Method and system for determining machining time and computer program product

Abstract

The invention provides a method and system for determining machining time and a computer program product, and relates to the technical field of intelligent manufacturing. The method for determining the machining time comprises the steps that a three-dimensional model of a to-be-machined part is obtained, and the to-be-machined part is a machined part; based on the three-dimensional model, matching a first historical model from a historical database, and taking related data of the first historical model as template data; identifying to-be-processed features of the three-dimensional model through a part identification model; based on the template data, processing requirements of the to-be-processed features are determined; identifying feature parameters of the to-be-processed features by using a CAD module; and determining the machining time of the to-be-machined part based on the machining requirements, the feature parameters and the machining process knowledge base of the to-be-machined features. According to the method, the defects of inconsistent man-hour estimation and the like caused by disjunction of the process and calculation, lack of intelligent decision and incapability of self-optimization of the system in the traditional method are overcome.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing, and in particular to a machining working hour determination method and system and computer program product. BACKGROUND

[0002] In the field of mechanical manufacturing, the traditional machining working hour determination method mainly relies on process engineers to manually estimate based on two-dimensional engineering drawings and personal experience. This method is not only inefficient and time-consuming, but also its accuracy highly depends on the personal technical level and subjective judgment of engineers, resulting in a huge difference in working hour estimation results of the same part by different personnel, which seriously affects the accuracy of enterprise cost accounting, the competitiveness of pricing and the executability of production planning. In addition, with the popularization of three-dimensional CAD design, the traditional two-dimensional drawing interpretation method has been difficult to fully utilize the rich geometric and manufacturing information contained in the three-dimensional model, and manual extraction and identification of machining features are prone to errors and incomplete, which has become the main bottleneck restricting the development of working hour estimation towards automation and intelligence.

[0003] At present, some computer-aided working hour estimation methods have improved efficiency to some extent, but still have significant deficiencies. For example, some methods use simple historical data analogy or fixed formula calculation, which fail to intelligently associate the specific three-dimensional geometric characteristics of the part with dynamic machining process parameters (such as cutting parameters and tool selection), resulting in a disconnection between process planning and working hour calculation, and a large deviation between the estimation results and actual machining requirements. At the same time, existing systems generally lack effective self-learning and optimization capabilities, and their core knowledge base and calculation model are difficult to automatically update and continuously improve with the accumulation of new machining cases and the evolution of production conditions, limiting the practicality and long-term accuracy of the system.

[0004] Therefore, there is an urgent need in the art for a new solution that can deeply utilize three-dimensional model information, realize intelligent linkage of process and working hour, and have self-evolution ability. SUMMARY

[0005] The embodiments of the present specification aim to solve the problems in the prior art that the machining working hour determination method excessively relies on manual experience, is inefficient, has poor accuracy, and is difficult to fully utilize three-dimensional model information. Specifically, the embodiments of the present specification provide an automatic, intelligent and continuously evolving machining working hour determination method, system and computer program product.

[0006] The one or more embodiments of the specification provide a method for determining machining time, comprising: obtaining a three-dimensional model of a part to be machined, the part to be machined being a machining part; matching a first historical model from a historical database based on the three-dimensional model, and taking relevant data of the first historical model as template data; identifying a feature to be machined of the three-dimensional model by a part recognition model; determining machining requirements of the feature to be machined based on the template data; identifying feature parameters of the feature to be machined by a CAD module; and determining machining time of the part to be machined based on the machining requirements, the feature parameters, and a machining process knowledge base of the feature to be machined.

[0007] In some embodiments, matching a first historical model from a historical database based on the three-dimensional model comprises: obtaining point cloud data of the three-dimensional model; and matching a model with the highest similarity from the historical database as the first historical model based on the point cloud data, the similarity being determined according to at least one of material, size, shape, and machining accuracy requirement of the part.

[0008] In some embodiments, identifying a feature to be machined of the three-dimensional model by a part recognition model comprises: obtaining point cloud data of the three-dimensional model; and identifying the feature to be machined based on the point cloud data by the part recognition model, the part recognition model being a trained PointNet network model.

[0009] In some embodiments, the template data comprises historical machining requirements of historical features of the first historical model; and determining machining requirements of the feature to be machined based on the template data comprises: determining a target historical feature corresponding to the feature to be machined among the historical features of the first historical model; and determining the machining requirements of the feature to be machined based on historical machining requirements of the target historical feature.

[0010] In some embodiments, the machining requirements comprise machining accuracy and / or surface roughness of the feature to be machined.

[0011] In some embodiments, determining machining time of the part to be machined based on the machining requirements, the feature parameters, and the machining process knowledge base of the feature to be machined comprises: calculating machining time for each feature to be machined respectively, specifically comprising: determining machining process sequence, recommended machining tool, and tool machining parameter of the feature to be machined based on the machining requirements and the feature parameters, and combining the machining process knowledge base; and calculating preparation time and machining time of the feature to be machined according to the machining process sequence and the tool machining parameter, and combining the machining process knowledge base; wherein the preparation time at least comprises machining tool switching time, and the machining time is calculated based on the tool machining parameter and the feature parameter; and summing up the preparation time and the machining time of all features to be machined to determine the machining time of the part to be machined.

[0012] In some embodiments, the preparation time further includes at least one of a machine tool idle time, a workpiece clamping and alignment time.

[0013] In some embodiments, the method for determining the machining time further includes: displaying at least one of machining requirements, feature parameters, machining process sequences, recommended machining tools and tool machining parameters of the to-be-machined feature in the interactive interface; obtaining user feedback; and adjusting the machining requirements, feature parameters, machining process sequences, recommended machining tools and / or tool machining parameters of the to-be-machined feature based on the user feedback.

[0014] In some embodiments, the method for determining the machining time further includes: adding the three-dimensional model, the to-be-machined feature, the machining requirements, the feature parameters, the machining process sequences, the tool machining parameters and the machining time of the to-be-machined part to the historical database and the machining process knowledge base.

[0015] In some embodiments, the related data includes a historical machining time calculation scheme of the first historical model, and the template data includes a machining time calculation template; and the determining the machining time of the to-be-machined part based on the machining requirements, the feature parameters of the to-be-machined feature and the machining process knowledge base includes: determining the machining time of the to-be-machined part based on the machining requirements, the feature parameters of the to-be-machined feature, the machining process knowledge base and the machining time calculation template.

[0016] In some embodiments, the related data includes a historical machining time calculation scheme of the first historical model, and the template data includes a machining time calculation template; and the method for determining the machining time further includes: when the part recognition model fails to recognize the to-be-machined feature of the three-dimensional model, determining a volume of the to-be-machined part; and determining the machining time of the to-be-machined part based on the volume of the to-be-machined part, the volume of the first historical model and the machining time calculation template.

