A method for the tool to quickly find the model coordinates during CNC 3-axis processing

Abstract

The invention discloses a method for a tool to quickly find model coordinates during a numerically controlled 3-axis process, and relates to the technical field related to computer-aided processing. The invention sets an initial coordinate value and a processing path and uses preset parameters for initial positioning, so as to accurately position the tool to the initial position of a workpiece. In the path planning stage, the CAD software is used to pre-design and generate the tool path, and the processing path of the tool is determined according to the geometric shape and processing requirements of the workpiece. The CAD software is used for path planning, and in the tool motion control and prediction stage, the preloaded workpiece model data and prediction algorithm are used to predict the coordinate points of the next tool positioning, so as to improve the accuracy and efficiency of the tool positioning. The process constructs a prediction model, inputs a workpiece surface model and current tool position characteristics, outputs a predicted value of the next tool position, and dynamically adjusts the prediction model according to real-time sensor data to adapt to the actual changes of the workpiece surface.

Description

Technical Field

[0001] The present invention relates to the technical field related to computer-aided machining, and in particular to a method for a tool to quickly find model coordinates during a CNC 3-axis machining process. Background Art

[0002] The CNC machining process includes the most common milling and turning, followed by grinding, EDM, etc. Milling is the use of a rotating tool on the surface of a workpiece, moving along 3, 4 or 5 axes. Milling is basically cutting or trimming the workpiece. Complex geometric shapes and precision parts can be quickly machined from metal or thermoplastics. Turning is the use of a lathe to manufacture parts containing cylindrical features.

[0003] For example, the patent publication number CN110096034B discloses a method for reconstructing the lateral information of a three-axis tool trajectory surface based on a projection algorithm. The reconstruction method includes the following steps: (1) recording the index of each tool position point according to the tool position file and storing it according to the KD-Tree structure; (2) finding the adjacent track projection point of each tool position point based on the data stored in step (1), and pairing and storing the adjacent track projection point with the corresponding tool position point, thereby obtaining the lateral information of the tool trajectory surface composed of all tool positions and the corresponding adjacent track projection points; wherein the adjacent track projection point is the point on the adjacent tool trajectory that is closest to the seed point, and the tool trajectory surface is a surface containing all tool positions in the tool trajectory. The invention does not need to reconstruct the surface, reduces the calculation burden, takes into account the lateral information of the tool trajectory, has a good optimization effect, and is highly applicable.

[0004] Although the above scheme has the above advantages, the method for the tool to quickly find the model coordinates during the traditional CNC 3-axis process relies on manual adjustment with limited accuracy and low efficiency. At the same time, once the tool positioning is completed, when the path adjustment is required during the processing stage, there is often a lack of real-time monitoring and adjustment mechanism, and it is impossible to respond to changes in the workpiece surface in time. Therefore, a method for the tool to quickly find the model coordinates during the CNC 3-axis process is urgently needed to solve this problem. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method for a tool to quickly find the model coordinates during a CNC 3-axis process machining process, thereby solving the problem that the traditional method for a tool to quickly find the model coordinates during a CNC 3-axis process machining process in the prior art relies on manual adjustment with limited accuracy and low efficiency. At the same time, once the tool positioning is completed, when path adjustment is required in the machining stage, there is often a lack of real-time monitoring and adjustment mechanism, and the change of the workpiece surface cannot be responded to in time.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] The present invention provides a method for a tool to quickly find model coordinates during a CNC 3-axis process, comprising:

[0008] Step 1. Initial positioning: set the initial coordinate value of the tool and the processing path; use the preset parameters for initial positioning to position the tool to the initial position of the workpiece;

[0009] Step 2. Path planning: pre-design and generate tool paths in CAD software, and determine the tool processing path according to the workpiece geometry and processing requirements, including cutting path, cutting direction and cutting depth parameters;

[0010] Step 3. Initial tool motion control and prediction, control the tool motion trajectory and speed, including the control of the tool's X, Y, and Z axis coordinates, preload the workpiece model into the CNC system before processing; use the preloaded model data and prediction algorithm to predict the coordinate point of the next tool positioning, and use the prediction result as the initial positioning;

[0011] Step 4. Real-time monitoring and positioning adjustment of the workpiece during the processing phase: deploy displacement sensors and cameras to monitor the surface of the workpiece in real time and collect sensor data. Dynamically adjust the prediction model based on the real-time data, adjust the position and posture of the tool in real time according to the actual situation of the workpiece surface and sensor data, and accurately align the tool with the workpiece surface.

