CN108062564B - Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification - Google Patents
Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification Download PDFInfo
- Publication number
- CN108062564B CN108062564B CN201711318123.XA CN201711318123A CN108062564B CN 108062564 B CN108062564 B CN 108062564B CN 201711318123 A CN201711318123 A CN 201711318123A CN 108062564 B CN108062564 B CN 108062564B
- Authority
- CN
- China
- Prior art keywords
- vector machine
- support vector
- particle
- kernel
- imf
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (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)
- Control Of Electric Motors In General (AREA)
Abstract
The invention relates to a method for optimizing a multi-core multi-feature fusion support vector machine for bearing fault identification, which comprises the steps of selecting a bearing vibration signal collected under a single sensor; carrying out EMD decomposition on bearing vibration signals at different rotating speeds to obtain IMF energy entropy and IMF arrangement entropy; extracting IMF energy entropy and IMF permutation entropy at different rotating speeds for fusion to obtain fusion characteristics containing different rotating speed information, and using the fusion characteristics to support vector machine training samples to obtain a multi-core multi-characteristic fusion support vector machine suitable for fault recognition at different rotating speeds; the performance of a Gaussian radial basis function kernel and the performance of a polynomial function kernel are integrated, training samples are subjected to linear regression which is mapped to a high-dimensional space through a nonlinear function space, so that the training samples are classified according to different characteristics to form a multi-core least square support vector machine, and the support vector machine can identify fault characteristics under variable loads; and (3) performing parameter optimization on the training sample by adopting a self-adjusting particle swarm algorithm with strong convergence, and then comparing the training sample with the test sample to perform bearing fault identification.
Description
Technical Field
The invention relates to a rotary mechanical fault identification method, in particular to a self-adjusting particle swarm optimization multi-core multi-feature fusion support vector machine for high-precision bearing fault identification.
Background
Large rotating machinery is usually placed in severe environments and in areas with rare people, and a general fault detection method needs personnel to periodically obtain characteristic information and perform complex analysis and processing, so that a large amount of manpower and material resources are consumed. The support vector machine is used as an intelligent mode identification method, and does not need personnel to keep on for a long time or manually analyze fault positions, so that the support vector machine is widely used for rotary mechanical fault diagnosis. For example, the invention patent with publication number CN107065568A proposes a transformer fault diagnosis method based on a particle swarm support vector machine, and the invention patent with publication number CN103679263A proposes a lightning approach prediction method based on a particle swarm support vector machine; the invention patent with the publication number of CN101655456A provides a particle swarm support vector machine-based high-voltage insulator equivalent salt deposit density optical fiber detection method.
Firstly, a polynomial kernel function kernel in a kernel function of a support vector machine is a global kernel, has strong generalization capability, but has weak learning capability; the Gaussian radial basis function kernel is a local kernel, and has good learning ability but weak generalization ability. Secondly, the pattern recognition information of the support vector machine under a single feature is not sufficient, for example, the pattern recognition under variable load, and more complete information must be obtained by fusing the features. Yu Kun fuses energy entropies obtained by EEMD decomposition of information under a plurality of sensors as feature vectors. The multi-sensor collects fault signals comprehensively, but a multi-position sensor is required to be arranged. When the acceleration sensor is used for detecting the vibration signal of the bearing, the installation position is close to the bearing as much as possible. Multiple sensors are cumbersome to install and typically do not have multiple available measurement locations.
Therefore, further optimization improvement is needed to be performed on the particle swarm optimization support vector machine so as to obtain a multi-core multi-feature fusion support vector machine with higher fault identification precision.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a method for optimizing a multi-core multi-feature fusion support vector machine for bearing fault identification, which integrates the performances of Gaussian radial basis kernel functions and polynomial kernel functions, forms a multi-core multi-feature fusion Least Square Support Vector Machine (LSSVM), and optimizes the parameters thereof by a Self-regulation particle swarm optimization (SRPSO) with strong convergence. Compared with the information fusion of the multi-position sensor, the fusion of various entropy values does not need a complicated sensor measurement system, and more comprehensive characteristic information is obtained.
