阅读说明：本技术 一种机械设计方案分步推荐方法 (Mechanical design scheme step-by-step recommendation method ) 是由 郑锐 吴泳荣 杨庆保 郜清科 高小城 国磊 于 2021-09-24 设计创作，主要内容包括：本发明公开了一种机械设计方案分步推荐方法,步骤包括：建立机械设计方案推荐模型；在机械设计方案推荐模型中导入待加工部件数模；确定待加工单元的第一个零件,根据待加工部件数模信息在机械设计方案推荐模型中推荐、选择第一个零件的设计方案；根据零件设计顺序确定待加工单元的第N个零件,N≥2,并在机械设计方案推荐模型中导出第N个零件的推荐结果；直至零件设计顺序结束,所有单个零件推荐设计方案的集合形成机械结构的整体推荐设计方案。本发明单步顺序推荐零件结构及其约束,方法简单、易行,可极大地提升机械的设计效率、准确率,节省设计周期,减少对设计人员的经验要求,具有良好的工程推广应用价值。(The invention discloses a method for recommending a mechanical design scheme step by step, which comprises the following steps: establishing a mechanical design scheme recommendation model; importing a part digital model to be processed into a mechanical design scheme recommendation model; determining a first part of a unit to be machined, recommending and selecting a design scheme of the first part in a mechanical design scheme recommendation model according to the digital-analog information of the part to be machined; determining the Nth part of the unit to be processed according to the design sequence of the parts, wherein N is more than or equal to 2, and deriving the recommendation result of the Nth part in a mechanical design scheme recommendation model; and until the design sequence of the parts is finished, the set of all the single part recommended design schemes forms the whole recommended design scheme of the mechanical structure. The method is simple and easy to implement, can greatly improve the design efficiency and accuracy of machinery, saves the design period, reduces the experience requirements on designers, and has good engineering popularization and application values.)

1. A mechanical design scheme step-by-step recommendation method is characterized by comprising the following steps:

1) establishing a mechanical design scheme recommendation model;

2) importing a part digital model to be processed into a mechanical design scheme recommendation model;

3) recommending and determining a first part design scheme according to a numerical model of a part to be machined: deriving a recommendation result of a first part in a mechanical design scheme recommendation model according to a part digital model to be processed, wherein the recommendation result of the first part comprises at least one group of recommended design schemes and part design sequences, and a group of recommended design schemes is selected;

4) recommending and determining the Nth part design scheme according to the part design sequence: n is more than or equal to 2, a recommendation result of the Nth part is derived from the mechanical design scheme recommendation model, the recommendation result of the Nth part comprises at least one set of recommendation design scheme, part design sequence and geometric constraint relation with the previous part of the Nth part, and a set of recommendation design scheme is selected;

5) and until the design sequence of the parts is finished, the set of all the single part recommended design schemes forms the whole recommended design scheme of the mechanical structure.

2. The method of claim 1, wherein the recommended design information includes: the structure, the spatial position and the geometric constraint relation with the part to be processed.

3. The method according to claim 1 or 2, wherein the set of recommended designs selected in steps 3) and 4) includes structural size constraints, order constraints and geometric relationship constraints.

4. The method according to claim 1, wherein the step 1) comprises the following steps:

1.1) creating a part database of a historical mechanical structure, a constraint relation library among parts and between the parts and parts to be processed;

1.2) establishing mapping relations between each part of the mechanical structure and the constraint relation;

1.3) training the algorithm to recommend a mechanical design proposal recommendation model.

5. The method for recommending mechanical design schemes step by step according to claim 4, wherein the recommended design schemes and the selection information are collected in step 3) and step 4) and fed back to the part database of the mechanical structure, the constraint relationship library among the parts and the parts to be machined.

6. The method according to claim 4 or 5, wherein the constraint relation library between the parts and the parts to be machined comprises a sequential constraint relation, a structural constraint relation and a geometric constraint relation.

