Scheduling management system and method for hardware mold production process

Abstract

The invention discloses a scheduling management system and method for a hardware mold production process. The scheduling management method for the hardware mold production process comprises the steps of obtaining a hardware mold order uploaded by a sales department; extracting mold design image data from the hardware mold order; inputting the mold design image data into a processing procedure analysis engine based on AI drive to obtain a recommended processing procedure list and sequence; after a confirmation instruction which is provided by a technician and aims at the recommended processing procedure list and sequence is received, extracting a material demand list from the hardware mold order; inputting the recommended processing procedure list and sequence and the material demand list into a productivity evaluation module to obtain an estimated feasible delivery period of the hardware mold order; and an expected feasible delivery period is extracted from the hardware mold order, and whether a delay delivery early warning prompt is generated is determined based on comparison between the estimated feasible delivery period and the expected feasible delivery period.

Description

Technical Field

[0001] This application relates to the field of hardware mold production management, specifically to a scheduling and management method for the hardware mold production process. Background Technology

[0002] As a key component of industrial production, the manufacturing process of metal molds typically involves complex design, high-precision machining, and close collaboration among multiple processes. In the face of fierce market competition, efficiently and accurately completing mold production and ensuring on-time delivery are crucial for enterprises to control costs and enhance competitiveness. Therefore, refined scheduling and management of the metal mold production process has become a goal pursued by modern manufacturing enterprises. Accurate acquisition and planning of the processing steps required for mold production is the cornerstone of the entire scheduling and management process, directly determining the accuracy and feasibility of subsequent material procurement, equipment scheduling, personnel allocation, and final delivery date estimation.

[0003] However, traditional hardware mold production scheduling and management, especially in the acquisition and planning of processing steps, generally suffers from numerous technical problems. Existing solutions typically rely on experienced technicians manually interpreting orders and making judgments based on their experience to initially determine the list and sequence of processing steps. This highly experience-dependent model leads to two main problems: First, process path planning relies on experience and lacks optimization, making it difficult to quickly and systematically generate optimal or suboptimal process solutions when faced with increasingly complex and varied mold structures. Second, the prediction of manufacturing difficulties is insufficient and delayed. Potential manufacturing challenges (such as machining thin walls, deep cavities, complex curved surfaces, or problems caused by special material properties) are often not fully identified and warned of in the early stages of design review, leading to problems only surfacing during actual production and resulting in uncontrollable risks such as rework, delayed delivery, and escalating costs. Furthermore, this manual approach also faces challenges such as low information extraction efficiency, insufficient standardization, difficulties in knowledge accumulation and reuse, and inaccurate time estimation, seriously affecting the accuracy and scheduling efficiency of the entire production plan.

[0004] Therefore, a new scheduling and management method for the hardware mold production process is needed. Summary of the Invention

[0005] One advantage of this application is that it provides a scheduling and management method for the production process of hardware molds. This method can solve the problem that process path planning relies on experience and lacks optimization, and can make early and forward-looking predictions and warnings about whether delivery can be made on time, effectively avoiding the problems of insufficient and delayed prediction of delivery dates.

[0006] According to one aspect of this application, a scheduling and management method for a hardware mold production process is provided, comprising: Obtain hardware mold orders uploaded by the sales department; Extract mold design image data from the hardware mold order; The mold design image data is input into an AI-driven processing procedure analysis engine to obtain a recommended list and sequence of processing procedures; Upon receiving confirmation instructions from technical personnel regarding the recommended processing procedure list and sequence, a material requirement list is extracted from the hardware mold order. Input the recommended processing procedure list and sequence and the material requirements list into the capacity assessment module to obtain the estimated feasible delivery date of the hardware mold order; Extract the expected feasible delivery date from the hardware mold order, and determine whether to generate a delayed delivery warning based on the comparison between the estimated feasible delivery date and the expected feasible delivery date.

[0007] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, the mold design image data is input into an AI-driven processing procedure analysis engine to obtain a recommended list and sequence of processing procedures. This includes: performing geometric analysis and recognition on the mold design image data to obtain geometric information analysis results; extracting dimensional and tolerance information from the mold design image data to obtain dimensional information analysis results and tolerance information analysis results; extracting technical text information from the drawing annotation area in the mold design image data to obtain technical text information analysis results; performing feature engineering and structuring on the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results to obtain structured mold feature data; and using the structured mold feature data through the AI-driven processing procedure analysis engine to obtain the recommended list and sequence of processing procedures.

[0008] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, geometric analysis and recognition are performed on the mold design image data to obtain geometric information analysis results, including: analyzing the geometric entities in the mold design image data through a computer vision module; and recognizing the standard mold structure and basic geometric features in the mold design image data through a computer vision module.

[0009] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, technical text information is extracted from the drawing annotation area in the mold design image data to obtain the technical text information parsing result, including: performing template matching on the mold design image data to obtain the ROI image of the drawing annotation area; and performing OCR text recognition on the ROI image of the drawing annotation area to obtain the technical text information parsing result.

[0010] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, the structured mold feature data is processed through the AI-driven processing procedure parsing engine to obtain the recommended processing procedure list and sequence, including: retrieving historical successful cases that best match the structured mold feature data from the historical mold case library based on a feature similarity algorithm; and extracting the processing procedure list and sequence of the best-matching historical successful cases as the recommended processing procedure list and sequence.

