bodies from each layer (extruded bodies and any retained bodies) can then be combined 1034 using B-Rep Boolean operations. It should be noted that the extrusion 1032 and combination 1034 performed using the parametric feature history can be performed in a positive manner (adding the extruded bodies of each layer together) or in a negative manner (subtracting the extruded bodies of each layer from the extruded blank profile). In addition, in some implementations, using the parametric feature history, holes whose axes are not parallel to the milling direction can be created on the final body, and / or fillets can be added to concave edges on the final body.

[0192] The parametric feature history may also include a sketch feature associated with one or more of the discrete height layers (e.g., a sketch user interface that can be used for each height layer profile). A user interface element that displays the sketch feature associated with the editable model may be presented 1040 to a display screen, for example, by one or more programs 116, 904. This allows a user to easily select a height layer and edit the profile of the layer using the sketch feature in the user interface as needed. Upon receiving user input, a subset of the smooth curves in at least one of the discrete height layers is reshaped 1050, for example, by one or more programs 116, 904, to change the solid 3D model. After shaping 1050, a series of construction steps represented by the parametric feature history are replayed 1055, for example, by one or more programs 116, 904, so as to construct a solid 3D model of the object as changed by the user input in a boundary representation format, and the process returns to presenting 1040 revised sketch user interface elements, potentially in conjunction with presenting a visualization of the updated 3D model.

[0193] Thus, the user can freely modify the different profiles of the corresponding height layers of the object to be manufactured using the 2.5-axis subtractive manufacturing process, and in response to these user edits, the parametric feature history automatically reconstructs the modified 3D model. This creates an easy-to-use editor interface for the user that allows rapid changes to the 3D model while also ensuring that the reconstructed 3D model after each edit will remain manufacturable using the 2.5-axis subtractive manufacturing process. It should also be noted that this sketch user interface easily uses the retention bodies (and potentially obstruction areas) that also exist in the 3D model. The shaping 1050 may include constraining the changes to at least a portion of the smooth curve that is nearly coincident with the retention body to remain tangent to and in contact with the retention body, and a series of construction steps may include combining an extruded version of the smooth curve with one or more retention body entities in response to each user input received to form a solid 3D model of the object in a boundary representation.

[0194] Once the design is approved, Fig.10 The process may be, for example, provided 1060 by one or more programs 116, 904 based on the user's editing of a 3D model of the object for 2.5-axis subtractive manufacturing of a physical structure. Providing 1060 may involve sending or saving the 3D model to a permanent storage device for use in manufacturing a physical structure corresponding to the object using an SM manufacturing system. In some implementations, providing 1060 involves using the 3D model to generate 1062 tool path instructions for one or more computer-controlled SM manufacturing systems, and using the tool path instructions generated for the SM machine by one or more computer-controlled SM manufacturing systems to manufacture 1064 at least a portion of the physical structure corresponding to the object. In some implementations, providing 1060 may include using the generated 1062 tool path instructions by a subtractive manufacturing machine to manufacture a mold of the physical structure.

[0195] Fig.11E A graphical representation of an example of the process of preparing 1030 an editable model of an object, receiving user input to a sketcher user interface element to reshape layers of a 3D model of the object, and replaying a parametric feature history to reconstruct the 3D model of the object is shown. In this example, the outermost contours of the geometry (when viewed along the milling direction) are first extracted. Fig.11E , an outline 1140 of a mesh geometry 1142 with a pin hole is shown (when viewed along the milling direction) that is included with the B-Rep retaining body. This outermost outline 1140 can be understood as a "blank" outline because it represents the blank shape from which material can be subtracted to obtain the final shape.

[0196] FIG. 11 also shows a B-Rep extrusion 1144 of a stock profile 1140 of a mesh geometry 1142. Thus, a strategy could be to simulate how a 2.5 axis machine tool would go about cutting geometry out of a stock. The profiles of each height layer could be extracted, and curves could be fit to these profiles (as described in this document). The fitted curves form a ring, and each ring could be made into a face, and each face could be extruded and cut out 1146 from the stock 1144. Thus, as Fig.11E As shown, layers can be cut from the stock shape using solid extrusions of the fitted contours (i.e., negative assembly using parameterized feature history) to construct the 3D model 1148. In some implementations, the solid extrusions of the fitted contours are added together (i.e., positive assembly using parameterized feature history) to achieve the 3D model 1148.

[0197] To align the curves, the profile can be aligned to a curve previously fitted in the process. For example, when fitting the profile 1150 of the first layer, portions of the profile are identified as coinciding with a) retained geometry 1152 that already has an accurate cross-sectional curve and b) a profile 1154 that has already been fitted (when projected onto the same cross-sectional plane). It should be noted that using accurate retained profiles of retained geometry 1152 and sharing curve geometry with previously fitted profiles 1154 facilitates 2.5-axis subtractive manufacturing. By detecting when a profile is nearly coincident with retained geometry or a previously fitted profile, the curve can be reused to ensure that the fitted profile is aligned with the other profiles or retained geometry. Continuing to use this fitting and cutting strategy for all layers will produce the final B-Rep conversion model 1160.

[0198] Although this B-Rep output can be edited using conventional direct editing tools, for example by one or more programs 116, 904, this will involve directly manipulating the edges and faces of the B-Rep to modify its shape. By decomposing the B-Rep model into a series of construction steps and representing these steps in a parametric feature history, an output in a more conveniently editable form can be produced. One such construction step can be extruding the blank profile into an initial solid body. This step can be represented in the parametric feature history as a sketch feature 1170, which is completed by editable curves and constraints to maintain tangency and contact with the retained body. It should be noted that the sketch feature 1170 can be used with any of the contours of the 3D model, and preparing the parametric editable model in this way helps the user to make changes to the 3D model by editing the contour without affecting the retained body. When the user edits the sketch, the parametric feature history replays the construction of the entire modified B-Rep. In addition, a concave fillet can also be automatically added as a parametric fillet feature, which represents the envelope created by the tool shape when the tool cuts into the concave corner.

