CNC part automatic pricing method based on artificial intelligence and geometric feature library

Abstract

The application provides a CNC part automatic pricing method based on artificial intelligence and a geometric feature library, and is based on a geometric feature library of pre-stored standard feature machining time parameters, and is executed by an artificial intelligence model throughout, and comprises the following steps: S1, drawing uploading and preprocessing; S2, drawing recognition and feature extraction; S3, geometric feature recognition; S4, intelligent matching of process resources; S5, process simulation and collision detection; S6, intelligent calculation of machining time; S7, full-dimension cost accounting; S8, multi-scheme pricing generation; S9, pricing result output and feedback; and S10, data self-learning and model optimization. The application can solve the problems of low efficiency, large error, inability to dynamically adjust and low standardization of traditional manual pricing, and realizes the automation, precision, dynamic and standardization of CNC part pricing.

Description

Technical Field

[0001] This invention relates to the fields of intelligent manufacturing and numerical control machining technology, and in particular to an automatic quotation method for CNC parts based on artificial intelligence and geometric feature library. Background Technology

[0002] Currently, as the manufacturing industry transforms towards flexibility and intelligence, CNC machining companies are facing an increasing demand for multi-variety, small-batch orders.

[0003] In existing technologies, traditional CNC part pricing relies primarily on the manual experience of process engineers, which has the following significant drawbacks: (1) Inefficiency: Manually identifying drawings, calculating processes, and calculating costs usually takes several hours or even days, making it difficult to respond quickly to customer inquiries, especially when order volume is large, which can easily lead to quotation delays; (2) Large error: The difference in experience of different engineers will lead to the deviation in the quotation, and manual calculation is prone to missing hidden costs, resulting in loss of corporate profits or loss of customers; (3) Low standardization: There is a lack of unified pricing standards and processes. The same parts are quoted at different times and by different people with inconsistent results, making it difficult to achieve refined cost control for enterprises. (4) Cannot be dynamically adjusted: Traditional quotations are static quotations and cannot be dynamically adjusted according to real-time equipment production schedules, tool inventory, material price fluctuations, etc., nor can they provide multi-dimensional quotation options.

[0004] In addition, although there are some automatic quotation systems in the existing technology, most of them only estimate based on the weight or simple size of the parts. They cannot accurately identify the complex geometric features, tolerance requirements and process difficulty of the parts, and they are not deeply integrated with the geometric feature library, tool library and process library of intelligent flexible machining lines. The accuracy and practicality of quotation still need to be improved. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic pricing method for CNC parts based on artificial intelligence and geometric feature library, which aims to solve or at least partially solve the shortcomings of the above-mentioned background technology. It can solve the problems of low efficiency, large error, inability to dynamically adjust, and low standardization of traditional manual pricing, and realize the automation, accuracy, dynamism and standardization of CNC part pricing.

[0006] This invention provides an automatic pricing method for CNC parts based on artificial intelligence and a geometric feature library. The method constructs a geometric feature library based on pre-stored standard feature machining time parameters, and the entire process is executed by an artificial intelligence model. The method includes the following steps: S1. Image Upload and Preprocessing: Receive images from clients and perform format verification, clarity enhancement, and noise reduction on the images. S2. Drawing recognition and feature extraction: Recognize the preprocessed drawings and extract relevant information and geometric features of the parts in the drawings; S3. Geometric Feature Recognition: The extracted geometric features are matched with the geometric feature library. Geometric features that match successfully are standard features, and geometric features that do not match successfully are non-standard features. Non-standard features are further divided into adjustable features and forced features. S4. Intelligent matching of process resources: Based on the relevant information of the parts, the corresponding process resources are automatically matched for standard features. For adjustable features, after customer confirmation, they are converted into corresponding standard features and the same matching process is executed. For mandatory features, the corresponding process resources are matched, and the processing difficulty of the mandatory features is graded and a mandatory feature process difficulty coefficient is generated. S5. Process Simulation and Collision Detection: Based on the matched process resources, generate the CNC machining path of the part, perform virtual machining simulation and collision interference detection, and optimize the process parameters and machining path of the part according to the simulation results. S6. Intelligent calculation of processing time: Based on the standard feature processing time parameters and the process difficulty coefficient of the forced feature pre-stored in the geometric feature library, calculate the total processing time of standard features and adjustable features as well as the total processing time of forced features. Add the auxiliary time to obtain the processing time of a single part and the total processing time of the part. S7. Full-dimensional cost accounting: Calculate multiple cost items one by one to obtain the basic processing cost of the parts; S8. Multiple Quotation Generation: Based on the basic processing cost and combined with different process routes for the parts, at least three quotation options are generated. S9. Quotation Result Output and Feedback: Output the quotation result and cost details to the enterprise and receive confirmation or modification feedback from the enterprise; S10. Data self-learning and model optimization: Upload actual processing data, cost data, and quotation feedback data to the artificial intelligence model to iteratively optimize feature recognition accuracy, process matching parameters, and cost accounting model.

