CN112800682B - Feedback optimization fan blade fault monitoring method - Google Patents

Feedback optimization fan blade fault monitoring method Download PDF

Info

Publication number
CN112800682B
CN112800682B CN202110154925.1A CN202110154925A CN112800682B CN 112800682 B CN112800682 B CN 112800682B CN 202110154925 A CN202110154925 A CN 202110154925A CN 112800682 B CN112800682 B CN 112800682B
Authority
CN
China
Prior art keywords
fan blade
particle
optimization
monitoring method
support vector
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
Application number
CN202110154925.1A
Other languages
Chinese (zh)
Other versions
CN112800682A (en
Inventor
于傲
张亚平
王方政
汤鹏
邹祖冰
朱小毅
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.)
China Three Gorges Corp
Original Assignee
China Three Gorges Corp
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 China Three Gorges Corp filed Critical China Three Gorges Corp
Priority to CN202110154925.1A priority Critical patent/CN112800682B/en
Publication of CN112800682A publication Critical patent/CN112800682A/en
Priority to PCT/CN2022/070901 priority patent/WO2022166527A1/en
Application granted granted Critical
Publication of CN112800682B publication Critical patent/CN112800682B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • 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]
    • 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
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/72—Wind turbines with rotation axis in wind direction

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Feedback Control In General (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A fan blade fault monitoring method for feedback optimization comprises the following steps: the method comprises the following steps: constructing a data set for monitoring the state of the fan blade; step two: constructing a support vector machine model to divide the characteristic space of the fan blade data; step three: optimizing and solving the parameters of the support vector machine model; step four: the support vector machine model was evaluated. The invention aims to solve the technical problems that the conventional fan blade fault monitoring method has single consideration factor, the single factor does not necessarily have absolute correlation with the fault of the fan blade, so that the fault of the fan blade cannot be effectively and accurately monitored, the problem of the fan blade fault is complex, the conventional algorithm has more iteration times and large calculated amount, the algorithm is easy to fall into a local optimal solution, and the requirement of fan blade fault monitoring cannot be well met.