[0199] Fig.12 Another example of a process for generating a watertight boundary representation for an object to be manufactured using a 2.5-axis subtractive manufacturing process is shown. A first 3D model of an object is obtained 1200, for example, by one or more programs 116, 904. As described above, the first 3D model may include meshes in various formats, may be output from one or more generative design processes or other processes (e.g., from a 3D scanner), may be associated with one or more modeling entities. For example, the first 3D model may be a triangular mesh 1202 (or level set geometry, such as level set voxel data) generated by a generative design solver, the triangular mesh having been constrained to produce a 2.5-axis millable shape, where such output from the generative design solver needs to be converted into an editable B-Rep volume.

[0200] The goal can be to obtain a polygonal mesh representation of a solid body in 2.5 dimensions (2.5D) and convert it into a boundary representation (B-Rep) format with a smooth fitting NURBS (Non-Uniform Rational B-Spline) surface. The goal is to produce a B-Rep whose shape can be efficiently manufactured using a CNC milling machine capable of 2.5-axis milling. Obtaining 1200 can involve receiving or generating a 3D model, for example, any polygonal mesh geometry or level set geometry that can be plane-sliced ​​to produce closed polyline loops. Therefore, the input can also include pre-existing B-Rep bodies (referred to as "retained" bodies 1204) representing sub-areas of the input mesh to be included in the output B-Rep without approximation. The input can also include pre-existing B-Rep bodies (referred to as "obstacle" bodies 1206) representing areas that must be free of material. The retained bodies allow the user to precisely specify the geometry that needs to be present in the final body, rather than having the algorithm include its approximation. This is important in manufacturing to ensure that the manufactured parts fit accurately with other parts in the assembly.

[0201] Obtaining 1200 may include the generative design process itself, which may employ the systems and techniques described herein or other methods of generating designs that are more easily machined on 2.5-axis CNC machines. In addition, other sources of the first 3D model are also possible; the input does not need to come from a generative design topology optimization process with a 2.5-axis constraint. For example, the input to the conversion process may include a depth map output from a machine learning (ML) generative adversarial network (GAN) image (or a similar output from another process) or a point cloud output from a 3D scan performed by an object scanner (or a mesh converted therefrom) for reverse engineering. However, the result of the conversion process will still be a 2.5D shape. A 2.5D shape is a subset of a 3D shape, in which the geometry can be constructed as a union of 3D regions each formed by linearly extruding the faces of a plane in a common bidirectional parallel direction (referred to as a 2.5D axis), and in which the faces of the plane have normals that are bidirectionally parallel to the 2.5D axis.

[0202] Obtaining 1200 may also include receiving or determining other information associated with the object to be manufactured and therefore the first 3D model, such as described above with respect to obtaining 1010, for example, one or more milling directions, milling height layer positions, and at least one milling tool radius (e.g., a minimum milling tool radius or a minimum milling tool radius available when cutting a part to form the object during manufacturing) for a 2.5-axis subtractive manufacturing process. The received milling tool radius information may be provided per layer, but the algorithm may be designed to operate without receiving any milling tool radius information, in which case a default minimum tool radius (or no minimum tool radius) may be used during the conversion.

[0203] Regardless of how the 3D model is obtained 1200, the shape defined by the 3D model may be divided 1210, for example, by one or more programs 116, 904, into layers 1212 as determined by levels (e.g., as Fig.12 Each 2.5D shape can be decomposed into a union of layers, each of which is of constant thickness everywhere (but not necessarily the same thickness for each layer). If the application provides the layer levels 1214, the layering is straightforward. If the layer levels are not provided, the 2.5D axis is oriented in the z direction, and the layer levels are obtained by performing a clustering algorithm on the z coordinates of the mesh vertices.

[0204] Next, mesh layering 1220 may be performed, for example, by one or more programs 116, 904. The goal is to produce a shape that can be manufactured using a 2.5-axis milling tool path with an attack direction that can be aligned along the z-axis. Depending on the settings, the allowed attack directions may include +z, -z, and +z or -z (bidirectional). The output of mesh layering 1220 may be a layered mesh in which any face must be vertical or horizontal, any point of a horizontal face must be visible (no undercuts) from the allowed attack directions, and any horizontal face must lie on a set of predefined and finite planes. It should be noted that the last of these conditions is an additional constraint, and it may be that the input mesh does not comply with these constraints.

[0205] Fig.13A An example of mesh layering is shown. An input mesh 1300 has associated layer levels 1302, horizontal planes 1304 that are not aligned with any layer level, undercuts 1306, and non-horizontal planes 1308. The goal is to transform the input mesh 1300 into an output mesh 1310 that is the smallest mesh that meets all the necessary conditions for the rest of the algorithm to work. Note that volume 1312 has been filled by mesh layering 1220.

