CN113369994A - Cutter state monitoring method in high-speed milling process - Google Patents
Cutter state monitoring method in high-speed milling process Download PDFInfo
- Publication number
- CN113369994A CN113369994A CN202110738211.5A CN202110738211A CN113369994A CN 113369994 A CN113369994 A CN 113369994A CN 202110738211 A CN202110738211 A CN 202110738211A CN 113369994 A CN113369994 A CN 113369994A
- Authority
- CN
- China
- Prior art keywords
- monitoring
- sample
- simulation
- milling process
- speed milling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
- B23Q17/00—Arrangements for observing, indicating or measuring on machine tools
- B23Q17/09—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool
- B23Q17/0952—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool during machining
- B23Q17/0971—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool during machining by measuring mechanical vibrations of parts of the machine
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention belongs to the technical field of machining, and particularly discloses a method for monitoring the state of a cutter in a high-speed milling process, which comprises the following steps: (1) acquiring an actually measured vibration signal, (2) acquiring a vibration simulation signal, (3) checking the effectiveness of a simulation sample, (4) optimizing the simulation model, (5) preparing a complete sample, (6) preparing a complete sample, (7) training a high-precision monitoring model, and (8) monitoring the tool classification. Compared with the prior art, the method provides a method for completing the cutter state monitoring sample in the high-speed milling process based on finite element simulation and a generating type countermeasure network, so that the problems of sample loss and sample shortage in the research of the high-speed milling process are solved, the problem of selecting a diagnosis model can be avoided, and the difficulty in adjusting and optimizing model parameters is greatly reduced.
Description
Technical Field
The invention belongs to the technical field of machining, and particularly discloses a method for monitoring the state of a cutter in a high-speed milling process.
Background
Milling refers to a machining method for machining the surface of an object by using a rotating multi-edge cutter (milling cutter), and the milling is generally performed on a milling machine or a boring machine, can machine planes, grooves, gear teeth, threads and spline shafts, and can machine more complicated special shapes. The cutter is used as the most easily damaged part in the milling process, and is vital to timely and effective state monitoring and fault identification.
The high-speed milling is a small-diameter milling cutter, the milling operation is carried out by adopting high rotating speed and small periodic feeding amount, compared with the common milling, the high-speed milling has the advantages of high efficiency, high precision, low surface roughness and the like, and the high-speed milling becomes an important processing means for ensuring the high-efficiency and high-precision processing requirements of a plurality of difficult-to-process materials. In high-speed milling, the damage of the cutter easily causes the insufficient processing precision, and the processing performance is seriously restricted. How to effectively carry out the cutter state monitoring (TCM) of high-speed milling, accurately and timely identify the damage degree of the cutter, monitor the running state of the cutter, become the problem that the intelligent development of high-speed milling is urgent to solve, and is one of the main research and development directions of the current intelligent processing technology.
In recent years, scholars at home and abroad make a great deal of research work on the aspects of identifying and monitoring the state of a cutter in the milling process, and provide a plurality of effective high-precision and high-reliability diagnosis methods such as a neural network, a support vector machine, a convolutional neural network and the like, so that a rich technical basis is provided for TCM research in the milling process. However, for high-speed milling, the identification of the tool state faces two problems:
(1) due to the complexity and dynamic randomness of the high-speed milling process, the situation that the tool state types appearing in a plurality of actual machining are not collected in an experiment easily occurs, namely the problem that a sample data set used for model training has sample loss. Particularly, when the cutting parameters are changed, that is, the cutting parameters corresponding to the test sample and the experimental sample are different, the trained model has poor monitoring performance, so that the monitoring model is difficult to apply.
(2) At present, a plurality of TCM models can obtain a good identification effect only by a large amount of sample data, and the sample collection cost in the high-speed milling process is high and time-consuming, so that it is very difficult to collect sufficient tool state samples under the condition. In the case of insufficient samples, the monitoring performance of many TCM models is greatly compromised.
Disclosure of Invention
The invention aims to provide a method for monitoring the state of a cutter in a high-speed milling process, which aims to solve the problems that the cutter state monitoring samples are few and incomplete in the high-speed milling process.
In order to achieve the purpose, the basic scheme of the invention is as follows: a method for monitoring the state of a cutter in a high-speed milling process,
(1) acquiring an actually measured vibration signal: determining a used cutter, a workpiece and cutting parameters according to actual machining requirements, and acquiring actual measurement vibration signals in a high-speed milling process under a plurality of cutter states;
(2) acquiring a vibration simulation signal: according to the Johnson-Cook constitutive model, a tool, a workpiece and cutting parameters adopted by an experiment, performing high-speed milling process dynamics modeling by using simulation software to obtain a simulation model, and simulating to obtain a vibration simulation signal in a normal tool state;
(3) checking the validity of the simulation sample: comparing cosine similarity values cos (theta) of the vibration simulation signal and the actually measured vibration signal; if cos (theta) is not less than 0.6, judging that the vibration simulation signal is effective, and entering the step (5), otherwise, executing the step (4);
(4) optimizing a simulation model: performing orthogonal test on parameters of the Johnson-Cook constitutive model in the simulation model in the step (2), and analyzing the test result to obtain the optimal parameter combination; if the cosine similarity value cos (theta) of the vibration simulation signal and the actually measured vibration signal corresponding to the optimal parameter combination is not less than 0.6, judging that the vibration simulation signal is effective, and entering the step (5), otherwise, iteratively executing the step (4) by using the optimal parameter combination as a reference;
(5) preparation of a complete sample: simulating the cutter state except for the step (1) by adopting a simulation model corresponding to the effective vibration simulation signal to obtain a simulation data sample of the cutter state except for the step (1), and expanding the simulation data sample into an experimental data sample to obtain a complete sample;
(6) preparing a complete sample: carrying out generative confrontation network training on the complete samples to generate a large number of synthetic samples; combining the synthesized sample and the complete sample to form a complete sample;
(7) training a high-precision monitoring model: calculating 10 time domain and frequency domain statistical parameters of a complete sample by using a monitoring algorithm, inputting the statistical parameters into the monitoring algorithm for training, and obtaining a high-precision monitoring model;
(8) and (3) cutter classification monitoring: and (3) periodically collecting actually-measured vibration signals in the high-speed milling process, calculating 10 time domain and frequency domain statistical parameters of the sample to be measured, and carrying out tool classification monitoring on the tool state by adopting the trained state monitoring model in the step (7).
The tool state in the scheme comprises normal wear, medium wear, serious wear, breakage, tipping and the like.
Further, the simulation software in step (2) uses Abaqus or Deform.
Further, the formula for calculating the cosine similarity value in step (3) is as follows:
further, 80%, 100% and 120% of parameter values in the simulation model are divided into three levels in step (4) to perform orthogonal experiments.
Further, the discriminator model and the generator model of the countermeasure network in the step (6) are both five-layer one-dimensional convolutional neural networks, and the hidden layer neuron nodes are 1800-.
Further, the training of the countermeasure network in the step (6) requires inputting a noise signal, and setting gaussian distribution and uniform distribution of the input noise. To improve the variability of the synthesized samples.
Further, the monitoring algorithm in the step (7) selects four classification algorithms: support vector machines, random forests, decision trees, and generalized recurrent neural networks.
Further, the four classification algorithms in step (7) are respectively set as follows: the support vector machine selects a radial basis kernel function, and the penalty factor and the kernel function radius are respectively set to be 3 and 1; the random forest and decision tree classifier uses default parameters in the Matlab toolbox; the value of SPREAD in the generalized recurrent neural network classifier is set to 0.1.
Further, the tool state in step (1) includes normal, medium wear and severe wear.
Further, the tool state other than the step (1) in the step (5) includes moderate wear, breakage, chipping.
The beneficial effect of this basic scheme lies in:
1. aiming at the problem of sample missing in the cutter state monitoring (TCM), a cutter state monitoring sample completion method based on finite element simulation is provided; according to the method, a high-precision dynamic simulation model based on a Johnson-Cook constitutive model in the high-speed milling process is constructed on the basis of a low-cost finite element simulation method, and then vibration signals in various cutter states are acquired by the high-precision simulation model to complete the sample data of the cutter states.
2. Aiming at the problem of insufficient samples in the cutter state monitoring (TCM), a cutter state monitoring sample enhancement method based on a generative countermeasure network (GAN) is provided, and the method utilizes the GAN on the basis of an experimental sample and a simulation sample to generate more synthetic samples so as to achieve the purpose of sample enhancement.
3. The data are collected according to the processing requirements, software simulation is adopted, a relatively comprehensive sample can be obtained, and the cutter state monitoring method can rapidly identify the effect at low cost, so that the cutter state monitoring sample is complete.
4. Multiple tests prove that 100% classification precision is obtained in a plurality of diagnosis models by using the method, so that the confusion of selecting the diagnosis models from a plurality of machine learning methods by a user is avoided, and the difficulty of optimizing the parameters of the diagnosis models is greatly reduced.
Compared with the prior art, the method provides a method for completing the cutter state monitoring sample in the high-speed milling process based on finite element simulation and a generating countermeasure network, so that the problems of sample loss and sample shortage in the research of the high-speed milling process are solved.
Drawings
FIG. 1 is a graph of the classification accuracy of each wear category obtained by an SVM classifier under three training data sets;
FIG. 2 is a graph of the classification accuracy for each wear category obtained by the RF classifier for three training data sets;
FIG. 3 is a graph of the classification accuracy for each wear category obtained by the DT classifier under three training data sets;
fig. 4 shows the classification accuracy of each wear category obtained by the GRNN classifier under three training data sets.
Detailed Description
The following is further detailed by way of specific embodiments:
the invention relates to a method for monitoring the state of a cutter in a high-speed milling process, which comprises the following steps (taking an end mill as an example in the embodiment):
(1) acquiring an actually measured vibration signal: determining a used cutter, a workpiece and cutting parameters according to actual machining requirements, and acquiring actual measurement vibration signals in a high-speed milling process under a plurality of cutter states;
in the embodiment, a high-speed milling experiment is carried out, vibration time domain signals in three directions under C cutter states (normal, slight abrasion, severe abrasion and the like) are collected and recorded as1,2, Z, C, Z is the number of collected signal points (in this embodiment, Z is 24000), the sampling frequency is 12KHz, and C is the C-th tool state.
(2) Acquiring a vibration simulation signal: according to the Johnson-Cook constitutive model, a tool, a workpiece and cutting parameters adopted by an experiment, performing high-speed milling process dynamics modeling by using simulation software to obtain a simulation model, and obtaining a vibration simulation signal under a normal tool state;
in this embodiment, a milling experiment is simulated based on finite element analysis software Abaqus to obtain vibration simulation signal data in three directions, and a J-C constitutive model selected by the model is as follows:
wherein A is the initial yield stress (MPa); b is a strain hardening constant (MPa); c is a strain rate coefficient; n is a strain hardening index; m is a temperature softening index; C. n and m are material characteristic coefficients T, Troom、TmeltRespectively deformation temperature, room temperature (typically 20 deg.c) and material melting point.
(3) Checking the validity of the simulation sample: comparing cosine similarity values cos (theta) of the vibration simulation signal and the actually measured vibration signal; if cos (theta) is not less than 0.6, the simulation signal can be judged to be effective, the step (5) is carried out, otherwise, the step (4) is carried out;
in this embodiment, the obtained simulated cutting force signal of the tool in the normal state is compared with the experimental cutting force signal, and 24000 data points are obtained from the previous 2 seconds when the tool completely cuts into the workpiece; table 1 shows cosine similarity values under eight sets of cutting parameters, and as can be seen from table 1, the cosine similarity values in three directions of the eight sets of cutting parameters are all greater than 0.6. It is considered that the simulation data has good similarity to the measurement data, and the simulation signal effectively enters step (5).
TABLE 1 cosine similarity values between simulation and experimental signals under eight cutting experiments
Cutting tool | Vibration signal in X direction | Vibration signal in Y direction | Vibration signal in |
1 | 0.847 | 0.741 | 0.681 |
2 | 0.835 | 0.736 | 0.698 |
3 | 0.846 | 0.763 | 0.702 |
4 | 0.852 | 0.765 | 0.714 |
5 | 0.844 | 0.747 | 0.706 |
6 | 0.836 | 0.754 | 0.687 |
7 | 0.848 | 0.737 | 0.732 |
8 | 0.830 | 0.761 | 0.707 |
(4) Optimizing a simulation model: performing orthogonal test on parameters of the Johnson-Cook constitutive model in the simulation model, and optimizing the simulation model; dividing 80%, 100% and 120% of parameter values in the simulation model into three levels to perform orthogonal experiments, and analyzing experiment results to obtain an optimal parameter combination; if the cosine similarity value cos (theta) of the simulation data corresponding to the optimal parameter combination and the actually measured vibration signal is more than or equal to 0.6, judging that the simulation signal is effective, and entering the step (5); otherwise, the step (4) is executed iteratively with the best parameter combination as the reference.
(5) Preparation of a complete sample: and (2) simulating other tool states (such as medium abrasion, breakage, tipping and the like) except the C tool states appearing in the step (1) by adopting the simulation model corresponding to the effective vibration simulation signal to obtain a simulation data sample in a missing state, and expanding the simulation data sample into an experimental data sample to obtain a complete sample.
