Numerical control machining simulation and process monitoring method based on digital twinning
Through digital twin technology, a three-dimensional digital twin model and deep learning model is constructed for CNC machining simulation and monitoring, solving the problems of low efficiency and unstable quality in traditional CNC machining, achieving an efficient and stable processing process and rapid response to market demand.
Patent Information
- Application Number
- CN202410983171.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
There are problems of low production efficiency, unstable product quality and frequent equipment failures in traditional CNC processing. It is necessary to introduce highly intelligent and refined technical means to improve the transparency and controllability of the processing process.
Using CNC machining simulation and process monitoring methods based on digital twins, the three-dimensional digital twin model with high fidelity is constructed, pre-process simulation verification and real-time monitoring are performed, and data analysis and prediction are combined with deep learning models to achieve real-time monitoring of the machining process. and abnormal warning.
It improves the consistency of production efficiency and product quality, reduces trial and error costs and equipment failures, extends the service life of the equipment, and enables manufacturing companies to respond quickly to market changes and shortens product development cycles.
Smart Images

Figure CN118940337A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of numerical control machining simulation and process monitoring, and in particular to a numerical control machining simulation and process monitoring method based on digital twins. Background Art
[0002] With the continuous development of the manufacturing industry, traditional processing processes are facing increasingly complex challenges, including low production efficiency, unstable product quality, frequent equipment failures and other problems. In order to meet these challenges, the manufacturing industry urgently needs to introduce advanced technical means to improve the intelligence and refinement of the processing process. As a representative of the new generation of information technology, digital twin technology is gradually becoming a key way to solve the above problems.
[0003] Digital twin technology creates virtual models of physical entities to achieve seamless integration of the physical world and the digital world, making it possible to simulate, monitor, optimize and predict the entire life cycle of physical entities. During the processing, the application of digital twin technology can effectively improve the transparency and controllability of the production process, and timely discover and solve potential problems through real-time monitoring and simulation analysis, thereby improving production efficiency and product quality. The processing process simulation of digital twins can evaluate the impact of different processing parameters on production results in advance through modeling and simulation in a virtual environment, optimize the processing path, and reduce trial and error costs. Process monitoring collects various data in the physical processing process in real time and compares them with the virtual model to achieve real-time monitoring of the processing status and abnormal warning, effectively preventing sudden failures and quality problems in the processing process.
[0004] Therefore, in-depth research on digital twin processing simulation and monitoring not only has important theoretical significance, but also has practical application value for promoting intelligent manufacturing and improving the competitiveness of the manufacturing industry. Summary of the invention
[0005] The purpose of the present invention is to provide a method for simulation verification before CNC machining and virtual-real monitoring during CNC machining, so as to ensure the quality and efficiency of CNC machining. The contents of the invention are as follows:
[0006] A method for numerical control machining simulation and process monitoring based on digital twins, comprising the following steps:
[0007] S10, Twin data modeling: By building a high-fidelity 3D digital twin model of CNC machine tools, fixtures, tools, and workpieces, the motion information and state information of the machining process can be dynamically reflected;
[0008] S20, three-dimensional cutting simulation before processing; based on the three-dimensional digital twin model constructed in step S10, the G code program and workpiece blank required for processing are imported, so as to simulate and verify the G code program before processing;
[0009] S30, data collection during machining process: by communicating with the CNC system, the motion information and switch quantity information of each motion axis of the CNC machine tool are obtained in real time, and the real-time driving of the three-dimensional digital twin model is realized; by installing an acceleration sensor on the spindle of the CNC machine tool, the vibration information of the spindle is collected in real time, thereby indirectly monitoring the tool, machining quality, chatter and other machining process status information;
[0010] S40, data processing and analysis: Based on the sensor data collected in step S30, data cleaning, filtering, feature extraction, and feature fusion are performed to build a state monitoring and prediction model based on deep learning, and analyze and predict problems such as tools, chatter, and processing quality during the processing;
[0011] S50, three-dimensional visualization display: according to the state information about the processing process obtained in step S40, the predicted results are reflected on the three-dimensional digital twin model to provide intuitive result display for the process and operators;
[0012] S60, abnormal warning and decision control; based on the prediction results driven by twin data, use sound, color and other methods to warn and alarm abnormal conditions such as collision and chipping; and based on the prediction results, provide data analysis reports and decision support tools to help process and operating personnel adjust the processing parameters and improve processing efficiency and quality.