Description

Feedback optimization fan blade fault monitoring method
Technical Field
The invention belongs to the technical field of wind power generation, and particularly relates to a method for monitoring faults of a fan blade, which is integrated with technologies such as a support vector machine and a particle swarm algorithm.
Background
Under the background of accelerating adjustment and optimization of industrial structures and energy structures at present and developing new energy resources vigorously, the wind power loading capacity of China is increasing in a saving mode. The wind power generation does not generate greenhouse gas in the operation process, does not damage the ecological environment, and is a new energy industry vigorously developed in China. At present, timely and reliable finding of the fault of the fan blade is an important work of operation and maintenance personnel of a power plant.
A Support Vector Machine (SVM) can solve the nonlinear segmentation problem in a high-dimensional feature space, a particle swarm algorithm is a bionic optimization algorithm, a mathematical model is established through the foraging process of simulated birds, and the algorithm is high in global search capacity.
The main research on the monitoring of the fan blade fault is as follows:
liu Xiaobo and the like, the method needs to monitor a vibration signal of the blade through a sensor, and detect the crack fault of the fan blade by performing wavelet packet decomposition on the vibration signal and reconstructing and analyzing characteristic changes of each signal frequency domain.
According to the method, principal component analysis is adopted to perform dimensionality reduction processing on optical fiber load signals collected at the root of an original wind driven generator blade, then unsupervised feature extraction is performed by adopting a multilayer deep convolution self-encoder to obtain feature vectors with fault information, and then the feature vectors are input into the XGboost to perform feature learning and classification, so that intelligent detection of the faults of the wind driven generator blade is finally achieved.
Li Shaohui and the like, by analyzing the relationship between the aerodynamic signal and the failure of the fan blade, it is considered that the crack of the blade causes the change of a partial frequency band of the aerodynamic signal. Car M et al detect fan blades by a LiDAR-loaded drone. Joshuva A et al heuristically detect fan blade faults by building a decision tree for the fan blade vibration signals.
The above research only considers a single factor, the single factor and the fault of the fan blade do not necessarily have absolute correlation, meanwhile, the problem of the fault of the fan blade is complex, the iteration times of the existing algorithm are many, the calculated amount is large, and the algorithm is easy to fall into a local optimal solution.
Disclosure of Invention
The invention aims to solve the technical problems that the conventional fan blade fault monitoring method has single consideration factor, the single factor does not necessarily have absolute correlation with the fault of the fan blade, so that the fault of the fan blade cannot be effectively and accurately monitored, the problem of the fan blade fault is complex, the conventional algorithm has more iteration times and large calculated amount, the algorithm is easy to fall into a local optimal solution, and the requirement of fan blade fault monitoring cannot be well met.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
a method for monitoring faults of fan blades of a support vector machine based on improved learning factor particle swarm optimization comprises the following steps:
the method comprises the following steps: constructing a data set for monitoring the state of the fan blade;
step two: constructing a support vector machine model to divide the characteristic space of the fan blade data;
step three: optimizing and solving the parameters of the support vector machine model;
step four: the support vector machine model was evaluated.
In step one, the data set consists of D samples (X) i ,Y i ) Can be formed, wherein X i ∈R n ，Y i E { -1, +1}, wherein operational data of the fan blade is collected as a feature vector X i And meanwhile, acquiring a fan running state label Y (normal is +1, and fault is-1)), and standardizing each characteristic.
The standardization process comprises the following steps:
Figure BDA0002934320700000021
wherein μ is X i Sample mean of features, σ being X t Sample standard deviation of features.
The operation data comprises one or more of wind speed, generator rotating speed, ambient temperature, wind direction angle, yaw position, horizontal direction acceleration, vertical direction acceleration, angles, speeds and switch temperatures of the blades 1,2 and 3, direct current of the switch direct current device and temperature in the cabin.
In the second step, a nonlinear segmented support vector machine model is constructed by adopting the radial basis functions to segment the characteristic space of the fan blade data.
The classification decision function of the support vector machine model constructed above is:
Figure BDA0002934320700000022
where K (x, y) is chosen as the radial basis function:
Figure BDA0002934320700000023
α i is [0,C]C is a penalty coefficient; b is offset, sgn is sign function
Figure BDA0002934320700000031
The specific optimization objective function is as follows:
Figure BDA0002934320700000032
the constraint conditions are as follows: alpha is more than or equal to 0 i ≤C，i＝1,2,....D
Figure BDA0002934320700000033
In the third step, the alpha of the support vector machine model is optimized and solved by utilizing the particle swarm optimization of the improved learning factor i σ parameter (i =1,2.