[0007] Further, in step S1, the drawing file is a two-dimensional drawing or a three-dimensional model file; the two-dimensional drawing includes CAD format drawing and PDF format drawing, and the three-dimensional model file includes a three-dimensional STEP model file and a three-dimensional IGES model file.

[0008] Furthermore, in step S2, the relevant information of the part includes: basic information of the part: drawing number, name, material grade, processing quantity, drawing scale, and unit of measurement; process requirements of the part: heat treatment requirements, surface treatment requirements, inspection requirements, and specific processing instructions; tolerance information of the part: form tolerance, orientation tolerance, position tolerance, and runout tolerance, as well as the tolerance grade and accuracy requirements corresponding to each tolerance.

[0009] Furthermore, in step S2, the geometric features include the dimensions, positions, and quantities of outer circles, inner holes, planes, steps, grooves, through holes, blind holes, chamfers, fillets, threads, center holes, bosses, and grooves.

[0010] Furthermore, in step S3, adjustable features can be converted into corresponding standard features by adjusting dimensional parameters without affecting the core function and assembly requirements of the part, and with customer confirmation; mandatory features are features that are explicitly marked by the customer as unmodifiable or that affect the core function, assembly relationship, or performance of the part, and must be processed strictly in accordance with the customer's drawing requirements.

[0011] Furthermore, in step S4, the process resources include machine tools, tooling fixtures and cutting tools. The forced feature process difficulty coefficient is determined based on the feature's geometric complexity, machining accuracy requirements, material cutting difficulty and equipment occupancy time, and in combination with historical processing data of similar features.

[0012] Furthermore, in step S5, the specific steps include: S51. Automatically generate CNC machining paths for roughing, semi-finishing, and finishing. S52. Construct a digital twin virtual machining environment that includes process resources and parts; S53. Real-time detection of collisions, interference, overtravel, and cutting parameter overload during the machining process; S54. Automatically optimize the processing path or adjust the process parameters, and re-simulate until there are no abnormalities; S55. Record the optimized process parameters and processing path for subsequent calculations.

[0013] Furthermore, in step S7, the various cost items include material costs, unit processing costs, management costs, special process costs, testing costs, shared and operating costs, and capital costs; among them, the costs of customized tools and special tooling corresponding to the mandatory features and the process adjustment costs generated by simulation optimization are accounted for separately.

[0014] Further, in step S8, the specific pricing schemes are as follows: Lowest cost scheme: Select the process route and equipment with the lowest processing cost, prioritize the standard feature conversion scheme for adjustable features, adopt conventional production scheduling, and set the lowest profit coefficient; Fastest delivery scheme: Prioritize currently idle equipment, adopt expedited process routes, compress the production cycle, prioritize process resources for mandatory features, set the highest profit coefficient, and include expedited production costs; Standard scheme: Adopt standard process routes and equipment, produce according to conventional scheduling, provide two options for adjustable features: conversion scheme and original scheme, and set the median profit coefficient.