[0011] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, the method retrieves the historical successful case that best matches the structured mold feature data from the historical mold case library based on a feature similarity algorithm, including: calculating the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices; and taking the historical successful case corresponding to the largest one in the set of Jaccard indices as the best-matching historical successful case.

[0012] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, calculating the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices includes: extracting multiple current feature values ​​from the structured mold feature data of the current order and multiple historical feature values ​​from the structured mold feature data of each historical successful case; performing correlation embedding processing on each current feature value to obtain a set of embedded correlation feature values; and calculating the Jaccard index between the set of embedded correlation feature values ​​and the set of historical feature values ​​in the structured mold feature data of each historical successful case to obtain a set of Jaccard indices.

[0013] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, the method performs correlation embedding processing on each current feature value to obtain a set of embedded correlation feature values, including: performing intersection / union cardinality modeling on each of the current feature values ​​to obtain a first correlation parameter and a second correlation parameter; and performing similarity correlation embedding on the current feature values ​​based on the first correlation parameter and the second correlation parameter to obtain multiple embedded correlation feature values.

[0014] In one embodiment of the scheduling and management method for the hardware mold production process according to this application, the intersection / union cardinality relationship model is performed on each of the current feature values ​​to obtain a first association parameter and a second association parameter, including: calculating the minimum value among the structured mold feature data of the current order. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the first association parameter; the minimum value in the structured mold feature data of the current order is calculated. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the second correlation parameter, where, ,and ,in, It represents the total number of current feature values.

[0015] According to another aspect of this application, this application proposes a scheduling and management system for the hardware mold production process, which includes: The order acquisition unit is used to acquire hardware mold orders uploaded by the sales department; The image extraction unit is designed to extract mold design image data from the hardware mold order. The processing procedure acquisition unit is used to input the mold design image data into an AI-driven processing procedure parsing engine to obtain a recommended list and sequence of processing procedures; The material requirements list acquisition unit is used to extract the material requirements list from the hardware mold order after receiving a confirmation instruction from a technician regarding the recommended processing procedure list and sequence. The estimated feasible delivery date acquisition unit is used to input the recommended processing procedure list and sequence and the material requirement list into the capacity assessment module to obtain the estimated feasible delivery date of the hardware mold order; The delivery assessment unit is used to extract the expected feasible delivery date from the hardware mold order and determine whether to generate a delayed delivery warning based on the comparison between the estimated feasible delivery date and the expected feasible delivery date.

[0016] The further objectives and advantages of this application will become fully apparent from the following description and accompanying drawings. Attached Figure Description

[0017] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 The illustration shows a flowchart of a scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0019] Figure 2The illustration shows a flowchart of S3 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0020] Figure 3 The illustration shows a flowchart of S31 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0021] Figure 4 The illustration shows a flowchart of S33 in another embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0022] Figure 5 The illustration shows a flowchart of S35 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0023] Figure 6 The illustration shows a flowchart of S351 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0024] Figure 7 The illustration shows a flowchart of S3511 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0025] Figure 8 The illustration shows a flowchart of S35112 in one embodiment of the scheduling and management method for the hardware mold production process according to an embodiment of this application.

[0026] Figure 9 The diagram illustrates a structural block diagram of a scheduling and management system for the hardware mold production process according to an embodiment of this application. Detailed Implementation

[0027] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0028] It is understood that the term "a" should be understood as "at least one" or "one or more," meaning that in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple. The term "a" should not be construed as a limitation on the quantity. "Multiple" means two or more.

[0029] While ordinal numbers such as “first,” “second,” etc., will be used to describe various components, there is no limitation on those components herein. The term is used only to distinguish one component from another; for example, a first component may be referred to as a second component, and similarly, a second component may be referred to as a first component, without departing from the teachings of this application. The term “and / or” as used herein includes any and all combinations of one or more of the associated listed items.

[0030] The terminology used herein is for the purpose of describing various embodiments only and is not intended to be limiting. As used herein, the singular form also includes the plural form, unless the context clearly indicates otherwise. It will also be understood that the terms “comprising” and / or “having” as used in this specification specify the presence of the described features, numbers, operations, components, elements or combinations thereof, without excluding the presence or addition of one or more other features, numbers, operations, components, elements or combinations thereof.

[0031] As mentioned above, existing solutions typically rely on experienced technicians to manually interpret mold design orders and make judgments based on their experience to initially determine the list and sequence of processing steps. This highly experience-dependent model leads to two main problems: first, process path planning relies on experience and lacks optimization; second, the prediction of manufacturing difficulties is insufficient and delayed. Furthermore, this manual processing method faces challenges such as low information extraction efficiency, insufficient standardization, difficulties in knowledge accumulation and reuse, and inaccurate time estimation, seriously affecting the accuracy and scheduling efficiency of the entire production plan.

[0032] Therefore, a new scheduling and management method for the hardware mold production process is needed.