The speed and position of the particle in the upper particle swarm optimization are vectors of the optimization function definition domain space, namely v and x are alpha i Iterative updating of vector, velocity and position of the σ (i =1,2,.. D) parameter space is as follows:
Figure BDA0002934320700000034
x (t+1) ＝x (t) +v (t)
wherein C is 1 And C 2 As a learning factor, v is the velocity of the particle, x is the position of the particle, p best Finding the optimal solution, g, for the particle individual in the objective optimization function best And (3) finding an optimal solution of the particle population in the target optimization function, wherein t represents the t-th iteration, and rand (0,1) represents an arbitrary value in a (0,1) interval.
The method specifically comprises the following optimization steps:
1) Setting an iteration step length m of a particle swarm algorithm, and setting a particle number N;
2) Setting the initial velocity of the particles
Figure BDA0002934320700000035
And an initial position
Figure BDA0002934320700000036
Wherein
Figure BDA0002934320700000037
Representing the velocity value of the j-th iteration of the ith particle,
Figure BDA0002934320700000038
σ = [0.01,1000 ] of radial basis function representing position of jth iteration of ith particle]The penalty factor C is equal to [0.1,500 ∈ ]]Learning factor C 1 And C 2 L of all particle velocity and position vectors for each iteration 2 A certain multiple of the Cauchy probability of the sum of squares of the norms;
Figure BDA0002934320700000041
Figure BDA0002934320700000042
3) Calculating an adaptive value of the objective function according to the initialization condition;
4) According to the adaptive value, the speed and the position of the particles are updated circularly, and the optimal solution p of the individual is obtained best Optimal solution g of the population best Wherein j is not less than 0 and not more than mFor the t-th (0. Ltoreq. T. Ltoreq.N) iteration of the particle
Figure BDA0002934320700000043
Figure BDA0002934320700000044
Wherein the t-th iteration (0. Ltoreq. T. Ltoreq.N)
Figure BDA0002934320700000045
min()
Represents the minimum value.
5) And according to the termination condition, finishing the algorithm optimizing process: l to maximum number of iterations or two generations of speed and position 2 And if the norm is smaller than the set value: | v | (V) (t+1) -v (t) Less than or equal to kappa and x (t+1) -x (t) And | is less than or equal to gamma, wherein kappa and gamma are precision set values.
In the fourth step, the optimal SVM model parameter value is obtained, the SVM model is verified, if the target accuracy is not met, the SVM model is retrained, and at the moment, the initial speed and the position of the particles are set to be the speed and the position corresponding to the optimal SVM model value and multiplied by a certain weight.
And the optimization result of the particle swarm optimization is divided in a feature space through a support vector machine, and after evaluation, if the result is not ideal, the evaluation result is fed back to the particle swarm optimization again to continue optimization and enlarge a search range, so that the situation that the optimization solution falls into a local optimal solution is avoided.
Compared with the prior art, the invention has the following technical effects:
1. the Cauchy probability density is used as a learning factor, so that the Cauchy probability density is not easy to fall into a local optimal solution, furthermore, a negative feedback means is adopted to avoid the situation that the optimization process falls into the local optimal solution, the optimization result of the particle swarm optimization is segmented in a characteristic space through a support vector machine, after the evaluation, if the result is not ideal, the evaluation result is fed back to the particle swarm optimization again, the optimization is continued, and the search range is enlarged, so that the situation that the local optimal solution falls into is avoided;
2. after cross validation of SVM models, if the accuracy rate does not reach a target value, the SVM models need to be retrained again, and the initial speed and the position of particles are set as the speed and the position of the particles corresponding to the training of the current optimal SVM model value and multiplied by a certain multiple of weight, so that the particles jump out of a local optimal solution when continuously searching for a global optimal solution, and local oscillation is prevented;
3. the invention uses the characteristics of various fan blades as the input of the model, thereby greatly ensuring the accuracy of the model;
4. the invention adopts the support vector machine model optimized by the particle swarm optimization for improving the learning factors to detect the faults of the fan blade, solves the problems that the nonlinear SVM model is difficult to optimize and is easy to fall into a local optimal area, and simultaneously improves the reliability of fault diagnosis of electrical equipment.
Drawings
The invention is further illustrated with reference to the following figures and examples:
FIG. 1 is a flow chart of the present invention.
Detailed Description
As shown in fig. 1, a method for monitoring a fault of a fan blade of a feedback optimization support vector machine includes the following steps:
the method comprises the following steps: constructing a data set for monitoring the state of the fan blade;
step two: constructing a support vector machine model to divide the characteristic space of the fan blade data;
step three: optimizing and solving the parameters of the support vector machine model;
step four: the support vector machine model is evaluated.
In step one, the data set consists of D samples (X) i ,Y i ) Can be formed, wherein X i ∈R n ，Y i E { -1, +1}, wherein operational data of the fan blade is collected as a feature vector X i And meanwhile, acquiring a fan running state label Y (normal is +1, and fault is-1)), and standardizing each characteristic.