[0015] Furthermore, in step S10, the specific steps include: S101. Collect actual processing time, tool wear, quality and cost data; S102. Compare actual production data with estimated quotation data, calculate the deviation value, and focus on analyzing the deviation of non-standard characteristic costs. S103. Adjust the standard feature processing time, non-standard feature process difficulty coefficient, and cost accounting model parameters according to the deviation value; S104. Convert high-frequency adjustable features into corresponding standard features and update the geometric feature library, and incorporate mature forced feature processing technology into the dedicated process resource library. S105. Update the corresponding database to achieve continuous iterative optimization of the artificial intelligence model.

[0016] This invention provides an automatic pricing method for CNC parts based on artificial intelligence and a geometric feature library. By automatically identifying geometric features in customer drawings and matching them with standardized processes using an AI model, the pricing time for a single part can be reduced from hours to minutes, enabling rapid response to customer inquiries and effectively improving order conversion rates. By strictly adopting a comprehensive standardized cost accounting system covering all explicit and implicit cost items, it eliminates subjective biases caused by human experience, improving the accuracy of CNC part pricing and preventing profit loss or customer churn. By establishing unified pricing standards and processes, it eliminates pricing discrepancies between different personnel and at different times, achieving refined cost control and standardized production management. By dynamically calculating based on real-time equipment scheduling, material price fluctuations, and delivery requirements, it generates differentiated multi-solution pricing, flexibly adapting to different customer needs and improving customer satisfaction. Through a closed-loop feedback mechanism between actual processing data and pricing data, it continuously iterates and optimizes feature recognition accuracy, process matching parameters, and cost accounting models, achieving dynamic alignment between pricing results and actual production conditions, forming a virtuous cycle. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of an automatic quotation method for CNC parts based on artificial intelligence and a geometric feature library, according to the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. Based on the description of the present invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present invention.

[0020] Unless otherwise explicitly specified and limited, the terms "setup," "installation," and "connection" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of these terms based on the specific circumstances.

[0021] The terms “upper,” “lower,” “left,” “right,” “front,” “back,” “top,” “bottom,” “inner,” and “outer,” etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use. They are only for the convenience of description and simplification, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0022] The terms “first,” “second,” “third,” etc., are used merely to distinguish elements with similar properties, not to indicate or imply relative importance or a specific order.

[0023] The terms “include,” “comprising,” or any other variation thereof are intended to cover non-exclusive inclusion, which includes not only the elements listed but also other elements not expressly listed.

[0024] This invention provides an automatic pricing method for CNC parts based on artificial intelligence and a geometric feature library. Its key feature is the construction of a geometric feature library based on pre-stored standard feature machining time parameters, executed entirely by an artificial intelligence model, including the following steps: S1. File Upload and Preprocessing: Receive files from clients via mobile app or computer. First, verify file integrity to check for corruption and the presence of necessary layer information. Then, perform format verification, sharpness enhancement, and noise reduction. Specifically, an adaptive histogram equalization algorithm is used to enhance the sharpness of blurred annotation areas, and a deep learning-based edge detection algorithm is used to remove watermarks, unwanted lines, and irrelevant annotations.

[0025] More specifically, the drawings are either two-dimensional drawings or three-dimensional model files; two-dimensional drawings include CAD format drawings and PDF format drawings, and three-dimensional model files include three-dimensional STEP model files and three-dimensional IGES model files.

[0026] S2. Drawing Recognition and Feature Extraction: The pre-processed drawings are identified, and relevant information and geometric features of the parts within the drawings are extracted. Geometric features include the dimensions, location, and quantity of outer circles, inner holes, planes, steps, grooves, through holes, blind holes, chamfers, fillets, threads, center holes, bosses, and grooves.

[0027] Furthermore, the relevant information for the parts includes: (1) Basic information of the part: drawing number, name, material grade, processing quantity, drawing scale, and unit of measurement.

[0028] (2) Process requirements for parts: heat treatment requirements, surface treatment requirements, inspection requirements and specific processing instructions.

[0029] (3) Tolerance information of parts: form tolerance, orientation tolerance, position tolerance and runout tolerance, as well as the tolerance grade and accuracy requirements corresponding to each tolerance.

