CN110000610A - A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network - Google Patents
A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network Download PDFInfo
- Publication number
- CN110000610A CN110000610A CN201910306988.7A CN201910306988A CN110000610A CN 110000610 A CN110000610 A CN 110000610A CN 201910306988 A CN201910306988 A CN 201910306988A CN 110000610 A CN110000610 A CN 110000610A
- Authority
- CN
- China
- Prior art keywords
- tool wear
- depth confidence
- sensor information
- value
- confidence network
- 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
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
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network that the present invention relates to a kind of, comprise the following steps that acquisition numerical-controlled machine tool machining process in multiple sensors signal, respectively extract sensor information time domain, frequency domain, time-frequency domain characteristic parameter.The characteristic parameter that can characterize cutter wear is effectively extracted by limitation Boltzmann machine (RBM)；By stacking multiple limitation Boltzmann machine (RBM) construction depth confidence networks (DBN), the monitoring of tool wear is realized.The present invention can get rid of the dependence to signal processing expertise, obtain cutting-tool wear state details by multiple sensors, fast and accurately identify the cutter wear state under different processing conditions, have the features such as monitoring accuracy is high, adaptable.
Description
Technical field
The invention belongs to numerical control machine tool wear monitoring technical fields, more specifically, being related to a kind of based on more sensings
The Tool Wear Monitoring method of the fusion of device information and depth confidence network.
Technical background
Tool wear is to influence the key factor of workpiece quality in processing industry, efficiently and accurately predicts that tool wear can
So that cutter is replaced in time, to avoid unnecessary waste；Studies have shown that CNC machine can be reduced after being equipped with tool monitoring system
The 75% of downtime, production efficiency improve 10-60%, and machine tool utilization rate improves 50%, stable, accurate tool monitoring
System is that modernization processing is essential.
When monitoring cutting tool state, the behaviour in service of cutter can be monitored using multiple sensors according to workplace, it is existing
Some methods are usually certain characteristic parameter using single-sensor signal to indicate the state of wear of cutter, monitoring it is accurate
Property be limited to the precision of a certain sensor, monitoring stability is poor, cannot effectively realize the monitoring of cutting tool state.
Instantly in manufacturing industry, sensor information is pre-processed, feature extraction, feature selecting depend on technology people
The signal processing technology and diagnostic experiences of member, is much not achieved intelligentized requirement, this field needs a kind of stabilization, accurately intelligence
Tool Wear Monitoring system can be changed.
Summary of the invention
Technical problems to be solved:
In view of the drawbacks of the prior art or Improvement requirement, the invention proposes one kind based on information fusion and depth confidence net
The Tool Wear Monitoring method of network is based on existing tool condition monitoring feature, studies and devises and is a kind of based on letter
The Tool Wear Monitoring method of breath fusion and depth confidence network；The monitoring method has merged the effective of multiple sensors signal
Information, and go out using depth confidence network online recognition the state of wear of cutter, it realizes and the stabilization of cutting tool state is supervised online
It surveys；This method can get rid of the dependence to signal processing technology and diagnostic experiences, realize the extracted in self-adaptive of tool wear feature,
And monitoring accuracy and flexibility are improved, monitoring is no longer limited by some sensor signal.
Technical solution:
Step 1: under a certain operating condition, material being processed using constant cutting parameter, cutter milling on side
Processing measures Cutting Force Signal, vibration signal during this, while measuring the tool flank wear of cutter after processing every time,
And using the knife face attrition value after normalized as the output valve of neural network；
Step 2: extracting characteristic parameter of each sensor signal on time domain, frequency domain and time-frequency domain respectively, and carry out
Normalized；
Step 3: using the characteristic parameter after normalized as the input vector of depth confidence network (DBN) to identification mould
Type is trained, and in the last layer plus softmax classifier, is finely adjusted by BP algorithm, final output cutting-tool wear state.
Further, the sensor signal includes vibration signal and Cutting Force Signal.
Further, the sensor signal of extraction the characteristic parameter of time domain include mean value, mean-square value, root mean value, absolutely
To mean value, absolute value summation, maximum value self-contained mean value, wave height rate, waveform rate, standard deviation, flexure.
