CN115351601A - Tool wear monitoring method based on transfer learning - Google Patents
Tool wear monitoring method based on transfer learning Download PDFInfo
- Publication number
- CN115351601A CN115351601A CN202211199862.2A CN202211199862A CN115351601A CN 115351601 A CN115351601 A CN 115351601A CN 202211199862 A CN202211199862 A CN 202211199862A CN 115351601 A CN115351601 A CN 115351601A
- Authority
- CN
- China
- Prior art keywords
- working condition
- monitoring
- wear
- value
- data set
- 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.)
- Granted
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 118
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000013526 transfer learning Methods 0.000 title claims abstract description 23
- 230000006870 function Effects 0.000 claims abstract description 22
- 238000013508 migration Methods 0.000 claims abstract description 21
- 230000005012 migration Effects 0.000 claims abstract description 21
- 238000009826 distribution Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000005299 abrasion Methods 0.000 claims abstract description 8
- 238000003754 machining Methods 0.000 claims abstract description 6
- 238000005520 cutting process Methods 0.000 claims description 33
- 238000003379 elimination reaction Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 230000008030 elimination Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000001172 regenerating effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000007514 turning Methods 0.000 claims description 2
- 238000004590 computer program Methods 0.000 claims 3
- 238000003860 storage Methods 0.000 claims 3
- 238000001514 detection method Methods 0.000 claims 1
- 238000013135 deep learning Methods 0.000 abstract description 3
- 239000000463 material Substances 0.000 description 12
- 238000004519 manufacturing process Methods 0.000 description 9
- 238000011161 development Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 229910052731 fluorine Inorganic materials 0.000 description 2
- 125000001153 fluoro group Chemical group F* 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
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
- B23Q17/0957—Detection of tool breakage
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Mechanical Engineering (AREA)
- Machine Tool Sensing Apparatuses (AREA)
- Numerical Control (AREA)
Abstract
A tool wear monitoring method based on transfer learning solves the problem that a current tool wear monitoring model based on deep learning can only be modeled facing a single working condition, and belongs to the technical field of tool wear state prediction of a numerical control machining center. The invention comprises the following steps: s1, collecting monitoring signals of a machining process under an original working condition and a new working condition; s2, acquiring cutter wear value labels under an original working condition and a new working condition, and integrating the cutter wear value labels with monitoring signals to obtain an original working condition data set with the labels and a new working condition data set; s3, constructing a migration learning model based on edge distribution self-adaption; s4, training by using a data set and taking a minimum migration learning model loss function as a target to obtain a new working condition cutter abrasion monitoring model; and S5, acquiring a monitoring signal of the new working condition at the current moment, and inputting the monitoring signal into the new working condition cutter abrasion monitoring model for prediction.
Description
Technical Field
The invention belongs to the technical field of tool wear state prediction of a numerical control machining center, and particularly relates to a tool wear monitoring method based on transfer learning.
Background
The performance of the cutter is the basic guarantee of the production efficiency and the processing quality in the actual production of enterprises, and supports the production technical level and the economic benefit of the whole manufacturing field. The accurate cutter wearing and tearing value of prediction can guarantee the part machining precision when maximize cutter rate of utilization, reduction in production cost, avoids equipment to damage. In the actual production of mill, the cutter of same model often can be towards different work piece materials cutting process, because the physicochemical property of work piece material is different, the degradation process of cutter in cutting process also can present different trends, but the wearing and tearing label acquisition process is loaded down with trivial details, therefore how to utilize the monitoring model that historical data and corresponding wearing and tearing value label training obtained to adapt to the new work piece material course of working who only has a small amount of data sample rapidly and carry out the application and be the current problem that needs to solve urgently.
Under the condition of having a large amount of historical data and corresponding abrasion label values and backgrounds, monitoring data of a small amount of new workpiece material cutting processes are collected, a rapid adaptation migration method of a model is researched, the requirement that monitoring data of a large amount of new workpiece material cutting processes must be collected for reestablishing the monitoring model can be avoided, time is saved, the development efficiency of the model is improved, and the method is a development requirement of the current cutter state online monitoring technology.