[0030] S3. Geometric Feature Recognition: The extracted geometric features are matched with the geometric feature library. Geometric features that match successfully are standard features, and geometric features that do not match successfully are non-standard features. Non-standard features are further divided into adjustable features and forced features.

[0031] Furthermore, adjustable features can be transformed into corresponding standard features by adjusting dimensional parameters, provided that the core function and assembly requirements of the part are not affected and the customer has confirmed them. Mandatory features are features that the customer has clearly marked as unmodifiable or that have a significant impact on the core function, assembly relationship, and performance of the part. They must be processed strictly in accordance with the customer's drawing requirements.

[0032] S4. Intelligent matching of process resources: Based on the relevant information of the parts, the corresponding process resources are automatically matched for standard features. For adjustable features, after customer confirmation, they are converted into corresponding standard features and the same matching process is executed. For mandatory features, the corresponding process resources are matched, and the processing difficulty of the mandatory features is graded and a mandatory feature process difficulty coefficient is generated.

[0033] More specifically, process resources include machine tools, tooling fixtures, and cutting tools. The matching of process resources employs a multi-objective optimization algorithm, simultaneously considering three objectives: processing cost, processing efficiency, and equipment load rate, to generate the optimal matching solution.

[0034] More specifically, the process difficulty coefficient of the forced feature is determined based on the geometric complexity of the feature, the machining accuracy requirements, the material cutting difficulty, and the equipment occupancy time, combined with historical processing data of similar features.

[0035] S5. Process Simulation and Collision Detection: Based on the matched process resources, generate the CNC machining path of the part, perform virtual machining simulation and collision interference detection, and optimize the process parameters and machining path of the part according to the simulation results.

[0036] More specifically, the specific steps of step S5 include: S51: Automatically generates CNC machining paths for roughing, semi-finishing, and finishing.

[0037] S52. Construct a digital twin virtual machining environment that includes process resources (machine tools, tooling fixtures, cutting tools) and parts.

[0038] S53. Real-time detection of collisions, interference, overtravel, and cutting parameter overload during the machining process.

[0039] S54. Automatically optimize the processing path or adjust the process parameters, and re-simulate until there are no abnormalities.

[0040] S55. Record the optimized process parameters and processing path for subsequent calculations.

[0041] S6. Intelligent Calculation of Processing Time: Based on the standard feature processing time parameters and the process difficulty coefficient of the forced feature pre-stored in the geometric feature library, the total processing time of standard features and adjustable features, as well as the total processing time of forced features, are calculated. Auxiliary time is added to obtain the processing time of a single part and the total processing time of the part.

[0042] It should be noted that the above standard feature machining time parameters include the time for each stage of roughing, semi-finishing, and finishing, as well as correction coefficients for different materials and machine tools; the above auxiliary time is automatically calculated based on the number of times the part is clamped, the number of tool changes, and the number of operations, while also taking into account the time allocation coefficient for batch production.

[0043] S7. Full-Dimensional Cost Accounting: Calculates various cost items item by item to obtain the basic processing cost of the part. These cost items include material costs, unit processing costs, management costs, special process costs, testing costs, shared and operating costs, and capital costs. The costs of customized tools and special fixtures corresponding to mandatory features, as well as the process adjustment costs generated by simulation optimization, are calculated separately.

[0044] Establish a dynamic cost database to synchronize data such as material prices, equipment hourly rates, and labor costs in real time; for bulk orders, adopt a tiered cost accounting method, where the larger the batch, the lower the unit cost; for urgent orders, additional calculations are required for equipment occupancy costs and overtime labor costs.

[0045] S8. Multiple Quotation Generation: Based on the basic processing cost and combined with different process routes for the parts, generate at least three quotation options.

[0046] The following are three pricing options: (1) Lowest cost scheme: Select the process route and equipment with the lowest processing cost, prioritize the standard feature conversion scheme for adjustable features, adopt conventional production scheduling, and set the minimum profit coefficient; (2) Fastest delivery plan: Prioritize currently idle equipment, adopt expedited process routes, compress production cycle, prioritize process resources for mandatory features, set the highest profit coefficient, and include expedited production costs; (3) Standard Scheme: Adopt standard process routes and equipment, and produce according to the regular schedule. For adjustable features, provide two options: conversion scheme and original scheme, and set the intermediate profit coefficient.