7. A method of machine design solution step recommendation according to claim 4 or 5, characterized in that said step 1.2): according to historical design data, the mutual mapping relations of the structure constraint of each part in the mechanical structure and the structure size thereof, the mutual position constraint and the mutual posture constraint among the parts, the parts and the parts to be processed are established, and a mapping relation library of the structure shape of the parts, the structure size of the parts, the positions and the postures among the parts, the sizes, the positions and the postures of the parts and the parts to be processed is formed to be used as a sample.

8. The method of claim 4 or 5, wherein the step 1.3) is a deep learning neural network model, comprising three parts: the first part is a part recommendation design sequence, the second part is a part structure recommendation, and the third part is a constraint relation matching.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any one of claims 1-8 when executing the computer program.

10. A computer-readable storage medium, having stored thereon a computer program, characterized in that the computer program, when being executed by a processor, is adapted to carry out the method of any one of claims 1-8.

Technical Field

The invention relates to a computer technology, in particular to a step-by-step recommendation method for a mechanical design scheme.

Background

Computer Aided Design (CAD) technology has become the subject technology and method of the current design industry, and the software of CAD is also numerous, such as CATIA, PRO/E, UG, Auto CAD, etc. The design software realizes the design of products under the interactive design application of designers by depending on the technology of a geometric engine and a rendering engine, and has higher requirements on the experience, the capability and the like of the designers. With the rise of artificial intelligence technology, the artificial intelligence technology is integrated into the field of traditional mechanical design, and the improvement of the current situation of mechanical design technology has become a key focus of industrial technicians.

CN202011167540.0 discloses a mechanical design proposal recommendation method and device, the method includes the following steps: determining design requirement data, and performing feature extraction on the design requirement data to obtain features to be recommended; determining a category to be recommended corresponding to the feature to be recommended based on predetermined classification logic; and determining a recommended machine design scheme of the design requirement data based on the historical machine design scheme corresponding to the category to be recommended.

CN202110022171.4 discloses a mechanical design scheme recommendation method, a device and electronic equipment, relating to the technical field of computers, the method firstly obtains design requirement information, including digital-analog information and process information; then determining the target working scene type corresponding to the design requirement information; and then recommending a corresponding target design scheme.

In both of the above two patent documents, a mechanical structure design scheme is recommended according to mechanical design requirement data, and both of the two patents are recommended as an integral design scheme. However, the whole design recommendation method has a great demand on historical design data, and the accuracy of the recommended scheme is low. In addition, the recommendation is only a structural data model, and information such as constraint relation among parts is lacked, so that the workload of modifying a subsequent design scheme is also extremely large. Therefore, a design method with stronger specialization and higher design capability and design efficiency is urgently needed in the industry, particularly in the field of automobile welding clamp design.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a step-by-step mechanical design scheme recommending method which can effectively improve the design speed and the accuracy and provide a basis for efficient modification of mechanical structure design.

The purpose of the invention is realized by the following technical scheme.

A mechanical design scheme step-by-step recommendation method comprises the following steps:

1) establishing a mechanical design scheme recommendation model;

2) importing a part digital model to be processed into a mechanical design scheme recommendation model;

3) recommending and determining a first part design scheme according to a numerical model of a part to be machined: deriving a recommendation result of a first part in a mechanical design scheme recommendation model according to a part digital model to be processed, wherein the recommendation result of the first part comprises at least one group of recommended design schemes and part design sequences, and a group of recommended design schemes is selected;

4) recommending and determining the Nth part design scheme according to the part design sequence: n is more than or equal to 2, a recommendation result of the Nth part is derived from the mechanical design scheme recommendation model, the recommendation result of the Nth part comprises at least one set of recommendation design scheme, part design sequence and geometric constraint relation with the previous part of the Nth part, and a set of recommendation design scheme is selected;

5) and until the design sequence of the parts is finished, the set of all the single part recommended design schemes forms the whole recommended design scheme of the mechanical structure.