With the development and application of industrial big data, the traditional cutter wear monitoring method is gradually replaced by a deep network prediction method with characteristic autonomous learning, and the advantage of deep learning on big data processing is more prominent under the background of massive machining process monitoring data. However, the current deep learning model can only be modeled for a single working condition, such as when the material of the workpiece changes, the existing model usually fails.
Disclosure of Invention
The invention provides a tool wear monitoring method based on transfer learning, aiming at the problem that the current tool wear monitoring model based on deep learning can only be modeled facing a single working condition.
The invention discloses a tool wear monitoring method based on transfer learning, which comprises the following steps:
s1, collecting monitoring signals of a machining process under an original working condition and a new working condition, wherein the monitoring signals comprise a cutting force signal, a vibration signal and an acoustic emission signal;
s2, acquiring cutter wear value labels under an original working condition and a new working condition, and integrating the cutter wear value labels with monitoring signals to obtain an original working condition data set with the labels and a new working condition data set;
s3, constructing a migration learning model based on edge distribution self-adaption, wherein the migration learning model comprises a feature adapter, two original working condition cutter wear monitoring models and two regression networks;
the original working condition cutter wear monitoring model 1 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in an original working condition data set, and the extracted features are respectively input into the regression network 1;
the original working condition cutter wear monitoring model 2 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in a new working condition data set, and the extracted features are respectively input into the two regression networks 2;
the regression network 1 is used for fitting a nonlinear relation between the input characteristics and the tool wear values in the original working condition data set;
the regression network 2 is used for fitting a nonlinear relation between the input characteristics and the tool wear value in the new working condition data set;
the characteristic adapter is used for acquiring the distribution difference of the characteristics in the original working condition data set and the characteristics in the new working condition data set;
s4, training the migration learning model based on the original working condition data set with the label and the new working condition data set by taking the minimized migration learning model loss function as a target, and taking the original working condition cutter wear monitoring model 2 and the regression network 2 obtained after training as the new working condition cutter wear monitoring model;
and S5, acquiring a monitoring signal of the new working condition at the current moment, and inputting the monitoring signal into the new working condition cutter abrasion monitoring model for prediction.
Preferably, in S4, the transfer learning model loss function is:
in the formula, D S Representing the characteristics extracted by the tool wear monitoring model 1 under the original working condition;
D T representing the characteristics extracted by the tool wear monitoring model 2 under the original working condition;
MMD 2 (D S ,D T ) Representing the difference in the distribution obtained by the feature adapter;
y S data set D representing original operating conditions S A corresponding tool wear value;
y T data set representing new conditions and D T A corresponding tool wear value;
l represents the total loss.
Preferably, the method further comprises:
and S6, forming a time sequence array by the predicted cutter wear value at the current moment predicted by the S5 and the predicted cutter wear value in a plurality of previous time periods, performing least square fitting on the formed time sequence array to obtain the predicted value at the current moment, comparing the predicted value of the least square fitting with the predicted cutter wear value of the S5, if the difference value is greater than a set threshold value, using the predicted value of the least square fitting as the predicted wear value at the current moment, otherwise, keeping the predicted cutter wear value of the S5, and turning to the S5 to predict the next moment.
Preferably, in S6, a least squares fit is performed on the time series of the composition, and the objective function and the constraint condition include;
s.t.2θ 2 x i +θ 1 ≥0
wherein, y i Representing the wear predicted value at a certain moment in the time sequence array of the wear predicted value of the cutter;
x i representing the corresponding cutting row program number; theta 0 、θ 1 、θ 2 Representing the coefficients of a quadratic polynomial.
Preferably, in S1, the cutting force signal is acquired by arranging a cutting force sensor at the bottom of the workpiece, the vibration signal is acquired by arranging a vibration sensor at the workpiece and the machine tool workbench, and the acoustic emission signal is acquired by arranging an acoustic emission sensor at the machine tool workbench.