The standardization process is as follows:
Figure BDA0002934320700000051
wherein μ is X i Sample mean of features, σ being X t Sample standard deviations of features.
The operation data comprises one or more of wind speed, generator rotating speed, ambient temperature, wind direction angle, yaw position, horizontal direction acceleration, vertical direction acceleration, angles, speeds and switch temperatures of the blades 1,2 and 3, direct current of the switch direct current device and temperature in the cabin.
In the second step, a nonlinear segmented support vector machine model is constructed by adopting the radial basis functions to segment the characteristic space of the fan blade data.
The classification decision function of the constructed support vector machine model is as follows:
Figure BDA0002934320700000061
where K (x, y) is chosen as the radial basis function:
Figure BDA0002934320700000062
α i is [0,C]C is a penalty coefficient, b is an offset, sgn is a sign function
Figure BDA0002934320700000063
The specific optimization objective function is as follows:
Figure BDA0002934320700000064
the constraint conditions are as follows: alpha is more than or equal to 0 i ≤C，i＝1,2,....D
Figure BDA0002934320700000065
In the step ofThirdly, the alpha of the support vector machine model is optimized and solved by utilizing the particle swarm optimization of the improved learning factor i σ parameter (i =1,2.
Wherein the speed and position of the particle in the particle swarm optimization is the vector of the optimization function definition domain space, namely v and x are alpha i σ (i =1,2,.. D) a vector of the parameter space;
the velocity and position are iteratively updated as follows:
Figure BDA0002934320700000066
x (t+1) ＝x (t) +v (t) ；
wherein C 1 And C 2 As a learning factor, v is the velocity of the particle, x is the position of the particle, p best Finding the optimal solution, g, for the particle individual in the objective optimization function best And (3) finding an optimal solution of the particle population in the target optimization function, wherein t represents the t-th iteration, and rand (0,1) represents an arbitrary value in a (0,1) interval.
When the optimization solution is carried out, firstly, the number of particles is set, the speed and the position of each particle are vectors of an optimization function definition domain space, the adaptive value of each particle is calculated, the speed v and the position x of each particle are updated, and the optimal solution p of the speed, the position and the individual of each particle is obtained best And group optimal solution g best In this connection, for the purpose of smoothing the solution process, cauchy probability density is used as a learning factor.
The method uses the Cauchy probability density as a learning factor, so that the Cauchy probability density is not easy to fall into the local optimal solution, furthermore, the method adopts a negative feedback means to avoid the optimization process from falling into the local optimal solution, the optimization result of the particle swarm optimization algorithm is segmented in a feature space through a support vector machine, and after the evaluation, if the result is not ideal, the evaluation result is fed back to the particle swarm algorithm again to continue the optimization, and the search range is enlarged, so that the local optimal solution is avoided from falling into;
as implementable, the following optimization steps are provided:
1) Setting an iteration step length m =1500 of a particle swarm algorithm, and setting a particle number N =20;
2) Setting the initial velocity of the particles
Figure BDA0002934320700000071
And an initial position
Figure BDA0002934320700000072
Wherein
Figure BDA0002934320700000073
Representing the velocity value of the j-th iteration of the ith particle,
Figure BDA0002934320700000074
represents the position of the jth iteration of the ith particle, σ e [0.01,1000 of the radial basis function]The penalty factor C is equal to [0.1,500 ∈ ]]Learning factor C 1 And C 2 L of all particle velocity and position vectors for each iteration 2 1.5 times and 2.5 times the cauchy probability of the sum of the squares of the norms,
Figure BDA0002934320700000075
Figure BDA0002934320700000076
learning factor C 1 And C 2 L of all particle velocity and position vectors for each iteration 2 A certain multiple of the Cauchy probability of the sum of squares of the norms;
the Cauchy probability is used as a learning factor, and the particle swarm optimization is not easy to enter a local optimal solution when the Cauchy probability density function is flat.
3) Calculating an adaptive value of the objective function according to the initialization condition;
4) According to the adaptive value, the speed and the position of the particles are updated circularly, and the optimal solution p of the individual is obtained best Optimal solution g of the population best Wherein j is not less than 0 and not more than jm) th iteration of particles (t is more than or equal to 0 and less than or equal to N)
Figure BDA0002934320700000077
Figure BDA0002934320700000078
Wherein the t-th iteration (0. Ltoreq. T. Ltoreq.N)
Figure BDA0002934320700000079
5) And according to the termination condition, finishing the algorithm optimizing process: l to maximum number of iterations or two generations of speed and position 2 If the norm is smaller than a set value: | v | (V) (t+1) -v (t) Less than or equal to kappa and x (t+1) -x (t) And | ≦ γ, and k =0.4 and γ =0.8 may be selected.
In the fourth step, the optimal SVM model parameter value is obtained, 10-fold cross validation is carried out on the SVM model, if the target accuracy is not met, the SVM model is retrained, and at the moment, the initial speed and the position of the particles are set as the speed and the position corresponding to the optimal SVM model value and multiplied by the weight of 1.5 times.