The technical scheme adopted by the invention is as follows:
a method for optimizing a multi-core multi-feature fusion support vector machine for bearing fault identification comprises the following steps:
s1, selecting a bearing vibration signal collected under a single sensor;
s2, carrying out EMD decomposition on the bearing vibration signals at different rotating speeds to obtain an IMF energy entropy and an IMF arrangement entropy;
s3, extracting IMF energy entropies at different rotating speeds and fusing the IMF energy entropies with IMF arrangement entropies to obtain fusion features containing information at different rotating speeds, wherein the fusion features are used for training samples of the support vector machine to obtain a multi-core multi-feature fusion support vector machine suitable for fault recognition at different rotating speeds;
s4, integrating the performance of a Gaussian radial basis function kernel and a polynomial function kernel, and performing linear regression on training samples mapped to a high-dimensional space through a nonlinear function space to enable the training samples to be classified according to different characteristics to form a multi-core least square support vector machine, so that the support vector machine can identify fault characteristics under variable loads;
and S5, performing parameter optimization on the training sample by adopting a self-adjusting particle swarm algorithm with strong convergence, and then comparing the training sample with the test sample to perform bearing fault identification.
In the method for optimizing the multi-core multi-feature fusion support vector machine for bearing fault identification, in step S4, the construction process and the parameter optimization process of the multi-core least squares support vector machine are as follows:
selecting proper Gaussian radial basis function kernel parameters g, polynomial kernel parameters c and combined kernel coefficients p,
radial basis of GaussFunction kernel: kg=exp(-g·||x-xi||2);
The polynomial kernel is defined as Kd=(xTxi+1)c;
A combined kernel function: k ═ p · (x)Txi+1)c+(1-p)·exp(-g·||x-xi||2);
Optimizing parameters p, c and g through the SRPSO, and using the identification precision of the support vector machine as a fitness index, wherein the higher the fitness index is, the higher the fault identification precision is; setting parameters p, c and g as SRPSO particles, and optimizing the SRPSO particles by the following steps:
1) initializing the position and speed of a particle swarm;
2) calculating the fitness value of each particle by taking the precision of the support vector machine as the fitness value of the particle (the historical optimal position Pbest of the initial particle, and comparing the historical optimal position fitness to obtain the global optimal position Gbest of the initial history);
3) calculating self-adjusting inertia weight w of each particle; for the best particles: w ═ w + η Δ w; the other particles are: w ═ w- Δ w, wherewIIs an initial value of inertial weight, wFInertial weight end value, NIterThe number of iterations, η controls the acceleration rate constant;
4) updating the particle velocity v and position;
speed updateWherein, c1,c2Is a coefficient of acceleration r1;r2Is (0,1) random number; p is the social cognition of the particle:a is a random number, and λ is a set threshold, which is 0.5. Updating the historical best position and the particle swarm global best position of each particle; and judging whether the ending condition is reached or not, and returning to the step 2) if the ending condition is not reached.
The invention has the beneficial effects that:
1. the method for optimizing the multi-core multi-feature fusion support vector machine for bearing fault identification optimizes the parameters of the support vector machine by self-adjusting the particle swarm, and fuses two kinds of entropy values at different rotating speeds, so that the multi-core multi-feature fusion support vector machine with higher fault identification precision can be obtained, and the higher fault identification precision is achieved.
2. The method for optimizing the multi-core multi-feature fusion support vector machine for bearing fault identification integrates the performance of a Gaussian radial basis function kernel and a polynomial function kernel to form the self-adjusting particle swarm optimization multi-core multi-feature fusion support vector machine, is used for identifying bearing faults with high precision, and has higher identification precision under different rotating speeds and different fault features. The invention has reasonable design, high precision, simplicity and reliability; the support vector machine can effectively identify various faults of the bearing at different rotating speeds.
Drawings
FIG. 1 is a diagram: a self-adjusting particle swarm optimization multi-core multi-feature fusion support vector machine is used for a bearing fault identification process block diagram;
FIG. 2 is a diagram of: supporting a vector machine classification diagram;
FIG. 3 is a diagram of: a flow block diagram for supporting multi-feature fusion of a vector machine for pattern recognition;
FIG. 4 is a diagram of: an IMF energy entropy and IMF permutation entropy fusion mode flow block diagram under different faults;
FIG. 5 is a diagram: self-adjusting particle swarm cognitive decision diagrams;
FIG. 6 is a diagram of: vibration signals of bearing inner ring faults under different rotating speeds;
FIG. 7 is a diagram of: vibration signals of bearing outer ring faults at different rotating speeds;
FIG. 8 is a diagram of: vibration signals of bearing ball faults at different rotating speeds;
FIG. 9 is a diagram of: optimizing a training parameter optimization fitness value curve by SRPSO of the multi-core LSSVM;
FIG. 10 is a diagram: and optimizing a fitness value curve by the PSO optimization training parameter of the multi-core LSSVM.