[0012] Step 5. Redundant control: deploy a highly redundant and fault-tolerant system architecture, equipped with multiple backup and redundant control mechanisms, and automatically switch to backup mode under abnormal circumstances.

[0013] The present invention is further configured as follows: the initial positioning step in step 1 includes:

[0014] Set the initial coordinate value of the tool and the processing path, and set the initial coordinate value of the tool in the CNC system, including the X, Y, and Z axis coordinates;

[0015] Set the tool processing path, including cutting path, cutting direction and cutting depth parameters;

[0016] Use preset parameters for initial positioning, and position the tool in the CNC system according to the preset initial coordinate values;

[0017] Use the motion control function provided by the CNC system to move the tool to the preset initial position;

[0018] According to the workpiece CAD model, determine the initial processing position, use the CNC system to convert the coordinate system, and position the tool to the initial position of the workpiece;

[0019] At the same time, record the tool coordinate values ​​and machining path parameters used for initial positioning;

[0020] The present invention is further configured as follows: the path planning method in step 2 is:

[0021] Use CAD software to import the CAD model file of the workpiece;

[0022] According to the geometric shape and processing requirements of the workpiece, determine the basic requirements of the processing path, including surface smoothness and dimensional accuracy;

[0023] According to the workpiece material, processing requirements and tool performance, the corresponding processing technology is adopted, and the corresponding process function is used in the CAD software to draw the cutting path;

[0024] Determine the direction, shape and spacing of the cutting path according to the processing requirements and workpiece geometry;

[0025] Setting the cutting direction, including the feed direction and the cutting direction, based on the determined cutting path direction, shape and spacing;

[0026] The present invention is further configured as follows: the workpiece model building step comprises:

[0027] Use a 3D laser scanner to scan the surface of the workpiece to obtain 3D point cloud data of the surface;

[0028] Process the scanned point cloud data, including filtering and point cloud registration, to obtain a three-dimensional model of the workpiece surface;

[0029] Using point cloud data, a grid-based reconstruction algorithm is used to generate a three-dimensional surface model of the workpiece;

[0030] Construct a prediction model, the input parameters include , as well as ,

[0031] in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position;

[0032] The input of the prediction model is and , the output is the value ;

[0033] The prediction model is trained with processing data to learn the relationship between the workpiece surface and tool motion, using back propagation Algorithms and optimizers to optimize model parameters and reduce prediction errors;

[0034] The present invention is further configured as follows: in step 3, the prediction algorithm is constructed in the following manner:

[0035] Prepare training sets and validation sets, build a prediction model based on the long short-term memory network LSTM, and define input parameters, including , as well as ,in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position;

[0036] Use the back propagation algorithm and gradient descent optimization algorithm to train the neural network model, and the optimization loss function formula is: ,in is the actual tool position, is the sample size, represents the time step;

[0037] Use the validation set to validate the trained neural network model and adjust the model structure and hyperparameters;

[0038] Before machining, the feature vector of the tool’s current position is and 3D model of the workpiece surface Input the trained neural network model;

[0039] The model outputs the predicted value of the next tool position ;

[0040] The present invention is further configured as follows: the workpiece real-time monitoring and positioning adjustment step in step 4 includes:

[0041] Install displacement sensors and cameras on processing equipment to monitor the surface of the workpiece in real time; continuously monitor the surface of the workpiece, including surface shape, roughness, and temperature;

[0042] Collect sensor data, process and analyze it; dynamically adjust the prediction model based on real-time sensor data to predict the next tool position;

[0043] The present invention is further configured as follows: the workpiece real-time monitoring and positioning adjustment step in step 4 further includes:

[0044] Adopt adaptive algorithms to update model parameters according to new data to adapt to the actual changes of the workpiece surface;

[0045] According to the tool position prediction value output by the prediction model and the real-time sensor data, the tool position and posture are adjusted in real time;

[0046] Use the control functions provided by the CNC system to adjust the tool's machining path and motion trajectory to accurately align the tool with the workpiece surface;

[0047] According to the real-time adjusted tool position and posture, the tool is precisely aligned with the workpiece surface, and the tool position and posture are adjusted to make it fit the workpiece surface perfectly;

[0048] The present invention is further configured as follows: in step 4, the method for predicting the next tool position is:

[0049] Use displacement sensors and cameras to continuously monitor the surface conditions of the workpiece, collect relevant sensor data, and pre-process the collected data, including noise filtering, data alignment and calibration;

[0050] A neural network model based on deep learning is used to take real-time sensor data as the input feature of the model to predict the output value of the next tool position;

[0051] Use gradient descent to update model parameters based on the actual prediction error;

[0052] Define the loss function to measure the difference between the model's predicted value and the actual value;

[0053] According to the tool position prediction value output by the updated prediction model and combined with the real-time sensor data, the position and posture of the tool are adjusted;

[0054] The present invention is further configured as follows: the redundant control system architecture includes:

[0055] The main control module is responsible for monitoring the operating status of the processing equipment, controlling the processing process, and processing various instructions;

[0056] Redundant main control module for active / standby switching and fault-tolerant control;

[0057] The backup control module, as a backup of the main control module, keeps in sync with the main control module, is ready to take over control at any time, and communicates with the main control module in real time through a redundant communication link;

[0058] The power supply module provides the power supply required by the equipment. It is equipped with a redundant power supply module, which can automatically switch to the backup power supply when the main power supply fails;

[0059] The present invention is further configured as follows: the redundant control system architecture further includes:

[0060] Sensor modules are responsible for monitoring workpiece and equipment parameters, including temperature, pressure, and speed; they are equipped with redundant sensor modules to continue to acquire data in the event of a sensor failure;

[0061] The execution module controls the movement and posture of the tool;

[0062] The communication module is responsible for communicating with external devices and systems, including industrial computers and monitoring systems;

[0063] The fault detection and automatic switching module implements abnormality detection and automatic switching logic, monitors the operating status of each module of the system, and automatically switches to the backup mode when a fault is detected.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] The present invention realizes accurate prediction of tool position by preloading workpiece model and using prediction algorithm, thereby improving positioning accuracy and efficiency;

[0066] The present invention deploys displacement sensors and cameras to monitor the surface of the workpiece in real time and collect sensor data, and dynamically adjusts the position and posture of the tool in real time according to the real-time data to ensure accurate alignment of the tool with the workpiece surface;

[0067] The present invention is equipped with multiple backup and redundant control mechanisms to ensure that the system automatically switches to the standby mode under abnormal circumstances, thereby improving the stability and safety of the processing process;

[0068] The present invention sets the initial coordinate value and the processing path and uses the preset parameters for initial positioning, accurately positioning the tool to the initial position of the workpiece. In the path planning stage, the CAD software is used to pre-design and generate the tool path, and the processing path of the tool is determined according to the geometric shape of the workpiece and the processing requirements. The CAD software is used for path planning. In the tool motion control and prediction stage, the preloaded workpiece model data and prediction algorithm are used to predict the coordinate point of the next tool positioning, thereby improving the accuracy and efficiency of tool positioning. The process constructs a prediction model, inputs the workpiece surface model and the current tool position characteristics, outputs the predicted value of the next tool position, and dynamically adjusts the prediction model according to the real-time sensor data to adapt to the actual changes of the workpiece surface.