Claims (8)

1. A fan blade fault monitoring method for feedback optimization is characterized by comprising the following steps:
the method comprises the following steps: constructing a data set for monitoring the state of the fan blade;
in step one, the data set consists of D samples (X) i ,Y i ) Can be formed, wherein X i ∈R n ，Y i E { -1, +1}, wherein operational data of the fan blade is collected as a feature vector X i Meanwhile, collecting a fan running state label Y (normal is +1, fault is-1)), and standardizing each characteristic;
step two: constructing a support vector machine model to divide the characteristic space of the fan blade data;
step three: optimizing and solving the parameters of the support vector machine model;
the method specifically comprises the following optimization steps:
1) Setting an iteration step length m of a particle swarm algorithm, and setting a particle number N;
2) Setting the initial velocity of the particles
Figure FDA0003763652240000011
And an initial position
Figure FDA0003763652240000012
Wherein
Figure FDA0003763652240000013
A velocity parameter representing the jth iteration of the ith particle,
Figure FDA0003763652240000014
represents the jth iteration position parameter of the ith particle, and the sigma epsilon of the radial basis function [0.01,1000 ]]The penalty factor C is equal to [0.1,500 ∈ ]]Learning factor C 1 And C 2 L of all particle velocity and position vectors for each iteration 2 A certain multiple of Cauchy probability of a sum of squares of the norms;
Figure FDA0003763652240000015
Figure FDA0003763652240000016
3) Calculating an adaptive value of the objective function according to the initialization condition;
4) According to the adaptive value, the speed and the position of the particles are updated circularly, and the optimal solution p of the individual is obtained best Optimal solution g of the population best Wherein j (0. Ltoreq. J. Ltoreq.m) th particle is iterated t times (0. Ltoreq. T. Ltoreq.N)
Figure FDA0003763652240000017
Figure FDA0003763652240000018
Wherein the t-th iteration (0. Ltoreq. T. Ltoreq.N)
Figure FDA0003763652240000019
Figure FDA00037636522400000110
min () represents the minimum value;
5) And according to the termination condition, finishing the algorithm optimizing process: l to maximum number of iterations or two generations of speed and position 2 If the norm is smaller than a set value: | | v (t+1) -v (t) Less than or equal to kappa and x (t+1) -x (t) | | is less than or equal to gamma, wherein kappa and gamma are precision set values;
step four: the support vector machine model was evaluated.
2. The feedback-optimized fan blade fault monitoring method of claim 1, wherein the normalization process is:
Figure FDA0003763652240000021
wherein μ is X i Sample mean of features, σ being X t Sample standard deviations of features.
3. The feedback-optimized fan blade fault monitoring method of claim 1, wherein the operational data includes one or more of wind speed, generator speed, ambient temperature, wind direction angle, yaw position, horizontal acceleration, vertical acceleration, angle, speed, switch temperature of blade 1, blade 2, blade 3, dc current of a switch dc, and in-cabin temperature.
4. The feedback-optimized fan blade fault monitoring method of claim 1, comprising: in the second step, a nonlinear segmented support vector machine model is constructed by adopting the radial basis functions to segment the characteristic space of the fan blade data.
5. The feedback-optimized fan blade fault monitoring method of claim 4, wherein the classification decision function of the constructed support vector machine model is:
Figure FDA0003763652240000022
where K (x, y) is chosen as the radial basis function:
Figure FDA0003763652240000023
α i is [0,C]C is a penalty coefficient, and b is an offset;
sgn is a sign function
Figure FDA0003763652240000024
The specific optimization objective function is as follows:
Figure FDA0003763652240000025
the constraint conditions are as follows: alpha is more than or equal to 0 i ≤C，i＝1,2,....D
Figure FDA0003763652240000031
6. The feedback-optimized fan blade fault monitoring method according to claim 5, wherein in step three, the alpha of the support vector machine model is optimized and solved by using the particle swarm optimization of the improved learning factor i σ parameter (i =1,2.
7. The feedback optimized fan blade fault monitoring method of claim 6, whereinThe velocity and position of the particle in the particle swarm optimization is the vector of the space of the optimization function definition domain, i.e. v and x are alpha i Iterative updating of vector, velocity and position of the σ (i =1,2,.. D) parameter space is as follows:
Figure FDA0003763652240000032
x (t+1) ＝x (t) +v (t)
wherein C 1 And C 2 As a learning factor, v is the velocity of the particle, x is the position of the particle, p best Finding the optimal solution, g, for the particle individuals in the objective optimization function best And (3) finding an optimal solution of the particle population in the target optimization function, wherein t represents the t-th iteration, and rand (0,1) represents an arbitrary value in a (0,1) interval.
8. The feedback optimization fan blade fault monitoring method according to any one of claims 1 to 7, wherein in the fourth step, an optimal SVM model parameter value is obtained, the SVM model is verified, if the target accuracy is not met, the SVM model is retrained, and at this time, the initial velocity and position of the particle are set as the velocity and position corresponding to the optimal SVM model value, and are multiplied by a certain weight.
CN202110154925.1A 2021-02-04 2021-02-04 Feedback optimization fan blade fault monitoring method Active CN112800682B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202110154925.1A CN112800682B (en) 2021-02-04 2021-02-04 Feedback optimization fan blade fault monitoring method
PCT/CN2022/070901 WO2022166527A1 (en) 2021-02-04 2022-01-10 Feedback optimization-based wind turbine blade fault monitoring method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110154925.1A CN112800682B (en) 2021-02-04 2021-02-04 Feedback optimization fan blade fault monitoring method

