Cutter wear monitoring method, device and equipment and readable storage medium

Abstract

The invention provides a tool wear monitoring method, device and equipment and a readable storage medium, and relates to the technical field of production and manufacturing, the tool wear monitoring method comprises the steps of obtaining cutting process parameters and surface roughness of a target workpiece when the target workpiece is machined by a target tool; the cutting process parameters and the surface roughness are input into the trained tool abrasion loss model, and the predicted tool abrasion loss output by the trained tool abrasion loss model is obtained; inputting the cutting process parameters and the predicted tool abrasion loss into the trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model; and obtaining a tool wear monitoring result of the target tool according to a comparison relationship between the predicted spindle power value and the real-time spindle power value. According to the scheme, the tool wear state can be monitored in real time in the actual machining process, the actual machining process is not hindered, cost is low, and efficiency is high.

Description

Technical Field

[0001] The present invention relates to the field of production and manufacturing technology, and in particular to a tool wear monitoring method, device, equipment and readable storage medium. Background Art

[0002] In traditional parts processing, tools are a very important part, but currently, most of the tool replacement and part surface accuracy judgment are based on the production staff's own experience. Especially when the tool wear reaches the critical value, the production staff needs to stop the machine several times to check whether the tool can still be used. This reduces the processing efficiency in actual production and processing, and also increases the probability of part damage.

[0003] Among the tool wear prediction methods currently under study, most of the data signals used (such as cutting force signals, acoustic emission signals, vibration signals, etc.) are not only expensive to obtain, but the installation of the signal acquisition equipment will also cause unnecessary interference to the actual production and processing.

[0004] In the selection of prediction models, different prediction model combinations have emerged, such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-term Memory Networks (LSTMN), Support Vector Machines (SVM), Deep Belief Networks (DBN), Generative Adversarial Networks (GAN), Back Propagation Networks (BP), etc. Most researchers only focus on the accuracy of the model and use complex input signals (signals fused from multiple sensor signals) for training, ignoring the training speed and prediction speed of the model, which greatly increases the time for training and prediction, making it difficult to monitor the wear status of the tool in real time in actual processing production. Summary of the invention

[0005] The embodiments of the present invention provide a tool wear monitoring method, device, equipment and readable storage medium to solve the problem in the prior art that it is difficult to monitor the wear state of the tool in real time during the actual machining process.

[0006] In order to solve the above technical problems, the embodiments of the present invention provide the following technical solutions:

[0007] An embodiment of the present invention provides a tool wear monitoring method, comprising:

[0008] Obtaining cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece;

[0009] Inputting the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model;

[0010] Inputting the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model;

[0011] According to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the tool wear monitoring result of the target tool is obtained.

[0012] Optionally, the method further comprises:

[0013] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear of the historical tools are used as the input of the preset BP neural network model, and the historical surface roughness of the historical workpieces is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

[0014] Optionally, the method further comprises:

[0015] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear amounts of the historical tools are used as the inputs of the preset BP neural network model, and the historical spindle power values ​​when the historical tools processed the historical workpieces are used as the outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

[0016] Optionally, historical cutting process parameters when the historical tool was used to process the historical workpiece and historical tool wear of the historical tool are used as inputs of a preset BP neural network model, and historical surface roughness of the historical workpiece is used as output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model, including:

[0017] Determining a first influence matrix according to the influence of the historical cutting process parameters and the historical tool wear of the historical tool on the historical surface roughness of the historical workpiece;

[0018] The historical cutting process parameters, the historical tool wear and the first influence matrix are used as inputs of a preset BP neural network model, and the historical surface roughness is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

[0019] Optionally, historical cutting process parameters and historical tool wear values ​​of historical tools when processing historical workpieces are used as inputs of a preset BP neural network model, and historical spindle power values ​​when processing historical workpieces are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model, including:

[0020] Determining a second influence matrix according to the influence of the historical cutting process parameters and the historical tool wear amount of the historical tool on the historical spindle power value;

[0021] The historical cutting process parameters, the historical tool wear and the second influence matrix are used as inputs of a preset BP neural network model, and the historical spindle power values ​​are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

[0022] Optionally, obtaining a tool wear monitoring result of the target tool according to a comparison relationship between the predicted spindle power value and the real-time spindle power value includes:

[0023] When the real-time spindle power value minus the predicted spindle power value is greater than a first preset threshold, determining that the target tool is in a tool damage state;

[0024] When the real-time spindle power value minus the predicted spindle power value is greater than a second preset threshold, and the real-time spindle power value minus the predicted spindle power value is less than or equal to the first preset threshold, determining that the target tool is in a tool wear state;

[0025] When the real-time spindle power value minus the predicted spindle power value is less than or equal to the second preset threshold, determining that the target tool is in a normal machining state;

[0026] The second preset threshold is smaller than the first preset threshold.

[0027] Optionally, the method further comprises:

[0028] Establishing a mapping relationship between the historical surface roughness of the historical workpiece and the historical tool wear amount of the historical tool when the historical tool processes the historical workpiece;

[0029] According to the mapping relationship, the trained tool wear model is verified to obtain a verification result;

[0030] When the verification result indicates that when the historical surface roughness is the same, the absolute value of the difference between the historical tool wear and the predicted tool wear output by the trained tool wear model is less than a third preset threshold, it is determined that the trained tool wear model is available.