Further, the sensor signal of extraction the characteristic parameter of time domain include power spectrum mean value, power spectrum degree of skewness,
Power spectrum kurtosis, power spectrum pulse crest value, power spectrum variance, power spectrum pulse index.
Further, when extracting characteristic parameter on frequency domain, by vibration signal by discrete Fourier transform to obtain function
Rate spectrum；Wherein, the change of the main band position of the gravity frequency, the square frequency and the root mean square frequency representation power spectrum
Change situation, the frequency variance and the frequency standard difference indicate the dispersion degree of energy spectrum.
Further, when time and frequency domain characteristics are extracted, Cutting Force Signal, vibration are believed using tri- layers of WAVELET PACKET DECOMPOSITION of db5
Number progress is decomposed and reconstituted, extracts its 8 wavelet-packet energy bands as time-frequency characteristics.
Further, normalized, formula are done to the characteristic parameter extracted using Min-Max Normalization
Are as follows:
In formula, x is input value；Y is normalized output value；Min is minimum value；Max is maximum value.
Further, depth confidence network (DBN) is made of several layers neuron, and constituent element is limited Boltzmann machine
(RBM), training process includes pre-training, fine tuning and prediction.
Further, it is limited in Boltzmann machine (RBM), setting aobvious layer has nvA neuron, hidden layer have nhA nerve
Member, v=(v1,v2,v3…,vnv)TIt is visible layer state vector (characteristic parameter of vibrating sensor and force snesor information), it is hidden
Layer state vector is h=(h1,h2,h3,...,hnh)T,There is vi, hj∈ { 0,1 },To show once
Bias vector,For the bias vector of hidden layer, W=(wi,j) weight matrix between hidden layer and aobvious layer,
Remember that θ=(W, a, b), the probability that each neuron is activated are shown below:
For training sampleWherein nsFor the feature ginseng of sensor information in S group cutting process
Number,Wherein S is characterized the spy for extracting sensor in tool cutting process
Parameter is levied, the purpose of training RBM is exactly to maximize following likelihood function:
The purpose for maximizing formula (1) is can to carry out stochastic gradient descent method to its negative to find optimal parameter θ
To determine the value:
<·>dataIndicate the mathematic expectaion to data distribution,<>modelIndicate the mathematic expectaion to model profile, <
>modelComputational complexity beWhen being sampled with the Gibbs method of sampling, with sample<>modelTo progress
Estimation, needs a large amount of sample to be just able to satisfy precision, greatly increases the complexity of RBM；The present invention is used to sdpecific dispersion
Algorithm (Contrastive Divergence, CD) trains RBM, Gibbs to sample from given visible layer in conjunction with Gibbs sampling
Data (v0), start the initial value (h for calculating hidden layer neuron0), then pass through (h0) calculate (v1), so circulation executes K times
Gibbs sampling can obtain<>when theoretically K approach is infinitemodelExact value, however only several steps in practice
Meet demand.
The utility model has the advantages that
The Tool Wear Monitoring method based on information fusion and depth confidence network that the invention proposes a kind of, acquires numerical control
The multiple sensors signal of lathe, the characteristic parameter on time domain, frequency domain and time-frequency domain, and the characteristic parameter extracted is done and is returned
One change processing improves flexibility and monitoring accuracy；DBN network overcomes conventional tool monitoring technology to signal processing technology
With the dependence of diagnostic experiences, can adaptive extraction can characterize the sensor information feature of cutting-tool wear state and need not rely on
Expertise, arithmetic speed is fast, reduces influence of the manual features to result to the full extent, has multi-level feature representation energy
Power obtains more abstract data characteristics, improves the validity of data characteristics.
Detailed description of the invention
Fig. 1 is tool wear on-line monitoring method flow chart.
Fig. 2 is WAVELET PACKET DECOMPOSITION structure chart.
Fig. 3 is RBM network structure.
Fig. 4 DBN neural network pre-training process.
Fig. 5 is cd-k algorithm.
Fig. 6 is DBN, ANN, SVM training error comparison diagram.
Fig. 7 is DBN, ANN, SVM training time comparison diagram.