(6) Preparing a complete sample: performing generative countermeasure network (GAN) training on the complete samples to generate a large number of synthetic samples; wherein, the discriminator model and the generator model of the GAN are both five-layer one-dimensional Convolutional Neural Networks (CNN), and the hidden layer neuron nodes are 1800-900-1800 and 450-900-1800 respectively; setting Gaussian distribution and uniform distribution for input noise in the antagonistic network training so as to improve the difference of synthesized samples; combining the synthesized sample and the complete sample to form a complete sample;
in this embodiment, the numbers of training sets of the actual measurement sample and the simulation sample are 900 and 560, respectively.
(7) Training a high-precision monitoring model: selecting a monitoring algorithm, further calculating 10 time domain and frequency domain statistical parameters (shown in table 2) of the complete sample, inputting the parameters into four monitoring algorithms for training, and obtaining a high-precision monitoring model;
to illustrate the effect of the present invention, the present embodiment adopts four classification algorithms: support Vector Machine (SVM), Random Forest (RF), Decision Tree (DT) and Generalized Recurrent Neural Network (GRNN), wherein:
the SVM selects a radial basis kernel function, and the penalty factor and the kernel function radius are respectively set to be 3 and 1;
the RF and DT classifiers use default parameters in the Matlab toolbox;
the value of read in the GRNN classifier is set to 0.1.
TABLE 2 time-Domain-frequency-Domain statistical parameter Table
(8) And (3) cutter classification monitoring: periodically and online collecting an actually measured vibration signal (sample to be measured) in the high-speed milling process on line, calculating 10 time domain and frequency domain statistical parameters (shown in table 2) of the sample to be measured, and classifying and monitoring the state of the tool by adopting the state monitoring model trained in the step (7);
in order to compare the advantages and disadvantages of the present invention with those of the currently popular sample enhancement method, the present embodiment selects a composite minority class oversampling technique (SMOTE) for comparison, as shown in table 3.
TABLE 3 Classification accuracy under different training sets
Fig. 1 to 4 show the detailed classification results of 5 tool classes obtained after training SVM, RF, DT and GRNN on three different training data sets after sample completion. As can be seen from the figure, on one hand, under the complete sample training obtained by the invention, the accuracy of the classification algorithms including SVM, RF and DT can reach 100%, and the accuracy of GRNN can also reach 95.2%. If more training samples are generated by GAN, the classification accuracy of GRNN is higher. SMOTE, on the other hand, obtains a somewhat larger training sample size (5500) than the present invention (5150), but with a much poorer classification accuracy. Therefore, the method provided by the invention is efficient and feasible.
The foregoing is merely an example of the present invention and common general knowledge in the art of designing and/or characterizing particular aspects and/or features is not described in any greater detail herein. It should be noted that, for those skilled in the art, without departing from the technical solution of the present invention, several variations and modifications can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicability of the patent. The scope of the claims of the present application shall be determined by the contents of the claims, and the description of the embodiments and the like in the specification shall be used to explain the contents of the claims.
Claims (10)
1. A cutter state monitoring method in a high-speed milling process is characterized by comprising the following steps:
(1) acquiring an actually measured vibration signal: determining a used cutter, a workpiece and cutting parameters according to actual machining requirements, and acquiring actual measurement vibration signals in a high-speed milling process under a plurality of cutter states;
(2) acquiring a vibration simulation signal: according to the Johnson-Cook constitutive model, a tool, a workpiece and cutting parameters adopted by an experiment, performing high-speed milling process dynamics modeling by using simulation software to obtain a simulation model, and simulating to obtain a vibration simulation signal in a normal tool state;
(3) checking the validity of the simulation sample: comparing cosine similarity values cos (theta) of the vibration simulation signal and the actually measured vibration signal; if cos (theta) is not less than 0.6, judging that the vibration simulation signal is effective, and entering the step (5), otherwise, executing the step (4);
(4) optimizing a simulation model: performing orthogonal test on parameters of the Johnson-Cook constitutive model in the simulation model in the step (2), and analyzing the test result to obtain the optimal parameter combination; if the cosine similarity value cos (theta) of the vibration simulation signal and the actually measured vibration signal corresponding to the optimal parameter combination is not less than 0.6, judging that the vibration simulation signal is effective, and entering the step (5), otherwise, iteratively executing the step (4) by using the optimal parameter combination as a reference;
(5) preparation of a complete sample: simulating the cutter state except for the step (1) by adopting a simulation model corresponding to the effective vibration simulation signal to obtain a simulation data sample of the cutter state except for the step (1), and expanding the simulation data sample into an experimental data sample to obtain a complete sample;
(6) preparing a complete sample: carrying out generative confrontation network training on the complete samples to generate a large number of synthetic samples; combining the synthesized sample and the complete sample to form a complete sample;
(7) training a high-precision monitoring model: calculating 10 time domain and frequency domain statistical parameters of a complete sample by using a monitoring algorithm, inputting the statistical parameters into the monitoring algorithm for training, and obtaining a high-precision monitoring model;
(8) and (3) cutter classification monitoring: and (3) periodically collecting actually-measured vibration signals in the high-speed milling process, calculating 10 time domain and frequency domain statistical parameters of the sample to be measured, and carrying out tool classification monitoring on the tool state by adopting the trained state monitoring model in the step (7).
2. The tool state monitoring method for the high-speed milling process according to claim 1, wherein the simulation software in the step (2) uses Abaqus or Deform.
4. the method for monitoring the cutter state in the high-speed milling process according to claim 3, wherein 80%, 100% and 120% of parameter values in the simulation model are divided into three levels in the step (4) to perform orthogonal experiments.
5. The method as claimed in claim 4, wherein the discriminator model and the generator model of the countermeasure network in step (6) are five-layer one-dimensional convolutional neural networks, and the hidden layer neuron nodes are 1800-900-1800 and 450-900-1800, respectively.
6. The tool state monitoring method for the high-speed milling process according to claim 5, wherein the anti-network training in the step (6) requires input of noise signals, and the input noise signals are set to be Gaussian distributed and uniformly distributed.
7. The tool state monitoring method for the high-speed milling process according to claim 6, wherein the monitoring algorithm in the step (7) adopts four classification algorithms: support vector machines, random forests, decision trees, and generalized recurrent neural networks.
8. The tool state monitoring method for the high-speed milling process according to claim 7, wherein the four classification algorithms in the step (7) are respectively set as follows: the support vector machine selects a radial basis kernel function, and the penalty factor and the kernel function radius are respectively set to be 3 and 1; the random forest and decision tree classifier uses default parameters in the Matlab toolbox; the value of SPREAD in the generalized recurrent neural network classifier is set to 0.1.
9. A method for monitoring tool state in high-speed milling process according to any one of claims 1-8, characterized in that the tool state in step (1) includes normal, medium and heavy wear.
10. The method for monitoring the cutter state in the high-speed milling process according to claim 9, wherein the cutter state in step (5) except step (1) comprises medium wear, breakage and chipping.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110738211.5A CN113369994A (en) | 2021-06-30 | 2021-06-30 | Cutter state monitoring method in high-speed milling process |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110738211.5A CN113369994A (en) | 2021-06-30 | 2021-06-30 | Cutter state monitoring method in high-speed milling process |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113369994A true CN113369994A (en) | 2021-09-10 |
Family
ID=77580253
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110738211.5A Pending CN113369994A (en) | 2021-06-30 | 2021-06-30 | Cutter state monitoring method in high-speed milling process |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113369994A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114161227A (en) * | 2021-12-28 | 2022-03-11 | 福州大学 | Cutter wear loss monitoring method based on simulation feature and signal feature fusion |
CN114453630A (en) * | 2022-01-20 | 2022-05-10 | 湖北文理学院 | Method and device for controlling machine tool to mill non-stick tool, electronic equipment and storage medium |
CN114659785A (en) * | 2021-12-27 | 2022-06-24 | 三一重能股份有限公司 | Fault detection method and device for transmission chain of wind driven generator |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1084795A (en) * | 1992-09-29 | 1994-04-06 | 清华大学 | A kind of cutter failure comprehensive monitoring and controlling method and device |
CN109158953A (en) * | 2018-09-04 | 2019-01-08 | 温州大学激光与光电智能制造研究院 | A kind of cutting-tool wear state on-line monitoring method and system |
CN109318055A (en) * | 2018-09-11 | 2019-02-12 | 温州大学苍南研究院 | A kind of milling cutter state of wear feature extraction Multipurpose Optimal Method |
CN111037365A (en) * | 2019-12-26 | 2020-04-21 | 大连理工大学 | Cutter state monitoring data set enhancing method based on generative countermeasure network |
CN111975453A (en) * | 2020-07-08 | 2020-11-24 | 温州大学 | Numerical simulation driven machining process cutter state monitoring method |
-
2021
- 2021-06-30 CN CN202110738211.5A patent/CN113369994A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1084795A (en) * | 1992-09-29 | 1994-04-06 | 清华大学 | A kind of cutter failure comprehensive monitoring and controlling method and device |
CN109158953A (en) * | 2018-09-04 | 2019-01-08 | 温州大学激光与光电智能制造研究院 | A kind of cutting-tool wear state on-line monitoring method and system |
CN109318055A (en) * | 2018-09-11 | 2019-02-12 | 温州大学苍南研究院 | A kind of milling cutter state of wear feature extraction Multipurpose Optimal Method |
CN111037365A (en) * | 2019-12-26 | 2020-04-21 | 大连理工大学 | Cutter state monitoring data set enhancing method based on generative countermeasure network |
CN111975453A (en) * | 2020-07-08 | 2020-11-24 | 温州大学 | Numerical simulation driven machining process cutter state monitoring method |
Non-Patent Citations (2)
Title |
---|
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT: "Sample Augmentation for Intelligent Milling Tool Wear Condition Monitoring Using Numerical Simulation and Generative Adversarial Network", 《IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》 * |
QINSONG ZHU等: "A tool wear condition monitoring approach for end milling based on numerical simulation", 《EKSPLOATACJA I NIEZAWODNOSC – MAINTENANCE AND RELIABILITY》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114659785A (en) * | 2021-12-27 | 2022-06-24 | 三一重能股份有限公司 | Fault detection method and device for transmission chain of wind driven generator |
CN114161227A (en) * | 2021-12-28 | 2022-03-11 | 福州大学 | Cutter wear loss monitoring method based on simulation feature and signal feature fusion |
CN114161227B (en) * | 2021-12-28 | 2024-05-03 | 福州大学 | Cutter abrasion loss monitoring method based on simulation feature and signal feature fusion |
CN114453630A (en) * | 2022-01-20 | 2022-05-10 | 湖北文理学院 | Method and device for controlling machine tool to mill non-stick tool, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113369994A (en) | Cutter state monitoring method in high-speed milling process | |
Tran et al. | Effective feature selection with fuzzy entropy and similarity classifier for chatter vibration diagnosis | |
Krishnakumar et al. | Tool wear condition prediction using vibration signals in high speed machining (HSM) of titanium (Ti-6Al-4 V) alloy | |
Shi et al. | A reinforced k-nearest neighbors method with application to chatter identification in high-speed milling | |
CN110647943B (en) | Cutting tool wear monitoring method based on evolution data cluster analysis | |
CN111085898A (en) | Working condition self-adaptive high-speed milling process cutter monitoring method and system | |
Li et al. | Hybrid learning for tool wear monitoring | |
Liu et al. | On-line chatter detection using servo motor current signal in turning | |
Zhu et al. | Sample augmentation for intelligent milling tool wear condition monitoring using numerical simulation and generative adversarial network | |
CN110059442A (en) | A kind of turnery processing cutter replacing method based on Part Surface Roughness and power information | |
CN107194427A (en) | A kind of milling cutter malfunction monitoring and recognition methods and system | |
CN114619292B (en) | Milling cutter wear monitoring method based on fusion of wavelet denoising and attention mechanism with GRU network | |
Patra | Acoustic emission based tool condition monitoring system in drilling | |
CN109241849B (en) | Feature decomposition selection and fault diagnosis method for main engine of intelligent power plant steam turbine | |
Kęcik et al. | Stability lobes analysis of nickel superalloys milling | |
Wang et al. | Investigation on an in-process chatter detection strategy for micro-milling titanium alloy thin-walled parts and its implementation perspectives | |
Abu-Mahfouz et al. | Surface roughness prediction in turning using three artificial intelligence techniques; a comparative study | |
Deng et al. | A hybrid model using genetic algorithm and neural network for process parameters optimization in NC camshaft grinding | |
CN116070527A (en) | Milling cutter residual life prediction method based on degradation model | |
CN111975453A (en) | Numerical simulation driven machining process cutter state monitoring method | |
Gao et al. | New tool wear estimation method of the milling process based on multisensor blind source separation | |
CN105930602A (en) | Optimal weighted wavelet package entropy-based chattering detection method | |
CN112372371A (en) | Method for evaluating abrasion state of numerical control machine tool cutter | |
Radhika et al. | Pattern recognition based surface roughness prediction in turning hybrid metal matrix composite using random forest algorithm | |
CN114888635A (en) | Cutter state monitoring method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210910 |