[0033] This application proposes a method to obtain a recommended list and sequence of machining processes by deeply analyzing mold design image data from orders and inputting it into an AI-driven machining process analysis engine. Specifically, mold design image data, as the original carrier of mold design information, contains key information such as geometry, dimensional tolerances, material requirements, and various technical annotations from customer orders. However, traditional manual interpretation of these drawings and planning of machining processes is inefficient and prone to errors. By inputting this mold design image data into an AI-driven machining process analysis engine, the advantages of artificial intelligence in image recognition, feature extraction, pattern recognition, and knowledge reasoning can be fully utilized. This AI analysis engine can automatically perform geometric analysis and recognition of 2D / 3D mold drawings, extract dimensional and tolerance information, and extract technical text information from the drawing annotation area using natural language processing technology. Subsequently, this unstructured or semi-structured information undergoes feature engineering and structuring processing to form standardized mold feature data that can be understood and processed by machines. Based on this, the AI-driven analysis engine can intelligently reason and recommend the optimal or better list and sequence of machining processes by combining a built-in mold manufacturing knowledge graph and historical case library. This not only overcomes the blind reliance of traditional solutions on the experience of a single engineer and achieves process path optimization, but also combines manufacturability design rules to make early and forward-looking predictions and warnings of potential manufacturing difficulties, avoiding the lag of problems only emerging in the later stages of production. The recommendation characteristics of AI can provide technicians with data-driven, optimized preliminary solutions, while the final confirmation and fine-tuning are still completed by technicians, realizing human-machine collaboration and significantly improving the efficiency, accuracy, and robustness of process planning. This provides a solid and reliable foundation for subsequent capacity assessment, delivery forecasting, and overall production scheduling management.

[0034] Based on this, this application proposes a scheduling and management method for the hardware mold production process. For example... Figures 1 to 8 As shown, a scheduling and management method for the hardware mold production process according to an embodiment of this application is illustrated. Figure 1As shown, the scheduling and management method for the hardware mold production process includes: S1, obtaining hardware mold orders uploaded by the sales department; S2, extracting mold design image data from the hardware mold orders; S3, inputting the mold design image data into an AI-driven processing procedure analysis engine to obtain a recommended processing procedure list and sequence; S4, after receiving a confirmation instruction from a technician regarding the recommended processing procedure list and sequence, extracting a material requirement list from the hardware mold orders; S5, inputting the recommended processing procedure list and sequence and the material requirement list into a capacity assessment module to obtain an estimated feasible delivery date for the hardware mold orders; S6, extracting the expected feasible delivery date from the hardware mold orders, and determining whether to generate a delayed delivery warning based on a comparison between the estimated feasible delivery date and the expected feasible delivery date.

[0035] Specifically, in step S1, the hardware mold order uploaded by the sales department is obtained. It should be understood that the order contains all the fundamental and crucial information required for mold production. This information directly affects subsequent process route selection, material preparation cycle, equipment load, and the final delivery commitment. Obtaining the hardware mold order ensures that this critical information is accurately transmitted to the production stage, providing a basis for subsequent delivery date assessment and delay warnings.

[0036] The hardware mold order mainly includes basic order information, mold identification information, mold design image data, technical specifications or requirements, bill of materials (BOM), and expected feasible delivery date. Basic order information mainly includes customer name, order number, order creation date, and sales manager. Mold identification information mainly includes mold model, name, and expected production quantity (e.g., one mold or multiple molds). Mold design image data is the most crucial technical part, mainly including 2D / 3D CAD drawing files: such as STEP, IGES, DWG, DXF, and other commonly used formats, representing 3D models or 2D engineering drawings of the mold, which detail the overall structure of the mold, the geometry of each part, detailed dimensions, geometric tolerances, assembly relationships, etc. The technical specifications or requirements mainly describe in text form the material grade used in the mold (e.g., type of specific mold steel, hardness requirements), surface treatment requirements (e.g., nitriding, PVD coating), precision grade, mold life requirements, mold opening method, cooling system design, etc. The Bill of Materials (BOM) primarily specifies the main materials required for the mold (such as steel plate batch number and specifications) and standard parts (such as ejector pins, guide pillars, springs, screws, etc.). The Expected Feasible Delivery Date (EFD) is the customer's explicit requirement or expectation for mold delivery time and serves as important reference data for subsequent capacity assessment and delay warnings. The hardware mold order also includes other relevant information, such as special testing or acceptance requirements, packaging, and transportation requirements.

[0037] It is worth mentioning that in this application, the mold design image data in the order is used as the basic data to perform in-depth analysis of the order in order to obtain the list and sequence of processing steps. Accordingly, in step S2, the mold design image data is extracted from the hardware mold order.

[0038] Furthermore, to avoid the problems of reliance on experienced technicians for manual interpretation and judgment of orders, resulting in experience-dependent process path planning lacking optimization and insufficient and delayed prediction of manufacturing difficulties, this application proposes to use an AI-driven processing step analysis engine to analyze mold design image data and obtain a recommended list and sequence of processing steps. Accordingly, in step S3, the mold design image data is input into the AI-driven processing step analysis engine to obtain the recommended list and sequence of processing steps.

[0039] Specifically, in step S3, although the mold design image data contains complete geometric information of the mold, for the AI-driven machining process analysis engine, this raw data is merely a digital "graphic". The AI-driven machining process analysis engine cannot directly "understand" from these graphics which are holes, which are slots, which are cavities, and their dimensions, positions, and relationships. Therefore, in-depth analysis of the mold design image data is required. More specifically, this mainly involves geometric analysis, dimensional analysis, and text information analysis of the mold design image data, and then determining the machining process list and sequence based on the analysis results.

[0040] Accordingly, in one embodiment of this application, as Figure 2 As shown, step S3 includes: S31, performing geometric analysis and recognition on the mold design image data to obtain geometric information analysis results; S32, extracting dimensional and tolerance information from the mold design image data to obtain dimensional information analysis results and tolerance information analysis results; S33, extracting technical text information from the drawing annotation area in the mold design image data to obtain technical text information analysis results; S34, performing feature engineering and structuring on the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results to obtain structured mold feature data; S35, using the structured mold feature data through the AI-driven processing procedure analysis engine to obtain the recommended processing procedure list and sequence.

