A high-precision coordinate boring machine tool wear intelligent prediction method

Abstract

The application discloses a high-precision coordinate boring machine tool wear intelligent prediction method, and relates to the technical field of cutting tools, which comprises the following steps: step one, equipment building: determining sensor arrangement and communication scheme, and completing data acquisition system building according to the same; step two, testing: performing milling processing experiment, collecting machine tool vibration signals, measuring tool flank wear width and workpiece surface roughness, and establishing a tool wear prediction model according to the same; and step three, using: inputting the collected machine tool vibration signals in the processing process into the model, so that the current tool wear degree can be judged according to the actual vibration signals, and whether the tool needs to be replaced in time can be judged. The application identifies the vibration signal characteristics of the high-precision coordinate boring machine, can effectively improve the prediction accuracy, judges whether the tool wear degree prediction value meets the requirements, and thus prompts the engineers to replace the tool, so that the processing requirements are met.

Description

Technical Field

[0001] This invention relates to the field of cutting tool technology, and in particular to a method for intelligent prediction of tool wear on a high-precision coordinate boring machine. Background Technology

[0002] Machine tools are the core equipment in mechanical manufacturing, playing a vital role in modern manufacturing. Cutting tools are one of the core components of machine tools, serving as the primary tools for material removal and shape processing, and playing a crucial role in machining.

[0003] As a core component of machine tools, cutting tools can prevent problems such as decreased machining quality, workpiece damage, or equipment failure caused by using excessively worn tools by predicting their wear in real time. A high-performance cutting tool can improve machining accuracy and surface quality; effectively control vibration and noise, improving equipment stability and reliability; and meet the demands of high-speed, high-load machining, thereby increasing the production efficiency and economic benefits of mechanical manufacturing.

[0004] Predicting tool wear can help develop reasonable tool replacement plans, avoiding production interruptions and increased costs caused by tool failure. Furthermore, timely replacement of severely worn tools can protect machine tools and workpieces, extending their service life. Therefore, tool wear prediction has significant application value in machining processes. Currently, tool wear prediction methods are mainly based on machine learning, establishing a relationship model between tool wear and machining condition signals, and combining the model with real-time machining condition signals to predict tool wear. However, there is a lack of tool wear prediction methods specifically for high-precision coordinate boring machines; and the tool wear prediction methods lack a connection with actual machining results.

[0005] Therefore, those skilled in the art are dedicated to developing an intelligent prediction method for tool wear on high-precision coordinate boring machines. Summary of the Invention

[0006] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is how to predict the tool wear degree of a high-precision coordinate boring machine by relating it to the actual cutting results.

[0007] To achieve the above objectives, this invention provides a method for intelligent prediction of tool wear on a high-precision coordinate boring machine. It includes the following steps:

[0008] Step 1, Equipment Setup: Determine the sensor layout and communication scheme, and complete the data acquisition system setup accordingly;

[0009] Step 2, Testing: Conduct milling experiments, collect machine tool vibration signals, and measure the tool flank wear width and workpiece surface roughness to establish a tool wear prediction model.

[0010] Step 3, Usage: Input the machine tool vibration signals collected during the machining process into the model. The current tool wear level can be determined based on the actual vibration signals, and then it can be determined whether the tool needs to be replaced in time.

[0011] Furthermore, step 1 also includes the following steps:

[0012] Step 1.1: Determine the sensor layout scheme;

[0013] Step 1.2: Determine the sensor communication scheme;

[0014] Step 1.3: Set up the data acquisition system;

[0015] Step 1.4: Test the data acquisition system.

[0016] Furthermore,

[0017] In step 1.1, vibration sensors are installed at the motor mounts and lead screw nuts of key components such as the electric spindle and linear axis. Specifically, the vibration sensors are installed in the X, Y, and Z directions at the corresponding positions to achieve comprehensive acquisition of machine tool vibration signal data.

