CN114036973A - Series arc fault identification method of extreme learning machine based on dynamic online sequence - Google Patents

Series arc fault identification method of extreme learning machine based on dynamic online sequence Download PDF

Info

Publication number
CN114036973A
CN114036973A CN202111187942.1A CN202111187942A CN114036973A CN 114036973 A CN114036973 A CN 114036973A CN 202111187942 A CN202111187942 A CN 202111187942A CN 114036973 A CN114036973 A CN 114036973A
Authority
CN
China
Prior art keywords
fault
arc
data
elm
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
Application number
CN202111187942.1A
Other languages
Chinese (zh)
Inventor
薛鹏
潘国兵
欧阳静
赵继凯
钱浚杰
邓伟芳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang University of Technology ZJUT
Original Assignee
Zhejiang University of Technology ZJUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang University of Technology ZJUT filed Critical Zhejiang University of Technology ZJUT
Priority to CN202111187942.1A priority Critical patent/CN114036973A/en
Publication of CN114036973A publication Critical patent/CN114036973A/en
Pending legal-status Critical Current

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
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
    • G06F17/13—Differential equations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/141—Discrete Fourier transforms
    • G06F17/142—Fast Fourier transforms, e.g. using a Cooley-Tukey type algorithm
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G06F17/148—Wavelet transforms
    • 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
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Biomedical Technology (AREA)
  • Algebra (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Discrete Mathematics (AREA)
  • Evolutionary Biology (AREA)
  • Operations Research (AREA)
  • Locating Faults (AREA)

Abstract

A series arc fault identification method of an extreme learning machine based on a dynamic online sequence comprises the following steps: 1) denoising the sampled data of the electric energy waveform; 2) intercepting the real-time data in a segmented manner; 3) calculating and processing waveform data; 4) ELM fault arc identification: based on an ELM algorithm, arc identification is converted into a fault classification problem, weights from an input layer to a hidden layer are randomly given, and after the weights from the input layer to the hidden layer exist, the weights from the hidden layer to an output layer are obtained according to a least square method, so that fault arc identification is realized. The method provides an accurate and effective way for identifying the series arc faults of the power grid under different load conditions through the dynamic online ELM learning algorithm with high calculation efficiency and high universality.

