A lathe automatic machining control method and system based on numerical control device

Abstract

The present invention discloses a method and system for controlling automated machining of lathes based on numerical control devices, and relates to the technical field of numerical control lathes. A control system for automated machining of lathes based on numerical control devices comprises: a data acquisition module, an automatic control module, a tool analysis module, a dynamic control module, and a production control module. Through a tool condition analysis model, combined with multi-dimensional real-time production data, the present invention can assess the health of the tool in real time and predict the remaining life of the tool. This intelligent analysis can promptly detect problems such as tool wear and damage, avoid machining inaccuracies or production interruptions caused by tool failures, thereby effectively extending the tool life and improving machining stability.

Description

Technical Field

[0001] The present invention relates to the technical field of numerically controlled lathes, and in particular to a numerically controlled device-based lathe automated machining control method and system. Background Art

[0002] With the continuous development of the manufacturing industry, lathes, as an important processing equipment, are widely used in metalworking, mold manufacturing, and component production. Traditional lathes often rely on manual operation, with machining parameters manually adjusted to meet the processing requirements of different workpieces. However, with the increasing demand for machining precision, traditional lathes have gradually exposed some limitations, especially when processing high-precision and complex workpieces, which make it difficult to achieve precise control and efficient machining. This leads to low production efficiency, unstable quality, and unpredictable tool life, limiting the application range and production capacity of lathes.

[0003] Currently, CNC technology has been applied to the automated control of lathes, achieving precise control of cutting parameters through digital means, significantly improving the accuracy and automation level of lathe processing. Modern CNC lathes are usually equipped with multi-axis control systems that can automatically adjust parameters such as feed speed and cutting depth. However, existing CNC systems still have some shortcomings, such as the inability to monitor the health of the tool in real time, the lack of a dynamic assessment mechanism for tool wear, and the difficulty in adjusting cutting parameters in real time during the processing process to cope with changing processing conditions. In addition, most existing CNC systems rely on preset processing templates and lack adaptive adjustment functions, resulting in the inability to optimize the efficiency and accuracy of the processing process.

[0004] The shortcomings of traditional CNC lathes primarily lie in their ability to monitor tool condition and dynamically adjust tool conditions. Due to a lack of real-time analysis of tool health, tool wear often goes undetected during production, leading to reduced machining quality or tool damage, impacting production efficiency. Furthermore, existing CNC lathes rely solely on preset fixed parameters and machining templates when dynamically adjusting machining speed and cutting parameters. This makes it difficult to adjust to the actual machining task and tool wear, resulting in unstable machining accuracy and shortened tool life. Summary of the Invention

[0005] A lathe automatic machining control method based on a numerical control device comprises the following steps:

[0006] Step S1: Data acquisition, where the production data during the automated machining process is acquired through sensors on the lathe. The production data includes the temperature of the tool, the frequency, amplitude, and waveform of the tool vibration, the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture, an image of the tool, and the speed and position of the spindle.

[0007] Step S2: Automatic control, obtaining the cutting parameters of the workpiece processing template in time sequence, and controlling the position of the tool through the processing control system composed of a single chip microcomputer; adjusting the cutting parameters of the tool through real-time production data feedback;

[0008] Step S3: Tool analysis: inputting production data into the tool condition analysis model for calculation, and outputting a tool health assessment report, including estimated life, processing compatibility, and recommended tools, for prompting tool replacement. The tool replacement is automatically performed by a four-position automatic tool holder;

[0009] Step S4: Dynamic control, based on the workpiece processing task data and combined with the tool's historical wear data, dynamically controls the workpiece processing speed;

[0010] Step S5: Production control, automatically scheduling processing tasks according to the priority of workpiece processing tasks and production resource conditions, updating the production schedule in real time, optimizing production resource allocation, and improving production efficiency.

[0011] As a preferred technical solution of the present invention, the production data includes:

[0012] The temperature of the tool is obtained by the temperature sensor installed on the tool and stored in the form of real-time temperature value;

[0013] The frequency, amplitude and waveform of the tool vibration are obtained by a vibration sensor installed on the tool and stored in the form of an array of vibration frequency, amplitude and waveform parameters;

[0014] The pressure at the connection between the tool and the workpiece is obtained by a pressure sensor installed at the connection between the tool and the workpiece and stored in the form of a time series of pressure values;

[0015] The pressure at the connection between the tool and the fixture is obtained by a pressure sensor installed at the connection between the tool and the fixture and stored in the form of a time series of pressure values;

[0016] The image of the tool is acquired by an image acquisition sensor installed above or on the side of the lathe and stored in the form of image frames, with each frame being image data;

[0017] The speed and position of the spindle are obtained through the photoelectric encoder installed on the spindle. The speed data is stored in the form of a time series of speed values, and the position data is stored in the form of a time series of position values;

[0018] The above data is stored as time series data in a structured format, with the timestamp as the key field corresponding to the data item of each sensor; the corresponding position of each data item in the time series is used for subsequent data analysis, prediction and feedback adjustment.

