CN112706001A - Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network - Google Patents
Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network Download PDFInfo
- Publication number
- CN112706001A CN112706001A CN202011539912.8A CN202011539912A CN112706001A CN 112706001 A CN112706001 A CN 112706001A CN 202011539912 A CN202011539912 A CN 202011539912A CN 112706001 A CN112706001 A CN 112706001A
- Authority
- CN
- China
- Prior art keywords
- data
- bigru
- machine tool
- wear
- cnn
- 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
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000012545 processing Methods 0.000 title claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 9
- 238000013135 deep learning Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 16
- 230000004913 activation Effects 0.000 claims description 12
- 210000002569 neuron Anatomy 0.000 claims description 12
- 239000000284 extract Substances 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 10
- 238000005520 cutting process Methods 0.000 claims description 6
- 238000003754 machining Methods 0.000 claims description 6
- 238000013144 data compression Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 230000002123 temporal effect Effects 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000002759 z-score normalization Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 239000000126 substance Substances 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 230000005540 biological transmission Effects 0.000 abstract description 5
- 238000013527 convolutional neural network Methods 0.000 description 22
- 230000008901 benefit Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000005299 abrasion Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- SUBDBMMJDZJVOS-UHFFFAOYSA-N 5-methoxy-2-{[(4-methoxy-3,5-dimethylpyridin-2-yl)methyl]sulfinyl}-1H-benzimidazole Chemical compound N=1C2=CC(OC)=CC=C2NC=1S(=O)CC1=NC=C(C)C(OC)=C1C SUBDBMMJDZJVOS-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- PXDJXZJSCPSGGI-UHFFFAOYSA-N palmityl palmitate Chemical compound CCCCCCCCCCCCCCCCOC(=O)CCCCCCCCCCCCCCC PXDJXZJSCPSGGI-UHFFFAOYSA-N 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 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
Abstract
The invention relates to a machine tool cutter wear prediction method based on edge data processing and a BiGRU-CNN network, and belongs to the technical field of numerical control machine tool identification. The method comprises the following steps: firstly, data preprocessing and primary feature extraction are carried out on collected machine tool monitoring data at an edge end close to a machine tool, so that the data volume to be transmitted is reduced, the required transmission bandwidth is saved, the delay is reduced, and the prediction real-time performance is improved; secondly, carrying out standardized processing on the monitoring data at the cloud so as to input the data into a network for training and learning; and then, performing depth feature extraction by using the BiGRU-CNN at the cloud end, and constructing a tool wear monitoring model based on deep learning to predict tool wear. The invention combines the optimized BiGRU-CNN cutter wear prediction model of edge data processing, and effectively improves the real-time performance and accuracy of cutter wear prediction.
Description
Technical Field
The invention belongs to the technical field of numerical control machine tool identification, and relates to a machine tool cutter wear prediction method based on edge data processing and a BiGRU-CNN network.
Background
In recent years, machine tools are widely used for machining and manufacturing high-precision parts. The precision and surface finish of these parts can only be guaranteed with high-sharpness tools. However, the machine tool is inevitably worn due to repeated use, and the sharpness is reduced. Tool wear must be monitored and predicted in order to replace the tool before it becomes severely worn. The real-time monitoring of the wear state of the tool is mainly divided into a direct method and an indirect method. The direct method is to directly measure the abrasion value of the cutter by an external measuring method such as a microscope, and the method has the advantage of accurate measurement, but the method needs to be carried out off line, so that the continuity of the processing is interrupted and the processing time is prolonged. The indirect method is to predict the tool wear value by establishing a mapping relation with the tool wear value through physical quantities which are easy to collect and do not influence the machining process, such as vibration, force, spindle current, sound and the like, and at present, the indirect method is the mainstream selection.
Traditional indirect wear prediction methods include support vector machines, decision trees, markov models, clustering, and artificial neural networks. Xie and the like extract the characteristics of current and power signals by using principal component analysis and establish a wear state identification model of a support vector machine; liu and the like use a time domain, a frequency domain and a wavelet packet analysis method to extract machining signal characteristics and establish an autoregressive moving average model and a three-layer back propagation neural network combined model to predict cutter abrasion; ozgr Cetin et al propose a multi-scale modeling multi-ratio coupling hidden Markov model for tool wear classification; omid Geramifard et al use a multimodal hidden Markov model based tool wear monitoring; the methods are based on data driving to establish a regression model, mostly manually extract original data characteristic values reflecting the tool wear condition and input the original data characteristic values into a deep learning network for training, and the data preprocessing and the manual characteristic extraction are time-consuming and labor-consuming and can cause information loss. The deep learning method has strong self-adaptive learning capability and anti-noise capability, and can automatically extract deep features, so that a better effect tends to be achieved, and the method is more universal than the traditional machine learning method. LSTM (long short term memory network) can extract temporal features of the time series data, CNN (convolutional neural network) can extract spatial features of the data. An HDNN mixed model based on LSTM and CNN is designed by Ali Al-Dulaiim and the like; zhao et al introduced a convolutional two-way long short term memory network (CBLSTM) to address the tool wear prediction task.
