CN113221998A - Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM - Google Patents

Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM Download PDF

Info

Publication number
CN113221998A
CN113221998A CN202110491200.1A CN202110491200A CN113221998A CN 113221998 A CN113221998 A CN 113221998A CN 202110491200 A CN202110491200 A CN 202110491200A CN 113221998 A CN113221998 A CN 113221998A
Authority
CN
China
Prior art keywords
ssa
optimal value
model
value
kernel function
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
CN202110491200.1A
Other languages
Chinese (zh)
Inventor
罗奕
李安昊
王腾飞
程哲
何宇华
唐亮
徐浩天
宋明谦
殷豪
张筵凯
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.)
Guilin University of Electronic Technology
Original Assignee
Guilin University of Electronic Technology
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 Guilin University of Electronic Technology filed Critical Guilin University of Electronic Technology
Priority to CN202110491200.1A priority Critical patent/CN113221998A/en
Publication of CN113221998A publication Critical patent/CN113221998A/en
Pending legal-status Critical Current

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

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

A rare earth extraction stirring shaft fault diagnosis method and system based on an SSA-SVM belong to the field of rare earth extraction equipment fault diagnosis, and an updated SSA model is obtained by setting upper and lower limit thresholds of an initial SSA model through optimizing an acceleration signal; optimizing a kernel function and a penalty parameter in the initial SVM model by updating the SSA model to obtain a kernel function optimal value and a penalty parameter optimal value, and diagnosing the fault type of the rare earth extraction stirring shaft; the linear representation in the SVM model is subjected to linear optimization through SSA, the optimal linear representation is found out through the punishment parameters and the kernel function, and the accuracy of fault classification judgment is improved; in addition, the invention can construct an updated SSA model and an updated SVM model only by sampling the original vibration acceleration signals, and can achieve the effect of accurate fault type judgment under the condition of single and simple input quantity.