[0031] Optionally, before obtaining the tool wear monitoring result according to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the method further includes:

[0032] Determine a correlation coefficient between each characteristic quantity of the historical current when the historical tool processed the historical workpiece and the historical spindle power value when the historical tool processed the historical workpiece;

[0033] Determining, according to the correlation coefficient, a target feature quantity that is most correlated with the historical spindle power value among the feature quantities;

[0034] The real-time spindle power value is obtained according to the target characteristic quantity of the real-time current when the target tool processes the target workpiece.

[0035] Optionally, before obtaining the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece, the method further includes:

[0036] Obtain the teaching tool wear and teaching spindle power values ​​obtained by teaching machining;

[0037] According to the comparison relationship between the teaching spindle power value and the predicted spindle power value, a tool wear teaching monitoring result is obtained;

[0038] When the tool wear teaching monitoring result indicates that the teaching spindle power value is less than the predicted spindle power value, it is determined to obtain the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece.

[0039] The embodiment of the present invention further provides a tool wear monitoring device, comprising:

[0040] A first acquisition module is used to acquire cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece;

[0041] A first processing module, configured to input the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model;

[0042] A second processing module is used to input the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model;

[0043] The third processing module is used to obtain the tool wear monitoring result of the target tool according to the comparison relationship between the predicted spindle power value and the real-time spindle power value.

[0044] An embodiment of the present invention also provides a tool wear monitoring device, comprising: a transceiver, a processor, a memory, and a program or instruction stored in the memory and executable on the processor; when the processor executes the program or instruction, the tool wear monitoring method described above is implemented.

[0045] An embodiment of the present invention further provides a readable storage medium on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps in the tool wear monitoring method described in any one of the above are implemented.

[0046] The beneficial effects of the present invention are:

[0047] The tool wear monitoring method provided by the scheme of the present invention obtains the cutting process parameters when the target tool processes the target workpiece and the surface roughness of the target workpiece; inputs the cutting process parameters and the surface roughness into a trained tool wear model to obtain the predicted tool wear output by the trained tool wear model; inputs the cutting process parameters and the predicted tool wear into a trained tool spindle power model to obtain the predicted spindle power value output by the trained tool spindle power model; obtains the tool wear monitoring result based on the comparison relationship between the predicted spindle power value and the real-time spindle power value and the comparison relationship between the predicted tool wear and the tool wear threshold, which can realize real-time monitoring of the tool wear status during the actual processing process without interfering with the actual processing process, with low cost and high efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flow chart showing a tool wear monitoring method provided by an embodiment of the present invention;

[0049] Figure 2 A specific flow chart showing a tool wear monitoring method provided by an embodiment of the present invention;

[0050] Figure 3 A schematic diagram showing the connection of a tool wear monitoring system provided by an embodiment of the present invention;

[0051] Figure 4A schematic diagram showing an interface of a tool wear monitoring system in a teaching machining state and a comparison relationship between a predicted spindle power value and a real-time spindle power value provided by an embodiment of the present invention;

[0052] Figure 5 A schematic diagram showing an interface of a tool wear monitoring system under a normal machining state and a comparison relationship between a predicted spindle power value and a real-time spindle power value provided by an embodiment of the present invention;

[0053] Figure 6 A schematic diagram showing an interface of a tool wear monitoring system under a tool wear state provided by an embodiment of the present invention and a comparison relationship between a predicted spindle power value and a real-time spindle power value;

[0054] Figure 7 A schematic diagram showing an interface of a tool wear monitoring system under a tool breakage state and a comparison relationship between a predicted spindle power value and a real-time spindle power value provided by an embodiment of the present invention;

[0055] Figure 8 A schematic diagram showing the structure of a tool wear monitoring device provided by an embodiment of the present invention;

[0056] Fig. 9 A schematic diagram showing the structure of a tool wear monitoring device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] The present invention aims at the problem in the prior art that it is difficult to monitor the wear state of a tool in real time during an actual machining process, and provides a tool wear monitoring method, device, equipment and a readable storage medium.

[0059] like Figure 1 As shown, an embodiment of the present invention provides a tool wear monitoring method, comprising:

[0060] Step 101: Obtain cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece.

[0061] In an embodiment of the present invention, the target tool is a tool to be monitored, the target workpiece is a workpiece to be processed, the surface roughness of the target workpiece is known, and the cutting process parameters include at least one of the following: rotation speed, feed, cutting depth and cutting width.

[0062] Step 102: Input the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear of a target tool output by the trained tool wear model.

[0063] In this step, the input of the trained tool wear model is five: rotation speed, feed, cutting depth, cutting width and surface roughness of the target workpiece, and the output is the predicted tool wear.

[0064] Step 103: input the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model.

[0065] In this step, the trained tool spindle power model has five inputs: speed, feed, cutting depth, cutting width and tool wear of the target tool, and the output is the predicted spindle power value.

[0066] Step 104: Obtain tool wear monitoring results based on the comparison relationship between the predicted spindle power value and the real-time spindle power value.

[0067] The real-time spindle power value is the real-time spindle power value of the machining machine, which is obtained by collecting the current from the current sensor arranged at the master control of the machining machine, and does not affect the actual machining production.

[0068] In this step, the comparison relationship between the predicted spindle power value and the real-time spindle power value is monitored. Based on the above two comparison relationships, the wear condition of the target tool can be determined. Specifically, when the real-time spindle power value minus the predicted spindle power value is greater than the first preset threshold value, it is determined that the target tool is damaged (the tool is in a damaged state), and it is necessary to stop the operation to check the tool status and determine whether to replace it. When the real-time spindle power value minus the predicted spindle power value is greater than the second preset threshold value and less than or equal to the first preset threshold value, it is determined that the target tool is worn (the tool is in a worn state), and a tool wear prompt message is generated. When the real-time spindle power value minus the predicted spindle power value is less than or equal to the second preset threshold value, it is determined that the target tool is in normal use (the tool is in a normal processing state), wherein the second preset threshold value is less than the first preset threshold value.

