CN114970643A - High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm - Google Patents
High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm Download PDFInfo
- Publication number
- CN114970643A CN114970643A CN202210862028.0A CN202210862028A CN114970643A CN 114970643 A CN114970643 A CN 114970643A CN 202210862028 A CN202210862028 A CN 202210862028A CN 114970643 A CN114970643 A CN 114970643A
- Authority
- CN
- China
- Prior art keywords
- umap
- dimension reduction
- data
- algorithm
- fault identification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a UMAP dimension reduction algorithm, which is applied to a method for identifying faults of a high-speed motorized spindle.
Description
Technical Field
The patent relates to the field of high-speed electric spindle fault identification, in particular to high-speed electric spindle fault identification based on vibration signal characteristic dimension reduction.
Background
The high-speed electric spindle runs at high speed and high load for a long time, abrasion or fatigue is inevitably generated, and if the high-speed electric spindle cannot be diagnosed in time, a product machined by a high-end precision machine tool is unqualified, and even the whole system cannot work normally. Therefore, the fault can be found as soon as possible, and the method has important significance for avoiding catastrophic accidents and ensuring the safe operation of machinery.
Fault identification techniques can be mainly divided into three major categories: analytical model based methods, signal processing based methods and knowledge based intelligent fault diagnosis methods. Along with the rapid development of computer technology, the efficiency and the recognition accuracy of the intelligent fault recognition technology are continuously improved, and the method provided by the patent mainly reduces the dimensionality of data, so that fault types are more easily distinguished.
Disclosure of Invention
The existing intelligent fault identification technology has large data calculation amount, and aims to improve the identifiability and the simplicity of data, so that the invention provides a data dimension reduction method applied to high-speed electric spindle fault identification, and dimension reduction processing of high-speed electric spindle vibration signals is realized through a UMAP dimension reduction algorithm, so that data are easier to distinguish. The method can be used in any dimension reduction processing scene of the vibration signal, and comprises the following steps:
step 1: reading a data set production training set and a verification set of vibration signals of different faults in a database, acquiring a vibration signal generation test set of the high-speed motorized spindle to be detected, and generating an initial data set from the training set and the test set;
step 2: processing the initial data set by utilizing time domain analysis and frequency domain analysis to obtain an initial feature set;
and step 3: carrying out normalization processing on the initial feature set, and clustering the obtained data through a UMAP dimension reduction algorithm;
and 4, step 4: and identifying the clustered data through an artificial intelligence identification algorithm, thereby realizing the fault identification of the high-speed motorized spindle.
Further describing step 2, the time domain features obtained by using the time domain analysis include: peak-to-peak, root-mean-square, standard deviation, waveform index, peak index, pulse index, kurtosis index, margin factor, etc.; the frequency domain features obtained using the frequency domain analysis include center of gravity frequency, mean square frequency, root mean square frequency, frequency variance, frequency standard deviation, and the like. And combining the characteristic value set obtained by time domain analysis with the characteristic value set obtained by frequency domain analysis to generate an initial characteristic value set.
In order to simplify the initial characteristic value set, firstly, the characteristic value set is normalized, and the initial characteristic set normalization calculation formula:
wherein x * And x is a characteristic value for the normalized value.
Further describing step 3, the specific steps of the UMAP dimension reduction algorithm are as follows:
using the algorithm of finding nearest neighbor to obtain each x i K nearest neighbor set { x } i1 ,x i2 ，…，x ik Using k nearest neighbor set, x can be determined i Rho of i And σ i ，ρ i And σ i The calculation formula of (2):
ρ i ＝min{d(x i ,x ij )|1≤j≤k,d(x i ,x ij )＞0}
let N high dimensional data { x 1 ,x 2 ，…，x N In a high-dimensional space, similarity is a local fuzzy simple set membership v based on smooth nearest neighbor distance i|j The calculation formula of (2):
v i|j ＝exp[(-d(x i ,x j )-ρ j )/σ j ]
in the formula: d (x) i ,x j ) Denotes x i And x j The distance between them.
After using UMAP to stick together points with locally varying metrics, it may happen that the weights between two nodes are not equal, so the high-dimensional data needs to be symmetric, expressed as:
v ij ＝v j|i +v i|j -v j|i v i|j
let N low-dimensional spatial similarity points be { y 1 ,y 2 ,…,y N The similarity points of low-dimensional data can be expressed as:
in the formula: a and b are user-defined positive values, and the default values using this procedure in a UMAP implementation are a-1.929 and b-0.7915.
Simulation correctness of a simulation data point in a low-dimensional space corresponding to a high-dimensional space point takes cross entropy as a cost function, and the cross entropy is expressed as:
C UMAP the smaller the value, the higher the correctness of the high-dimensional data of the low-dimensional data fitting, if C UMAP When the probability distribution is equal to 0, the minimum value of the cost function is obtained for the higher correctness of the formula simulation, and the gradient is utilizedOptimizing the cost function by a descending method, and obtaining a result { y after dimension reduction 1 ,y 2 ,…,y N }. In order to improve the dimension reduction effect, repeated iterative operation can be carried out on the original data, and the correctness of the low-dimensional space simulation data is improved.