Description

Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM
Technical Field
The invention relates to the field of rare earth extraction equipment fault diagnosis, and particularly relates to a rare earth extraction stirring shaft fault diagnosis method and system based on an SSA-SVM.
Background
The development of the rare earth industry is greatly limited by the development of industrial equipment, and the development degree of industrial production equipment shows that the industrial development of China is not advanced on the whole, which is also the key for the development of the rare earth industry. For mineral resource development and production enterprises, the automation degree of the extraction production process in the powder material industry is still relatively backward, a large amount of work is completed by manual operation, and the production efficiency is low through off-line analysis, component detection and manual control. In the process of rare earth production technology, the key technology is whether the extraction agent and the extracted element are fully dissolved, and the result of insufficient dissolution is quality accidents and resource waste. Factors causing incomplete dissolution are many, such as slippage of the drive belt; the stirring motor is abnormally operated, and even the production is stopped. When the driving belt slips and the stirring motor carries out secondary extraction, the production efficiency is influenced, and the shutdown can be seriously caused. Therefore, effective fault monitoring in the production process is the guarantee of production quality and safe production.
The SVM is a supervised learning method and has the most ideal detection effect on the balance data. At present, the most common method for detecting rare earth faults is a neural network method, but the neural network method is easy to have the problems of dimension disaster, local extremum and the like, and the detection effect is not ideal. Compared with a neural network method, the SVM can better solve the problems of dimension disaster, small sample learning, nonlinearity, local extremum and the like, and particularly has good generalization performance on the aspect of small sample learning, so that the SVM has wide application in the field of fault detection. However, the selection of the kernel function and the penalty parameter of the SVM has great influence on the diagnosis result, and randomness and blindness of model parameter selection exist.
Disclosure of Invention
Aiming at the technical limitation that the fault diagnosis of the rare earth extraction stirring shaft cannot be achieved in the prior art, the invention provides a fault diagnosis method and a fault diagnosis system of the rare earth extraction stirring shaft based on an SSA-SVM.
The technical scheme provided by the invention is as follows:
a rare earth extraction stirring shaft fault diagnosis method based on SSA-SVM comprises the following steps:
acquiring an acceleration signal at the tail end of the stirring shaft, and performing singular value decomposition and normalization processing to obtain an optimized acceleration signal;
setting an upper limit threshold and a lower limit threshold of the initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter;
and constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining a fault type based on the optimization acceleration signal.
Preferably, the optimizing the kernel function and the penalty parameter in the initial SVM model by the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter includes:
constructing a real number vector based on a kernel function range and a penalty parameter range in the initial SVM model;
and substituting the real number vector into the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
Preferably, the substituting the real number vector into the updated SSA model to obtain a kernel function optimal value and a penalty parameter optimal value includes:
bringing the real number vector into a target function to obtain an adaptive value;
comparing the adaptive value with the upper and lower limit thresholds of the updated SSA model to obtain a kernel function optimal value and a penalty parameter optimal value;
the adaptation value is a single column matrix comprising a set number of elements.
Preferably, the objective function is represented by the following formula:
Figure BDA0003052089870000031
wherein: g is the penalty parameter range, k is the kernel function range, b is a constant, and a is a Lagrange multiplier.
Preferably, the comparing the adaptive value with the upper and lower threshold values of the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter includes:
judging whether the adaptive value is in the upper and lower limit threshold range of the updated SSA model, and if the adaptive value is in the upper and lower limit threshold range of the updated SSA model, obtaining the kernel function optimal value and the punishment parameter optimal value;
otherwise, resetting all elements of the adaptive value to preset initial values, setting the maximum iteration times of the updating SSA, and performing iterative optimization on the elements in the adaptive value at the same time to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
Preferably, the iteratively optimizing the elements in the adaptive value at the same time to obtain the optimal value of the kernel function and the optimal value of the penalty parameter includes:
performing iterative operation on all elements in the adaptive value through the updated SSA model at the same time, and updating and replacing the current element in the adaptive value with an optimal value when the element obtains the optimal value before the maximum iteration times;
and if the element does not obtain the optimal value before the maximum iteration times, keeping the current element unchanged from the preset initial value.
Preferably, the obtaining the fault type based on the optimized acceleration signal includes:
extracting the optimized acceleration with a set proportion as a training set, and taking the rest optimized acceleration signals as a data set;
training the training set through the updating SVM to obtain a fault type;
and comparing the data set with the fault type to obtain a fault type.
A rare earth extraction stirring shaft fault diagnosis system based on SSA-SVM, the diagnosis system comprises:
a signal acquisition module: acquiring an acceleration signal at the tail end of the stirring shaft, and performing singular value decomposition and normalization processing to obtain an optimized acceleration signal;
an optimal value operation module: setting an upper limit threshold and a lower limit threshold of the initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter;
a judging module: and constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining a fault type based on the optimization acceleration signal.
Preferably, the optimal value operation module includes:
and a real number vector construction submodule: constructing a real number vector based on a kernel function range and a penalty parameter range in the initial SVM model;
an optimal value operation submodule: and substituting the real number vector into the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