[0018] In step 1.2, an overall communication system scheme including Siemens PLC, Anybus module, CNC system and PC is established to realize the conversion of communication protocol and the transmission of machine tool vibration signal;

[0019] In step 1.3, the vibration sensor is connected to the Siemens PLC, and the signal transmitted from the Profinet bus protocol is converted into the Enthercat bus protocol signal through the Anybus module, so as to realize the transmission of the machine tool vibration signal from the sensor to the CNC system and the PC.

[0020] In step 1.4, the machine tool vibration is simulated to verify whether the sensor data can be accurately collected and reliably transmitted to the data processing system; at the same time, the stability of the communication connection is checked to ensure that the data acquisition system can operate normally in actual work.

[0021] Furthermore, step 2 also includes the following steps:

[0022] Step 2.1: Conduct milling machining tests;

[0023] Step 2.2: Collect machine tool vibration signals;

[0024] Step 2.3: Calculate the degree of tool wear;

[0025] Step 2.4: Measure the surface roughness of the workpiece;

[0026] Step 2.5: Preprocessing of machine tool vibration signal data;

[0027] Step 2.6: Establish an initial model for tool wear prediction;

[0028] Step 2.7: Optimize the parameters of the tool wear prediction model.

[0029] Furthermore,

[0030] In step 2.1, the cutting parameters are set as follows: spindle speed 1100 rpm, feed rate 700 mm / rpm, depth of cut 0.5 mm; the material to be cut is gray cast iron, and the test tool is a milling tool, model SEHT1204AFTN-ZNS4020.

[0031] In step 2.2, the machine tool vibration signals of the entire milling process from no wear to complete wear failure are collected by the data acquisition system.

[0032] In step 2.3, the tool is observed using a macroscopic microscope, and the wear width of the tool's flank face is measured. The degree of tool wear is equal to the wear width of the tool's flank face divided by the total width of the tool's flank face.

[0033] In step 2.4, nine measuring points are selected on the workpiece machining surface, and the surface roughness at the measuring points is measured by a roughness meter. The average roughness of the nine measuring points is taken as the workpiece surface roughness. The tool wear threshold is determined based on whether the workpiece surface roughness meets the machining accuracy.

[0034] In step 2.5, the machine tool vibration signal data is truncated to remove tool idle data; wavelet threshold denoising technology is used to process the truncated milling data to eliminate noise interference and extract effective information; time-frequency feature extraction is performed on the denoised data to obtain more representative and richer vibration signal features.

[0035] In step 2.6, a mapping relationship between machine tool vibration signals and tool wear degree is established based on machine learning neural network to form an initial tool wear prediction model. The node parameters in the initial tool wear prediction model are all initial values, and iterative optimization is performed based on the dataset obtained from machining.

[0036] In step 2.7, a dataset is established based on existing data, the loss function corresponding to the neural network is set, and the established dataset is used to perform parameter optimization iteration on the initial model for tool wear prediction. Whether the model has been optimized is determined by whether the loss function converges.

[0037] Furthermore, step 3 also includes the following steps:

[0038] Step 3.1: Perform milling machining;

[0039] Step 3.2: Collect machine tool vibration signals;

[0040] Step 3.3: Preprocessing of machine tool vibration signal data;

[0041] Step 3.4: Prediction of tool wear.

[0042] Furthermore,

[0043] In step 3.1, select the appropriate tool type according to the specific workpiece material and machining requirements, and determine the appropriate cutting parameters in combination with the machining process; adjust parameters such as spindle speed, feed rate and depth of cut for different materials and machining surface requirements to ensure cutting quality and machining efficiency.