[0069] Through the above steps, the tool wear during the processing can be monitored in real time without hindering the processing production. The cost is low and the calculation speed is fast. Moreover, the tool damage can be discovered in time, which improves the processing efficiency and ensures the processing accuracy.

[0070] In an optional embodiment of the present invention, the method further includes:

[0071] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear of the historical tools are used as the input of the preset BP neural network model, and the historical surface roughness of the historical workpieces is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

[0072] Specifically, in this embodiment, historical data of the process of processing historical workpieces by historical tools are collected, and a database is established. In the database, historical cutting process parameters and historical tool wear are selected as inputs of a preset BP neural network model (BP neural network model 1), and the historical surface roughness of the historical workpiece is used as the output of the preset BP neural network model. The preset BP neural network model is trained, and the trained tool wear model is obtained when the target error of the training is achieved. Optionally, the selected historical data (including historical cutting process parameters, historical tool wear and historical surface roughness) are divided according to a ratio of 4:1, and 4 parts of the historical data are used as training data sets for training, and 1 part of the historical data is used as a test data set for testing.

[0073] Among them, during training, the BP neural network model uses historical speed, historical feed, historical cutting depth, historical cutting width, and historical tool wear as input, and historical surface roughness as output. The BP neural network model has one input layer, one hidden layer, and one output layer, and the structure is 5-256-1.

[0074] In an optional embodiment of the present invention, the method further includes:

[0075] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear amounts of the historical tools are used as the inputs of the preset BP neural network model, and the historical spindle power values ​​when the historical tools processed the historical workpieces are used as the outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

[0076] Specifically, in this embodiment, historical data of the process in which historical tools process historical workpieces is collected, and a database is established. In the database, historical cutting process parameters and historical tool wear are selected as inputs of a preset BP neural network model (BP neural network model 2), and historical spindle power values ​​are used as outputs of the preset BP neural network model. The preset BP neural network model is trained, and the trained tool spindle power model is obtained when the target error of the training is achieved. Optionally, the selected historical data (including historical cutting process parameters, historical tool wear, and historical spindle power values) are divided in a ratio of 4:1, and 4 portions of the historical data are used as training data sets for training, and 1 portion of the historical data is used as a test data set for testing.

[0077] Among them, during training, the BP neural network model uses historical speed, historical feed, historical cutting depth, historical cutting width, and historical tool wear as input, and historical spindle power value as output. The BP neural network model consists of an input layer, a hidden layer, and an output layer. The structure is 5-128-1, with simple structure and fast calculation speed.

[0078] Furthermore, in an embodiment of the present invention, in order to increase the training speed, a gain matrix is ​​calculated according to the influence of the input on the output, which can enable the model to affect the distribution of weights during training, thereby achieving a state of rapid convergence and reducing the time for model training and learning.

[0079] In the model training process, based on the output function, an influence degree analysis calculation is performed on the input value of the model, that is, in an optional embodiment of the present invention, the historical cutting process parameters when the historical tool processes the historical workpiece and the historical tool wear of the historical tool are used as the input of the preset BP neural network model, and the historical surface roughness of the historical workpiece is used as the output of the preset BP neural network model, and the preset BP neural network model is trained to obtain the trained tool wear model, including:

[0080] According to the degree of influence of the historical cutting process parameters and the historical tool wear of the historical tool on the historical surface roughness of the historical workpiece, a first influence matrix is ​​determined, the historical cutting process parameters, the historical tool wear and the first influence matrix are used as inputs of a preset BP neural network model, the historical surface roughness is used as output of the preset BP neural network model, the preset BP neural network model is trained, and the trained tool wear model is obtained.

[0081] The specific operation is to design and conduct an analysis experiment on the influence of historical cutting process parameters and historical tool wear on the historical surface roughness of historical workpieces, and perform range analysis on the experimental data to obtain the primary and secondary influence of each historical cutting process parameter and historical tool wear on the historical surface roughness of historical workpieces, and calculate the enhanced diagonal matrix A (i.e., the first influence matrix) according to the primary and secondary influences, A=diag{a 11 ,a 22 ,……,a nn}, where a nnIt is calculated based on the influence of historical cutting process parameters and historical tool wear of historical tools on the historical surface roughness of historical workpieces. Then, the historical cutting process parameters, the historical tool wear and the first influence matrix are used as inputs of a preset BP neural network model, and the historical surface roughness is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model. The trained tool wear model formula is as follows:

[0082]

[0083] Among them, y j is the output of the jth neuron in the BP neural network model, σ is the activation function, A is the first influence matrix, w ij is the weight coefficient, specifically the connection weight from the i-th neuron to the j-th neuron in the BP neural network model, a i is the input of the jth neuron, b j is the bias, specifically the bias of the jth neuron.

[0084] It should also be noted that different historical cutting process parameters and historical tool wear of historical tools have different influences on the historical surface roughness of historical workpieces. The greater the influence, the greater the corresponding element a in the first influence matrix A. nn The larger the value, the greater the influence degree of the feed parameter in the historical cutting process parameters on the historical surface roughness of the historical workpiece. Therefore, the element a corresponding to the feed parameter in the historical cutting process parameters in the first influence matrix A is nn The value of is the largest.