Publications (2)

Publication Number Publication Date
CN112800682A CN112800682A (en) 2021-05-14
CN112800682B true CN112800682B (en) 2022-10-04

Family

ID=75814232

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110154925.1A Active CN112800682B (en) 2021-02-04 2021-02-04 Feedback optimization fan blade fault monitoring method

Country Status (2)

Country Link
CN (1) CN112800682B (en)
WO (1) WO2022166527A1 (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112800682B (en) * 2021-02-04 2022-10-04 中国长江三峡集团有限公司 Feedback optimization fan blade fault monitoring method
CN115791887B (en) * 2023-02-08 2023-04-18 北京汉飞航空科技有限公司 Self-adaptive measurement algorithm of turbine blade based on six-point measuring tool
CN116400585B (en) * 2023-06-06 2023-08-18 济南大学 Marine rescue flying wing control optimization method based on improved rime optimization algorithm
CN116776600B (en) * 2023-06-21 2024-04-12 安徽工业大学 Wind turbine blade optimal design method and system based on self-adaptive proxy model
CN117077064B (en) * 2023-10-13 2024-02-06 云南滇能智慧能源有限公司 Fault detection method, device and equipment for wind power equipment and storage medium
CN117290732B (en) * 2023-11-24 2024-03-01 山东理工昊明新能源有限公司 Construction method of fault classification model, wind power equipment fault classification method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104820142A (en) * 2015-03-20 2015-08-05 国家电网公司 Electric power system fault diagnosis method based on improved particle swarm optimization algorithm
CN105426954A (en) * 2015-08-20 2016-03-23 武汉科技大学 Particle swarm optimization method based on multi-strategy synergistic function
WO2018072351A1 (en) * 2016-10-20 2018-04-26 北京工业大学 Method for optimizing support vector machine on basis of particle swarm optimization algorithm