[0044] In step 3.2, the machine tool vibration signal during the milling process is collected through the constructed data acquisition system;

[0045] In step 3.3, the machine tool vibration signal is recorded, the recorded machine tool vibration signal data is truncated, and wavelet threshold denoising is performed on the truncated data to reduce the interference of noise on signal analysis; time-frequency feature extraction is performed on the denoised data to extract key vibration features, reduce the amount of data and provide more accurate input for subsequent model building;

[0046] In step 3.4, the preprocessed machine tool vibration signal data is input into the established tool wear prediction optimization model to obtain the predicted value of tool wear degree.

[0047] Furthermore, if the predicted value of tool wear exceeds the threshold determined during the testing phase, the tool is deemed to have failed and must be replaced promptly to avoid affecting machining quality and workpiece accuracy. Conversely, if the predicted value does not exceed the threshold, the tool can continue to be used for machining operations, thereby achieving more effective production planning and resource utilization.

[0048] Furthermore, the tool wear prediction model was tested and verified for its feasibility and scalability on a four-axis machining center.

[0049] Furthermore, the vibration sensor is an acceleration sensor, the test data acquisition terminal is a Donghua test data acquisition terminal, and the tool wear prediction model is built based on a Bayesian ridge neural network.

[0050] Compared with the prior art, the present invention has the following advantages:

[0051] (1) The method of the present invention can effectively improve the accuracy of prediction by identifying the vibration signal characteristics of high precision coordinate boring machine.

[0052] (2) The method of the present invention can be used to determine whether the predicted value of tool wear meets the requirements, thereby prompting the engineer to replace the tool to meet the processing requirements.

[0053] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0054] Figure 1 This is a schematic flowchart of the cutting tool wear prediction method of the present invention;

[0055] Figure 2 This is a schematic diagram of an acceleration sensor according to an embodiment of the present invention.

[0056] Wherein: 1-Acceleration signal transmission port, 2-Acceleration sensor body. Detailed Implementation

[0057] The preferred embodiments of the present invention are described below with reference to the accompanying drawings to make the technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0058] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0059] This invention proposes an intelligent prediction method for tool wear on high-precision coordinate boring machines. Considering the machining characteristics of high-precision coordinate boring machines, accelerometers are strategically placed to measure the machine tool's vibration signals during boring and milling operations. A tool wear prediction model is established based on a machine learning neural network. This model is then linked to actual machining accuracy requirements. As tool wear intensifies, the surface roughness gradually increases. The method determines whether tool wear has failed based on whether the surface roughness of the workpiece meets the machining accuracy requirements.

[0060] like Figure 1 The diagram shown is a flowchart illustrating the tool wear prediction method for high-precision coordinate boring machine processing according to the present invention. The method of the present invention mainly includes the following three steps:

[0061] The first step is equipment setup: it is necessary to determine the sensor layout and communication scheme, and then complete the data acquisition system setup accordingly.

[0062] The second step is testing: a milling experiment needs to be conducted to collect machine tool vibration signals and measure the wear width of the tool flank and the surface roughness of the workpiece, thereby establishing a tool wear prediction model.

[0063] The third step is to use the machine tool vibration signal during the machining process. The model can then be used to determine the current degree of tool wear based on the actual vibration signal, and thus determine whether the tool needs to be replaced in a timely manner.

[0064] The following details the content of each step:

[0065] Step 1: Equipment Setup

[0066] 1. Determine the sensor layout scheme

[0067] By analyzing the structure and working principle of the machine tool, vibration sensors were determined to be placed at key components such as the electric spindle, linear axis, motor mounts, and lead screw nuts. Considering that the vibration signals during machine tool processing will change in the X, Y, and Z directions, vibration sensors in the X, Y, and Z directions need to be installed at corresponding locations to achieve comprehensive acquisition of machine tool vibration signal data.

[0068] 2. Determine the sensor communication scheme

[0069] Establish an integrated communication system solution that includes Siemens PLC, Anybus module, CNC system and PC, to realize the conversion of communication protocols and the transmission of machine tool vibration signals.