Description

Series arc fault identification method of extreme learning machine based on dynamic online sequence
Technical Field
The invention relates to the field of arc fault identification, in particular to a series arc fault identification method of an Extreme Learning Machine (ELM) based on a dynamic online sequence.
Background
The arc fault identification technology analyzes one or more parameters by collecting various parameters in a line and applying different methods, so as to judge whether an arc fault is generated in the line. But arc faults are characterized by randomness, concealment, complexity and the like, so that the arc faults are difficult to identify.
At present, there are three types of methods for detecting series arc faults by researchers at home and abroad. The first is a detection method based on a mathematical model. The second is a detection method based on the physical characteristics of the arc. The third is a detection method based on arc current and voltage waveforms. Due to the complexity and randomness of the arc faults, the existing arc model has difficulty in accurately and comprehensively expressing the dynamic characteristics of the arc faults when the arc faults occur. When the characteristics of arc sound, arc light, temperature and the like during arc combustion are used for identifying the arc, the characteristics are limited by the installation quantity and the positions of the sensors, and the detection effect is greatly reduced for the hidden arc. At present, arc detection by using waveform characteristics is mainstream.
In a known conventional arc detection method, extracted arc current waveform feature data is input to a classifier such as a support vector machine to be classified, and a classification result is obtained. However, the load conditions of different lines are very different and can change frequently, resulting in poor training effect and failure to meet the requirements. Therefore, an Extreme Learning Machine (ELM) algorithm of dynamic online sequence is adopted to detect the arc fault in the line.
The ELM can be regarded as an improved feedforward neural network, and compared with a traditional machine learning network, the ELM of the dynamic online sequence has the following advantages in arc fault detection that (1) the training speed of the extreme learning machine is nearly thousands of times faster than that of the traditional neural network (such as BP algorithm). (2) Extreme learning machines have been verified to have better universality. (3) The ELM is combined with a dynamic online sequence, online learning can be realized, and the method has strong adaptability to power grid systems with different load intensities and real-time changes.
Disclosure of Invention
In order to overcome the defects that the existing arc fault identification process is low in identification precision, poor in adaptability when the load condition changes and more in false detection and missing detection situations, the invention provides the arc fault identification algorithm of the extreme learning machine of the dynamic online sequence, and the arc fault identification algorithm can adapt to different power grid load environments through an online learning method.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a method for series arc fault identification of an extreme learning machine based on a dynamic online sequence, the method comprising the steps of:
step 1) noise reduction of current waveform sampling data
Acquiring waveform sampling data of current in a power grid by using a high-speed sampling mutual inductance sensor, taking three adjacent sampling points as a group, adopting a pre-mean filtering algorithm, replacing the numerical value of a first point with the mean value of the group, adding a new point and discarding an old point;
step 2) real-time data segmentation interception
Intercepting the current real-time signal obtained in the step 1) by taking n current cycles as a period, namely intercepting data by adopting a window with an nT period as a length;
step 3) waveform data calculation and processing
After waveform data are obtained in the step 2), extracting time threshold characteristic root mean square error, integral energy value and period average value difference, extracting inter-harmonic factors and wavelet high-frequency components of frequency domain characteristic parts, and performing normalization processing on the characteristics;
step 4) ELM fault arc identification
Based on an ELM algorithm, arc identification is converted into a fault classification problem, weights from an input layer to a hidden layer are randomly given, and after the weights from the input layer to the hidden layer exist, the weights from the hidden layer to an output layer are obtained according to a least square method, so that fault arc identification is realized.
Further, the process of the step 4) is as follows:
4.1 selection of arc Fault database N at initialization stage0Group data
Figure RE-GDA0003455841870000021
Wherein
Figure RE-GDA0003455841870000022
Randomly generating an input weight matrix wiBias matrix biAn activation function f (), where i ═ 1, 2.,
Figure RE-GDA0003455841870000023
learning method calculation according to ELM
Figure RE-GDA0003455841870000024
Calculating an output weight matrix beta0＝M0H0T0Wherein M is0＝(H0H0)-1，
Figure RE-GDA0003455841870000031
Obtaining Y ═ Hkβ0Calculating the error between the normal output and the sample labeli＝|Yi-Tag |, with maximum error distance ∈ being countedmaxMinimum error distance of epsilonminSetting a margin parameter
Figure RE-GDA0003455841870000035
Calculating to obtain a judgment threshold value epsilonb＝k*(εmax-εmin)；
4.2, after the initialization stage shown in 4.1 is completed, accessing the system into a fault arc detection line, and inputting the normalized numerical value real-time data obtained in the step 3);
firstly, the output value Y is calculated as Hkβ0. At this time epsiloni＝|Y-Tag|；
When epsiloni≤εbJudging the sample to be in a normal running state, and marking a normal label on the sample to enter online learning;
when epsiloni＞εbWhen the power grid is in error, warning is given, the power grid condition is manually checked, if the power grid condition is in error, a normal label is marked, and the power grid condition is used as a new sample to enter online learning;
recording a new sample set