Compared with the prior art, the invention has the beneficial effects that: setting an upper limit threshold and a lower limit threshold of an initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter; the linear representation in the SVM model is subjected to linear optimization through SSA, the optimal linear representation is found out through the punishment parameters and the kernel function, and the accuracy of fault classification judgment is improved; in addition, the invention can construct an updated SSA model and an updated SVM model only by sampling the original vibration acceleration signals, and can achieve the effect of accurate fault type judgment under the condition of single and simple input quantity.
Drawings
FIG. 1 is a flow chart of a method for diagnosing faults of a rare earth extraction stirring shaft provided by the invention;
FIG. 2 is a flow chart of the acceleration signal processing at the end of the stirring shaft according to the present invention;
FIG. 3 is a flow chart of the sparrow search algorithm optimizing SVM.
Detailed Description
For the purpose of enhancing the understanding of the present invention, the present invention will be described in further detail with reference to the accompanying drawings and examples, which are provided for the purpose of illustration only and are not intended to limit the scope of the present invention.
The first embodiment is as follows:
the embodiment provides a rare earth extraction stirring shaft fault diagnosis method based on an SSA-SVM, and a flow chart of the method is shown in FIG. 1.
The method comprises the following steps: acquiring a stirring shaft terminal acceleration signal, decomposing through singular values and carrying out normalization processing to obtain an optimized acceleration signal, and specifically comprising the following steps:
singular Value Decomposition (SVD) noise reduction processing is carried out on the measured vibration signals, firstly, the measured original data is matrixed, an n x n matrix is constructed according to the method that signals of each line are circularly arranged, elements on a row vector are completely repeated with the previous row of elements, and the matrix is defined as B, then:
Figure BDA0003052089870000061
where x (i) represents the acquired signal.
The effective order size of the SVD-constructed signal matrix affects the noise cancellation effect, so it is necessary to determine the effective order of the signal matrix B.
The method comprises the steps of determining the effective order of a signal matrix by using a singular value ratio method based on MMRR, firstly selecting larger p (p is more than or equal to n), setting the larger p as the order of a signal matrix B, decreasing the p value and constructing a corresponding special matrix, and taking the eigenvalue ratio as a target function of effective order estimation. Defining the ratio parameter of the maximum eigenvalue to the minimum eigenvalue of the matrix B as follows:
Figure BDA0003052089870000062
in the formula, the numerator is the maximum eigenvalue of the matrix B under the order l, and the denominator is the minimum eigenvalue of the matrix B under the order l. The effective order of the signal matrix B is set to be L, and as the acquired signal matrix is a finite length sequence containing noise, a certain error exists between the reconstructed signal matrix B and the matrix B, the minimum characteristic value of the matrix B is not equal to 0 but approaches to 0 when p is larger than or equal to L. The degree of mutation of MMR values at p ═ L was reduced. To enhance the mutation effect of MMR, the MMR neighborhood data ratio MMRR is defined as:
Figure BDA0003052089870000063
the objective function is:
Figure BDA0003052089870000064
in the above equation, L is the effective order of the signal matrix B and the diagonal matrix Σ. According to the effective order Z, reserving the middle and front Z singular values in the diagonal matrix and setting other singular values to be 0 to obtain a new diagonal matrix; and (3) utilizing the new diagonal matrix to restore to obtain the Frobenious norm approximation of the original signal, and recovering the noise-canceling signal, namely the noise-canceling matrix. Thereby completing the SVD denoising process.
Normalizing the denoised signal data to obtain an optimized acceleration signal, wherein the normalization processing formula is as follows:
Figure BDA0003052089870000071
wherein x isiRepresenting the measured speed vibration signal, xmaxRepresents the maximum value of the data, xminRepresenting the minimum value of the data.
Step two: setting an upper limit threshold and a lower limit threshold of the initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model by the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter, which specifically comprises the following steps:
and constructing an SSA initial model, and initializing an upper limit threshold value, a lower limit threshold value and a maximum iteration number, wherein the upper limit and the lower limit are determined by the introduced optimized acceleration signal, the maximum iteration number is determined by the dimensionality of the optimized acceleration signal, and the higher the dimensionality is, the more the iteration number is, and the updated SSA model is obtained.
Combining a parameter penalty parameter g of a support vector machine in an initial SVM model with a kernel function k in a real vector form and substituting the parameter penalty parameter g into an updated SSA model, wherein the real vector is introduced into the model in a matrix form for the convenience of calculation, and the expression is as follows:
Figure BDA0003052089870000072
where d represents the dimension and n represents the order of the matrix.
Calculating an adaptive value after the substituting, wherein the adaptive value is calculated as:
Figure BDA0003052089870000081
then an adaptive matrix of n rows and 1 column is constructed, which is expressed as:
Figure BDA0003052089870000082
the adaptive value is brought into the updated SSA model for iterative solution, and the operation process in the updated SSA model is as shown in fig. 3. Taking the optimal value solved by each iteration as a finder and the rest as followers, wherein after each iteration, the updating position formula of the finder is as follows:
Figure BDA0003052089870000083
where t represents the current iteration number, j represents a constant representing the maximum number of iterations, and j is 1,2,3 … … itremax.
Figure BDA0003052089870000084
Representing the position information of the ith sparrow in the jth dimension. Alpha is (0, 1)]Is a random number. R2And ST represents the warning value and the safety value, respectively, Q is a random number following a normal distribution, and L represents a 1 × L matrix in which each element in the matrix is 1 in its entirety. When R is2If ST, the finder will move randomly around the current position in a normal distribution. Otherwise, the mobile terminal is farther away from the current position.
After the location of the finder is updated, the location of the follower is also updated accordingly, and the update formula is as follows:
Figure BDA0003052089870000091
Figure BDA0003052089870000092
is the optimum position occupied by the finder at present, XworstRepresenting the current global worst position, A represents a 1 x d matrix in which each element is randomly assigned a value of 1 or-1, when i>n/2, this indicates that the fitness is not satisfactory.