[0070] 3. Build a data acquisition system

[0071] According to the sensor layout scheme, vibration sensors are installed at key component locations such as the motor mounts and lead screw nuts of the electric spindle and linear axis. The sensors are connected to the Siemens PLC, and the signals transmitted from the Profinet bus protocol are converted into signals from the Enthercat bus protocol through the Anybus module, thus realizing the transmission of machine tool vibration signals from the sensors to the CNC system and PC.

[0072] 4. Test data acquisition system

[0073] After the equipment is set up, machine tool vibration is simulated to verify whether the sensor data can be accurately collected and reliably transmitted to the data processing system. Simultaneously, the stability of the communication connection is checked to ensure the data acquisition system can operate normally in actual work.

[0074] Step Two: Testing

[0075] 1. Conduct milling machining tests

[0076] The cutting parameters were set to a spindle speed of 1100 rpm, a feed rate of 700 mm / rpm, and a depth of cut of 0.5 mm. The material to be cut was gray cast iron, and the cutting tool was a milling insert, model SEHT1204AFTN-ZNS4020.

[0077] 2. Collect machine tool vibration signals

[0078] The data acquisition system was built to collect machine tool vibration signals throughout the entire milling process, from the point where the tool is unworn to when it is completely worn and fails.

[0079] 3. Calculate the degree of tool wear.

[0080] The cutting tool is observed using a macroscopic microscope, and the wear width of the tool's flank face is measured. The degree of tool wear is equal to the wear width of the tool's flank face divided by the total width of the tool's flank face.

[0081] 4. Measure the surface roughness of the workpiece

[0082] Nine measuring points are selected on the machined surface of the workpiece. The surface roughness at each measuring point is measured using a roughness tester. The average roughness of the nine measuring points is taken as the workpiece surface roughness. The tool wear threshold is determined based on whether the workpiece surface roughness meets the machining accuracy requirements. This step is to evaluate the quality of the machined surface, thereby indirectly reflecting the tool wear situation and providing an important basis for subsequent analysis.

[0083] 5. Preprocessing of machine tool vibration signal data

[0084] The machine tool vibration signal data is truncated to remove tool idle data; wavelet threshold denoising technology is used to process the truncated milling data to eliminate noise interference and extract effective information; time-frequency feature extraction is performed on the denoised data to obtain more representative and richer vibration signal features.

[0085] 6. Establish an initial model for tool wear prediction

[0086] A mapping relationship between machine tool vibration signals and tool wear is established based on machine learning neural networks to form an initial model for tool wear prediction. The node parameters in this model are all initial values ​​and need to be iteratively optimized based on the dataset collected during machining.

[0087] 7. Optimization of tool wear prediction model parameters

[0088] A dataset is built based on existing data, and a loss function is set for the neural network. The established dataset is used to perform parameter optimization iterations on the initial model for tool wear prediction. Whether the model has been optimized is determined by whether the loss function converges. This optimization process will improve the accuracy of the model's predictions.

[0089] Step 3, Use

[0090] 1. Carry out milling machining

[0091] In actual milling operations, it is necessary to select the appropriate tool type based on the specific workpiece material and machining requirements, and determine suitable cutting parameters in conjunction with the machining process. For different materials and surface finish requirements, parameters such as spindle speed, feed rate, and depth of cut need to be adjusted to ensure cutting quality and machining efficiency.

[0092] 2. Collect machine tool vibration signals

[0093] The vibration signals of the machine tool during the milling process were collected by the data acquisition system that was built.

[0094] 3. Preprocessing of machine tool vibration signal data

[0095] To obtain effective machining signals, the recorded machine tool vibration signal data is truncated; wavelet threshold denoising is performed on the truncated data to reduce noise interference with signal analysis; time-frequency feature extraction is performed on the denoised data to extract key vibration features, thereby reducing the amount of data and providing more accurate input for subsequent model building.