[0019] As a preferred technical solution of the present invention, the position control includes:

[0020] Obtaining time-series cutting parameters in a machining template of a workpiece, wherein the cutting parameters include spindle speed, spindle position, feed rate, and cutting depth, and the cutting parameters are arranged in time sequence according to the requirements of the machining path;

[0021] The machining template also includes standard feedback data corresponding to the sequential cutting parameters, including the frequency, amplitude and waveform of the tool vibration, and standard values ​​of the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture;

[0022] Through the processing control system composed of a single-chip microcomputer, the position and motion trajectory of the tool are controlled in real time according to the timing cutting parameters and standard feedback data in the processing template to ensure that the tool cuts according to the specified processing path;

[0023] The position control adjusts the position of the tool according to the standard values ​​of the spindle speed and the spindle position to achieve a precise cutting operation.

[0024] As a preferred technical solution of the present invention, the feedback regulation includes:

[0025] By comparing real-time production data with the standard feedback data in the current corresponding processing template, it is determined whether the processing is in place; if the processing is not in place, the tool will continue to cut at the original position until it meets the standard;

[0026] By taking the difference between real-time production data and standard feedback data and performing weighted summation, anomaly scores are calculated for anomaly assessment.

[0027] The anomaly score is calculated by taking the difference and weighted sum of the data items of the real-time production data with standard feedback data, and using the remaining data items in the real-time production data to obtain the deviation weights to calculate the comprehensive anomaly score. , the calculation formula is: ,in is the total number of the remaining data items, is the total number of data items of real-time production data for which standard feedback data exists, For the i The remaining data items, is the corresponding normal standard value; For the j There are data items of real-time production data with standard feedback data, Feedback data for the corresponding standards;

[0028] If the comprehensive abnormality score exceeds the preset step threshold, it is judged as poor processing or abnormal processing, and the system will make corresponding adjustments; for poor processing, the cutting parameters will be adjusted; for abnormal processing, the processing will be stopped and an alarm will be issued;

[0029] Abnormal assessment also includes setting abnormal extreme values ​​for each data item. If the deviation value of the corresponding data item exceeds the set single abnormal extreme value, it is also judged as abnormal processing.

[0030] As a preferred technical solution of the present invention, the structure of the tool condition analysis model includes:

[0031] Image analysis layer: used to process tool image data. The convolutional neural network analyzes the tool surface image and identifies the tool's physical condition. Image data processing includes image preprocessing, feature extraction, image classification, and defect detection. It is used to identify the tool's external damage and deformation and provide a visual assessment of the tool's health. The tool's physical condition includes wear, damage, and cracks.

[0032] Vibration and pressure data analysis layer: This layer is used to extract vibration characteristics corresponding to the frequency, amplitude, and waveform of tool vibration, and pressure characteristics corresponding to the pressure data at the connection between the tool, workpiece, and fixture. Vibration characteristics are obtained through fast Fourier transform; pressure data is obtained through data fluctuation analysis. This layer provides predictions of tool failure or excessive wear.

[0033] Data fusion layer: This layer fuses the results of the image analysis layer and the vibration and pressure data analysis layer, combining the analysis results of each data to comprehensively evaluate the overall condition of the tool. By weighted fusion of the outputs of the two analysis layers, the tool condition is quantified into a tool health score.

[0034] Life prediction and matching analysis layer: This layer uses regression analysis to calculate tool life estimates based on the fused tool health score, combined with historical data and the time-series cutting parameters of the workpiece. It also evaluates the matching degree between the current tool and the workpiece, determines whether the tool is suitable for the current machining task, and outputs the decision basis for recommending tool replacement.

[0035] Decision output layer: used to output tool health status assessment report based on life prediction and matching analysis results, including tool estimated life, processing matching and recommended tools for reference by system control and operators.

[0036] As a preferred technical solution of the present invention, the training of the tool condition analysis model includes:

[0037] Dataset construction: Historical production data is obtained and labeled according to the actual condition of the tool to construct a labeled training dataset. The training dataset includes input data and corresponding tool health labels for supervised learning. Before use, the training dataset is divided into training, validation, and test sets.

[0038] Data preprocessing and feature extraction: For image data, convolutional neural networks are used for image processing, including image enhancement, denoising, and normalization operations, to extract the surface features of the tool. For vibration and pressure data, fast Fourier transform is used to extract frequency domain features and analyze the fluctuations of vibration and pressure signals.

[0039] Training process: A structure combining a deep convolutional neural network and a long short-term memory network is used for training using the training set. The deep convolutional neural network is used to process image data and extract the surface features of the tool from the image; the long short-term memory network is used to process the time series data of vibration and pressure and capture the dynamic characteristics of the data over time. The loss function of the entire network is defined as: ,in is the predicted value of the output, is the corresponding tool health label, is the total number of samples is the weight parameter, is the regularization coefficient, which is used to prevent overfitting and optimize the network parameters by minimizing the loss function;

[0040] Model evaluation and adjustment: During the training process, the model performance is comprehensively evaluated by combining cross-validation on the validation set, accuracy and precision. Based on the training results, the hyperparameters of the tool condition analysis model during the training process are analyzed to further optimize the network structure and performance.