Preferably, the monitoring signal in S2 is a preprocessed signal, and the preprocessing includes:
s21, after cutting feeding starts, carrying out monitoring signal offset elimination, setting a time period, calculating the average value of each sensor signal, taking the average value as an offset value to be subtracted from the signal, and meanwhile, calculating the standard deviation of the signal, taking the standard deviation as a standard deviation basic value, and representing the fluctuation degree of the monitoring signal when cutting is not carried out;
s22,: periodically calculating the standard deviation of the signals in the feeding process, marking the moment as a cutting starting point when the standard deviation is more than 3 times of a basic value, and clearing monitoring signals before the starting point to finish the elimination of the blank sections;
s23, intercepting the monitoring signals of the stable cutting section, and carrying out Z fraction standardization on each path of monitoring signals, wherein the expression is as follows:
wherein x represents the value of a sample point on a channel signal in a sample;
μ represents the mean of all sample points of the signal data of x channel;
σ represents the standard deviation of all sample points of the x channel signal data;
s24, performing down-sampling processing on the data after Z fraction standardization in a sliding window mode, and calculating the first four-order moment statistic of the data in the sliding window: mean, variance, skewness, and kurtosis, i.e.: and converting the monitoring signal into four-dimensional data.
Preferably, in S2, a portable microscope photographing mode is adopted to record the wear change trend of the cutter in the cutting process, the wear value of the cutter is marked by combining the size, resolution and magnification of a microscope pixel, and the wear value is integrated with the cutter monitoring signal in the corresponding time period to obtain an original working condition data set and a new working condition data set with the marks.
Preferably, in S4, the feature adapter calculates the distribution difference by using a maximum mean difference method:
wherein φ (·) represents a mapping for mapping original variables into a regenerative kernel Hilbert space;
D Si n characteristics are total to represent the ith characteristic extracted by the cutter wear monitoring model 1 under the original working condition;
D Tj representing j characteristics extracted by the original working condition cutter wear monitoring model 2, wherein the j characteristics are m characteristics;
k (-) represents a kernel function;
The invention has the beneficial effects that: the method and the device achieve the purpose of quickly adapting to the tool wear value prediction of the new workpiece material processing process by using the prediction model obtained by training under the original working condition and combining a small amount of monitoring data under the new working condition, effectively reduce the data volume requirement of the new model training, improve the development efficiency of the monitoring model and ensure the production continuity. The invention realizes the predictive maintenance of the numerical control cutter and has important significance for the actual production.
Drawings
Fig. 1 is a flowchart of a tool wear monitoring model based on transfer learning according to an embodiment of the present invention.
FIG. 2 is a schematic diagram of source domain vibration signal cancellation shear-cut according to the present invention.
Fig. 3 is a schematic diagram of a deep convolution residual network structure, where Input represents 4-dimensional time domain statistical characteristic data of an Input model, conv represents a convolution layer, BN represents a batch normalization layer, reLU represents a Linear rectification function, max Pool represents a maximum pooling layer, adaptive AvgPool represents an Adaptive average pooling layer, flatten represents a flattening layer, linear represents a Linear layer, and Output represents a predicted tool wear value Output by the model.
Fig. 4 is a schematic diagram of edge distribution adaptation used in the present invention.
FIG. 5 is a schematic diagram of a transfer learning model according to the present invention.
FIG. 6 is a flow chart illustrating smoothing of predicted wear values according to the present invention.
FIG. 7 is a graph of the results of the transfer learning model training of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the embodiments and features of the embodiments of the present invention may be combined with each other without conflict.
The invention is further described with reference to the following drawings and specific examples, which are not intended to be limiting.
The tool wear monitoring method based on transfer learning of the embodiment comprises the following steps:
step 3, constructing a migration learning model based on edge distribution self-adaption, wherein the migration learning model comprises a feature adapter, two original working condition cutter wear monitoring models and two regression networks;
the original working condition cutter wear monitoring model 1 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in an original working condition data set, and the extracted features are respectively input into the regression network 1;
the original working condition cutter abrasion monitoring model 2 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in a new working condition data set, and the extracted features are respectively input into the two regression networks 2;
the regression network 1 is used for fitting a nonlinear relation between the input characteristics and the tool wear values in the original working condition data set;
the regression network 2 is used for fitting a nonlinear relation between the input characteristics and the tool wear value in the new working condition data set;
the characteristic adapter is used for acquiring the distribution difference of the characteristics in the original working condition data set and the characteristics in the new working condition data set;
step 4, training the migration learning model based on the original working condition data set with the label and the new working condition data set by taking the minimized migration learning model loss function as a target, and taking the original working condition cutter wear monitoring model 2 and the regression network 2 obtained after training as the new working condition cutter wear monitoring model;
and 5, acquiring a monitoring signal of the new working condition at the current moment, and inputting the monitoring signal into the new working condition cutter abrasion monitoring model for prediction.