[0017] The one or more embodiments of the specification also provide a machining time determination system, comprising a part acquisition module, a template matching module, a feature recognition module, a machining requirement determination module, a feature parameter recognition module and a time determination module; the part acquisition module is configured to acquire a three-dimensional model of a machining part; the template matching module is configured to match a first historical model from a historical database based on the three-dimensional model, and take relevant data of the first historical model as template data; the feature recognition module is configured to recognize a machining feature of the three-dimensional model through a part recognition model; the machining requirement determination module is configured to determine machining requirements of the machining feature based on the template data; the feature parameter recognition module is configured to recognize feature parameters of the machining feature by using a CAD module; and the time determination module is configured to determine machining time of the machining part based on the machining requirements, the feature parameters of the machining feature and a machining process knowledge base.

[0018] The one or more embodiments of the specification also provide a computer program product, comprising a computer program, when at least part of the computer program is executed by a processor, the computer program can implement the above-mentioned machining time determination method. BRIEF DESCRIPTION OF DRAWINGS

[0019] The specification will be further illustrated in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. The same reference numbers in the drawings represent the same structures or steps.

[0020] Figure 1 is an application scenario diagram of a machining time determination system according to some embodiments of the specification.

[0021] Figure 2 is an exemplary flowchart of a machining time determination method according to some embodiments of the specification.

[0022] Figure 3 is a model structure diagram of a PointNet network model according to some embodiments of the specification.

[0023] Figure 4 is an exemplary flowchart of determining machining time of a machining part according to some embodiments of the specification.

[0024] Figure 5 is a functional module diagram of a machining time determination system according to some embodiments of the specification. DETAILED DESCRIPTION

[0025] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0026] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way to distinguish different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0027] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0028] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0029] In the actual operation of machinery manufacturing enterprises, one of the core pain points faced by the process planning and cost accounting departments when quoting prices for new parts or preparing for production is how to quickly and accurately estimate machining time. Traditional operating models heavily rely on the individual capabilities of process engineers. Engineers need to repeatedly review two-dimensional or three-dimensional drawings, mentally "deconstruct" each machining feature of the part (such as drilling, milling, tapping, etc.), manually select machining methods, tools, and cutting parameters for each feature, and finally consult manuals or add up the total machining time based on experience. This process not only takes several hours or even days, but also, due to differences in the experience and understanding of different engineers, the estimation results for the same part often vary greatly, leading to the failure of enterprise cost control and distorted market quotations.

[0030] Although some enterprises introduce computer-aided systems, these systems mostly only realize the electronicization of formulas or rely on simple historical data analogy, and fail to fundamentally solve the problems of process decision automation and work time calculation refinement. The rich machining information contained in the three-dimensional model is not effectively utilized, and process planning and work time calculation are still two separate links. In addition, in actual application, many parts to be machined only have a three-dimensional model without corresponding machining requirements, or the machining requirements are embodied in the form of paper or PDF version, which is difficult to be read by the system. Therefore, developing a method capable of intelligently interpreting three-dimensional models, automatically planning processes, and accurately calculating work time has become an urgent need to promote the digital transformation of manufacturing industry and achieve cost reduction and efficiency improvement.

[0031] In view of this, some embodiments of the present specification provide a machining work time determination method, system and computer program product.

[0032] Figure 1 is an application scenario diagram of a machining work time determination system according to some embodiments of the present specification. As shown in Figure 1 , the machining work time determination system 100 (hereinafter referred to as system 100) can include a server 110, a network 120, a client 130, a database 140 and a part to be machined 150.

[0033] The server 110 can be configured to process information and / or data related to machining work time calculation. In some embodiments, the server 110 is the intelligent computing hub and business logic core of the system. It is usually one or more high-performance computers deployed in the enterprise data center or cloud platform, responsible for executing the most computationally intensive algorithm steps. For example, the server 110 can load and run a trained PointNet part recognition model, process the uploaded three-dimensional model point cloud data at high speed, complete the part type confirmation and machining feature recognition; and perform intelligent matching of template data (or work time calculation template), derivation of machining process sequence, and dynamic work time calculation based on physical formula. It is embedded with query and reasoning logic of machining process knowledge base. In some embodiments, the server 110 can be regional or remote. For example, the server 110 can access information and / or data stored in the client 130 and / or the database 140 through the network 120. For another example, the server 110 can be directly connected to the client 130 and / or the database 140 to access the stored information and / or data. In some embodiments, the server 110 can be implemented on a cloud platform. For example, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a dispersed cloud, an internal cloud, a multi-cloud, or the like or any combination of the above examples.

[0034] The network 120 can facilitate the exchange of information and / or data, and is the artery that connects the components of the system, ensuring the flow of data. In some embodiments, one or more components in the system 100 (e.g., the server 110, the client 130, the database 140, and the machined part 150) can send information and / or data to other components in the system 100 through the network 120. For example, the server 110 can exchange data with the client 130 through the network 120. In some embodiments, the network 120 can be any of a wired network or a wireless network, or a combination thereof. In enterprise applications, it usually refers to an enterprise internal local area network (LAN), a virtual private network (VPN), or a secure wide area network (WAN). The network 120 is mainly responsible for transmitting three-dimensional model data, reliably transmitting large three-dimensional model files uploaded by the client 130 to the server 110; carrying interactive instructions and results, real-time transmission of parameter adjustment instructions, process schemes, calculation results, and visualization data between the interactive interface of the client 130 and the calculation engine of the server 110; ensuring data security and integrity: through encryption transmission protocols (such as HTTPS / SSL) and enterprise firewall policies, ensuring the security of core process data and model assets during transmission, preventing leakage and tampering.

[0035] The client 130 is the core operation terminal and visualization window for users (such as process engineers) to interact with the system. It usually appears as a dedicated software application installed on a workstation computer or a Web interface accessed through a browser. In some embodiments, the client 130 can include a smart mobile device, a smart tablet, a computer, a notebook computer, or the like, or any combination thereof. In some embodiments, the client 130 can provide a friendly operation interface, allowing engineers to select a three-dimensional model file of a machined part 150 from a local or integrated system and initiate a time analysis task. In some embodiments, the client 130 can receive and clearly display the automatic identification results, recommended machining processes, tools, and parameters returned by the server 110. Engineers can review, fine-tune, or confirm on this interface, realizing the organic combination of artificial intelligence and expert experience. In some embodiments, the client 130 can display the final total time, cost analysis, and detailed process route in a graphical and structured manner (such as 3D highlighted features, time decomposition tree, process card), and support one-key export of standard reports.

[0036] The database 140 can store data and / or instructions. In some embodiments, the database 140 can be a high-performance, structured data storage system, typically an enterprise-grade relational database or a time-series database. The database 140 persistently stores a historical database, a machining process knowledge base, a labor calculation template library. In some embodiments, the database 140 can include mass storage, removable storage, volatile read-write memory, read-only memory (ROM), etc., or any combination thereof. In some embodiments, the database 140 can be implemented on a cloud platform. By way of example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, inter-cloud, multi-cloud, etc., or any combination thereof. The database 140 not only provides instant data support for system operation, but also provides a data foundation for system closed-loop optimization and self-learning.