[0041] In step S31, the mold design image data is subjected to geometric analysis and recognition to obtain geometric information analysis results. For example... Figure 3As shown, geometric analysis is performed through the following steps: S311, the geometric entities in the mold design image data are analyzed by the computer vision module; S312, the standard mold structure and basic geometric features in the mold design image data are identified by the computer vision module.

[0042] Specifically, in step S311, in one embodiment of this application, the computer vision module first converts the mold design image data into a unified pixel image format; then it performs preprocessing such as noise removal, contrast enhancement, binarization, and image correction to improve the accuracy of subsequent recognition and obtain a preprocessed image; next, it performs geometric feature extraction (point, line, and surface recognition), for example, by detecting all edges and contours in the image using algorithms such as Canny and Sobel; and by recognizing straight lines and arcs in the image using algorithms such as Hough transform to obtain geometric feature vectors.

[0043] In step S312, the preprocessed image and the geometric feature vector are identified by the trained deep convolutional neural network to identify standard mold structures (such as cavities, cores, sliders, angled ejectors, gates, runners, cooling water channels, ejector pin holes, etc.) and basic geometric features (holes, slots, bosses, fillets, draft angles, etc.).

[0044] In step S32, dimensional and tolerance information is extracted from the mold design image data to obtain dimensional information analysis results and tolerance information analysis results. Specifically, various dimensional data (vertical deviation, limit deviation, symmetry deviation, etc.) and tolerance symbols (such as flatness, perpendicularity, position, concentricity, etc.) in the mold design image data are identified by a computer vision module.

[0045] In step S33, technical text information is extracted from the drawing annotation area of ​​the mold design image data to obtain the technical text information parsing result. Specifically, as follows... Figure 4 As shown, technical text information is extracted through the following steps: S331, template matching is performed on the mold design image data to obtain the ROI image of the drawing annotation area; S332, OCR text recognition is performed on the ROI image of the drawing annotation area to obtain the parsing result of the technical text information. ROI refers to the region of interest. OCR refers to Optical Character Recognition.

[0046] In step S331, the preprocessed image can be identified by a trained deep convolutional neural network to identify areas in the mold design drawings that contain important technical information, such as the title block, technical requirements block, revision records, and precautions. These areas are collectively referred to as the drawing annotation area.

[0047] In step S332, the trained deep convolutional neural network is used to identify glyphs (letters, numbers, and symbols) in the ROI image of the drawing annotation area, i.e., OCR text recognition is performed. After the glyphs are initially identified, corrections can be made through dictionary matching, spell checking, and context judgment, and noise reduction processing such as removing redundant spaces, line breaks, and unnecessary punctuation marks can be performed to obtain the final OCR text recognition result.

[0048] Furthermore, the OCR text recognition results need to be further analyzed to obtain the meaning and classification of each character in the OCR text recognition results. Specifically, regular expressions or keyword matching technology can be used for keyword recognition and information extraction, to identify common technical terms in drawings and extract their corresponding values. For example, identifying the text following the material to extract the material number (e.g., SKD11, P20); identifying the hardness value and treatment method following heat treatment to extract heat treatment requirements (e.g., HRC58, carburizing and quenching); identifying the numerical value and symbol following surface roughness to extract surface treatment requirements (e.g., Ra0.8, mirror polishing); identifying free text under remarks to extract general precautions (e.g., deburring, chamfering); and identifying key information such as drawing number, part number, version number, and revision history.

[0049] In step S34, feature engineering and structuring are performed on the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results to obtain structured mold feature data. It should be understood that the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results are discrete. For example, the geometric information analysis result may include: there is a circle at a first position; the dimensional information analysis result may include: the diameter of the circle is 10mm; the dimensional information analysis result may include: upper and lower deviations ±0.02mm; and the technical text information analysis result may include: a threaded hole. To better understand the information in the mold design image data, the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results need to be correlated to form a complete structured information set. For example, there may be a threaded hole at a first position with a diameter of 10mm and a tolerance of ±0.02mm.

[0050] Specifically, the results of the geometric information parsing, the dimensional information parsing, the tolerance information parsing, and the technical text information parsing are first aligned and correlated. For example, for each identified geometric entity (such as a circle), the dimensional information parsing results are traversed to find its corresponding dimensional information (such as diameter and radius) annotations; the tolerance information parsing results are traversed to find the dimensional tolerances and geometric tolerances corresponding to the geometric entity; and the technical text information parsing results are traversed to find related supplementary explanations (such as if the hole is a threaded hole). The correlation is achieved using spatial location matching and feature ID matching mechanisms. Then, feature classification is performed based on the associated information. For example, low-level geometric entities are elevated to features with manufacturing semantics (such as combining a circle with its depth to determine whether it is a through hole or a blind hole). Next, attribute filling and derivation can be performed, filling all associated data as attributes of the feature object. New attributes (such as aspect ratio, accuracy level, and suggested processing technology group) can be derived according to preset rules and calculation formulas. Subsequently, a hierarchical structure is constructed to organize the hierarchical relationships between features (such as classifying multiple holes under a certain mold part), forming a complete mold feature tree or graph structure. Finally, data formatting is performed, outputting the final feature data according to a predefined JSON, XML, or database schema.