Figure RE-GDA0003455841870000032
At this point, a new sample set hidden layer output matrix is calculated:
Figure RE-GDA0003455841870000033
calculating a new output weight matrix:
Figure RE-GDA0003455841870000034
when k is k +1, completing one-time online sequence training;
4.3 repeat step 4.2 after new data is obtained, and perform a new training round.
Compared with the prior series arc fault detection technology, the invention has the following advantages:
the invention adopts an Extreme Learning Machine (ELM) algorithm with a dynamic sequence to identify the fault, and has the advantages of high learning and processing speed, good normalization performance and high identification accuracy compared with the traditional BP neural network algorithm.
The invention adopts a dynamic time sequence method on the basis of an Extreme Learning Machine (ELM), has stronger adaptability aiming at different loads of a power grid in the operation process and strong anti-interference capability, and is suitable for occasions needing long-time data monitoring.
The features of the root mean square error, the integral energy value, the period mean value difference, the partial inter-harmonic factor and the wavelet high-frequency component extracted by the method can be obtained by integrating, Fast Fourier Transform (FFT) and wavelet transform through the terminal, the calculation has low requirement on calculation power, and the calculation power can be integrated in a terminal collector to save cloud calculation power and transmission bandwidth.
The method tests the effect of series arc fault detection, can adaptively adjust the identification sensitivity under the conditions of different loads (1A and 5A), and has stronger anti-interference capability and real-time property under different conditions. The method can accurately identify the series arc fault and send out warning, and can utilize the real-time data online optimization model in normal operation to improve the accuracy.
Drawings
FIG. 1 is a flow chart of a series arc fault detection method in an embodiment of the invention.
Detailed Description
The invention will be described in detail with reference to the accompanying drawings: the present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments.
Referring to fig. 1, a method for series arc fault identification of an extreme learning machine based on a dynamic online sequence, the method comprising the steps of:
step 1) noise reduction of current waveform sampling data
The method comprises the steps that a mutual inductance sensor for high-speed sampling is used for obtaining waveform sampling data of current in a power grid, burrs exist in the data due to interference and the like, adjacent three sampling points are used as a group, a pre-mean filtering algorithm is adopted, the mean value of the group replaces the numerical value of a first point, a new point is added, and the old point is discarded, and the like;
step 2) real-time data segmentation interception
In order to facilitate the rapid processing of data and improve the real-time detection capability, 10(n is 10) current cycles are adopted as a period to intercept the current real-time signal obtained in the step 1), namely a window with the length of 10T period is adopted to intercept data;
step 3) waveform data calculation and processing
After waveform data are obtained in the step 2), extracting time threshold characteristic root mean square error, integral energy value and period average value difference, extracting inter-harmonic factors and wavelet high-frequency components of frequency domain characteristic parts, and performing normalization processing on the characteristics;
step 4) ELM fault arc identification
Based on an ELM algorithm, arc identification is converted into a fault classification problem, weights from an input layer to a hidden layer are randomly given, and after the weights from the input layer to the hidden layer exist, the weights from the hidden layer to an output layer can be obtained according to a least square method, so that fault arc identification is realized.
The process of the step 4) is as follows:
4.1 selection of arc Fault database N at initialization stage0Group data
Figure RE-GDA0003455841870000051
Wherein
Figure RE-GDA0003455841870000052
Randomly generating an input weight matrix wiBias matrix biAn activation function f (), where i ═ 1, 2.,
Figure RE-GDA0003455841870000053
learning method calculation according to ELM
Figure RE-GDA0003455841870000054
Calculating an output weight matrix beta0＝M0H0T0Wherein M is0＝(H0H0)-1，
Figure RE-GDA0003455841870000055
Obtaining Y ═ Hkβ0Calculating between normal output and sample labelError epsiloni＝|Yi-Tag |, with maximum error distance ∈ being countedmaxMinimum error distance of epsilonminSetting a margin parameter
Figure RE-GDA0003455841870000057
Calculating to obtain a judgment threshold value epsilonb＝k*(εmax-εmin) The threshold parameter is not suitable to be set too small;
4.2, after the initialization stage shown in 4.1 is completed, accessing the system into a fault arc detection line, and inputting the normalized numerical value real-time data obtained in the step 3);
firstly, the output value Y is calculated as Hkβ0At this time εi＝|Y-Tag|；
When epsiloni≤εbJudging the sample to be in a normal running state, and marking a normal label on the sample to enter online learning;
when epsiloni＞εbWhen the power grid is in error, warning is given, the power grid condition is manually checked, if the power grid condition is in error, a normal label is marked, and the power grid condition is used as a new sample to enter online learning;
recording a new sample set
Figure RE-GDA0003455841870000056
At this point, a new sample set hidden layer output matrix is calculated:
Figure RE-GDA0003455841870000061
calculating a new output weight matrix:
Figure RE-GDA0003455841870000062
when k is k +1, completing one-time online sequence training;
4.3 repeat step 4.2 after new data is obtained, and perform a new training round.
Finally, it should also be noted that the above-mentioned list is only one specific embodiment of the invention. It is obvious that the invention is not limited to the above embodiments, but that many variations are possible. All modifications which can be derived or suggested by a person skilled in the art from the disclosure of the present invention are to be considered within the scope of the invention.

Claims (2)

1. A series arc fault identification method of an extreme learning machine based on a dynamic online sequence is characterized by comprising the following steps:
step 1) noise reduction of current waveform sampling data
Acquiring waveform sampling data of current in a power grid by using a high-speed sampling mutual inductance sensor, taking three adjacent sampling points as a group, adopting a pre-mean filtering algorithm, replacing the numerical value of a first point with the mean value of the group, adding a new point and discarding an old point;
step 2) real-time data segmentation interception
Intercepting the current real-time signal obtained in the step 1) by taking n current cycles as a period, namely intercepting data by adopting a window with an nT period as a length;
step 3) waveform data calculation and processing
After waveform data are obtained in the step 2), extracting time threshold characteristic root mean square error, integral energy value and period average value difference, extracting inter-harmonic factors and wavelet high-frequency components of frequency domain characteristic parts, and performing normalization processing on the characteristics;
step 4) ELM fault arc identification
Based on an ELM algorithm, arc identification is converted into a fault classification problem, weights from an input layer to a hidden layer are randomly given, and after the weights from the input layer to the hidden layer exist, the weights from the hidden layer to an output layer are obtained according to a least square method, so that fault arc identification is realized.
2. The method for identifying series arc faults of the extreme learning machine based on the dynamic online sequence as claimed in claim 1, wherein the process of the step 4) is as follows:
4.1 initialization phase selectionArc fault database N0Group data
Figure RE-FDA0003455841860000011
Wherein
Figure RE-FDA0003455841860000012
Randomly generating an input weight matrix wiBias matrix biActivating a function f (), wherein
Figure RE-FDA0003455841860000013
Learning method calculation according to ELM
Figure RE-FDA0003455841860000014
Calculating an output weight matrix beta0＝M0H0T0Wherein M is0＝(H0H0)-1，
Figure RE-FDA0003455841860000015
Obtaining Y ═ Hkβ0Calculating the error between the normal output and the sample labeli＝|Yi-Tag |, with maximum error distance ∈ being countedmaxMinimum error distance of epsilonminSetting a margin parameter
Figure RE-FDA0003455841860000021
Calculating to obtain a judgment threshold value epsilonb＝k*(εmax-εmin)；
4.2, after the initialization stage shown in 4.1 is completed, accessing the system into a fault arc detection line, and inputting the normalized numerical value real-time data obtained in the step 3);
firstly, the output value Y is calculated as Hkβ0At this time εi＝|Y-Tag|；
When epsiloni≤εbWhen it is positive, it is judged asIn a normal running state, the sample is marked with a normal label to enter online learning;
when epsiloni＞εbWhen the power grid is in error, warning is given, the power grid condition is manually checked, if the power grid condition is in error, a normal label is marked, and the power grid condition is used as a new sample to enter online learning;
recording a new sample set
Figure RE-FDA0003455841860000022
At this point, a new sample set hidden layer output matrix is calculated:
Figure RE-FDA0003455841860000023
calculating a new output weight matrix:
Figure RE-FDA0003455841860000024
when k is k +1, completing one-time online sequence training;
4.3 repeat step 4.2 after new data is obtained, and perform a new training round.
CN202111187942.1A 2021-10-12 2021-10-12 Series arc fault identification method of extreme learning machine based on dynamic online sequence Pending CN114036973A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111187942.1A CN114036973A (en) 2021-10-12 2021-10-12 Series arc fault identification method of extreme learning machine based on dynamic online sequence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111187942.1A CN114036973A (en) 2021-10-12 2021-10-12 Series arc fault identification method of extreme learning machine based on dynamic online sequence

Publications (1)

Publication Number Publication Date
CN114036973A true CN114036973A (en) 2022-02-11

Family

ID=80141301

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111187942.1A Pending CN114036973A (en) 2021-10-12 2021-10-12 Series arc fault identification method of extreme learning machine based on dynamic online sequence

Country Status (1)

Country Link
CN (1) CN114036973A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116256592A (en) * 2022-11-28 2023-06-13 国网山东省电力公司德州供电公司 Medium-voltage distribution cable latent fault detection method and system
CN116908632A (en) * 2023-07-19 2023-10-20 常熟理工学院 Low-voltage alternating-current series arc fault characteristic self-adaptive optimizing detection method and system
CN117540288A (en) * 2023-11-09 2024-02-09 广东石油化工学院 Reinforced extreme learning machine fault diagnosis method based on small sample unbalance learning

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858509A (en) * 2018-11-05 2019-06-07 杭州电子科技大学 Based on multilayer stochastic neural net single classifier method for detecting abnormality
WO2019218263A1 (en) * 2018-05-16 2019-11-21 深圳大学 Extreme learning machine-based extreme ts fuzzy inference method and system
CN113033632A (en) * 2021-03-10 2021-06-25 西北工业大学 High-altitude platform fault diagnosis method based on wavelet analysis and multi-layer overrun learning machine

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019218263A1 (en) * 2018-05-16 2019-11-21 深圳大学 Extreme learning machine-based extreme ts fuzzy inference method and system
CN109858509A (en) * 2018-11-05 2019-06-07 杭州电子科技大学 Based on multilayer stochastic neural net single classifier method for detecting abnormality
CN113033632A (en) * 2021-03-10 2021-06-25 西北工业大学 High-altitude platform fault diagnosis method based on wavelet analysis and multi-layer overrun learning machine

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HAIYANG YU: "Ensemble OS-ELM based on combination weight for data stream classification", APPLIED INTELLIGENCE, 14 January 2019 (2019-01-14), pages 1 - 3 *
任永功等: "基于机器学习的数据挖掘技术研究", 31 May 2021, 哈尔滨工业大学出版社, pages: 179 - 180 *
李扬: "低压故障电弧识别方法研究", 中国优秀硕士学位论文全文数据库, 8 April 2019 (2019-04-08), pages 12 - 54 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116256592A (en) * 2022-11-28 2023-06-13 国网山东省电力公司德州供电公司 Medium-voltage distribution cable latent fault detection method and system
CN116256592B (en) * 2022-11-28 2023-09-26 国网山东省电力公司德州供电公司 Medium-voltage distribution cable latent fault detection method and system
CN116908632A (en) * 2023-07-19 2023-10-20 常熟理工学院 Low-voltage alternating-current series arc fault characteristic self-adaptive optimizing detection method and system
CN117540288A (en) * 2023-11-09 2024-02-09 广东石油化工学院 Reinforced extreme learning machine fault diagnosis method based on small sample unbalance learning

Similar Documents

Publication Publication Date Title
CN114036973A (en) Series arc fault identification method of extreme learning machine based on dynamic online sequence
CN112257530B (en) Rolling bearing fault diagnosis method based on blind signal separation and support vector machine
CN110259648B (en) Fan blade fault diagnosis method based on optimized K-means clustering
CN112580267A (en) Aero-engine surge prediction method based on multi-branch feature fusion network
CN112378660A (en) Intelligent fault diagnosis method for aero-engine bearing based on data driving
CN117668751B (en) High-low voltage power system fault diagnosis method and device
CN109190598A (en) A kind of rotating machinery monitoring data noise detection method based on SES-LOF
CN111275570A (en) Wind turbine generator set power abnormal value detection method based on iterative statistics and hypothesis test
CN115220396B (en) Intelligent monitoring method and system for numerical control machine tool
CN117093947B (en) Power generation diesel engine operation abnormity monitoring method and system
CN113962264A (en) Aero-engine rotor system fault diagnosis algorithm based on deep learning
CN110701087A (en) Axial flow compressor pneumatic instability detection method based on single-classification overrun learning machine
CN117093938A (en) Fan bearing fault detection method and system based on deep learning
CN112033656A (en) Mechanical system fault detection method based on broadband spectrum processing
CN111046790A (en) Pump bearing fault diagnosis method
CN110632443A (en) Diagnosis method for IGBT open circuit fault of traction rectifier of CRH3 motor train unit
CN108594156B (en) Improved current transformer saturation characteristic identification method
CN118242331B (en) Mechanical state detection method for hydraulic motor
CN107505337B (en) Microwave reflection signal characteristic database for composite insulator defect detection
CN113436645A (en) Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment
CN114330430A (en) Elevator fault judgment method and system based on big data characteristic analysis
CN116717461B (en) Intelligent monitoring method and system for operating state of vacuum pump
Zhao et al. A fault diagnosis approach for rolling element bearing based on S-transform and artificial neural network
CN112525336B (en) Automatic detection method for continuous vibration growth of mechanical equipment
CN116227172A (en) Rolling bearing performance degradation evaluation method based on convolutional neural network

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