[0037] In some embodiments, the database 140 can be connected to the network 120 and communicate with one or more components in the system 100 (e.g., the server 110, the client 130, etc.). One or more components in the system 100 can access data or instructions stored in the database 140 through the network 120. In some embodiments, the database 140 can be directly connected to or communicate with one or more components in the system 100 (e.g., the server 110, the client 130, etc.). In some embodiments, the database 140 can be part of the server 110.

[0038] The part to be machined 150 is the starting point of the physical object processed by the system. In the digitization process, it first participates in system interaction in the form of its three-dimensional digitized model. In some embodiments, the model is a CAD file (such as STEP, PRT format) containing complete geometric information and material properties. It serves as the data source of the entire method flow, and its information enters the digital system through scanning, direct uploading, etc., and is then parsed by the algorithm of the server 110, and finally converted into an executable process plan and quantitative labor data. In some embodiments, the part to be machined 150 can be a machined part, or other different types of parts.

[0039] Those of ordinary skill in the art will appreciate that when an element of system 100 performs, the element can do so through signals and / or electromagnetic signals. For example, when client 130 sends a service request to server 110, a processor of client 130 can generate a signal encoding the request. The processor of client 130 can then send the signal to an output port. If client 130 communicates with server 110 through a wired network, the output port can be physically connected to a cable that further sends the signal to an input port of server 110. If client 130 communicates with server 110 through a wireless network, the output port of client 130 can be one or more antennas that convert the signal to an electromagnetic signal. Within an electronic device such as client 130 and / or server 110, when a processor processes instructions, issues instructions, and / or performs actions, the instructions and / or actions are done through signals. For example, when a processor retrieves or saves data from a storage medium, it can send a signal to a read / write device of the storage medium that can read or write structured data in the storage medium. The structured data can be transmitted to the processor through a bus of the electronic device in the form of a signal. Here, a signal can refer to one signal, a series of signals, and / or at least two discrete signals.

[0040] Figure 2 is an exemplary flowchart of a method of determining machining labor hours according to some embodiments of the present specification. In some embodiments, Figure 2 Flow 200 can be performed by server 110. In some embodiments, flow 200 can be implemented by a machining labor hours determination system 500 deployed on server 110. In some embodiments, as Figure 5 illustrated, machining labor hours determination system 500 can include a part acquisition module 510, a template matching module 520, a feature recognition module 530, a machining requirement determination module 540, a feature parameter recognition module 550, and a labor hours determination module 560.

[0041] In some embodiments, as Figure 2 illustrated, flow 200 can include the following steps.

[0042] Step 210: Acquire a three-dimensional model of a part to be machined. In some embodiments, step 210 can be implemented by part acquisition module 510 in Figure 5 .

[0043] In some embodiments, in actual industrial environments, three-dimensional model data can be stored in different systems and locations. Therefore, the present embodiment supports flexible reading of three-dimensional model files from multiple sources by constructing a unified data access interface. In some embodiments, the system can directly access a specified file server in the intranet or the local hard disk of an engineer's workstation, and read the model file through a standard file path (such as \\NAS\Project X\CAD\Models or C:\Users\Workspace). This approach is suitable for enterprises that have not yet centralized data on the cloud or for small batch and temporary analysis. In some embodiments, for enterprises that have adopted cloud storage, the system calls the API interface of a cloud service provider (such as Aliyun OSS or Tencent COS), uses a secure access key (Access Key) and token (Token) for identity verification, and downloads the target model file from a specified cloud storage bucket (Bucket). This approach facilitates centralized management and remote collaboration of data. In some embodiments, the system can also use the unique identifier of a part, such as the material code or drawing number, to initiate a data request to the PLM system through the standard RESTful API or SOAP Web Service interface provided by the PLM / PDM system (such as Teamcenter or Windchill). After permission verification, the PLM system returns the corresponding three-dimensional model file and metadata (such as material and version number). This approach ensures that the used model is the latest and valid version approved, and achieves seamless connection with the product data source.

[0044] In some embodiments, the part to be processed is a machined part. In some embodiments, the machined part can include: shaft parts such as transmission shafts, stepped shafts, and crankshafts, whose main processing features are cylindrical surfaces, threads, key grooves, and center holes; disc and sleeve parts such as flange discs, end covers, and gear blanks, whose main features are inner and outer rotary surfaces, hole systems, and end planes; box and shell parts such as machine tool spindle boxes, engine cylinder blocks, and transmission case shells, which have complex structures and features such as hole systems (coaxial, parallel, or perpendicular holes), assembly planes, and cavities; special-shaped parts and precision structural parts such as thin-walled shells, impellers in the aerospace field, and complex cavity parts of automobile molds, whose features are usually complex free-form surfaces, deep cavities, and small holes.

[0045] Step 220: matching a first historical model from the historical database based on the three-dimensional model, and taking the related data of the first historical model as template data. In some embodiments, step 220 can be implemented by the template matching module 520 in the historical database 500. Figure 5

[0046] ​In some embodiments, matching the first historical model from the historical database based on the three-dimensional model comprises: obtaining point cloud data of the three-dimensional model; and matching a model with the highest similarity from the historical database as the first historical model based on the point cloud data, the similarity being determined according to at least one of a material, a size, a shape, and a machining precision requirement of the part.

[0047] In some embodiments, the system first performs point cloud data collection on the uploaded three-dimensional model, converts model surface geometric information into a set of standard number of three-dimensional space point sets through uniform sampling, and performs normalization processing to eliminate the influence of absolute size and position. Subsequently, the system calculates the feature descriptor (such as the fast point feature histogram) of the point cloud, and takes this as the core, combines the multi-dimensional attributes such as material, key size, and precision requirement parsed from the model metadata or associated process file, and constructs a comprehensive retrieval vector. The system compares the vector with the successfully processed case models in the historical database in multiple dimensions for similarity, wherein the shape similarity can be calculated by comparing the metric distance (such as the chamfer distance) of the point cloud descriptor; the material, size, and precision similarity are quantitatively scored according to the preset rules. Finally, the system adopts a weighted fusion strategy to comprehensively evaluate the similarity of each dimension, determines the historical case model with the highest overall similarity score as the first historical model, and automatically extracts the related data (including the machining features and machining requirements of the first historical model) of the first historical model as the template data for the current part work time estimation. In some embodiments, the related data of the first historical model further includes the historical work time calculation scheme of the first historical model (including the standard preparation time, the specific process parameter mapping relationship, and the calculation coefficient, etc.). In some embodiments, the template data includes the historical machining requirements of the historical features of the first historical model. The historical machining requirements are used to determine the machining requirements of the features to be machined in step 240. In some embodiments, the template data includes a work time calculation template, and the historical work time calculation scheme of the first historical model can be used as the work time calculation template.