[0047] S9. Quotation Result Output and Feedback: Outputs quotation results and cost details to the enterprise and receives confirmation or modification feedback from the enterprise. More specifically, it supports exporting quotation sheets in multiple formats such as PDF and Excel, including part drawings, process flow diagrams, and cost details; it establishes a quotation feedback processing flow, automatically recalculating and generating new quotation schemes for customer modification suggestions.

[0048] S10. Data Self-Learning and Model Optimization: Upload actual processing data, cost data, and quotation feedback data to the artificial intelligence model for iterative optimization of feature recognition accuracy, process matching parameters, and cost accounting models. Establish a data quality assessment mechanism to filter out abnormal data and employ incremental learning algorithms to continuously absorb new processing data without affecting the performance of the original model.

[0049] The sub-steps of this step include: S101. Collect actual processing time, tool wear, quality and cost data.

[0050] S102. Compare actual production data with estimated quotation data, calculate the deviation value, and focus on analyzing the deviation of non-standard characteristic costs.

[0051] S103. Adjust the processing time of standard features, the process difficulty coefficient of non-standard features, and the parameters of the cost accounting model according to the deviation value.

[0052] S104. Convert the high-frequency adjustable features into corresponding standard features and update the geometric feature library, and incorporate the mature machining processes of forced features into the dedicated process resource library. The dedicated process resource library refers to a database specifically storing mature machining processes for forced features, including specific tool parameters, tooling design schemes, and combinations of machining parameters for specific geometric complexities and accuracy requirements.

[0053] S105. Update the corresponding database to achieve continuous iterative optimization of the artificial intelligence model.

[0054] In summary, this invention provides an automatic pricing method for CNC parts based on artificial intelligence and a geometric feature library. By automatically identifying geometric features in customer drawings and matching them with standardized processes using an AI model, the pricing time for a single part can be reduced from hours to minutes, enabling rapid response to customer inquiries and effectively improving order conversion rates. By strictly adopting a comprehensive standardized cost accounting system covering all explicit and implicit cost items, subjective biases caused by human experience can be eliminated, improving the accuracy of CNC part pricing and preventing profit loss or customer churn. By establishing unified pricing standards and processes, price discrepancies between different personnel and at different times can be eliminated, achieving refined cost control and standardized production management. By dynamically calculating based on real-time equipment scheduling, material price fluctuations, and delivery requirements, differentiated multi-scheme pricing can be generated, flexibly adapting to different customer needs and improving customer satisfaction. Through a closed-loop feedback mechanism between actual processing data and pricing data, the accuracy of feature recognition, process matching parameters, and cost accounting models can be continuously iterated and optimized, achieving dynamic alignment between pricing results and actual production conditions, forming a virtuous cycle.

[0055] The present invention will be further described in detail below with reference to specific embodiments.

[0056] This embodiment provides an automatic pricing method for CNC parts based on artificial intelligence and a geometric feature library, applied to an intelligent flexible machining line. This machining line is equipped with six three-axis CNC machine tools, a geometric feature library, a standard tool library, a standard process library, and an intelligent manufacturing control system. The geometric feature library pre-stores standard parameters, tool matching information, process parameters, and standard machining time parameters for various core geometric features.

[0057] This embodiment uses an aluminum alloy flange part as an example for further detailed explanation.

[0058] S1. File Upload and Preprocessing: The customer uploads the 3D STEP model file of the aluminum alloy flange via computer. The AI ​​model first performs file integrity verification to confirm that the file is not damaged and contains all necessary layers. Then, it performs format verification, clarity enhancement, and removes watermarks and irrelevant lines from the drawing.

[0059] S2. Drawing Recognition and Feature Extraction: Recognize the pre-processed drawings and extract relevant information and geometric features of the aluminum alloy flange in the drawings.

[0060] The relevant information for the parts includes: (1) Basic information: The part name is FL-001, the material grade is 6061 aluminum alloy, the processing quantity is 100 pieces, the drawing scale is 1:1, and the unit of measurement is mm.