Specific embodiment
Now in conjunction with embodiment, attached drawing, the invention will be further described:
The present invention proposes a kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network, attached
Fig. 1 shows the Tool Wear Monitoring method detailed processes, mainly include three steps:
Step 1: under a certain operating condition, part being processed using constant cutting parameter, milling cutter milling on workpiece
Radial distance 1mm, the side of axial cutting-in 2mm 70 times are processed, cutter becomes blunt abrasion from running-in wear, measures processing every time
The tool flank wear of cutter afterwards, Kistler acceleration transducer are taken the vibration signal in three directions, are used simultaneously
KIStler9123C rotation dynamometer measures the cutting force in three directions during this, extracts wherein 50 groups of tool wear measurements
Value is trained as the output valve of DNN neural network；Deflection can be reduced using constant cut parameter, reduces and calculates hardly possible
Degree, and rotate dynamometer with dynamometry signal stabilization, it is easy for installation, the advantages that strong antijamming capability.
Step 2: extracting characteristic parameter of each sensor signal on time domain, frequency domain and time-frequency domain respectively, and carry out
Normalized.
Specifically, multiple sensors signal is acquired, to each sensor signal respectively on time domain, frequency domain and time-frequency domain
Characteristic parameter is extracted, and all characteristic parameters after extraction are done into normalized.In present embodiment, in the spy that time domain is extracted
Levying parameter includes mean value, mean-square value, root mean value, absolute mean, absolute value summation, maximum value self-contained mean value, wave height rate, wave
Form quotient, standard deviation, flexure；When time domain extracts characteristic parameter using the complete empirical mode decomposition method based on adaptive noise
(CEEMDAN), the temporal signatures of extraction are the energy value of mode function.
Wherein, (1) mean value:
(2) mean-square value:
(3) root mean value:
(4) absolute value summation:
(5) maximum:
max(Si)
(6) self-contained mean value:
(7) wave height rate:
(8) waveform rate:
(9) standard deviation:
(10) flexure:
(11) absolute mean:
When extracting feature on frequency domain, each sensor following six characteristic parameter is extracted as frequency domain character.
(1) power spectrum mean value:
(2) power spectrum degree of skewness:
(3) power spectrum kurtosis:
(4) power spectrum pulse crest value:
max(S(f))
(5) power spectrum variance:
(6) power spectrum pulse index:
When time and frequency domain characteristics are extracted, Cutting Force Signal, vibration signal are decomposed using tri- layers of WAVELET PACKET DECOMPOSITION of db5
Reconstruct:
Wherein, n is Frequency Index, and k is positioning index, and j is scale index, wnReferred to as about the Orthogonal Wavelet Packet of ψ (t)
Base, by the scaling function ψ (t) of an orthonormalization, w0=ψ (t), by double scale difference recursion equation groups, generating function group:
Wherein, hk,gkFor a pair of conjugate quadrature mirror filter coefficient derived from ψ (t).
Wavelet packet can be with the increase of resolution ratio 2j, and it is excellent that there is the spectral window to broaden further segmentation to attenuate
Quality combines, signal can be divided to any frequency by orthogonal filter hk, gk given signal by one group of low high pass
Duan Shang makes it in low frequency and high frequency time with higher and frequency resolution；The invention patent is first by processed
Cutting Force Signal in journey carries out three layers of WAVELET PACKET DECOMPOSITION, and cutting force relevant to abrasion is extracted from wavelet packet coefficient reconstruct image
The structure of feature and cutting vibration feature, WAVELET PACKET DECOMPOSITION process such as attached drawing 2 is divided, and j indicates Decomposition order；It is acted on
Be: the Cutting Force Signal of sample frequency 8kHZ can be subdivided into 8 sections by three layers of WAVELET PACKET DECOMPOSITION on frequency domain, in the time domain will
The wavelet packet coefficient of each frequency range is reconstructed, and using the energy value of this signal as tool wear characteristic value.
In present embodiment, normalized is done to the characteristic parameter extracted using Min-Max Normalization,
Formula are as follows:
In formula, x is input value；Y is normalized output value；Min is minimum value；Max is maximum value.