[0041] Model testing and validation: After training is complete, the trained model is validated using the test set. The model's prediction accuracy is evaluated by comparing the predicted results with the actual labels, and subsequent adjustments are made to address the errors. Ultimately, a tool condition analysis model is developed that can comprehensively assess tool health based on production data.

[0042] As a preferred technical solution of the present invention, the dynamic control includes:

[0043] Machining speed control: Machining speed control is achieved through a speed stage table, which includes different preset speed stages. The speed stages are dynamically adjusted based on the workpiece processing task data and the tool wear history data. Specifically, the matching speed stage is selected according to the workpiece processing type, material properties and tool wear, and the speed stage is maintained within the dynamic range.

[0044] Speed ​​stage adjustment: According to the workpiece processing requirements and tool health assessment report, select the most appropriate speed range from the stage table and adjust the speed stage to optimize the cutting process;

[0045] Application of historical wear data: During the dynamic control process, the historical wear data of the tool is obtained in real time, including the tool wear curve and over-cutting conditions. By comparing the historical wear data with the real-time processing conditions, the speed stage is adjusted to adapt to the current processing task requirements of the tool and workpiece.

[0046] As a preferred technical solution of the present invention, the production control includes:

[0047] Task priority determination: Comprehensively evaluate the priority of each task based on the type, urgency, and production goals of the workpiece processing task; allocate resources based on priority to ensure that critical or urgent tasks are handled first and avoid production bottlenecks caused by task delays;

[0048] Production resource status monitoring: Real-time monitoring of resource status on the production line, including equipment operating status, tool health, and workpiece supply. By collecting real-time data on various resources, assess whether production resources are sufficient and whether there is any waste or shortage of resources.

[0049] Automatic scheduling of processing tasks: Automatically schedule processing tasks based on the priority of workpiece processing tasks, production resource status and historical processing data; dynamically adjust processing sequence and task allocation based on real-time data and task priority;

[0050] Real-time update of production schedule: Dynamically update the production schedule based on the completion progress of processing tasks and the real-time status of the production line to ensure the accuracy and timeliness of scheduling information; optimize production plans by adjusting the start and completion times of processing tasks to avoid conflicts between tasks and waste of resources;

[0051] Optimize production resource allocation: Based on real-time updates of production scheduling, adjust the allocation plan of production resources to ensure the efficient configuration and utilization of equipment, tools, and workpiece resources.

[0052] A lathe automatic machining control system based on a numerical control device includes the following modules:

[0053] Data acquisition module: used to obtain production data in the automated processing process through sensors on the lathe;

[0054] Automatic control module: used to control the tool position by combining the cutting parameters with the machining control system composed of a single-chip microcomputer; and to adjust the tool cutting parameters through real-time production data feedback;

[0055] Tool analysis module: This module is used to input production data into the tool condition analysis model for calculation and output a tool health assessment report, including estimated tool life, processing compatibility, and recommended tools. This module is used to remind users to replace tools. Tool replacement is automatically performed by the four-position automatic tool holder.

[0056] Dynamic control module: Based on the workpiece processing task data and the historical wear data of the tool, the processing speed of the workpiece is dynamically controlled;

[0057] Production control module: used to automatically schedule processing tasks according to the priority of workpiece processing tasks and production resource conditions, update the production schedule in real time, optimize production resource allocation, and improve production efficiency.

[0058] The present invention has the following advantages:

[0059] The present invention integrates multiple sensors to comprehensively collect production data, covering multiple indicators. Compared with traditional lathes that only rely on a small amount of data for processing control, the collected multi-dimensional data can provide more accurate and richer information for subsequent analysis and decision-making, ensuring the accuracy and quality of the processing process.

[0060] The automatic control system of the present invention can dynamically adjust the cutting parameters of the tool based on real-time production data feedback. In this way, the lathe can respond to changes in the workpiece processing conditions in real time, accurately control the cutting depth and feed speed during the processing, effectively improve the processing accuracy and efficiency, and avoid errors caused by human intervention.

[0061] Through the tool condition analysis model, combined with multi-dimensional real-time production data, the present invention can evaluate the health of the tool in real time and predict the remaining life of the tool. This intelligent analysis can promptly detect problems such as tool wear and damage, avoiding inaccurate processing or production interruptions caused by tool failure, thereby effectively extending the tool life and improving processing stability.

[0062] Based on the workpiece processing task data and the historical tool wear data, the present invention dynamically adjusts the processing speed and the tool usage parameters to ensure the optimal cutting efficiency and tool life during the processing. By flexibly adjusting the speed stage and cutting parameters during the processing, the processing accuracy is maximized while reducing tool wear and improving the overall production efficiency.