[0061] (2) Process requirements: The surface treatment is anodizing, the tolerance is IT12 unless otherwise specified, the finished product inspection is carried out by three coordinate measuring machine, and the sharp edges are blunted and deburred.

[0062] (3) Tolerance information: The tolerance of the φ30mm inner hole is H7, the flatness of the end face is 0.02mm, and the position tolerance of the through hole is φ0.1mm.

[0063] (4) Geometric features: includes one φ80mm outer circle, one φ30mm inner hole, one end face, four φ7.5mm through holes, four M6 threaded holes, and four C1.2 chamfers.

[0064] S3, Geometric Feature Recognition: Match the extracted geometric features with the geometric feature library.

[0065] (1) Standard features: φ80mm outer circle, φ30mm inner hole, end face, M6 threaded hole, C1.2 chamfer.

[0066] (2) Non-standard features: φ7.5mm through hole.

[0067] (3) Judgment of non-standard features: The φ7.5mm through hole is a non-fitting hole, which does not affect the core function of the part and the assembly requirements, and is judged as an adjustable feature.

[0068] S4. Intelligent matching of process resources: The artificial intelligence model matches process resources with standard features and adjustable features based on relevant information of the parts.

[0069] (1) Standard features: Matches BF-V8 three-axis CNC machine tool, standard three-jaw chuck equipped with locating pin tool, external turning tool, internal boring tool, M6 machine tap, and chamfering tool; (2) Adjustable features: Automatically generate a solution to convert a φ7.5mm through hole into a standard φ8mm through hole, evaluate that it has no impact on the function of the part, push it to the customer for confirmation, and after the customer agrees to the conversion, execute the standard feature matching process, and match the φ8mm drill bit on the basis of the above standard feature process resources.

[0070] S5, Process Simulation and Collision Detection S51 automatically generates CNC machining paths for roughing, semi-finishing, and finishing, optimizes toolpaths, and reduces idle travel by 15%.

[0071] S52. Construct a high-precision digital twin virtual machining environment that includes a BF-V8 machine tool kinematic model, a three-jaw chuck, cutting tools, and a flange.

[0072] S53. Real-time collision detection was performed, and no collision, interference, overtravel, or cutting parameter overload issues were found.

[0073] S54. The processing technology is normal, and there is no need to optimize the processing path or adjust the process parameters.

[0074] S55. Record the existing process parameters and processing paths for subsequent calculations.

[0075] S6, Intelligent Calculation of Processing Time (1) Statistics on processing time for geometric features: φ80mm outer diameter: 2 min / piece; φ30mm inner diameter: 3 min / piece; end face: 1 min / piece; M6 threaded hole: 0.5 min / piece; C1.2 chamfer: 0.2 min / piece; φ8mm through hole: 0.3 min / piece.

[0076] Total standard feature time: 2min + 3min + 1min + 4 × 0.5min + 4 × 0.2min + 4 × 0.3min = 10min.

[0077] (2) Auxiliary working hours statistics: Total clamping time: 2 min; Total tool change time: 1 min; Process auxiliary time: 1 min.

[0078] Auxiliary working time: 2min + 1min + 1min = 4min.

[0079] (3) Processing time for a single flange: 10min + 4min = 14min.

[0080] (4) Total processing time: 14min × 100 pieces + 30min (batch preparation time) = 1430min ≈ 23.83h.

[0081] S7, Comprehensive Cost Accounting (1) Material cost: 6061 aluminum alloy is a commonly used material. The cost is calculated using the enterprise's unified batch purchase cost accounting, which includes material purchase cost (including transportation management fee) and 5% cutting loss cost, totaling RMB 806.4.

[0082] (2) Cost of processing a single part: 23.83h (total processing time) × 80 yuan / hour + 200 yuan (tool wear cost) + 50 yuan (tooling depreciation cost) = 2156.4 yuan.

[0083] (3) Management cost: 23.83h (total processing hours) × 30 yuan / hour = 714.9 yuan.

[0084] (4) Special process cost: 5 yuan / piece × 100 pieces (anodizing cost) = 500 yuan.