Detailed Description
The technical solution of the present invention is further described in detail by the following embodiments.
Example 1
Referring to fig. 1, fig. 2 and fig. 3, the method for optimizing the multi-core multi-feature fusion support vector machine for bearing fault identification of the invention comprises the following steps: s1, selecting a bearing vibration signal collected under a single sensor;
s2, decomposing the bearing vibration signals at different rotating speeds by an EMD (Empirical Mode Decomposition) algorithm to obtain IMF (Intrinsic Mode Function) energy entropy and IMF array entropy;
s3, extracting IMF energy entropies at different rotating speeds and fusing the IMF energy entropies with IMF arrangement entropies to obtain fusion features containing information at different rotating speeds, wherein the fusion features are used for training samples of the support vector machine to obtain a multi-core multi-feature fusion support vector machine suitable for fault recognition at different rotating speeds;
s4, integrating the performance of a Gaussian radial basis function kernel and a polynomial function kernel, and performing linear regression on training samples mapped to a high-dimensional space through a nonlinear function space to enable the training samples to be classified according to different characteristics to form a Multi-core least square support vector machine (MK-LSSVM), so that the MK-LSSVM can identify fault characteristics under variable loads;
and S5, performing parameter optimization on the training sample by adopting a self-adjusting particle swarm algorithm with strong convergence, and then comparing the training sample with the test sample to perform bearing fault identification.
A polynomial kernel function kernel in the kernel function of the support vector machine is a global kernel and has stronger generalization capability but weaker learning capability; the Gaussian radial basis function kernel is a local kernel, and has good learning ability but weak generalization ability. The self-adjusting particle swarm optimization multi-core multi-feature fusion support vector machine is formed by integrating the performance of the Gaussian radial basis function kernel and the performance of the polynomial function kernel, is used for identifying bearing faults with high precision, and has higher identification precision under different rotating speeds and different fault features.
Example 2
The method for identifying the bearing fault by optimizing the multi-core multi-feature fusion support vector machine in the embodiment is different from the method in the embodiment 1 in that: in step S4, the multi-core least squares support vector machine construction process and the parameter optimization process are as follows:
selecting proper Gaussian radial basis function kernel parameters g, polynomial kernel parameters c and combined kernel coefficients p, wherein the Gaussian radial basis function kernel: kg=exp(-g·||x-xi||2) (ii) a The polynomial kernel is defined as Kd=(xTxi+1)c(ii) a A combined kernel function: k ═ p · (x)Txi+1)c+(1-p)·exp(-g·||x-xi||2) (ii) a Through the SRPSO optimization parameters p, c and g, the identification accuracy of the support vector machine is used as a fitness index, and the higher the fitness index is, the higher the fault identification accuracy is; setting parameters p, c and g as SRPSO particles, and optimizing the SRPSO particles by the following steps:
1) initializing the position and speed of a particle swarm;
2) calculating the fitness value of each particle by taking the precision of the support vector machine as the fitness value of the particle (the historical optimal position Pbest of the initial particle, and comparing the historical optimal position fitness to obtain the global optimal position Gbest of the initial history);
3) calculating self-adjusting inertia weight w of each particle; for the best particles: w ═ w + η Δ w; the other particles are: w ═ w- Δ w, wherewIIs an initial value of inertial weight, wFInertial weight end value, NIterThe iteration times, eta, control the acceleration rate constant;
4) updating the particle velocity v and position;
speed updateWherein, c1,c2Is a coefficient of acceleration r1;r2Is (0,1) random number; p is the social cognition of the particle:a is a random number, and λ is a set threshold, which is 0.5. Updating the historical best position and the particle swarm global best position of each particle; and judging whether the ending condition is reached or not, and returning to the step 2) if the ending condition is not reached.
Example 3
The process of optimizing the multi-core multi-feature fusion support vector machine for bearing fault identification is shown in fig. 1. The nature of the self-adjusting particle swarm optimization multi-core multi-feature fusion support vector machine for the support vector machine is that a sample is subjected to linear regression which is mapped to a high-dimensional space through a nonlinear function space, so that the sample is classified according to different characteristics, as shown in fig. 2. The support vector machine is used for the pattern recognition process as shown in fig. 3, and it can be seen that the important indexes of the difference of the recognition accuracy of the support vector machine are feature selection and parameter selection.
