Case retrieval method based on product parameter configuration information
Through the case search method based on product parameter configuration information, Boolean logic and hierarchical analysis methods are used to evaluate similarity, the problem of low retrieval efficiency of three-dimensional design data in manufacturing is solved, and the reference design is quickly obtained, which improves design efficiency and employees. efficiency.
Patent Information
- Application Number
- CN202410921521.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-10
- Publication Date
- 2024-11-08
AI Technical Summary
In the manufacturing industry, it is difficult for the existing technology to efficiently retrieve three-dimensional design data, resulting in inefficient design efficiency, excessive redundant drawings, and difficult to query order information, which affects the order completion time.
A case search method based on product parameter configuration information is adopted, and a standardized order information is constructed, subjective similarity evaluation is performed using Boolean logic and hierarchical analysis method, objective similarity evaluation is performed in combination with Manhattan distance, and similar cases are output to achieve rapid acquisition of reference designs.
It improves design efficiency, shortens product R&D cycle, improves employee work efficiency, effectively utilizes PDM function, and reduces manual labor.
Smart Images

Figure CN118917910A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of three-dimensional model similarity evaluation and retrieval, and in particular to a case retrieval method based on product parameter configuration information. Background Art
[0002] In the manufacturing industry, 3D design software has become very popular, such as UG, Solidworks, etc. Now with the trend of customization, enterprises are designing more and more product categories, and data redundancy is serious. At present, most design data is stored in PDM, but there is no efficient method to retrieve it. At present, the needs of customers in some mechanical design fields such as automobile manufacturing and shipbuilding are diverse. If historical product information cannot be reasonably used and designers rely on their experience to design from scratch, the design efficiency will be greatly reduced, affecting the order completion time, indirectly aggravating the situation of designing and producing at the same time, and a large number of redundant drawings and model designs will appear. In addition, the company has not fully formulated the order information standards. Each order is represented by a document. It is difficult to query the order information content and cannot be retrieved according to the rules. It is impossible to accurately retrieve it only by the order number. These are the problems that need to be solved by my country's medium and small manufacturing enterprises at this stage. Summary of the invention
[0003] The purpose of the present invention is to provide a case retrieval method based on product parameter configuration information to solve the problem.
[0004] Technical solution: The case retrieval method based on product parameter configuration information described in the present invention comprises the following steps:
[0005] (1) Construct standardized order information including product structure configuration, basic product parameters, and order information notes;
[0006] (2) Search using Boolean logic;
[0007] (3) Subjective similarity evaluation of order information;
[0008] (4) Objective similarity evaluation of order information;
[0009] (5) Output of similar cases.
[0010] Furthermore, step (1) is as follows: first, extract the product type, product model, and equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the existing models of the enterprise, import their corresponding information into the data table and share it with the MES system database.
[0011] Furthermore, step (2) is specifically as follows: an index is constructed between each product order case data table in the SQL server database and the historical case model, the type and structure configuration information in the product is used as a constraint condition, and Boolean logic is used to retrieve and filter out cases that do not meet the conditions.
[0012] Furthermore, step (3) is specifically as follows: one-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state register bits to encode N states, and only one bit is valid, to establish a coded data set; use the hierarchical analysis method to obtain the weight of the keyword according to the expert's subjective scoring method, and use the text information corresponding to the code to perform one-hot encoding in the order of the keyword table of the coding data set to digitize the data; calculate the cosine similarity between the codes to obtain the product note text similarity.
[0013] Furthermore, step (4) is as follows: the parameter information obtained from the order information is used and normalized, the similarity between the parameters is obtained using Manhattan distance, a comprehensive similarity evaluation between the new order information and the historical order information is obtained, and finally, a number of cases and models that can be used for reference are obtained by sorting the similarities.
[0014] The case retrieval system based on product parameter configuration information of the present invention comprises:
[0015] Order information module: used to build standardized order information including product structure configuration, basic product parameters, and order information notes;
[0016] Retrieval module: used to search using Boolean logic;
[0017] Subjective module: used for subjective similarity evaluation of order information;
[0018] Objective module: used for objective similarity evaluation of order information;
[0019] Output module: used for similar case output.
[0020] Furthermore, in the order information module, the details are as follows: First, extract the product type, product model, and equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the company's existing models, import their corresponding information into the data table and share it with the MES system database.
[0021] Furthermore, in the objective module, the details are as follows: an index is constructed between each product order case data table in the SQLserver database and the historical case model, the type and structural configuration information in the product is used as a constraint, and Boolean logic is used to retrieve and filter cases that do not meet the conditions.
[0022] Furthermore, in the subjective module, the details are as follows: One-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state registers to encode N states, and only one bit is valid, to establish a coded data set; use the hierarchical analysis method to obtain the weight of the keyword according to the expert's subjective scoring method, and use the text information corresponding to the encoding to perform One-hot encoding in the order of the keyword table of the encoding data set to digitize the data; use the cosine similarity calculation method to calculate the codes between each other to obtain the product note text similarity.
[0023] Furthermore, in the output module, the details are as follows: the parameter information obtained from the order information is used and normalized, the similarity between the parameters is obtained using the Manhattan distance, a comprehensive similarity evaluation between the new order information and the historical order information is obtained, and finally, similarity sorting is used to obtain several cases and models that can be used for reference.
[0024] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: under the premise that product configuration information is not sufficient to accurately retrieve product cases as reference designs, by utilizing SolidWorks secondary development, hierarchical analysis method, case-based reasoning (CBR), etc., and improving the information of case similarity evaluation, the retrieval mechanism allows designers to quickly obtain reference cases, saving a lot of manual labor, making more efficient use of PDM functions, shortening the product R&D cycle, and improving employee work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION
[0026] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.
[0027] like Figure 1 As shown, an embodiment of the present invention provides a case retrieval method based on product parameter configuration information, comprising the following steps:
[0028] (1) Construct standardized order information including product structure configuration, basic product parameters, and order information notes; the details are as follows: First, extract the product type, product model, and equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the company's existing models, import their corresponding information into the data table and share it with the MES system database.
[0029] (2) Use Boolean logic for retrieval; specifically, build an index between each product order case data table in the SQL server database and the historical case model, use the type and structure configuration information in the product as constraints, and use Boolean logic retrieval to filter out cases that do not meet the conditions.
[0030] (3) Subjective similarity evaluation of order information; specifically as follows: One-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state registers to encode N states, and only one bit is valid, to establish a coded data set; use the hierarchical analysis method to obtain the weight of the keyword according to the expert's subjective scoring method, and use the text information corresponding to the encoding to perform One-hot encoding in the order of the keyword table of the encoding data set to digitize the data; calculate the cosine similarity between the encodings to obtain the product note text similarity. The specific process is as follows:
[0031] The impact of each keyword on the case is different. A scale is developed as the importance of the keyword to the product and experts score it. The scale is as follows:
[0032] Scale Configuration I is better than Configuration J 1 Equally important 3 Slightly important 5 Clearly important 7 Strongly important 9 Extremely important reciprocal I is more important than J on a scale of 9, and J is more important than I on a scale of 1 / 9.
[0033] Before weight allocation, a judgment matrix needs to be performed. The judgment matrix is scored by experts as shown in the following table:
[0034]
[0035]
[0036] The arithmetic mean method is used to calculate the weight of the judgment matrix. The formula is as follows:
[0037]
[0038] Where i=j=n, where n represents the number of configurations.
[0039] In order to avoid the error that keyword 2 is more important than keyword 1, keyword 1 is more important than keyword 3, and keyword 3 is more important than keyword 2 in the scoring results, a consistency check is required before weight calculation. The consistency check index CI is calculated as follows:
[0040]
[0041] Where n represents the matrix order and a represents the matrix eigenvalue. The closer the CI is to 0, the more reasonable the scoring result is, and less than 0.1 means that the weights are consistent.
[0042] Keywords and examples in the order information notes for cabin products:
[0043] Keyword Examples Weight coefficient Feature word category Modified double compression bracket 0.6 Entity Object Rear, left side 0.2 Entity Object Position Three, one group 0.1 Number of entity objects Equipped with 0.1 Other types of vocabulary
[0044] The results of cosine similarity calculation are combined with the hierarchical analysis method to obtain the keyword pair
[0045]
[0046] A and B represent the order information text information. i ,b i are the elements using one-hot coding numbers in the two text information, and k represents the weight coefficient w of each element obtained by the hierarchical analysis method to find the matrix.
[0047] (4) Objective similarity evaluation of order information; specifically as follows: Utilize the parameter information obtained from the order information and perform normalization processing, use Manhattan distance to obtain the similarity between the parameters, obtain a comprehensive similarity evaluation between the new order information and the historical order information, and finally use similarity sorting to obtain several reference cases and their models. The specific process is as follows: The product case information contains not only configuration text information but also model size digital information. The difference between the same type of description dimensions between different products is large. Therefore, the dimension information needs to be normalized when performing dimension similarity evaluation. Different product models require different parameter information to be extracted. Taking the basic dimensions of the model, length, width, and height, as an example, the normalization method is as follows:
[0048] sum1=a1+a2+a3
[0049]
[0050] Where a1, a2, a3 represent the length, width, and height of the model respectively, sum1 represents the sum of the three, and A1, A2, A3 represent the normalized results of the three. After normalizing the component parameters, the parameter vector is put into the set A and represented as: A = (A1, A2, ..., A n ), where n is the number of parameters extracted.
[0051] The distance between case parameters is obtained by Manhattan distance to express their similarity.
[0052]
[0053] where ‖AB‖ represents the Manhattan distance A={a1,a2,…,a n} represents the parameter set of case A, B={b1,b2,…,b n} represents the parameter set of case B. Manhattan distance is normalized.
[0054] After obtaining the size similarity and text similarity, weight distribution is performed. The similarity evaluation after weight distribution is processed as follows:
[0055] Sim=w1*Sim1+w2*Sim2
[0056] Among them, w1=0.62 is the product configuration weight, w2=0.28 is the product parameter weight, the values of w1 and w2 depend on the similarity evaluation results, and the parameters of different types of products will be different. And w1+w2=1.
[0057] (5) Output of similar cases. Sort the similarity calculation results obtained, and output the N cases with the highest similarity as reference.
[0058] Order information module: used to build standardized order information including product structure configuration, basic product parameters, and order information notes; the details are as follows: first extract the product type, product model, and the equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the company's existing models, import their corresponding information into the data table and share it with the MES system database.
[0059] Retrieval module: used to search using Boolean logic; specifically, build an index between each product order case data table in the SQL server database and the historical case model, use the type and structure configuration information in the product as constraints, and use Boolean logic to search and filter out cases that do not meet the conditions.
[0060] Subjective module: used for subjective similarity evaluation of order information; specifically as follows: One-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state register bits to encode N states, and only one bit is valid, and establish a coded data set; after word segmentation of product names and configurations to obtain keywords, One-hot encode them according to the keyword table order of the coded data set to digitize the data; calculate the cosine similarity between the codes to obtain the product note text similarity.
[0061] Objective module: used for objective similarity evaluation of order information; in the output module, the details are as follows: the parameter information obtained from the order information is used and normalized, the Manhattan distance is used to obtain the similarity between the parameters, and the comprehensive similarity evaluation between the new order information and the historical order information is obtained. Finally, the similarity ranking is used to obtain several cases and models that can be used for reference.
[0062] Output module: used for similar case output.
Claims
1. A case retrieval method based on product parameter configuration information, characterized in that: The following steps are involved: (1) Construct standardized order information including product structure configuration, basic product parameters, and order information notes; (2) Search using Boolean logic; (3) Subjective similarity evaluation of order information; (4) Objective similarity evaluation of order information; (5) Output of similar cases.
2. A case retrieval method based on product parameter configuration information according to claim 1, characterized in that: Step (1) is as follows: First, extract the product type, product model, and equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the company's existing models, import their corresponding information into the data table and share it with the MES system database.
3. A case retrieval method based on product parameter configuration information according to claim 1, characterized in that: Step (2) is as follows: construct an index between each product order case data table in the SQL server database and the historical case model, use the type and structure configuration information in the product as constraints, and use Boolean logic retrieval to filter out cases that do not meet the conditions.
4. The case retrieval method based on product parameter configuration information according to claim 1, characterized in that: Step (3) is as follows: One-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state registers to encode N states, and only one bit is valid, to establish a coded data set; use the hierarchical analysis method to obtain the weight of the keyword according to the expert's subjective scoring method, and use the text information corresponding to the code to perform One-hot encoding in the order of the keyword table of the coding data set to digitize the data; calculate the cosine similarity between the codes to obtain the product note text similarity.
5. The case retrieval method based on product parameter configuration information according to claim 1, characterized in that: Step (4) is as follows: use the parameter information obtained from the order information and perform normalization processing, use Manhattan distance to obtain the similarity between the parameters, obtain a comprehensive similarity evaluation between the new order information and the historical order information, and finally use similarity sorting to obtain several cases and models that can be used for reference.
6. A case retrieval system based on product parameter configuration information, characterized in that: include: Order information module: used to build standardized order information including product structure configuration, basic product parameters, and order information notes; Retrieval module: used to search using Boolean logic; Subjective module: used for subjective similarity evaluation of order information; Objective module: used for objective similarity evaluation of order information; Output module: used for similar case output.
7. A case retrieval system based on product parameter configuration information according to claim 5, characterized in that: In the order information module, the details are as follows: First, extract the product type, product model, and equipment name and model information contained in the order information. Create a new SQL server database, create a data table with <order number, delivery date, product type, product model, key component name, key component type, key component size> as the column name, collect and organize the company's existing models, import their corresponding information into the data table and share it with the MES system database.
8. The case retrieval system based on product parameter configuration information according to claim 5, characterized in that: In the objective module, the details are as follows: an index is built between each product order case data table in the SQLserver database and the historical case model, the type and structural configuration information in the product is used as a constraint, and Boolean logic is used to retrieve and filter cases that do not meet the conditions.
9. The case retrieval system based on product parameter configuration information according to claim 5, characterized in that: In the subjective module, the details are as follows: One-hot encode the order information notes according to the keyword table obtained after word segmentation, that is, use N-bit state registers to encode N states, and only one bit is valid, to establish a coded data set; use the hierarchical analysis method to obtain the weight of the keyword according to the expert's subjective scoring method, and use the text information corresponding to the encoding to perform One-hot encoding in the order of the keyword table of the encoding data set to digitize the data; use the cosine similarity calculation method to calculate the codes between each other to obtain the product note text similarity.
10. The case retrieval system based on product parameter configuration information according to claim 5, characterized in that: In the output module, the details are as follows: the parameter information obtained from the order information is used and normalized, the Manhattan distance is used to obtain the similarity between the parameters, and the comprehensive similarity evaluation between the new order information and the historical order information is obtained. Finally, the similarity sorting is used to obtain several cases and models that can be used for reference.
Citation Information
Cited By
Order remark information processing method, medium, equipment and program product
CN120374237A
Scheduling management system and method for hardware mold production process
CN120875429A