[0206] The grid layering 1220 can receive a level set representation of an object, or receive a mesh representation of an object (e.g., from a generative design process) and convert the mesh representation into a level set representation of the object. A level set representation is a volume representation that defines an interior and an exterior, thereby creating a grid in the space of the grid and storing the distance to the grid surface at each point of the grid. For example, consider a grid with integer spacing in three dimensions and a sphere with a radius R=10 centered at the origin. Any point of the grid (i,j,k) will store the distance

[0207]

[0208] Modify one or more level set values in the level set representation of an object in each of one or more milling directions specified for a 2.5-axis subtractive manufacturing process to remove undercuts. For example, the level set values for each z-column can be modified to remove undercuts. Undercuts exist only if the distance function has multiple local minima along z (i.e., the milling direction). To avoid undercuts, the distance function must have only one global minimum along any z-column. The algorithm can pick the global minimum and make the function not increase before that point and not decrease after that point. To make the function not increase to achieve z < z_min, the lower limit value can be propagated as follows:

[0209] if(D(z = i)>D(z = i+1)) D(z = i+1) = D(z = i); / / if i+1 < i_min

[0210] And to make the function not decrease to achieve z > z_min, the upper limit value can be propagated:

[0211] if(D(z = i)<D(z = i+1)) D(z = i) = D(z = i+1); / / if i > i_min.

[0212] Fig. 13B An example of modifying level set values to remove undercuts in a grid is shown. As shown, the distance to the volume boundary is modified to remove unwanted local minima and protrusions while maintaining the global minimum; compare the distance values in the top part of Fig. 13B with the distance values in the bottom part of Fig. 13B , which shows protrusion correction.

[0213] Next, modify one or more additional level set values in the level set representation of the object to move the faces of one or more planes of the object up to the height level of the corresponding one in the discrete height levels of the faces of the planes. For example, considering the z-column distance values, it may be required that the distance be uniform between layer levels. This means that the value of the distance function must not change within a layer and must maintain the minimum value so that material can be added all the time. Fig. 13C An example of modifying level set values to move the face of a plane up to a layer level is shown. Fig. 13C The upper part of Fig. 13C shows the unwanted slope generated by the distance values. The bottom part of

[0214] shows how to move the distance values to make the distance function uniform between the layers of the layer levels. Thereafter, the modified level set representation of the object can be converted back to the mesh representation of the object, for example, using the marching cubes algorithm. Fig.12, mesh layering 1220 may be followed, for example, by one or more programs 116, 904, by volume-retaining layering 1225. Layering 1225 of any retained B-Rep has the same goal as layering 1220 of the mesh, but for the B-Rep, accurate geometry should be produced, so the same volumetric tools as for the mesh cannot be used. Volume layering 1225 may consist of two parts: first, homogenizing the layer; and second, removing undercuts. Fig.13D An example of a process of homogenizing layers of an input modeling entity preserving regions is shown. For ease of illustration, this example of volume stratification is shown in 2D.

[0215] For each initial volume 1340 (each modeled retained entity), the layer of the volume 1340 is selected 1342 using the Boolean intersection of the volume 1340 with the layer (layer 4 in the example shown). This intersection produces a portion of the modeled entity 1340. This portion is then moved 1344 completely below the layer in the milling direction, and this portion of the volume 1340 is then swept 1346 vertically (opposite to the milling direction) through the layer to produce the swept entity. The layer of this swept entity is selected 1348 using the Boolean intersection of the swept entity with the layer (layer 4 in the example shown). This intersection produces a portion of the swept volume that is located inside the layer, which is used as a replacement for the portion of the modeled entity 1340 in the layer. These operations can be applied to all layers, and the output volumes can be combined to form a replacement entity volume 1350. Fig.13E Shown is applied to 3D preserved volume Fig.13D The input preserved volume modeling entity 1352 is homogenized 1354 within three layers, thereby producing an output volume modeling entity 1356.

[0216] After homogenizing the layers between layers (body_0), the result may include undercuts, which should be removed. Fig.13FAn example of the process of removing undercuts from a retained body that has been homogenized within a defined layer is shown (in 2D for ease of illustration). One or more upper side faces of a modeled entity 1360 (having a homogenized layer) are swept upward 1362 in the direction opposite to the milling direction to the topmost level of the discrete height layers to produce one or more upper side swept entities. All top faces of body_0 are swept in the +z direction (sweeped upward to the end of the layer) to produce a consistent (body_1). Any lower side faces of the modeled entity are swept downward 1364 in the milling direction to the bottommost level of the discrete height layers to produce one or more lower side swept entities. All bottom faces are swept downward in the -z direction (sweeped downward to the end of the layer) to produce a consistent (body_2). The body with the filled undercuts is then the Boolean intersection 1366 of the swept upward and swept downward, which is then combined 1368 with the modeled entity 1360: body_0+(body_1*body_2), where "+" is a Boolean union and "*" is a Boolean intersection.

[0217] Fig.13E Figure 2 shows the effect of applying a 3D retention volume that has been homogenized across layers. Fig.13F The undercuts of the uniform retention volume 1354 have been removed 1370, resulting in an output volume modeling entity 1372. Thus, when one or more undercut fill entities (generated by intersecting one or more up-swept entities with one or more down-swept entities) are combined with the cross-layer homogenized modeling entity, the retention volume layering applied to the 3D retention volume is complete.

[0218] Return to Fig.12 After mesh layering 1220 and volume layering 1225, polyline slices are extracted 1230 for each layer, for example by one or more programs 116, 904. This extraction can be done from the mesh or from a level set representation, as described above, for example in conjunction with Fig. 11B Each of the polylines is then segmented 1235, for example by one or more programs 116, 904, into regions associated with retained geometry and regions not associated with retained geometry (e.g., portions of organic shapes generated by a generative design process). The input shape data may be a mesh representation of an object or a level set representation of an object, which may be converted into a mesh representation for processing to generate polylines, for example, by slicing the mesh representation of the object along a milling direction with planes located within corresponding ones of the discrete height layers, each of the planes being perpendicular to the milling direction. It should be noted that the layered mesh and layered retained volume (as described above) may be used as input for contour slicing to generate each polyline.