[0069] In the real-time monitoring and adjustment stage, displacement sensors and cameras are deployed to monitor the surface of the workpiece in real time and collect sensor data. The prediction model is dynamically adjusted according to the real-time data to accurately align the tool with the workpiece surface. The tool position prediction value output by the prediction model is combined with the real-time sensor data to dynamically adjust the position and posture of the tool to achieve accurate alignment between the tool and the workpiece surface.

[0070] Finally, in the redundant control and safety assurance stage, a highly redundant and fault-tolerant system architecture is deployed, equipped with multiple backup and redundant control mechanisms, which can automatically switch to backup mode or safe state under abnormal circumstances to ensure the stability and safety of the processing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The present invention is a flow chart of a method for a tool to quickly find model coordinates during a CNC 3-axis machining process. DETAILED DESCRIPTION

[0072] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0073] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0074] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0075] Example 1

[0076] See also Figure 1 The present invention provides a method for a tool to quickly find model coordinates during a CNC 3-axis process, comprising:

[0077] Step 1. Initial positioning: set the initial coordinate value of the tool and the processing path; use the preset parameters for initial positioning to position the tool to the initial position of the workpiece;

[0078] The initial positioning steps in step 1 include:

[0079] Set the initial coordinate value of the tool and the processing path, and set the initial coordinate value of the tool in the CNC system, including the X, Y, and Z axis coordinates;

[0080] Set the tool processing path, including cutting path, cutting direction and cutting depth parameters;

[0081] Use preset parameters for initial positioning, and position the tool in the CNC system according to the preset initial coordinate values;

[0082] Use the motion control function provided by the CNC system to move the tool to the preset initial position;

[0083] According to the workpiece CAD model, determine the initial processing position, use the CNC system to convert the coordinate system, and position the tool to the initial position of the workpiece;

[0084] At the same time, record the tool coordinate values ​​and machining path parameters used for initial positioning;

[0085] Implementing initial positioning through step 1 to accurately position the tool to the initial position of the workpiece;

[0086] Step 2. Path planning: pre-design and generate tool paths in CAD software, and determine the tool processing path according to the workpiece geometry and processing requirements, including cutting path, cutting direction and cutting depth parameters;

[0087] The path planning method in step 2 is:

[0088] Use CAD software to import the CAD model file of the workpiece;

[0089] According to the geometric shape and processing requirements of the workpiece, determine the basic requirements of the processing path, including surface smoothness and dimensional accuracy;

[0090] According to the workpiece material, processing requirements and tool performance, the corresponding processing technology is adopted, and the corresponding process function is used in the CAD software to draw the cutting path;

[0091] Determine the direction, shape and spacing of the cutting path according to the processing requirements and workpiece geometry;

[0092] Setting the cutting direction, including the feed direction and the cutting direction, based on the determined cutting path direction, shape and spacing;

[0093] Perform path planning in CAD software to determine the tool processing path;

[0094] Step 3. Initial tool motion control and prediction, control the tool motion trajectory and speed, including the control of the tool's X, Y, and Z axis coordinates, and preload the workpiece model into the CNC system before processing; use the preloaded model data and prediction algorithm to predict the coordinate point of the next tool positioning, and use the prediction result as the initial positioning to improve the accuracy and efficiency of tool positioning;

[0095] The steps of building the artifact model include:

[0096] Use a 3D laser scanner to scan the surface of the workpiece to obtain 3D point cloud data of the surface;

[0097] Process the scanned point cloud data, including filtering and point cloud registration, to obtain a three-dimensional model of the workpiece surface;

[0098] Using point cloud data, a grid-based reconstruction algorithm is used to generate a three-dimensional surface model of the workpiece;

[0099] Construct a prediction model, the input parameters include , as well as ,

[0100] in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position;

[0101] The input of the prediction model is and , the output is the value, ;