[0048] Step 230: identifying the features to be machined of the three-dimensional model through a part recognition model. In some embodiments, step 230 can be implemented through the feature recognition module 530 in the part recognition model 200. Figure 5

[0049] ​In some embodiments, identifying the to-be-processed feature of the three-dimensional model by the part recognition model comprises: acquiring point cloud data of the three-dimensional model; and identifying the to-be-processed feature based on the point cloud data by using the part recognition model, the part recognition model being a trained PointNet network model. In some embodiments, the system can perform point cloud data acquisition on the acquired three-dimensional model. For more information about acquiring the point cloud data of the three-dimensional model, please refer to step 220 and the related description thereof, which will not be repeated here. In some embodiments, the part recognition model can perform instance segmentation on the point cloud data of the three-dimensional model to identify the to-be-processed feature of the three-dimensional model.

[0050] Subsequently, the point cloud data can be input to a pre-trained part recognition model, which preferably adopts a PointNet network architecture, because it can directly process unordered point cloud data and effectively learn global features. Figure 3 FIG. 3 is a schematic diagram of a model structure of a PointNet network model according to some embodiments of the present specification. The PointNet network model 300 can be divided into an input layer, a hidden layer, and an output layer according to the functional hierarchy of the network structure. The hidden layer further refines a plurality of functionally explicit sub-layers, including a feature extraction layer, a max-pooling layer, and a fully connected layer. These sub-layers play an indispensable role in data processing and conversion.

[0051] (1) The input layer is the starting point of the entire network, which is responsible for receiving and processing the original three-dimensional point cloud data. These data are usually presented in the form of a fixed number of points, each point containing three-dimensional coordinate information. In order to facilitate network processing, these data are uniformly converted into an n x 3 tensor form, which provides convenience for subsequent processing. In some alternative embodiments, each point can contain three-dimensional coordinate information and normal vector information, and these data can be uniformly converted into an n x 6 tensor form.

[0052] (2) The hidden layer includes multiple sub-layers with different functions to complete complex data conversion and feature extraction. Among them, the input transformation layer plays a particularly important role. The core is a parameterized 256x9 matrix, through which a 3x3 transformation matrix can be obtained as output. After convolution operation on the input point cloud data, 1024-dimensional features can be extracted. These features are further processed by pooling and fully connected layers to obtain a 256-dimensional feature vector. This vector is multiplied by the parameterized 256x9 matrix, and then a 9-dimensional bias vector is added to finally convert it into a 3x3 transformation matrix. This matrix is multiplied by the original point cloud data to obtain the transformed point cloud data, which has the same dimension as the original data. The value of this matrix can be adjusted through learning, and its value is updated by minimizing the loss function. This transformation can map the point cloud data to a canonical space, making the model better adapt to rigid body transformation and facilitating subsequent feature extraction.

[0053] (3) The feature extraction layer is responsible for extracting more representative features from the transformed point cloud data. This layer is composed of special multi-layer neurons, which use different size convolution kernels to convolve the point cloud data. The first feature extraction layer uses 1x3 and 1x1 convolution kernels to extract features of each point through two convolution layers. The first convolution layer uses 64 1x3 convolution kernels to connect the three-dimensional coordinates of each point. The subsequent convolution layer uses a 1x1 convolution kernel to further connect the features of each point. After convolution operation, the ReLU function is used for nonlinear processing of the results to extract 64-dimensional features of each point. The feature transformation layer is similar in principle to the input transformation layer, but the size of the first convolution kernel is 1x1. This layer learns a 64x64 matrix, which is multiplied by the previously learned 64-dimensional feature vector to align the features. The purpose of this step is to normalize the extracted features to a reasonable space, making these features more effective for the model's classification task. The second feature extraction layer is similar to the first, but the size of the convolution kernel is 1x1. It gradually increases the dimension of the features through a 64x128x1024 neuron, finally extracting 1024-dimensional information for each point.

[0054] (4) The role of the max pooling layer is to solve the unordered problem of point cloud data. It processes the n x 1024-dimensional features extracted previously, and this layer takes the maximum value in each of the 1024 dimensions to obtain the 1024-dimensional global features of the point cloud model. It can ensure that the calculation result is independent of the order of the data, effectively handling the unordered nature of point cloud data.

[0055] (5) Fully connected layer: A fully connected layer with a structure of 512 x 256 x k, where k is the number of classes for classification. In some embodiments, to control the overfitting phenomenon, dropout mechanism is added in the fully connected layer with a default setting of 0.7. This means that only 70% of neurons will be activated in each training process, thereby reducing the amount of parameters and improving the generalization ability of the model.

[0056] (6) Output layer: A softmax function is used as the feature classifier. The output of the previous layer of neurons is mapped to the interval (0, 1), thereby obtaining the predicted probability of each class. This enables the model to output the likelihood of each class, providing a basis for the final classification decision.

[0057] In some embodiments, the construction and training of the part recognition model can be carried out in the following steps.

[0058] Sample library construction and division: A large number of three-dimensional model samples of parts covering machined parts, die-cast parts, sheet metal parts, etc. are collected and labeled, and their corresponding standardized point cloud data is generated according to the aforementioned method to form a labeled data set. This data set is randomly divided into training set, validation set and test set according to a predetermined ratio (such as 7:2:1).

[0059] Network training and feature learning.

[0060] Input: The point cloud data in the training set (N x 3 matrix, N is the number of points, 3 is the coordinate) is input into the PointNet network.

[0061] Feature extraction: The network first extracts features independently for each point through a shared multi-layer perceptron (MLP). Then, through a symmetric function (such as Max Pooling), the features of all points are aggregated to generate a global feature vector that can represent the shape of the entire part. This design makes the network invariant to the input order of points.

[0062] Classification output: The global feature vector is then classified through a fully connected layer, outputting the probability distribution of the part belonging to each predefined class (the primary class is "machined part").

[0063] Loss calculation and optimization: The cross-entropy loss function is used to calculate the error between the predicted class and the true label, and the network weight parameters are updated through the backpropagation algorithm (such as the Adam optimizer) to minimize the loss.

[0064] Model tuning and verification.

[0065] During the training process, the validation set is used regularly to evaluate the performance of the model and monitor its accuracy and loss changes to prevent overfitting.

[0066] The network's learning rate, batch size, MLP layer number, and other hyperparameters are repeatedly tuned and optimized using methods such as grid search or Bayesian optimization.

[0067] When the model's performance on the validation set stabilizes and the accuracy reaches the expected value (e.g., over 95%), the final evaluation is performed using the test set that did not participate in training. Only when the accuracy of the test results exceeds the pre-set reliability threshold (e.g., 98%), the model is considered to be trained and can be used in production environment.