[0063] The production control module of the present invention can automatically schedule processing tasks according to the priority of workpiece processing tasks and the production resource status, and update the production schedule in real time to optimize the allocation of production resources. This not only improves production efficiency, but also ensures the optimal utilization of resources, avoids production bottlenecks and resource waste, and further improves the flexibility and automation level of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only schematic diagrams of the present invention. Those skilled in the art can also derive other drawings based on the provided drawings without inventive effort.

[0065] Figure 1 This is a schematic structural diagram of a lathe automated machining control system based on a numerical control device adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] To make the objectives, technical solutions, and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] Example 1, a lathe automated machining control method based on a numerical control device, comprising the following steps:

[0068] Step S1: Data acquisition, where the production data during the automated machining process is acquired through sensors on the lathe. The production data includes the temperature of the tool, the frequency, amplitude, and waveform of the tool vibration, the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture, an image of the tool, and the speed and position of the spindle.

[0069] The production data includes:

[0070] The temperature of the tool is obtained by the temperature sensor installed on the tool and stored in the form of real-time temperature value;

[0071] The frequency, amplitude and waveform of the tool vibration are obtained by a vibration sensor installed on the tool and stored in the form of an array of vibration frequency, amplitude and waveform parameters;

[0072] The pressure at the connection between the tool and the workpiece is obtained by a pressure sensor installed at the connection between the tool and the workpiece and stored in the form of a time series of pressure values;

[0073] The pressure at the connection between the tool and the fixture is obtained by a pressure sensor installed at the connection between the tool and the fixture and stored in the form of a time series of pressure values;

[0074] The image of the tool is acquired by an image acquisition sensor installed above or on the side of the lathe and stored in the form of image frames, with each frame being image data;

[0075] The speed and position of the spindle are obtained through the photoelectric encoder installed on the spindle. The speed data is stored in the form of a time series of speed values, and the position data is stored in the form of a time series of position values;

[0076] The above data is stored as time series data in a structured format, with the timestamp as the key field corresponding to the data item of each sensor; the corresponding position of each data item in the time series is used for subsequent data analysis, prediction and feedback adjustment.

[0077] Step S2: Automatic control, obtaining the cutting parameters of the workpiece processing template in time sequence, and controlling the position of the tool through the processing control system composed of a single chip microcomputer; adjusting the cutting parameters of the tool through real-time production data feedback;

[0078] The position control includes:

[0079] Obtaining time-series cutting parameters in a machining template of a workpiece, wherein the cutting parameters include spindle speed, spindle position, feed rate, and cutting depth, and the cutting parameters are arranged in time sequence according to the requirements of the machining path;

[0080] The machining template also includes standard feedback data corresponding to the sequential cutting parameters, including the frequency, amplitude and waveform of the tool vibration, and standard values ​​of the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture;

[0081] Through the processing control system composed of a single-chip microcomputer, the position and motion trajectory of the tool are controlled in real time according to the timing cutting parameters and standard feedback data in the processing template to ensure that the tool cuts according to the specified processing path;

[0082] The position control adjusts the position of the tool according to the standard values ​​of the spindle speed and the spindle position to achieve a precise cutting operation.

[0083] In actual use, a stepper motor was installed on the existing lathe and connected to the X, Y, and Z axes via a gear reduction mechanism. The stepper motor's drive precisely controls motion in all three directions, enabling digital control of feed motion, which previously relied on manual operation. By installing a high-precision stepper motor and reduction mechanism, more precise longitudinal feed motion is achieved, improving overall machining accuracy.

[0084] A ball screw was installed in the lathe's longitudinal feed section and connected to a stepper motor via a coupling. The ball screw provides low-friction, smooth, and efficient motion during the feed process, ensuring higher repeatability and a smoother motion trajectory during machining. This modification not only improved feed accuracy but also enhanced the lathe's dynamic response.

[0085] To achieve high-precision thread machining, the spindle is equipped with a photoelectric encoder. This encoder precisely detects the spindle's rotational position and speed, and uses feedback signals to digitally control the spindle speed. During thread machining, the encoder provides accurate speed adjustment and position feedback, resulting in more precise thread cutting.

[0086] An interface conversion circuit and a high-voltage control circuit were added to the lathe's electrical control box. These modifications enabled the original lathe to connect to the new CNC system. The interface conversion circuit converts the CNC system's control signals into signals recognizable by the lathe, ensuring the system can effectively drive devices such as the stepper motor and automatic tool holder. The high-voltage control circuit is responsible for controlling the power supply to key components such as the lathe's main power supply and spindle, ensuring efficient operation.

[0087] The feedback regulation includes:

[0088] By comparing real-time production data with the standard feedback data in the current corresponding processing template, it is determined whether the processing is in place; if the processing is not in place, the tool will continue to cut at the original position until it meets the standard;

[0089] By taking the difference between real-time production data and standard feedback data and performing weighted summation, anomaly scores are calculated for anomaly assessment.

[0090] The anomaly score is calculated by taking the difference and weighted sum of the data items of the real-time production data with standard feedback data, and using the remaining data items in the real-time production data to obtain the deviation weights to calculate the comprehensive anomaly score. , the calculation formula is: ,in is the total number of the remaining data items, is the total number of data items of real-time production data for which standard feedback data exists, For the i The remaining data items, is the corresponding normal standard value; For the j There are data items of real-time production data with standard feedback data, Feedback data for the corresponding standards;

[0091] If the comprehensive abnormality score exceeds the preset step threshold, it is judged as poor processing or abnormal processing, and the system will make corresponding adjustments; for poor processing, the cutting parameters will be adjusted; for abnormal processing, the processing will be stopped and an alarm will be issued;

[0092] Abnormal assessment also includes setting abnormal extreme values ​​for each data item. If the deviation value of the corresponding data item exceeds the set single abnormal extreme value, it is also judged as abnormal processing.

[0093] Step S3: Tool analysis: inputting production data into the tool condition analysis model for calculation, and outputting a tool health assessment report, including estimated life, processing compatibility, and recommended tools, for prompting tool replacement. The tool replacement is automatically performed by a four-position automatic tool holder;

[0094] In actual use, tool changes were automated by replacing the original manual tool holder with a four-position automatic tool holder. The four-position automatic tool holder has multiple tool mounting positions and automatically switches tools based on processing requirements, eliminating manual tool changes and improving processing efficiency. This modification shortens tool change time and reduces machining inaccuracies caused by errors during tool changes.

[0095] The structure of the tool condition analysis model includes:

[0096] Image analysis layer: used to process tool image data. The convolutional neural network analyzes the tool surface image and identifies the tool's physical condition. Image data processing includes image preprocessing, feature extraction, image classification, and defect detection. It is used to identify the tool's external damage and deformation and provide a visual assessment of the tool's health. The tool's physical condition includes wear, damage, and cracks.

[0097] Vibration and pressure data analysis layer: This layer is used to extract vibration characteristics corresponding to the frequency, amplitude, and waveform of tool vibration, and pressure characteristics corresponding to the pressure data at the connection between the tool, workpiece, and fixture. Vibration characteristics are obtained through fast Fourier transform; pressure data is obtained through data fluctuation analysis. This layer provides predictions of tool failure or excessive wear.

[0098] Data fusion layer: This layer fuses the results of the image analysis layer and the vibration and pressure data analysis layer, combining the analysis results of each data to comprehensively evaluate the overall condition of the tool. By weighted fusion of the outputs of the two analysis layers, the tool condition is quantified into a tool health score.

[0099] Life prediction and matching analysis layer: This layer uses regression analysis to calculate tool life estimates based on the fused tool health score, combined with historical data and the time-series cutting parameters of the workpiece. It also evaluates the matching degree between the current tool and the workpiece, determines whether the tool is suitable for the current machining task, and outputs the decision basis for recommending tool replacement.

[0100] The decision output layer generates a tool health assessment report based on life prediction and tool compatibility analysis results. This report includes the tool's estimated lifespan, tool compatibility, and recommended tools for system control and operator reference. This layer provides tool replacement recommendations based on the predictions, which are implemented using a four-position automatic toolholder. The tool condition analysis model combines image data with other production data (such as vibration and pressure) using a hierarchical structure for separate data analysis and processing. Image and data analysis are performed independently at different levels, and the data fusion layer integrates the analysis results from these different data sources to ensure accurate and efficient tool condition analysis. The model's hierarchical design meets the functional requirements of real-world applications, providing accurate tool condition assessments and scientific predictions of tool life and replacement timing.

[0101] The training of the tool condition analysis model includes:

[0102] Dataset construction: Historical production data is obtained and labeled according to the actual condition of the tool to construct a labeled training dataset. The training dataset includes input data and corresponding tool health labels for supervised learning. Before use, the training dataset is divided into training, validation, and test sets.

[0103] Data preprocessing and feature extraction: For image data, convolutional neural networks are used for image processing, including image enhancement, denoising, and normalization, to extract the surface features of the tool. For vibration and pressure data, fast Fourier transform is used to extract frequency domain features and analyze the fluctuations of vibration and pressure signals.

[0104] Training process: A structure combining a deep convolutional neural network and a long short-term memory network is used for training using the training set. The deep convolutional neural network is used to process image data and extract the surface features of the tool from the image; the long short-term memory network is used to process the time series data of vibration and pressure and capture the dynamic characteristics of the data over time. The loss function of the entire network is defined as: ,in is the predicted value of the output, is the corresponding tool health label, is the total number of samples is the weight parameter, is the regularization coefficient, which is used to prevent overfitting and optimize the network parameters by minimizing the loss function;

[0105] Model evaluation and adjustment: During the training process, the model performance is comprehensively evaluated by combining cross-validation on the validation set, accuracy and precision. Based on the training results, the hyperparameters of the tool condition analysis model during the training process are analyzed to further optimize the network structure and performance.