In order to prevent the occurrence of false optimal solution in the updating SSA model, namely the optimal adaptive value is met but the range of the upper limit and the lower limit is exceeded, an alarm value is introduced, and the updating is disclosed as follows:
Figure BDA0003052089870000093
wherein the content of the first and second substances,
Figure BDA0003052089870000094
the current global optimal position is, beta is taken as a step length control parameter, and is a random number which follows normal distribution with the mean value of 0 and the variance of 1. K is a random number between (-1, +1), fi is the fitness value of the current individual, fgAnd fwRespectively the current global best and worst fitness value. When f isi＞fgThis indicates that the individual at this time is at the upper and lower boundary edges. When f isi＝fgThis indicates that the individual is trying to narrow the distance to the upper and lower bounds, and K represents the step size control parameter.
Judging whether elements in the adaptive value are in the upper and lower limit threshold range of the updated SSA model, and if the adaptive value is in the upper and lower limit threshold range of the updated SSA model, obtaining a kernel function optimal value and a punishment parameter optimal value;
otherwise, resetting all elements of the adaptive value to preset initial values, setting the maximum iteration times for updating the SSA, and performing iteration optimization on the elements in the adaptive value at the same time to obtain a kernel function optimal value and a penalty parameter optimal value;
updating all elements in the adaptive value by an SSA model and carrying out iterative operation at the same time, and updating and replacing the current element in the adaptive value to be the optimal value when the element obtains the optimal value before the maximum iteration times;
if the element does not obtain the optimal value before the maximum iteration number, the current element is kept unchanged from the preset initial value, which is 1 in this embodiment.
And outputting the global optimal position and the global optimal fitness value of the population, wherein the penalty parameter g and the kernel function k are the optimal parameters of the SVM. And then the optimal parameters are brought into the SVM to classify and identify the data.
In step one and step two, the processing procedure for the acquired raw acceleration signal is shown in fig. 2.
Step three: constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining a fault type based on the optimization acceleration signal, wherein the method specifically comprises the following steps:
and inputting 80% of the preprocessed data serving as a test set and 20% serving as a training set into the updated SVM model, and constructing a rare earth extraction stirring shaft fault diagnosis model based on the SSA-SVM. The method comprises the steps of constructing a matrix of n x 1 by vibration acceleration signals, randomly extracting 80% of the vibration acceleration signals as a training set, inputting the vibration acceleration signals into an updated SVM model for training in a form of codes according to fault types, classifying the same labels into one class in a clustering mode, inputting the same labels as original data, outputting the same labels as codes, constructing a rare earth extraction stirring shaft fault diagnosis model based on SSA-SVM, taking the rest 20% of data sets as test sets to be brought into the fault diagnosis model, making the labels in the form of codes according to the fault types, verifying whether the results classified by the SVM are consistent with the labels or not, and obtaining the fault diagnosis accuracy.
Example two:
based on the same invention idea, the embodiment provides a rare earth extraction stirring shaft fault diagnosis system based on an SSA-SVM, and the system includes:
a signal acquisition module: the device is used for acquiring acceleration signals at the tail end of the stirring shaft, and obtaining optimized acceleration signals through singular value decomposition and normalization processing;
an optimal value operation module: the upper and lower limit thresholds of the initial SSA model are set based on the optimized acceleration signal, and an updated SSA model is obtained; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter;
a judging module: and the fault type obtaining module is used for constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining the fault type based on the optimization acceleration signal.
The optimal value operation module comprises:
and a real number vector construction submodule: the method comprises the steps of constructing a real number vector based on a kernel function range and a penalty parameter range in an initial SVM model;
an optimal value operation submodule: and the system is used for substituting the real number vector into the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
The optimal value operation submodule comprises:
an adaptation value construction unit: the real number vector is brought into a target function to obtain an adaptive value;
an adaptive value comparison unit: and comparing the adaptive value with the upper and lower limit thresholds of the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
The adaptive value construction unit constructs an adaptive value according to the following formula:
Figure BDA0003052089870000111
wherein: g is the penalty parameter range, k is the kernel function range, b is a constant, and a is a Lagrange multiplier.
The adaptive value comparison unit comprises:
a comparison subunit: judging whether the adaptive value is in the upper and lower limit threshold range of the updated SSA model, and if the adaptive value is in the upper and lower limit threshold range of the updated SSA model, obtaining the kernel function optimal value and the punishment parameter optimal value;
otherwise, resetting all elements of the adaptive value to preset initial values, and setting the maximum iteration times of the updating SSA; performing iterative operation on all elements in the adaptive value through the updated SSA model at the same time, and updating and replacing the current element in the adaptive value with an optimal value when the element obtains the optimal value before the maximum iteration times; if the element does not obtain the optimal value before the maximum iteration times, the operation is carried out through a resetting subunit;
the reset subunit: for keeping the current element unchanged from the preset initial value.
The judging module comprises:
a data set planning submodule: the system is used for extracting the optimized acceleration in a set proportion to serve as a training set, and taking the rest optimized acceleration signals as a data set;
training a submodule: the training set is used for training through the updating SVM to obtain a fault type;
a comparison submodule: and the fault type comparison module is used for comparing the data set with the fault type to obtain the fault type.
It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The present invention is not limited to the above embodiments, and any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are included in the scope of the claims of the present invention which are filed as the application.