[0068] Step 240: Based on the template data, determine the machining requirements of the to-be-machined feature. In some embodiments, step 240 can be implemented by a machining requirement determination module 540 in the feature recognition system 500. Figure 5

[0069] In some embodiments, based on the template data, determining the machining requirements of the to-be-machined feature includes: determining a target historical feature corresponding to the to-be-machined feature in the historical features of the first historical model; and determining the machining requirements of the to-be-machined feature based on the historical machining requirements of the target historical feature. In some embodiments, the machining requirements include the machining accuracy and / or surface roughness of the to-be-machined feature.

[0070] In some embodiments, after completing the part-level template matching, the system needs to further perform accurate mapping at the feature granularity. The core of this process is to determine the corresponding "target historical feature" of each "to-be-machined feature" identified on the current three-dimensional model in the selected "first historical model". This process is not a simple geometric dimension comparison, but a comprehensive judgment combining feature type, geometric properties, and contextual relationships. In some embodiments, the system first filters the feature set of the same type in the first historical model according to the identification results, and then calculates the similarity of the two in key parameters (such as for hole features, compare diameter, depth, position). For complex features, the system can also analyze the topological relationship between features (such as hole series on a certain plane) to match feature groups with similar machining sequences and clamping constraints. Through this multi-dimensional evaluation, the system determines the historical feature with the highest similarity as the "target historical feature", thereby establishing a precise correspondence from historical experience to new feature points.

[0071] ​In some embodiments, after determining the target feature history, the system intelligently extracts and adapts the machining requirements for the current feature from its associated, proven-in-practice "history machining requirements". These parameters primarily and critically include machining accuracy and surface roughness. In some embodiments, the system not only directly reads these parameter values, but also combines the specific geometric size and position of the current feature to be machined for applicability verification. For example, for a smaller diameter hole, the system may automatically tighten its tolerance recommendation value according to the rules. At the same time, the complete process context inherited from the historical feature, such as the machine tool accuracy level used, the measurement method, and even the cutting parameters, are also referred to as implicit knowledge to ensure that the recommended accuracy and roughness requirements have process accessibility. The final machining requirements will be clearly presented on the user interaction interface for the process engineer to make the final confirmation and calibration, ensuring the reliability of experience transfer.

[0072] Step 250: Identify feature parameters of the feature to be machined using the CAD module. In some embodiments, step 250 can be implemented by the feature parameter identification module 550 in Figure 5 .

[0073] In some embodiments, the system calls the integrated CAD module to perform accurate feature parameter identification. This process can be achieved by calling the secondary development API (such as NX Open, SolidWorks API) of the target CAD software (such as Siemens NX, Dassault SolidWorks) to silently start a non-interface CAD session in the background and load the original three-dimensional model. In some embodiments, the system traverses the model entity, according to the obtained feature type and spatial position information, accurately locates the accurate face and edge set constituting the specific feature (for example, a "through hole" feature corresponds to two end planes and a cylindrical surface between them) through geometric topological relationship, and performs high-precision geometric calculation and attribute reading accordingly: for hole features, directly query the diameter attribute of the cylindrical surface and calculate the axial distance between the two end faces as the depth; for thread features, analyze the decorative line or helical surface attribute to obtain the nominal diameter, pitch and thread length; for plane features, calculate the contour boundary envelope area and normal vector; for cavity or chamfer features, extract key dimensions such as depth, width, and radius by analyzing the angle between adjacent faces and the type of transition face, and finally output a set of structured and quantitative feature parameter sets for each feature to be machined, providing accurate geometric parameter input for subsequent process matching and machining time calculation.

[0074] Step 260: Determine the machining time of the part to be machined based on the machining requirements of the feature to be machined, the feature parameters, and the machining process knowledge base. In some embodiments, step 260 can be implemented by the machining time calculation module 560 in Figure 5The man-hour determination module 560 in the server 110 is implemented. In some embodiments, the relevant data includes a historical man-hour calculation scheme of the first historical model, and the template data includes a man-hour calculation template; and the machining man-hour of the part to be machined is determined based on the machining requirements of the features to be machined, the feature parameters, and the machining process knowledge base, and includes: determining the machining man-hour of the part to be machined based on the machining requirements of the features to be machined, the feature parameters, the machining process knowledge base, and the man-hour calculation template.

[0075] In some embodiments, the step 260 can be implemented by the flow 400 shown in FIG. 4. Figure 4 The flow 400 shown in FIG. 4 can be implemented by the server 110. In some embodiments, the flow 400 can be implemented by the machining man-hour determination system 500 deployed on the server 110. Further, the flow 400 can be implemented by the man-hour determination module 560 in the server 110. Figure 4 The man-hour determination module 560 in the server 110 is implemented. In some embodiments, the relevant data includes a historical man-hour calculation scheme of the first historical model, and the template data includes a man-hour calculation template; and the machining man-hour of the part to be machined is determined based on the machining requirements of the features to be machined, the feature parameters, and the machining process knowledge base, and includes: determining the machining man-hour of the part to be machined based on the machining requirements of the features to be machined, the feature parameters, the machining process knowledge base, and the man-hour calculation template. Figure 4 The man-hour determination module 560 in the server 110 is implemented. In some embodiments, the relevant data includes a historical man-hour calculation scheme of the first historical model, and the template data includes a man-hour calculation template; and the machining man-hour of the part to be machined is determined based on the machining requirements of the features to be machined, the feature parameters, and the machining process knowledge base, and includes: determining the machining man-hour of the part to be machined based on the machining requirements of the features to be machined, the feature parameters, the machining process knowledge base, and the man-hour calculation template. Figure 5 The man-hour determination module 560 in the server 110 is implemented. In some embodiments, the relevant data includes a historical man-hour calculation scheme of the first historical model, and the template data includes a man-hour calculation template; and the machining man-hour of the part to be machined is determined based on the machining requirements of the features to be machined, the feature parameters, and the machining process knowledge base, and includes: determining the machining man-hour of the part to be machined based on the machining requirements of the features to be machined, the feature parameters, the machining process knowledge base, and the man-hour calculation template.

[0076] In some embodiments, the step 260 can further include the following steps.

[0077] Step 410: Calculate the machining man-hour for each feature to be machined respectively. The step 410 can be implemented by the man-hour determination module 560 in the server 110. In some embodiments, the step 410 can further include the step 411 and the step 412. Figure 5 The man-hour determination module 560 in the server 110 is implemented. In some embodiments, the relevant data includes a historical man-hour calculation scheme of the first historical model, and the template data includes a man-hour calculation template; and the machining man-hour of the part to be machined is determined based on the machining requirements of the features to be machined, the feature parameters, and the machining process knowledge base, and includes: determining the machining man-hour of the part to be machined based on the machining requirements of the features to be machined, the feature parameters, the machining process knowledge base, and the man-hour calculation template.

