CN112507790B - Fault diagnosis method and system of complementary classification regression tree based on differential evolution - Google Patents

Fault diagnosis method and system of complementary classification regression tree based on differential evolution Download PDF

Info

Publication number
CN112507790B
CN112507790B CN202011211079.4A CN202011211079A CN112507790B CN 112507790 B CN112507790 B CN 112507790B CN 202011211079 A CN202011211079 A CN 202011211079A CN 112507790 B CN112507790 B CN 112507790B
Authority
CN
China
Prior art keywords
regression tree
classification regression
complementary
individual
population
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
CN202011211079.4A
Other languages
Chinese (zh)
Other versions
CN112507790A (en
Inventor
马速良
李建林
李金林
李雅欣
李穷
谭宇良
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.)
Beijing Lianzhi Huineng Technology Co ltd
North China University of Technology
Anhui Lvwo Recycling Energy Technology Co Ltd
Original Assignee
Beijing Lianzhi Huineng Technology Co ltd
North China University of Technology
Anhui Lvwo Recycling Energy Technology Co Ltd
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 Beijing Lianzhi Huineng Technology Co ltd, North China University of Technology, Anhui Lvwo Recycling Energy Technology Co Ltd filed Critical Beijing Lianzhi Huineng Technology Co ltd
Priority to CN202011211079.4A priority Critical patent/CN112507790B/en
Publication of CN112507790A publication Critical patent/CN112507790A/en
Application granted granted Critical
Publication of CN112507790B publication Critical patent/CN112507790B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—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
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Signal Processing (AREA)
  • Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)

Abstract

The invention relates to a fault diagnosis method and system of a complementary classification regression tree based on differential evolution. The method comprises the following steps: acquiring a sample set; the sample set comprises sample signals corresponding to various fault categories, and each sample signal is an operation signal of the equipment under the corresponding fault type; analyzing each sample signal in the sample set to obtain a sample feature vector set consisting of all sample feature vectors; according to the sample feature vector set, a complementary classification regression tree model is obtained by taking a genetic algorithm as a differential evolution basis; the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree; determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all the leaf nodes of the classification regression tree and the number of the leaf nodes to obtain a fault diagnosis model of the equipment; and carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment. The invention can improve the performance of equipment fault diagnosis.

Description

Fault diagnosis method and system of complementary classification regression tree based on differential evolution
Technical Field
The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method and system of a complementary classification regression tree based on differential evolution.
Background
With the development of machine learning and intelligent optimization technology, the demand for intelligent monitoring service for identifying the state of running equipment is increasing. The classification regression tree algorithm has the advantages of simple and easy realization, strong interpretation, graphical property and the like, and is often used for monitoring the health condition of the operation equipment. However, due to the influence of data noise and human experience, the characteristics of the monitoring signals of the operation equipment are often extracted, and the diagnosis performance of the model is easily affected by invalid and redundant characteristics. To obtain better diagnostic model performance, intelligent optimization techniques are beginning to be applied to optimize diagnostic model parameters or analyze feature importance. At present, more research is conducted on optimizing model parameters based on an intelligent optimization algorithm, evaluation on the performance of a diagnosis model is generally designed based on additional verification data or a cross verification mode, a directional iterative optimization process is realized on the diagnosis model parameters, and a high-quality strong diagnosis model is formed. However, the way to add additional validation data beyond the training data set test data set is often only meaningful in large data sample scale applications; the cross-validation method for splitting the training data set into a plurality of subsets has the problems of complex process, large calculated amount, large influence of sample set division, easy reduction of training data capacity and the like. In another aspect, the fault diagnosis performance is affected by multiple aspects such as data, features, models and the like, and the influence caused by the feature advantages and disadvantages is ignored by the optimization of the parameters of the diagnosis model due to the over emphasis, so that the improvement of the diagnosis performance is limited.
In order to reduce the time and cost of manually checking the faults of the components one by one and improve the application performance and applicability of the fault diagnosis technology, the fault diagnosis method of the running equipment and the intelligent optimization method are flexibly combined, and the research and development of the fault detection method of the running equipment with high efficiency and high precision has great and profound significance.
Disclosure of Invention
The invention aims to provide a fault diagnosis method and system for a complementary classification regression tree based on differential evolution, so as to improve the performance of equipment fault diagnosis.
In order to achieve the above object, the present invention provides the following solutions:
a fault diagnosis method of complementary classification regression tree based on differential evolution comprises the following steps:
acquiring a sample set; the sample set comprises sample signals corresponding to various fault types, and each sample signal is an operation signal of equipment under the corresponding fault type;
analyzing each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtaining a sample feature vector set formed by all sample feature vectors;
according to the sample feature vector set, a complementary classification regression tree model is obtained by taking a genetic algorithm as a differential evolution basis; the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population;
Determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all leaf nodes of the classification regression tree and the number of the leaf nodes to obtain a fault diagnosis model of the equipment;
and carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment.
Optionally, the obtaining a complementary classification regression tree model based on a genetic algorithm as a differential evolution basis according to the sample feature vector set specifically includes:
initializing parameters of a genetic algorithm; the parameters of the genetic algorithm include: the method comprises the steps of an initial population and iteration stopping conditions, wherein each individual in the initial population is represented by binary codes, and the binary code length of each individual is the feature quantity of a sample feature vector;
for each iteration, obtaining an original population and a complementary population of the current iteration; the binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration, the original population of the current iteration is the initial population;
for the jth individual of the original population, selecting partial features of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual;
Selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individuals in the complementary population for the jth individuals in the complementary population, and constructing a classification regression tree corresponding to the jth individuals to obtain a complementary classification regression tree of the jth individuals;
sequentially obtaining an original classification regression tree and a complementary classification regression tree of each individual;
calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual;
determining the optimal individuals of the current iteration according to the fitness values of all the individuals; the optimal individual in the current iteration is the individual with the highest fitness value in all individuals;
according to the optimal individual of the current iteration, determining the historical optimal individual of the current iteration;
judging whether an iteration stop condition is reached;
when the iteration stopping condition is reached, determining an original classification regression tree and a complementary classification regression tree corresponding to the historical optimal individual of the current iteration as the complementary classification regression tree model;
and when the iteration stopping condition is not reached, updating the original population by adopting a genetic algorithm to obtain the original population of the next iteration, adding 1 to the iteration times, returning to the step of acquiring the original population and the complementary population of the current iteration, and entering the next iteration.
Optionally, for the jth individual in the original population, selecting a part of features of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual, which specifically includes:
obtaining P coding bits with the value of 1 in the binary coding of the jth individual in the original population;
according to the coding bit with the value of 1, P features corresponding to the P coding bits in each sample feature vector in the sample feature vector set are obtained;
and generating a classification regression tree according to the P features of each sample feature vector to obtain the original classification regression tree of the jth individual.
Optionally, for the jth individual in the complementary population, selecting a part of features of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the complementary population, and constructing a classification regression tree corresponding to the jth individual to obtain a complementary classification regression tree of the jth individual, which specifically includes:
acquiring Q coding bits with the value of 1 in the binary coding of the jth individual in the complementary population;
According to the coding bit with the value of 1, Q characteristics corresponding to the Q coding bits in each sample characteristic vector in the sample characteristic vector set are obtained;
and generating a classification regression tree according to the Q features of each sample feature vector to obtain a complementary classification regression tree of the jth individual.
Optionally, the calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual specifically includes:
for the jth individual, obtaining the sum of all leaf node base indexes in the original classification regression tree of the jth individualAnd total leaf node number->
Obtaining the sum of all leaf node base indexes in the complementary classification regression tree of the jth individualAnd total leaf node number->
Using the formulaCalculating fitness value +.>Where t represents the number of iterations.
Optionally, the determining the historical optimal individual of the current iteration according to the optimal individual of the current iteration specifically includes:
judging whether the fitness value of the optimal individual in the current iteration is larger than that of the historical optimal individual in the previous iteration;
if the fitness value of the optimal individual of the current iteration is larger than that of the historical optimal individual of the previous iteration, determining the optimal individual of the current iteration as the historical optimal individual of the current iteration;
And if the fitness value of the optimal individual of the current iteration is not greater than that of the historical optimal individual of the previous iteration, determining the historical optimal individual of the previous iteration as the historical optimal individual of the current iteration.
Optionally, determining the optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all the leaf nodes and the number of the leaf nodes to obtain a fault diagnosis model of the equipment specifically includes:
obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number->
Obtaining the sum of all leaf node base indexes in the complementary classification regression tree modelAnd total leaf node number->
Judging whether or not to meet
If it meetsDetermining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment;
if it does not meetAnd determining an original classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment.
The invention also provides a fault diagnosis system of the complementary classification regression tree based on differential evolution, which comprises:
the sample set acquisition module is used for acquiring a sample set; the sample set comprises sample signals corresponding to various fault types, and each sample signal is an operation signal of equipment under the corresponding fault type;
The sample analysis module is used for analyzing each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtaining a sample feature vector set formed by all sample feature vectors;
the complementary classification regression tree model acquisition module is used for acquiring a complementary classification regression tree model by taking a genetic algorithm as a differential evolution basis according to the sample feature vector set; the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population;
the fault diagnosis model acquisition module is used for determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all leaf nodes of the classification regression tree and the number of the leaf nodes to obtain a fault diagnosis model of the equipment;
and the fault diagnosis module is used for carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment.
Optionally, the complementary classification regression tree model obtaining module specifically includes:
An initializing unit for initializing parameters of a genetic algorithm; the parameters of the genetic algorithm include: the method comprises the steps of an initial population and iteration stopping conditions, wherein each individual in the initial population is represented by binary codes, and the binary code length of each individual is the feature quantity of a sample feature vector;
the population acquisition unit is used for acquiring an original population and a complementary population of the current iteration for each iteration; the binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration, the original population of the current iteration is the initial population;
the original classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the original population for the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual; sequentially obtaining an original classification regression tree of each individual;
the complementary classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the complementary population for the jth individual in the complementary population, and constructing a classification regression tree corresponding to the jth individual to obtain a complementary classification regression tree of the jth individual; sequentially obtaining complementary classification regression trees of each individual;
The fitness value calculating unit is used for calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual;
the optimal individual determining unit is used for determining the optimal individual of the current iteration according to the fitness values of all the individuals; the optimal individual in the current iteration is the individual with the highest fitness value in all individuals;
the historical optimal individual determining unit is used for determining the historical optimal individual of the current iteration according to the optimal individual of the current iteration;
an iteration stop judging unit for judging whether an iteration stop condition is reached;
the complementary classification regression tree model determining unit is used for determining an original classification regression tree and a complementary classification regression tree corresponding to the current iterative historical optimal individual as the complementary classification regression tree model when the iteration stopping condition is reached;
and the iteration unit is used for updating the original population by adopting a genetic algorithm when the iteration stopping condition is not reached, obtaining the original population of the next iteration, adding 1 to the iteration times, returning to the step of obtaining the original population and the complementary population of the current iteration, and entering the next iteration.
Optionally, the fault diagnosis model obtaining module specifically includes:
A leaf node base index sum and leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number->
A leaf node base index sum and leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the complementary classification regression tree modelAnd total leaf node number->
A judging unit for judging whether or not the condition is satisfied
A fault diagnosis model determining unit for, when meetingWhen the fault diagnosis method is used, determining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment; when do not meet->And determining an original classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
1. the invention utilizes the binary coding process of individuals in the genetic algorithm, carries out complementary feature selection based on the binary coding of the individuals and the 0-1 mutual complement thereof, establishes a group of complementary classification regression tree models, carries out the difference of characterization diagnosis performance, and provides an evaluation standard for the genetic evolution process. Compared with other genetic algorithm optimization and classification regression tree diagnosis methods, the method provides a new thought for the combination of genetic algorithm optimization and classification regression tree diagnosis, and the optimization process is beneficial to reducing the influence of invalid and redundant characteristic variables and improving the performance of the classification regression tree diagnosis model.
2. The invention generates a new population which is opposite and complementary with the original population in a genetic algorithm, further generates a complementary feature space to establish a group of classified regression tree models, and evaluates features and models by utilizing the base index sum of classified regression leaf nodes. Compared with a method for evaluating the performance of a diagnosis model by using an additional verification data set or a cross verification mode diagnosis result, the method reduces the requirement on the sample data scale, is suitable for solving the problem of fault diagnosis in a small sample environment, and improves the applicability of a diagnosis scheme.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a fault diagnosis method of a complementary classification regression tree based on differential evolution;
FIG. 2 is a diagram of a circuit breaker vibration signal acquisition site of the energy storage system of the present invention;
FIG. 3 is a flow chart of genetic differential evolution of the present invention;
FIG. 4 is a schematic illustration of the selection of features by the jth individual in the original population of the t generation of the present invention;
FIG. 5 is a schematic representation of a complementary classification regression tree generated by a jth individual in a t-th generation of original population according to the present invention;
fig. 6 is a schematic structural diagram of a fault diagnosis system based on a complementary classification regression tree of differential evolution according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
FIG. 1 is a flow chart of a fault diagnosis method of a complementary classification regression tree based on differential evolution. As shown in fig. 1, the fault diagnosis method of the complementary classification regression tree based on differential evolution of the invention comprises the following steps:
Step 100: a sample set is obtained. The invention uses the measuring device to collect the operation signal which can reflect the state of the monitored equipment under various different fault conditions as the sample signal, and marks each sample signal with the fault type, thus obtaining the sample set containing the fault type. Taking a breaker with monitored equipment as an energy storage system as an example, collecting vibration signals of the breaker under S different fault conditions by using a vibration information measuring device, wherein the total number of the vibration signals is m, and the vibration signals are V (i) (t) denotes i=1, 2, …, m. Then, labeling fault categories for each signal to obtain a sample set A containing the fault categories,wherein (1)>Indicating that the vibration signal of the ith measurement belongs to fault C s 。
Step 200: and analyzing each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtaining a sample feature vector set formed by all sample feature vectors. For example, n key features capable of reflecting differences among samples measured under different faults to a certain extent are defined by adopting methods such as statistics, time domain energy, fourier spectrum analysis, wavelet transformation, empirical mode decomposition, information entropy and the like, so that feature vectors of each sample signal are obtained, and a sample feature vector set B containing fault types is formed. For example, the calculation of the eigenvector of the ith sample signal can be expressed as
Step 300: and obtaining a complementary classification regression tree model by taking a genetic algorithm as a differential evolution basis according to the sample feature vector set. The complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population. The specific process is shown in fig. 3, and comprises the following steps:
step1: parameters of the genetic algorithm are initialized.
The parameters of the genetic algorithm include: the number k of individuals in the initial population, the length of the binary codes of the individuals are equal to the vector dimension m in the feature set B (namely the number of features in the feature vector), the selectivity Ps, the crossover rate Pm, the mutation rate Px and the iteration stop condition (the maximum iteration number G). Each individual in the initial population is represented by a randomly generated binary code.
Step2: the original population and the complementary population of the current iteration are obtained.
The binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration The original population of the current iteration is the initial population. For example, in the t iteration, the binary code of the jth individual in the original population isMutually arranging 0 and 1 in binary codes to generate binary code of jth individual in complementary population as +.>Thus, the binary code of each individual in the complementary population can be obtained according to the binary code of each individual in the original population, and the complementary population is obtained.
Step3: and sequentially obtaining an original classification regression tree and a complementary classification regression tree of each individual.
And for the jth individual of the original population, selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to the binary code of the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual. As shown in fig. 4, the coded bits of the binary codes of the individuals are in one-to-one correspondence with the features of the sample feature vectors, and the invention screens the features of the sample feature vectors by using the coded bits with the value of 1 in the binary codes of each individual.
Specifically, as shown in fig. 5, for the jth individual in the original population, according to P coding bits with a value of 1 in binary codes of the jth individual in the original population, selecting P features corresponding to the P coding bits in each sample feature vector in the sample feature vector set, training and generating a classification regression tree to obtain an original classification regression tree of the jth individual
For the jth individual of the complementary population, selecting the sample feature vector set and Q codes in each sample feature vector according to the Q code bits with the value of 1 in the binary codes of the jth individual in the complementary populationTraining and generating Q features corresponding to the code bits to obtain a classification regression tree to obtain a complementary classification regression tree of the jth individual
A pair of complementary classification regression trees corresponding to each individual are sequentially obtained by adopting the method
Step4: and calculating the leaf node base index sum in the complementary classification regression tree model and dividing the leaf node base index sum by the absolute value of the difference of the respective leaf node number according to the original classification regression tree and the complementary classification regression tree of each individual, and further evaluating the adaptability of each individual.
Original classification regression tree generated for the jth generation, jth individual of original populationThe base index calculation of the L-th leaf node of (2) is shown in the formula (1):
wherein p represents the proportion of various fault samples in the L-th leaf node, N s (L) represents the number of class s failure samples in the L-th leaf node,representing the total number of samples in the L-th leaf node.
And the original classification regression tree generated by the jth generation of the jth individual of the original populationThe sum of the base indexes of all leaf nodes is calculated as shown in formula (2):
Based on the same method, the sum of the base indexes of all leaf nodes of the complementary classification regression tree generated by the jth generation of individuals of the complementary population can be obtained
Defining the fitness value of the jth generation and the jth individual asAnd further, the advantages and disadvantages of individuals in the population are evaluated, and the larger the fitness value is, the better the individuals are. Wherein->Leaf node number of original classification regression tree generated for the t th generation, the jth individual of the original population,/-, and>the number of leaf nodes of the complementary classification regression tree generated for the t-th generation, the jth individual of the complementary population.
Sorting the fitness values of all individuals in the t generation, and according to the fitness function values of the individualsFind the maximum (assuming->) And determining the individual with the largest fitness value as the optimal individual of the current iteration. Then judging whether the optimal individual of the t generation is better than the historical optimal individual obtained in the t-1 generation, namely, judging the fitness function value J of the optimal individual of the t generation and the historical optimal individual OPT Comparison (if t=1, fitness value J of history optimal individual OPT =0). If the optimal individual of the t th generation is better than the history optimal individual obtained in the t-1 th generation, namely +.>The history optimal individual is updated, the optimal individual of the t generation is determined as the history optimal individual of the t generation, and the fitness value of the history optimal individual of the t generation is +. >If the optimal individual of the t generation is not better than the historical optimal individual obtained in the t-1 generation, the historical optimal individual is unchanged, and the historical optimal individual obtained in the t-1 generation is determined to be the historical optimal individual of the t generation.
Step5: and judging whether an iteration stop condition is reached. For example, it is determined whether the maximum number of iterations G is reached. If yes, executing Step6; if not, step7 is performed.
Step6: determining an original classification regression tree and a complementary classification regression tree corresponding to a history optimal individual in the current iteration as the complementary classification regression tree model;
step7: updating individuals in the t generation of the original population by using the traditional operation processes of the selectivity Ps, the crossover rate Pm, the mutation rate Px and the like in the classical genetic optimization algorithm to form the t+1th generation original population, returning to Step2 for the iteration times t=t+1, and entering the next iteration.
Step 400: and determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all the leaf nodes of the classification regression tree and the number of the leaf nodes to obtain a fault diagnosis model of the equipment.
Based on the complementary classification regression tree model optimized by the genetic algorithm in step 300, namely a group of optimal classification regression treesThe minimum calculated value of the classification regression Tree is selected as the optimal classification regression Tree by calculating the sum of the base indexes of all leaf nodes in each Tree divided by the number of the leaf nodes, namely the final fault diagnosis model Tree OPT As shown in formula (3):
wherein Q is min Index represents the Index (position) of the minimum value.
Or determining a fault diagnosis model of the equipment in a judging mode, wherein the specific process is as follows:
obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number->
Obtaining the sum of all leaf node base indexes in the complementary classification regression tree modelAnd total leaf node number->
Judging whether or not to meet
If it meetsDetermining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment;
if it does not meetAnd determining an original classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment.
Step 500: and carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment. In particular, the newly measured device operating signal to be diagnosed is advantageously usedForming feature vectors for diagnosis by using the key feature calculation method designed in the step 200(i=1, 2, …, m) represents the eigenvector of the new measured, i-th sample to be diagnosed; the feature vectors are then input into a final fault diagnosis model Tree OPT The fault diagnosis model returns the predicted fault type Cs (s=1, 2, …, S) of the running equipment under the measurement signal, and the diagnosis test is completed.
Based on the method, the invention also provides a fault diagnosis system of the complementary classification regression tree based on differential evolution. Fig. 6 is a schematic structural diagram of a fault diagnosis system based on a differential evolution complementary classification regression tree according to the present invention, and as shown in fig. 6, the fault diagnosis system based on a differential evolution complementary classification regression tree according to the present invention includes the following structures:
a sample set acquisition module 601, configured to acquire a sample set; the sample set comprises sample signals corresponding to various fault types, and each sample signal is an operation signal of equipment under the corresponding fault type.
The sample analysis module 602 is configured to analyze each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtain a sample feature vector set formed by all sample feature vectors.
The complementary classification regression tree model obtaining module 603 is configured to obtain a complementary classification regression tree model based on a genetic algorithm as a differential evolution basis according to the sample feature vector set; the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population.
The fault diagnosis model obtaining module 604 is configured to determine an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all the leaf nodes and the number of the leaf nodes, and obtain a fault diagnosis model of the device.
The fault diagnosis module 605 is configured to perform fault diagnosis on the device by using a fault diagnosis model of the device based on the operation signal of the device.
As a specific embodiment, the complementary classification regression tree model obtaining module 603 of the present invention specifically includes:
an initializing unit for initializing parameters of a genetic algorithm; the parameters of the genetic algorithm include: an initial population and an iteration stop condition, wherein each individual in the initial population is represented by a binary code, and the binary code length of each individual is the characteristic quantity of the sample characteristic vector.
The population acquisition unit is used for acquiring an original population and a complementary population of the current iteration for each iteration; the binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration, the original population of the current iteration is the initial population.
The original classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the original population for the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual; and sequentially obtaining the original classification regression tree of each individual.
The complementary classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the complementary population for the jth individual in the complementary population, and constructing a classification regression tree corresponding to the jth individual to obtain a complementary classification regression tree of the jth individual; and obtaining the complementary classification regression tree of each individual in turn.
And the fitness value calculating unit is used for calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual.
The optimal individual determining unit is used for determining the optimal individual of the current iteration according to the fitness values of all the individuals; the optimal individual in the current iteration is the individual with the highest fitness value among all the individuals.
And the historical optimal individual determining unit is used for determining the historical optimal individual of the current iteration according to the optimal individual of the current iteration.
And the iteration stop judging unit is used for judging whether the iteration stop condition is reached.
And the complementary classification regression tree model determining unit is used for determining the original classification regression tree and the complementary classification regression tree corresponding to the historical optimal individual of the current iteration as the complementary classification regression tree model when the iteration stopping condition is reached.
And the iteration unit is used for updating the original population by adopting a genetic algorithm when the iteration stopping condition is not reached, obtaining the original population of the next iteration, adding 1 to the iteration times, returning to the step of obtaining the original population and the complementary population of the current iteration, and entering the next iteration.
As a specific embodiment, the fault diagnosis model obtaining module 604 of the present invention specifically includes:
a leaf node base index sum and leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number->
A leaf node base index sum and leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the complementary classification regression tree model And total leaf node number->
A judging unit for judging whether or not the condition is satisfied
A fault diagnosis model determining unit for, when meetingWhen the fault diagnosis method is used, determining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment; when do not meet->And determining an original classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment. />
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.

Claims (6)

1. The fault diagnosis method of the complementary classification regression tree based on differential evolution is characterized by comprising the following steps of:
acquiring a sample set; the sample set comprises sample signals corresponding to various fault types, and each sample signal is an operation signal of equipment under the corresponding fault type;
analyzing each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtaining a sample feature vector set formed by all sample feature vectors;
according to the sample feature vector set, a complementary classification regression tree model is obtained by taking a genetic algorithm as a differential evolution basis, and the method specifically comprises the following steps:
initializing parameters of a genetic algorithm; the parameters of the genetic algorithm include: the method comprises the steps of an initial population and iteration stopping conditions, wherein each individual in the initial population is represented by binary codes, and the binary code length of each individual is the feature quantity of a sample feature vector;
for each iteration, obtaining an original population and a complementary population of the current iteration; the binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration, the original population of the current iteration is the initial population;
For the jth individual of the original population, selecting partial features of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual;
selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individuals in the complementary population for the jth individuals in the complementary population, and constructing a classification regression tree corresponding to the jth individuals to obtain a complementary classification regression tree of the jth individuals;
sequentially obtaining an original classification regression tree and a complementary classification regression tree of each individual;
calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual;
determining the optimal individuals of the current iteration according to the fitness values of all the individuals; the optimal individual in the current iteration is the individual with the highest fitness value in all individuals;
according to the optimal individual of the current iteration, determining the historical optimal individual of the current iteration;
judging whether an iteration stop condition is reached;
When the iteration stopping condition is reached, determining an original classification regression tree and a complementary classification regression tree corresponding to the historical optimal individual of the current iteration as the complementary classification regression tree model;
when the iteration stopping condition is not reached, updating the original population by adopting a genetic algorithm to obtain an original population of the next iteration, adding 1 to the iteration times, returning to the step of acquiring the original population and the complementary population of the current iteration, and entering the next iteration;
the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population;
determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all the leaf nodes and the number of the leaf nodes to obtain a fault diagnosis model of the equipment, wherein the method specifically comprises the following steps:
obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number->
Obtaining the sum of all leaf node base indexes in the complementary classification regression tree model And total leaf node number->
Judging whether or not to meet
If it meetsDetermining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment;
if it does not meetDetermining an original classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment;
and carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment.
2. The fault diagnosis method for complementary classification regression tree based on differential evolution according to claim 1, wherein for the jth individual of the original population, selecting a part of characteristics of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain the original classification regression tree of the jth individual, specifically comprising:
obtaining P coding bits with the value of 1 in the binary coding of the jth individual in the original population;
according to the coding bit with the value of 1, P features corresponding to the P coding bits in each sample feature vector in the sample feature vector set are obtained;
And generating a classification regression tree according to the P features of each sample feature vector to obtain the original classification regression tree of the jth individual.
3. The fault diagnosis method for complementary classification regression tree based on differential evolution according to claim 1, wherein for the jth individual of the complementary population, selecting a part of characteristics of each sample feature vector in the sample feature vector set according to the binary code of the jth individual in the complementary population, and constructing a classification regression tree corresponding to the jth individual to obtain a complementary classification regression tree of the jth individual, which specifically comprises:
acquiring Q coding bits with the value of 1 in the binary coding of the jth individual in the complementary population;
according to the coding bit with the value of 1, Q characteristics corresponding to the Q coding bits in each sample characteristic vector in the sample characteristic vector set are obtained;
and generating a classification regression tree according to the Q features of each sample feature vector to obtain a complementary classification regression tree of the jth individual.
4. The fault diagnosis method for complementary classification regression tree based on differential evolution according to claim 1, wherein the calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual specifically comprises:
For the jth individual, obtaining the sum of all leaf node base indexes in the original classification regression tree of the jth individualAnd total leaf node number->
Obtaining the sum of all leaf node base indexes in the complementary classification regression tree of the jth individualAnd total leaf node number->
Using the formulaCalculating fitness value +.>Where t represents the number of iterations.
5. The fault diagnosis method based on complementary classification regression tree of differential evolution according to claim 1, wherein the determining the historical optimal individual of the current iteration according to the optimal individual of the current iteration specifically comprises:
judging whether the fitness value of the optimal individual in the current iteration is larger than that of the historical optimal individual in the previous iteration;
if the fitness value of the optimal individual of the current iteration is larger than that of the historical optimal individual of the previous iteration, determining the optimal individual of the current iteration as the historical optimal individual of the current iteration;
and if the fitness value of the optimal individual of the current iteration is not greater than that of the historical optimal individual of the previous iteration, determining the historical optimal individual of the previous iteration as the historical optimal individual of the current iteration.
6. A fault diagnosis system based on a complementary classification regression tree of differential evolution, comprising:
The sample set acquisition module is used for acquiring a sample set; the sample set comprises sample signals corresponding to various fault types, and each sample signal is an operation signal of equipment under the corresponding fault type;
the sample analysis module is used for analyzing each sample signal in the sample set to obtain a sample feature vector corresponding to each sample signal, and further obtaining a sample feature vector set formed by all sample feature vectors;
the complementary classification regression tree model acquisition module is used for acquiring a complementary classification regression tree model by taking a genetic algorithm as a differential evolution basis according to the sample feature vector set; the complementary classification regression tree model comprises an original classification regression tree and a complementary classification regression tree, wherein the original classification regression tree is a classification regression tree corresponding to an original population, and the complementary classification regression tree model is a classification regression tree corresponding to a complementary population;
the complementary classification regression tree model acquisition module specifically comprises:
an initializing unit for initializing parameters of a genetic algorithm; the parameters of the genetic algorithm include: the method comprises the steps of an initial population and iteration stopping conditions, wherein each individual in the initial population is represented by binary codes, and the binary code length of each individual is the feature quantity of a sample feature vector;
The population acquisition unit is used for acquiring an original population and a complementary population of the current iteration for each iteration; the binary codes of the jth individual in the complementary population of the current iteration are the coding sequences obtained by mutually arranging 0 and 1 for the binary codes of the jth individual in the original population; when the current iteration is the 1 st iteration, the original population of the current iteration is the initial population;
the original classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the original population for the jth individual in the original population, and constructing a classification regression tree corresponding to the jth individual to obtain an original classification regression tree of the jth individual; sequentially obtaining an original classification regression tree of each individual;
the complementary classification regression tree construction unit is used for selecting partial characteristics of each sample characteristic vector in the sample characteristic vector set according to binary codes of the jth individual in the complementary population for the jth individual in the complementary population, and constructing a classification regression tree corresponding to the jth individual to obtain a complementary classification regression tree of the jth individual; sequentially obtaining complementary classification regression trees of each individual;
The fitness value calculating unit is used for calculating the fitness value of each individual according to the original classification regression tree and the complementary classification regression tree of each individual;
the optimal individual determining unit is used for determining the optimal individual of the current iteration according to the fitness values of all the individuals; the optimal individual in the current iteration is the individual with the highest fitness value in all individuals;
the historical optimal individual determining unit is used for determining the historical optimal individual of the current iteration according to the optimal individual of the current iteration;
an iteration stop judging unit for judging whether an iteration stop condition is reached;
the complementary classification regression tree model determining unit is used for determining an original classification regression tree and a complementary classification regression tree corresponding to the current iterative historical optimal individual as the complementary classification regression tree model when the iteration stopping condition is reached;
the iteration unit is used for updating the original population by adopting a genetic algorithm when the iteration stopping condition is not reached, obtaining the original population of the next iteration, adding 1 to the iteration times, returning to the step of obtaining the original population and the complementary population of the current iteration, and entering the next iteration;
the fault diagnosis model acquisition module is used for determining an optimal classification regression tree in the complementary classification regression tree model based on the sum of the base indexes of all leaf nodes of the classification regression tree and the number of the leaf nodes to obtain a fault diagnosis model of the equipment;
The fault diagnosis model acquisition module specifically comprises:
a leaf node base index sum and leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the original classification regression tree in the complementary classification regression tree modelAnd total leaf node number
Complementary to each otherA leaf node base index sum leaf node number calculation unit for obtaining the sum of all leaf node base indexes in the complementary classification regression tree modelAnd total leaf node number
A judging unit for judging whether or not the condition is satisfied
A fault diagnosis model determining unit for, when meetingWhen the fault diagnosis method is used, determining a complementary classification regression tree in the complementary classification regression tree model as a fault diagnosis model of the equipment; when do not meet->When the fault diagnosis method is used, an original classification regression tree in the complementary classification regression tree model is determined to be a fault diagnosis model of the equipment;
and the fault diagnosis module is used for carrying out fault diagnosis on the equipment by adopting a fault diagnosis model of the equipment based on the operation signal of the equipment.
CN202011211079.4A 2020-11-03 2020-11-03 Fault diagnosis method and system of complementary classification regression tree based on differential evolution Active CN112507790B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011211079.4A CN112507790B (en) 2020-11-03 2020-11-03 Fault diagnosis method and system of complementary classification regression tree based on differential evolution

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011211079.4A CN112507790B (en) 2020-11-03 2020-11-03 Fault diagnosis method and system of complementary classification regression tree based on differential evolution

Publications (2)

Publication Number Publication Date
CN112507790A CN112507790A (en) 2021-03-16
CN112507790B true CN112507790B (en) 2023-09-29

Family

ID=74955214

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011211079.4A Active CN112507790B (en) 2020-11-03 2020-11-03 Fault diagnosis method and system of complementary classification regression tree based on differential evolution

Country Status (1)

Country Link
CN (1) CN112507790B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106067066A (en) * 2016-05-31 2016-11-02 西安工程大学 Method for diagnosing fault of power transformer based on genetic algorithm optimization pack algorithm
CN109597401A (en) * 2018-12-06 2019-04-09 华中科技大学 A kind of equipment fault diagnosis method based on data-driven
CN109840541A (en) * 2018-12-05 2019-06-04 国网辽宁省电力有限公司信息通信分公司 A kind of network transformer Fault Classification based on XGBoost
CN110082136A (en) * 2019-04-19 2019-08-02 电子科技大学 Rotary machinery fault diagnosis method based on Retrieval method Support Vector Machines Optimized
CN111242302A (en) * 2019-12-27 2020-06-05 冶金自动化研究设计院 XGboost prediction method of intelligent parameter optimization module
CN111562541A (en) * 2020-05-31 2020-08-21 宁夏隆基宁光仪表股份有限公司 Software platform for realizing electric energy meter detection data management by applying CART algorithm

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2882171A1 (en) * 2005-02-14 2006-08-18 France Telecom METHOD AND DEVICE FOR GENERATING A CLASSIFYING TREE TO UNIFY SUPERVISED AND NON-SUPERVISED APPROACHES, COMPUTER PROGRAM PRODUCT AND CORRESPONDING STORAGE MEDIUM
US11500056B2 (en) * 2015-07-17 2022-11-15 Origin Wireless, Inc. Method, apparatus, and system for wireless tracking with graph-based particle filtering

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106067066A (en) * 2016-05-31 2016-11-02 西安工程大学 Method for diagnosing fault of power transformer based on genetic algorithm optimization pack algorithm
CN109840541A (en) * 2018-12-05 2019-06-04 国网辽宁省电力有限公司信息通信分公司 A kind of network transformer Fault Classification based on XGBoost
CN109597401A (en) * 2018-12-06 2019-04-09 华中科技大学 A kind of equipment fault diagnosis method based on data-driven
CN110082136A (en) * 2019-04-19 2019-08-02 电子科技大学 Rotary machinery fault diagnosis method based on Retrieval method Support Vector Machines Optimized
CN111242302A (en) * 2019-12-27 2020-06-05 冶金自动化研究设计院 XGboost prediction method of intelligent parameter optimization module
CN111562541A (en) * 2020-05-31 2020-08-21 宁夏隆基宁光仪表股份有限公司 Software platform for realizing electric energy meter detection data management by applying CART algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Differential Evolution and Perceptron Decision Trees for Fault Detection in Power Transformers;Freitas A R R等;《Soft Computing Models in Industrial and Environmental Applications》;第143-152页 *
采用遗传算法优化装袋分类回归树组合算法的变压器故障诊断;黄新波等;《高电压技术》;第42卷(第5期);第1617-1623页 *

Also Published As

Publication number Publication date
CN112507790A (en) 2021-03-16

Similar Documents

Publication Publication Date Title
CN111914873B (en) Two-stage cloud server unsupervised anomaly prediction method
CN109000930B (en) Turbine engine performance degradation evaluation method based on stacking denoising autoencoder
CN111914883B (en) Spindle bearing state evaluation method and device based on deep fusion network
CN107423414B (en) Information transfer model-based process industry complex electromechanical system fault tracing method
CN116306320B (en) Method and system for managing production quality of non-vascular lumen guide wire
CN112257263B (en) Equipment residual life prediction system based on self-attention mechanism
CN113259216B (en) CAN bus signal analysis method and system
CN114239718B (en) High-precision long-term time sequence prediction method based on multi-element time sequence data analysis
CN112147432A (en) BiLSTM module based on attention mechanism, transformer state diagnosis method and system
CN114529228A (en) Risk early warning method and system for power monitoring system supply chain
CN115169430A (en) Cloud network end resource multidimensional time sequence anomaly detection method based on multi-scale decoding
CN116401532B (en) Method and system for recognizing frequency instability of power system after disturbance
CN114266289A (en) Complex equipment health state assessment method
CN116665483A (en) Novel method for predicting residual parking space
CN113485863A (en) Method for generating heterogeneous unbalanced fault samples based on improved generation countermeasure network
CN115184054A (en) Mechanical equipment semi-supervised fault detection and analysis method, device, terminal and medium
CN114548493A (en) Method and system for predicting current overload of electric energy meter
CN112507790B (en) Fault diagnosis method and system of complementary classification regression tree based on differential evolution
CN117477544A (en) LSTM ultra-short-term photovoltaic power prediction method and system integrating time mode features
CN117194219A (en) Fuzzy test case generation and selection method, device, equipment and medium
CN112767692A (en) Short-term traffic flow prediction system based on SARIMA-GA-Elman combined model
CN116910559A (en) Index anomaly detection method for intelligent operation and maintenance application of power grid supercomputer center
CN116304849A (en) Two-dimensional piston pump fault diagnosis method based on local cut space arrangement and gating circulation network
CN112488805B (en) Long-renting market early warning method based on multiple regression time series analysis
CN113821401A (en) WT-GA-GRU model-based cloud server fault diagnosis method

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