[0096] 4. Prediction of tool wear

[0097] The preprocessed machine tool vibration signal data is input into the established tool wear prediction optimization model to obtain the predicted value of tool wear. When the predicted value of tool wear exceeds the threshold determined in the testing phase, the tool is judged to be in failure and needs to be replaced in time to avoid affecting the machining quality and workpiece accuracy; conversely, if the predicted value does not exceed the threshold, the tool can continue to be used for machining operations, thereby achieving more effective production planning and resource utilization.

[0098] To verify the feasibility and scalability of the tool wear prediction model, it was tested on an existing four-axis machining center, using methods such as... Figure 2 The accelerometer shown collects vibration signals from the spindle and linear axes of a four-axis machining center via its main body 2. The signals are then input to the Donghua test data acquisition terminal via acceleration signal transmission port 1, and a CSV file is output. A tool wear prediction model is built based on a Bayesian ridge neural network. The model structure can be represented by the following pseudocode:

[0099] 1. Import the required function libraries for the code.

[0100] 2. Import the training dataset

[0101] 3. Establishing an initial model for predicting tool wear (input signal format)

[0102] 3.1 Create an empty neural network model

[0103] 3.2 Add fully connected layers and other hidden layers to the neural network model.

[0104] 3.3 Setting forward and reverse parameter optimization formulas

[0105] 4. Instantiate a tool wear prediction model

[0106] 5. Define the loss function and optimizer

[0107] 6. Set the number of loop iterations and begin loop optimization.

[0108] 6.1 Optimizing the Node Parameters of a First-Round Neural Network Using the Training Set

[0109] 6.2 Calculate the loss function and its gradient

[0110] 6.3 Determine if the loss function has converged. If it has converged, exit the loop; otherwise, continue the loop.

[0111] 7. Evaluate the model on the test set.

[0112] 8 Output training and test set accuracy

[0113] The method of this invention can effectively improve the accuracy of prediction by identifying the vibration signal characteristics of high-precision coordinate boring machines. By judging whether the predicted value of tool wear meets the requirements, it can prompt engineers to replace the tool to meet the machining requirements.

[0114] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for intelligent prediction of tool wear on a high-precision coordinate boring machine, characterized in that, Includes the following steps: Step 1, Equipment Setup: Determine the sensor layout and communication scheme, and complete the data acquisition system setup accordingly; Vibration sensors are installed at the motor mounts and lead screw nuts of key components of the electric spindle and linear axis. Specifically, the vibration sensors are installed in the X, Y, and Z directions at the corresponding positions to achieve comprehensive acquisition of machine tool vibration signal data. Establish an integrated communication system solution that includes Siemens PLC, Anybus module, CNC system and PC, to realize the conversion of communication protocol and the transmission of machine tool vibration signals; The vibration sensor is connected to the Siemens PLC, and the signal transmitted from the Profinet bus protocol is converted into the Enthercat bus protocol signal through the Anybus module, so as to realize the transmission of machine tool vibration signal from the sensor to the CNC system and the PC. Step 2, Testing: Conduct milling experiments, collect machine tool vibration signals, and measure the wear width of the tool flank and the surface roughness of the workpiece. Based on this, establish a tool wear prediction model. The cutting parameters were set as follows: spindle speed 1100 rpm, feed rate 700 mm / rpm, depth of cut 0.5 mm; the material to be cut was gray cast iron, and the test tool was a milling cutter, model SEHT1204AFTN-ZNS4020. The machine tool vibration signals were collected by a data acquisition system to track the entire milling process from no wear to complete wear failure. The tool was observed using a macroscopic microscope, and the wear width of the tool's flank face was measured. The degree of tool wear was equal to the wear width of the tool's flank face divided by the total width of the tool's flank face. Nine measuring points are selected on the workpiece surface. The surface roughness at the measuring points is measured by a roughness meter. The average roughness of the nine measuring points is taken as the workpiece surface roughness. The tool wear threshold is determined based on whether the workpiece surface roughness meets the machining accuracy. The machine tool vibration signal data is truncated to remove tool idle data; wavelet threshold denoising technology is used to process the truncated milling data to eliminate noise interference and extract effective information; time-frequency feature extraction is performed on the denoised data to obtain more representative and richer vibration signal features. A mapping relationship between machine tool vibration signals and tool wear degree is established based on machine learning neural networks to form an initial tool wear prediction model. The node parameters in the initial tool wear prediction model are all initial values, and are iteratively optimized based on the dataset collected during machining. The tool wear prediction model is built based on a Bayesian ridge neural network. A dataset is established based on existing data, a loss function is set for the neural network, and the established dataset is used to perform parameter optimization iteration on the initial model for tool wear prediction. Whether the model has been optimized is determined by whether the loss function converges. Step 3, Usage: Input the machine tool vibration signals collected during the machining process into the model. The current tool wear level can be determined based on the actual vibration signals, and then it can be determined whether the tool needs to be replaced in time.

2. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in claim 1, characterized in that, Step 1 further includes the following steps: Step 1.1: Determine the sensor layout scheme; Step 1.2: Determine the sensor communication scheme; Step 1.3: Set up the data acquisition system; Step 1.4: Test the data acquisition system.

3. The intelligent prediction method for tool wear of high-precision coordinate boring machines as described in claim 2, characterized in that, In step 1.4, the machine tool vibration is simulated to verify whether the sensor data can be accurately collected and reliably transmitted to the data processing system; at the same time, the stability of the communication connection is checked to ensure that the data acquisition system can operate normally in actual work.

4. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in claim 1, characterized in that, Step 2 also includes the following steps: Step 2.1: Conduct milling machining tests; Step 2.2: Collect machine tool vibration signals; Step 2.3: Calculate the degree of tool wear; Step 2.4: Measure the surface roughness of the workpiece; Step 2.5: Preprocessing of machine tool vibration signal data; Step 2.6: Establish an initial model for tool wear prediction; Step 2.7: Optimize the parameters of the tool wear prediction model.

5. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in claim 1, characterized in that, Step 3 further includes the following steps: Step 3.1: Perform milling machining; Step 3.2: Collect machine tool vibration signals; Step 3.3: Preprocessing of machine tool vibration signal data; Step 3.4: Prediction of tool wear.

6. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in claim 5, characterized in that, In step 3.1, select the appropriate tool type according to the specific workpiece material and machining requirements, and determine the appropriate cutting parameters in combination with the machining process; adjust parameters such as spindle speed, feed rate and depth of cut for different materials and machining surface requirements to ensure cutting quality and machining efficiency. In step 3.2, the machine tool vibration signal during the milling process is collected through the constructed data acquisition system; In step 3.3, the machine tool vibration signal is recorded, the recorded machine tool vibration signal data is truncated, and wavelet threshold denoising is performed on the truncated data to reduce the interference of noise on signal analysis; time-frequency feature extraction is performed on the denoised data to extract key vibration features, reduce the amount of data and provide more accurate input for subsequent model building; In step 3.4, the preprocessed machine tool vibration signal data is input into the established tool wear prediction optimization model to obtain the predicted value of tool wear degree.

7. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in claim 6, characterized in that, If the predicted value of tool wear exceeds the threshold determined during the testing phase, the tool is considered to have failed and must be replaced promptly to avoid affecting machining quality and workpiece accuracy. Conversely, if the predicted value does not exceed the threshold, the tool can continue to be used for machining operations, thereby achieving more effective production planning and resource utilization.

8. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in any one of claims 1-7, characterized in that, The tool wear prediction model was tested and verified for its feasibility and scalability on a four-axis machining center.

9. The intelligent prediction method for tool wear of a high-precision coordinate boring machine as described in any one of claims 1-7, characterized in that, The test data acquisition terminal is the Donghua test data acquisition terminal.

Images