[0078] Step 411: Based on the machining requirements and the feature parameters, and in combination with the machining process knowledge base, determine the machining process sequence, the recommended machining tool, and the tool machining parameter of the feature to be machined. In some embodiments, the machining tool can include, but is not limited to, one or a combination of a cutter, a drill bit, a grinding tool, a fixture, etc. The tool machining parameter can include, but is not limited to, one or a combination of a cutting speed, a feed per tooth, a spindle speed, and a cutting depth, etc. In some embodiments, for each feature to be machined for which the accurate feature parameters and machining requirements have been identified and extracted, the system takes it as a structured query request, and submits it to the machining process knowledge base for intelligent process decision. The knowledge base is an expert system that integrates process rules, a tool database, and a tool machining parameter library. The system first matches the most suitable machining process sequence in the knowledge base according to the feature type and the key parameters. Then, the system recommends the machining tool for each process in the sequence from the tool database, and calls and instantiates the specific tool machining parameter from the tool machining parameter library according to the workpiece material, the tool material, and the feature geometric size. The whole process realizes the automatic conversion from the abstract feature to the executable and quantifiable detailed process scheme.

[0079] Step 412: Calculate the preparation time and machining time of the feature to be machined according to the machining process sequence and tool machining parameters, combined with the machining process knowledge base. The preparation time includes at least tool switching time, and the machining time is calculated based on tool machining parameters and feature parameters. In some embodiments, the preparation time also includes at least one of machine idle time, workpiece clamping and alignment time.

[0080] In some embodiments, after obtaining the detailed machining process, the system starts the dynamic time calculation engine. In some embodiments, the calculation is divided into preparation time and machining time. The preparation time is directly related to the machining process sequence, and its calculation includes at least tool switching time: according to the number of tool changes in the sequence, multiply the average tool changing time of the machine tool obtained from the machining process knowledge base; in addition, the machine idle positioning time can also be calculated according to the feature space position. The calculation of machining time can be based on a physical model, and the core formula is: machining time = (machining path length / feed speed) + (machining volume / material removal rate). Wherein, the machining path length and machining volume are directly derived from the feature parameters (such as hole depth, plane area) of the feature; the feed speed and material removal rate are calculated from the tool machining parameters (such as cutting parameters) determined in the previous step.

[0081] In some embodiments, step 412 can include: calculating the preparation time and machining time of the feature to be machined according to the machining process sequence and tool machining parameters, combined with the time calculation template. Wherein, the average tool changing time of the machine tool can be obtained according to the time calculation template.

[0082] Step 420: Summarize the preparation time and machining time of all features to be machined to determine the machining time of the part to be machined. Step 420 can be implemented by the time determination module 560 in Figure 5

[0083] ​In some embodiments, after the calculation of the machining time for all the features to be machined is completed, the system performs a final machining time aggregation and integration to determine the total machining time of the part. This process is not a simple arithmetic addition, but a structured aggregation following the manufacturing logic and optimized intelligently. In some embodiments, the system first performs an internal aggregation for each feature, adding the machining time of each operation in its process sequence and the special preparation time directly related to the feature (such as the time for changing the tool for the feature alone) to obtain a subtotal of the independent time consumption of each feature; then, the system enters the global integration at the part level, the core of which is to remove the duplicate and optimize the calculation of the common preparation time - the system intelligently analyzes the machining sequence of all features, generates a globally optimal machining route through an algorithm (such as based on tool grouping and shortest path planning), thereby combining the scattered multiple tool changes and compressing the empty travel path, to calculate the optimized total tool change time and machine empty travel time; at the same time, the one-time clamping and alignment time of the workpiece based on the overall size and complexity of the part obtained from the machining time calculation template is also included. Finally, the machining time of the part to be machined is dynamically determined by the following formula: total machining time = Σ (machining time of each feature + preparation time of each feature) + globally optimized tool change time + optimized empty travel time + workpiece clamping and alignment time. After the aggregation is completed, the system automatically generates a structured machining time report, which clearly shows the total machining time and its multi-level decomposition by feature and by time type, providing accurate data basis for production scheduling and cost accounting.

[0084] In some embodiments, the process 200 further includes: displaying at least one of the machining requirements, the feature parameters, the machining process sequence, the recommended machining tool, and the tool machining parameters of the feature to be machined in the interactive interface; obtaining user feedback; and adjusting the machining requirements, the feature parameters, the machining process sequence, the recommended machining tool, and / or the tool machining parameters of the feature to be machined based on the user feedback.

[0085] In some embodiments, the system can implement human-machine collaborative calibration through a graphical user interface that pops up automatically or presents in a separate view after the background computation is completed. In some embodiments, the system presents key process data of the current part in a structured and visualized manner: for each highlighted feature to be machined in the 3D lightweight model, its feature parameters (e.g. diameter, depth), the system-automatically determined machining requirements, the intelligently planned machining process sequence, the recommended machining tools (including tool model and specifications) for each process, and the calculated tool machining parameters (e.g. rotation speed, feed rate) are all clearly listed in the corresponding panels or tables. Engineers can review the system’s automatic decisions comprehensively in this interface and provide user feedback by changing tools in the drop-down menu, correcting values in the input boxes, or adjusting parameters through sliders, etc. In some embodiments, the system captures these interactive instructions in real time, and its background logic immediately adjusts the relevant data based on the feedback: for example, when the user selects another brand of reamer with the same specifications from the tool library list, the system will update the recommended machining tool record synchronously, and may fine-tune the recommended cutting speed based on the material properties of the new tool; if the user manually increases the surface roughness requirement, the system may add a fine milling or grinding process to the machining process sequence and recalculate all related parameters. All adjustments take effect immediately and trigger partial recalculation, ensuring that the final adopted solution integrates the efficiency of artificial intelligence and the experience of human experts.

[0086] In some embodiments, the process 200 further includes adding the 3D model, the features to be machined, the machining requirements, the feature parameters, the machining process sequence, the tool machining parameters, and the machining hours of the part to be machined to a historical database and a machining process knowledge base.

[0087] In some embodiments, after the machining time of the current part to be processed is determined and the user is calibrated, the system can package and associate all key data generated in this task, including the three-dimensional model file as the source, the list of all features to be processed, the feature parameters (geometric dimensions) and processing requirements (precision requirements) corresponding to each feature, the final determined processing sequence, the selected recommended processing tool for each process and its specific tool processing parameters, and the final calculated machining time of the part to be processed, into a highly structured package with a unique task ID. This data package not only stores the original data, but also records the logical relationship between the data (such as different features corresponding to different process sequences). In some embodiments, the system stores this data package in the historical database and the machining process knowledge base through a secure interface. In the historical database, it serves as a complete traceable case to support similarity matching and template retrieval for new parts in the future. In the feature knowledge base, the new feature-process-parameter combination expands the coverage of the knowledge base as an example sample, while the actual tool machining parameters and the final work time result are used as optimization data to gradually calibrate and iteratively update the existing process rules in the knowledge base, tool recommendation priority, and cutting parameter recommendation value through a background statistical analysis model, so as to drive the system to make more accurate and more production capacity-oriented decisions in the next task, realizing a closed loop from single application to continuous self-enhancement.