[0085] (5) Inspection cost: 1 yuan / piece × 100 pieces (three-coordinate inspection fee) = 100 yuan.

[0086] (6) Shared and operating costs: 23.83h (total processing hours) × 10 yuan / hour = 238.3 yuan.

[0087] (7) Cost of capital: Calculated based on a 30-day monthly settlement method, totaling RMB 50.

[0088] (8) Add up all cost items to get the basic processing cost of the whole batch of parts. 806.4+2156.4+714.9+500+100+238.3+50=4566.0 yuan.

[0089] S8, Multiple Quotation Generation The following are three pricing options: (1) Lowest cost option: adopt the standard process route, arrange production during periods of low equipment load, profit coefficient 0.15, batch price: 4566.0×(1+0.15)=5250.9 yuan, single piece price is about 52.51 yuan, delivery time is 7 days, and the quotation is valid for 15 days.

[0090] (2) Fastest delivery plan: Prioritize production using currently available machine tools, adopt the expedited process route, profit coefficient 0.25, expedited fee 500 yuan, total price for the whole batch: 4566.0×(1+0.25)+500=6207.5 yuan, about 62.08 yuan per piece, delivery time 3 days, quotation validity period 7 days; (3) Standard Scheme: Using standard process route and regular scheduling, profit coefficient 0.20, the price of the whole batch is 4566.0×(1+0.2)+500=5479.2 yuan, the price of a single piece is about 54.79 yuan, the delivery time is 5 days, and the quotation is valid for 10 days.

[0091] S9. Quotation Result Output and Feedback: The AI ​​model outputs three quotation schemes and detailed cost breakdowns to the enterprise, while also highlighting cost savings resulting from adjustable feature transformations. It supports exporting quotation sheets in PDF and Excel formats, including part drawings, process flow diagrams, and cost details. The enterprise selects and confirms the standard scheme.

[0092] S10. Data Self-Learning and Model Optimization: After the batch of parts was processed, the AI ​​model automatically collected the actual processing time as 23.2 hours and the actual cost as 4850 yuan. Comparing this with the estimated quote data, the deviation was calculated, revealing that the actual processing time for the φ8mm through-hole was 0.05 minutes less per hole than estimated. The system uses an incremental learning algorithm to fine-tune the processing time parameters and cost calculation model for the φ8mm through-hole in the geometric feature library and update it to the corresponding database.

[0093] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for automatic pricing of CNC parts based on artificial intelligence and geometric feature library, characterized in that, The geometric feature library is constructed based on pre-stored standard feature processing time parameters. The entire process is executed by an artificial intelligence model, including the following steps: S1. Image Upload and Preprocessing: Receive the client's image files and perform format verification, clarity enhancement, and noise reduction on the image files; S2. Drawing recognition and feature extraction: Recognize the preprocessed drawing and extract relevant information and geometric features of the parts in the drawing; S3. Geometric Feature Recognition: The extracted geometric features are matched with a geometric feature library. Geometric features that match successfully are considered standard features, while those that do not match successfully are considered non-standard features. Non-standard features are further divided into adjustable features and forced features. S4. Intelligent matching of process resources: Based on the relevant information of the part, the corresponding process resources are automatically matched for the standard features. For the adjustable features, after being confirmed by the customer and converted into the corresponding standard features, the same matching process is executed. For the mandatory features, the corresponding process resources are matched, and the processing difficulty of the mandatory features is graded and a mandatory feature process difficulty coefficient is generated. S5. Process Simulation and Collision Detection: Based on the matched process resources, generate the CNC machining path of the part, perform virtual machining simulation and collision interference detection, and optimize the process parameters and machining path of the part according to the simulation results. S6. Intelligent calculation of processing time: Based on the standard feature processing time parameters pre-stored in the geometric feature library and the process difficulty coefficient of the forced feature, calculate the total processing time of the standard feature and the adjustable feature, as well as the total processing time of the forced feature. Add the auxiliary time to obtain the processing time of a single part and the total processing time of the part. S7. Full-dimensional cost accounting: Calculate multiple cost items one by one to obtain the basic processing cost of the part; S8. Multiple quotation generation: Based on the basic processing cost and combined with different process routes of the part, at least three quotation schemes are generated. S9. Quotation Result Output and Feedback: Output the quotation result and cost details to the enterprise and receive confirmation or modification feedback from the enterprise; S10. Data self-learning and model optimization: Upload actual processing data, cost data, and quotation feedback data to the artificial intelligence model to iteratively optimize the feature recognition accuracy, process matching parameters, and cost accounting model.

2. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S1, the drawing file is a two-dimensional drawing or a three-dimensional model file; the two-dimensional drawing includes CAD format drawing and PDF format drawing, and the three-dimensional model file includes a three-dimensional STEP model file and a three-dimensional IGES model file.

3. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S2, the relevant information of the part includes: Basic information of the part: drawing number, name, material grade, quantity to be processed, drawing scale, and unit of measurement; The process requirements for the parts include: heat treatment requirements, surface treatment requirements, inspection requirements, and specific processing instructions. The tolerance information of the part includes: form tolerance, orientation tolerance, position tolerance and runout tolerance, as well as the tolerance grade and accuracy requirements corresponding to each tolerance.

4. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S2, the geometric features include the dimensions, positions, and quantities of outer circles, inner holes, planes, steps, grooves, through holes, blind holes, chamfers, fillets, threads, center holes, bosses, and grooves.

5. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S3, the adjustable feature is converted into the corresponding standard feature by adjusting the dimensional parameters, provided that it does not affect the core function and assembly requirements of the part and is confirmed by the customer. The mandatory feature is a feature that the customer has clearly marked as unmodifiable or that affects the core function, assembly relationship, or performance of the part. It must be processed strictly in accordance with the customer's drawing requirements.

6. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S4, the process resources include machine tools, tooling fixtures and cutting tools. The forced feature process difficulty coefficient is determined based on the geometric complexity of the feature, the machining accuracy requirements, the material cutting difficulty and the equipment occupancy time, and in combination with historical processing data of similar features.

7. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S5, the specific steps include: S51. Automatically generate CNC machining paths for roughing, semi-finishing, and finishing. S52. Construct a digital twin virtual machining environment that includes the process resources and the parts; S53. Real-time detection of collisions, interference, overtravel, and cutting parameter overload during the machining process; S54. Automatically optimize the processing path or adjust the process parameters, and re-simulate until there are no abnormalities; S55. Record the optimized process parameters and processing path for subsequent calculations.

8. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S7, the various cost items include material costs, unit processing costs, management costs, special process costs, testing costs, shared and operating costs, and capital costs; wherein the costs of customized tools and special tooling corresponding to the mandatory features and the process adjustment costs generated by simulation optimization are accounted for separately.

9. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S8, the pricing scheme is specifically as follows: Lowest cost solution: Select the process route and equipment with the lowest processing cost, prioritize the standard feature conversion scheme for the adjustable features, adopt conventional production scheduling, and set the minimum profit coefficient; Fastest delivery plan: Prioritize currently idle equipment, adopt expedited process routes, compress production cycles, prioritize process resources for the aforementioned mandatory features, set the highest profit coefficient, and include expedited production costs; Standard solution: Using standard process routes and equipment, and following standard production schedules, the adjustable features are provided with two options: conversion and original solutions, with a median profit coefficient set.

10. The automatic pricing method for CNC parts based on artificial intelligence and geometric feature library as described in claim 1, characterized in that, In step S10, the specific steps include: S101. Collect actual processing time, tool wear, quality and cost data; S102. Compare the actual production data with the estimated quotation data, calculate the deviation value, and focus on analyzing the cost deviation of the non-standard features. S103. Adjust the standard feature processing time, non-standard feature process difficulty coefficient, and cost accounting model parameters according to the deviation value; S104. Convert the high-frequency adjustable features into corresponding standard features and update the geometric feature library, and incorporate the mature processing technology of the forced features into the dedicated process resource library. S105. Update the corresponding database to achieve continuous iterative optimization of the artificial intelligence model.

Images