[0085] In another optional embodiment of the present invention, historical cutting process parameters and historical tool wear of historical tools when processing historical workpieces are used as inputs of a preset BP neural network model, and historical spindle power values ​​when processing historical workpieces are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model, including:

[0086] According to the degree of influence of the historical cutting process parameters and the historical tool wear of the historical tool on the historical spindle power value, a second influence matrix is ​​determined, the historical cutting process parameters, the historical tool wear and the second influence matrix are used as inputs of a preset BP neural network model, the historical spindle power value is used as the output of the preset BP neural network model, the preset BP neural network model is trained, and the trained tool spindle power model is obtained.

[0087] The specific operation is to design and conduct an analysis experiment on the influence of historical cutting process parameters and historical tool wear on historical spindle power values, and perform range analysis on the experimental data to obtain the primary and secondary influence of each historical cutting process parameter and historical tool wear on the historical spindle power value, and calculate the enhanced diagonal matrix A (i.e., the second influence matrix) according to the primary and secondary influences, A=diag{a 11 ,a 22 ,……,a nn}, where a nn It is calculated based on the influence of historical cutting process parameters and historical tool wear of historical tools on historical spindle power values. Then, the historical cutting process parameters, the historical tool wear and the second influence matrix are used as inputs of a preset BP neural network model, and the historical spindle power values ​​are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model, wherein the trained tool spindle power model is similar to the above-mentioned trained tool wear model formula, which will not be repeated here.

[0088] It should also be noted that different historical cutting process parameters and historical tool wear of historical tools have different influences on the historical spindle power value. The greater the influence, the greater the corresponding element a in the second influence matrix A. nn The larger the value, the greater the historical spindle power value. For the historical spindle power value, the cutting depth parameter in the historical cutting process parameters has the greatest influence on the historical spindle power value. Therefore, the element a corresponding to the cutting depth in the historical cutting process parameters in the second influence matrix A is nn The value of is the largest.

[0089] Furthermore, the process of training the tool wear model using the influence matrix in this optional embodiment is similar to the process of training the tool spindle power model using the influence matrix in the above optional embodiment, and will not be repeated here.

[0090] The improved model is trained using historical data as samples. By adding matrix A, the model can converge faster and increase the operation speed. The trained tool spindle power model and the trained tool wear model are saved for backup, and the real-time processing data is added to the historical database, the database is updated, and the updated database is used to further train the model to improve the training accuracy.

[0091] Optionally, the method further comprises:

[0092] A mapping relationship between the historical surface roughness of the historical workpiece and the historical tool wear of the historical tool when the historical tool processed the historical workpiece is established. According to the mapping relationship, the trained tool wear model is verified to obtain a verification result. Specifically, the established mapping relationship is based on an empirical formula to fit an equation to the historical surface roughness and the historical tool wear. Thereafter, the tool wear corresponding to the historical surface roughness is calculated according to the fitting equation and the tool wear corresponding to the historical surface roughness is calculated according to the trained tool wear model. The two tool wears are compared to obtain a verification result, that is, the verification result includes inputting the historical surface roughness and the corresponding historical cutting process parameters into the trained tool wear model under the same historical surface roughness, and the predicted tool wear output by the model. The predicted tool wear is compared with the historical tool wear to obtain a comparison result. If the verification result indicates that under the same historical surface roughness, the absolute value of the difference between the historical tool wear and the predicted tool wear output by the trained tool wear model is less than a third preset threshold value, then it is determined that the trained tool wear model is available.

[0093] After confirming that the trained tool wear model is available, actual processing is performed to ensure the accuracy of tool wear prediction using the trained tool wear model.

[0094] In an optional embodiment of the present invention, before obtaining the tool wear monitoring result according to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the method further includes:

[0095] Determine the correlation coefficient between each characteristic quantity of the historical current when the historical tool processed the historical workpiece and the historical spindle power value when the historical tool processed the historical workpiece, that is, in this optional example, collect historical data, establish a database, and calculate nine common characteristic vectors of the historical current based on the historical current in the database: mean value (MV), root mean square error (RMSE), square mean root (SMR), root mean square (RMS), maximum value (MA), skewness (SF), peak factor (Kurtosis Factor, KF), crest factor (CF) and margin factor (MF). The calculation formula of each characteristic vector is as follows:

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103] CF: y 8 =y 5 / y 4

[0104] MF: y 9 =y 5 / y 3

[0105] Among them, y 1 represents the mean MV, y 2 represents the mean square error RMSE, y 3 Indicates the mean square amplitude SMR, y 4 represents root mean square RMS, y 5 Indicates the maximum value MA, y 6 Indicates the skewness SF, y 7 Indicates the peak factor KF, y 8 represents the kurtosis factor CF, y 9 represents the margin factor MF, represents the function variable and N represents the number of data.

[0106] The solved characteristic quantity is correlated with the predicted target (historical spindle power value) and the Pearson correlation coefficient is calculated. The formula is as follows:

[0107]

[0108] Among them, r xy is the correlation coefficient, x i is the i-th data in sample X (feature quantity), is the mean value of sample X, y i is the i-th data in sample Y (prediction target), is the mean value of sample Y.

[0109] According to the correlation coefficient, a target feature quantity that is most relevant to the historical spindle power value among the feature quantities is determined; and according to a target feature quantity of real-time current when a target tool processes a target workpiece, the real-time spindle power value is obtained. The target feature quantity is one of the above-mentioned feature quantities.

[0110] It should also be noted that, during the model training process, the historical spindle power value can also be obtained based on the target characteristic quantity of the historical current through the above process, thereby improving the calculation efficiency and the accuracy of the model.