[0106] Model testing and validation: After training is complete, the trained model is validated using the test set. The model's prediction accuracy is evaluated by comparing the predicted results with the actual labels, and subsequent adjustments are made to address the errors. Ultimately, a tool condition analysis model is developed that can comprehensively assess tool health based on production data.

[0107] Through the complete process of dataset construction, data preprocessing, feature extraction, network training, model evaluation and verification, a network structure combining CNN and LSTM is adopted to ensure the joint analysis of image data and time series data, thereby optimizing the performance of the tool condition analysis model.

[0108] Step S4: Dynamic control, based on the workpiece processing task data and combined with the tool's historical wear data, dynamically controls the workpiece processing speed;

[0109] The dynamic control includes:

[0110] Machining speed control: Machining speed is controlled through a speed stage table, which includes preset speed stages. Based on the workpiece processing task data and the tool wear history, the speed stage is dynamically adjusted to ensure optimal cutting efficiency and tool life during the machining process. Specifically, the matching speed stage is selected based on the workpiece processing type, material properties and tool wear, and the speed stage is maintained within the dynamic range.

[0111] Speed ​​stage adjustment: Based on the workpiece processing requirements and the tool health assessment report, the most appropriate speed range is selected from the stage table and the speed stage is adjusted to optimize the cutting process. In this way, the speed stage changes dynamically during the processing without adjusting the overall processing speed, thereby maximizing processing accuracy and extending tool life.

[0112] Application of historical wear data: During the dynamic control process, the historical wear data of the tool is obtained in real time, including the tool wear curve and over-cutting conditions. By comparing the historical wear data with the real-time processing conditions, the speed stage is adjusted to adapt to the current processing task requirements of the tool and workpiece.

[0113] Step S5: Production control, automatically scheduling processing tasks according to the priority of workpiece processing tasks and production resource conditions, updating the production schedule in real time, optimizing production resource allocation, and improving production efficiency.

[0114] The production control includes:

[0115] Task priority determination: Comprehensively evaluate the priority of each task based on the type, urgency, and production goals of the workpiece processing task; allocate resources based on priority to ensure that critical or urgent tasks are handled first and avoid production bottlenecks caused by task delays;

[0116] Production resource status monitoring: Real-time monitoring of resource status on the production line, including equipment operating status, tool health, and workpiece supply. By collecting real-time data on various resources, assess whether production resources are sufficient and whether there is any waste or shortage of resources.

[0117] Automatic scheduling of processing tasks: Automatically schedule processing tasks based on the priority of workpiece processing tasks, production resource status, and historical processing data; dynamically adjust processing sequence and task allocation based on real-time data and task priority to optimize resource utilization in the production process;

[0118] Real-time update of production schedule: Dynamically update the production schedule based on the completion progress of processing tasks and the real-time status of the production line to ensure the accuracy and timeliness of scheduling information; optimize production plans by adjusting the start and completion times of processing tasks to avoid conflicts between tasks and waste of resources;

[0119] Optimize production resource allocation: Based on real-time updates to the production schedule, adjust the allocation of production resources to ensure efficient allocation and utilization of resources such as equipment, tools, and workpieces. By optimizing resource allocation, idle time and equipment idleness are reduced, improving overall production efficiency and reducing production costs.

[0120] Example 2, a lathe automatic processing control system based on a numerical control device, see Figure 1 As shown, it includes the following modules:

[0121] Data acquisition module: used to obtain production data in the automated processing process through sensors on the lathe;

[0122] Automatic control module: used to control the tool position by combining the cutting parameters with the machining control system composed of a single-chip microcomputer; and to adjust the tool cutting parameters through real-time production data feedback;

[0123] Tool analysis module: This module is used to input production data into the tool condition analysis model for calculation and output a tool health assessment report, including estimated tool life, processing compatibility, and recommended tools. This module is used to remind users to replace tools. Tool replacement is automatically performed by the four-position automatic tool holder.

[0124] Dynamic control module: Based on the workpiece processing task data and the historical wear data of the tool, the processing speed of the workpiece is dynamically controlled;

[0125] Production control module: used to automatically schedule processing tasks according to the priority of workpiece processing tasks and production resource conditions, update the production schedule in real time, optimize production resource allocation, and improve production efficiency.

[0126] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A lathe automatic processing control method based on a numerical control device, characterized in that: The following steps are involved: Step S1: Data acquisition, where the production data during the automated machining process is acquired through sensors on the lathe. The production data includes the temperature of the tool, the frequency, amplitude, and waveform of the tool vibration, the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture, an image of the tool, and the speed and position of the spindle. Step S2: Automatic control, obtaining the cutting parameters of the workpiece processing template in time sequence, and controlling the position of the tool through the processing control system composed of a single chip microcomputer; adjusting the cutting parameters of the tool through real-time production data feedback; Step S3: Tool analysis: inputting production data into the tool condition analysis model for calculation, and outputting a tool health assessment report, including estimated life, processing compatibility, and recommended tools, for prompting tool replacement. The tool replacement is automatically performed by a four-position automatic tool holder; Step S4: Dynamic control, based on the workpiece processing task data and combined with the tool's historical wear data, dynamically controls the workpiece processing speed; The dynamic control includes: Machining speed control: Machining speed control is achieved through a speed stage table, which includes different preset speed stages. The speed stages are dynamically adjusted based on the workpiece processing task data and the tool wear history data. Specifically, the matching speed stage is selected according to the workpiece processing type, material properties and tool wear, and the speed stage is maintained within the dynamic range. Speed ​​stage adjustment: According to the workpiece processing requirements and tool health assessment report, select the most appropriate speed range from the stage table and adjust the speed stage to optimize the cutting process; Application of historical wear data: During dynamic control, historical tool wear data, including tool wear curves and overcutting conditions, is acquired in real time. By comparing historical wear data with real-time machining conditions, the speed stage is adjusted to suit the machining requirements of the current tool and workpiece. Step S5: Production control, automatically scheduling processing tasks according to the priority of workpiece processing tasks and production resource conditions, updating the production schedule in real time, optimizing production resource allocation, and improving production efficiency.

2. The method for controlling automatic machining of a lathe based on a numerical control device according to claim 1, characterized in that: The production data includes: The temperature of the tool is obtained by the temperature sensor installed on the tool and stored in the form of real-time temperature value; The frequency, amplitude and waveform of the tool vibration are obtained by a vibration sensor installed on the tool and stored in the form of an array of vibration frequency, amplitude and waveform parameters; The pressure at the connection between the tool and the workpiece is obtained by a pressure sensor installed at the connection between the tool and the workpiece and stored in the form of a time series of pressure values; The pressure at the connection between the tool and the fixture is obtained by a pressure sensor installed at the connection between the tool and the fixture and stored in the form of a time series of pressure values; The image of the tool is acquired by an image acquisition sensor installed above or on the side of the lathe and stored in the form of image frames, with each frame being image data; The speed and position of the spindle are obtained through the photoelectric encoder installed on the spindle. The speed data is stored in the form of a time series of speed values, and the position data is stored in the form of a time series of position values; The above data is stored as time series data in a structured format, with the timestamp as the key field corresponding to the data item of each sensor; the corresponding position of each data item in the time series is used for subsequent data analysis, prediction and feedback adjustment.

3. The lathe automatic machining control method based on a numerical control device according to claim 1, characterized in that: The position control includes: Obtaining time-series cutting parameters in a machining template of a workpiece, wherein the cutting parameters include spindle speed, spindle position, feed rate, and cutting depth, and the cutting parameters are arranged in time sequence according to the requirements of the machining path; The machining template also includes standard feedback data corresponding to the sequential cutting parameters, including the frequency, amplitude and waveform of the tool vibration, and standard values ​​of the pressure at the connection between the tool and the workpiece and at the connection between the tool and the fixture; Through the processing control system composed of a single-chip microcomputer, the position and motion trajectory of the tool are controlled in real time according to the timing cutting parameters and standard feedback data in the processing template to ensure that the tool cuts according to the specified processing path; The position control adjusts the position of the tool according to the standard values ​​of the spindle speed and the spindle position to achieve a precise cutting operation.

4. The method for controlling automatic machining of a lathe based on a numerical control device according to claim 1, characterized in that: The feedback regulation includes: By comparing real-time production data with the standard feedback data in the current corresponding processing template, it is determined whether the processing is in place; if the processing is not in place, the tool will continue to cut at the original position until it meets the standard; By taking the difference between real-time production data and standard feedback data and performing weighted summation, anomaly scores are calculated for anomaly assessment. The anomaly score is calculated by taking the difference and weighted sum of the data items of the real-time production data with standard feedback data, and using the remaining data items in the real-time production data to obtain the deviation weights to calculate the comprehensive anomaly score. , the calculation formula is: ,in is the total number of the remaining data items, is the total number of data items of real-time production data for which standard feedback data exists, For the i The remaining data items, is the corresponding normal standard value; For the j There are data items of real-time production data with standard feedback data, Feedback data for the corresponding standards; If the comprehensive abnormality score exceeds the preset step threshold, it is judged as poor processing or abnormal processing, and the system will make corresponding adjustments; for poor processing, the cutting parameters will be adjusted; for abnormal processing, the processing will be stopped and an alarm will be issued; Abnormal assessment also includes setting abnormal extreme values ​​for each data item. If the deviation value of the corresponding data item exceeds the set single abnormal extreme value, it is also judged as abnormal processing.

5. The lathe automatic machining control method based on a numerical control device according to claim 1, characterized in that: The structure of the tool condition analysis model includes: Image analysis layer: used to process tool image data. The convolutional neural network analyzes the tool surface image and identifies the tool's physical condition. Image data processing includes image preprocessing, feature extraction, image classification, and defect detection. It is used to identify the tool's external damage and deformation and provide a visual assessment of the tool's health. The tool's physical condition includes wear, damage, and cracks. Vibration and pressure data analysis layer: This layer is used to extract vibration characteristics corresponding to the frequency, amplitude, and waveform of tool vibration, and pressure characteristics corresponding to the pressure data at the connection between the tool, workpiece, and fixture. Vibration characteristics are obtained through fast Fourier transform; pressure data is obtained through data fluctuation analysis. This layer provides predictions of tool failure or excessive wear. Data fusion layer: This layer fuses the results of the image analysis layer and the vibration and pressure data analysis layer, combining the analysis results of each data to comprehensively evaluate the overall condition of the tool. By weighted fusion of the outputs of the two analysis layers, the tool condition is quantified into a tool health score. Life prediction and matching analysis layer: This layer uses regression analysis to calculate tool life estimates based on the fused tool health score, combined with historical data and the time-series cutting parameters of the workpiece. It also evaluates the matching degree between the current tool and the workpiece, determines whether the tool is suitable for the current machining task, and outputs the decision basis for recommending tool replacement. Decision output layer: used to output tool health status assessment report based on life prediction and matching analysis results, including tool estimated life, processing matching and recommended tools for reference by system control and operators.

6. The method for controlling automatic machining of a lathe based on a numerical control device according to claim 5, characterized in that: The training of the tool condition analysis model includes: Dataset construction: Historical production data is obtained and labeled according to the actual condition of the tool to construct a labeled training dataset. The training dataset includes input data and corresponding tool health labels for supervised learning. Before use, the training dataset is divided into training, validation, and test sets. Data preprocessing and feature extraction: For image data, convolutional neural networks are used for image processing, including image enhancement, denoising, and normalization operations, to extract the surface features of the tool. For vibration and pressure data, fast Fourier transform is used to extract frequency domain features and analyze the fluctuations of vibration and pressure signals. Training process: A structure combining a deep convolutional neural network and a long short-term memory network is used for training using the training set. The deep convolutional neural network is used to process image data and extract the surface features of the tool from the image; the long short-term memory network is used to process the time series data of vibration and pressure and capture the dynamic characteristics of the data over time. The loss function of the entire network is defined as: ,in is the predicted value of the output, is the corresponding tool health label, is the total number of samples is the weight parameter, is the regularization coefficient, which is used to prevent overfitting and optimize the network parameters by minimizing the loss function; Model evaluation and adjustment: During the training process, the model performance is comprehensively evaluated by combining cross-validation on the validation set, accuracy and precision. Based on the training results, the hyperparameters of the tool condition analysis model during the training process are analyzed to further optimize the network structure and performance. Model testing and validation: After training is complete, the trained model is validated using the test set. The model's prediction accuracy is evaluated by comparing the predicted results with the actual labels, and subsequent adjustments are made to address the errors. Ultimately, a tool condition analysis model is developed that can comprehensively assess tool health based on production data.

7. The method for controlling automatic machining of a lathe based on a numerical control device according to claim 1, characterized in that: The production control includes: Task priority determination: Comprehensively evaluate the priority of each task based on the type, urgency, and production goals of the workpiece processing task; allocate resources based on priority to ensure that critical or urgent tasks are handled first and avoid production bottlenecks caused by task delays; Production resource status monitoring: Real-time monitoring of resource status on the production line, including equipment operating status, tool health, and workpiece supply. By collecting real-time data on various resources, assess whether production resources are sufficient and whether there is any waste or shortage of resources. Automatic scheduling of processing tasks: Automatically schedule processing tasks based on the priority of workpiece processing tasks, production resource status and historical processing data; dynamically adjust processing sequence and task allocation based on real-time data and task priority; Real-time update of production schedule: Dynamically update the production schedule based on the completion progress of processing tasks and the real-time status of the production line to ensure the accuracy and timeliness of scheduling information; optimize production plans by adjusting the start and completion times of processing tasks to avoid conflicts between tasks and waste of resources; Optimize production resource allocation: Based on real-time updates of production scheduling, adjust the allocation plan of production resources to ensure the efficient configuration and utilization of equipment, tools, and workpiece resources.

8. A lathe automatic processing control system based on a numerical control device, characterized in that: The system applies a lathe automated machining control method based on a numerical control device as described in any one of claims 1 to 7, and includes the following modules: Data acquisition module: used to obtain production data in the automated processing process through sensors on the lathe; Automatic control module: used to control the tool position by combining the cutting parameters with the machining control system composed of a single-chip microcomputer; and to adjust the tool cutting parameters through real-time production data feedback; Tool analysis module: This module is used to input production data into the tool condition analysis model for calculation and output a tool health assessment report, including estimated tool life, processing compatibility, and recommended tools. This module is used to remind users to replace tools. Tool replacement is automatically performed by the four-position automatic tool holder. Dynamic control module: Based on the workpiece processing task data and the historical wear data of the tool, the processing speed of the workpiece is dynamically controlled; Production control module: used to automatically schedule processing tasks according to the priority of workpiece processing tasks and production resource conditions, update the production schedule in real time, optimize production resource allocation, and improve production efficiency.

Images