[0051] In step S35, the structured mold feature data is processed by the AI-driven machining process analysis engine to obtain the recommended machining process list and sequence. It is worth noting that this application uses the AI-driven machining process analysis engine to draw on a vast number of historical successful cases for optimization, that is, to find the optimal or second-best process solution.

[0052] It should be understood that the higher the similarity between the structured mold feature data of the current order and the structured mold feature data in historical successful cases, the more instructive the processes of the historical successful cases will be for the processes of the current order. After finding successful cases in historical successful cases whose structured mold feature data is similar to that of the current order, the processing process list and sequence of the similar successful cases can be used as the processing process list and sequence of the current order, or the processing process list and sequence of the similar successful cases can be optimized to generate the processing process list and sequence of the current order.

[0053] Accordingly, in one embodiment of this application, as Figure 5 As shown, step S35 includes: S351, retrieving historical successful cases from the historical mold case library that best match the structured mold feature data based on the feature similarity algorithm; S352, extracting the processing procedure list and sequence of the best matching historical successful cases as the recommended processing procedure list and sequence.

[0054] In step S351, a historical successful case that best matches the structured mold feature data is retrieved from the historical mold case library based on a feature similarity algorithm. Specifically, a suitable similarity algorithm can be selected to calculate the similarity between the structured mold feature data of the current order and the structured mold feature data of each historical successful case, and the historical successful case that is most similar to the structured mold feature data is selected as the best matching historical successful case.

[0055] In one embodiment of this application, the Jaccard index algorithm is selected to calculate the structured mold feature data of the current order and the structured mold feature data of each historical successful case. The Jaccard index (Jaccard similarity coefficient) is used to compare the similarity and differences between a finite set of samples. The larger the Jaccard index, the higher the sample similarity. The Jaccard index of two sets is the ratio of the size of the intersection of the two sets to the size of the size of the union of the two sets.

[0056] Accordingly, such as Figure 6 As shown, step S351 includes: S3511, calculating the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices; S3512, taking the historical successful case corresponding to the largest Jaccard index in the set as the best matching historical successful case.

[0057] In step S3511, the Jaccard index is calculated between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices. It should be understood that the structured mold feature data of the current order includes multiple current feature values, which is a set of current feature values, such as geometric feature values, dimensional feature values, tolerance feature values, and text feature values. Similarly, the structured mold feature data of each historical successful case includes multiple historical feature values, which is a set of historical feature values. The calculation of the Jaccard index involves calculating the Jaccard index between the set of current feature values ​​and the set of historical feature values.

[0058] It is worth mentioning that when calculating the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library, the similarity between the sets is measured by the Jaccard index by abstracting the geometric analysis information, the quasi-geometric representation information of dimensions and tolerances and the textual semantic information together into a structured feature set through feature engineering. Therefore, it is necessary to consider the intersection / union correlation between discrete features within the set based on the Jaccard index.

[0059] In other words, during the intersection / union correlation mapping process for calculating the Jaccard index, while reconstructing the mapping space, it is also necessary to maintain the invariance of the correlation relationships (such as value adjacency, value connectivity, etc.) between discrete features of the sets, to avoid distortion of the structured feature set correlation caused by spatial mapping, which would affect the accuracy of the calculated Jaccard index. Therefore, before calculating the Jaccard index, this application performs correlation embedding processing on each current feature value to maintain the correlation relationships between discrete features of the sets.

[0060] Accordingly, such as Figure 7 As shown, step S3511 includes: S35111, extracting multiple current feature values ​​from the structured mold feature data of the current order and multiple historical feature values ​​from the structured mold feature data of each historical successful case; S35112, performing correlation embedding processing on each current feature value to obtain a set of embedded correlation feature values; S35113, calculating the Jaccard index between the set of embedded correlation feature values ​​and the set of historical feature values ​​in the structured mold feature data of each historical successful case to obtain a set of Jaccard indices.

[0061] In step S35112, each current feature value is subjected to association embedding processing to obtain a set of embedded association feature values. Specifically, in one embodiment of this application, for each current feature value in the structured mold feature data of the current order, for example, denoted as... First, based on unified geometric and semantic features, we model the cardinality relationship of intersection / union: ;in, Indicates the first associated parameter; Indicates the first The current feature values; This represents the smallest value in the set of current eigenvalues. The sum of eigenvalues; This represents the largest value in the set of current eigenvalues. The sum of eigenvalues.

[0062] ;in, Indicates the second associated parameter; This represents the smallest value in the set of current eigenvalues. The sum of eigenvalues; This represents the largest value in the set of current eigenvalues. The sum of eigenvalues; ,and ,in, This represents the total number of current feature values ​​in the structured mold feature data of the current order. Thus, for discontinuous geometric and semantic objects in set representations, modeling based on discrete intersection / union cardinality relationships avoids the smooth relation dependencies such as continuous value adjacency and continuous value connectivity found in feature manifold representations.

[0063] Then, similarity association embedding is performed from a set perspective: ;in, This represents the embedded association feature values. In this way, within the intersection / union similarity association mapping space, the association relationships of the feature values ​​remain unchanged while eliminating redundant association structures, satisfying the minimum orthogonality constraint requirement of similarity representation under set theory, i.e., avoiding spatial structured association intersections. Thus, by calculating the Jaccard index between the optimized structured mold feature data in the current order and the structured mold feature data of each historical successful case in the historical mold case library, the accuracy of the calculated Jaccard index can be improved.