Further describing step 4, the UMAP is used for obtaining the data after dimensionality reduction, and an intelligent identification algorithm (BP neural network, RBF, GRNN, PNN neural network, competitive neural network, support vector machine and the like) is used for training the data after dimensionality reduction and identifying the fault of the high-speed electric spindle. Therefore, the fault identification of the high-speed electric spindle is realized.
The dimension reduction processing of the high-speed motorized spindle vibration data provided by the invention can fully reserve the difference characteristics of different fault types and reserve the similarity characteristics of the same fault type, thereby improving the data conciseness and the data identifiability.
Drawings
The foregoing and other advantages of the invention will become more apparent from the following detailed description of the invention when taken in conjunction with the accompanying drawings.
FIG. 1 is a partial vibration data set used in the present invention.
Fig. 2 is a flow chart of the fault identification of the present invention.
FIG. 3 is a diagram of the effect of the features of the UMAP algorithm after dimensionality reduction. .
Detailed Description
The technical solution in the embodiment of the present invention will be described below with reference to the drawings in the embodiment of the present invention.
FIG. 2 is a flow chart of the present invention, comprising the steps of:
constructing a high-speed electric main shaft fault identification characteristic database
Extracting time domain characteristics and frequency domain characteristics of each vibration signal waveform from the similar vibration data vibration signal waveforms in FIG. 1, wherein the time domain characteristics and the frequency domain characteristics comprise peak-to-peak values, root-mean-square values, standard deviations, waveform indexes, peak indexes, pulse indexes, kurtosis indexes, margin factors, center-of-gravity frequencies, mean-square frequencies, root-mean-square frequencies, frequency variances, frequency standard deviations and the likeInto an initial feature set { x 1 ,x 2 ，…，x N }. The fault identification database of the high-speed motorized spindle is formed as { x i1 ,x i2 ，…，x ik And h, wherein i is the waveform number of the vibration signal, and k is the characteristic number of the high-speed electric spindle.
2. Intelligent fault identification technology for high-speed electric spindle
2.1, clustering and dividing multi-dimensional vibration characteristics of high-speed electric spindle vibration signals
Using the algorithm of finding nearest neighbor to obtain each x i K nearest neighbor set { x } i1 ,x i2 ，…，x ik Using k nearest neighbor set, x can be determined i Rho of i And σ i ，ρ i And σ i Represented by the formula:
ρ i ＝min{d(x i ,x ij )|1≤j≤k,d(x i ,x ij )＞0}
2.2 visualization of Fault identification
And a UMAP dimensionality reduction algorithm is selected to cluster the high-dimensional characteristic quantities and visualize the clustering result, and the UMAP is subjected to two-dimensional visualization research to find out the intrinsic membership of the data through the probability distribution of random walk on the field diagram. Through visual display, the division effect of the fault recognition can be clearly and intuitively displayed as shown in fig. 3.
2.3 Intelligent fault identification method
And inputting the training set in the data subjected to dimensionality reduction into a GA-SVM neural network for intelligent learning, acquiring a training set model to obtain a neural network classifier, and then inputting a verification sample set or a test sample set to realize fault identification of the high-speed motorized spindle.
The scheme can rapidly and clearly identify various fault states, and achieves rapid and accurate fault identification of the high-speed electric spindle.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (3)
1. A high-speed electric spindle fault identification method based on UMAP dimension reduction is characterized by comprising the following steps:
step 1: reading a data set production training set of vibration signals of different faults in a database, acquiring a vibration signal generation test set of the high-speed motorized spindle to be detected, and generating an initial data set from the training set and the test set;
and 2, step: processing the initial data set by utilizing time domain analysis and frequency domain analysis to obtain an initial feature set;
and step 3: carrying out normalization processing on the initial feature set, and clustering and reducing dimensions of the obtained data through a UMAP dimension reduction algorithm;
and 4, step 4: and identifying the clustered data through an artificial intelligence identification algorithm, thereby realizing the fault identification of the high-speed motorized spindle.
2. The method according to claim 1, characterized in that step 3, comprises the steps of:
and selecting UMAP dimension reduction, performing dimension reduction on a clustering result of the high-dimensional characteristic quantity by using a clustering algorithm, and performing two-dimensional visual research on UMAP to find out the intrinsic membership of data through probability distribution.
3. The method as claimed in claim 1, wherein the UMAP dimension reduction algorithm is applied to the fault recognition of the high-speed motorized spindle in combination with GA-SVM recognition technology.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2021111912627 | 2021-10-13 | ||
| CN202111191262.7A CN113935375A (en) | 2021-10-13 | 2021-10-13 | High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114970643A true CN114970643A (en) | 2022-08-30 |
Family
ID=79279098
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111191262.7A Pending CN113935375A (en) | 2021-10-13 | 2021-10-13 | High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm |
| CN202210862028.0A Pending CN114970643A (en) | 2021-10-13 | 2022-07-20 | High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111191262.7A Pending CN113935375A (en) | 2021-10-13 | 2021-10-13 | High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN113935375A (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115331752B (en) * | 2022-07-22 | 2024-03-05 | 中国地质大学（北京） | Method capable of adaptively predicting quartz forming environment |
| CN115204319A (en) * | 2022-09-15 | 2022-10-18 | 广东电网有限责任公司中山供电局 | Low-voltage distribution network topology parameter identification method and system |
| 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 (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106769049A (en) * | 2017-01-18 | 2017-05-31 | 北京工业大学 | A kind of Fault Diagnosis of Roller Bearings based on Laplce's score value and SVMs |
| CN108073158A (en) * | 2017-12-05 | 2018-05-25 | 上海电机学院 | Based on PCA and KNN density algorithm Wind turbines Method for Bearing Fault Diagnosis |
| CN109829402A (en) * | 2019-01-21 | 2019-05-31 | 福州大学 | Different operating condition lower bearing degree of injury diagnostic methods based on GS-SVM |
| CN110866502A (en) * | 2019-11-19 | 2020-03-06 | 安徽工业大学 | Fault diagnosis method based on linear discriminant analysis and particle swarm optimization support vector machine |
| CN113478477A (en) * | 2021-06-08 | 2021-10-08 | 上海交通大学 | Robot state monitoring method and system based on multiple sensors and data transmission |
-
2021
- 2021-10-13 CN CN202111191262.7A patent/CN113935375A/en active Pending
-
2022
- 2022-07-20 CN CN202210862028.0A patent/CN114970643A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106769049A (en) * | 2017-01-18 | 2017-05-31 | 北京工业大学 | A kind of Fault Diagnosis of Roller Bearings based on Laplce's score value and SVMs |
| CN108073158A (en) * | 2017-12-05 | 2018-05-25 | 上海电机学院 | Based on PCA and KNN density algorithm Wind turbines Method for Bearing Fault Diagnosis |
| CN109829402A (en) * | 2019-01-21 | 2019-05-31 | 福州大学 | Different operating condition lower bearing degree of injury diagnostic methods based on GS-SVM |
| CN110866502A (en) * | 2019-11-19 | 2020-03-06 | 安徽工业大学 | Fault diagnosis method based on linear discriminant analysis and particle swarm optimization support vector machine |
| CN113478477A (en) * | 2021-06-08 | 2021-10-08 | 上海交通大学 | Robot state monitoring method and system based on multiple sensors and data transmission |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113935375A (en) | 2022-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110132598B (en) | Fault noise diagnosis algorithm for rolling bearing of rotating equipment | |
| CN114970643A (en) | High-speed electric spindle fault identification method based on UMAP dimension reduction algorithm | |
| Zhang et al. | Feature selection for high-dimensional machinery fault diagnosis data using multiple models and radial basis function networks | |
| CN109781411B (en) | Bearing fault diagnosis method combining improved sparse filter and KELM | |
| Baarsch et al. | Investigation of internal validity measures for K-means clustering | |
| US7889914B2 (en) | Automated learning of model classifications | |
| Guo et al. | Online remaining useful life prediction of milling cutters based on multisource data and feature learning | |
| CN109582003A (en) | Based on pseudo label semi-supervised kernel part Fei Sheer discriminant analysis bearing failure diagnosis | |
| CN107590506A (en) | A kind of complex device method for diagnosing faults of feature based processing | |
| Arbin et al. | Comparative analysis between k-means and k-medoids for statistical clustering | |
| CN110647943A (en) | Cutting tool wear monitoring method based on evolutionary data cluster analysis | |
| CN111367777B (en) | Alarm processing method, device, equipment and computer readable storage medium | |
| CN106482967A (en) | A kind of Cost Sensitive Support Vector Machines locomotive wheel detecting system and method | |
| CN115221930A (en) | Fault diagnosis method for rolling bearing | |
| Guo et al. | Fault diagnosis for power system transmission line based on PCA and SVMs | |
| CN115587543A (en) | Federal learning and LSTM-based tool residual life prediction method and system | |
| CN109434562A (en) | Milling cutter state of wear recognition methods based on partition clustering | |
| Colliri et al. | A network-based high level data classification technique | |
| Lei et al. | Fault diagnosis of rotating machinery based on a new hybrid clustering algorithm | |
| CN108985462B (en) | Unsupervised feature selection method based on mutual information and fractal dimension | |
| Radovanović et al. | Application of agglomerative hierarchical clustering for clustering of time series data | |
| Feng et al. | Temporal local correntropy representation for fault diagnosis of machines | |
| Sánchez et al. | Applicability of cluster validation indexes for large data sets | |
| CN114781448A (en) | Bearing fault feature extraction method, system, medium and equipment | |
| CN112059725A (en) | Cutter wear monitoring method based on EMD-SVM |
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 |