[0111] The trained tool wear model and the trained tool spindle power model are integrated into the monitoring system for actual processing monitoring. Before the actual processing monitoring is performed, teaching processing is first performed, that is, before obtaining the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece, the method further includes:

[0112] The teaching tool wear amount and the teaching spindle power value obtained by the teaching processing are obtained, and the tool wear teaching monitoring result is obtained according to the comparative relationship between the teaching spindle power value and the predicted spindle power value. When the tool wear teaching monitoring result indicates that the teaching spindle power value is less than the predicted spindle power value, it is checked whether the tool is working normally. If there is no error in the teaching processing, normal processing can be carried out thereafter to determine the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool is used to process the target workpiece.

[0113] The teaching processing may be processing performed by a target tool at a position other than a target processing position on the target workpiece.

[0114] The process of obtaining the teaching tool wear amount and the teaching spindle power value obtained by teaching machining is basically the same as the process of obtaining the predicted tool wear amount and the predicted spindle power value in step 102 and step 103, and will not be repeated here.

[0115] Carry out teaching processing first and then normal processing. The current tool status is judged by different power responses, which can monitor the tool while ensuring the processing accuracy of the parts.

[0116] It should also be noted that after obtaining the trained tool wear model, a mapping relationship between historical tool wear and historical surface roughness is established, and the mapping relationship is used to compare the mapping relationship between the predicted tool wear and the predicted surface roughness output by the trained tool wear model. The trained tool wear model is verified based on the comparison relationship.

[0117] In the embodiment of the present invention, a set of historical data is used to train the two models. After the training is completed, the predicted tool wear and the predicted surface roughness can be predicted at one time, which is convenient for studying the relationship between tool wear and part surface quality. At the same time, for the establishment of the two models, the calculation of the fitting equation (mapping relationship) is added to perform a second verification on the predicted values ​​of the two models; with the increase in the number of processing productions, the more processing conditions are collected, and the more data is accumulated, the prediction ability and monitoring ability of the model will become stronger and stronger; finally, the data in the processing process is monitored by the monitoring system, which not only avoids tool damage but also ensures the processing accuracy of parts.

[0118] Combine the following Figure 2 , using a specific embodiment to illustrate the specific process of the tool wear monitoring method:

[0119] The MIK-HRI-50A Hall current sensor is used to measure the spindle current data of the machine tool, and a spindle current database is established to save all the collected data. The collected current data is filtered and denoised to obtain the spindle current data, and nine eigenvalues ​​are extracted. The Pearson correlation coefficient analysis method is used to analyze and find the eigenvalue with the strongest correlation as the input feature for standby. The spindle power storage database is calculated using the power calculation formula, and the calculated data is compared with the spindle power data changes inside the machine tool to verify the accuracy of the current. According to the following formula, the spindle power data is determined according to the spindle current data:

[0120]

[0121] The calculated spindle power data is stored in a database, and the spindle power values ​​in the database are used to input the tool spindle power model for training.

[0122] The orthogonal tool wear experiment was designed to collect the tool wear. Each group of orthogonal combination parameters was divided into 50 cutting processes, and the milling length of each process was 1000mm. The machine tool was VMC1008B CNC machine tool, and the tool was φ8 carbide four-edge milling cutter. After completing each set of cutting experiments, the tool was removed and the HY-H2100 portable electron microscope was used to take pictures of the width of the tool's back face wear zone and record them, and save them in the database. At the same time, the Mitutoyo surface roughness instrument was used to measure the roughness of the machined surface. Three positions were selected at equal distances for measurement, and then the average value was calculated as the surface roughness Ra value of the current tool cutting in the current state, and saved for later use.

[0123] The two models are trained with the historical processed data in the prepared database, and the processed data set is split in a 4:1 ratio, with 4 parts used as training data sets for training and 1 part used as a test data set for testing.

[0124] The tool wear data and surface roughness data are used to fit the equation to solve the mapping relationship, which is convenient for solving the threshold of tool wear VB according to the surface accuracy requirements. At the same time, the predicted values ​​of the two trained models can also be verified twice.

[0125] The two trained models are embedded in the tool wear prediction and monitoring system. Since the spindle current is collected in real time during the machine tool processing process (the internal spindle power value of the machine tool is collected through the data acquisition box, and the collected internal spindle power value of the machine tool is also stored in the database), the spindle power converted from the spindle current can be used as a real-time monitoring value for monitoring. By inputting the calculated tool wear threshold, the spindle power threshold is automatically calculated and compared with the real-time collected power curve to achieve the purpose of automatically monitoring the spindle power.

[0126] The data collected during real-time monitoring is still saved in the established database, and the training operation is repeated to optimize and upgrade the trained tool wear prediction model and surface roughness prediction model to enhance their prediction and monitoring capabilities.

[0127] like Figure 3 As shown, the embodiment of the present invention also provides a tool wear monitoring system, which integrates the tool wear amount model (prediction model 1) trained according to the process parameters and spindle current and the tool spindle power model (prediction model 2) trained into the tool wear prediction module and the processing accuracy prediction module, determines the accuracy requirements and preset wear of the target workpiece, performs real-time power monitoring, and integrates it into the monitoring module, performs real-time data collection, collects it into the database, and integrates it into the data acquisition module through the data processor. The tool wear prediction module, the processing accuracy prediction module, the monitoring module and the data acquisition module are integrated into the tool wear monitoring system. The tool wear monitoring system is used to monitor the tool during the processing. The monitored states include four, including the teaching processing state, the normal processing state, the tool wear state and the tool damage state. Optionally, it is ensured that the processing can be stopped when the tool wear state is reached or the processing is stopped in the tool wear state.

[0128] The four states are judged by converting the data collected from real-time processing into spindle power data and comparing it with the power threshold output by the tool wear prediction module. Processing is performed under the currently determined processing parameters, and the power threshold is automatically calculated to ensure the accuracy requirements of the processing accuracy calculation. When the power curve is initially generated, teaching processing is required to stabilize the system. The interface of the tool wear monitoring system in the teaching processing state and the schematic diagram of the comparison relationship between the predicted spindle power value and the real-time spindle power value are shown in the figure below. Figure 4As shown; below this threshold is the normal machining state, the interface of the tool wear monitoring system in the normal machining state and the schematic diagram of the comparison relationship between the predicted spindle power value and the real-time spindle power value are shown in Figure 5 As shown; the real-time power curve approaches the threshold power curve when the tool enters the wear state. The interface of the tool wear monitoring system in the tool wear state and the comparison relationship between the predicted spindle power value and the real-time spindle power value are shown in Figure 6 As shown; when the power mutation exceeds the threshold curve, it proves that the tool is damaged and it is in the tool damage state. It is necessary to stop processing immediately to check the tool status. The interface of the tool wear monitoring system in the tool damage state and the comparison relationship between the predicted spindle power value and the real-time spindle power value are shown in the figure. Figure 7 shown.

[0129] It should be noted that Figures 4 to 7 In the example, the actual power is the real-time spindle power value, and the power threshold is the predicted spindle power value.

[0130] The embodiment of the present invention saves all data during the use of the entire technology, establishes a database, and processes the data in the database in nine ways to extract characteristic values. Correlation analysis is performed based on the different inputs and outputs of different models, and the most highly correlated features are used as training inputs. The database data also receives and saves the real-time processed data, and uses the newly added data to further optimize and update the two models to make their predictions more accurate. The self-optimization and learning of the training model expand the functions of the model and can predict more processing conditions.

[0131] like Figure 8 As shown, the embodiment of the present invention also provides a tool wear monitoring device, including:

[0132] The first acquisition module 801 is used to acquire the cutting process parameters when the target tool processes the target workpiece and the surface roughness of the target workpiece;

[0133] A first processing module 802 is used to input the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model;

[0134] The second processing module 803 is used to input the cutting process parameters and the predicted tool wear amount into the trained tool spindle power model to obtain the predicted spindle power value output by the trained tool spindle power model;

[0135] The third processing module 804 is used to obtain the tool wear monitoring result of the target tool according to the comparison relationship between the predicted spindle power value and the real-time spindle power value.

[0136] Optionally, the device further comprises:

[0137] The first training module is used to use the historical cutting process parameters of the historical tools when processing the historical workpieces and the historical tool wear of the historical tools as the input of the preset BP neural network model, and use the historical surface roughness of the historical workpieces as the output of the preset BP neural network model, train the preset BP neural network model, and obtain the trained tool wear model.

[0138] Optionally, the device further comprises:

[0139] The second training module is used to use the historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear of the historical tools as the input of the preset BP neural network model, and use the historical spindle power values ​​when the historical tools processed the historical workpieces as the output of the preset BP neural network model, train the preset BP neural network model, and obtain the trained tool spindle power model.

[0140] Optionally, the first training module includes:

[0141] A first determining unit is used to determine an influence degree matrix according to the influence degree of the historical cutting process parameters on the historical spindle power value;

[0142] The first training unit is used to take the historical cutting process parameters, the historical tool wear and the influence matrix as inputs of a preset BP neural network model, take the historical surface roughness as the output of the preset BP neural network model, train the preset BP neural network model, and obtain the trained tool wear model.

[0143] Optionally, the second training module includes:

[0144] A second determining unit is used to determine an influence degree matrix according to the influence degree of the historical cutting process parameters on the historical spindle power value;

[0145] The second training unit is used to take the historical cutting process parameters, the historical tool wear and the influence matrix as the input of the preset BP neural network model, take the historical spindle power value as the output of the preset BP neural network model, train the preset BP neural network model, and obtain the trained tool spindle power model.

[0146] Optionally, the third processing module 804 includes:

[0147] A third determining unit is used to determine that the target tool is in a tool damage state when the real-time spindle power value minus the predicted spindle power value is greater than a first preset threshold value;

[0148] a fourth determining unit, configured to determine that the target tool is in a tool wear state when the real-time spindle power value minus the predicted spindle power value is greater than a second preset threshold value, and the real-time spindle power value minus the predicted spindle power value is less than or equal to the first preset threshold value;

[0149] A fifth determining unit, configured to determine that the target tool is in a normal machining state when the real-time spindle power value minus the predicted spindle power value is less than or equal to the second preset threshold;

[0150] The second preset threshold is smaller than the first preset threshold.

[0151] Optionally, the device further comprises:

[0152] A fourth processing module is used to establish a mapping relationship between the historical surface roughness of the historical workpiece and the historical tool wear amount of the historical tool when the historical tool processed the historical workpiece;

[0153] A verification module, used to verify the trained tool wear model according to the mapping relationship to obtain a verification result;

[0154] The first determination module is used to determine that the trained tool wear model is available when the verification result indicates that the historical surface roughness is the same and the absolute value of the difference between the historical tool wear and the predicted tool wear output by the trained tool wear model is less than a third preset threshold.

[0155] Optionally, the device further comprises:

[0156] A second determination module is used to determine a correlation coefficient between each characteristic quantity of a historical current when a historical tool processed a historical workpiece and a historical spindle power value when a historical tool processed a historical workpiece;

[0157] A third determination module is used to determine, according to the correlation coefficient, a target feature quantity that is most correlated with the historical spindle power value among the feature quantities;

[0158] The fifth processing module is used to obtain the real-time spindle power value according to the target characteristic quantity of the real-time current when the target tool processes the target workpiece.