For feature selection, S1, selecting a bearing vibration signal collected under a single sensor; s2, carrying out EMD decomposition on the bearing vibration signals at different rotating speeds to obtain an IMF energy entropy and an IMF arrangement entropy; s3, fusing the IMF energy entropy and the IMF permutation entropy under different faults; and S4, combining the feature fusion entropy values at different rotating speeds to obtain fusion features containing different rotating speed information. Because the speed change is often generated due to load fluctuation in real production, the method theoretically has stronger feature integrity, so that the support vector machine can identify the fault features under variable load. The fusion mode is shown in figure 4.
In the aspect of parameter optimization, the SRPSO is selected as a parameter optimization algorithm, and the decision process of the optimization is shown in fig. 5. The support vector machine needs to select proper Gaussian radial basis function kernel parameters g, polynomial kernel parameters c and combined kernel coefficients p. Wherein, the Gaussian radial basis function kernel: kg=exp(-g·||x-xi||2) (ii) a The polynomial kernel is defined as Kd=(xTxi+1)c(ii) a A combined kernel function: k ═ p · (x)Txi+1)c+(1-p)·exp(-g·||x-xi||2). Through SRPSO optimization parameters p, c and g, the identification accuracy of the support vector machine is used as a fitness index, and the higher the fitness index is, the higher the fault identification accuracy is representedThe higher. Setting parameters p, c and g as SRPSO particles, and optimizing the SRPSO particles by the following steps:
s1, initializing the position and speed of a particle swarm;
s2, calculating the fitness value of each particle (initial particle historical optimal position Pbest, and comparing the historical optimal position fitness to obtain an initial historical global optimal position Gbest);
s3, calculating self-regulation inertia weight w of each particle; for the best particles: w ═ w + η Δ w; the other particles are: w- Δ w (whereinwIIs an initial value of inertial weight, wFInertial weight end value, NIterIteration number, eta control acceleration rate constant)
S4, updating the particle speed v and the particle position;
speed updateWherein, c1,c2Is an acceleration coefficient r1;r2Is (0,1) random number; p is the social cognition of the particle:a is a random number and λ is a set threshold, typically 0.5.
Updating the historical best position and the particle swarm global best position of each particle; it is determined whether or not the end condition is reached, and if the end condition is not reached, the flow returns to S2.
Taking bearing experimental data of an electrical engineering laboratory of Kaiser university of storage as an example, the bearing key information 6205-2RS of a driving end is obtained; the number of the balls is Inner Ring + Outer Ring is 9 (one); the bearing Pitch Diameter (Pitch Diameter) is equal to 1.537 (inch) and is equal to 39.04 mm; ball Diameter (Ball Diameter) 0.3126(inch) is about 7.94 mm; the outer ring fault is 104.56 Hz; inner ring failure is 157.94 Hz; the rolling element failure is 137.48 Hz.
The vibration signal characteristics of different faults at different rotating speeds are shown in figures 6-8.
Partial entropy values of the signals obtained by EMD decomposition are shown in tables 1-8.
IMF energy entropy:
TABLE 1
TABLE 2
TABLE 3
TABLE 4
IMF arrangement entropy:
TABLE 5
TABLE 6
TABLE 7
TABLE 8
The support vector machine parameter optimization fitness value curves are shown in fig. 9 and fig. 10, and it can be seen that the SRPSO convergence is better than the PSO; the optimal solution can be obtained without exceeding 5 generations of updating. Therefore, the selection is updated to generation 6.
Table 9: bearing fault signal identification accuracy
As can be seen from table 9, the optimization types 1, 4, and 5 can improve the recognition accuracy of the support vector machine by fusing different features at different rotation speeds; as can be seen from table 9, the optimization types 2 and 3 indicate that the fault identification accuracy of the SRPSO optimized support vector machine is higher than that of the PSO optimized support vector machine. As can be seen from table 9, optimization types 1 and 2, self-adjusting the parameters of the ion swarm optimization support vector machine can improve the recognition accuracy of the support vector machine. In conclusion, the comparison shows that the classification accuracy of the support vector machine is obviously improved after the self-regulation particle swarm optimization multi-core least square support vector machine.