[0102] The prediction model is trained with machining data to learn the relationship between the workpiece surface and tool motion, and the model parameters are optimized using the back propagation algorithm and optimizer;

[0103] In step 3, the prediction algorithm is constructed as follows:

[0104] Prepare training sets and validation sets, build a prediction model based on the long short-term memory network LSTM, and define input parameters, including , as well as ,in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position;

[0105] Use the back propagation algorithm and gradient descent optimization algorithm to train the neural network model, and the optimization loss function formula is: ,in is the actual tool position, is the sample size, represents the time step;

[0106] Use the validation set to validate the trained neural network model and adjust the model structure and hyperparameters;

[0107] Before machining, the feature vector of the tool’s current position is and 3D model of the workpiece surface Input the trained neural network model;

[0108] The model outputs the predicted value of the next tool position ;

[0109] Use deep learning technology to build a prediction model to accurately predict the next tool positioning;

[0110] Step 4. Real-time monitoring and positioning adjustment of the workpiece during the processing phase: deploy displacement sensors and cameras to monitor the surface of the workpiece in real time and collect sensor data. Dynamically adjust the prediction model based on the real-time data, adjust the position and posture of the tool in real time according to the actual situation of the workpiece surface and sensor data, and accurately align the tool with the workpiece surface.

[0111] Specifically: Install displacement sensors and cameras on processing equipment to monitor the surface of the workpiece in real time; continuously monitor the surface of the workpiece, including surface shape, roughness, temperature,

[0112] Collect sensor data, process and analyze it; dynamically adjust the prediction model based on real-time sensor data to predict the next tool position;

[0113] Adopting adaptive algorithms to update model parameters according to new data, adapting to the actual changes of the workpiece surface, adjusting the position and posture of the tool in real time according to the tool position prediction value output by the prediction model and real-time sensor data, using the control function provided by the CNC system to adjust the processing path and motion trajectory of the tool so that the tool is accurately aligned with the workpiece surface, and adjusting the position and posture of the tool so that it fits the workpiece surface perfectly according to the real-time adjusted tool position and posture;

[0114] The next tool position is predicted as follows:

[0115] Use displacement sensors and cameras to continuously monitor the surface conditions of the workpiece, collect relevant sensor data, and pre-process the collected data, including noise filtering, data alignment and calibration;

[0116] A neural network model based on deep learning is used to take real-time sensor data as the input feature of the model to predict the output value of the next tool position;

[0117] Use gradient descent to update model parameters based on the actual prediction error;

[0118] Define the loss function to measure the difference between the model's predicted value and the actual value;

[0119] According to the tool position prediction value output by the updated prediction model and combined with the real-time sensor data, the position and posture of the tool are adjusted;

[0120] Step 5. Redundancy control: deploy a highly redundant and fault-tolerant system architecture, equipped with multiple backup and redundant control mechanisms, and automatically switch to backup mode in abnormal situations;

[0121] The redundant control system architecture includes:

[0122] The main control module is responsible for monitoring the operating status of the processing equipment, controlling the processing process, and processing various instructions;

[0123] Redundant main control module for active / standby switching and fault-tolerant control;

[0124] The backup control module, as a backup of the main control module, keeps in sync with the main control module, is ready to take over control at any time, and communicates with the main control module in real time through a redundant communication link;

[0125] The power supply module provides the power supply required by the equipment. It is equipped with a redundant power supply module, which can automatically switch to the backup power supply when the main power supply fails;

[0126] The redundant control system architecture also includes:

[0127] Sensor modules are responsible for monitoring workpiece and equipment parameters, including temperature, pressure, and speed; they are equipped with redundant sensor modules to continue to acquire data in the event of a sensor failure;

[0128] The execution module controls the movement and posture of the tool;

[0129] The communication module is responsible for communicating with external devices and systems, including industrial computers and monitoring systems;

[0130] The fault detection and automatic switching module implements abnormality detection and automatic switching logic, monitors the operating status of each module of the system, and automatically switches to the backup mode when a fault is detected.