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102944418B (en) * 2012-12-11 2015-06-03 东南大学 Wind turbine generator group blade fault diagnosis method
CN105550744A (en) * 2015-12-06 2016-05-04 北京工业大学 Nerve network clustering method based on iteration
CN110969096A (en) * 2019-11-07 2020-04-07 国电南京自动化股份有限公司 Motor fault mode diagnosis method based on particle swarm optimization support vector machine
CN111351668B (en) * 2020-01-14 2022-03-25 江苏科技大学 Diesel engine fault diagnosis method based on optimized particle swarm algorithm and neural network
CN112800682B (en) * 2021-02-04 2022-10-04 中国长江三峡集团有限公司 Feedback optimization fan blade fault monitoring method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104820142A (en) * 2015-03-20 2015-08-05 国家电网公司 Electric power system fault diagnosis method based on improved particle swarm optimization algorithm
CN105426954A (en) * 2015-08-20 2016-03-23 武汉科技大学 Particle swarm optimization method based on multi-strategy synergistic function
WO2018072351A1 (en) * 2016-10-20 2018-04-26 北京工业大学 Method for optimizing support vector machine on basis of particle swarm optimization algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于改进PSO-GSA算法的快速开关型电网故障限流器优化配置;王博等;《高压电器》;20200616(第06期);全文 *

Also Published As

Publication number Publication date
CN112800682A (en) 2021-05-14
WO2022166527A1 (en) 2022-08-11

Similar Documents

Publication Publication Date Title
CN112800682B (en) Feedback optimization fan blade fault monitoring method
Han et al. An adaptive spatiotemporal feature learning approach for fault diagnosis in complex systems
CN109582003B (en) Bearing fault diagnosis method based on pseudo label semi-supervised kernel local Fisher discriminant analysis
Wang et al. A two-stage data-driven approach for image-based wind turbine blade crack inspections
US20230094389A1 (en) Quantum computing based deep learning for detection, diagnosis and other applications
CN106951997B (en) Method and device for predicting fault of fan
CN109597401A (en) A kind of equipment fault diagnosis method based on data-driven
CN110766175A (en) Pitch system fault detection method and device based on optimal interval distribution machine
CN111815806B (en) Method for preprocessing flight parameter data based on wild value elimination and feature extraction
CN111680875A (en) Unmanned aerial vehicle state risk fuzzy comprehensive evaluation method based on probability baseline model
CN115112372A (en) Bearing fault diagnosis method and device, electronic equipment and storage medium
CN115314235A (en) System and method for network attack detection in wind turbine control systems
CN109325553B (en) Wind power gear box fault detection method, system, equipment and medium
Chen et al. Fault detection for turbine engine disk based on adaptive weighted one-class support vector machine
Qin et al. High-efficiency generative adversarial network model for chemical process fault diagnosis
CN112818819B (en) AUV state monitoring method based on dynamic model and complex network theory
Jiang et al. A SVDD and K-means based early warning method for dual-rotor equipment under time-varying operating conditions
Farahani et al. A novel method for designing transferable soft sensors and its application
CN117009841A (en) Model training method, motor fault diagnosis method and microcontroller
CN115577290A (en) Distribution network fault classification and source positioning method based on deep learning
Ghate et al. Induction machine fault detection using support vector machine based classifier
Chen et al. Learning to adapt to domain shifts with few-shot samples in anomalous sound detection
Liu et al. Adaptive robustness evaluation for complex system prognostics and health management software platform
Hübner et al. Wind Turbine Rotor aerodynamic imbalance detection using CNN
Saied et al. Supervised Learning Classification Applications in Fault Detection and Diagnosis: An Overview of Implementations in Unmanned Aerial Systems

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