[0064] Accordingly, such as Figure 8 As shown, step S35112 includes: S101, performing intersection / union cardinality modeling on each of the current feature values ​​to obtain a first association parameter and a second association parameter; S102, performing similarity association embedding on the current feature values ​​based on the first association parameter and the second association parameter to obtain multiple embedded association relationship feature values. In step S101, the smallest value among the structured mold feature data of the current order is calculated. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the first association parameter; the minimum value in the structured mold feature data of the current order is calculated. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the second correlation parameter, where, ,and ,in, It represents the total number of current feature values.

[0065] In step S35113, the Jaccard index is calculated between the set of embedded correlation feature values ​​and the set of historical feature values ​​in the structured mold feature data of each historical success case to obtain a set of Jaccard indices. Specifically, the Jaccard index between the set of embedded correlation feature values ​​and the set of historical feature values ​​in the structured mold feature data of a historical success case is calculated using the following formula: ;in, This represents the set of feature values ​​representing embedded associations; Indicates the first A collection of historical feature values ​​from structured mold feature data of historical successful cases; express and The size of the intersection; express and The size of the union of the sets.

[0066] In step S4, after receiving confirmation instructions from technicians regarding the recommended processing procedure list and sequence, a material requirement list is extracted from the hardware mold order. It should be understood that the recommended processing procedure list and sequence obtained by the AI-driven processing procedure analysis engine only represents potentially feasible processing procedures. The complexity of mold manufacturing, subtle differences in experience, and unforeseen circumstances (such as temporary equipment failures or new material properties) cannot guarantee on-time delivery when processing is carried out according to this procedure. Materials are the foundation of all production; therefore, a material requirement list needs to be further extracted from the hardware mold order to estimate feasible delivery dates based on the current mold manufacturing production line's capacity.

[0067] In step S5, the recommended processing procedure list and sequence, along with the material requirements list, are input into the capacity assessment module to obtain the estimated feasible delivery date for the hardware mold order. The recommended processing procedure list and sequence indicate the process by which an order is completed; the material requirements list indicates the materials used to complete an order; and the workshop's capacity information indicates how much the workshop can complete. By integrating the recommended processing procedure list and sequence and the material requirements list and matching them with the workshop's current capacity information, a rough completion date can be estimated, i.e., the estimated feasible delivery date.

[0068] Specifically, taking the production of a new mold MOLD-001 as an example, the recommended list and sequence of processing steps include, but are not limited to: Step 1: Rough milling of the cavity, requiring CNC milling machine No. 1 (model: VM-1000L), estimated processing time 8 hours, no preceding steps (only material arrival required); Step 2: Finish milling of the cavity, requiring CNC milling machine No. 1 (model: VM-1000L), estimated processing time 5 hours, preceding step step 1; Step 3: Wire cutting of holes, requiring wire cutting machine No. 1 (model: WEDM-600), estimated processing time 6 hours, preceding step step 1; Step 4: Heat treatment, requiring heat treatment furnace, estimated processing time 16 hours (including cooling), preceding steps steps 2 and 3; Step 5: Grinding, requiring grinding machine (model: Grinder-200), estimated processing time 4 hours, preceding step step 4; Step 6: Assembly and debugging, requiring assembly table and general tools, estimated processing time 12 hours, preceding step step 5.

[0069] The material requirements list includes, but is not limited to: Main material: H13 steel block, 300x200x100mm, quantity 1 piece, estimated delivery time: T+5 days (today is T+0; T is the transportation period); Standard parts: 1 set of guide post sleeve, in stock; Auxiliary materials: 20L of cutting fluid, in stock.

[0070] Workshop capacity information includes, but is not limited to: Today is Monday, 8:00 AM, T+0, 8 hours on weekdays, no work on weekends; CNC milling machine No. 1: Model VM-1000L, currently occupied by order "Product-XYZ" from Monday 8:00 AM to Wednesday 4:00 PM (T+0 to T+2.5 days), available after Wednesday 4:00 PM (T+2.5 days); Wire cutting machine No. 1: Model WEDM-600, currently occupied by order "Product-ABC" from Monday 8:00 AM to Tuesday 8:00 AM (T+0 to T+1 days), available after Tuesday 8:00 AM (T+1 day); Heat treatment furnace: currently empty, available at any time; Grinding machine: currently occupied by order "Product-DEF" from Monday 8:00 AM to Tuesday 12:00 PM (T+0 to T+1.5 days), available after Tuesday 12:00 PM (T+1.5 days); Assembly table: currently empty, available at any time.

[0071] After the recommended processing procedure list and sequence, along with the material requirements list, are input into the capacity assessment module, the module first analyzes the material delivery time: the main material, H13 steel blocks, will arrive no later than T+5 days; therefore, any process requiring main materials cannot start until at least T+5 days later. Standard parts and auxiliary materials are already in stock and will not be a bottleneck; the earliest project start time is T+5 days (8:00 AM next Monday).Then, the process timeline is arranged as follows: Process 1 (Rough Milling of Cavity): Requires CNC milling machine No. 1, earliest start time T+5 days (after all materials arrive). CNC milling machine No. 1 is available on T+2.5 days, so the equipment is not a bottleneck. Start time is Monday 8:00 AM (T+5 days), end time is Monday 8:00 AM + 8 hours = Monday 4:00 PM (T+5 days); Process 2 (Fine Milling of Cavity): Requires CNC milling machine No. 1. The preceding process, Process 1, ends on Monday 4:00 PM (T+5 days). CNC milling machine No. 1 is idle at the start of the fine milling process (because it has just completed Process 1). Start time is Monday 4:00 PM (T+5 days), end time is Monday 4:00 PM + 5 hours = Tuesday 9:00 AM (T+6 days, because 4 hours remain on Monday, and 1 hour is used on Tuesday, leaving only 8 hours of production that day); Process 3 (Wire EDM Hole): Required equipment: Wire EDM machine No. 1. The preceding process, Process 1, ends at 4:00 PM on Monday (T+5 days). Wire EDM machine No. 1 is available on T+1 day. After Process 1 is completed, the wire EDM machine is ready to use. Start time: Monday 4:00 PM (T+5 days). End time: Monday 4:00 PM + 6 hours = Tuesday 10:00 AM (T+6 days, 4 hours remain on Monday, 2 hours used on Tuesday); Process 4 (Heat Treatment): Required equipment: Heat treatment furnace. The preceding process, Process B, ends at 9:00 AM on Tuesday (T+6 days). Process 3 ends at 10:00 AM on Tuesday (T+6 days), whichever ends latest. The heat treatment furnace is available immediately. Start time: Tuesday 10:00 AM (T+6 days). End time: Tuesday 10:00 AM + 16 hours = Thursday 10:00 AM (T+8 days, because there are 6 hours on Tuesday, 8 hours on Wednesday, and 2 hours on Thursday); Process 5 (Grinding): Required equipment is a grinding machine. The preceding process, Process 4, ends at 10:00 AM on Thursday (T+8 days). The grinding machine is available in T+1.5 days, so its availability is not a bottleneck. Start time is Thursday 10:00 AM (T+8 days), and end time is Thursday 10:00 AM + 4 hours = Thursday 2:00 PM (T+8 days); Process 6 (Assembly and Debugging): Required equipment is an assembly table. The preceding process, Process 5, ends at 2:00 PM on Thursday (T+8 days). The assembly table is available at any time. Start time is Thursday 2:00 PM (T+8 days), and end time is Thursday 2:00 PM + 12 hours = Friday 4:00 PM (T+9 days, because there are 6 hours on Thursday and 6 hours on Friday). According to the calculation, all processes will be completed by Friday 4:00 PM (T+9 days). Therefore, the estimated feasible delivery date for order MOLD-001 is T+9 days, which is 4 PM the following Friday.

[0072] In step S6, the expected feasible delivery date is extracted from the hardware mold order, and based on the comparison between the estimated feasible delivery date and the expected feasible delivery date, it is determined whether to issue a delayed delivery warning. Specifically, the hardware mold order includes the expected feasible delivery date. By comparing the estimated feasible delivery date with the expected feasible delivery date, an early and forward-looking prediction and warning can be made regarding whether delivery can be made on time, effectively avoiding the problems of insufficient or delayed prediction of delivery dates.

[0073] For example, in one instance, the expected feasible delivery date is within 30 days of receiving the order; assuming today is November 20, 2023, then the expected delivery date is December 19, 2023; the estimated feasible delivery date is T+9 days; the estimated feasible delivery date is extrapolated to November 29, 2023; the estimated feasible delivery date is earlier than the expected feasible delivery date; therefore, it is determined that there is no risk of delay.

[0074] In another example, the expected feasible delivery date is November 25, 2023; the estimated feasible delivery date is November 29, 2023; the estimated feasible delivery date is later than the expected feasible delivery date; it is determined that there is no risk of delay, and the delay period is 4 days.

[0075] This application also proposes a scheduling and management system for the hardware mold production process. Specifically, such as... Figure 9 As shown, the scheduling and management system for the hardware mold production process includes an order acquisition unit 11, a design image extraction unit 12, a processing procedure acquisition unit 13, a material requirement list acquisition unit 14, an estimated feasible delivery date acquisition unit 15, and a delivery evaluation unit 16. The order acquisition unit 11 is used to acquire hardware mold orders uploaded by the sales department; the design image extraction unit 12 is used to extract mold design image data from the hardware mold orders; the processing procedure acquisition unit 13 is used to input the mold design image data into an AI-driven processing procedure parsing engine to obtain a recommended processing procedure list and sequence; the material requirement list acquisition unit 14 is used to extract a material requirement list from the hardware mold orders after receiving a confirmation instruction from a technician regarding the recommended processing procedure list and sequence; the estimated feasible delivery date acquisition unit 15 is used to input the recommended processing procedure list and sequence and the material requirement list into a capacity assessment module to obtain an estimated feasible delivery date for the hardware mold orders; the delivery assessment unit 16 is used to extract the expected feasible delivery date from the hardware mold orders and, based on a comparison between the estimated feasible delivery date and the expected feasible delivery date, determine whether to generate a delayed delivery warning.

[0076] In summary, the scheduling management system and method for the hardware mold production process have been clarified. The scheduling management method for the hardware mold production process can solve the problem of process path planning relying on experience and lacking optimization, and can provide early, forward-looking predictions and warnings regarding whether delivery can be made on time, effectively avoiding problems of insufficient or delayed predictions of delivery dates.

[0077] The present application and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present application. The actual structure is not limited to this. In conclusion, if a person skilled in the art is inspired by this description and designs a similar structure and embodiment without departing from the spirit of the present application, such design should fall within the protection scope of the present application.