[0131] The present invention sets the initial coordinate value and the processing path and uses the preset parameters for initial positioning, accurately positioning the tool to the initial position of the workpiece. In the path planning stage, the CAD software is used to pre-design and generate the tool path, and the processing path of the tool is determined according to the geometric shape of the workpiece and the processing requirements. The CAD software is used for path planning. In the tool motion control and prediction stage, the preloaded workpiece model data and prediction algorithm are used to predict the coordinate point of the next tool positioning, thereby improving the accuracy and efficiency of tool positioning. The process constructs a prediction model, inputs the workpiece surface model and the current tool position characteristics, outputs the predicted value of the next tool position, and dynamically adjusts the prediction model according to the real-time sensor data to adapt to the actual changes of the workpiece surface.

[0132] In the real-time monitoring and adjustment stage, displacement sensors and cameras are deployed to monitor the surface of the workpiece in real time and collect sensor data. The prediction model is dynamically adjusted according to the real-time data to accurately align the tool with the workpiece surface. The tool position prediction value output by the prediction model is combined with the real-time sensor data to dynamically adjust the position and posture of the tool to achieve accurate alignment between the tool and the workpiece surface.

[0133] Finally, in the redundant control and safety assurance stage, a highly redundant and fault-tolerant system architecture is deployed, equipped with multiple backup and redundant control mechanisms, which can automatically switch to backup mode or safe state under abnormal circumstances to ensure the stability and safety of the processing process.