[0088] In some embodiments, the process 200 further includes: when the part recognition model fails to recognize the features to be processed of the three-dimensional model, determining the volume of the part to be processed; determining the machining time of the part to be processed based on the volume of the part to be processed, the volume of the first historical model, and the work time calculation template.

[0089] In some embodiments, when the system processes the point cloud data of the three-dimensional model through the pre-trained PointNet part recognition model, if the confidence of the model output is lower than a preset threshold or the specific geometric region cannot be classified into any known machining feature category, it is determined that the feature cannot be recognized. At this time, the system will automatically start the backup labor time estimation process to ensure the robustness of the function. First, the system accurately calculates the volume of the entire part to be machined (V_new) through the CAD kernel interface, that is, the total volume of the space occupied by all entities of its three-dimensional model. Then, the system compares V_new with the volume of the first historical model (V_history) matched successfully in step 220 and calculates the volume ratio K = V_new / V_history. Finally, the system scales the labor time calculation template inherited from the first historical model based on the proportion coefficient K. Specifically, the total labor time is composed of the machining labor time which is strongly related to the volume and the relatively fixed preparation labor time: the machining labor time part will be scaled according to the material removal amount and the machining difficulty, according to the proportion K (or combined with the correction coefficient of shape complexity); and the preparation labor time (such as standard clamping time) may be adaptively adjusted according to the size interval to which the volume belongs, or the reference value in the template is retained. Through this analogy method based on overall geometric similarity (volume), the system can still provide a reasonable and traceable preliminary labor time estimation when facing new or special features, ensuring the completeness and practicality of the method in practice.

[0090] Some embodiments of the present specification also provide a machining labor time determination system. Figure 5 is a functional module schematic diagram of a machining labor time determination system according to some embodiments of the present specification. In some embodiments, as shown in Figure 5 the machining labor time determination system 500 can include a part acquisition module 510, a template matching module 520, a feature recognition module 530, a machining requirement determination module 540, a feature parameter recognition module 550, and a labor time determination module 560. In some embodiments, the various modules of the machining labor time determination system 500 can be configured in the server 110.

[0091] In some embodiments, the part acquisition module 510 is configured to acquire a three-dimensional model of a part to be machined, the part to be machined being a machining part.

[0092] In some embodiments, the template matching module 520 is configured to match a first historical model from a historical database based on the three-dimensional model, and use the related data of the first historical model as template data. In some embodiments, the template matching module 520 can also acquire point cloud data of the three-dimensional model; match a model with the highest similarity from the historical database as the first historical model based on the point cloud data, the similarity being determined according to at least one of the material, size, shape, and machining precision requirement of the part.

[0093] In some embodiments, the feature recognition module 530 is configured to identify the to-be-processed feature of the three-dimensional model by using the part recognition model. In some embodiments, the feature recognition module 530 can further acquire point cloud data of the three-dimensional model; and identify the to-be-processed feature by using the part recognition model based on the point cloud data, wherein the part recognition model is a trained PointNet network model.

[0094] In some embodiments, the machining requirement determination module 540 is configured to determine the machining requirement of the to-be-processed feature based on the template data. In some embodiments, the machining requirement determination module 540 can further determine a target historical feature corresponding to the to-be-processed feature in the historical features of the first historical model; and determine the machining requirement of the to-be-processed feature based on the historical machining requirement of the target historical feature. The machining requirement includes machining precision and / or surface roughness of the to-be-processed feature.

[0095] In some embodiments, the feature parameter recognition module 550 is configured to identify the feature parameter of the to-be-processed feature by using the CAD module.

[0096] In some embodiments, the machining time determination module 560 is configured to determine the machining time of the to-be-processed part based on the machining requirement, the feature parameter, and the machining process knowledge base of the to-be-processed feature. In some embodiments, the machining time determination module 560 can further calculate the machining time of each to-be-processed feature, specifically including: determining the machining process sequence, the recommended machining tool, and the tool machining parameter of the to-be-processed feature based on the machining requirement and the feature parameter, and in combination with the machining process knowledge base; and calculating the preparation time and the machining time of the to-be-processed feature according to the machining process sequence and the tool machining parameter, and in combination with the machining process knowledge base; wherein the preparation time at least includes the machining tool switching time, and the machining time is calculated based on the tool machining parameter and the feature parameter; and the machining time determination module 560 finally aggregates the preparation time and the machining time of all to-be-processed features to determine the machining time of the to-be-processed part.

[0097] For more information about the modules, please refer to the related description of Figures 2 to 4 , which will not be repeated here. It should be understood that Figure 5The illustrated system and its modules can be implemented in various ways. For instance, in some embodiments, the system and its modules can be implemented in hardware, software, or a combination of software and hardware. The hardware portion can be implemented with special logic, while the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art can understand that the above-mentioned methods and systems can be implemented using computer-executable instructions and / or control codes contained in a processor, such as a carrier medium, such as a disk, CD or DVD-ROM, memory of a programmable device, etc. The system and its modules of the present specification can not only be implemented in hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors, such as logic chips, transistors, etc., or programmable hardware devices, such as field programmable gate arrays, programmable logic devices, etc., but also in software, for example, by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuitry and software (for example, firmware).

[0098] Some embodiments of the present specification also provide a computer program product. The computer program product stores a computer program, and when at least a part of the computer program is executed by a processor, the computer program can implement the method for determining machining time as described in any embodiment of the present specification. If each component module of the above-mentioned system is implemented in the form of a software function unit and sold or used as an independent product, it can be stored in the computer program product.

[0099] The beneficial effects that the embodiments of the present specification can bring include but are not limited to: (1) high automation and intelligence: the AI model automatically identifies the part processing features, and intelligently matches the processing technology by using the knowledge base, realizing end-to-end automatic calculation from the three-dimensional model to the work result, greatly reducing manual intervention and improving efficiency. (2) high estimation accuracy and good consistency: the work hour calculation is decomposed to each specific processing feature, and dynamic calculation is carried out based on scientific calculation formula and real process parameters, avoiding the subjectivity and randomness of experience estimation, so that the estimation result is closer to the actual situation, and has high consistency among different personnel or different batches. (3) make full use of three-dimensional model information: through the CAD secondary development interface, the feature parameters of the three-dimensional model are directly read, ensuring the integrity and accuracy of information extraction, overcoming the shortcomings of incomplete and error-prone information in traditional two-dimensional drawings. (4) deep integration of process and calculation: feature recognition, process planning and work hour calculation are closely coupled to ensure that work hour calculation is based on a feasible processing scheme, fundamentally solving the problem of disconnection between process and cost accounting. (5) self-learning and continuous optimization capability: by feeding the result data of each calculation to the historical database and knowledge base, the system can continuously accumulate experience, and optimize the calculation template and knowledge base through subsequent data analysis or machine learning, so that the system becomes smarter and more accurate, forming a benign self-evolution closed loop. It should be noted that different embodiments can have different beneficial effects, and in different embodiments, the beneficial effects that can be produced can be any one or a combination of the above, or any other beneficial effects that can be obtained.