[0219] Fig.14AAn example of polylines for slicing of an object and the initial segmentation of the polylines is shown. A layer of the mesh 1202 has been sliced ​​to produce polylines 1400, shown with reference to the retention volume 1204. It should be noted that the final result should have a completely smooth connection between the organic area and the retention. In order to achieve this result, the polylines of the layer need to be properly segmented. By slicing the mesh with a plane parallel to the horizontal and located at the midpoint of each layer, the polyline cross-sections are extracted, as shown in FIG. Fig.14A Then, segmentation divides the polyline 1400 into different regions, where the segmentation is based on their proximity to the retention volume. For example, line segments 1402 may be created for the polyline 1400.

[0220] In this example, the segmentation does not satisfy an important constraint. At 1404, a small, undesirable line segment 1402 has been created on the polyline 1400 adjacent to the retention volume 1204. Such small areas due to mesh noise should be avoided. To this end, a large tolerance can be used during segmentation to assign polyline vertices and line segments to any retention volumes that exist. For example, if the basic tolerance is tol=diagonal(mesh bounding box) / 1000, the large tolerance can be large_tolerance=sqrt(10)*tol. Typically, the larger tolerance should be at least several times (2, 3, 4, or 5 times) the basic tolerance used. In addition, in this example, a gap 1406 has been formed between the organic region and the retention volume. This is also undesirable, and therefore the organic regions should be expanded until they actually intersect any retention volumes or until the polyline is tangent to the retention volume.

[0221] Fig. 14B and Fig. 14C The process of generating line segments of a polyline is shown. The polyline is segmented into a first line segment assigned to a reserved volume and a second line segment not assigned to the reserved volume, wherein the segmentation uses a tolerance value when assigning polyline vertices to line segments that are incorrectly assigned to the reserved volume. Therefore, if Fig. 14B As shown, the initial segmentation of the fold line creates segmentation regions 1410 that are fully associated with the retaining volume even if the slices of the mesh do not closely match the shape of the retaining volume, and the initial segmentation of the fold line creates segmentation regions 1412 for organic portions of the mesh that are not close to the retaining volume.

[0222] Therefore, the line segments in these regions 1412 are extended 1414, 1416 until they intersect the retention body. In one case shown, the line segment extension moves the endpoint to the intersection 1414 with the retention body, while in the other case shown, the line segment extension moves the endpoint to the tangency with the retention body. The organic portion of the polyline can then be approximated with the polyline, and the tangency condition is easily added when combined with the contour of the retention body. The result is an overall smooth contour, where the generated contour 1420 of the organic region smoothly meets the contour 1422 of the retention body (in G1 continuity).

[0223] Furthermore, during expansion, care should be taken to keep the ends aligned between layers. Fig. 14C As shown, the points 1424 that meet the retention body at three different levels should be on the same vertical line. To achieve this result, the algorithm for extending the organic region can check whether it has reached a point that is the endpoint of the contour at the previous layer, and in this case, the extension process can be stopped. Therefore, a check can be made to align the intersection of any extended line segments that intersect the same retention body. In addition, in some implementations, if the change in shape is reasonable, the organic region can be reduced to align to the endpoint defined by the previous layer.

[0224] Return to Fig.12 After the polyline is segmented 1235, a smooth spline curve may be fitted 1240 to the organic regions of the polyline, for example by one or more programs 116, 904. This may be done using the above combined Fig.10 The smooth curves 1245 may then be aligned between the layers, for example by one or more programs 116, 904. Since the extruded bodies of different layers will be combined together to form the final 3D model, if the bodies to be combined share some vertical walls, it is very necessary to have the walls perfectly aligned so that the shapes can be efficiently machined with a 2.5-axis milling tool path. To this end, portions of the profile curves used to extrude the bodies are aligned with each other. Fig.15A An example of an extruded profile 1500 of an organic portion of a mesh is shown, wherein the extruded profile 1500 shares sidewalls 1502 .

[0225] In some implementations, the curve alignment process begins by identifying one curve as the "parent" and the other curve as the "child". Note that both curves are planar and their planes are parallel. The child curve is translated to the parent plane to facilitate deviation checking. The alignment range between the parent and child can be found by checking the deviation of the parent and child against a reasonable tolerance. For example, a reasonable tolerance can be diagonal(mesh box) / 1000.

[0226] Splits the subcurve at the start / end of the alignment range, and this splits the subcurve into two subsections. The subsection corresponding to the alignment range is replaced with a copy of the subsection of the parent corresponding to the same alignment range. The other subsection is then slightly modified near its end so that the newly copied subsection can be smoothly joined into it. Note that the parent curve will not undergo any changes in this process, and the smoothness of the subcurve is maintained at the end of the alignment range in tangent continuity.

[0227] Fig. 15B Shown from Fig.15A An example of this process for a common sidewall 1502 of a polyline. A subcurve of a higher level is split into two sub-portions. Sub-portion 1504 is replaced by a copy from the parent curve, and sub-portion 1506 is refitted with the tangent constraint. Note that the tangent vector is derived from the parent curve. The algorithm can take as input: (1) a curve C_Organic (or a subcurve), i.e., an approximation of the organic portion of the polyline; (2) a reference profile, i.e., a series of curves from the previous level (call these curves C_Ref(i), I=1…n); (3) curves C_Design from adjacent C_Organic of segments of the preserved body (they can be 0 (in the case that C_Organic is closed), 1, or 2); (4) a tolerance that determines when curves overlap. Fig. 15C C_Organic and C_Ref are shown where the curve alignment creates perfect overlap and G1 continuity at the point where the shared sidewall diverges.