Step 3: using the 50 groups of features obtained from step 2 as the input of DNN network, by 50 groups of cutter flanks of measurement
State of wear is trained network as the output end of DNN neural network, real as Tool Wear Monitoring by remaining 20 groups
It tests；Cutting force, the vibration performance totally 150 of input carry out feature extraction by DBN neural network, and Fig. 3 is RBM network composition；
Using DBN neural network, the feature that can characterize cutter molding is successively extracted, completes pre-training, Fig. 4 is pre-training procedure chart；
Then entire network parameter, final output cutting-tool wear state are finely tuned by BP algorithm.
In limited Boltzmann machine (RBM), setting aobvious layer has nvA neuron, hidden layer have nhA neuron,It is visible layer state vector (characteristic parameter of vibrating sensor and force snesor information), hidden layer shape
State vector is h=(h1,h2,h3,...,hnh)T,There is vi,hj∈{0,1},For aobvious biasing once
Vector,For the bias vector of hidden layer, W=(wi,j) weight matrix between hidden layer and aobvious layer, note θ=
(W, a, b), the probability that each neuron is activated are shown below:
For training sampleWherein nsFor the feature ginseng of sensor information in S group cutting process
Number,Wherein S is characterized the spy for extracting sensor in tool cutting process
Parameter is levied, the purpose of training RBM is exactly to maximize following likelihood function:
The purpose for maximizing formula (1) is can to carry out stochastic gradient descent method to its negative to find optimal parameter θ
To determine the value:
<·>dataIndicate the mathematic expectaion to data distribution,<>modelIndicate the mathematic expectaion to model profile, <
>modelComputational complexity beWhen being sampled with the Gibbs method of sampling, with sample<>modelTo progress
Estimation, needs a large amount of sample to be just able to satisfy precision, greatly increases the complexity of RBM；The present invention is used to sdpecific dispersion
Algorithm (Contrastive Divergence, CD) trains RBM in conjunction with Gibbs sampling, as shown in Fig. 5, Gibbs sampling from
Given visible layer data (v0), start the initial value (h for calculating hidden layer neuron0), then pass through (h0) calculate (v1), such as
This circulation executes 3 Gibbs samplings, can obtain<>when theoretically K approach is infinitemodelExact value, however in practice
Only 3 steps can meet demand.
In conjunction with attached drawing 6, attached drawing 7, using two different monitoring method support vector machines (SVM) and artificial neural network
(ANN network) and the present invention are a kind of to be carried out based on multi-sensor information fusion and the Tool Wear Monitoring method of depth confidence network
Contrast verification, comparing result show that method accuracy rate and speed provided by the invention are all higher, are indicated above this method stability
More preferably, it is indicated above this patent stability and is more preferably better than other two methods.
In conclusion Tool Wear Monitoring method proposed by the present invention can be adaptive extraction sensor in tool wear
Condition information gets rid of the dependence to a large amount of signal processing knowledge and diagnosis engineering experience, and achieves higher monitoring accuracy,
And arithmetic speed can more accurately identify cutting-tool wear state when in face of complicated processing environment.
Claims (5)
1. a kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network, it is characterised in that step
It is as follows:
Step 1: under a certain operating condition, material is processed using constant cutting parameter, cutter Milling Process on side,
Cutting Force Signal, the vibration signal during this are measured, while measuring the tool flank wear of cutter after processing every time, and will return
One changes treated output valve of the knife face attrition value as neural network；
Step 2: extracting characteristic parameter of each sensor signal on time domain, frequency domain and time-frequency domain respectively, and carry out normalizing
Change processing；
Step 3: using the characteristic parameter after normalized as the input vector of depth confidence network (DBN) to identification model into
Row training, in the last layer plus softmax classifier, is finely adjusted, final output cutting-tool wear state by BP algorithm.
2. a kind of Tool Wear Monitoring side based on multi-sensor information fusion and depth confidence network as described in claim 1
Method, it is characterised in that: depth confidence network (DBN) is by several RBM " series connection ", wherein the hidden layer of a upper RBM is
For the aobvious layer of next RBM, the output of a upper RBM is the input of next RBM.
3. a kind of Tool Wear Monitoring side based on multi-sensor information fusion and depth confidence network as described in claim 1
Method, it is characterised in that: the sensor signal includes vibration signal, Cutting Force Signal.
4. a kind of Tool Wear Monitoring side based on multi-sensor information fusion and depth confidence network as described in claim 1
Method, it is characterised in that: the sensor signal of extraction includes in the characteristic parameter of time domain, mean value, mean-square value, root mean value, absolutely
Mean value, absolute value summation, maximum value self-contained mean value, wave height rate, waveform rate, standard deviation, flexure.
5. a kind of Tool Wear Monitoring side based on multi-sensor information fusion and depth confidence network as described in claim 1
Method, it is characterised in that: the sensor signal of extraction includes power spectrum mean value, power spectrum degree of skewness, function in the characteristic parameter of frequency domain
Rate composes kurtosis, power spectrum pulse crest value, power spectrum variance, power spectrum pulse index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910306988.7A CN110000610A (en) | 2019-04-17 | 2019-04-17 | A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910306988.7A CN110000610A (en) | 2019-04-17 | 2019-04-17 | A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110000610A true CN110000610A (en) | 2019-07-12 |
Family
ID=67172644
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910306988.7A Pending CN110000610A (en) | 2019-04-17 | 2019-04-17 | A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110000610A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110346130A (en) * | 2019-07-19 | 2019-10-18 | 北京理工大学 | A kind of boring flutter detection method based on empirical mode decomposition and time-frequency multiple features |
| 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 |
| CN111325112A (en) * | 2020-01-31 | 2020-06-23 | 贵州大学 | Cutter wear state monitoring method based on depth gate control circulation unit neural network |
| CN111488968A (en) * | 2020-03-03 | 2020-08-04 | 国网天津市电力公司电力科学研究院 | Method and system for extracting comprehensive energy metering data features |
| CN111761409A (en) * | 2020-07-09 | 2020-10-13 | 内蒙古工业大学 | Multi-sensor numerical control machine tool cutter wear monitoring method based on deep learning |
| CN111881860A (en) * | 2020-07-31 | 2020-11-03 | 重庆理工大学 | Modeling method of hob abrasion in-situ recognition model and hob abrasion in-situ recognition method |
| CN112192250A (en) * | 2020-10-20 | 2021-01-08 | 衢州学院 | Information acquisition and fusion device and method based on thermal design of spindle system |
| CN112365935A (en) * | 2020-10-20 | 2021-02-12 | 燕山大学 | Cement free calcium soft measurement method based on multi-scale depth network |
| WO2021174525A1 (en) * | 2020-03-06 | 2021-09-10 | 大连理工大学 | Parts surface roughness and cutting tool wear prediction method based on multi-task learning |
| CN114888635A (en) * | 2022-04-27 | 2022-08-12 | 哈尔滨理工大学 | Cutter state monitoring method |
| CN115157006A (en) * | 2022-09-08 | 2022-10-11 | 中科航迈数控软件(深圳)有限公司 | Component wear detection method and device based on information entropy, terminal and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB8323284D0 (en) * | 1982-09-03 | 1983-10-05 | Valeron Corp | Worn tool detector |
| JPH10143218A (en) * | 1996-11-08 | 1998-05-29 | Nissan Motor Co Ltd | Cycle time prediction device for robot |
| US20060089743A1 (en) * | 2004-10-25 | 2006-04-27 | Ford Motor Company | Data management and networking system and method |
| US7117050B2 (en) * | 2002-11-08 | 2006-10-03 | Toshiba Kikai Kabushiki Kaisha | Management supporting apparatus, management supporting system, management supporting method, management supporting program, and a recording medium with the program recorded therein |
| CN104157290A (en) * | 2014-08-19 | 2014-11-19 | 大连理工大学 | Speaker recognition method based on depth learning |
| CN105196114A (en) * | 2015-11-05 | 2015-12-30 | 西安科技大学 | Real-time online tool wear monitoring method based on wavelet analysis and neural network |
| CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Adaptive deep confidence network bearing fault diagnosis method based on Nesterov momentum method |
| CN108107838A (en) * | 2017-12-27 | 2018-06-01 | 山东大学 | A kind of numerical control equipment tool wear monitoring method based on cloud knowledge base and machine learning |
| CN108710771A (en) * | 2018-05-31 | 2018-10-26 | 西安电子科技大学 | Mechanized equipment service reliability appraisal procedure based on the integrated extraction of depth characteristic |
| CN108747590A (en) * | 2018-06-28 | 2018-11-06 | 哈尔滨理工大学 | A kind of tool wear measurement method based on rumble spectrum and neural network |
| CN108830308A (en) * | 2018-05-31 | 2018-11-16 | 西安电子科技大学 | A kind of Modulation Identification method that traditional characteristic signal-based is merged with depth characteristic |
| CN109434564A (en) * | 2018-12-21 | 2019-03-08 | 哈尔滨理工大学 | A kind of cutter wear state monitoring method based on deep neural network |
-
2019
- 2019-04-17 CN CN201910306988.7A patent/CN110000610A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB8323284D0 (en) * | 1982-09-03 | 1983-10-05 | Valeron Corp | Worn tool detector |
| JPH10143218A (en) * | 1996-11-08 | 1998-05-29 | Nissan Motor Co Ltd | Cycle time prediction device for robot |
| US7117050B2 (en) * | 2002-11-08 | 2006-10-03 | Toshiba Kikai Kabushiki Kaisha | Management supporting apparatus, management supporting system, management supporting method, management supporting program, and a recording medium with the program recorded therein |
| US20060089743A1 (en) * | 2004-10-25 | 2006-04-27 | Ford Motor Company | Data management and networking system and method |
| CN104157290A (en) * | 2014-08-19 | 2014-11-19 | 大连理工大学 | Speaker recognition method based on depth learning |
| CN105196114A (en) * | 2015-11-05 | 2015-12-30 | 西安科技大学 | Real-time online tool wear monitoring method based on wavelet analysis and neural network |
| CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Adaptive deep confidence network bearing fault diagnosis method based on Nesterov momentum method |
| CN108107838A (en) * | 2017-12-27 | 2018-06-01 | 山东大学 | A kind of numerical control equipment tool wear monitoring method based on cloud knowledge base and machine learning |
| CN108710771A (en) * | 2018-05-31 | 2018-10-26 | 西安电子科技大学 | Mechanized equipment service reliability appraisal procedure based on the integrated extraction of depth characteristic |
| CN108830308A (en) * | 2018-05-31 | 2018-11-16 | 西安电子科技大学 | A kind of Modulation Identification method that traditional characteristic signal-based is merged with depth characteristic |
| CN108747590A (en) * | 2018-06-28 | 2018-11-06 | 哈尔滨理工大学 | A kind of tool wear measurement method based on rumble spectrum and neural network |
| CN109434564A (en) * | 2018-12-21 | 2019-03-08 | 哈尔滨理工大学 | A kind of cutter wear state monitoring method based on deep neural network |
Non-Patent Citations (3)
| Title |
|---|
| 付洋: "切削加工过程中振动状态及刀具磨损的智能监测技术研究", 《中国优秀博士学位论文全文数据库工程科技Ⅰ辑》 * |
| 李威霖: "车铣刀具磨损状态监测及预测关键技术研究", 《中国优秀博士学位论文全文数据库工程科技Ⅰ辑》 * |
| 林杨: "基于深度学习的刀具磨损状态监测技术的研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110346130A (en) * | 2019-07-19 | 2019-10-18 | 北京理工大学 | A kind of boring flutter detection method based on empirical mode decomposition and time-frequency multiple features |
| 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 |
| CN111325112A (en) * | 2020-01-31 | 2020-06-23 | 贵州大学 | Cutter wear state monitoring method based on depth gate control circulation unit neural network |
| CN111325112B (en) * | 2020-01-31 | 2023-04-07 | 贵州大学 | Cutter wear state monitoring method based on depth gate control circulation unit neural network |
| CN111488968A (en) * | 2020-03-03 | 2020-08-04 | 国网天津市电力公司电力科学研究院 | Method and system for extracting comprehensive energy metering data features |
| WO2021174525A1 (en) * | 2020-03-06 | 2021-09-10 | 大连理工大学 | Parts surface roughness and cutting tool wear prediction method based on multi-task learning |
| US11761930B2 (en) | 2020-03-06 | 2023-09-19 | Dalian University Of Technology | Prediction method of part surface roughness and tool wear based on multi-task learning |
| CN111761409A (en) * | 2020-07-09 | 2020-10-13 | 内蒙古工业大学 | Multi-sensor numerical control machine tool cutter wear monitoring method based on deep learning |
| CN111881860B (en) * | 2020-07-31 | 2022-05-03 | 重庆理工大学 | Modeling method of hob abrasion in-situ recognition model and hob abrasion in-situ recognition method |
| CN111881860A (en) * | 2020-07-31 | 2020-11-03 | 重庆理工大学 | Modeling method of hob abrasion in-situ recognition model and hob abrasion in-situ recognition method |
| CN112365935A (en) * | 2020-10-20 | 2021-02-12 | 燕山大学 | Cement free calcium soft measurement method based on multi-scale depth network |
| CN112365935B (en) * | 2020-10-20 | 2022-08-30 | 燕山大学 | Cement free calcium soft measurement method based on multi-scale depth network |
| CN112192250A (en) * | 2020-10-20 | 2021-01-08 | 衢州学院 | Information acquisition and fusion device and method based on thermal design of spindle system |
| CN114888635A (en) * | 2022-04-27 | 2022-08-12 | 哈尔滨理工大学 | Cutter state monitoring method |
| CN114888635B (en) * | 2022-04-27 | 2023-07-25 | 哈尔滨理工大学 | Cutter state monitoring method |
| CN115157006A (en) * | 2022-09-08 | 2022-10-11 | 中科航迈数控软件(深圳)有限公司 | Component wear detection method and device based on information entropy, terminal and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110000610A (en) | A kind of Tool Wear Monitoring method based on multi-sensor information fusion and depth confidence network | |
| Li et al. | The entropy algorithm and its variants in the fault diagnosis of rotating machinery: A review | |
| CN105834835B (en) | A kind of tool wear on-line monitoring method based on Multiscale Principal Component Analysis | |
| CN108356606B (en) | Tool wear online monitoring method based on wavelet packet analysis and RBF neural network | |
| CN105275833B (en) | CEEMD (Complementary Empirical Mode Decomposition)-STFT (Short-Time Fourier Transform) time-frequency information entropy and multi-SVM (Support Vector Machine) based fault diagnosis method for centrifugal pump | |
| CN109916628B (en) | Rolling bearing fault diagnosis method based on improved multi-scale amplitude perception permutation entropy | |
| Wang et al. | A new tool wear monitoring method based on multi-scale PCA | |
| CN109434564A (en) | A kind of cutter wear state monitoring method based on deep neural network | |
| CN110070060B (en) | Fault diagnosis method for bearing equipment | |
| CN111761409A (en) | Multi-sensor numerical control machine tool cutter wear monitoring method based on deep learning | |
| CN106002490B (en) | Milling workpiece roughness monitoring method based on Path and redundant eliminating | |
| CN111413089A (en) | Gear fault diagnosis method based on combination of VMD entropy method and VPMCD | |
| CN105134619A (en) | Failure diagnosis and health evaluation method based on wavelet power, manifold dimension reduction and dynamic time warping | |
| CN113177537B (en) | Fault diagnosis method and system for rotary mechanical equipment | |
| CN113420691A (en) | Mixed domain characteristic bearing fault diagnosis method based on Pearson correlation coefficient | |
| CN106842922A (en) | A kind of NC Machining Error optimization method | |
| CN110346130B (en) | Boring flutter detection method based on empirical mode decomposition and time-frequency multi-feature | |
| CN112101227A (en) | Mechanical state monitoring method based on CELMDAN and SSKFDA | |
| CN114346761B (en) | Cutter abrasion condition detection method based on improved condition generation countermeasure network | |
| CN114492527A (en) | Fuzzy neural network and principal component analysis based surface roughness online prediction method | |
| CN109590805A (en) | A kind of determination method and system of turning cutting tool working condition | |
| CN109434562A (en) | Milling cutter state of wear recognition methods based on partition clustering | |
| CN115375026A (en) | Method for predicting service life of aircraft engine in multiple fault modes | |
| CN114714145A (en) | Method for enhancing, comparing, learning and monitoring tool wear state by using Gelam angular field | |
| Wang et al. | Cutting force embedded manifold learning for condition monitoring of vertical machining center |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190712 |
|
| WD01 | Invention patent application deemed withdrawn after publication |