[0134] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for a tool to quickly find model coordinates during CNC 3-axis machining, characterized in that: include: Step 1. Initial positioning, setting the initial coordinate value of the tool and the processing path; Use preset parameters for initial positioning to position the tool to the initial position of the workpiece; The initial positioning steps include: Set the initial coordinate value of the tool and the processing path, and set the initial coordinate value of the tool in the CNC system, including the X, Y, and Z axis coordinates; Set the tool processing path, including cutting path, cutting direction and cutting depth parameters; Use preset parameters for initial positioning, and position the tool in the CNC system according to the preset initial coordinate values; Use the motion control function provided by the CNC system to move the tool to the preset initial position; According to the workpiece CAD model, determine the initial processing position, use the CNC system to convert the coordinate system, and position the tool to the initial position of the workpiece; At the same time, record the tool coordinate values ​​and machining path parameters used for initial positioning; Step 2. Path planning: pre-design and generate tool paths in CAD software, and determine the tool processing path according to the workpiece geometry and processing requirements, including cutting path, cutting direction and cutting depth parameters; The path planning method is: Use CAD software to import the CAD model file of the workpiece; According to the geometric shape and processing requirements of the workpiece, determine the basic requirements of the processing path, including surface smoothness and dimensional accuracy; According to the workpiece material, processing requirements and tool performance, the corresponding processing technology is adopted, and the corresponding process function is used in the CAD software to draw the cutting path; Determine the direction, shape and spacing of the cutting path according to the processing requirements and workpiece geometry; Setting the cutting direction, including the feed direction and the cutting direction, based on the determined cutting path direction, shape and spacing; Step 3. Initial tool motion control and prediction, control the tool motion trajectory and speed, including the control of the tool's X, Y, and Z axis coordinates, preload the workpiece model into the CNC system before processing; use the preloaded model data and prediction algorithm to predict the coordinate point of the next tool positioning, and use the prediction result as the initial positioning; The steps of building the artifact model include: Use a 3D laser scanner to scan the surface of the workpiece to obtain 3D point cloud data of the surface; Process the scanned point cloud data, including filtering and point cloud registration, to obtain a three-dimensional model of the workpiece surface; Using point cloud data, a grid-based reconstruction algorithm is used to generate a three-dimensional surface model of the workpiece; Construct a prediction model, the input parameters include , as well as , in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position; The input of the prediction model is and , the output is the value ; The prediction model is trained with machining data to learn the relationship between the workpiece surface and tool motion, and the model parameters are optimized using the back propagation algorithm and optimizer; In step 3, the prediction algorithm is constructed as follows: Prepare training sets and validation sets, build a prediction model based on the long short-term memory network LSTM, and define input parameters, including , as well as ,in, is the three-dimensional model of the workpiece surface, representing the three-dimensional grid data. is the characteristic vector of the current position of the tool, including position coordinates, cutting speed, and acceleration. Indicates the predicted value of the next tool position; Use the back propagation algorithm and gradient descent optimization algorithm to train the neural network model, and the optimization loss function formula is: ,in is the actual tool position, is the sample size, represents the time step; Use the validation set to validate the trained neural network model and adjust the model structure and hyperparameters; Before machining, the feature vector of the tool’s current position is and 3D model of the workpiece surface Input the trained neural network model; The model outputs the predicted value of the next tool position ; Step 4. Real-time monitoring and positioning adjustment of the workpiece during the processing phase: deploy displacement sensors and cameras to monitor the surface of the workpiece in real time and collect sensor data. Dynamically adjust the prediction model based on the real-time data, adjust the position and posture of the tool in real time according to the actual situation of the workpiece surface and sensor data, and align the tool with the workpiece surface. Specifically: Install displacement sensors and cameras on processing equipment to monitor the surface of the workpiece in real time and continuously monitor the surface of the workpiece, including surface shape, roughness, and temperature. Collect sensor data, process and analyze it. According to real-time sensor data, dynamically adjust the prediction model to predict the next tool position. Adopting adaptive algorithms to update model parameters according to new data, adapting to the actual changes of the workpiece surface, adjusting the position and posture of the tool in real time according to the tool position prediction value output by the prediction model and real-time sensor data, using the control function provided by the CNC system to adjust the processing path and motion trajectory of the tool so that the tool is aligned with the workpiece surface, and adjusting the position and posture of the tool so that it fits the workpiece surface completely according to the real-time adjusted tool position and posture; The way to predict the next tool position is: Use displacement sensors and cameras to continuously monitor the surface of the workpiece, collect relevant sensor data, and pre-process the collected data, including noise filtering, data alignment and calibration. A neural network model based on deep learning is used, and real-time sensor data is used as the input feature of the model to predict the output value of the next tool position. The gradient descent method is used to update the model parameters according to the actual prediction error. A loss function is defined to measure the difference between the model prediction value and the actual value. The tool position and posture are adjusted based on the tool position prediction value output by the updated prediction model and combined with the real-time sensor data. Step 5. Redundant control: deploy a highly redundant and fault-tolerant system architecture, equipped with multiple backup and redundant control mechanisms, and automatically switch to backup mode under abnormal circumstances.

2. The method for a tool to quickly find model coordinates during a CNC 3-axis process according to claim 1, characterized in that: The redundant control system architecture includes: The main control module is responsible for monitoring the operating status of the processing equipment, controlling the processing process, and processing various instructions; Redundant main control module for active / standby switching and fault-tolerant control; The backup control module, as a backup of the main control module, keeps in sync with the main control module, is ready to take over control at any time, and communicates with the main control module in real time through a redundant communication link; The power supply module provides the power supply required by the equipment. It is equipped with a redundant power supply module, which can automatically switch to the backup power supply when the main power supply fails.

3. The method for a tool to quickly find model coordinates during CNC 3-axis machining according to claim 2, characterized in that: The redundant control system architecture also includes: Sensor modules are responsible for monitoring workpiece and equipment parameters, including temperature, pressure, and speed; they are equipped with redundant sensor modules to continue to acquire data in the event of a sensor failure; The execution module controls the movement and posture of the tool; The communication module is responsible for communicating with external devices and systems, including industrial computers and monitoring systems; The fault detection and automatic switching module implements abnormality detection and automatic switching logic, monitors the operating status of each module of the system, and automatically switches to the backup mode when a fault is detected.

Images