The recommending first part design information comprises: the structure, the spatial position and the geometric constraint relation with the part to be processed.

The selection constraint conditions for selecting a set of recommended design solutions in the steps 3) and 4) comprise structural size constraint, sequence constraint and geometric relationship constraint.

The step 1) specifically comprises the following steps:

1.1) creating a part database of a historical mechanical structure, a constraint relation library among parts and between the parts and parts to be processed;

1.2) establishing mapping relations between each part of the mechanical structure and the constraint relation;

1.3) training the algorithm to recommend a mechanical design proposal recommendation model.

And 3) collecting the recommended design scheme and the selection information in the step 3) and the step 4), and feeding back the recommended design scheme and the selection information to a part structure database in the mechanical structure, a constraint relation library among parts and a constraint relation library between the parts and the parts to be processed.

The constraint relation library among all parts and between the parts and the parts to be processed comprises a sequential constraint relation, a structural constraint relation and a geometric relation constraint relation.

Said step 1.2): according to historical design data, the mutual mapping relations of the structure constraint of each part and the structure size thereof, the mutual position constraint and the mutual posture constraint of the parts, the parts and the parts to be processed are established, and a mapping relation library of the structure shape of the parts, the structure size of the parts, the positions and the postures of the parts, the sizes, the positions and the postures of the parts and the parts to be processed is formed as a sample.

The step 1.3) is a deep learning neural network model, which comprises three parts: the first part is a part recommendation design sequence, the second part is a part structure recommendation, and the third part is a constraint relation matching.

An electronic device comprising a memory, a processor, a computer program stored in the memory and executable on the processor, the processor implementing the method of any one of claims 1-8 when executing the computer program.

A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, performs the method of any one of claims 1-8.

Compared with the prior art, the invention has the advantages that: the invention recommends and selects the design scheme of a single part step by step based on the structure size, the design sequence and the interrelation among the parts in the mechanical structure, combines the whole design scheme of the unit to be processed, and trains the intelligent algorithm model.

Drawings

FIG. 1 shows a model of a part to be machined (a side panel of an automobile);

FIG. 2 shows a first part (dowel) of the locating unit;

FIG. 3 is a second detail (pillar) recommendation of the positioning unit;

FIG. 4 shows the overall positioning unit design;

FIG. 5 shows a first part (cradle block) recommendation for the clamping unit;

FIG. 6 shows a second part of the clamping unit (column and cylinder) recommended;

fig. 7 overall clamping unit design, wherein: 1-single step recommendation of the entire positioning unit of the design; 2-one step recommends the design of the entire clamping unit.

Detailed Description

The invention is described in detail below with reference to the drawings and specific examples.

A mechanical design scheme step-by-step recommendation method comprises the following steps:

1) establishing a mechanical design scheme recommendation model;

2) importing a part digital model to be processed into a mechanical design scheme recommendation model;

3) recommending and determining a first part design scheme according to a numerical model of a part to be machined: deriving a recommendation result of a first part in a mechanical design scheme recommendation model according to a part digital model to be processed, wherein the recommendation result of the first part comprises at least one group of recommended design schemes and part design sequences, and a group of recommended design schemes is selected;

4) recommending and determining the Nth part design scheme according to the part design sequence: n is more than or equal to 2, a recommendation result of the Nth part is derived from the mechanical design scheme recommendation model, the recommendation result of the Nth part comprises at least one set of recommendation design scheme, part design sequence and geometric constraint relation with the previous part of the Nth part, and a set of recommendation design scheme is selected;

5) and until the design sequence of the parts is finished, the set of all the single part recommended design schemes forms the whole recommended design scheme of the mechanical structure.

The recommending first part design information comprises: the structure, the spatial position and the geometric constraint relation with the part to be processed.

The selection constraint conditions for selecting a set of recommended design solutions in the steps 3) and 4) comprise structural size constraint, sequence constraint and geometric relationship constraint.