Claims (9)

1. A rare earth extraction stirring shaft fault diagnosis method based on SSA-SVM is characterized by comprising the following steps:
acquiring an acceleration signal at the tail end of the stirring shaft, and performing singular value decomposition and normalization processing to obtain an optimized acceleration signal;
setting an upper limit threshold and a lower limit threshold of the initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter;
and constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining a fault type based on the optimization acceleration signal.
2. The method of claim 1, wherein the optimizing the kernel function and the penalty parameter in the initial SVM model by the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter comprises:
constructing a real number vector based on a kernel function range and a penalty parameter range in the initial SVM model;
and substituting the real number vector into the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
3. The method of claim 2, wherein the substituting the real vectors into the updated SSA model to obtain an optimal value for a kernel function and an optimal value for a penalty parameter comprises:
bringing the real number vector into a target function to obtain an adaptive value;
comparing the adaptive value with the upper and lower limit thresholds of the updated SSA model to obtain a kernel function optimal value and a penalty parameter optimal value;
the adaptation value is a single column matrix comprising a set number of elements.
4. The fault diagnosis method according to claim 3, characterized in that the objective function is represented by the following formula:
Figure FDA0003052089860000021
wherein: g is the penalty parameter range, k is the kernel function range, b is a constant, and a is a Lagrange multiplier.
5. The method of claim 3, wherein the comparing the fitness value to the upper and lower threshold values of the updated SSA model to obtain an optimum value of a kernel function and an optimum value of a penalty parameter comprises:
judging whether the adaptive value is in the upper and lower limit threshold range of the updated SSA model, and if the adaptive value is in the upper and lower limit threshold range of the updated SSA model, obtaining the kernel function optimal value and the punishment parameter optimal value;
otherwise, resetting all elements of the adaptive value to preset initial values, setting the maximum iteration times of the updating SSA, and performing iterative optimization on the elements in the adaptive value at the same time to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
6. The method of claim 5, wherein the iterative optimization of the elements in the fitness value to obtain the optimal value of the kernel function and the optimal value of the penalty parameter comprises:
performing iterative operation on all elements in the adaptive value through the updated SSA model at the same time, and updating and replacing the current element in the adaptive value with an optimal value when the element obtains the optimal value before the maximum iteration times;
and if the element does not obtain the optimal value before the maximum iteration times, keeping the current element unchanged from the preset initial value.
7. The fault diagnosis method according to claim 1, wherein said deriving a fault type based on said optimized acceleration signal comprises:
extracting the optimized acceleration with a set proportion as a training set, and taking the rest optimized acceleration signals as a data set;
training the training set through the updating SVM to obtain a fault type;
and comparing the data set with the fault type to obtain a fault type.
8. A rare earth extraction stirring shaft fault diagnosis system based on SSA-SVM is characterized by comprising the following components:
a signal acquisition module: acquiring an acceleration signal at the tail end of the stirring shaft, and performing singular value decomposition and normalization processing to obtain an optimized acceleration signal;
an optimal value operation module: setting an upper limit threshold and a lower limit threshold of the initial SSA model based on the optimized acceleration signal to obtain an updated SSA model; optimizing a kernel function and a penalty parameter in the initial SVM model through the updated SSA model to obtain an optimal value of the kernel function and an optimal value of the penalty parameter;
a judging module: and constructing an updated SVM model based on the optimization kernel function and the optimization penalty parameter, and obtaining a fault type based on the optimization acceleration signal.
9. The fault diagnosis system according to claim 8, wherein the optimal value operation module comprises:
and a real number vector construction submodule: constructing a real number vector based on a kernel function range and a penalty parameter range in the initial SVM model;
an optimal value operation submodule: and substituting the real number vector into the updated SSA model to obtain the optimal value of the kernel function and the optimal value of the penalty parameter.
CN202110491200.1A 2021-05-06 2021-05-06 Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM Pending CN113221998A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110491200.1A CN113221998A (en) 2021-05-06 2021-05-06 Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110491200.1A CN113221998A (en) 2021-05-06 2021-05-06 Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM

Publications (1)

Publication Number Publication Date
CN113221998A true CN113221998A (en) 2021-08-06

Family

ID=77090989

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110491200.1A Pending CN113221998A (en) 2021-05-06 2021-05-06 Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM

Country Status (1)

Country Link
CN (1) CN113221998A (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1850589A1 (en) * 2006-03-29 2007-10-31 Sony Deutschland Gmbh Method for video mode detection
CN108414226A (en) * 2017-12-25 2018-08-17 哈尔滨理工大学 Fault Diagnosis of Roller Bearings under the variable working condition of feature based transfer learning
CN109871758A (en) * 2019-01-14 2019-06-11 南京航空航天大学 Fault-signal SVD noise-reduction method based on Multiscale Morphological optimization
CN110153801A (en) * 2019-07-04 2019-08-23 西南交通大学 A kind of cutting-tool wear state discrimination method based on multi-feature fusion
CN110487499A (en) * 2019-09-03 2019-11-22 大连理工大学 A kind of bolt pretightening discrimination method based on singular value decomposition and support vector machines
CN112016624A (en) * 2020-08-28 2020-12-01 西北农林科技大学 Spectrum data dimension reduction method for optimizing kernel independent components based on sparrow search algorithm
US20210125066A1 (en) * 2019-10-28 2021-04-29 Lightmatter, Inc. Quantized architecture search for machine learning models
CN112733603A (en) * 2020-12-11 2021-04-30 江苏大学 Frequency conversion scroll compressor fault diagnosis method based on improved VMD and SVM