[0228] The algorithm can compare C_Organic with C_Ref(i), I=1...n, and perform the following operations for each comparison:

[0229] 1) Find a list of ranges where C_Organic and C_Ref(i) overlap; this can be achieved by finding the intersection point and proceeding from the intersection point using small steps until the curves diverge; in this way, a suitable range can be found and the algorithm for expanding the range can take into account both the distance between curves and the angle between tangents.

[0230] 2) For each overlapping range, split the curve C_Organic into 3 parts: C_before_range, C_inside_range and C_after_range. It should be noted that the before and after ranges can be empty.

[0231] 3) Replace C_inside_range with the corresponding part C_Ref(i).

[0232] 4) Modify C_before_range and C_after_range to be G1 continuous with C_inside_range; these curves are NURBS curves that can be modified by moving control points; it should be noted that in the special case where the curve is a 1 degree NURBS (a line), the degree should be increased to be able to impose the G1 condition.

[0233] 5) C_Organic is now replaced with the curve list and the process continues for each curve in the list.

[0234] After this process, C_Organic is replaced with a list of curves. These curves may not be connected. This usually happens when C_Organic coincides with two consecutive curves from the previous layer. In this case, small gaps may be created when trying to avoid creating very small curves.

[0235] This happens when the overlap range that matches C_Ref(0) is a little shorter than the length of C_Ref(0). Fig.15D An example of a gap that has been created is shown. To solve this problem, the algorithm can use a variety of solutions to fill such gaps, such as by using a reference contour, extending and intersecting the curves, or adding line segments. Another problem with curve alignment is the creation of self-intersecting contours, i.e., when modifying a curve, it may be possible to make it intersect itself or other curves. Fig.15E An example of this problem and a solution based on curve trimming is shown. Graph 1520 shows the input C_Organic (solid line) and the adjacent C_Preserves (dashed line). Graph 1522 shows the possible result after the curves are aligned. Graph 1524 shows the result of curve trimming to correct the self-intersection. It should be noted that G1 continuity may be lost using this method.

[0236] Return to Fig.12 After aligning the curves 1245 between the layers, a profile is created 1250 from the curves, for example by one or more programs 116, 904. This can be done using known techniques. Milling tool accessibility can then be imposed 1255 on the profile, for example by one or more programs 116, 904. In some implementations, this involves, for each set of smooth curves within one of the discrete height layers, offsetting the set of smooth curves outwardly by an amount based on the curvature of the smallest milling tool available, and then offsetting the set of smooth curves inwardly by that amount, while using arcs to close any resulting gaps. This is another example of how concave curvature compliance can be imposed.

[0237] The goal of this stage of the algorithm is to ensure that the shape of the profile allows a milling tool of a given radius to reach and cut the geometry without undesirably gouging distant areas. Geometrically, this means not exceeding the curvature of a given layer, as determined by the tool radius, and ensuring that the gaps between surfaces are never too small for the tool to pass. Different layers can be assigned different tool radii and therefore different minimum curvature radii and gaps.

[0238] The strategy is to offset the contours, which are closed planar rings of the curve, outward and then inward, trimming off any inversions or self-intersections of the curve to remove areas of excessive curvature and narrow gaps. For each layer, there are one or more contours and a maximum allowed radius of curvature K, each of which is a closed planar ring of the curve. It is important that the contours include the entire input geometry (including the preserved area). Each ring is processed as follows:

[0239] 1) Create a directed WIRE body from the loop. This simply adds topology to the curve so that there are vertices that join the curve and have a consistent direction going clockwise around the loop so that the material is on the left side of the curve.

[0240] 2) Apply the PLANAR_WIRE_OFFSET operation to WIRE, giving it an outward offset distance of 1 / K. This is a common algorithm found in many geometric modeling kernels and CAD applications. This algorithm should do the following:

[0241] a. With the plane normal N of a directed WIRE wrapped around N in a right-handed direction, shift each point P on each curve by a given distance 1 / K in the direction of the vector T x N, where T is the tangent vector from P to the new point P'. The union of all P's forms a new curve.

[0242] b. Any protruding sharp corners in the WIRE will separate, leaving a gap. The algorithm should close the gap by at least one method (usually by the natural extension of an arc or curve).

[0243] c. The new curve forms a new, closed, offset WIRE.

[0244] d. Detect and trim self-intersections in WIRE, discarding the inverted portion of the curve (inverted means the direction of the curve is reversed due to an offset) in case of high curvature areas, and discarding the inner portion of the curve (the portion inside the ring) to achieve global intersection.

[0245] 3) This ensures that the offset WIRE does not contain any curve geometry with a curvature exceeding K.

[0246] 4) Use the same algorithm to offset the WIRE back the same distance 1 / K, but this time make sure that any gap created by the offset between two adjacent curve endpoints is closed using a circular arc.

[0247] 5) In the final WIRE, the areas with excessively high curvature in the original WIRE have been replaced by arcs with a curvature of K, and the gaps that are too small (<2 / K) have been reduced and removed.

[0248] As a result, the final profile has replaced areas of high curvature with arcs. The effect is to add some material to make the part machineable with a milling tool of a specific radius. It should be noted that this process of shifting outwards and inwards will also eliminate small holes that the tool cannot reach.

[0249] After the profile has been modified to impose 1255 milling tool accessibility, the profile is extruded 1260 to the correct height for its respective layer, e.g., by the one or more programs 116, 904. The extruded body is joined 1270 with the retained body, e.g., by Boolean operations, by the one or more programs 116, 904, to form a 3D model 1272. A parametric editable model 1282 is constructed 1280, e.g., by the one or more programs 116, 904, and B-Rep solid bodies and construction geometry supporting the parametric recipe are output 1290, e.g., by the one or more programs 116, 904.

[0250] When presenting the results of the B-Rep conversion in a parametric CAD application, the output can be made in a highly editable form using the construction geometry provided by the algorithm. Each layer of the 2.5D model can be represented by an extrusion of a set of contours defined in the 2D sketch. Fig.16 An example of a 2D sketch user interface element showing a parametric editable model. When the final solid is constructed as a union of extrusions, if the user edits any of the sketch curves, the parametric feature history will replay the construction of the entire modified solid body. The construction of a parametric feature recipe can include the following:

[0251] 1) A set of planar faces representing the contour of each layer is transferred to the sketch environment of the target CAD application. Each curve of the contour boundary is converted into a corresponding curve type supported by the sketch environment.

[0252] 2) Display any simple curve (e.g., line, arc) with endpoints available in the UI for the user to select and move in the sketch plane. Edit the spline curve by making control points available in the UI for the user to move in the sketch plane.

[0253] 3) Curves originating from the intersection of a layer with a preserved volume are "frozen" in the sketch so that the user cannot edit them. If the user were to edit them, the user would run the risk of creating small, undesirable gaps between the organic and preserved regions that would cause difficulties with Boolean operations or divide the volume into multiple blocks.

[0254] 4) Use the CAD application's EXTRUDE feature to extrude each profile into a solid with the extents required for each layer. The extents are stored as parameters in the Extrude feature, which provides a way for the user to edit the extents for each layer if needed. Add each of these EXTRUDE features to the feature recipe.

[0255] 5) Union the solids from each extrusion layer via a Boolean operation to form a single solid representing the organic region. Add this COMBINE feature to the recipe.

[0256] 6) Union the organic regions with all retained bodies via Boolean operations to form a complete solid. Add this COMBINE feature to the recipe.

[0257] 7) Subtract the obstacle volume from the complete solid via Boolean operations to form the final solid. Add this SUBTRACT feature to the recipe. Note: If the retain and obstacle regions overlap, whether you first retain and then subtract the obstacle (or vice versa) will affect which region takes precedence in the result.

[0258] 8) A parametric feature recipe is then created in the CAD application: EXTRUDE (sketch), COMBINE (extrude), COMBINE (retain), SUBTRACT (obstruction) and presented to the user, each feature being editable.

[0259] Editability is achieved by creating not only the final B-Rep solid but also the construction geometry that leads to this final solid. This editability includes profile geometry that can be extruded to different extents to produce components that can then be joined through Boolean operations to form the final desired body. The profile geometry is in the form of splines that can be edited in the CAD application, giving the user the ability to change the shape of the final body. The returned construction geometry can be assembled as a feature recipe in any modern parametric CAD application.

[0260] Additionally, since the milling process relies on a cylindrical tool to cut through the material, the shape of the object should be such that all surfaces of the object can be reached by the selected tool without gouging. This is equivalent to requiring that surfaces should not have a radius of curvature smaller than the tool radius, and that any gaps between surfaces cannot be too narrow to be reached by the tool. As detailed above, these issues have been addressed and an algorithm ensures that: with the minimum tool radius, the B-Rep geometry does not exceed this curvature limit and that all gaps between surfaces can be reached by the tool, making the object suitable for manufacturing by a milling machine. As a result, the output geometry can be edited in parametric CAD applications, is easy and efficient to manufacture using 2.5-axis milling, and directly includes any existing B-Rep retained volumes without loss of accuracy.

[0261] The embodiments and functional operations of the present subject matter described in this specification can be implemented in digital electronic circuits, or in the form of computer software, firmware or hardware (including the structures disclosed in this specification and their structural equivalents), or in a combination of one or more of them. The embodiments of the subject matter described in this specification can be implemented using one or more modules of computer program instructions encoded on a non-temporary computer-readable medium for execution by a data processing device or for controlling the operation of the data processing device. Computer-readable media can be manufactured products, such as hard disk drives in computer systems, or optical disks sold through retail channels, or embedded systems. Computer-readable media can be obtained separately and then encoded with one or more modules of computer program instructions, for example, after transmitting one or more modules of computer program instructions through a wired or wireless network. Computer-readable media can be a combination of a machine-readable storage device, a machine-readable storage substrate, a memory device, or one or more of them.

[0262] The term "data processing apparatus" encompasses all apparatus, devices and machines for processing data, including by way of example a programmable processor, a computer or a plurality of processors or computers. In addition to hardware, an apparatus may include hardware, code that creates an execution environment for the computer program in question, for example, code constituting processor firmware, a protocol stack, a database management system, an operating system, a runtime environment, or a combination of one or more of them. In addition, an apparatus may employ a variety of different computing model infrastructures, such as network services, distributed computing, and grid computing infrastructures.

[0263] A computer program (also referred to as a program, module, software application, script or code) can be written in any suitable form of programming language, including compiled or interpreted languages, declarative or procedural languages, and can be deployed in any suitable form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program may be stored in a portion of a file that stores other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple collaborative files (e.g., files storing one or more modules, subroutines or code portions). A computer program may be deployed to execute on one computer or on multiple computers, which are located at one site or distributed across multiple sites and interconnected by a communication network.

[0264] The processes and logic flows described in this specification may be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows may also be performed by, and the apparatus may be implemented as, a special purpose logic circuit, such as an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

[0265] The processor that is suitable for executing computer programs includes any one or more processors of general and special two kinds of microprocessors and any kind of digital computer by way of example.Usually, the processor will receive instructions and data from a read-only memory or a random access memory or both.The basic element of a computer is a processor for executing instructions and one or more memory devices for storing instructions and data.The computer will also include one or more large-capacity storage devices (for example, disks, magneto-optical disks, or optical disks) for storing data, or be operably coupled to receive data from it, or to transmit data to it, or to perform both.However, the computer does not need to have such devices.In addition, the computer can be embedded in another device, for example, a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device (for example, a universal serial bus (USB) flash drive), just to name a few. Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media and storage devices, including by way of example semiconductor memory devices such as EPROM (erasable programmable read-only memory), EEPROM (electrically erasable programmable read-only memory) and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and memory may be supplemented by, or incorporated in, special purpose logic circuitry.

[0266] To provide interaction with a user, embodiments of the subject matter described in this description may be implemented on a computer having a display device (e.g., an LCD (liquid crystal display) display device, an OLED (organic light emitting diode) display device, or another monitor) for displaying information to the user, and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other types of devices may also be used to provide interaction with a user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and the input from the user (including voice, language, or tactile input) may be received in any form.

[0267] The computing system may include a client and a server. The client and the server are generally far away from each other, and usually interact through a communication network. The relationship between the client and the server is due to the computer program running on the corresponding computer and having a client-server relationship with each other. The embodiment of the theme described in this specification can be realized in a computing system, and the computing system includes a back-end component (for example, as a data server), or includes a middleware component (for example, an application server), or includes a front-end component (for example, a client computer with a graphical user interface or a browser user interface that a user can use to interact with the implementation of the theme described in this specification), or any combination of one or more such back ends, middleware or front-end components. The components of the system can be interconnected by digital data communication (for example, a communication network) of any suitable form or medium. The example of a communication network includes a local area network ("LAN") and a wide area network ("WAN"), an internetwork (for example, the Internet) and a peer-to-peer network (for example, a self-organized (ad hoc) peer-to-peer network).

[0268] Although this specification includes many implementation details, these details should not be interpreted as limitations on the scope of the content that is claimed or can be claimed, but should be interpreted as descriptions of features specific to a particular embodiment of the disclosed subject matter. Certain features described in the context of a separate embodiment in this specification may also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any suitable sub-combination. In addition, although the features may be described above as acting in the form of certain combinations, and even initially claimed in this way, one or more features from the claimed combination may be removed from the combination in some cases, and the claimed combination may be directed to a sub-combination or a variation of the sub-combination.

[0269] Similarly, although operations are shown in a particular order in the accompanying drawings, this should not be understood as requiring that such operations be performed in the particular order shown or in a sequential order, or that all of the operations shown be performed to achieve the desired results. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of various system components in the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0270] Thus, certain embodiments of the invention have been described. Other embodiments are within the scope of the following claims. Additionally, the actions recited in the claims can be performed in a different order and still achieve desirable results.

Claims

1. A method for computer-aided design of a physical structure having flat regions in discrete height layers that facilitate 2.5-axis subtractive manufacturing, the method comprising: obtaining a first model of an object to be manufactured using a 2.5-axis subtractive manufacturing process, wherein the first model comprises smooth curves fit to contours representing the discrete height layers of the object to be manufactured using the 2.5-axis subtractive manufacturing process; modifying at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process, the modifying comprising detecting at least a portion of a first smooth curve of a first one of the discrete height levels that substantially coincides with at least a portion of a second smooth curve of a second one of the discrete height levels, and replacing the at least a portion of the first smooth curve of the first one of the discrete height levels with the at least a portion of the second smooth curve of the second one of the discrete height levels; as well as Providing an editable three-dimensional model of the object for use in fabricating a physical structure corresponding to the object using one or more computer-controlled manufacturing systems employing the 2.5-axis subtractive manufacturing process, wherein the editable three-dimensional model comprises a combination of extruded versions of the smooth curves.

2. The method of claim 1 , wherein obtaining the first model of the object comprises: receiving shape data corresponding to the object to be manufactured using the 2.5-axis subtractive manufacturing process; processing the shape data to generate polylines matching the contours representing the discrete height layers of the object; fitting the smooth curve to the broken line; as well as At least one line segment of at least one of the smooth curves is replaced for at least one of the discrete height levels to reduce a curvature of the at least one line segment to be less than or equal to a curvature of a smallest milling tool usable in the 2.5-axis subtractive manufacturing process.

3. The method of claim 2, wherein replacing at least one line segment of at least one of the smooth curves comprises: offsetting a set of smooth curves outward by an amount based on the curvature of the smallest milling tool available; as well as Offsets the set of smooth curves inward by this amount, using circular arcs to close any resulting gaps.

4. The method of claim 2, wherein the shape data comprises a level set distance field, and processing the shape data comprises: resampling the level set distance field to produce a two-dimensional level set grid at a slice plane corresponding to a current one of the discrete height layers, the slice plane being perpendicular to a milling direction to be used in the 2.5-axis subtractive manufacturing process; as well as A current layer polyline is extracted from the two-dimensional level set grid, the current layer polyline matching a current layer contour of the object at the slice plane.

5. The method of claim 2 , wherein obtaining the first model of the object comprises replacing at least another line segment of at least one of the smooth curves, the smooth curve comprising a point closer to a medial axis of a corresponding contour of the curve than a radius of the smallest milling tool that can be used in the 2.5-axis subtractive manufacturing process.

