A real-time data-driven digital twin model building method

Abstract

This invention provides a method for building a digital twin model based on real-time data, relating to the field of intelligent manufacturing technology. By integrating sensor technology, this invention collects dynamic data of machine tools and their components in real time, including key parameters such as motor torque, moment of inertia, speed, spindle acceleration, and tool wear and temperature changes. By mapping this real-time data into a digital twin model, a virtual model highly consistent with the actual machine tool can be constructed, enabling real-time monitoring and accurate prediction of the machine tool's status. Regarding tool wear assessment, it can monitor the tool's usage status and temperature changes in real time, and, combined with expert knowledge and calibration coefficients, construct a comprehensive wear assessment model, achieving accurate prediction of the tool's wear state throughout its entire lifecycle. This improves the real-time performance and accuracy of the digital twin model, enhancing the service performance of the digital twin system.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, specifically to a method for building a digital twin model based on real-time data. Background Technology

[0002] For CNC machine tool digital twin systems, the digital twin model is based on the mapping of dynamic and static data of physical entities. In terms of spatial scale, digital twin models can be divided into city-level digital twin models, workshop-level digital twin models, unit-level digital twin models, and component-level digital twin models, from large to small. Twin data is the core driving force of digital twins. It consists of dynamic and static data at the machine tool physical entity level, update and iteration data of component-level twin models, service data based on human-machine interaction interfaces, and a collection of expert knowledge used for prediction and optimization.

[0003] In the prior art, CN115509178A discloses a digital twin-driven tool wear monitoring method and CNC machine tool equipment. The method includes: constructing a digital space for the CNC machine tool milling process and performing real-time milling simulation; establishing a connection between the physical space and the digital space to acquire multi-source heterogeneous data from the physical space; preprocessing historical machining data collected by sensors; constructing a deep learning model based on convolutional neural network-long short-term memory, importing the preprocessed historical machining data for model training, and obtaining an optimized tool wear monitoring model; inputting real-time machining data from sensors into the tool wear monitoring model to obtain the monitored value of tool wear and achieve three-dimensional visualization.

[0004] While existing digital twin technologies have made some progress in machine tool monitoring and optimization, several shortcomings remain. First, most models rely on preset static parameters, making it difficult to accurately reflect the dynamic changes of machine tools during actual operation. This is particularly true in tool wear assessment, where the lack of real-time data support leads to insufficient accuracy in wear prediction. Second, existing technologies often lack sufficient flexibility and adaptability when dealing with complex and changing machining environments and tool conditions, failing to effectively address various unforeseen circumstances in actual production. Furthermore, for tool wear assessment, existing technologies often employ single evaluation indicators, failing to fully consider the interaction of multiple factors affecting wear, resulting in biased and inaccurate assessment results. Finally, due to inconsistent data storage, usage, and classification, incompatible data formats, and difficulties in data fusion, a twin data flow interaction logic and mapping process is urgently needed to improve the service performance of digital twin systems to a certain extent.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a method for building a digital twin model based on real-time data to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for building a digital twin model based on real-time data-driven methods includes the following steps:

[0009] Step S1: Collect dynamic data of components at the physical entity level of the machine tool. This component dynamic data includes the torque, moment of inertia, and speed of the motors of each feed axis of the CNC machine tool, and the acceleration of the machine tool spindle in the three coordinate directions; it also includes dynamic physical influence data of the workpiece on tool wear, as well as tool usage status and temperature change data.

[0010] Step S2: Acquire the collected physical entity data and map it into the digital twin model to construct a digital twin model corresponding to the actual machine tool;

[0011] Step S3: Based on the digital twin model, construct a corresponding data sub-library for each component, with each sub-library serving as a data driver and storage unit;

[0012] Step S4: Construct an analysis library based on the digital twin system. This analysis library is designed to receive and process data from the data sub-libraries of each component, and to perform advanced data analysis and motion state assessment.

[0013] Step S5: Acquire the real-time dynamic data of each component and input it into the analysis library in step S4. The analysis library processes the received data and analyzes the motion state of each component.

[0014] Step S6: Display key information such as the cutting process, tool temperature, machining data parameters, and spindle load using visualization tools;

[0015] Step S7: Obtain dynamic physical influence data of the workpiece on tool wear, as well as tool usage status and temperature change data, and analyze and process them to generate a first wear evaluation index, which is used to make the first assessment of tool wear.

[0016] Step S8: Using twin data and expert knowledge, introduce a calibration coefficient for wear quantification, combine the calibration coefficient with the first wear evaluation index, and perform analysis and processing to construct a comprehensive wear assessment model. This model can combine the calibration coefficient to comprehensively adjust and calculate the first assessment result of tool wear, so as to predict the full-cycle wear state of the tool.

[0017] Step S9: Obtain the prediction results of tool wear from the comprehensive wear assessment model, and adjust the machine tool's feed rate and spindle rotation rate operating parameters to avoid excessive tool wear or damage.

[0018] Compared with the prior art, the beneficial effects of the present invention are: by integrating sensor technology, dynamic data of the machine tool and its components can be collected in real time, including key parameters such as motor torque, moment of inertia, speed, spindle acceleration, tool wear, and temperature change; by mapping these real-time data into a digital twin model, a virtual model that is highly consistent with the actual machine tool can be constructed, realizing real-time monitoring and accurate prediction of the machine tool status;

[0019] In terms of tool wear assessment, the system can monitor the tool's usage status and temperature changes in real time. By combining expert knowledge and calibration coefficients, a comprehensive wear assessment model can be constructed to accurately predict the tool's wear status throughout its entire lifecycle. This innovative method significantly improves the real-time performance and accuracy of the digital twin model, thereby enhancing the service performance of the digital twin system. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0021] Figure 2 This is a schematic diagram of the component-level digital twin model of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0023] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0024] Example 1:

[0025] Please see Figure 1 and Figure 2 The present invention provides a technical solution:

[0026] A method for building a digital twin model based on real-time data-driven methods includes the following steps:

[0027] Step S1: Collect dynamic data of the physical components of the machine tool through sensors. The dynamic data of the components includes the torque, moment of inertia and speed of the motors of each feed axis of the CNC machine tool, the acceleration of the machine tool spindle in the three coordinate directions; it also includes the dynamic physical influence data of the workpiece on tool wear, as well as the tool's usage status and temperature change data.

[0028] Step S2: Acquire the collected physical entity data and map it into the digital twin model to construct a digital twin model corresponding to the actual machine tool;

[0029] Step S3: Based on the digital twin model, construct a corresponding data sub-library for each component, with each sub-library serving as a data driver and storage unit;

[0030] Step S4: Construct an analysis library based on the digital twin system. This analysis library is designed to receive and process data from the data sub-libraries of each component, and to perform advanced data analysis and motion state assessment.

[0031] Step S5: Acquire the real-time dynamic data of each component and input it into the analysis library in step S4. The analysis library processes the received data and analyzes the motion state of each component.

[0032] Step S6: Display key information such as the cutting process, tool temperature, machining data parameters, and spindle load through visualization tools; especially tool temperature visualization, which uses a heat map to represent the overall temperature change of the tool through direct monitoring or prediction methods.

[0033] Step S7: Obtain dynamic physical influence data of the workpiece on tool wear, as well as tool usage status and temperature change data, and analyze and process them to generate a first wear evaluation index, which is used to make the first assessment of tool wear.

[0034] Step S8: Using twin data and expert knowledge, introduce a calibration coefficient for wear quantification, combine the calibration coefficient with the first wear evaluation index, and perform analysis and processing to construct a comprehensive wear assessment model. This model can combine the calibration coefficient to comprehensively adjust and calculate the first assessment result of tool wear, so as to predict the full-cycle wear state of the tool.

[0035] Step S9: Obtain the prediction results of tool wear from the comprehensive wear assessment model, adjust the machine tool's feed rate and spindle rotation rate operating parameters, and execute an emergency stop of the machine tool if necessary to avoid excessive tool wear or damage, and ensure machining quality and safe operation of the machine tool.

[0036] Example 2:

[0037] Based on Example 1, the construction of the digital twin model corresponding to the actual machine tool specifically includes the following construction contents:

[0038] Data preprocessing, time series analysis, mapping function construction, data mapping, validation and adjustment, and updating of digital twin models;

[0039] Data preprocessing: First, the physical entity data collected from the sensors is preprocessed, including data cleaning and formatting, to ensure data quality and consistency. All the collected real-time data is input into the MySQL database at the edge, which serves as the source database from which the data sub-database of the twin system extracts real-time data.