The step 1) specifically comprises the following steps:

1.1) creating a part database of a historical mechanical structure, a constraint relation library among parts and between the parts and parts to be processed;

1.2) establishing mapping relations between each part of the mechanical structure and the constraint relation;

1.3) training the algorithm to recommend a mechanical design proposal recommendation model.

And 3) collecting the recommended design scheme and the selection information in the step 3) and the step 4), and feeding back the recommended design scheme and the selection information to a part structure database in the mechanical structure, a constraint relation library among parts and a constraint relation library between the parts and the parts to be processed.

The constraint relation library among all parts and between the parts and the parts to be processed comprises a sequential constraint relation, a structural constraint relation and a geometric relation constraint relation.

Said step 1.2): according to historical design data, the mutual mapping relations of the structure constraint of each part and the structure size thereof, the mutual position constraint and the mutual posture constraint of the parts, the parts and the parts to be processed are established, and a mapping relation library of the structure shape of the parts, the structure size of the parts, the positions and the postures of the parts, the sizes, the positions and the postures of the parts and the parts to be processed is formed as a sample.

The step 1.3) is a deep learning neural network model, which comprises three parts: the first part is a part recommendation design sequence, the second part is a part structure recommendation, and the third part is a constraint relation matching.

An electronic device comprising a memory, a processor, a computer program stored in the memory and executable on the processor, the processor implementing the method of any one of claims 1-8 when executing the computer program.

A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, performs the method of any one of claims 1-8.

Example 1: design recommendation of positioning unit of automobile side wall

Firstly, designing single step recommendation of a positioning unit of the automobile side wall.

1. And importing information into the trained algorithm model. Importing digital-to-analog (shown in figure 1) information of the automobile side wall part to be processed.

2. The design of the first part (positioning pin) of the positioning unit is recommended. At least one design scheme is recommended for the positioning pin, and preferably 1-3, and 3 design schemes are recommended in the embodiment. The recommended positioning pin design scheme automatically designs a positioning pin structure according to the information of the positioning hole of the side wall part to be processed, and corrects the selected positioning pin spatial position according to the geometric constraint relationship between the positioning pin and the positioning hole (the axial lead of the positioning pin is superposed with the central line of the positioning hole) (as shown in fig. 2).

3. The design of the second part (support) of the positioning unit is recommended. The recommendation result comprises a design sequence of the part, a design scheme, a geometric constraint relation between the recommended part and the previous part and the like. In this embodiment, the second part of the positioning unit is preferably a support, and the design scheme is 3 (as shown in fig. 3), the support is placed with a distance of 10-40mm from the part to be processed, and the L-shaped bottom of the lower part of the support is placed towards the side of the part to be processed.

4. The remaining parts of the positioning unit (the connection block in this embodiment) are recommended individually. The selected design information includes structural size constraints, order constraints, and geometric relationship constraints. After the whole unit part is designed, the design scheme is shown in fig. 4.

And secondly, training an algorithm recommendation model of the positioning unit.

1. And creating a constraint relation library among the historical positioning unit design parts, all parts and the parts to be processed. The positioning unit roughly comprises a positioning pin, a vertical column and a connecting block. And establishing a constraint relation library among all parts in the positioning unit structure according to all part libraries of the positioning unit, wherein the constraint relation library comprises three constraint relations of sequence, structure and geometric relation. Wherein: after the design of the positioning pins is finished, recommending the upright columns and the auxiliary structures, and then recommending connecting blocks and the like; structural constraint is the structural design size of each part; the geometrical relationship constraint is the mutual position and posture constraint among all parts and between the parts and the parts to be processed. The data sample size of each part and constraint relation library is about 1200, and the data samples are classified and stored.

2. And establishing mapping relations between the parts of the positioning unit and the constraint relations. According to historical design data, the structure of each part of the positioning unit is restricted with the structure size of the positioning unit, and the mutual mapping relations of the position and the posture restriction among the parts, the parts and the parts to be processed are established, so that a mapping relation library of the structure shape of the parts, the structure size of the parts, the position and the posture among the parts, the size, the position and the posture of the parts and the parts to be processed is formed as a sample.

3. And recommending algorithm model training. The algorithm recommendation model is a deep learning neural network model and is divided into three parts: the first part is a part recommendation design sequence, the second part is a part structure recommendation, and the third part is a constraint relation matching. And (3) constraining the structure of each part of the positioning unit and the structure size of the positioning unit, and inputting the mutual mapping relation of the mutual position and posture constraint among the parts, the parts and the parts to be processed into a recommendation algorithm model, and learning the mapping relation among the structure, the structure size, the position and the posture of the parts, the size, the position and the posture of the parts and the parts to be processed through the model.

4. And outputting the recommended scheme. After the recommendation algorithm model is trained, according to the imported information of the part to be processed and the reference positioning point, the recommended scheme comprises a part structure and a constraint relation between the parts. And the information of the recommended and selected schemes and the like is automatically stored in the part library and the constraint relation library.

Example 2: design recommendation of clamping unit of automobile side wall

Firstly, the design of the clamping unit of the side wall of the automobile is recommended in a single step.

1. And importing information. And importing a digital model (the digital model is designed by adopting a clamping unit, namely shown in figure 4) of the automobile side wall part to be processed and datum positioning point information.

2. The design of the first part of the clamping unit (prolonged viewing) is recommended. At least one design scheme of the supporting and pressing block is recommended, and 1-3 are preferred, and 3 design schemes of the supporting and pressing block are recommended in the embodiment. The recommended pressure supporting block design scheme automatically designs a pressure supporting block structure according to the structural information of the clamping position of the side wall part to be processed, and corrects the selected space position of the pressure supporting block according to the geometric constraint relation between the pressure supporting block and the part clamping position (the direction of a main shaft of the pressure supporting block is vertical to the reference plane of the part to be processed) (as shown in figure 5).

3. The design of the second part of the clamping unit (column and cylinder) is recommended. The second part of the clamping unit is preferably a support and a cylinder, the design scheme is 3 (figure 6), the support is arranged at a distance of 10-40mm from the part to be processed, and the L-shaped bottom of the lower part of the support is arranged towards the side of the part to be processed.

4. The remaining parts of the clamping unit (in this embodiment, the connecting block) are recommended individually. The selected design information includes structural size constraints, order constraints, and geometric relationship constraints. After the whole unit part is designed, the design scheme is shown in fig. 7.

And secondly, training an algorithm recommendation model of the clamping unit.

1. And creating a constraint relation library among the historical clamping unit design parts, all parts and the parts to be processed. The clamping unit mainly comprises a supporting pressing block, an upright column (including a cylinder), a connecting block and the like. The parts of the clamping unit and the constraint relation library are established close to the positioning unit, the data sample size is about 1500 respectively, and the data samples are classified and stored.

2. And establishing a mapping relation between each part of the clamping unit and the constraint relation. And taking a mapping relation library among the structure of the part, the size of the structure of the part, the position and the posture among the parts, the size, the position and the posture of the part and the part to be processed as a sample according to the historical design data of the clamping unit.

3. And recommending algorithm model training. The algorithm recommendation model of the clamping unit is similar to that of the positioning unit, and only the data samples learned by the algorithm model are different. And (3) constraining the structure of each part of the clamping unit and the structure size of the clamping unit, and inputting the mutual mapping relation between the parts, the mutual positions between the parts and the parts to be processed and the posture constraint into a recommendation algorithm model, and learning the mapping relation between the structure of the parts, the structure size of the parts, the positions and postures between the parts and the sizes, the positions and the postures between the parts and the parts to be processed by the model.

4. And outputting and selecting a recommended scheme. After the recommendation algorithm model is trained, according to the imported information of the part to be processed and the reference positioning point, the recommended scheme comprises a part structure and a constraint relation between the parts. And the information of the recommended and selected schemes and the like is automatically stored in the part library and the constraint relation library.