[0100] Unless the claim specifies otherwise, the order of processing elements and sequences, the use of numerical letters, or the use of other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments are discussed in the above disclosure through various examples, it should be understood that such details are only for the purpose of illustration, and the additional claims are not limited to the disclosed embodiments, on the contrary, the claims are intended to cover all modifications and equivalent combinations that meet the spirit and scope of the embodiments. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on existing servers or mobile devices.

[0101] Similarly, it should be noted that, in order to simplify the description of the present disclosure and to help understand one or more embodiments of the specification, sometimes multiple features are combined into one embodiment, figure or description thereof in the foregoing description of the embodiments of the present application. However, this disclosure method does not mean that the features required by the present application are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the above disclosed single embodiment.

[0102] Each of the patents, patent applications, patent application publications, and other materials, such as articles, books, specifications, publications, documents, and the like, cited

[0103] Finally, it should be understood that the embodiments described herein are merely exemplary of the application. Other variations of the embodiments can also be possible and are within the scope of the application. Thus, for example, an alternative configuration of the embodiments of the application can be considered as being within the teachings of the application. Accordingly, the embodiments of the application are not to be restricted, except in light of the attached claims and their equivalents.

Claims

1. A method for determining machining time, characterized in that, include: Obtain a three-dimensional model of the part to be processed, wherein the part to be processed is a machined part; Based on the three-dimensional model, a first historical model is matched from the historical database, and the relevant data of the first historical model is used as template data. The parts identification model identifies the features to be processed in the 3D model; Based on the template data, the processing requirements of the feature to be processed are determined; The feature parameters of the feature to be processed are identified using the CAD module; Based on the processing requirements, feature parameters, and machining process knowledge base of the features to be processed, the machining time of the part to be processed is determined.

2. The method for determining machining time according to claim 1, characterized in that, The step of matching the first historical model from the historical database based on the three-dimensional model includes: Obtain the point cloud data of the three-dimensional model; Based on the point cloud data, the model with the highest similarity is matched from the historical database as the first historical model. The similarity is determined according to at least one of the material, size, shape and machining accuracy requirements of the part.

3. The method for determining machining time according to claim 1, characterized in that, The process of identifying the features to be processed in the 3D model through the part identification model includes: Obtain the point cloud data of the three-dimensional model; Based on the point cloud data, the part recognition model is used to identify the features to be processed. The part recognition model is a trained PointNet network model.

4. The method for determining machining time according to claim 1, characterized in that, The template data includes the historical processing requirements of the historical features of the first historical model; The process of determining the processing requirements of the feature to be processed based on the template data includes: Determine the target historical feature in the historical features of the first historical model that corresponds to the feature to be processed; Based on the historical processing requirements of the target historical features, the processing requirements of the features to be processed are determined.

5. The method for determining machining time according to claim 4, characterized in that, The processing requirements include the processing accuracy and / or surface roughness of the feature to be processed.

6. The method for determining machining time according to claim 1, characterized in that, The determination of machining time for the part to be processed based on the processing requirements, feature parameters, and machining process knowledge base of the features to be processed includes: The processing time is calculated separately for each feature to be processed, specifically including: Based on the processing requirements and the characteristic parameters, and in conjunction with the machining process knowledge base, the processing process sequence, recommended processing tools, and tool processing parameters for the feature to be processed are determined; and Based on the processing sequence and the tool processing parameters, and in conjunction with the machining process knowledge base, the preparation time and processing time of the feature to be processed are calculated; wherein, the preparation time includes at least the tool switching time, and the processing time is calculated based on the tool processing parameters and feature parameters; Summarize the preparation time and processing time of all features to be processed, and determine the machining time of the part to be processed.

7. The method for determining machining time according to claim 6, characterized in that, The preparation time also includes at least one of the machine tool idle travel time and workpiece clamping and alignment time.

8. The method for determining machining time according to claim 6, characterized in that, Also includes: The interactive interface displays at least one of the following: processing requirements, feature parameters, processing sequence, recommended processing tools, and tool processing parameters for the feature to be processed. Obtain user feedback; Based on the user feedback, adjust the processing requirements, feature parameters, processing sequence, recommended processing tools and / or tool processing parameters of the feature to be processed.

9. The method for determining machining time according to claim 6, characterized in that, Also includes: The three-dimensional model, features to be processed, processing requirements, feature parameters, processing process sequence, tool processing parameters, and machining time of the parts to be processed are added to the historical database and the machining process knowledge base.

10. The method for determining machining time according to claim 1, characterized in that, The relevant data includes the historical working hour calculation scheme of the first historical model, and the template data includes a working hour calculation template; The determination of machining time for the part to be processed based on the processing requirements, feature parameters, and machining process knowledge base of the features to be processed includes: Based on the processing requirements, feature parameters, machining process knowledge base, and time calculation template of the feature to be processed, the machining time of the part to be processed is determined.

11. The method for determining machining time according to claim 1, characterized in that, The relevant data includes the historical working time calculation scheme of the first historical model, and the template data includes a working time calculation template; the method for determining machining working time further includes: When the part recognition model cannot recognize the features to be processed in the 3D model, the volume of the part to be processed is determined; Based on the volume of the part to be processed, the volume of the first historical model, and the time calculation template, the machining time of the part to be processed is determined.

12. A system for determining machining time, characterized in that it includes a part acquisition module, a template matching module, a feature recognition module, a machining requirement determination module, a feature parameter recognition module, and a time determination module; The part acquisition module is used to acquire a three-dimensional model of the part to be processed, which is a machined part; The template matching module is used to match a first historical model from the historical database based on the three-dimensional model, and use the relevant data of the first historical model as template data; The feature recognition module is used to identify the features to be processed in the three-dimensional model through the part recognition model; The processing requirement determination module is used to determine the processing requirements of the feature to be processed based on the template data; The feature parameter recognition module is used to identify the feature parameters of the feature to be processed using the CAD module; The time determination module is used to determine the machining time of the part to be processed based on the processing requirements, feature parameters, and machining process knowledge base of the feature to be processed.

13. A computer program product, characterized in that, The method includes a computer program that, when at least a portion of the computer program is executed by a processor, enables the determination of machining time as described in any one of claims 1 to 11.

Images