[0159] Optionally, the device further comprises:

[0160] The second acquisition module is used to obtain the teaching tool wear value and the teaching spindle power value obtained by the teaching processing;

[0161] A sixth processing module, configured to obtain a tool wear teaching monitoring result according to a comparison relationship between the teaching spindle power value and the predicted spindle power value;

[0162] The fourth determination module is used to determine the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece when the tool wear teaching monitoring result indicates that the teaching spindle power value is less than the predicted spindle power value.

[0163] It should be noted that the tool wear monitoring device provided in the embodiment of the present invention is a device capable of executing the above-mentioned tool wear monitoring method. All embodiments of the above-mentioned tool wear monitoring method are applicable to the device and can achieve the same or similar technical effects.

[0164] The embodiment of the present invention also provides a tool wear monitoring device, such as Fig. 9 As shown, it includes: a processor 901; and a memory 902 connected to the processor 901 through a bus interface, the memory 902 is used to store programs and data used by the processor 901 when performing operations, and the processor 901 calls and executes the programs and data stored in the memory 902.

[0165] The transceiver 903 is connected to the bus interface and is used to receive and send data under the control of the processor 901 .

[0166] Among them, Fig. 9 In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically one or more processors represented by processor 901 and various circuits of memory represented by memory 902 are linked together. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides a user interface 904. The transceiver 903 may be a plurality of components, namely, a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium. For different user devices, the user interface 904 may also be an interface capable of externally and internally connecting required devices, and the connected devices include but are not limited to a keypad, a display, a speaker, a microphone, a joystick, and the like.

[0167] The processor 901 is responsible for managing the bus architecture and general processing, and the memory 902 can store data used by the processor 901 when performing operations.

[0168] The processor 901 performs the following processes:

[0169] Obtaining cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece;

[0170] Inputting the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model;

[0171] Inputting the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model;

[0172] According to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the tool wear monitoring result of the target tool is obtained.

[0173] Optionally, the processor 901 is further configured to:

[0174] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear of the historical tools are used as the input of the preset BP neural network model, and the historical surface roughness of the historical workpieces is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

[0175] Optionally, the processor 901 is further configured to:

[0176] The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear amounts of the historical tools are used as the inputs of the preset BP neural network model, and the historical spindle power values ​​when the historical tools processed the historical workpieces are used as the outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

[0177] Optionally, the processor 901 is specifically configured to:

[0178] Determining an influence matrix according to the influence degree of the historical cutting process parameters on the historical spindle power value;

[0179] The historical cutting process parameters, the historical tool wear and the influence matrix are used as inputs of a preset BP neural network model, and the historical surface roughness is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

[0180] Optionally, the processor 901 is specifically configured to:

[0181] Determining an influence matrix according to the influence degree of the historical cutting process parameters on the historical spindle power value;

[0182] The historical cutting process parameters, the historical tool wear and the influence matrix are used as inputs of a preset BP neural network model, and the historical spindle power values ​​are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

[0183] Optionally, the processor 901 is specifically configured to:

[0184] When the real-time spindle power value minus the predicted spindle power value is greater than a first preset threshold, determining that the target tool is in a tool damage state;

[0185] When the real-time spindle power value minus the predicted spindle power value is greater than a second preset threshold, and the real-time spindle power value minus the predicted spindle power value is less than or equal to the first preset threshold, determining that the target tool is in a tool wear state;

[0186] When the real-time spindle power value minus the predicted spindle power value is less than or equal to the second preset threshold, determining that the target tool is in a normal machining state;

[0187] The second preset threshold is smaller than the first preset threshold.

[0188] Optionally, the processor 901 is further configured to:

[0189] Establishing a mapping relationship between the historical surface roughness of the historical workpiece and the historical tool wear amount of the historical tool when the historical tool processes the historical workpiece;

[0190] According to the mapping relationship, the trained tool wear model is verified to obtain a verification result;

[0191] When the verification result indicates that when the historical surface roughness is the same, the absolute value of the difference between the historical tool wear and the predicted tool wear output by the trained tool wear model is less than a third preset threshold, it is determined that the trained tool wear model is available.

[0192] Optionally, the processor 901 is specifically configured to:

[0193] Determine a correlation coefficient between each characteristic quantity of the historical current when the historical tool processed the historical workpiece and the historical spindle power value when the historical tool processed the historical workpiece;

[0194] Determining, according to the correlation coefficient, a target feature quantity that is most correlated with the historical spindle power value among the feature quantities;

[0195] The real-time spindle power value is obtained according to the target characteristic quantity of the real-time current when the target tool processes the target workpiece.

[0196] Optionally, the processor 901 is further configured to:

[0197] Obtain the teaching tool wear and teaching spindle power values ​​obtained by teaching machining;

[0198] According to the comparison relationship between the teaching spindle power value and the predicted spindle power value, a tool wear teaching monitoring result is obtained;

[0199] When the tool wear teaching monitoring result indicates that the teaching spindle power value is less than the predicted spindle power value, it is determined to obtain the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece.

[0200] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or may be accomplished by instructing related hardware through a program, wherein the program includes instructions for executing part or all of the steps of the above method; and the program may be stored in a readable storage medium, which may be any form of storage medium.

[0201] An embodiment of the present invention further provides a readable storage medium, wherein a program is stored on the readable storage medium, and when the program is executed by a processor, the tool wear monitoring method as described in any one of the above items is implemented.

[0202] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0203] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may be physically included separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0204] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform some steps of the sending and receiving methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0205] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary personnel in this technical field, several improvements and modifications can be made without departing from the principles described in the present invention. These improvements and modifications are also within the scope of protection of the present invention.

Claims

1. A tool wear monitoring method, characterized in that: include: Obtaining cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece; Inputting the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model; Inputting the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model; According to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the tool wear monitoring result of the target tool is obtained.

2. The method according to claim 1, characterized in that The method further comprises: The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear of the historical tools are used as the input of the preset BP neural network model, and the historical surface roughness of the historical workpieces is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

3. The method according to claim 1, characterized in that The method further comprises: The historical cutting process parameters when the historical tools processed the historical workpieces and the historical tool wear amounts of the historical tools are used as the inputs of the preset BP neural network model, and the historical spindle power values ​​when the historical tools processed the historical workpieces are used as the outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

4. The method according to claim 2, characterized in that: The method uses historical cutting process parameters when a historical tool processes a historical workpiece and historical tool wear of the historical tool as inputs of a preset BP neural network model, uses historical surface roughness of the historical workpiece as outputs of the preset BP neural network model, trains the preset BP neural network model, and obtains the trained tool wear model, including: Determining a first influence matrix according to the influence of the historical cutting process parameters and the historical tool wear of the historical tool on the historical surface roughness of the historical workpiece; The historical cutting process parameters, the historical tool wear and the first influence matrix are used as inputs of a preset BP neural network model, and the historical surface roughness is used as the output of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool wear model.

5. The method according to claim 3, characterized in that: The method uses historical cutting process parameters and historical tool wear of historical tools when processing historical workpieces as inputs of a preset BP neural network model, uses historical spindle power values ​​when processing historical workpieces by historical tools as outputs of the preset BP neural network model, and trains the preset BP neural network model to obtain the trained tool spindle power model, including: Determining a second influence matrix according to the influence of the historical cutting process parameters and the historical tool wear amount of the historical tool on the historical spindle power value; The historical cutting process parameters, the historical tool wear and the second influence matrix are used as inputs of a preset BP neural network model, and the historical spindle power values ​​are used as outputs of the preset BP neural network model. The preset BP neural network model is trained to obtain the trained tool spindle power model.

6. The method according to claim 1, characterized in that According to the comparison relationship between the predicted spindle power value and the real-time spindle power value, a tool wear monitoring result of the target tool is obtained, including: When the real-time spindle power value minus the predicted spindle power value is greater than a first preset threshold, determining that the target tool is in a tool damage state; When the real-time spindle power value minus the predicted spindle power value is greater than a second preset threshold, and the real-time spindle power value minus the predicted spindle power value is less than or equal to the first preset threshold, determining that the target tool is in a tool wear state; When the real-time spindle power value minus the predicted spindle power value is less than or equal to the second preset threshold, determining that the target tool is in a normal machining state; The second preset threshold is smaller than the first preset threshold.

7. The method according to claim 1, characterized in that The method further comprises: Establishing a mapping relationship between the historical surface roughness of the historical workpiece and the historical tool wear amount of the historical tool when the historical tool processes the historical workpiece; According to the mapping relationship, the trained tool wear model is verified to obtain a verification result; When the verification result indicates that when the historical surface roughness is the same, the absolute value of the difference between the historical tool wear and the predicted tool wear output by the trained tool wear model is less than a third preset threshold, it is determined that the trained tool wear model is available.

8. The method according to claim 1, characterized in that Before obtaining the tool wear monitoring result according to the comparison relationship between the predicted spindle power value and the real-time spindle power value, the method further includes: Determine a correlation coefficient between each characteristic quantity of the historical current when the historical tool processed the historical workpiece and the historical spindle power value when the historical tool processed the historical workpiece; Determining, according to the correlation coefficient, a target feature quantity that is most correlated with the historical spindle power value among the feature quantities; The real-time spindle power value is obtained according to the target characteristic quantity of the real-time current when the target tool processes the target workpiece.

9. The method according to claim 1, characterized in that: Before obtaining the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece, the method further includes: Obtain the teaching tool wear and teaching spindle power values ​​obtained by teaching machining; According to the comparison relationship between the teaching spindle power value and the predicted spindle power value, a tool wear teaching monitoring result is obtained; When the tool wear teaching monitoring result indicates that the teaching spindle power value is less than the predicted spindle power value, it is determined to obtain the cutting process parameters of the target tool and the surface roughness of the target workpiece when the target tool processes the target workpiece.

10. A tool wear monitoring device, characterized in that: include: A first acquisition module is used to acquire cutting process parameters when a target tool processes a target workpiece and the surface roughness of the target workpiece; A first processing module, configured to input the cutting process parameters and the surface roughness into a trained tool wear model to obtain a predicted tool wear output by the trained tool wear model; A second processing module is used to input the cutting process parameters and the predicted tool wear amount into a trained tool spindle power model to obtain a predicted spindle power value output by the trained tool spindle power model; The third processing module is used to obtain the tool wear monitoring result of the target tool according to the comparison relationship between the predicted spindle power value and the real-time spindle power value.

11. A tool wear monitoring device comprising: A transceiver, a processor, a memory, and a program or instruction stored in the memory and executable on the processor; characterized in that when the processor executes the program or instruction, the tool wear monitoring method as described in any one of claims 1 to 9 is implemented.

12. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by a processor, the steps of the tool wear monitoring method according to any one of claims 1 to 9 are implemented.

Images