Claims

1. A scheduling and management method for the production process of hardware molds, characterized in that, include: Obtain hardware mold orders uploaded by the sales department; Extract mold design image data from the hardware mold order; The mold design image data is input into an AI-driven processing procedure analysis engine to obtain a recommended list and sequence of processing procedures; Upon receiving confirmation instructions from technical personnel regarding the recommended processing procedure list and sequence, a material requirement list is extracted from the hardware mold order. Input the recommended processing procedure list and sequence and the material requirements list into the capacity assessment module to obtain the estimated feasible delivery date of the hardware mold order; Extract the expected feasible delivery date from the hardware mold order, and determine whether to generate a delayed delivery warning based on the comparison between the estimated feasible delivery date and the expected feasible delivery date.

2. The scheduling and management method for the hardware mold production process according to claim 1, characterized in that, The mold design image data is input into an AI-driven machining process analysis engine to obtain a recommended list and sequence of machining processes, including: The mold design image data is subjected to geometric analysis and recognition to obtain geometric information analysis results; The dimensions and tolerances of the mold design image data are extracted to obtain the dimension information analysis results and the tolerance information analysis results; Technical text information is extracted from the drawing annotation area in the mold design image data to obtain the technical text information parsing result; Feature engineering and structuring are performed on the geometric information analysis results, the dimensional information analysis results, the tolerance information analysis results, and the technical text information analysis results to obtain structured mold feature data; The structured mold feature data is used by the AI-driven processing procedure parsing engine to obtain the recommended processing procedure list and sequence.

3. The scheduling and management method for the hardware mold production process according to claim 2, characterized in that, The geometric analysis and recognition of the mold design image data are performed to obtain geometric information analysis results, including: The geometric entities in the mold design image data are analyzed using a computer vision module; The computer vision module identifies the standard mold structure and basic geometric features in the mold design image data.

4. The scheduling and management method for the hardware mold production process according to claim 2, characterized in that, Technical text information is extracted from the drawing annotation area of ​​the mold design image data to obtain the technical text information parsing results, including: Template matching is performed on the mold design image data to obtain the ROI image of the drawing annotation area; OCR text recognition is performed on the ROI image of the drawing annotation area to obtain the parsing result of the technical text information.

5. The scheduling and management method for the hardware mold production process according to claim 2, characterized in that, The structured mold feature data is processed by the AI-driven machining process analysis engine to obtain the recommended machining process list and sequence, including: Based on the feature similarity algorithm, retrieve the historical successful cases from the historical mold case library that best match the structured mold feature data; Extract the processing procedure list and sequence from the most matching historical successful cases as the recommended processing procedure list and sequence.

6. The scheduling and management method for the hardware mold production process according to claim 5, characterized in that, Based on a feature similarity algorithm, historical successful cases that best match the structured mold feature data are retrieved from the historical mold case library, including: Calculate the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices; The historical success case corresponding to the largest Jaccard index in the set is taken as the best matching historical success case.

7. The scheduling and management method for the hardware mold production process according to claim 6, characterized in that, Calculate the Jaccard index between the structured mold feature data and the structured mold feature data of each historical successful case in the historical mold case library to obtain a set of Jaccard indices, including: Extract multiple current feature values ​​from the structured mold feature data of the current order and multiple historical feature values ​​from the structured mold feature data of each historical successful case; Perform correlation embedding processing on each current feature value to obtain a set of feature values ​​with embedded correlation relationships; The Jaccard index is calculated between the set of embedded association feature values ​​and the set of historical feature values ​​in the structured mold feature data of each historical success case to obtain the set of Jaccard indices.

8. The scheduling and management method for the hardware mold production process according to claim 7, characterized in that, Perform association embedding processing on each current feature value to obtain a set of embedded association feature values, including: Model the intersection / union cardinality relationship of each of the current feature values ​​to obtain the first association parameter and the second association parameter; Based on the first association parameter and the second association parameter, the current feature value is embedded with similarity association to obtain multiple embedded association relationship feature values.

9. The scheduling and management method for the hardware mold production process according to claim 8, characterized in that, Modeling the intersection / union cardinality relationship of each of the current feature values ​​to obtain the first association parameter and the second association parameter includes: Calculate the minimum value among the structured mold feature data of the current order. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the first correlation parameter; Calculate the minimum value among the structured mold feature data of the current order. The sum of the current eigenvalues ​​and the largest The ratio of the sum of the current feature values ​​is used to obtain the second correlation parameter, where, ,and ,in, It represents the total number of current feature values.

10. A scheduling and management system for the production process of hardware molds, characterized in that, include: The order acquisition unit is used to acquire hardware mold orders uploaded by the sales department; The image extraction unit is designed to extract mold design image data from the hardware mold order. The processing procedure acquisition unit is used to input the mold design image data into an AI-driven processing procedure parsing engine to obtain a recommended list and sequence of processing procedures; The material requirements list acquisition unit is used to extract the material requirements list from the hardware mold order after receiving a confirmation instruction from a technician regarding the recommended processing procedure list and sequence. The estimated feasible delivery date acquisition unit is used to input the recommended processing procedure list and sequence and the material requirement list into the capacity assessment module to obtain the estimated feasible delivery date of the hardware mold order; The delivery assessment unit is used to extract the expected feasible delivery date from the hardware mold order and determine whether to generate a delayed delivery warning based on the comparison between the estimated feasible delivery date and the expected feasible delivery date.

Images