6. The method of claim 2, wherein the shape data comprises representations of one or more modeled entities to be retained in the editable three-dimensional model of the object, the mesh representation of the object is the shape data or is generated by processing the shape data, processing the shape data comprises slicing the mesh representation of the object along a milling direction by planes located within corresponding ones of the discrete height layers, each of the planes being perpendicular to the milling direction, and obtaining the first model of the object comprises: segmenting the polyline into a first set of line segments assigned to the one or more modeled entities and a second set of line segments not assigned to the one or more modeled entities, wherein the segmenting uses a tolerance value when assigning polyline vertices to line segments that were erroneous in assignment to the one or more modeled entities; as well as One or more line segments in the second group of line segments connected to one or more line segments in the first group of line segments are extended until the one or more line segments in the second group of line segments meet at an intersection point or a tangent point of at least one of the one or more modeling entities, while also checking to align the intersection point of any extended line segments that intersects at least one same modeling entity of the one or more modeling entities.

7. A system comprising: a non-transitory storage medium having instructions of a computer-aided design program stored thereon; as well as One or more data processing devices configured to run the instructions of the computer-aided design program to perform the method of any one of claims 1-6.

8. The system of claim 7, comprising a 2.5 axis subtractive manufacturing machine.

9. A non-transitory computer-readable medium encoding a computer-aided design program operable to cause one or more data processing devices to perform the method of any one of claims 1-6.

10. A method for computer-aided design of a physical structure having flat regions in discrete height layers that facilitate 2.5-axis subtractive manufacturing, the method comprising: Obtaining a first model of an object to be manufactured using a 2.5-axis subtractive manufacturing process, wherein the first model comprises: (i) smooth curves fitted to contours representing discrete height layers of the object to be manufactured using the 2.5-axis subtractive manufacturing process; and (ii) one or more modeled entities to be retained in the solid 3D model of the object; modifying at least one of the smooth curves to facilitate the 2.5-axis subtractive manufacturing process; preparing an editable model of the object, the editable model comprising a series of construction steps represented by a parametric feature history, wherein the series of construction steps comprises combining an extruded version of the smooth curve and the one or more modeled solids to form a solid three-dimensional model of the object, and the parametric feature history comprises a sketch feature associated with at least one of the discrete height layers; presenting a user interface element displaying said sketch features associated with said editable model; receiving user input via the user interface element; shaping a subset of the smooth curves in the at least one of the discrete height levels to alter the solid 3D model of the object in response to the user input; and The series of construction steps represented by the parametric feature history are replayed to construct the solid 3D model of the object as modified by the user input.

11. The method of claim 10, wherein modifying the at least one of the smooth curves comprises: detecting at least a portion of at least one first smooth curve of at least one of the discrete altitude layers that substantially coincides with at least a portion of the one or more modeled entities; as well as The at least a portion of the at least one first smooth curve is aligned with the at least a portion of the one or more modeled solids.

12. The method of claim 11, wherein shaping the subset of the smooth curves comprises constraining changes to the at least a portion of the at least one first smooth curve to maintain tangency and contact with the at least a portion of the one or more modeled solids.

13. The method of claim 10, wherein obtaining the first model of the object comprises: receiving a level set representation of the object, or receiving a mesh representation of the object and converting the mesh representation into a level set representation of the object; modifying one or more level set values ​​in the level set representation of the object in each of one or more milling directions specified for the 2.5-axis subtractive manufacturing process to remove undercuts; modifying one or more further level set values ​​in the level set representation of the object to move one or more planar faces of the object upward to a height level of a corresponding one of discrete height layers of the planar faces; as well as The modified level set representation of the object is converted into the smooth curve of the first model.

14. The method of claim 13, comprising: For each modeling entity of the one or more modeling entities: For each of the discrete height layers along the milling direction: intersecting the layer with the modeled entity to produce a portion of the modeled entity, moving said portion of said modeled entity completely below said layer in said milling direction, sweeping the portion of the modeled solid upward through the layer opposite to the milling direction to produce a swept solid, and intersecting the layer with the swept solid to produce a replacement of the portion of the modeled solid; sweeping one or more upper side surfaces of the modeled solid upwardly along the milling direction in the opposite direction to the milling direction to the topmost level of the discrete height layers to generate one or more upper side swept solids; sweeping any lower side surface of the modeled solid in the milling direction downward to the lowest level of the discrete height layers in the milling direction to generate one or more lower side swept solids; intersecting the one or more upper swept solids with the one or more lower swept solids to produce one or more undercut fill solids; and The one or more undercut fill entities are combined with the modeled entity.

15. The method of claim 10, comprising: receiving the one or more modeling entities; as well as One or more milling directions for the 2.5-axis subtractive manufacturing process are detected from the one or more modeled entities.

16. The method of claim 10, comprising: receiving the one or more modeling entities; as well as At least one of the discrete altitude levels is extracted from the one or more modeled entities.

17. The method of claim 10, wherein: The series of construction steps represented by the parametric feature history includes adding fillets to concave edges of the solid three-dimensional model of the object.

18. A system comprising: a non-transitory storage medium having instructions of a computer-aided design program stored thereon; as well as One or more data processing devices configured to run the instructions of the computer-aided design program to perform the method of any one of claims 10-17.

19. The system of claim 18, comprising a 2.5 axis subtractive manufacturing machine.

20. A non-transitory computer readable medium encoding a computer-aided design program operable to cause one or more data processing devices to perform the method of any one of claims 10-17.

Citation Information

Cited By

  • Mechanical digital design three-dimensional model building method and system

    CN120805565A

  • Mechanical digital design three-dimensional model building method and system

    CN120805565B