The method utilizes the idea of transfer learning, maximally utilizes the internal similarity between prediction tasks, finds the internal relation and difference of tool degradation trends under two working conditions, and utilizes an algorithm to perform characteristic edge distribution adaptation under the two working conditions, thereby obtaining a prediction model applied to a newly processed material. The method comprises the steps of collecting monitoring signals under the working conditions of original workpiece materials and new workpiece materials, obtaining tool wear labels under the original working conditions and a small number of new working conditions, and integrating to obtain an original working condition data set with the labels and a new working condition data set; a transfer learning model is built, and then a monitoring model under a new working condition is built through transfer learning, so that the requirement on labeled data of a target domain is reduced, and meanwhile, the problem of generalization application of an original cutter abrasion monitoring model when the material of a workpiece changes is solved.
In the preferred embodiment, in step 1, the cutting force signal is acquired by arranging a cutting force sensor at the bottom of the workpiece, the vibration signal is acquired by arranging a vibration sensor on the workpiece and the machine tool workbench, and the acoustic emission signal is acquired by arranging an acoustic emission sensor on the machine tool workbench.
In the preferred embodiment, in step 2, the monitoring signal is a preprocessed signal, a wear change trend of the cutter in the cutting process is recorded by adopting a portable microscope photographing mode, the wear value of the cutter is marked by combining the size, resolution and magnification of a microscope pixel, and the wear value is integrated with the cutter monitoring signal in the corresponding time period to obtain an original working condition data set with a label and a new working condition data set.
The pretreatment step comprises:
and step 21, monitoring signal offset elimination is carried out after the cutting feed is started. Calculating the average value of the sensor signals in a 120ms time period after the start of the feeding motion, taking the average value as an offset value to be subtracted from the signals, and meanwhile, calculating the standard deviation of the signals in the process, taking the standard deviation as a standard deviation basic value, and representing the fluctuation degree of the monitoring signals when the cutting is not performed;
step 22, calculating the standard deviation once every 400ms in the feeding process, marking the moment as a cutting starting point when the standard deviation is 3 times larger than a basic value, and clearing monitoring signals before the starting point to finish the elimination of the blank cut segment, wherein the elimination process of the blank cut segment is shown in fig. 2;
step 23, intercepting the monitoring signal of the stable cutting section, reserving a sampling point with the monitoring signal length of 100000, and performing Z fraction standardization on each path of monitoring signal, wherein the expression is as follows:
wherein x represents the value of a sample point on a channel signal in a sample;
μ represents the mean of all sample points of x this channel signal data;
σ represents the standard deviation of all sample points of the x channel signal data;
step 24, performing down-sampling processing on the data after Z fraction standardization in a sliding window mode, and calculating the first four-order moment statistic of the data in the sliding window: mean, variance, skewness, and kurtosis, i.e.: converting the monitoring signal into four-dimensional data, thereby reducing the length of the sample and reducing the calculated amount of the model, wherein the calculation formula of the statistic comprises:
wherein,
representing the mean of the data; x is a radical of a fluorine atom ms Representing the variance of the data; α represents a data skewness; β represents the kurtosis of the data; x is a radical of a fluorine atom i An ith sample point representing a certain path of sensing signal; n representsThe sample length of the sliding window.
In the preferred embodiment, the original working condition cutter wear monitoring model in step 3 adopts a deep convolution residual network to perform high-dimensional embedding space feature extraction on the input signal, the network is mainly formed by stacking residual basic blocks, the parameter setting of the whole framework of the residual convolution neural network is shown in table 1, and the network structure is shown in fig. 3;
TABLE 1 residual neural network architecture settings
The feature adapter adopts the maximum mean difference as the measurement of the distribution difference of the source domain feature and the target domain feature, and the adopted edge distribution self-adaption principle is shown in FIG. 4;
the feature adapter calculates the distribution difference by adopting a maximum mean difference method:
wherein φ (·) represents a mapping for mapping original variables into a regenerating kernel Hilbert space;
D Si n characteristics are total to represent the ith characteristic extracted by the cutter wear monitoring model 1 under the original working condition;
D Tj representing j characteristics extracted by the original working condition cutter wear monitoring model 2, wherein the j characteristics are m characteristics; k (-) represents a kernel function;
the calculation of the sum of squares of the functions in hilbert space is indicated.
The kernel function selects a Gaussian radial basis function, limited dimensional data are mapped to a high dimensional space, and a calculation formula of the Gaussian radial basis function comprises the following steps:
in the formula, x' represents a kernel function central point;
x represents any point in space;
σ denotes the bandwidth, which is used to control the radial reach of the function.
The regression network is formed by a fully connected neural network and is used for fitting the nonlinear relation between the high-dimensional embedding characteristics and the tool wear value.
The principle of the transfer learning model is shown in fig. 5.
In a preferred embodiment, the migration learning model loss function in step 4 is:
in the formula D S Representing the characteristics extracted by the tool wear monitoring model 1 under the original working condition;
D T representing the characteristics extracted by the tool wear monitoring model 2 under the original working condition;
MMD 2 (D S ,D T ) Representing the difference in the distribution obtained by the feature adapter;
y S representing the sum of the original operating conditions data set and D S A corresponding tool wear value;
y T data set D representing new operating conditions T A corresponding tool wear value;
l represents the total loss.
In a preferred embodiment, the method of this embodiment further includes:
as shown in fig. 6, step 6, the predicted value of the wear of the tool at the current time predicted in step 5 and the predicted value of the wear of the tool in the previous time periods are combined into a time sequence number sequence, the combined time sequence number sequence is subjected to least square fitting to obtain the predicted value at the current time, the predicted value of the least square fitting is compared with the predicted value of the wear of the tool in step 5, if the difference value is greater than a set threshold value, the predicted value of the least square fitting is used as the predicted wear value at the current time, otherwise, the predicted value of the wear of the tool in step 5 is retained, and the next time prediction is performed in step 5.
Step 6, performing least square fitting on the formed time sequence series, wherein the target function and the constraint condition comprise;
s.t.2θ 2 x i +θ 1 ≥0
wherein, y i Representing the wear predicted value at a certain moment in the time sequence series of the wear predicted value of the cutter;
x i representing the corresponding cutting row program number; theta.theta. 0 、θ 1 、θ 2 Representing the coefficients of a quadratic polynomial.
The tool wear state monitoring model obtained by the migration learning is used for monitoring the tool wear state in the new workpiece material cutting process, as shown in fig. 7, the prediction result approaches to the true value of tool wear, which shows that the tool wear state monitoring model provided by the invention has the capability of working condition migration, can accurately predict the tool wear state under the new working condition, and has important significance for actual production.
Although the invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that features described in different dependent claims and herein may be combined in ways different from those described in the original claims. It is also to be understood that features described in connection with individual embodiments may be used in other described embodiments.