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1850589A1 (en) * 2006-03-29 2007-10-31 Sony Deutschland Gmbh Method for video mode detection
CN108414226A (en) * 2017-12-25 2018-08-17 哈尔滨理工大学 Fault Diagnosis of Roller Bearings under the variable working condition of feature based transfer learning
CN109871758A (en) * 2019-01-14 2019-06-11 南京航空航天大学 Fault-signal SVD noise-reduction method based on Multiscale Morphological optimization
CN110153801A (en) * 2019-07-04 2019-08-23 西南交通大学 A kind of cutting-tool wear state discrimination method based on multi-feature fusion
CN110487499A (en) * 2019-09-03 2019-11-22 大连理工大学 A kind of bolt pretightening discrimination method based on singular value decomposition and support vector machines
US20210125066A1 (en) * 2019-10-28 2021-04-29 Lightmatter, Inc. Quantized architecture search for machine learning models
CN112016624A (en) * 2020-08-28 2020-12-01 西北农林科技大学 Spectrum data dimension reduction method for optimizing kernel independent components based on sparrow search algorithm
CN112733603A (en) * 2020-12-11 2021-04-30 江苏大学 Frequency conversion scroll compressor fault diagnosis method based on improved VMD and SVM

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
JIANKAI XUE等: "A novel swarm intelligence optimization approach: sparrow search algorithm", 《SYSTEMS SCIENCE & CONTROL ENGINEERING AN OPEN ACCESS JOURNAL》 *
罗奕等: "动态变权层次分析法在本科教学效果评估中的应用_罗奕", 《大众科技》 *
马晨佩等: "基于麻雀搜索算法优化支持向量机的滚动轴承故障诊断", 《科学技术与工程》 *

Similar Documents

Publication Publication Date Title
CN108737406B (en) Method and system for detecting abnormal flow data
CN109741292A (en) The method for detecting abnormal image in the first image data set with confrontation self-encoding encoder
Wahono et al. Neural network parameter optimization based on genetic algorithm for software defect prediction
CN113687972B (en) Processing method, device, equipment and storage medium for abnormal data of business system
CN112765003B (en) Risk prediction method based on APP behavior log
CN110956278A (en) Method and system for retraining machine learning models
CN116756594A (en) Method, system, equipment and medium for detecting abnormal points of power grid data
CN115759033A (en) Method, device and equipment for processing track data
CN111881159A (en) Fault detection method and device based on cost-sensitive extreme random forest
CN115705413A (en) Method and device for determining abnormal log
CN112488202A (en) Distributed process monitoring method based on mutual information and projection dictionary pair learning
CN115758086A (en) Method, device and equipment for detecting faults of cigarette cut-tobacco drier and readable storage medium
CN113221998A (en) Rare earth extraction stirring shaft fault diagnosis method and system based on SSA-SVM
CN116188445A (en) Product surface defect detection and positioning method and device and terminal equipment
CN113705624B (en) Intrusion detection method and system for industrial control system
CN114385472A (en) Abnormal data detection method, device, equipment and storage medium
CN114741690A (en) Network security monitoring method, device, equipment and computer storage medium
CN114553468A (en) Three-level network intrusion detection method based on feature intersection and ensemble learning
CN114429166A (en) Method, device and equipment for acquiring high-dimensional features of data and computer storage medium
CN109583590B (en) Data processing method and data processing device
CN112232419A (en) Rotating equipment fault monitoring method and system based on big data and readable storage medium
CN111522750A (en) Method and system for processing function test problem
CN117349770B (en) Structural health monitoring multi-strain sensor data anomaly detection and repair method
Oliker et al. Classification-optimization model for contamination event detection in water distribution systems
CN117152550A (en) Neuron coverage guide type test method for automatic driving system

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210806