[0013] Furthermore, in step S10, the twin data modeling first uses 3D CAD software such as SolidWorks to build a 3D geometric model, and then uses 3D MAX to optimize the model and import it into Unity 3D.
[0014] Furthermore, the three-dimensional digital twin model after modeling needs to map motion characteristics and behavioral characteristics. The motion characteristics include the position and speed of operations such as translation and rotation of each motion axis; the behavioral characteristics include switching behaviors such as CNC machine tool lighting, coolant, and safety doors.
[0015] Furthermore, the three-dimensional cutting simulation in step S20 includes the removal of blank material and collision detection. The blank material removal adopts mesh Boolean operation based on dynamic octree; and the collision detection adopts AABB bounding box algorithm.
[0016] Furthermore, in step S30, the communication with the numerical control system is carried out in the form of OPC-UA, and the acquired data includes parameters such as the position and speed of each motion axis, the rotation speed of the spindle, and the feed speed.
[0017] Furthermore, in step S40, the data is filtered using a comb filter to eliminate irrelevant frequency signals and their harmonic components; feature extraction includes extracting time domain, frequency domain, and time-frequency domain features, wherein the time domain and frequency domain features are statistical features, and the time-frequency domain features are features obtained based on wavelet packet decomposition.
[0018] Furthermore, in step S40, the state monitoring and prediction model is a deep learning model based on GAN-CNN, the GAN model is used to generate training data in cases of missing samples and small samples, and the CNN model is used for fault diagnosis and state prediction.
[0019] Furthermore, the three-dimensional visualization display and abnormal warning in steps S50 and S60 include the display of the PC, mobile terminal, and enhanced implementation equipment, which monitors the state changes and abnormal information in the processing process in real time and displays them with high fidelity through vivid colors and sounds.
[0020] Furthermore, the displayed results include three-dimensional geometric information during the processing, statistical information such as time and energy consumption during the processing, abnormal status and alarm information during the processing.
[0021] The present invention has the following beneficial effects:
[0022] (1) The method of the present invention can optimize the processing parameters and processing paths in advance through detailed simulation of the processing process, reduce trial and error time and material waste, and improve overall production efficiency.
[0023] (2) The method of the present invention monitors the key status parameters of the processing process in real time, promptly discovers and corrects problems in the processing process, ensures the stability of the processing process, and significantly improves the consistency and quality of the products; at the same time, real-time data and historical data can be used for analysis and modeling, which can accurately predict possible equipment failures and perform maintenance in advance, thereby reducing unplanned equipment downtime and extending the service life of the equipment.
[0024] (3) The digital twin-based technology in the method of the present invention enables manufacturing companies to respond to market changes more quickly, verify and adjust new products and new processes through virtual simulation, and shorten the product development cycle.
[0025] (4) The method of the present invention provides data-based decision support by collecting and analyzing a large amount of data during the processing in real time, helping process and operation personnel to formulate production plans and optimize resource allocation more scientifically. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 This is a flow chart of a CNC machining simulation and process monitoring method based on digital twin proposed in the present invention.
[0027] Figure 2 Schematic diagram of three-dimensional digital twin model construction and processing simulation.
[0028] Figure 3 This is the processing simulation process and result diagram.
[0029] Figure 4 This is a flow chart of data processing and analysis. DETAILED DESCRIPTION
[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0031] Reference Figure 1 , an embodiment provided by the present invention: a numerical control machining simulation and process monitoring method based on digital twin, comprising the following steps:
[0032] Step S10, twin data modeling: by constructing a high-fidelity three-dimensional digital twin model of the CNC machine tool, fixture, tool, and workpiece, the motion information and state information of the machining process are dynamically reflected. The constructed three-dimensional digital twin model of CNC machining is as follows: Figure 2 As shown;
[0033] Specifically, in step S10, the twin data modeling first uses 3D CAD software such as SolidWorks to build a 3D geometric model, then uses 3D MAX to optimize the model and import it into Unity 3D. After modeling, the 3D digital twin model needs to map motion characteristics and behavior characteristics. The motion characteristics include the position and speed of operations such as translation and rotation of each motion axis; the behavior characteristics include the switch quantity behaviors such as CNC machine tool lighting, coolant, and safety door.
[0034] S20, 3D cutting simulation before processing; based on the 3D digital twin model constructed in step S10, the G code program and workpiece blank required for processing are imported to simulate and verify the G code program before processing. The processing simulation process and results of a five-axis aviation impeller are as follows: Figure 3 As shown;
[0035] Specifically, the three-dimensional cutting simulation in step S20 includes the removal of blank material and collision detection. The blank material removal adopts mesh Boolean operation based on dynamic octree; and the collision detection adopts AABB bounding box algorithm.
[0036] S30, data collection during machining process: by communicating with the CNC system, the motion information and switch quantity information of each motion axis of the CNC machine tool are obtained in real time, and the real-time driving of the three-dimensional digital twin model is realized; by installing an acceleration sensor on the spindle of the CNC machine tool, the vibration information of the spindle is collected in real time, thereby indirectly monitoring the tool, machining quality, chatter and other machining process status information;
[0037] Specifically, in step S30 , the communication with the numerical control system is carried out in the form of OPC-UA, and the acquired data includes parameters such as the position and speed of each motion axis, the rotation speed of the spindle, and the feed speed.
[0038] S40, data processing and analysis: Based on the sensor data collected in step S30, data cleaning, filtering, feature extraction, and feature fusion are performed to build a state monitoring and prediction model based on deep learning, and analyze and predict problems such as tools, chatter, and processing quality during the processing;
[0039] Specifically, the data processing and analysis process in step S40 is shown in Figure 4, and includes the following steps:
[0040] Step S41: data cleaning and filtering, using a comb filter to eliminate irrelevant frequency signals and their harmonic components;
[0041] Step S42: feature extraction, using statistical methods to extract time domain and frequency domain features, and using wavelet packet decomposition to obtain time-frequency domain features;
[0042] Step S43: feature dimensionality reduction and fusion, using an autoencoder to reduce the dimensionality of features, and selecting the most representative features for fusion to reduce redundant features and interference;
[0043] Step S44: Sample generation, using the GAN algorithm to generate training data in the case of missing samples and small samples;
[0044] Step S45: Model training, inputting the original sample data and the samples generated by GAN into the CNN model for training;
[0045] Step S46: state identification and prediction, inputting real-time sample data into the trained CNN model, and inputting real-time prediction results for tool wear state, vibration state, and processing quality.
[0046] S50, three-dimensional visualization display: according to the state information about the processing process obtained in step S40, the predicted results are reflected on the three-dimensional digital twin model to provide intuitive result display for the process and operators;
[0047] S60, abnormal warning and decision control; based on the prediction results driven by twin data, use sound, color and other methods to warn and alarm abnormal conditions such as collision and chipping; and based on the prediction results, provide data analysis reports and decision support tools to help process and operating personnel adjust the processing parameters and improve processing efficiency and quality.
[0048] Specifically, the three-dimensional visualization display and abnormal warning in steps S50 and S60 include the display of the PC terminal, mobile terminal, and enhanced realization equipment, and real-time monitoring and high-fidelity display of state changes and abnormal information in the processing process through vivid colors and sounds. The displayed results include three-dimensional geometric information in the processing process, statistical information such as time and energy consumption in the processing process, abnormal status and alarm information in the processing process.
Claims
1. A CNC machining simulation and process monitoring method based on digital twins, It is characterized in that The following steps are involved: S10, twin data modeling; By building a high-fidelity 3D digital twin model of CNC machine tools, fixtures, tools, and workpieces, the motion and status information of the machining process can be dynamically reflected; S20, three-dimensional cutting simulation before processing; based on the three-dimensional digital twin model constructed in step S10, the G code program and workpiece blank required for processing are imported, so as to simulate and verify the G code program before processing; S30, data collection during machining process: by communicating with the CNC system, the motion information and switch quantity information of each motion axis of the CNC machine tool are obtained in real time, and the real-time driving of the three-dimensional digital twin model is realized; by installing an acceleration sensor on the spindle of the CNC machine tool, the vibration information of the spindle is collected in real time, thereby indirectly monitoring the tool, machining quality, chatter and other machining process status information; S40, data processing and analysis: Based on the sensor data collected in step S30, data cleaning, filtering, feature extraction, and feature fusion are performed to build a state monitoring and prediction model based on deep learning, and analyze and predict problems such as tools, chatter, and processing quality during the processing; S50, three-dimensional visualization display: according to the state information about the processing process obtained in step S40, the predicted results are reflected on the three-dimensional digital twin model to provide intuitive result display for the process and operators; S60, abnormal warning and decision control; Based on the prediction results driven by twin data, early warnings and alarms are issued by means of sound, color, etc. for abnormal conditions such as collision and chipping. Based on the prediction results, data analysis reports and decision support tools are provided to help process and operating personnel adjust the processing parameters of the machining process and improve machining efficiency and quality.
2. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that Step S10: Twin data modeling first uses 3D CAD software such as SolidWorks to build a 3D geometric model, then uses 3D MAX to optimize the model and import it into Unity 3D.
3. Twin data modeling according to claim 2, It is characterized in that After modeling, the 3D digital twin model needs to map motion characteristics and behavior characteristics. The motion characteristics include the position and speed of operations such as translation and rotation of each motion axis. Behavioral characteristics include switch behaviors such as CNC machine lighting, coolant, and safety doors.
4. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that Step S20 3D cutting simulation includes blank material removal and collision detection. Blank material removal adopts mesh Boolean operation based on dynamic octree; collision detection adopts AABB bounding box algorithm.
5. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that In step S30, the communication with the numerical control system is carried out in the form of OPC-UA, and the acquired data includes parameters such as the position and speed of each motion axis, the rotation speed of the spindle, and the feed speed.
6. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that In step S40, the data is filtered using a comb filter to eliminate irrelevant frequency signals and their harmonic components; feature extraction includes extracting time domain, frequency domain, and time-frequency domain features, wherein the time domain and frequency domain features are statistical features, and the time-frequency domain features are features obtained based on wavelet packet decomposition.
7. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that The state monitoring and prediction model in step S40 is a deep learning model based on a generative adversarial network (GAN) and a convolutional neural network (CNN). The GAN model is used to generate training data in cases of missing samples and small samples, and the CNN model is used for fault diagnosis and state prediction.
8. The numerical control machining simulation and process monitoring method according to claim 1, It is characterized in that The three-dimensional visualization display and abnormal warning in steps S50 and S60 include the display of the PC, mobile terminal, and enhanced implementation equipment, which monitors the state changes and abnormal information of the processing process in real time and displays them with high fidelity through vivid colors and sounds.
9. The three-dimensional visual display and abnormal warning according to claim 8, It is characterized in that The displayed results include 3D geometric information of the machining process, statistical information such as time and energy consumption during the machining process, abnormal status and alarm information during the machining process.
Citation Information
Cited By
Workpiece machining digital twin system and method based on machining path simulation
CN122613894A
Workpiece machining digital twin system and method based on machining path simulation
CN122613894B