[0040] The digital twin system continuously requests data from the source data, with the three coordinate data being input into each feed axis. Before this, the parent-child relationship and limit of the machine tool multibody system are constructed.

[0041] Time series analysis: Applying time series analysis techniques, such as the autoregressive moving average model ARMA, to analyze the preprocessed data. The purpose of this step is to identify and extract the time dependencies and trends in the data.

[0042] Mapping function construction: Based on the results of time series analysis, a mapping function f is constructed that can convert physical entity data X(t) into data Y(t) in the digital twin model; this function includes polynomial fitting, exponential smoothing or other mathematical models to ensure accurate mapping of the data;

[0043] Data mapping: Using the constructed mapping function f, the physical entity data X(t) at each time point t is mapped to the corresponding data Y(t) in the digital twin model; this step ensures that the digital model can accurately reflect the dynamic behavior of the actual machine tool.

[0044] Verification and Adjustment: Verify the mapped data by comparing the results predicted by the digital model with the observation data of the actual machine tool to evaluate the accuracy of the mapping; if necessary, adjust the mapping function f to optimize the accuracy of the data mapping.

[0045] Update the digital twin model: Integrate the validated and adjusted data into the digital twin model to ensure that the model's data foundation is consistent with the actual machine tool's status. This step is crucial to ensuring that the digital twin model can reflect the machine tool's status in real time.

[0046] The construction of a corresponding data sub-library for each component based on the digital twin model specifically includes the following:

[0047] Data sub-database design, data structure definition, distributed database configuration, data interface development, data storage and synchronization, performance optimization and testing, and monitoring and maintenance;

[0048] Data Sub-library Design: An independent data sub-library is designed for each machine tool component; in this embodiment, the machine tool component data sub-library consists of dynamic data of each feed axis of the machine tool, dynamic physical data of tool wear on the workpiece, and tool usage status and temperature change data.

[0049] Each data sub-database will contain the dynamic data for that component;

[0050] Data structure definition: Define a data structure to store the dynamic data of each component, where each data point contains a timestamp and the corresponding physical quantity value;

[0051] Define the data structure formula as D i =(x i1 ,t1),(x i2 ,t2),…,(x in ,t n ); where D i It is the data sub-database of the i-th component, x ij At time t j Data points collected;

[0052] Distributed database configuration: Select a distributed database system. This example uses Apache Cassandra or Google Cloud Spanner.

[0053] Configure the database to support high-concurrency read and write operations, ensuring data real-time performance and reliability;

[0054] Data Interface Development: Develop API interfaces to allow the analysis library and other systems to access and query data in the data sub-libraries;

[0055] The interface supports real-time data stream processing, ensuring the immediacy of data analysis;

[0056] Data storage and synchronization: Store the physical entity data mapped in step S2 into the corresponding data sub-database;

[0057] Implement a data synchronization mechanism to ensure that the data in the data sub-database is consistent with the actual machine tool status;

[0058] Performance optimization and testing: Optimize database performance, including index optimization and query optimization;

[0059] Conduct system testing to ensure the stability of the data sub-database and the efficiency of data access;

[0060] Monitoring and maintenance: Deploy monitoring tools to monitor the running status and performance indicators of the data sub-databases in real time;

[0061] Perform regular database maintenance, including data backup and fault recovery;

[0062] Technical solution selection: Please refer to Figure 2 We chose to use distributed database technology to build the data sub-database. This technology can efficiently process large amounts of real-time data and supports parallel processing, which is crucial for real-time data analysis.

[0063] The analysis library built based on the digital twin system specifically includes the following:

[0064] Data preprocessing, model selection and configuration, model training, model validation and optimization, real-time data processing module development, result output and feedback;

[0065] Data preprocessing: Data cleaning: Removing noise and outliers from the data sub-database;

[0066] Normalization: Scaling data to a uniform scale to facilitate processing by machine learning algorithms;

[0067] Feature extraction: Extracting useful features from raw data for use in model training;

[0068] Model selection and configuration:

[0069] Algorithm selection: Select a machine learning algorithm based on data characteristics and prediction requirements. In this embodiment, random forest or support vector machine is selected.

[0070] Configuration parameters: Set the hyperparameters of the algorithm, such as the number of trees, depth, or regularization parameters. Specifically, set the parameters of the random forest, such as the number of trees to 100 and the maximum depth to 5, to ensure that the model is neither too complex nor too simple.

[0071] Model training: Dataset partitioning: Divide the dataset into training and validation sets;

[0072] Training the model: Using the training set data to train a machine learning model;

[0073] Model validation and optimization: Cross-validation: Evaluate the model's generalization ability through cross-validation;

[0074] Parameter tuning: Optimize model parameters using grid search or random search methods;

[0075] Real-time data processing module development:

[0076] Interface Design: Design the API interface to receive data from the sub-database in real time;

[0077] Real-time prediction: Develop real-time prediction capabilities to use trained models to predict new data;

[0078] Results and feedback:

[0079] Visualization Interface: Develop a visualization interface to display prediction results and key performance indicators; specifically;

[0080] Based on the component-level digital twin model, a data sub-library is constructed for each component level, serving as the data driver and storage unit for each component. Since not all components have drivers, the component-level digital twin model is divided into three types: active body, variable body, and basic body. Active body components are paired with data sub-libraries, and each sub-library sends the driving data to the analysis library of the digital twin system. The analysis library performs motion analysis on each active body and then inputs the data into the variable body. The variable body generates voxel updates and texture changes based on the analysis data.

[0081]

[0082] The data from the driving body is input into the analysis library. The input data includes the three coordinates of the machine tool spindle and the torque, moment of inertia, speed, spindle current, and acceleration in all directions of the motors of each feed axis. For the variable body (workpiece), a voxel modeling approach is adopted. The unit order of the last digit of the coordinate is used as the side length of the smallest voxel, such as a unit voxel with a side length of 0.01mm. Then, based on the radius of the tool, a cutting unit body is enclosed with the coordinate as the center and the tool radius as the radius. All voxels swept by this cutting unit body are eliminated. This is used as the visual output of the cutting process.

[0083] Feedback mechanism: Establish a feedback mechanism to adjust machine tool operating parameters based on prediction results;

[0084] The process involves acquiring real-time dynamic data from each component and inputting it into the analysis library in step S4. The analysis library then processes the received data and analyzes the motion state of each component, specifically including the following:

[0085] The preprocessed data is input into the analysis library in real time, and the steps of real-time data analysis, instant feedback report and feedback report transmission are performed in sequence. The stability and efficiency of the data input interface are ensured so that the analysis library can receive the data in a timely manner.

[0086] Real-time data analysis: The analysis library uses pre-trained machine learning models to analyze the input real-time data through random forests or support vector machines. The model will predict the motion state and performance indicators of components based on the input data.

[0087] Formula presentation and explanation: in, It is a real-time prediction result based on the current time t, where P is the machine learning prediction model and D'(t) is the preprocessed real-time data;

[0088] Based on the results of real-time data analysis, generate instant feedback reports that include component status assessments, performance metrics, and any necessary operational recommendations.

[0089] Real-time feedback reports are delivered to operators or control systems in a way that is intuitive and easy to understand, such as through dashboards or alarm systems.

[0090] The visualization tool displays key information such as the cutting process, tool temperature, machining data parameters, and spindle load; specifically, it includes the following:

[0091] The visualization steps are as follows: determining visualization metrics, selecting visualization tools and platforms, data preprocessing, real-time data access, designing the visualization interface, and real-time monitoring and interactive functions.

[0092] Determine visualization metrics: First, determine the key performance indicators (KPIs) that need to be visualized: In this example, the component dynamic data contains several parameters;

[0093] In this embodiment, visualization software and platforms are selected. MATLAB, Tab Leau, or professional machine tool monitoring software are used. These tools can process real-time data streams and perform dynamic visualization.

[0094] Data preprocessing: Before visualization, the collected data needs to be preprocessed.

[0095] Data cleaning: Remove outliers and irrelevant data to ensure data accuracy and consistency;

[0096] Data transformation: Converting raw data into a format that visualization tools can process;

[0097] Real-time data access: Set up data flow pipelines to ensure that data is received in real time from the analysis library in the digital twin system, and push the real-time data to the visualization platform using API or database connection;

[0098] Design a visual interface: Design the user interface (UI) to ensure it is intuitive and easy to use, and can clearly display all key information;

[0099] Use charts and graphs to represent key data, such as using a dashboard to display spindle load and a heat map to display tool temperature;

[0100] Real-time animation simulates the cutting process, showing the interaction between the tool and the material;

[0101] Real-time monitoring is enabled, allowing operators to see the current machine tool status and tool wear.

[0102] Add interactive features, such as adjusting the view angle, zooming the chart, and querying specific data points.

[0103] Example 3:

[0104] Based on Example 2, the advanced data analysis is further explained as follows: The dynamic physical influence data includes cutting force, acoustic emission signal index, vibration data, and tool coating performance degradation index, and the cutting force is labeled as QXl, the acoustic emission signal index is labeled as SFs, the vibration data is labeled as ZDs, and the tool coating performance degradation index is labeled as DJTc.

[0105] Cutting force: The three-dimensional cutting force of the tool during the cutting process is measured by sensors, namely axial, radial and tangential forces.

[0106] Reason: Cutting force is a key parameter that directly reflects the interaction between the tool and the workpiece, and its changes can indicate the degree of tool wear;

[0107] The cutting force is measured by a sensor and is divided into axial force F. axial Radial force F radial and tangential force F tangential The mathematical expression is as follows:

[0108] QXl=[F axial F radial F tangential ]

[0109] Among them, F axial It is the axial force, measured in Newtons (N); F radial It is radial force, measured in Newtons (N); F tangential It is tangential force, measured in Newtons (N);

[0110] Acoustic emission signals are generated by monitoring the acoustic emission signals produced by the cutting tool during the cutting process. These signals can reflect microcracks and material peeling when the tool comes into contact with the workpiece.

[0111] The acoustic emission signal includes the signal amplitude A, frequency f, and duration t1, and its mathematical expression is as follows:

[0112]

[0113] Where A is the amplitude of the acoustic emission signal, measured in volts (V) or decibels (dB); f is the frequency of the acoustic emission signal, measured in hertz (Hz); and t1 is the duration of the acoustic emission signal, measured in seconds (s).

[0114] Vibration data measures the frequency and amplitude of vibrations in cutting tools and machine tools. This data can reflect the dynamic stability and wear of the cutting tools.

[0115] Vibration data can reflect the overall dynamic performance of the cutting tool, and abnormal vibration is often related to tool wear; vibration data typically includes the vibration frequency f. v and vibration amplitude A v The mathematical expression is as follows:

[0116]

[0117] Among them, A v This refers to the vibration amplitude, measured in meters (m) or millimeters (mm); f v It is the vibration frequency, and the unit is Hertz (Hz);

[0118] The performance degradation indicators of tool coatings are monitored by using nanoindentation testing or microhardness testing to detect changes in the hardness and elastic modulus of the tool coating; performance degradation of the tool coating, such as reduced hardness and weakened adhesion, will affect the overall performance and wear resistance of the tool.

[0119] The performance degradation index of the tool coating is obtained through nanoindentation testing or microhardness testing, including the coating's hardness H and elastic modulus E; the mathematical expression is as follows:

[0120]

[0121] Where H is the hardness of the coating, in Pascals (Pa) or Gipascals (GPa); E is the elastic modulus of the coating, in Pascals (Pa) or Gipascals (GPa).

[0122] The usage status data includes tool rotation speed, feed rate, tool usage time, and chemical composition change index of the tool-workpiece contact area. The temperature change data monitors the temperature difference between the cutting point and the exposed point. The tool rotation speed is labeled as DJv, the feed rate as JGv, the tool usage time as SYt, and the chemical composition change index of the tool-workpiece contact area as HXBh.

[0123] The tool temperature change data is obtained by directly measuring the temperature difference between the monitored cutting point and the exposed point of the tool using a temperature sensor.

[0124] The exposed point is determined as follows: The exposed point refers to the part of the tool that does not directly participate in cutting, i.e., the non-cutting area of ​​the tool. Methods for determining the exposed point include:

[0125] Visual observation: During the cutting process, the cutting area and non-cutting area of ​​the tool are determined by visual observation or by using a high-speed camera;

[0126] Thermal imager: An infrared thermal imager is used to capture the temperature distribution on the tool surface, and the cutting point and exposed point are determined by the temperature distribution map;

[0127] Sensor positioning: A temperature sensor is installed on the cutting tool, and the cutting point and exposed point are determined by the positioning of the sensor;

[0128] The assessment of temperature difference changes is as follows:

[0129] The wear characteristics of the cutting tool are evaluated by monitoring the temperature difference between the cutting point and the exposed point:

[0130] Thermal stress analysis: Temperature difference reflects the thermal stress experienced by the tool during cutting. A large temperature difference means that there is a large thermal stress inside the tool material, which can lead to premature wear or cracks in the tool.

[0131] Thermal conductivity characteristics: Temperature difference can also reflect the thermal conductivity characteristics of the tool material. If the temperature difference is large, it indicates that the thermal conductivity of the tool is low, which will lead to heat accumulation in the cutting area and aggravate wear.

[0132] Wear mechanism identification: Different wear mechanisms, such as adhesive wear, abrasive wear, and fatigue wear, will lead to different temperature distributions. By analyzing the changes in temperature difference, the dominant wear mechanism can be identified.

[0133] Tool life prediction: The trend of temperature difference can be used to predict the remaining tool life. If the temperature difference continues to increase, it indicates that the tool is about to reach the end of its service life.

[0134] Tool spindle speed and feed rate: Monitor the actual working parameters of the tool, which directly affect the tool wear rate;

[0135] The tool rotation speed is measured by a tachometer and is expressed in revolutions per minute (RPM). The feed rate is the speed at which the tool moves relative to the workpiece and is measured by a feed rate meter and is expressed in millimeters per minute (mm / min) or inches per minute (in / min).

[0136] Tool usage time is the time recorded from the start of tool use to the present, serving as a basic parameter for assessing tool wear.

[0137] Temperature change data is integrated into the calculation of tool usage time, and a temperature change threshold ΔT is introduced. threshold To generate coefficients affecting tool life, a mathematical expression for tool life SYt is generated, and a new coefficient is introduced to reflect the effect of temperature changes on tool life.

[0138]

[0139] If the tool's usage time t is less than or equal to the upper limit of continuous usage time T max And the temperature difference ΔT is less than or equal to the threshold ΔT threshold Then the effective usage time SYt of the tool is equal to the actual usage time t;

[0140] If the tool's usage time t exceeds the upper limit of continuous usage time T max Then the effective service time of the tool, SYt, is equal to T. max ;

[0141] If the temperature difference ΔT exceeds the threshold ΔT threshold Then the effective usage time SYt of the tool will be reduced, and the reduction ratio will be determined by K. temp (ΔT) determines; K temp For every 5% increase or decrease in (ΔT), SYt increases or decreases by 3% accordingly.

[0142] This means that temperature changes negatively impact tool life, reducing the actual effective tool life; and the total effective tool life T is calculated. total for:

[0143]

[0144] in, This represents the total effective usage time function under a specific usage time τ, which is based on the current usage time τ and the upper limit of continuous usage time T. max The effective service time of the tool is calculated using the temperature difference ΔT.

[0145] τ is the integral variable from time 0 to the current time t, representing every instant on the time axis; dτ represents an infinitesimal interval or increment on the time axis. During integration, the effective usage time SYt(τ,T) within these infinitesimal time intervals is used. max Summing ΔT(τ) and ΔT(τ) is used to calculate the total effective usage time from time 0 to time t;

[0146] If the temperature difference ΔT is less than or equal to the threshold ΔT threshold Then the temperature influence coefficient K temp A value of 0 indicates that temperature changes have no effect on tool life.

[0147] If the temperature difference ΔT exceeds the threshold ΔT threshold Then K temp The value will be between 0 and 1, representing a discount on tool life; if ΔT is ΔT threshold +50%, then K temp The value is 0.5, which means that the effective usage time of the tool will be reduced by 50%.

[0148] And K temp The formula for calculating (ΔT) is as follows:

[0149]

[0150] This integral function calculates the total effective usage time from the start of tool use to the current time t. As time progresses, the effect of temperature changes is considered in each calculation of SYt. If the temperature difference exceeds a threshold, the effective tool usage time is reduced in each SYt calculation, thereby reducing T. total The value reflects the cumulative negative impact of temperature changes on tool life;

[0151] The specific parameters are explained below:

[0152] t is the actual time from the start of the tool's use to the present, in hours (h) or minutes (min);

[0153] T max This is the upper limit of the continuous use time of the cutting tool, expressed in hours (h) or minutes (min);

[0154] ΔT is the temperature difference between the cutting point and the exposed point, measured in degrees Celsius (°C) or Kelvin (K).

[0155] ΔT threshold This is the temperature change threshold; once this threshold is exceeded, temperature changes begin to affect tool life.

[0156] ΔT max It is the maximum permissible temperature difference, used for standardization of K. temp The value of T; total This refers to the total effective usage time of the cutting tool.

[0157] K temp (ΔT) is a function that calculates a coefficient affecting tool life based on the temperature difference ΔT; if ΔT exceeds a threshold ΔT... threshold K temp The value will be between 0 and 1, representing a discount on tool life.

[0158] Example 4:

[0159] Based on Example 3, the chemical composition change index of the tool-workpiece contact area is monitored using spectroscopic analysis techniques, such as X-ray fluorescence spectroscopy (XRF) and laser-induced breakdown spectroscopy (LIBS).

[0160] An adjustment factor is introduced, which is related to the area of ​​the chemical composition change point covering the cutting point of the tool. Furthermore, a scoring system is defined to quantify the degree of chemical composition change. The mathematical expression is as follows:

[0161] HXBh=C X ×A cover ×S scale

[0162] Among them, C X It represents the change in the concentration of element X, expressed as a percentage (%) or parts per million (ppm).

[0163] A cover It is the area covered by the point of chemical composition change over the cutting point of the tool, and the unit is square millimeters (mm).

[0164] S scale It is a rating scale coefficient used to convert the degree of change in chemical composition into a rating value, ranging from 0 to 1, where 0 represents no change and 1 represents the maximum change;

[0165] To calculate S scale Define the following function, which is based on C X and A cover The value is used to adjust the score:

[0166]

[0167] In this function, C X , max This is the maximum permissible concentration variation of element X, expressed as a percentage (%) or parts per million (ppm); A max It is the maximum coverage area, measured in square millimeters (mm).

[0168] This embodiment proposes S scale The formula can calculate HXBh based on the degree of change in chemical composition and the area covered by the change point;

[0169] When the change point does not cover the cutting point or the coverage area is smaller, A cover The smaller the value, the less HXBh there is; as the area covered by the change point increases, HXBh also increases; in this way, the impact of chemical composition changes on tool performance can be evaluated more accurately.

[0170] During the cutting process, a chemical reaction occurs between the tool and the workpiece, causing changes in the chemical composition of the tool surface material. These changes affect the wear characteristics of the tool.

[0171] Tool temperature variation data is an important factor affecting tool wear; high temperature will accelerate the fatigue and wear of tool materials.

[0172] The rotational speed and feed rate of the cutting tool directly affect the cutting force and the load on the cutting tool, which in turn affects wear.

[0173] Tool usage time is a fundamental parameter for assessing tool life and wear condition, and can serve as a reference for wear evaluation.

[0174] Example 5:

[0175] Based on Example 4, the first wear evaluation index is defined as E. wear The calculation formula is:

[0176] A first wear evaluation index is constructed by combining indicators such as temperature difference changes, tool usage time, and chemical composition changes.

[0177]

[0178] Wherein, N(QXl,SFs,ZDs,DJTc) is a normalization function used to combine cutting force, acoustic emission signal indicators, vibration data, and tool coating performance degradation indicators into a single index, expressed as follows:

[0179]

[0180] This function aims to unify various types of data to the same scale and extract key features;

[0181] W(DJv,JGv,SYt,HXBh) is a complex information filtering function used to handle the influence of tool rotation speed, feed rate, tool usage time, and chemical composition changes on wear assessment. Its expression is:

[0182]

[0183] The purpose of this function is to filter and weight different influencing factors to extract information that has a significant impact on wear;

[0184] K temp (ΔT) is a temperature influence coefficient used to describe the effect of temperature changes on tool life, and its definition is as described above;

[0185] QXl is the magnitude of the cutting force vector, reflecting the triaxial cutting force of the tool during the cutting process; SFs is the acoustic emission signal index, which reflects the micro-cracks and material peeling when the tool contacts the workpiece; ZDs is vibration data, reflecting the dynamic stability and wear of the tool; DJTc is the performance degradation index of the tool coating, characterizing the changes in coating hardness and elastic modulus; DJv represents the tool rotation speed, which is the rotational speed of the tool during operation; JGv is the feed rate, which refers to the speed at which the tool moves relative to the workpiece; SYt is the tool service time, reflecting the total effective service time of the tool from the start of use to the present; HXBh is the chemical composition change index of the tool-workpiece contact area, reflecting the degree of chemical composition change; ΔT is the temperature difference change data, which is expressed by the temperature influence coefficient K. temp (ΔT) adjustment is used to describe the effect of temperature changes on the first wear evaluation index;

[0186] The range of this formula depends on the design of the normalization function and the complex information filtering function, as well as the range of the temperature influence coefficient. Theoretically, the range of this formula is designed to be 0 ≤ E. wear ≤M, where M is a positive real number, ensuring that the evaluation index has an upper limit to avoid infinite growth; different E wear The value E will reflect the different degrees of tool wear. wear The lower the value, the less tool wear, while E wear The higher the value, the more severe the wear and tear, requiring appropriate maintenance or replacement measures.

[0187] The first wear assessment index E wear The range of values ​​is 0 ≤ E wear ≤M is divided into several intervals, specifically,

[0188] Assuming M = 100, it is divided into the following three intervals: 0-33, 33-66, and 66-100; each interval is described as follows:

[0189] Interval 1: 0 ≤ E wear≤33 indicates relatively light tool wear, suitable for tools in the initial use stage or well-maintained tools. The above description of tool condition is determined according to the manufacturer's guidelines or relevant experts. Within this range, cutting force, acoustic emission signal, vibration data, and coating degradation index all remain at low levels. For example, the three components of cutting force QXl—axial force, radial force, and tangential force—fluctuate within the normal range; the amplitude, frequency, and duration of acoustic emission signal SFs are all below 50% of the threshold; the frequency and amplitude of vibration data ZDs are low; and the changes in coating hardness H and elastic modulus E are within 3%. Tool usage time SYt remains close to the actual usage time under the influence of temperature changes, specifically within ±6% of the actual usage time. The chemical composition change index HXBh also remains at a low level. At this time, the temperature difference ΔT is small, and the temperature influence coefficient K... temp (ΔT) is close to 0;

[0190] Set E respectively wear And the specific threshold ΔT, E wear =20 and ΔT=5℃ are used as the criteria to evaluate the condition of the tool within this range; E wear =20 is used as the judgment standard; when the first wear evaluation index E of the tool is... wear When E is less than 20, it indicates that the tool is in optimal working condition and requires no maintenance; when E wear When the value is between 20 and 33, it indicates that the tool is beginning to show slight wear, but it is still within the acceptable range. Routine inspection should be performed, and immediate replacement is not necessary. ΔT = 5℃ is the judgment standard. When ΔT is less than 5℃, the temperature effect is not significant. When ΔT is greater than 5℃ but less than 10℃, attention should be paid to the potential impact of temperature on tool life.

[0191] Within this range, the interaction relationships between the parameters are as follows:

[0192] If the cutting force QXl increases by 10%, the amplitude of the acoustic emission signal SFs will increase by 5%, the frequency of the vibration data ZDs will increase by 3%, the change in the coating hardness H and elastic modulus E will not exceed 2%, and the increase in tool speed DJv and feed rate JGv will lead to E wear Increase, but due to the temperature effect coefficient K temp (ΔT) remains low, having a relatively small impact on the overall first wear assessment index; the change in the chemical composition variation index HXBh within this range has a significant impact on E. wear The impact does not exceed 5%; these changes collectively reflect the stable state of the cutting tool in the initial stage of use, which helps to determine the health status of the cutting tool and carry out preventive maintenance;

[0193] Interval 2, 33 <E wear≤66 indicates that the tool is in a state of moderate wear, applicable to tools after a period of use. "A period of use" means the tool's service life is between 10% and 25% of its current life. Within this range, the three components of the cutting force QXl—axial force, radial force, and tangential force—increase. The amplitude, frequency, and duration of the acoustic emission signal SFs all reach the threshold of 50% to 75%. The frequency and amplitude of the vibration data ZDs increase significantly by less than 30%. The hardness H and elastic modulus E of the coating begin to show significant changes of less than 9%. The tool service time SYt begins to shorten under the influence of temperature changes, with a reduction of less than 10% of the current service time. The chemical composition change index HXBh also increases. At this time, the temperature difference ΔT is large, and the temperature influence coefficient K... temp (ΔT) is between 0.2 and 0.5;

[0194] Within this interval, E is set respectively. wear With respect to the specific threshold ΔT, E wear =50 and ΔT=20℃ are used as the judgment criteria; when the first wear evaluation index E of the tool is... wear When the value is less than 50, it indicates that the tool wear is still within a controllable range and can continue to be used, but more frequent monitoring and appropriate maintenance are required; when E wear When the temperature is greater than 50, it indicates that the tool wear is relatively severe, requiring detailed inspection and tool replacement; when ΔT is between 10℃ and 20℃, the temperature effect begins to be significant, and measures must be taken to reduce the temperature; when ΔT exceeds 20℃, the tool life will be greatly shortened, requiring immediate action.

[0195] Within this range, the interaction relationships between the parameters are as follows:

[0196] If the cutting force QXl increases by 15%, the amplitude of the acoustic emission signal SFs will increase by 10%, the frequency of the vibration data ZDs will increase by 8%, and the changes in the coating hardness H and elastic modulus E will reach 5%; increasing the tool speed DJv and feed rate JGv will significantly improve E. wear Meanwhile, the temperature influence coefficient K temp An increase in (ΔT) will significantly increase the overall first wear evaluation index; the change in the chemical composition change index HXBh within this range will affect E. wear The impact is significant, reaching over 10%. These changes reflect the wear condition of the cutting tool during the mid-term use stage, helping users determine whether the cutting tool needs maintenance or replacement.

[0197] Interval 3, 66 <E wear ≤100 indicates that the tool is in a state of severe wear, which applies to tools that are about to reach the end of their service life or have already exceeded their service life; about to reach the end of their service life means that the current tool service life is less than 5% of the total service life.

[0198] Within this range, the three components of the cutting force QXl—axial force, radial force, and tangential force—significantly increase; the amplitude, frequency, and duration of the acoustic emission signal SFs all exceed the 75% threshold; the frequency and amplitude of the vibration data ZDs increase significantly by more than 30%; the hardness H and elastic modulus E of the coating decrease significantly by more than 9%; the tool service time SYt is significantly shortened by more than 10% under the influence of temperature changes; and the chemical composition change index HXBh also increases by more than 16%. At this point, the temperature difference ΔT is extremely large, and the temperature influence coefficient K... temp (ΔT) is between 0.5 and 1.0;

[0199] Within this interval, E is set respectively. wear With respect to the specific threshold ΔT, E wear =80 and ΔT=25℃ are used as the judgment criteria. When the first wear evaluation index E of the tool is... wear A value less than 80 indicates that the tool is severely worn and needs to be stopped and replaced immediately; when E wear When the temperature is above 80°C, it indicates that the tool has lost its normal function and must be replaced immediately to avoid affecting the machining quality and equipment safety. When the temperature exceeds 25°C, the temperature effect is extremely significant, and the tool life is almost exhausted. When the temperature reaches 30°C, the tool is in a scrapped state and must be replaced immediately.

[0200] Within this range, the interaction relationships between the parameters are as follows:

[0201] If the cutting force QXl increases by 20%, the amplitude of the acoustic emission signal SFs will increase by 15%, the frequency of the vibration data ZDs will increase by 12%, and the changes in the coating's hardness H and elastic modulus E will reach 10%; increasing the tool speed DJv and feed rate JGv will significantly improve E. wear Meanwhile, the temperature influence coefficient K temp An increase in (ΔT) will significantly increase the overall first wear assessment index; the change in the chemical composition index HXBh within this range will affect E. wear The impact is significant, with changes exceeding 20% ​​of the current value; these changes reflect the wear condition of the cutting tool in the final stage of use, helping users determine whether the tool needs to be replaced immediately to ensure machining quality and equipment safety.

[0202] Example 6:

[0203] Building upon Example 5, the method further utilizes twin data and expert knowledge to introduce a calibration coefficient for wear quantification. This calibration coefficient is then combined with a first wear evaluation index and analyzed to construct a comprehensive wear assessment model. This model, incorporating the calibration coefficient, comprehensively adjusts and calculates the initial assessment results of tool wear to predict the tool's full-cycle wear state. Specifically, this includes the following:

[0204] The calibration coefficient for introducing wear quantization is defined as follows:

[0205] Define calibration coefficients: Calibration coefficients are a set of parameters derived from actual operation and experimental data, used to adjust the deviation of the theoretical model. Several key calibration coefficients are defined.

[0206] C temp It is the temperature calibration coefficient; C force It is the cutting force calibration factor; C vibration It is the vibration calibration coefficient; C time It is the time calibration factor; C chemistry It is a calibration coefficient for changes in chemical composition;

[0207] These calibration coefficients are calibrated using experimental and historical data to improve the accuracy of the model;

[0208] The calibration coefficients, combined with the first wear evaluation index, are as follows:

[0209] The calibration factor is introduced into the first wear evaluation index E. wear In the calculation, the formula is adjusted as follows:

[0210]

[0211] Among them, N(QXl·C) force ,SFs·C vibration ,ZDs·C vibration DJTc is the adjusted normalization function;

[0212] W(DJv,JGv,SYt·C time ,HXBh·C chemistry ) is the adjusted complex information filtering function;

[0213] K temp (ΔT)·C temp This is the adjusted temperature influence coefficient;

[0214] Dynamic physical data is continuously collected from sensors and mapped into a digital twin model; an initial first wear assessment index E is calculated using an uncalibrated formula. wear Based on the initial first wear evaluation index E wearThe calibration coefficient is applied for adjustment to obtain the calibrated first wear evaluation index E. wear,calibrated ;

[0215] The predicted full-cycle wear state of the tool is as follows:

[0216] Based on the calibrated first wear evaluation index E wear,calibrated A time series model based on the ARIMA model is established to predict future wear trends; by combining historical data and the prediction model, the wear status of the tool throughout its entire life cycle is evaluated, and the key stages of wear and the predicted failure time points are identified.

[0217] Real-time monitoring of tool usage status and wear, and dynamic adjustment of calibration coefficient C. temp C force C vibration C time C chemistry To improve the accuracy of the model, manual verification and adjustment are carried out in conjunction with expert knowledge to confirm the rationality of the model's prediction results.

[0218] The accuracy and reliability of the model were evaluated through experimental verification and practical application verification. Based on the verification results, the calibration coefficients and model parameters were optimized to further improve the accuracy of wear prediction.

[0219] Through the above steps, combined with calibration coefficients for analysis and processing, the constructed comprehensive wear assessment model can not only comprehensively adjust and calculate the initial assessment results of tool wear, but also predict the full-cycle wear state of the tool, ensuring the tool's efficiency and safety. The prediction of the full-cycle wear state specifically includes the following:

[0220] The specific steps for collecting real-time data are as follows:

[0221] 1. Install sensors on key parts of machine tools and cutting tools, including cutting force sensors, acoustic emission sensors, vibration sensors, temperature sensors, and sensors for related parameters;

[0222] Through the Internet of Things (IoT) system, sensor data is collected in real time, including cutting force QXl, acoustic emission signal SFs, vibration data ZDs, tool coating performance degradation index DJTc, tool speed DJv, feed rate JGv, tool usage time SYt, chemical composition change index HXBh, and temperature difference change data ΔT.

[0223] 2. The specific steps for calculating the initial first wear evaluation index are as follows:

[0224] The collected raw data is cleaned, filtered, and denoised to obtain accurate input data.

[0225] The initial first wear assessment index E was calculated using an uncalibrated formula. wear ;

[0226]

[0227] 3. The specific steps for adjusting using calibration coefficients are as follows:

[0228] Based on historical experimental data and actual usage data, the calibration coefficient C was determined. temp C force C vibration C time C chemistry ;

[0229] Substituting the calibration coefficient into the formula for the first wear evaluation index, the adjusted formula is:

[0230]

[0231] Based on the adjusted formula, the calibrated first wear evaluation index E is calculated in real time. wear,calibrated ;

[0232] 4. The specific steps for establishing a time series model are as follows:

[0233] Collect historical wear and tear data over a certain period of time. wear,calibrated And form a time series dataset;

[0234] Choose an appropriate time series forecasting model, such as the ARIMA model or the LSTM neural network;

[0235] Use historical data to train time series models so that the models can learn the time dependence and trends of wear and tear data;

[0236] Use historical data that was not used in training to verify the accuracy of the model, adjust the model parameters, and improve prediction accuracy;

[0237] 5. The specific steps for full-cycle wear condition assessment are as follows:

[0238] Using a trained time-series model, the first wear assessment index E is calibrated based on current and historical data. wear,calibrated Predict future wear and tear conditions;

[0239] Based on the prediction results, the key stages of tool wear are identified, including the initial wear stage, the stable stage, and the rapid wear stage.

[0240] Determining the predicted failure time of the cutting tool provides a reference for preventive maintenance and replacement;

[0241] 6. The specific steps for real-time monitoring and feedback adjustments are as follows:

[0242] The dynamic data of the cutting tool is continuously monitored through an IoT system, and the first wear assessment index E is updated in real time. wear,calibrated ;

[0243] The calibration coefficient C is dynamically adjusted based on real-time monitoring data and model prediction results. temp C force C vibration C time C chemistry This improves the accuracy of the model;

[0244] By combining the experience of field engineers and experts, the model prediction results are manually verified and adjusted to ensure the accuracy of wear condition assessment;

[0245] 7. Comprehensive evaluation of model verification and optimization, the specific steps are as follows:

[0246] The accuracy and reliability of the comprehensive wear assessment model were evaluated through practical use and experimental verification.

[0247] Based on the validation results, the calibration coefficients and model parameters were optimized to further improve the accuracy of wear prediction;

[0248] Establish a feedback mechanism to continuously improve the model based on usage, ensuring its adaptability and accuracy in different usage environments;

[0249] By implementing the above detailed steps, a comprehensive evaluation model is constructed. Combined with calibration coefficients for adjustment calculations, the full-cycle wear status of the tool is predicted. This model can achieve the first assessment of the tool wear amount and provide real-time monitoring and dynamic adjustment functions to ensure tool usage efficiency and production safety.

[0250] The comprehensive adjustment calculation of the first assessment result of tool wear includes the following:

[0251] For the interval 1, 0 ≤ E wear ≤33:

[0252] Initial wear condition: The tool wear is relatively light, suitable for the initial use stage or well-maintained tools;

[0253] Temperature difference change ΔT, ΔT≤5℃;

[0254] The cutting force QXl has three components: axial force, radial force, and tangential force, which fluctuate within the normal range of 15-25N.

[0255] Acoustic emission signals (SFs) with amplitude, frequency, and duration all below 50% of the threshold, for example, amplitude in the range of 2-4 mV;

[0256] Vibration data (ZDs) are low in frequency and amplitude, with frequencies ranging from 0.5 to 1.5 kHz and amplitudes from 0.01 to 0.03 mm.

[0257] The coating has a hardness of DJTc, with minimal changes in hardness and elastic modulus, and a hardness of 1500-2000HV.

[0258] The first wear rating index E after calibration wear,calibrated After adjusting the wear index by calibration coefficient, it remains between 0 and 33, indicating that the tool wear is stable and it can continue to be used.

[0259] Future wear prediction: The prediction model shows that the wear rate is slow, and it can be predicted that the tool can continue to be used for at least 100 hours under this condition;

[0260] Real-time monitoring and feedback adjustment

[0261] Data monitoring continuously monitors cutting force, acoustic emission signals, vibration, and temperature data;

[0262] The calibration coefficient is adjusted in real time, dynamically based on real-time data. For example, if the temperature difference ΔT exceeds 5℃, then C is adjusted. temp 5% of the current value;

[0263] An early warning mechanism is in place, setting warning thresholds such as when the cutting force exceeds 25N or the acoustic emission signal amplitude exceeds 4mV, to issue a maintenance reminder; expert optimization involves combining expert knowledge to regularly verify data and models to ensure monitoring accuracy;

[0264] Interval 2, 33 <E wear ≤66; Mid-term wear condition, the tool is in a moderate wear condition, suitable for tools that have been used for a period of time;

[0265] The temperature difference varies greatly, with ΔT ranging from 10℃ to 20℃.

[0266] The cutting force QXl increases by 25-40N;

[0267] Acoustic emission signals (SFs) with amplitude, frequency, and duration all within the 50%–75% threshold, for example, amplitude of 4–6 mV;

[0268] Vibration data ZDs show increased frequency and amplitude, with frequencies ranging from 1.5 to 2.5 kHz and amplitudes from 0.03 to 0.05 mm.

[0269] The coating hardness DJTc changed significantly, with the hardness decreasing to 1000-1500 HV;

[0270] The first wear rating index E after calibration wear,calibratedA wear index between 33 and 66 indicates that continued monitoring and appropriate maintenance are required.

[0271] Future wear prediction: The prediction model shows that the wear rate is accelerating, and it can be predicted that the tool can continue to be used for about 50 hours under this condition.

[0272] Feedback adjustment is implemented to adjust the calibration coefficient in real time. If the temperature difference ΔT exceeds 20℃, C is adjusted promptly. temp 5% of the current value and C time 7% of the current value;

[0273] An early warning mechanism is in place to issue a replacement reminder if the cutting force exceeds 40N or the acoustic emission signal amplitude exceeds 6mV; based on expert opinions, tool adjustments or replacements are made to improve safety.

[0274] Interval 3, 66 <E wear ≤100; End-stage wear condition, the tool is in a state of severe wear, applicable to tools that are about to reach the end of their service life or have exceeded their service life;

[0275] The temperature difference change ΔT is extremely large, exceeding 25℃; the cutting force QXl increases significantly, by 40-60N; the acoustic emission signal SFs exceeds the 75% threshold in amplitude, frequency, and duration, for example, the amplitude is 6-8mV; the vibration data ZDs show a significant increase in frequency and amplitude, with the frequency at 2.5-4kHz and the amplitude at 0.05-0.1mm; the coating hardness DJTc decreases significantly, with the hardness dropping to 500-1000HV; the calibrated first wear evaluation index E... wear,calibrated A wear index between 66 and 100 indicates that the tool is about to fail and needs to be replaced immediately. The predictive model shows that the wear rate is very fast, and the tool can only be used for a maximum of 10 hours under this condition.

[0276] Data monitoring is conducted, closely monitoring all key parameters, especially temperature and cutting force; calibration coefficients are adjusted in real time: if the temperature difference ΔT exceeds 30℃, C is adjusted promptly. temp 20% of the current value and C force If the current value is 15%; the cutting force exceeds 60N; or the acoustic emission signal amplitude exceeds 8mV, a replacement instruction is immediately issued; based on the prediction results, a tool replacement plan is arranged in advance to ensure uninterrupted production; through the above detailed quantitative explanation and zoning management, the full-cycle wear status of the tool can be more accurately assessed, data can be monitored in real time and feedback adjustments can be made to ensure tool utilization efficiency and production safety.

[0277] Example 7:

[0278] Building upon Example 6, this example aims to verify the calibration of the first wear evaluation index E. wear,calibratedAssess the tool wear condition and adjust machine tool operating parameters based on the prediction results to avoid excessive tool wear or damage, and ensure machining quality and safe machine tool operation;

[0279] Machine tool and cutting tool selection: A certain model of five-axis CNC machine tool and coated carbide cutting tools were selected, and various sensors were installed, including cutting force sensors, acoustic emission sensors, vibration sensors, temperature sensors, and other sensors required for acquiring relevant parameters; steel with certain hardness and toughness was selected as the processing material to ensure the harshness of the test conditions; all sensors were calibrated to ensure the accuracy of data acquisition;

[0280] The machine tool and cutting tool are collected in real time by sensors to collect various dynamic data during the machining process, including cutting force QXl, acoustic emission signal SFs, vibration data ZDs, performance degradation index of tool coating DJTc, tool speed DJv, feed rate JGv, tool usage time SYt, chemical composition change index HXBh, and temperature difference change data ΔT.

[0281] Continuously monitor cutting force, acoustic emission signals, vibration, temperature, and related data; set up an early warning mechanism to promptly issue maintenance reminders when the cutting force exceeds 40N or the acoustic emission signal amplitude exceeds 6mV; dynamically adjust the calibration coefficient, such as adjusting C if the temperature difference ΔT exceeds 20℃. temp and C time ;

[0282] The time series model is used to predict the first wear evaluation index after calibration and identify the state of the tool at different wear stages. Based on the prediction results, the feed rate and spindle rotation rate operation parameters of the machine tool are adjusted, and the machine tool is stopped urgently when necessary to avoid excessive wear or damage to the tool.

[0283] Record the first wear assessment index, sensor data, calibration coefficient adjustments, and machine tool operating parameters at different time points; analyze the data to verify the accuracy and effectiveness of the comprehensive wear assessment model, ensuring the safe operation of the tool under various wear conditions; the table below shows the experimental data records, and the Excel spreadsheet is as follows.

[0284] Table 1

[0285]

[0286]

[0287] The adjustment strategies for each calibration factor are shown in the table below:

[0288] Table 2:

[0289]

[0290] The adjustment strategy is explained in detail below:

[0291] For the temperature calibration coefficient C temp Tool A: The temperature difference change ΔT is small, and the calibration coefficient remains at the default value of 1.00;

[0292] Tool B: As the temperature difference change ΔT increases, the calibration coefficient is adjusted to 1.05 to compensate for the effect of temperature on wear;

[0293] Tool C: The temperature difference change ΔT increased significantly, and the calibration coefficient was adjusted to 1.10 to significantly adjust the model's response to temperature;

[0294] Tool D: The temperature difference change ΔT is large, so the calibration coefficient is adjusted to 1.08 to appropriately adjust the model's sensitivity to temperature;

[0295] Cutting force calibration factor C force As the cutting force increases, the calibration coefficient gradually increases to reflect the direct impact of the cutting force on wear.

[0296] Vibration calibration factor C vibration As vibration data increases, the calibration coefficient gradually increases to reflect the impact of vibration on tool wear.

[0297] Time calibration factor C time As the tool's usage time increases, the calibration coefficient gradually increases to reflect the impact of time on wear accumulation.

[0298] Chemical composition calibration factor C chemistry As the chemical composition changes, the calibration coefficient gradually increases to reflect the impact of the chemical composition change on the tool performance;

[0299] By comparing and analyzing the data in the tables, the beneficial effects of this invention on tool wear assessment and machining parameter adjustment can be clearly seen. The detailed analysis process is as follows:

[0300] Interval 1 (0≤E) wear ≤33), the data for tool A indicates relatively light wear, and the specific analysis is as follows:

[0301] 1. Wear assessment: The cutting force QXl of tool A is 20N, the acoustic emission signal SFs is 3mV, the vibration data ZDs is 1.0kHz, the coating hardness DJTc is 1800HV, and the temperature difference change ΔT is only 3℃.

[0302] 2. First Wear Evaluation Index: The calibrated first wear evaluation index E wear,calibrated The calculation shows that the index of tool A is 25, indicating that the tool wear is stable;

[0303] 3. Future wear prediction: The remaining life is predicted to be 100 hours, indicating that the tool can continue to be used for a long time under the condition that the wear rate is slow.

[0304] By introducing calibration coefficients, such as C temp and C force It accurately reflects the actual wear condition of the tool, improving accuracy by 20% compared to the uncalibrated model; Parameter adjustment: When the wear is light, the machine tool operating parameters do not need to be adjusted significantly, and the current settings can be maintained, reducing unnecessary downtime and increasing production efficiency by 10%;

[0305] Interval 2(33) <E wear ≤66), in interval 2, the data for tool B indicates moderate wear, as detailed below:

[0306] 1. Wear assessment: The cutting force QXl of tool B is 35N, the acoustic emission signal SFs is 5mV, the vibration data ZDs is 2.0kHz, the coating hardness DJTc is 1300HV, and the temperature difference change ΔT is 15℃.

[0307] 2. First Wear Evaluation Index: The calibrated first wear evaluation index E wear,calibrated A value of 50 indicates that continued monitoring is required; 3. Future wear prediction: The predicted remaining lifespan is 50 hours, indicating that the wear rate is accelerating;

[0308] In the calibrated model, the newly added temperature influence factor K temp (ΔT) improves the model’s sensitivity to temperature changes, increasing accuracy by 25%. Based on the wear condition, the feed rate JGv and spindle speed DJv are reduced to control the wear rate, avoid accelerating tool failure, and extend tool life by 30%.

[0309] Interval 3 (66) <E wear ≤100), in interval 3, the data for tool C and tool D indicate severe wear, as detailed below:

[0310] 1. Wear assessment: Tool C: cutting force QXl is 50N, acoustic emission signal SFs is 7mV, vibration data ZDs is 3.0kHz, coating hardness DJTc is 800HV, temperature difference change ΔT is 28℃;

[0311] Tool D: Cutting force QXl is 45N, acoustic emission signal SFs is 6mV, vibration data ZDs is 2.5kHz, coating hardness DJTc is 900HV, and temperature difference change ΔT is 22℃.

[0312] 2. First Wear Evaluation Index: Tool C: Calibrated First Wear Evaluation Index E wear,calibrated85; Tool D: First wear evaluation index E after calibration wear,calibrated It is 75;

[0313] 3. Future Wear Prediction: Tool C: Predicted remaining lifespan is 10 hours; Tool D: Predicted remaining lifespan is 15 hours; Technical Effects:

[0314] Comprehensive adjustment calculation: With the help of real-time monitoring and feedback adjustment functions, the model can reflect the severe wear state of the tool in a timely manner, avoid emergency shutdown, and improve the prediction accuracy by 30%.

[0315] Parameter adjustment: In cases of severe wear, by reducing the feed rate and spindle speed, and performing emergency stops when necessary, more serious tool damage can be avoided, ensuring machining quality and safe machine operation, and improving production safety by 40%.

[0316] The relationship and mutual influence of changes in formula parameters:

[0317] 1. Increased cutting force QXl: This leads to an increase in N(QXl,SFs,ZDs,DJTc); the acoustic emission signal SFs and vibration data ZDs will also increase, reflecting increased tool load and accelerated wear;

[0318] Increase the first wear assessment index E wear Operating parameters need to be adjusted to reduce the wear rate;

[0319] 2. An increase in the temperature difference change ΔT directly increases the temperature influence coefficient K. temp (ΔT); By increasing the denominator in the formula, the first wear evaluation index E is affected. wear The increase is faster; ΔT needs to be adjusted in a timely manner to reflect the impact of real temperature on wear and ensure the accuracy of model predictions.

[0320] 3. A decrease in coating hardness DJTc indicates that the tool coating has degraded, with reduced hardness and elastic modulus;

[0321] This leads to a decrease in N(QXl,SFs,ZDs,DJTc), but actually increases the first comprehensive wear index E. wear ;

[0322] Monitoring needs to be strengthened, and C needs to be adjusted. vibration and C force To improve the accuracy of the model;

[0323] Data analysis in the table reveals that the invention has significant beneficial effects in the following aspects:

[0324] 1. Comprehensive adjustment calculation: By introducing calibration coefficients, the accuracy of wear assessment is significantly improved, especially under conditions of significant temperature changes and high load, the model's prediction accuracy is improved by 20%-30%;

[0325] 2. Real-time monitoring and feedback adjustment: Continuously monitor cutting force, acoustic emission signals, vibration and temperature data, dynamically adjust calibration coefficients to ensure the real-time accuracy of the model, reduce unplanned downtime, and improve production efficiency and safety;

[0326] 3. Critical Stage Identification and Failure Time Prediction: By using time series models to predict wear status, the critical stages of tool wear are accurately identified, and the failure time is predicted, providing a scientific basis for preventive maintenance and replacement, extending tool life and reducing production costs;

[0327] 4. Parameter adjustment: Adjust the machine tool operating parameters according to the wear condition to avoid excessive tool wear or damage and ensure machining quality. This is especially important in the middle and late stages of wear, which improves the safety and efficiency of overall production.

[0328] In summary, the present invention demonstrates significant innovation and novelty in tool wear assessment, parameter adjustment, and wear state prediction, and can provide strong technical support and assurance for actual production processes.

[0329] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0330] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0331] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0332] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for building a digital twin model based on real-time data-driven methods, characterized in that, The specific steps include: Step S1: Collect dynamic data of components at the physical entity level of the machine tool. This component dynamic data includes the torque, moment of inertia, and speed of the motors of each feed axis of the CNC machine tool, and the acceleration of the machine tool spindle in the three coordinate directions; it also includes dynamic physical influence data of the workpiece on tool wear, as well as tool usage status and temperature change data. Step S2: Acquire the collected physical entity data and map it into the digital twin model to construct a digital twin model corresponding to the actual machine tool; Step S3: Based on the digital twin model, construct a corresponding data sub-library for each component, with each sub-library serving as a data driver and storage unit; Step S4: Construct an analysis library based on the digital twin system. This analysis library is designed to receive and process data from the data sub-libraries of each component, and to perform advanced data analysis and motion state assessment. Step S5: Acquire the real-time dynamic data of each component and input it into the analysis library in step S4. The analysis library processes the received data and analyzes the motion state of each component. Step S6: Display key information such as the cutting process, tool temperature, machining data parameters, and spindle load using visualization tools; Step S7: Obtain dynamic physical influence data of the workpiece on tool wear, as well as tool usage status and temperature change data, and analyze and process them to generate a first wear evaluation index, which is used to make the first assessment of tool wear. Step S8: Using twin data and expert knowledge, introduce a calibration coefficient for wear quantification, combine the calibration coefficient with the first wear evaluation index, and perform analysis and processing to construct a comprehensive wear assessment model. This model can combine the calibration coefficient to comprehensively adjust and calculate the first assessment result of tool wear, so as to predict the full-cycle wear state of the tool. Step S9: Obtain the prediction results of tool wear from the comprehensive wear assessment model, and adjust the machine tool's feed rate and spindle rotation rate operating parameters to avoid excessive tool wear or damage. The dynamic physical influence data includes cutting force, acoustic emission signal index, vibration data, and tool coating performance degradation index, and the cutting force is labeled as QXl, the acoustic emission signal index is labeled as SFs, the vibration data is labeled as ZDs, and the tool coating performance degradation index is labeled as DJTc. The usage status data includes tool rotation speed, feed rate, tool usage time, and chemical composition change index of the tool-workpiece contact area. The temperature change data monitors the temperature difference between the cutting point and the exposed point. The tool rotation speed is labeled as DJv, the feed rate as JGv, the tool usage time as SYt, and the chemical composition change index of the tool-workpiece contact area as HXBh. The tool temperature change data is obtained by directly measuring the temperature difference between the monitored cutting point and the exposed point of the tool using a temperature sensor. The exposed point refers to the part of the tool that does not directly participate in cutting, i.e., the non-cutting area of ​​the tool. Tool usage time is the time recorded from the start of tool use to the present. Temperature change data is integrated into the calculation of tool usage time, and a temperature change threshold is introduced. To generate coefficients affecting tool life, and thus generate a mathematical expression for tool life SYt: in, It is with the temperature difference The change in temperature effect coefficient, This represents the effective usage time function under usage time t, which is based on the current time t and the upper limit of continuous usage time. and temperature difference To calculate the effective usage time of the cutting tool; The temperature influence coefficient Specifically, it is expressed as follows: If the tool's usage time t is less than or equal to the upper limit of continuous usage time And temperature difference Less than or equal to the threshold Then the effective usage time SYt of the tool is equal to the actual usage time t; If the tool's usage time t exceeds the upper limit of continuous usage time... Then the effective service time of the tool, SYt, is equal to ; If temperature difference Exceeded the threshold If this happens, the effective usage time of the tool, SYt, will be reduced. If temperature difference Less than or equal to the threshold The temperature influence coefficient A value of 0 indicates that temperature changes have no effect on tool life.

2. The method for building a digital twin model based on real-time data-driven approach according to claim 1, characterized in that: The construction of the digital twin model corresponding to the actual machine tool specifically includes the following construction contents: Data preprocessing, time series analysis, mapping function construction, data mapping, validation and adjustment, and updating of digital twin models; The process of constructing a corresponding data sub-library for each component based on the digital twin model includes the following steps in sequence: Data sub-database design, data structure definition, distributed database configuration, data interface development, data storage and synchronization, performance optimization and testing, and monitoring and maintenance; Design an independent data sub-library for each machine tool component; the data sub-library consists of dynamic data of each feed axis of the machine tool, dynamic physical data of tool wear on the workpiece, and tool usage status and temperature change data. The analysis library is constructed based on the digital twin system. The steps for constructing the analysis library include the following: Data preprocessing, model selection and configuration, model training, model validation and optimization, development of real-time data processing modules, and result output and feedback.

3. The method for building a digital twin model based on real-time data-driven approach according to claim 2, characterized in that: For the chemical composition change index in the tool-workpiece contact area, spectral analysis technology is used to monitor the chemical composition change in the tool-workpiece contact area; a scoring system is defined to quantify the degree of chemical composition change, and the mathematical expression is as follows: in, It is the change in the concentration of element X; It is the area covered by the point of chemical composition change at the cutting point of the tool; It is a rating scale coefficient used to convert the degree of change in chemical composition into a rating value, ranging from 0 to 1, where 0 represents no change and 1 represents the maximum change; Define the following function to calculate The function is based on and The value is used to adjust the score: In this function, It is the maximum permissible concentration change of element X; It is the maximum coverage area; When the change point does not cover the cutting point or the coverage area is smaller The smaller the value, the less HXBh will be; as the area covered by the change point increases, HXBh will also increase.

4. The method for building a digital twin model based on real-time data as described in claim 3, characterized in that: The first wear evaluation index is defined as: The calculation formula is: in, It is a normalization function used to combine cutting force, acoustic emission signal indicators, vibration data, and tool coating performance degradation indicators into a single index; the specific expression is as follows: This is a complex information filtering function used to process the influence of tool rotation speed, feed rate, tool usage time, and chemical composition changes on wear assessment; the specific expression is: It is a temperature influence coefficient used to describe the effect of temperature changes on tool life. The first wear assessment index value range is designed as follows: Where M is a positive real number, The lower the value, the less the tool wear, and The higher the value, the more severe the wear and tear, requiring appropriate maintenance or replacement measures. The first wear assessment index value range Divided into several intervals, specifically, Assuming M=100, it is divided into the following three intervals: 0-33, 33-66, and 66-100; each interval is described as follows: Interval 1, This range indicates relatively light tool wear, suitable for initial use or well-maintained tools; (Separate settings are needed). and Specific thresholds are used as judgment criteria to evaluate the state of the tool within that range; Interval 2, This range indicates that the tool is in a state of moderate wear and is used for tools after a period of use; [The text then abruptly shifts to a different topic:] ...set separately... and A specific threshold is used as a judgment standard; this judgment standard is used to adjust the degree of tool wear and maintenance content. Interval 3, This range indicates that the tool is in a severely worn state, and is used for tools that are about to reach the end of their service life or have already exceeded their service life; [The following is a separate setting:] ... and The specific threshold is used to assess whether the cutting tool needs to be replaced.

5. The method for building a digital twin model based on real-time data-driven approach according to claim 4, characterized in that: The method utilizes twin data and expert knowledge to introduce a calibration coefficient for wear quantification. This calibration coefficient is then combined with a first wear evaluation index and analyzed to construct a comprehensive wear assessment model. This model, incorporating the calibration coefficient, comprehensively adjusts and calculates the initial assessment results of tool wear to predict the tool's full-cycle wear state. Specifically, this includes the following: The calibration coefficient for introducing wear quantization is defined as follows: It is the temperature calibration coefficient; It is the cutting force calibration coefficient; It is the vibration calibration coefficient; It is the time calibration factor; It is a calibration coefficient for changes in chemical composition; Incorporate calibration coefficients into the first wear assessment index. In the calculation, the formula is adjusted as follows: in, It is the adjusted normalized function; It is the adjusted complex information filtering function; This is the adjusted temperature influence coefficient; It is the first wear evaluation index after calibration.

6. The method for building a digital twin model based on real-time data-driven approach according to claim 5, characterized in that: The comprehensive adjustment calculation of the first assessment result of tool wear includes the following: For interval 1, Temperature difference change , 5°C; First wear rating index after calibration Maintaining a value between 0 and 33 indicates stable tool wear and continued usability; the calibration coefficient is dynamically adjusted based on real-time data, depending on temperature differences. If the temperature exceeds 5°C, adjust accordingly. 5% of the current value; Interval 2, Temperature difference change Larger Between 10°C and 20°C; the calibrated first wear rating index A value between 33 and 66 indicates that continued monitoring and appropriate maintenance are required. If temperature difference If the temperature exceeds 20°C, adjust 5% of the current value and 7% of the current value; Interval 3, Temperature difference change great, Exceeding 25°C; First wear rating index after calibration A reading between 66-100 indicates the tool is about to fail and needs immediate replacement; if the temperature difference... If the temperature exceeds 30°C, adjust 20% of the current value and 15% of the current value.

Images