Although the above method introduces various deep learning models and improves the accuracy of tool state monitoring to some extent, there is still room for improvement.
First, the complex structure of the LSTM makes it require longer training times. GRU (gated-loop unit) based on LSTM improvement overcomes this problem, with simpler unit structure and fewer parameters making GRU convergence faster. The GRU has less computation time while performing equivalent performance to the LSTM. Secondly, edge computing is used as a supplement of cloud computing, and a novel solution is provided for solving the problems of high delay, data security guarantee, bandwidth cost saving and the like. The action of data acquisition and processing of the edge nodes deployed near the edge side of the machine tool greatly reduces delay and transmission bandwidth.
Disclosure of Invention
In view of the above, the present invention aims to provide a machine tool wear prediction method based on edge data processing and a BiGRU-CNN network, which solves the problem of long training time caused by the complex LSTM structure in the conventional method. First, this problem is overcome with GRU (gated-loop unit), a simpler unit structure and fewer parameters making GRU convergence faster. The GRU has less computation time while performing equivalent performance to the LSTM. Secondly, edge computing is used as a supplement of cloud computing, and a novel solution is provided for solving the problems of high delay, data security guarantee, bandwidth cost saving and the like. The action of data acquisition and processing of the edge nodes deployed near the edge side of the machine tool greatly reduces delay and transmission bandwidth.
In order to achieve the purpose, the invention provides the following technical scheme:
a machine tool cutter wear prediction method based on edge data processing and a BiGRU-CNN network comprises the following steps:
s1: acquiring sensing data and a tool wear value in the machining process of a machine tool through a sensor and a microscope;
s2: uploading the data collected in the step S1 to an edge end, and performing data preprocessing and data decomposition on the edge end;
s3: uploading the data primarily processed in the step S2 to a cloud, performing optimal feature selection and data standardization processing, constructing a deep learning data set, and performing feature extraction by using a BiGRU-CNN network;
s4: and building a learning network, building a BiGRU-CNN network model, optimizing and predicting the wear value of the machine tool cutter.
Further, in step S1, the sensing data are cutting force, vibration and high frequency sound data, wherein the cutting force and vibration data respectively include x-axis, y-axis and z-axis data.
Further, in step S2, the data preprocessing includes: denoising, outlier rejection, data structure defragmentation and data compression; the data compression is to compress the original data by using a Huffman coding mode so as to reduce the data transmission quantity. The data decomposition is to decompose the cleaned data through four-level Discrete Wavelet Transform (DWT) so as to obtain a time-frequency signal; then, a series of statistical features such as mean, median, variance, mean square error, root mean square, kurtosis and the like are extracted according to the time-frequency signals.
Further, in step S2, each level of wavelet transform is expressed as:
y[n]＝x[Qn]
wherein x [ N ] and y [ N ] represent discrete input and output signals, and have a length of N; g [ n ], h [ n ] and Q are respectively a low-pass filter, a high-pass filter and a down-sampling filter; α represents the number of layers, L and H represent low and high frequencies, respectively, and K represents the discrete domain size.
Further, in step S3, performing optimal feature selection using the spearman correlation coefficient to obtain a feature set with better performance;
the data normalization process is a z-score normalization process, and the expression is as follows:
wherein Z isijRepresenting normalized data, XijRepresenting the original data, XiMeans, S, representing the raw dataiRepresenting the variance of the original data.
Further, in step S3, performing depth feature extraction on the input using a BiGRU-CNN network, where the CNN extracts spatial features and the BiGRU extracts temporal features;
the first structure is CNN, which contains three convolutional layers in total, and zero padding operation is used to change the dimension of feature mapping; the features output by the last convolutional layer contain useful information, which is used as input for the next BiGRU model; the convolution process is represented as:
wherein the content of the first and second substances, 
Is an input;
is the convolution kernel weight;
is an offset; f () is the activation function, all convolutional layers using ReLU as the activation function;
is the output of the l-th convolutional layer j kernel;
the second structure is BiGRU, one BiGRU layer containing 100 hidden neurons; introducing a Dropout algorithm to randomly shield the neurons in the BiGRU layer, and setting the parameters of the Dropout layer to be 0.25; discarding neurons randomly in a training phase to prevent frequent extraction of the same features; the main advantage of using BiGRU instead of the GRU layer is to process each data sequence in two opposite directions, including forward and backward, so that the complete information before and after each time step in each sequence can be accessed. In the forward direction, the information is predicted by the GRU unit, while the backward GRU is predicted smoothly and noise is removed;
the weights in the hybrid network are updated using back-propagation and stable convergence is achieved with root mean square propagation (RMSprop) as the optimization function, with the learning rate set to 0.0001. To avoid overfitting, an early stopping technique of regularization was introduced, where the maximum number of training cycles was 2000.
Further, in step S4, two full-connectivity layers are added at the end of the BiGRU-CNN network model, where the full-connectivity layer is represented as:
Xl＝σ(WlXl-1+bl)
Wherein, σ (·) is an activation function, and a Sigmoid activation function is used; wlIs a weight, blFor the bias of the l layer, the model only has one neuron and outputs a specific wear prediction value.
The invention has the beneficial effects that: the data of the method is preprocessed and compressed at the edge side, so that the data volume uploaded to the cloud is reduced; the BiGRU-CNN model is constructed at the cloud side to extract features, so that the problems of complicated task quantity of feature extraction and incomplete feature extraction are solved; and finally, better real-time prediction is carried out on the abrasion of the cutter through two full connecting layers.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.
Drawings
For the purposes of promoting a better understanding of the objects, aspects and advantages of the invention, reference will now be made to the following detailed description taken in conjunction with the accompanying drawings in which:
FIG. 1 is a block diagram of a machine tool wear prediction method of the present invention;
FIG. 2 is a flow chart of a program of a machine tool wear prediction method according to the present invention;
FIG. 3 is a schematic diagram of a BiGRU-CNN network structure according to the present invention;
fig. 4 is a diagram of four-level discrete wavelet transform according to the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention in a schematic way, and the features in the following embodiments and examples may be combined with each other without conflict.
Referring to fig. 1 to 4, fig. 1 is a frame diagram of a method for predicting wear of a tool of a machine tool according to the present invention, and fig. 2 is a flow chart of a program of the method for predicting wear of a tool of a machine tool according to the present invention; as shown in fig. 1 and 2, sensing data and a tool wear value during a machining process of a machine tool are collected through a sensor and a microscope, and the sensing data are cutting force, vibration and high-frequency sound data, wherein the cutting force and the vibration data respectively comprise x-axis data, y-axis data and z-axis data. Uploading the data to an edge side for preprocessing and compressing, wherein the preprocessing comprises denoising, outlier rejection and data structure fragment sorting; the data compression is to compress original data by using a Huffman coding mode so as to reduce the data transmission quantity. Uploading the primarily processed data to a cloud, performing data feature selection and standardization processing, constructing a deep learning data set, and performing deep feature extraction by using a BiGRU-CNN network, wherein FIG. 3 is a BiGRU-CNN network structure schematic diagram; and finally, constructing a regression model and optimizing the regression model for predicting the wear of the tool of the machine tool. The method comprises the following specific steps:
1) Carrying out data preprocessing including denoising, outlier rejection and data structure defragmentation;
2) as shown in fig. 4, decomposing the cleaned data by four-level Discrete Wavelet Transform (DWT) to obtain time-frequency signals;
each level of wavelet transform may be represented as:
y[n]＝x[Qn]
wherein x [ N ] and y [ N ] represent discrete input and output signals, and have a length of N; g [ n ], h [ n ] and Q are respectively a low-pass filter, a high-pass filter and a down-sampling filter; α represents the number of layers, L and H represent low and high frequencies, respectively, and K represents the discrete domain size.
3) Extracting time-frequency signal statistical characteristics including mean, median, variance, mean square error, root mean square and kurtosis;
4) the method comprises the steps of performing optimal feature selection by using a spearman correlation coefficient to obtain a feature set with better performance;
the spearman correlation coefficient can be expressed as:
wherein x isiRepresenting the ith extracted statistical feature, N representing the number of features, tjDenotes xiThe dependency characteristics of (a);
5) a z-score normalization process was used. The specific method comprises the following steps:
wherein Z isijRepresenting normalized data, XijRepresenting the original data, XiMeans, S, representing the raw dataiRepresenting the variance of the original data.
6) Performing feature extraction on original data by using a BiGRU-CNN network, wherein the CNN extracts spatial features and the BiGRU extracts temporal features;
7) The first structure is CNN, which contains a total of three convolutional layers, and zero-padding operations are used to change the dimension of the feature map. The features output by the last convolutional layer contain useful information that will be input to the next BiGRU model. The convolution process can be expressed as:
wherein
Is an input;
is the convolution kernel weight;
is an offset; the f () function is an activation function, all convolutional layers using the ReLU as an activation function;
is the output of the l-th convolutional layer j kernel;
8) the second structure is a BiGRU with input of CNN layer output X0～XiOne BiGRU layer contains 100 hidden neurons. The Dropout algorithm was introduced to randomly mask neurons in the BiGRU layer with the Dropout layer parameter set to 0.25. Neurons are randomly discarded during the training phase to prevent frequent extraction of the same features. The main advantage of using BiGRU instead of the GRU layer is to process each data sequence in two opposite directions, including forward and backward, so that each time step in each sequence can be accessedComplete information before and after. In the forward direction M0～MiInformation is predicted by the GRU unit, and M' is reversed0～M＇iIn the above, the GRU then smoothly predicts and removes noise.
9) The weights in the hybrid network are updated using back-propagation and trained to converge steadily at a learning rate set at 0.0001 with root mean square propagation (RMSprop) as the optimization function. To avoid overfitting, an early stopping technique of regularization was introduced, where the maximum number of training cycles was 2000.
10) Adding two full connection layers at the tail end of the BiGRU-CNN mixed model, wherein the full connection layers are expressed as
Xl＝σ(WlXl-1+bl)
Where σ (-) is the activation function, Sigmoid activation function, W is usedlIs a weight, blIs the bias of the l layer, the model only has one neuron, and outputs a specific wear prediction value.
Finally, the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all of them should be covered by the claims of the present invention.
Claims (7)
1. A machine tool cutter wear prediction method based on edge data processing and a BiGRU-CNN network is characterized by specifically comprising the following steps:
s1: acquiring sensing data and a tool wear value in the machining process of a machine tool through a sensor and a microscope;
s2: uploading the data collected in the step S1 to an edge end, and performing data preprocessing and data decomposition on the edge end;
s3: uploading the data primarily processed in the step S2 to a cloud, performing optimal feature selection and data standardization processing, constructing a deep learning data set, and performing feature extraction by using a BiGRU-CNN network;
S4: and building a learning network, building a BiGRU-CNN network model, optimizing and predicting the wear value of the machine tool cutter.
2. The method of predicting wear of a machine tool according to claim 1, wherein the sensed data is cutting force, vibration and high frequency sound data in step S1, wherein the cutting force and vibration data includes x-axis, y-axis and z-axis data, respectively.
3. The method of predicting machine tool wear of claim 1, wherein in step S2, the data preprocessing includes: denoising, outlier rejection, data structure defragmentation and data compression; the data decomposition is to decompose the cleaned data through four-level Discrete Wavelet Transform (DWT) so as to obtain a time-frequency signal; then, the statistical characteristics are extracted according to the time-frequency signals.
4. The method for predicting machine tool wear according to claim 3, wherein in step S2, each level of wavelet transform is represented as:
y[n]＝x[Qn]
wherein x [ N ] and y [ N ] represent discrete input and output signals, and have a length of N; g [ n ], h [ n ] and Q are respectively a low-pass filter, a high-pass filter and a down-sampling filter; α represents the number of layers, L and H represent low and high frequencies, respectively, and K represents the discrete domain size.
5. The method for predicting wear of a machine tool according to claim 1, wherein in step S3, a feature set is obtained by performing optimum feature selection using the spearman correlation coefficient;
the data normalization process is a z-score normalization process, and the expression is as follows:
wherein Z isijRepresenting normalized data, XijRepresenting the original data, XiMeans, S, representing the raw dataiRepresenting the variance of the original data.
6. The method of predicting machine tool wear according to claim 1, wherein in step S3, the BiGRU-CNN network is used to perform a depth feature extraction on the input, wherein the CNN extracts a spatial feature and the BiGRU extracts a temporal feature;
the first structure is CNN, which contains three convolutional layers in total, and zero padding operation is used to change the dimension of feature mapping; the features output by the last convolutional layer contain useful information, which is used as input for the next BiGRU model; the convolution process is represented as:
wherein the content of the first and second substances,
is an input;
is the convolution kernel weight;
is an offset; f () is the activation function, all convolutional layers using ReLU as the activation function;
is the output of the l-th convolutional layer j kernel;
the second structure is BiGRU, one BiGRU layer containing 100 hidden neurons; introducing Dropout algorithm to randomly shield the neurons in the BiGRU layer; discarding neurons randomly in a training phase to prevent frequent extraction of the same features; in the forward direction, the information is predicted by the GRU unit, while the backward GRU is predicted smoothly and noise is removed;
And updating the weights in the hybrid network by adopting back propagation, and taking root mean square propagation as an optimization function.
7. The method for predicting wear of a machine tool bit according to claim 1, wherein in step S4, two fully connected layers are added at the end of the BiGRU-CNN network model, and the fully connected layers are represented as:
Xl＝σ(WlXl-1+bl)
wherein, σ (·) is an activation function, and a Sigmoid activation function is used; wlIs a weight, blFor the bias of the l layer, the model only has one neuron and outputs a specific wear prediction value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011539912.8A CN112706001A (en) | 2020-12-23 | 2020-12-23 | Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011539912.8A CN112706001A (en) | 2020-12-23 | 2020-12-23 | Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112706001A true CN112706001A (en) | 2021-04-27 |
Family
ID=75543786
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011539912.8A Pending CN112706001A (en) | 2020-12-23 | 2020-12-23 | Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112706001A (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160091393A1 (en) * | 2014-09-26 | 2016-03-31 | Palo Alto Research Center Incorporated | Computer-Implemented Method And System For Machine Tool Damage Assessment, Prediction, And Planning In Manufacturing Shop Floor |
| CN108873813A (en) * | 2018-06-25 | 2018-11-23 | 山东大学 | Tool wear degree detection method based on main shaft of numerical control machine tool servo motor current signal |
| KR20190043232A (en) * | 2017-10-18 | 2019-04-26 | 주식회사 스마트랩 | Tool life maintenance system and tool life maintenance method for machine tools |
| 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 |
| CN110303380A (en) * | 2019-07-05 | 2019-10-08 | 重庆邮电大学 | A kind of cutting tool for CNC machine method for predicting residual useful life |
| CN110509109A (en) * | 2019-07-16 | 2019-11-29 | 西安交通大学 | Tool Wear Monitoring method based on multiple dimensioned depth convolution loop neural network |
| CN110561193A (en) * | 2019-09-18 | 2019-12-13 | 杭州友机技术有限公司 | Cutter wear assessment and monitoring method and system based on feature fusion |
| CN111475921A (en) * | 2020-03-13 | 2020-07-31 | 重庆邮电大学 | Tool residual life prediction method based on edge calculation and L STM network |
| CN111687689A (en) * | 2020-06-23 | 2020-09-22 | 重庆大学 | Cutter wear state prediction method and device based on LSTM and CNN |
| CN111791090A (en) * | 2020-07-02 | 2020-10-20 | 重庆邮电大学 | Cutter life abrasion judgment method based on edge calculation and particle swarm optimization |
| CN112022125A (en) * | 2020-09-28 | 2020-12-04 | 无锡博智芯科技有限公司 | Intelligent blood pressure prediction method based on CNN-BiGRU model and PPG |
| CN112100777A (en) * | 2020-11-16 | 2020-12-18 | 杭州景业智能科技股份有限公司 | Tool life prediction method and device based on edge calculation and computer equipment |
-
2020
- 2020-12-23 CN CN202011539912.8A patent/CN112706001A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160091393A1 (en) * | 2014-09-26 | 2016-03-31 | Palo Alto Research Center Incorporated | Computer-Implemented Method And System For Machine Tool Damage Assessment, Prediction, And Planning In Manufacturing Shop Floor |
| KR20190043232A (en) * | 2017-10-18 | 2019-04-26 | 주식회사 스마트랩 | Tool life maintenance system and tool life maintenance method for machine tools |
| CN108873813A (en) * | 2018-06-25 | 2018-11-23 | 山东大学 | Tool wear degree detection method based on main shaft of numerical control machine tool servo motor current signal |
| 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 |
| CN110303380A (en) * | 2019-07-05 | 2019-10-08 | 重庆邮电大学 | A kind of cutting tool for CNC machine method for predicting residual useful life |
| CN110509109A (en) * | 2019-07-16 | 2019-11-29 | 西安交通大学 | Tool Wear Monitoring method based on multiple dimensioned depth convolution loop neural network |
| CN110561193A (en) * | 2019-09-18 | 2019-12-13 | 杭州友机技术有限公司 | Cutter wear assessment and monitoring method and system based on feature fusion |
| CN111475921A (en) * | 2020-03-13 | 2020-07-31 | 重庆邮电大学 | Tool residual life prediction method based on edge calculation and L STM network |
| CN111687689A (en) * | 2020-06-23 | 2020-09-22 | 重庆大学 | Cutter wear state prediction method and device based on LSTM and CNN |
| CN111791090A (en) * | 2020-07-02 | 2020-10-20 | 重庆邮电大学 | Cutter life abrasion judgment method based on edge calculation and particle swarm optimization |
| CN112022125A (en) * | 2020-09-28 | 2020-12-04 | 无锡博智芯科技有限公司 | Intelligent blood pressure prediction method based on CNN-BiGRU model and PPG |
| CN112100777A (en) * | 2020-11-16 | 2020-12-18 | 杭州景业智能科技股份有限公司 | Tool life prediction method and device based on edge calculation and computer equipment |
Non-Patent Citations (5)
| Title |
|---|
| 刘秉儒等: "《生态数据分析与建模》", 30 December 2019, 宁夏人民教育出版社 * |
| 孙华魁: "《数据图像处理与识别技术研究》", 30 May 2019, 天津科学技术出版社 * |
| 张传雷等: "《基于图像分析的植物及其病虫害识别方法研究》", 30 October 2018, 中国经济出版社 * |
| 王丽亚等: "CNN-BiGＲU 网络中引入注意力机制的中文文本情感分析", 《计算机应用》 * |
| 陈林等: "《"互联网+"智慧校园技术与工程设施》", 30 September 2017, 电子科技大学出版社 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108319962B (en) | Tool wear monitoring method based on convolutional neural network | |
| CN109948647B (en) | Electrocardiogram classification method and system based on depth residual error network | |
| CN110502991B (en) | Internal combustion engine health monitoring method and system based on random convolutional neural network structure | |
| CN111325112A (en) | Cutter wear state monitoring method based on depth gate control circulation unit neural network | |
| WO2022057262A1 (en) | Image recognition method and device, and computer-readable storage medium | |
| CN109743356B (en) | Industrial internet data acquisition method and device, readable storage medium and terminal | |
| CN111325342A (en) | Model compression method and device, target detection equipment and storage medium | |
| CN111832432B (en) | Cutter wear real-time prediction method based on wavelet packet decomposition and deep learning | |
| CN113537063B (en) | Forest environment remote sensing monitoring method and terminal based on comprehensive remote sensing technology | |
| CN114742211B (en) | Convolutional neural network deployment and optimization method facing microcontroller | |
| CN115438714A (en) | Clustering federal learning driven mechanical fault diagnosis method, device and medium | |
| CN115204035A (en) | Generator set operation parameter prediction method and device based on multi-scale time sequence data fusion model and storage medium | |
| CN114897025A (en) | Human body posture recognition model establishing method and human body posture recognition method | |
| CN116029435A (en) | Environmental comfort early warning system is bred to live pig facility | |
| CN113313198B (en) | Cutter wear prediction method based on multi-scale convolution neural network | |
| CN114722923A (en) | Light electromechanical equipment fault diagnosis method | |
| CN114297912A (en) | Tool wear prediction method based on deep learning | |
| CN112706001A (en) | Machine tool cutter wear prediction method based on edge data processing and BiGRU-CNN network | |
| CN111523572A (en) | Real-time bridge structure damage state identification method and system | |
| CN115186574A (en) | Tool residual life prediction method based on gated cyclic residual error network | |
| CN114841412A (en) | Method for predicting pH value of sea cucumber growing water | |
| CN107374609A (en) | A kind of rhythm abnormality dynamic realtime diagnostic system | |
| CN116861204B (en) | Intelligent manufacturing equipment data management system based on digital twinning | |
| CN113344169B (en) | Novel tractor fault diagnosis system and fault diagnosis method | |
| CN115859201B (en) | Chemical process fault diagnosis method and system |
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 |
Application publication date: 20210427 |
|
| RJ01 | Rejection of invention patent application after publication |