Claims (10)
1. A tool wear monitoring method based on transfer learning, the method comprising:
s1, collecting monitoring signals of a machining process under an original working condition and a new working condition, wherein the monitoring signals comprise a cutting force signal, a vibration signal and an acoustic emission signal;
s2, acquiring cutter wear value labels under the original working condition and the new working condition, and integrating the labels with the monitoring signals to obtain an original working condition data set with the labels and a new working condition data set;
s3, constructing a migration learning model based on edge distribution self-adaption, wherein the migration learning model comprises a feature adapter, two original working condition cutter wear monitoring models and two regression networks;
the original working condition cutter wear monitoring model 1 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in an original working condition data set, and the extracted features are respectively input into the regression network 1;
the original working condition cutter wear monitoring model 2 is used for carrying out high-dimensional embedding space feature extraction on monitoring signals in a new working condition data set, and the extracted features are respectively input into the two regression networks 2;
the regression network 1 is used for fitting a nonlinear relation between the input characteristics and the tool wear values in the original working condition data set;
the regression network 2 is used for fitting a nonlinear relation between the input characteristics and the tool wear value in the new working condition data set;
the characteristic adapter is used for acquiring the distribution difference of the characteristics in the original working condition data set and the characteristics in the new working condition data set;
s4, training the migration learning model based on the original working condition data set with the label and the new working condition data set by taking the minimized migration learning model loss function as a target, and taking the original working condition cutter wear monitoring model 2 and the regression network 2 obtained after training as the new working condition cutter wear monitoring model;
and S5, acquiring a monitoring signal of the new working condition at the current moment, and inputting the monitoring signal into the new working condition cutter abrasion monitoring model for prediction.
2. The tool wear monitoring method based on the transfer learning of claim 1, wherein in S4, the transfer learning model loss function is:
in the formula, D S Representing the characteristics extracted by the tool wear monitoring model 1 under the original working condition;
D T representing the characteristics extracted by the tool wear monitoring model 2 under the original working condition;
MMD 2 (D S ,D T ) Representing the difference in the distribution obtained by the feature adapter;
y S representing the sum of the original operating conditions data set and D S A corresponding tool wear value;
y T data set representing new conditions and D T A corresponding tool wear value;
l represents the total loss.
3. The tool wear monitoring method based on transfer learning of claim 1, further comprising:
and S6, forming a time sequence array by the predicted value of the cutter wear at the current moment predicted by the S5 and the predicted value of the cutter wear in a plurality of previous time periods, performing least square fitting on the formed time sequence array to obtain the predicted value at the current moment, comparing the predicted value of the least square fitting with the predicted value of the cutter wear of the S5, if the difference value is greater than a set threshold value, using the predicted value of the least square fitting as the predicted wear value at the current moment, otherwise, keeping the predicted value of the cutter wear of the S5, and turning to the S5 to predict the next moment.
4. The tool wear monitoring method based on the transfer learning of claim 1, wherein in S6, least squares fitting is performed on the time series of the composition, and the objective function and the constraint condition include;
s.t.2θ 2 x i +θ 1 ≥0
wherein, y i Representing the wear predicted value at a certain moment in the time sequence array of the wear predicted value of the cutter;
x i representing the corresponding cutting row program number; theta 0 、θ 1 、θ 2 Representing the coefficients of a quadratic polynomial.
5. The tool wear monitoring method based on the migration learning of claim 1, wherein in S1, the cutting force signal is acquired by arranging a cutting force sensor at the bottom of the workpiece, the vibration signal is acquired by arranging a vibration sensor at the workpiece and the machine tool worktable, and the acoustic emission signal is acquired by arranging an acoustic emission sensor at the machine tool worktable.
6. The tool wear monitoring method based on the transfer learning of claim 5, wherein the monitoring signal in S2 is a preprocessed signal, and the preprocessing comprises:
s21, after cutting feeding starts, carrying out monitoring signal offset elimination, setting a time period, calculating the average value of each sensor signal, taking the average value as an offset value to be subtracted from the signal, and meanwhile, calculating the standard deviation of the signal, taking the standard deviation as a standard deviation basic value, and representing the fluctuation degree of the monitoring signal when cutting is not carried out;
s22 and: periodically calculating the standard deviation of the signals in the feeding process, marking the moment as a cutting starting point when the standard deviation is more than 3 times of a basic value, and clearing monitoring signals before the starting point to finish the elimination of the blank cutting section;
s23, intercepting the monitoring signals of the stable cutting section, and carrying out Z fraction standardization on each path of monitoring signals, wherein the expression is as follows:
wherein x represents the value of a sample point on a channel signal in a sample;
μ represents the mean of all sample points of x this channel signal data;
σ represents the standard deviation of all sample points of the x channel signal data;
s24, performing down-sampling processing on the data after Z fraction standardization in a sliding window mode, and calculating the first four-order moment statistic of the data in the sliding window: mean, variance, skewness, and kurtosis, i.e.: and converting the monitoring signal into four-dimensional data.