Claims (1)
1. A method for optimizing a multi-core multi-feature fusion support vector machine for bearing fault identification comprises the following steps:
s1, selecting a bearing vibration signal collected under a single sensor;
s2, carrying out EMD decomposition on the bearing vibration signals at different rotating speeds to obtain an IMF energy entropy and an IMF arrangement entropy;
s3, extracting IMF energy entropies at different rotating speeds and fusing the IMF energy entropies with IMF arrangement entropies to obtain fusion features containing information at different rotating speeds, wherein the fusion features are used for training samples of the support vector machine to obtain a multi-core multi-feature fusion support vector machine suitable for fault recognition at different rotating speeds;
s4, integrating the performance of a Gaussian radial basis function kernel and a polynomial function kernel, and performing linear regression on training samples mapped to a high-dimensional space through a nonlinear function space to enable the training samples to be classified according to different characteristics to form a multi-core least square support vector machine, so that the support vector machine can identify fault characteristics under variable loads; the construction and parameter optimization process of the multi-core least square support vector machine is as follows:
selecting proper Gaussian radial basis function kernel parameters g, polynomial kernel parameters c and combined kernel coefficients p,
gaussian radial basis function kernel: kg=exp(-g·||x-xi||2);
The polynomial kernel is defined as Kd=(xTxi+1)c;
A combined kernel function: k ═ p · (x)Txi+1)c+(1-p)·exp(-g·||x-xi||2);
Optimizing parameters p, c and g through the SRPSO, and using the identification precision of the support vector machine as a fitness index, wherein the higher the fitness index is, the higher the fault identification precision is; setting parameters p, c and g as SRPSO particles, and optimizing the SRPSO particles by the following steps:
1) initializing the position and speed of a particle swarm;
2) calculating the fitness value of each particle and the historical optimal position Pbest of the initial particle by taking the precision of the support vector machine as the fitness value of the particle, and comparing the historical optimal position fitness to obtain the initial historical global optimal position Gbest;
3) calculating self-adjusting inertia weight w of each particle; for the best particles: w ═ w + η Δ w; the other particles are: w ═ w- Δ w, wherewIIs an initial value of inertial weight, wFInertial weight end value, NIterThe iteration times, eta, control the acceleration rate constant;
4) updating the particle velocity v and position;
speed updateWherein, c1,c2Is the coefficient of acceleration, r1、r2Is (0,1) random number; p is the social cognition of the particle:a is a random number, lambda is a set threshold value, and 0.5 is selected;
updating the historical best position and the particle swarm global best position of each particle; judging whether the ending condition is reached, if not, returning to the step 2);
and S5, performing parameter optimization on the training sample by adopting a self-adjusting particle swarm algorithm with strong convergence, and then comparing the training sample with the test sample to perform bearing fault identification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711318123.XA CN108062564B (en) | 2017-12-12 | 2017-12-12 | Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711318123.XA CN108062564B (en) | 2017-12-12 | 2017-12-12 | Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108062564A CN108062564A (en) | 2018-05-22 |
CN108062564B true CN108062564B (en) | 2021-07-09 |
Family
ID=62138351
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711318123.XA Active CN108062564B (en) | 2017-12-12 | 2017-12-12 | Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108062564B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110161035B (en) | 2019-04-26 | 2020-04-10 | 浙江大学 | Structural surface crack detection method based on image feature and Bayesian data fusion |
CN112085084B (en) * | 2020-08-24 | 2023-12-15 | 宁波大学 | Transformer fault diagnosis method based on multi-feature fusion common vector |
CN113076516A (en) * | 2021-04-16 | 2021-07-06 | 上海欣影电力科技股份有限公司 | PSO-EMMD-ICA-based transformer vibration signal online separation method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104156580A (en) * | 2014-08-03 | 2014-11-19 | 内蒙古科技大学 | Optimal noise parameter selection based ELMD (ensemble local mean decomposition) algorithm for fault diagnosis of bearings |
CN105548862A (en) * | 2016-01-25 | 2016-05-04 | 合肥工业大学 | Simulation circuit fault diagnosis method on the basis of generalized multi-nuclear support vector machine |
CN105628425A (en) * | 2016-01-01 | 2016-06-01 | 三峡大学 | Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine |
CN105784340A (en) * | 2016-03-03 | 2016-07-20 | 电子科技大学 | Air valve fault diagnosis method based on hybrid intelligent technology |
CN106017879A (en) * | 2016-05-18 | 2016-10-12 | 河北工业大学 | Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals |
CN106952183A (en) * | 2017-04-24 | 2017-07-14 | 武汉大学 | A kind of short-term load forecasting method based on particle group optimizing least square method supporting vector machine |
-
2017
- 2017-12-12 CN CN201711318123.XA patent/CN108062564B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104156580A (en) * | 2014-08-03 | 2014-11-19 | 内蒙古科技大学 | Optimal noise parameter selection based ELMD (ensemble local mean decomposition) algorithm for fault diagnosis of bearings |
CN105628425A (en) * | 2016-01-01 | 2016-06-01 | 三峡大学 | Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine |
CN105548862A (en) * | 2016-01-25 | 2016-05-04 | 合肥工业大学 | Simulation circuit fault diagnosis method on the basis of generalized multi-nuclear support vector machine |
CN105784340A (en) * | 2016-03-03 | 2016-07-20 | 电子科技大学 | Air valve fault diagnosis method based on hybrid intelligent technology |
CN106017879A (en) * | 2016-05-18 | 2016-10-12 | 河北工业大学 | Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals |
CN106952183A (en) * | 2017-04-24 | 2017-07-14 | 武汉大学 | A kind of short-term load forecasting method based on particle group optimizing least square method supporting vector machine |
Non-Patent Citations (4)
Title |
---|
Self regulating particle swarm optimization algorithm;M.R. Tanweer等;《Information Sciences》;20141031;第182-202页 * |
基于 EMD 能量熵和支持向量机的齿轮故障诊断方法;张超等;《振动与冲击》;20101231;第29卷(第10期);第216-220页 * |
基于EMD熵特征融合的滚动轴承故障诊断方法;向丹等;《航空动力学报》;20150531;第30卷(第5期);第1149-1155页 * |
多尺度变异粒子群优化MK-LSSVM的轴承寿命预测;张焱等;《仪器仪表学报》;20161130;第37卷(第11期);第2489-2496页 * |
Also Published As
Publication number | Publication date |
---|---|
CN108062564A (en) | 2018-05-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sobie et al. | Simulation-driven machine learning: Bearing fault classification | |
CN108062564B (en) | Method for optimizing multi-core multi-feature fusion support vector machine for bearing fault identification | |
CN112798280B (en) | Rolling bearing fault diagnosis method and system | |
Harmouche et al. | Statistical approach for nondestructive incipient crack detection and characterization using Kullback-Leibler divergence | |
CN113935406A (en) | Mechanical equipment unsupervised fault diagnosis method based on counter current model | |
Pan et al. | A method for mechanical fault recognition with unseen classes via unsupervised convolutional adversarial auto-encoder | |
Nguyen et al. | Improving intrusion detection in SCADA systems using stacking ensemble of tree-based models | |
Garcia et al. | Machine learning-based scheme for multi-class fault detection in turbine engine disks | |
CN110956331A (en) | Method, system and device for predicting operation state of digital factory | |
Jiang et al. | A SVDD and K-means based early warning method for dual-rotor equipment under time-varying operating conditions | |
CN114356897A (en) | Electromechanical actuator fault diagnosis method based on data fusion | |
Jianhui et al. | New fault diagnosis method for rolling bearing based on PCA | |
CN116861232A (en) | Air quality data anomaly detection model based on DBN-OCSVM | |
Abedin et al. | Vibration signal for bearing fault detection using random forest | |
Borhana et al. | Fault detection of bearing using support vector machine-SVM | |
Hariharan et al. | New approach of classification of rolling element bearing fault using artificial neural network | |
Jiang et al. | A novel ensemble fuzzy model for degradation prognostics of rolling element bearings | |
Lee et al. | Anomaly Detection based on 1D-CNN-LSTM Auto-Encoder for Bearing Data | |
CN114021275A (en) | Rolling bearing fault diagnosis method based on deep convolution fuzzy system | |
Boukra et al. | Identifying new prognostic features for remaining useful life prediction | |
CN113516162A (en) | OCSVM and K-means algorithm based industrial control system flow abnormity detection method and system | |
Pan et al. | Support vector data description with model selection for condition monitoring | |
Li et al. | Fuzzy feature fusion and multimodal degradation prognosis for mechanical components | |
Wang et al. | Three-dimensional acoustic emission source localisation in concrete based on sparse least-squares support vector regression | |
Zhao et al. | Health indicator selection and health assessment of rolling element bearing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |