Metal additive manufacturing self-supporting structure intelligent generation system and self-supporting structure generation method

Abstract

The present application belongs to the technical field of metal additive manufacturing intelligent design and computer aided engineering, in particular to a metal additive manufacturing self-supporting structure intelligent generation system and a self-supporting structure generation method, the system comprising a model import and preprocessing module, a to-be-supported region identification module, a dual-modal knowledge base engine, a support scheme generation and parameterization combination module, a constraint consistency checking and optimization module, and a user interaction and result output module; by constructing a knowledge system for self-supporting design, and fusing rule reasoning, case reasoning and constraint checking into a unified process, to-be-supported region identification, support scheme rapid matching, parameterization generation and result optimization are realized, thereby improving support design efficiency, reducing material consumption, and improving post-processing friendliness, realizing rapid design of support structure without changing the original function geometry and overall shape of the part.

Description

Technical Field

[0001] This invention belongs to the field of intelligent design and computer-aided engineering technology in metal additive manufacturing, and in particular, it relates to an intelligent generation system and method for generating self-supporting structures in metal additive manufacturing. Background Technology

[0002] In metal additive manufacturing, especially in SLM (Selective Laser Melting) forming, auxiliary supports are typically required in overhanging areas, bridging areas, thin-walled areas, and areas with poor local heat dissipation to ensure forming stability, suppress thermal deformation, and mitigate the risk of collapse. While existing industrial software has the capability to automatically generate supports, its implementation largely relies on geometric recognition, preset templates, and empirical parameters. Designers often need to repeatedly adjust parameters such as support spacing, contact width, top offset, and support type based on model features. For parts with complex structures and significant local feature differences, this approach still relies heavily on manual intervention.

[0003] Existing technologies mainly fall into two categories: one is a design method that reduces support requirements by modifying the local geometry of parts or introducing topology optimization constraints; the other is a method that, while keeping the original geometry of parts unchanged, uses software-built-in templates (block-like, line-like, tree-like, or grid-like supports) combined with human experience for local correction. Among existing publicly available solutions, technologies can achieve overhang surface recognition, build file generation, and additive manufacturing parameter configuration.

[0004] In practical engineering applications, a common implementation scheme is as follows: First, import the 3D model into software environments such as NX or Magics, and identify potential areas to be supported based on the overhang angle threshold; then, designers, combining the part's posture, local span, wall thickness, and experience, select the form of support, such as block support or line support; finally, adjust parameters such as support density, wall thickness, contact points, and offset through trial and error, and output a support model for slicing or printing. This type of scheme can complete the basic support construction, but its rules, experience, and historical success cases are mostly scattered, and a unified knowledge organization and dynamic retrieval mechanism has not yet been formed.

[0005] For complex support components, components with limited internal cavities, or components with high functional surface requirements, existing solutions often rely solely on a single overhang angle for judgment, without fully considering factors such as bridging span, local minimum wall thickness, thermal sensitivity, post-processing accessibility, and functional surface protection requirements. As a result, although the generated results can meet basic support requirements, there is still significant room for optimization in terms of material consumption, ease of removal, and surface quality.

[0006] In summary, existing technologies primarily rely on manual experience and single geometric threshold judgments for support design, lacking a systematic retrieval and reuse of successful historical cases. Rule knowledge and instance knowledge are stored in a scattered manner, and an integrated rapid support design mechanism encompassing "historical case reuse—rule correction—constraint verification—parametric generation" is missing. Therefore, a self-supporting structure intelligent generation method is needed that can integrate case-based reasoning and rule-based reasoning, achieving the aforementioned integrated processing. Summary of the Invention

[0007] The purpose of this invention is to address the problems of high reliance on manual labor, difficulty in reusing experiential knowledge, low efficiency in parameter configuration, and high redundancy in the support design process in the prior art. Therefore, this invention proposes an intelligent generation system and method for generating self-supporting structures in metal additive manufacturing.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The present invention discloses an intelligent generation system for self-supporting structures in metal additive manufacturing, comprising: The model import and preprocessing module is used to read the 3D model of the part, obtain the construction direction, and complete the geometric integrity check. The module for identifying areas to be supported is used to identify risk areas to be supported based on information such as the normal of the model triangular facets, the local construction direction, the bridging span, and the wall thickness. A dual-modal knowledge base engine, with a rule base and a case base as its core, and linked with a parameter base and a constraint base to complete knowledge retrieval; The support scheme generation and parametric combination module is used to combine basic support units such as base units, connecting trusses and contact interfaces according to rules to generate an editable support structure. The constraint consistency verification and optimization module is used to perform feasibility verification on candidate solutions at the geometric, process and quality levels, and to provide feedback for adjustments when the verification fails. The user interaction and result output module is used to display candidate solutions, local parameters, material consumption information and stress / risk warnings, and output supporting files independent of the original model; Each module is embedded in or runs within the Siemens NX (Siemens NX 3D design and manufacturing platform) environment, and information is transmitted through a unified data interface.

[0009] As a further preferred option, the rule base adopts a production rule representation and is stored in the manner of "IF geometry / process conditions THEN support type and key parameters"; The case library stores historically successful design examples and enables retrieval and reuse based on the similarity of local features of parts; The parameter library is used to maintain the geometric parameters and applicable range of different support units; The constraint library is used to limit the boundary conditions of the support scheme in terms of geometry, process formability, and post-processing accessibility.

[0010] The present invention discloses a self-supporting structure generation method for a metal additive manufacturing self-supporting structure intelligent generation system, comprising the following steps: S1: Import the part model and obtain the construction direction, perform geometric checks and feature preprocessing on the model; S2: Input the feature parameters of the region to be supported into the dual-modal knowledge base engine, and call the rule base and case base respectively; S3: Based on the results of rule-based reasoning and case retrieval, determine the supporting organizational structure, contact methods, and main parameters; S4: Use parameterized support templates to instantiate and combine support units; S5: Verify the geometric consistency, process adaptability, and quality constraints of the candidate solutions; S6: Outputs the final support file and displays the support structure, local parameters, candidate schemes and related results in the interactive interface, supporting designers to further fine-tune and confirm the results.

[0011] As a further preferred option, in S1, the local overhang angle is extracted based on the relationship between the normal of the triangular facet of the model surface and the construction direction, and the risk area to be supported is determined by combining the bridging span, local wall thickness, thermal sensitivity and accessibility to be removed.

[0012] As a further preferred option, in S2, the rule base triggers support design rules based on local geometry and process conditions; the case base retrieves and reuses historically successful solutions based on similar characteristics.

[0013] As a further preferred option, in S3, block supports are preferred for large-span, low-angle overhang areas with high support requirements; For thin-walled areas or areas near the smallest local cross-section, lightweight supports with smaller contact areas are preferred. Reduce the support spacing in areas where heat tends to accumulate. For areas adjacent to critical functional surfaces, reduce the contact width or increase the top offset; For areas with narrow cavities and poor post-processing accessibility, easily removable support forms should be preferred.

[0014] As a further preferred option, in S4, the system combines the base unit, connecting truss and contact interface according to the regional requirements, and generates independent support entities while keeping the original part model unchanged.

[0015] As a further preferred option, in S5, if the solution has problems such as spatial interference, process incompatibility, excessive contact risk, or unattainable post-processing, it will be fed back to the parameter adjustment and rule correction stage to form a closed loop of "generation-verification-correction".

[0016] Compared with the prior art, the present invention has the following technical effects: (1) Without changing the original functional geometry and overall shape of the parts, the integrated processing of the area to be supported, knowledge retrieval, support matching, parameter generation and constraint verification is realized, reducing the reliance on personal experience and repeated trial and error in traditional support design.

[0017] (2) By coordinating rule knowledge, case knowledge, parameter knowledge and constraint knowledge, the structured storage and reusable retrieval of supporting design knowledge can be realized, thereby improving the consistency and pertinence of supporting design for complex parts.

[0018] (3) In the comparative experiment of typical support parts, the knowledge reuse-based solution showed a smaller support volume and less material consumption compared with the traditional manual support solution. The internal opening / lightweight features of the support were more obvious, and the number of contact points in the post-processing was expected to be less.

[0019] (4) In terms of post-processing, the optimized support scheme forms a more continuous powder discharge channel, which can reduce the powder trapping phenomenon. Although the removal time of the two schemes is similar during the support removal process, the scheme of the present invention can significantly reduce the labor intensity of dismantling and improve the operation safety. The surface morphology of the support contact area is more uniform, and the degree of protrusion residue, spherical adhesion and mechanical scratches is less, while the surface quality of the non-support area is basically unaffected.

[0020] (5) Since the current experiment uses the same substrate for synchronous printing, the printing time of a single piece is not quantified independently. However, from the perspective of supporting lightweighting and material utilization, the solution of the present invention is beneficial to improving the overall forming efficiency. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall framework structure of the system of the present invention; Figure 2 This is a schematic diagram illustrating the intelligent generation process supporting the present invention; Figure 3 A schematic diagram of a bimodal knowledge base engine and its collaborative operation; Figure 4 The diagram shows the system operation interface of the present invention, wherein (a) is the part addition interface, (b) is the overhang area identification interface, (c) is the instance retrieval result interface, and (d) is the support parameter configuration interface. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0023] This invention discloses an intelligent generation system for self-supporting structures in metal additive manufacturing, specifically a system for generating self-supporting structures in selective laser melting (SLM) scenarios. This system relies on the Siemens NX platform and its secondary development interface to automatically identify, call upon, generate solutions, optimize parameters, and output results for the areas to be supported in the part model. Without altering the original geometry of the part (original functional geometry and overall shape), the system constructs a knowledge system oriented towards self-supporting design and integrates rule-based reasoning, case-based reasoning, and constraint verification into a unified process. This enables the identification of areas to be supported, rapid matching of support solutions, parameterized generation, and result optimization, thereby improving support design efficiency, reducing material consumption, and enhancing post-processing friendliness, thus achieving rapid design of support structures.

[0024] The system of this invention generally includes a model import and preprocessing module, a region-to-be-supported identification module, a bimodal knowledge base engine, a support scheme generation and parameterized combination module, a constraint consistency verification and optimization module, and a user interaction and result output module. For example... Figure 1 The diagram shown is a schematic of the overall framework of the system of the present invention, illustrating the relationship between the model import and preprocessing module, the region to be supported identification module, the dual-modal knowledge base engine, the support generation and parameterization combination module, the constraint consistency verification and optimization module, and the user interaction and result output module.

[0025] All of the above modules are embedded in or run within the Siemens NX platform environment, and complete information transmission through a unified data interface, including: The model import and preprocessing module is used to read the 3D model of the part, obtain the construction direction, and complete the geometric integrity check. The module for identifying areas to be supported is used to identify risk areas to be supported based on information such as the normal of the model triangular facets, the local construction direction, the bridging span, and the wall thickness. A dual-modal knowledge base engine, with a rule base and a case base as its core, and linked with a parameter base and a constraint base to complete knowledge retrieval; The support scheme generation and parametric combination module is used to combine basic support units such as base units, connecting trusses and contact interfaces according to rules to generate an editable support structure. The constraint consistency verification and optimization module is used to perform feasibility verification on candidate solutions at the geometric, process and quality levels, and to provide feedback for adjustments when the verification fails. The user interaction and result output module is used to display candidate solutions, local parameters, material consumption information and stress / risk warnings, and output supporting files independent of the original model.

[0026] like Figure 2 The diagram shown illustrates the intelligent generation process supported by this invention, illustrating the flow of "model input—area identification to be supported—rule / case invocation—parameterized generation—constraint verification—result output". The specific intelligent generation method of this invention includes the following steps: S1: Import the part model and obtain the construction direction. Perform geometric checks and feature preprocessing on the model. Based on the relationship between the normal of the triangular facets on the model surface and the construction direction, extract the local overhang angles, and determine the risk areas to be supported by combining the bridging span, local wall thickness, thermal sensitivity, and accessibility removal.

[0027] S2: Input the feature parameters of the area to be supported into the dual-modal knowledge base engine, and call the rule base and case base respectively. The rule base triggers support design rules based on local geometry and process conditions; the case base retrieves and reuses historical successful solutions based on similar features.

[0028] S3: Based on the rule-based reasoning results and case retrieval results, determine the organization method, contact form, and main parameters of the supports. For large-span, low-angle overhang areas with high support requirements, block supports are preferred; for thin-walled areas or areas near the smallest local cross-section, lightweight supports with smaller contact areas are preferred; for areas where heat easily accumulates, reduce the support spacing; for areas near critical functional surfaces, reduce the contact width or increase the top offset; for areas with narrow cavities and poor post-processing accessibility, easily removable support forms are preferred.

[0029] S4: Instantiate and assemble support units using parametric support templates. The system combines base units, connecting trusses, and contact interfaces according to regional requirements, and generates independent support entities while keeping the original part models unchanged.

[0030] S5: Verify the geometric consistency, process adaptability, and quality constraints of the candidate solutions. If a solution has issues such as spatial interference, process incompatibility, excessive contact risk, or unachievable post-processing, it will be fed back to the parameter adjustment and rule correction stage, forming a closed loop of "generation-verification-correction".

[0031] S6: Outputs the final support file and displays the support structure, local parameters, candidate schemes and related results in the interactive interface, supporting designers to further fine-tune and confirm the results.

[0032] Figure 3 This diagram illustrates the dual-modal knowledge base engine and its collaborative operation, showing the calling relationships between the rule base, case base, parameter base, and constraint base.

[0033] Figure 4 This is a schematic diagram of the system operation interface of the present invention, which shows the interfaces for importing parts, identifying overhanging areas, retrieving instances, configuring support parameters, and outputting results.

[0034] To achieve intelligent transformation from "feature input" to "parameterized support scheme" in the functional layer, this invention introduces a multi-dimensional algorithm collaboration mechanism in the system background operation. The specific transformation process and algorithm logic are as follows: (1) Algorithm implementation in the knowledge retrieval and fine matching stage: After the risk area to be supported is locked, the system extracts the process and geometric features of the current task and constructs a multi-level feature vector (covering overhang angle, span, wall thickness and material type). In order to accurately measure the reusability of historical experience, the algorithm adopts a subjective and objective comprehensive weighting method to determine the weight of each feature. The formula is as follows:

[0035] In the formula, The subjective weights are determined based on the process mechanism. The objective weights are determined based on the discriminative power of the candidate set. This is the weighting balance coefficient.

[0036] Next, the algorithm calculates the local similarity of each attribute using normalized distance and enumerated matching, and then calculates the comprehensive similarity using a hierarchical weighted method. :

[0037] In the formula, and The similarity is between the part layer and the feature layer, respectively. and This represents the corresponding hierarchical weight.

[0038] The backend uses a KD-tree to perform fast nearest neighbor retrieval, selecting high-quality cases with a similarity greater than 85% for direct reuse or minor modification.

[0039] (2) Algorithm implementation for rule correction and conflict resolution stages: To address potential global spatial conflicts or process mismatches that may arise when combining local support strategies across multiple regions, the constraint consistency verification and optimization module is equipped with a conflict resolution algorithm. Arbitration and parameter correction are strictly performed according to the priority logic chain of "constraint priority -- process adaptation priority -- local manufacturability priority -- instance reuse secondary". Simultaneously, to prevent extreme deformation of historical experience parameters in new tasks, the algorithm introduces a parameter range boundary projection function for any initial support parameter. (e.g., support spacing, thickness, etc.), its normalized projection mapping relationship is as follows:

[0040] Through this projection algorithm, the system dynamically forces the initial solution to be limited to the current process safety range, thereby guiding the support scheme generation and parameterized combination module to complete the final entity construction.

[0041] The technical solution of the present invention will be further described below with reference to a typical bracket-type component. This embodiment is only used to illustrate the principle and implementation process of the present invention and does not constitute a limitation on the scope of protection of the present invention.

[0042] Taking a support-type component with local protrusions, stepped transitions, varying overhang angles, multiple wall thicknesses, and local minimum cross-sections as an example, a 3D model was imported into the Siemens NX platform, and a unified construction direction was set. Considering the AlSi10Mg material and a layer thickness of 0.03 mm, the geometric integrity of the component and the identification of areas requiring support were performed. Based on factors such as overhang angle, bridging span, local wall thickness, and thermal sensitivity, the system marked areas prone to collapse, warping, or difficult post-processing.

[0043] The system inputs the aforementioned regional features into the knowledge base engine. For large, low-angle overhanging areas, the rule "preferentially use block supports when the overhang angle is less than 45° and the local span is greater than 15 mm" is triggered; for thin-walled areas and areas adjacent to the smallest local cross-section, the rule "use lightweight supports with smaller contact areas when the local wall thickness is less than 3 mm" is triggered; for continuous overhanging areas with high heat sensitivity, the rule "reduce support spacing" is triggered; for areas adjacent to critical functional surfaces, the rule "reduce contact width or increase top offset" is triggered; for areas with narrow cavities and poor post-processing accessibility, the rule "preferably use easily removable support forms" is triggered. Simultaneously, the system retrieves historical solutions similar to the current local features from the case library and reuses the initial solution based on similar support forms and parameter configurations.

[0044] After the initial scheme is generated, the system combines base units, connecting trusses, and contact interfaces using parameterized templates to form independent support entities and outputs candidate support schemes. Next, the constraint verification module verifies the candidate schemes from three aspects: geometric space, process adaptability, and post-processing quality. If local supports interfere with adjacent structures, or if parameter combinations do not meet current material and layer thickness conditions, or if the support form may lead to inaccessibility during dismantling, damage to functional surfaces, or redundant material consumption, the system automatically feeds back to the rule correction or parameter adjustment stage until a final support scheme that meets the constraints is obtained.

[0045] To verify the effectiveness of this invention, under the same equipment, materials, process parameters, and placement conditions, the traditional manual support scheme and the scheme of this invention were compared and printed on the same substrate. The results show that the traditional manual design of the support scheme for this bracket part takes an average of about 20 minutes, while the intelligent matching generation using this bimodal knowledge base only takes 5 minutes, improving design efficiency. The support structure generated by the scheme of this invention has a more open internal structure, with more obvious openings and lightweight features. While ensuring the forming requirements of the overhanging area, the support volume and material consumption are reduced. The support volume of the traditional scheme is 68.5 cm³, while the volume of the self-supporting structure generated by this system is only 45.2 cm³, saving approximately 34% of the support material. Powder cleaning is more convenient, and the labor intensity of support removal is lower, reducing the actual post-processing support removal time from 20 minutes to 8 minutes. The surface residue and mechanical damage of the support contact area are reduced, demonstrating good engineering application value.

[0046] This invention can also be extended to thin-walled shell parts, locally heat-sensitive parts, and metal additive manufacturing parts with complex internal cavities. For different material systems and equipment conditions, only the rule thresholds, parameter library value ranges, and case selection conditions need to be adapted and adjusted to complete the intelligent generation of support solutions within a unified framework.

[0047] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A metal additive manufacturing self-supporting structure intelligent generation system, characterized in that, include: The model import and preprocessing module is used to read the 3D model of the part, obtain the construction direction, and complete the geometric integrity check. The module for identifying areas to be supported is used to identify risk areas to be supported based on information such as the normal of the model triangular facets, the local construction direction, the bridging span, and the wall thickness. A dual-modal knowledge base engine, with a rule base and a case base as its core, and linked with a parameter base and a constraint base to complete knowledge retrieval; The support scheme generation and parametric combination module is used to combine basic support units such as base units, connecting trusses and contact interfaces according to rules to generate an editable support structure. The constraint consistency verification and optimization module is used to perform feasibility verification on candidate solutions at the geometric, process and quality levels, and to provide feedback for adjustments when the verification fails. The user interaction and result output module is used to display candidate solutions, local parameters, material consumption information and stress / risk warnings, and output supporting files independent of the original model; Each module is either embedded or runs within the Siemens NX platform environment and transmits information through a unified data interface.

2. The intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 1, characterized in that: The rule base uses production rule representation and is stored in the format of "IF geometric / process conditions THEN support type and key parameters"; The case library stores historically successful design examples and enables retrieval and reuse based on the similarity of local features of parts; The parameter library is used to maintain the geometric parameters and applicable range of different support units; The constraint library is used to limit the boundary conditions of the support scheme in terms of geometry, process formability, and post-processing accessibility.

3. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 1, characterized in that, Includes the following steps: S1: Import the part model and obtain the construction direction, perform geometric checks and feature preprocessing on the model; S2: Input the feature parameters of the region to be supported into the dual-modal knowledge base engine, and call the rule base and case base respectively; S3: Based on the results of rule-based reasoning and case retrieval, determine the supporting organizational structure, contact methods, and main parameters; S4: Use parameterized support templates to instantiate and combine support units; S5: Verify the geometric consistency, process adaptability, and quality constraints of the candidate solutions; S6: Outputs the final support file and displays the support structure, local parameters, candidate schemes and related results in the interactive interface, supporting designers to further fine-tune and confirm the results.

4. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 3, characterized in that: In S1, based on the relationship between the normal of the triangular facets on the model surface and the construction direction, the local overhang angle is extracted, and the risk area to be supported is determined by combining the bridging span, local wall thickness, thermal sensitivity and accessibility to be removed.

5. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 3, characterized in that: In S2, the rule base triggers support design rules based on local geometry and process conditions; the case base retrieves and reuses historical successful solutions based on similar features.

6. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 3, characterized in that: In S3, block supports are preferred for large-span, low-angle overhang areas with high support requirements; For thin-walled areas or areas near the smallest local cross-section, lightweight supports with smaller contact areas are preferred. Reduce the support spacing in areas where heat tends to accumulate. For areas adjacent to critical functional surfaces, reduce the contact width or increase the top offset; For areas with narrow cavities and poor post-processing accessibility, easily removable support forms should be preferred.

7. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 3, characterized in that: In S4, the system combines base units, connecting trusses, and contact interfaces according to regional requirements, and generates independent support entities while keeping the original part models unchanged.

8. The self-supporting structure generation method of the intelligent generation system for self-supporting structures in metal additive manufacturing according to claim 3, characterized in that: In S5, if the solution has problems such as spatial interference, process incompatibility, excessive contact risk, or unattainable post-processing, it will be fed back to the parameter adjustment and rule correction stage, forming a closed loop of "generation-verification-correction".

Images