7. The tool wear monitoring method based on the migration learning of claim 6, wherein in S2, a portable microscope photographing mode is adopted to record the wear change trend of the tool during the cutting process, the tool wear value is marked by combining the size, resolution and magnification of a microscope pixel, and then the wear value is integrated with the tool monitoring signal of the corresponding time period to obtain an original working condition data set and a new working condition data set with labels.
8. The tool wear monitoring method based on the transfer learning of claim 1, wherein in S4, the feature adapter calculates the distribution difference by a maximum mean difference method:
wherein φ (·) represents a mapping for mapping original variables into a regenerative kernel Hilbert space;
D Si representing the ith characteristic extracted by the tool wear monitoring model 1 under the original working condition, wherein the ith characteristic is n characteristics;
D Tj representing j characteristics extracted by the original working condition cutter wear monitoring model 2, wherein the j characteristics are m characteristics; k (-) represents a kernel function;
the calculation of the sum of squares of the functions in hilbert space is indicated.
9. A computer-readable storage device, in which a computer program is stored, which, when being executed, implements a tool wear monitoring method based on migration learning according to any one of claims 1 to 8.
10. A brake adjuster control rod round pin and round pin cotter pin loss detection device, comprising a storage device, a processor and a computer program stored in the storage device and capable of running on the processor, wherein the processor executes the computer program to realize the tool wear monitoring method based on the transfer learning of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211199862.2A CN115351601B (en) | 2022-09-29 | 2022-09-29 | Tool wear monitoring method based on transfer learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211199862.2A CN115351601B (en) | 2022-09-29 | 2022-09-29 | Tool wear monitoring method based on transfer learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115351601A true CN115351601A (en) | 2022-11-18 |
| CN115351601B CN115351601B (en) | 2024-05-28 |
Family
ID=84008301
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211199862.2A Active CN115351601B (en) | 2022-09-29 | 2022-09-29 | Tool wear monitoring method based on transfer learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115351601B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115847187A (en) * | 2023-02-27 | 2023-03-28 | 成都大金航太科技股份有限公司 | Real-time monitoring system for deep and narrow groove turning |
| CN116540627A (en) * | 2023-02-07 | 2023-08-04 | 广东工业大学 | Machine tool thermal error prediction compensation group control method and system based on deep transfer learning |
| CN116992954A (en) * | 2023-09-26 | 2023-11-03 | 南京航空航天大学 | UMAP data dimension reduction-based similarity measurement transfer learning method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005125480A (en) * | 2003-09-30 | 2005-05-19 | Allied Material Corp | Cutting method having function for diagnosing tool life |
| CN102091972A (en) * | 2010-12-28 | 2011-06-15 | 华中科技大学 | Numerical control machine tool wear monitoring method |
| CN102765010A (en) * | 2012-08-24 | 2012-11-07 | 常州大学 | Cutter damage and abrasion state detecting method and cutter damage and abrasion state detecting system |
| CN108942409A (en) * | 2018-08-26 | 2018-12-07 | 西北工业大学 | The modeling and monitoring method of tool abrasion based on residual error convolutional neural networks |
| CN109822399A (en) * | 2019-04-08 | 2019-05-31 | 浙江大学 | Cutting tool for CNC machine state of wear prediction technique based on parallel deep neural network |
| CN112192318A (en) * | 2020-09-28 | 2021-01-08 | 北京天泽智云科技有限公司 | Machining tool state monitoring method and system |
| CN112518425A (en) * | 2020-12-10 | 2021-03-19 | 南京航空航天大学 | Intelligent machining cutter wear prediction method based on multi-source sample migration reinforcement learning |
-
2022
- 2022-09-29 CN CN202211199862.2A patent/CN115351601B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005125480A (en) * | 2003-09-30 | 2005-05-19 | Allied Material Corp | Cutting method having function for diagnosing tool life |
| CN102091972A (en) * | 2010-12-28 | 2011-06-15 | 华中科技大学 | Numerical control machine tool wear monitoring method |
| CN102765010A (en) * | 2012-08-24 | 2012-11-07 | 常州大学 | Cutter damage and abrasion state detecting method and cutter damage and abrasion state detecting system |
| CN108942409A (en) * | 2018-08-26 | 2018-12-07 | 西北工业大学 | The modeling and monitoring method of tool abrasion based on residual error convolutional neural networks |
| CN109822399A (en) * | 2019-04-08 | 2019-05-31 | 浙江大学 | Cutting tool for CNC machine state of wear prediction technique based on parallel deep neural network |
| CN112192318A (en) * | 2020-09-28 | 2021-01-08 | 北京天泽智云科技有限公司 | Machining tool state monitoring method and system |
| CN112518425A (en) * | 2020-12-10 | 2021-03-19 | 南京航空航天大学 | Intelligent machining cutter wear prediction method based on multi-source sample migration reinforcement learning |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116540627A (en) * | 2023-02-07 | 2023-08-04 | 广东工业大学 | Machine tool thermal error prediction compensation group control method and system based on deep transfer learning |
| CN116540627B (en) * | 2023-02-07 | 2024-04-12 | 广东工业大学 | Machine tool thermal error prediction compensation group control method and system based on deep transfer learning |
| CN115847187A (en) * | 2023-02-27 | 2023-03-28 | 成都大金航太科技股份有限公司 | Real-time monitoring system for deep and narrow groove turning |
| CN115847187B (en) * | 2023-02-27 | 2023-05-05 | 成都大金航太科技股份有限公司 | Real-time monitoring system for deep and narrow groove turning |
| CN116992954A (en) * | 2023-09-26 | 2023-11-03 | 南京航空航天大学 | UMAP data dimension reduction-based similarity measurement transfer learning method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115351601B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115351601A (en) | Tool wear monitoring method based on transfer learning | |
| CN108907896B (en) | A kind of cutter remaining life on-line prediction method and system | |
| CN107877262B (en) | A kind of numerical control machine tool wear monitoring method based on deep learning | |
| CN115035120B (en) | Machine tool control method and system based on Internet of things | |
| CN110488754B (en) | Machine tool self-adaptive control method based on GA-BP neural network algorithm | |
| CN113780153B (en) | Cutter wear monitoring and predicting method | |
| DE102020001169A1 (en) | Device and system for setting machining conditions | |
| CN111113150B (en) | Method for monitoring state of machine tool cutter | |
| CN112051799A (en) | Self-adaptive control method for machining | |
| CN113798920A (en) | Cutter wear state monitoring method based on variational automatic encoder and extreme learning machine | |
| US10890890B2 (en) | Smart adjustment system and method thereof | |
| CN114749996A (en) | Tool residual life prediction method based on deep learning and time sequence regression model | |
| CN110109431B (en) | Intelligent acquiring system for OEE information of die casting machine | |
| CN116187725A (en) | Forging equipment management system for forging automatic line | |
| CN115716218B (en) | Numerical control cutter state on-line monitoring system | |
| Xiao et al. | Deep learning based modeling for cutting energy consumed in CNC turning process | |
| CN117150235A (en) | Cutter life prediction method by cooperation of monitoring data and degradation model | |
| CN113293270B (en) | Blade surface ultrasonic rolling reinforcement self-adaptive control system and method | |
| CN116992920A (en) | Short-term load prediction method of power system | |
| Li et al. | A deep learning based method for cutting parameter optimization for band saw machine | |
| CN111611946A (en) | Micro milling cutter abrasion on-line monitoring method based on principal component analysis and BP neural network | |
| CN118092314B (en) | Method for adjusting operation parameters of machine tool, operation parameter adjusting device and storage medium | |
| CN116912241B (en) | CNC machine adjustment optimization method and system based on machine learning | |
| CN118276507B (en) | Method and system for optimizing turning tool path of large vertical lathe | |
| CN117921197B (en) | Method, system, equipment and medium for manufacturing special-shaped groove plate by laser precision cutting |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |