CN111191727A - Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ - Google Patents
Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ Download PDFInfo
- Publication number
- CN111191727A CN111191727A CN201911420812.0A CN201911420812A CN111191727A CN 111191727 A CN111191727 A CN 111191727A CN 201911420812 A CN201911420812 A CN 201911420812A CN 111191727 A CN111191727 A CN 111191727A
- Authority
- CN
- China
- Prior art keywords
- fault
- gas pressure
- pressure regulator
- data
- neural network
- 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
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application provides a gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ. The method comprises the following steps: acquiring original fault data of a gas pressure regulator with a known fault type; performing dimensionality reduction on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data; and carrying out fault classification on the sample data subjected to dimensionality reduction by using a particle swarm algorithm and a kernel principal component analysis method by using a learning vector quantization neural network. According to the method, the original fault sample data of the gas pressure regulator is preprocessed by using a principal component analysis method, and the new sample data after dimensionality reduction is used as the input of a learning vector quantization neural network, so that the dependence on experience is effectively reduced, the occurrence of uncertainty caused by artificial participation is reduced, the fault early warning and intelligent fault diagnosis of main vulnerable parts of the gas pressure regulator can be realized, and the method plays a great role in stable, safe and reliable operation of gas equipment.
Description
Technical Field
The application relates to the technical field of fault diagnosis of gas pressure regulators, in particular to a fault diagnosis method, a fault diagnosis system, a fault diagnosis terminal and a computer storage medium of a gas pressure regulator based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network.
Background
In the twenty-first century, natural gas is used as a clean and pollution-free energy source and becomes an essential part for the life and development of cities at present under the market condition of the rapid development of the natural gas industry in China. Natural gas is used as a green clean energy source, on one hand, the pollution to the environment can be reduced, and on the other hand, the natural gas is more economical and practical compared with artificial gas. The application of natural gas in the industry can reduce the application of a large amount of fossil fuels such as coal and the like, and can greatly improve the yield and the quality of produced products. Under the condition that society is continuously developing forward, the current situation of tension presented by energy use and ecological problems brought by environmental pollution have attracted high attention of all the world, how to reasonably utilize coal and petroleum resources under the current situation to develop new energy, deal with global greenhouse effect and protect the ecological environment on which we rely to live has long been a major topic which is commonly concerned and needs to be paid high attention by all the countries in the world.
The gas supply system can be called as the life line engineering of the city, and plays a role in economic development and social safety and stability. Along with the continuous development of cities, the scale of a gas supply system of the cities is gradually enlarged, and the problems of stability and safety of the gas supply system are increasingly highlighted in the daily operation process. Through data search and actual on-site investigation, the natural gas pipeline transmission and distribution system generally comprises a gate station, a storage and distribution station, a transmission and distribution pipe network, a pressure regulating chamber, operation management operation, control equipment, a dispatching and maintenance center and the like. In a general town gas supply system, the gas pressure at the user side is always required to be kept at a lower and stable pressure, which requires that medium-high pressure natural gas in a gas pipe network can be transmitted and distributed to each user only by reducing and stabilizing the pressure through a pressure regulator.
The problems faced by the current gas pressure regulator are as follows: in the past, a maintenance technician judges whether a gas pressure regulator is in a normal operation state or not through data collected by pressure meter paper and an automatic instrument according to experience, and generally, the maintenance technician judges whether the gas pressure regulator is in the normal operation state or not through the data collected by the pressure meter paper and the automatic instrument according to the experience, so that the artificial subjectivity is high, the time is very long, and the intelligent degree is not high; in addition, at present, a regular maintenance strategy is adopted for all the gas pressure regulators, namely, the gas pressure regulators are disassembled and maintained no matter whether the gas pressure regulators operate well or not in a maintenance period, so that the maintenance task amount and the maintenance cost are greatly increased, the original well-operated gas pressure regulators are possibly damaged in maintenance, and meanwhile, excessive maintenance also causes a great deal of waste of resources such as manpower and material resources.
With the progress and development of scientific technology, the development of automation technology enriches the functions of modern equipment and complicates the system. Due to the complexity of the system, its equipment problems caused by various factors will also be unavoidable. Regardless of the country and abroad, a plurality of major accidents are caused by the fact that the fault discovery of equipment is not timely. The process from the non-regular maintenance to the non-regular maintenance of the equipment by human beings reflects that people seek safety in the process of carrying out the operation and the maintenance of the equipment, and simultaneously, the requirement on the economy is also made. The scheduled maintenance is easy to generate higher surplus expenditure, and the condition of untimely maintenance and the like can occur when the scheduled maintenance is not performed. At present, most of the domestic and foreign researches on the gas pressure regulator are concentrated on the static characteristic curve of the gas pressure regulator, and the corresponding characteristic curve can be found according to the instructions of manufacturers of the gas pressure regulator. Other scholars also have corresponding researches on the dynamic characteristics and model simulation of the gas pressure regulator, however, researches on online fault early warning and diagnosis of the gas pressure regulator are still basically in the primary germination stage, and a perfect and effective intelligent fault diagnosis system of the gas pressure regulator is not yet available. Under such a multifaceted demand, the necessity of performing intelligent fault diagnosis technical research is easily predicted.
Therefore, in order to better solve the problem of fault diagnosis and troubleshooting of the gas pressure regulator, it is necessary to provide an intelligent fault diagnosis method for the gas pressure regulator.
Disclosure of Invention
Aiming at the defects of the prior art, the application provides a fault diagnosis method, a fault diagnosis system, a fault diagnosis terminal and a computer storage medium for a gas pressure regulator based on a particle swarm, a kernel principal component analysis method and a learning vector quantization neural network, and solves the problems of low intelligence degree, low identification precision and the like in the fault diagnosis process of the gas pressure regulator in the prior art.
In order to solve the technical problem, the application provides a fault diagnosis method for a gas pressure regulator based on a particle swarm, a kernel principal component analysis method and a learning vector quantization neural network, which comprises the following steps:
acquiring original fault data of a gas pressure regulator with a known fault type;
performing dimensionality reduction on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method based on the combination of a particle swarm algorithm (PSO), the Kernel Principal Component Analysis (KPCA) and a learning vector quantization neural network (LVQ), and generating low-dimensional irrelevant sample data;
and carrying out fault classification on the sample data subjected to dimensionality reduction by using a particle swarm algorithm and a kernel principal component analysis method by using a learning vector quantization neural network.
Optionally, the raw fault data of the gas pressure regulator includes:
the original fault data of the pilot, the membrane, the input end pressure and the valve port valve position of the gas pressure regulator.
Optionally, the performing, by using a particle swarm algorithm and a kernel principal component analysis method, dimension reduction processing on the original fault data to generate low-dimensional irrelevant sample data includes:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
Optionally, the classifying the fault of the sample data subjected to the dimensionality reduction processing by using the particle swarm algorithm and the kernel principal component analysis method by using the learning vector quantization neural network includes:
randomly distributing the sample data into a training data set and a testing data set;
establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained gas pressure regulator fault diagnosis model;
and (4) utilizing the learning vector quantization neural network finished by the test sample set test training to analyze the fault diagnosis result.
Optionally, the establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained fault diagnosis model of the gas pressure regulator includes:
initializing a learning vector quantization neural network weight, including determining the number of nodes of an input layer, a hidden layer and an output layer of the learning vector quantization neural network, connecting the weight and a neuron excitation function, and the like;
giving a sample vector, calculating the output of each layer of neuron, and calculating the total error of an output layer;
and judging whether the total error of the output layer of the iteration end condition meets the requirement, if not, recalculating the error and adjusting the weight, returning to the step of executing and calculating the output of each layer of neurons, and sequentially performing the steps downwards until the end of meeting the requirement is finished, thereby finishing the fault matching training of the learning vector quantization neural network.
In a second aspect, the present application further provides a gas pressure regulator fault diagnosis system based on a particle swarm, a kernel principal component analysis method and a learning vector quantization neural network, including:
the data acquisition unit is configured for acquiring original fault data of the gas pressure regulator with known fault type;
the preprocessing unit is configured to perform dimension reduction processing on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data;
and the fault classification unit is configured and used for classifying the faults of the sample data subjected to dimensionality reduction processing by the particle swarm algorithm and the kernel principal component analysis method by utilizing the learning vector quantization neural network.
Optionally, the data acquired by the data acquiring unit includes:
the original fault data of the pilot, the membrane, the input end pressure and the valve port valve position of the gas pressure regulator.
Optionally, the preprocessing unit is specifically configured to:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
Optionally, the fault classification list includes:
the data set setting unit is configured to randomly distribute the sample data into a training data set and a testing data set;
the model generation unit is configured for establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set and generating a trained gas pressure regulator fault diagnosis model;
and the diagnosis testing unit is configured to utilize the test sample set to test the trained learning vector quantization neural network and analyze a fault diagnosis result.
Optionally, the modeling generation unit includes:
the initialization unit is configured for initializing the learning vector quantization neural network weight, and comprises the steps of determining the number of nodes of an input layer, a hidden layer and an output layer of the learning vector quantization neural network, connecting the weight and a neuron excitation function and the like;
the error calculation unit is configured for giving a sample vector, calculating the output of each layer of neuron and calculating the total error of an output layer;
the error judgment unit is configured for judging whether the total error of the iteration ending condition output layer meets the requirement or not;
and the weight value adjusting unit is configured and used for recalculating the error and adjusting the weight value if the requirement is not met, returning to the step of executing the output of calculating each layer of neurons, and sequentially performing each step downwards until the end of meeting the requirement is finally reached, thereby completing the fault matching training of the learning vector quantization neural network.
In a third aspect, the present application provides a terminal, comprising:
a processor, a memory, wherein,
the memory is for storing a computer program.
The processor is configured to call and run the computer program from the memory, so that the terminal performs the method of the terminal.
In a fourth aspect, the present application provides a computer storage medium having instructions stored thereon that, when executed on a computer, cause the computer to perform the method of the above aspects.
The gas pressure regulator fault diagnosis method based on the combination of the Particle Swarm Optimization (PSO), the Kernel Principal Component Analysis (KPCA) and the learning vector quantization neural network (LVQ) is provided. The original gas pressure regulator fault sample data is preprocessed by a particle swarm algorithm and a kernel principal component analysis method, so that the dimension of the original data is reduced, the correlation existing between the original samples is eliminated, the network structure is simplified, and the fault identification precision is effectively improved; and the new sample data after the dimensionality reduction processing is used as the input of the learning vector quantization neural network, so that the dependence on experience is effectively reduced, the occurrence of uncertainty caused by manual participation is reduced, the complex corresponding relation between data characteristics and the operation condition can be effectively reflected, the fault early warning and intelligent fault diagnosis of main vulnerable parts of the pressure regulator can be realized, the current unscientific method for carrying out regular disassembly maintenance and maintenance on the pressure regulator is finally solved, the purposes of saving manpower and material resources and avoiding maintainability damage are achieved, and the method plays a great role in the stable, safe and reliable operation of gas equipment.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a flowchart of a fault diagnosis method for a gas pressure regulator based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a gas pressure regulator provided in an embodiment of the present application;
FIG. 3 is a basic process flow diagram of a gas pressure regulating station according to an embodiment of the present disclosure;
FIG. 4 is a basic flowchart of a kernel principal component analysis method provided in an embodiment of the present application;
FIG. 5 is a basic flowchart of a particle swarm algorithm combined with kernel principal component analysis (PSO-KPCA algorithm) provided in an embodiment of the present application;
fig. 6 is a schematic structural diagram of a learning vector quantization neural network according to an embodiment of the present disclosure;
FIG. 7 is a basic flowchart of a learning vector quantization neural network training provided by an embodiment of the present application;
fig. 8 is a flowchart of a fault diagnosis method for a gas pressure regulator based on a particle swarm optimization, a kernel principal component analysis and a learning vector quantization neural network according to another embodiment of the present application;
fig. 9 is a schematic structural diagram of a gas pressure regulator fault diagnosis system based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network according to an embodiment of the present application;
fig. 10 is a schematic structural diagram of a gas pressure regulator fault diagnosis terminal based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but 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 application.
Referring to fig. 1, fig. 1 is a flowchart of a fault diagnosis method for a gas pressure regulator based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network according to an embodiment of the present application, where the method 100 includes:
s101, acquiring original fault data of a gas pressure regulator with a known fault type;
s102, performing dimensionality reduction on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data;
and S103, carrying out fault classification on the sample data subjected to dimension reduction processing by a particle swarm algorithm and a kernel principal component analysis method by utilizing a learning vector quantization neural network.
Based on the above embodiment, as an optional embodiment, the original fault data of the gas pressure regulator in step S101 includes:
the original fault data of the pilot, the membrane, the input end pressure and the valve port valve position of the gas pressure regulator.
It should be noted that the gas pressure regulator has an automatic regulating function, and can perform the functions of reducing and stabilizing pressure on the gas. When a certain pressure fluctuation range is set for the gas pressure regulator, the gas pressure regulator can automatically regulate the opening degree of the valve port according to the gas pressure change in the pipe network, so that the gas pressure conveyed to a user side is kept in a required stable range. According to the working environment of the gas pressure regulator, the faults of the gas pressure regulator are divided into four types, namely low peak gas pressure, high closing pressure, surge alarm and normal mode.
As shown in fig. 2-3, fig. 2 is a schematic structural diagram of a gas pressure regulator, and fig. 3 is a basic process flow diagram of a gas pressure regulating station. Gas is fed from an inlet of the gas pressure regulating station, the temperature and the pressure of the gas are detected at the gas inlet, and the gas is cut off when the temperature and the pressure at the gas inlet are ultrahigh; then the gas flow is filtered by a filter, and the instantaneous flow and the accumulated flow of the gas are monitored at the outlet of the filter; and then the gas flows through a gas pressure regulator to realize pressure regulation, the pressure of the gas is monitored at the outlet of the pressure regulator, when the pressure of the gas at the outlet of the pressure regulator is ultrahigh, the gas is cut off and protected, and finally the gas is discharged from the outlet of the gas pressure regulating station.
When the gas flows through the gas pressure regulator, the director function of the gas pressure regulator is to monitor and control the pressure, so as to ensure the normal operation of the gas pressure regulator. One of the reasons causing the fault of the gas pressure regulator is a fault part of a director, and the director is used for amplifying a signal of rising or lowering of outlet pressure P2, so that the action of the pressure regulator is accelerated, and the precision and the sensitivity of the gas pressure regulator are improved. When the outlet pressure p2 is lower than the given value, the thin film of the director is lowered to open the director valve, and the gas with the pressure p3 is supplemented to the space below the membrane of the main pressure regulator after throttling. Since p3 is greater than p2, the main regulator valve is opened up, the flow increases and p2 returns to the set point. Conversely, when p2 exceeds a given value, the pilot membrane rises, causing the valve to close. When a fault occurs, the director cannot well coordinate the main pressure regulator to regulate the pressure. At the moment, in order to monitor the fault of the director in time, the user interface sends out a monitoring instruction, the diagnosis subsystem applies for data access to the monitoring subsystem after receiving the monitoring instruction, and the monitoring subsystem sends diagnosis data to the monitoring subsystem after receiving the data access application. The diagnostic system extracts the distribution characteristics of the original data at the bottommost layer of the network architecture by using the kernel principal component analysis method and the learning vector quantization neural network algorithm, recombines the distribution characteristics into more compact high-level distribution characteristics to realize the characteristic extraction of the data, and then completes the matching training of the faults and the identification and classification of the faults by using supervised learning.
In addition, the diaphragm is a sensing element and is used for measuring the difference between the pressure of the downstream working condition and the required pressure, and if a fault occurs, the difference cannot be detected. The valve position is abnormal, so that the opening degree of the valve port cannot be controlled, and the pressure cannot be controlled. And the input end pressure is too low, so that the fuel quantity is directly influenced. The three parts apply the same principle to collect the collected data in real time and carry out fault diagnosis.
Based on the foregoing embodiment, as an optional embodiment, the step S102 performs dimension reduction processing on the original fault data by using a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data, including:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
As shown in fig. 4, fig. 4 is a basic flow chart of the kernel principal component analysis KPCA algorithm. The method comprises the steps of utilizing principal component analysis to conduct dimensionality reduction processing on original fault data of a gas pressure regulator, conducting preprocessing on original fault data samples X to obtain a standard data matrix Z, calculating all variable mean values of the standard data matrix Z, obtaining a covariance matrix R through all variable mean values of the standard data matrix Z, conducting eigenvalue decomposition on the covariance matrix R to obtain an eigenvalue matrix and an eigenvector matrix of the covariance matrix R, calculating principal component cumulative variance contribution rate according to the matrix, determining a principal component number m according to the ranking and the proportion of the principal component cumulative variance contribution rate, and further determining a principal component U.
It should be noted that the kernel principal component analysis method is generally regarded as a linear transformation, and transforms the original data into another new coordinate system, resulting in a first large variance of all data projections falling on a first coordinate, i.e. a so-called first principal component; correspondingly, a second large variance will fall on a second coordinate, which is called a second principal component; the third to nth principal components are recurred in this manner. Kernel principal component analysis is often used to reduce the dimensionality of the data and to ensure that the feature of the dataset that contributes most to variance is preserved by removing higher-order principal components and preserving lower-order principal components, which typically preserve most of the important information of the data. The fundamental relationship between the data obtained after the principal component analysis processing of the raw data and the raw data can be summarized as follows: most of the information of the original data is still retained by the pivot; all the principal elements obtained by transformation can be represented by linear combination of variables of the original data; the dimensionality of the data obtained after processing is far smaller than that before transformation; the obtained main elements have no correlation among the main elements.
The algorithm of the kernel principal component analysis method is adopted to carry out dimensionality reduction on data without depending on an accurate mathematical model, the data in a high-dimensional space can be projected to a low-dimensional space, the dimensionality of the data is greatly reduced, the correlation between variables is reduced, some redundant information is removed, main change information and statistical characteristics are extracted from a new low-dimensional space, the characteristic of original data is understood, and the fact that the data with fewer components fully reflects and represents the whole of the original more data component depiction is achieved. Therefore, the original fault data of the gas regulator is subjected to dimensionality reduction by adopting the kernel principal component analysis method, so that the error caused by redundant information is reduced, the fault identification effect is improved, and the intrinsic structural characteristics in the data can be searched.
As shown in fig. 5, fig. 5 is a basic flowchart of a particle swarm optimization combined with a kernel principal component analysis (PSO-KPCA algorithm). The algorithm process firstly initializes kernel function parameters, PSO algorithm parameters and particle fitness values. Calculating the fitness value of each particle by using the fitness function to obtain an individual extremum and global extremum optimal initialization (namely comparing the fitness value of each particle with the fitness of the individual extremum of each particle, updating the individual extremum if the fitness value of each particle is better, otherwise, keeping the original value, comparing the updated individual extremum of each particle with the global extremum, updating the individual extremum if the fitness value is better, or keeping the original value), continuously performing population iteration to obtain kernel function parameter decoding, calculating the target function through the kernel function decoding until the maximum iteration times are reached to complete iteration, and obtaining the optimal parameter combination of the learning vector quantization neural network model. This is followed by a dimensionality reduction of the original fault data.
The PSO and KPCA combined algorithm is recorded as PSO-KPCA algorithm, the PSO algorithm is used for global optimization and determining the number of neurons in the intermediate layer of the subsequent neural network so as to simplify the network structure; the kernel principal component analysis KPCA is used for reducing the dimension of data, and the combination is to carry out the preprocessing of the structural parameters and the preprocessing of the input parameters in the early stage for the subsequent learning vector quantization neural network LVQ.
Based on the foregoing embodiment, as an optional embodiment, in step S103, performing fault classification on the sample data subjected to dimensionality reduction processing by using a particle swarm algorithm and a kernel principal component analysis method by using a learning vector quantization neural network includes:
randomly distributing the sample data into a training data set and a testing data set;
establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained gas pressure regulator fault diagnosis model;
and (4) utilizing the learning vector quantization neural network finished by the test sample set test training to analyze the fault diagnosis result.
Specifically, the dimensionality of the data sample is reduced after the data sample is analyzed by a kernel principal component analysis method, and the data sample after dimensionality reduction is randomly divided into a training data set and a testing data set; constructing a learning vector quantization neural network model, determining output variables and hidden layer numbers of the learning vector quantization neural network, training the learning vector quantization neural network by using a training data set subjected to dimensionality reduction by a kernel principal component analysis method as input and a fault type as output, and generating a trained gas pressure regulator fault diagnosis model; and inputting the test data set into a trained gas pressure regulator fault diagnosis model, performing fault classification test, and outputting the fault type of the gas pressure regulator.
Based on the foregoing embodiment, as an optional embodiment, the establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained fault diagnosis model of the gas pressure regulator includes:
initializing a learning vector quantization neural network weight, including determining the number of nodes of an input layer, a hidden layer and an output layer of the learning vector quantization neural network, connecting the weight and a neuron excitation function, and the like;
giving a sample vector, calculating the output of each layer of neuron, and calculating the total error of an output layer;
and judging whether the total error of the output layer of the iteration end condition meets the requirement, if not, recalculating the error and adjusting the weight, returning to the step of executing and calculating the output of each layer of neurons, and sequentially performing the steps downwards until the end of meeting the requirement is finished, thereby finishing the fault matching training of the learning vector quantization neural network.
As shown in fig. 6-7, fig. 6 is a schematic structural diagram of a learning vector quantization neural network, and fig. 7 is a basic flow chart of learning vector quantization neural network training. The number of input layer nodes of the learning vector quantization neural network can be determined according to the input sample, the dimension of the feature vector of the input sample is equal to the number of the input layer nodes, and the number (4) of the fault classification modes of the sample is the number of the output nodes of the learning vector quantization neural network. The number of hidden layers of the learning vector quantization neural network needs to be calculated and determined according to the training and optimization of the network: firstly, the number of nodes of the middle layer is taken as 0, training is carried out after a sample vector is given, the output of each layer of neuron and the total error of the output layer are calculated, whether the total error of the output layer meets the requirement or not is judged, meanwhile, the number of the nodes of the middle layer is automatically increased, the input error is repeatedly checked, and the criterion for stopping training is that the required error requirement is met or the number of the nodes of the middle layer is up-limited.
It should be noted that the fault diagnosis method based on the learning vector quantization neural network (LVQ) has a simple network structure, and the LVQ is a supervised training method, but adopts an unsupervised data clustering technique to process a data set, thereby allowing a clustering center to be obtained. The LVQ network consists of an input layer, a contention layer and an output layer. When fault data is input into the LVQ, the neurons of the competition layer follow the principle of 'winner is king' to generate winning neurons, the output is 1, and the rest of the neurons failing in competition are 0. The output neurons with 1 are classified into the same type as the input neurons, and different output neurons represent different types, so that the purpose of classification and identification of output samples is achieved.
As shown in fig. 8, fig. 8 is a flowchart of a gas pressure regulator fault diagnosis method based on a particle swarm optimization, a kernel principal component analysis method and a learning vector quantization neural network according to another embodiment of the present application. The method comprises the following specific steps:
the method comprises the steps that a fault diagnosis system of the gas pressure regulator obtains original fault data of the gas pressure regulator;
carrying out normalization processing on original fault data; in order to reduce errors and avoid smaller values being "eaten" by larger values;
performing dimension reduction processing on the original fault data after the normalization processing by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data;
randomly distributing the sample data into a training data set and a testing data set;
establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained gas pressure regulator fault diagnosis model;
and utilizing the test sample set to perform diagnosis test on the trained learning vector quantization neural network, and analyzing a fault diagnosis result.
Referring to fig. 9, fig. 9 is a schematic structural diagram of a gas pressure regulator fault diagnosis system based on a particle swarm, a kernel principal component analysis method and a learning vector quantization neural network according to an embodiment of the present application, where the system 900 includes:
a data obtaining unit 910, configured to obtain original fault data of a gas pressure regulator with a known fault type;
the preprocessing unit 920 is configured to perform dimension reduction processing on the original fault data by using a particle swarm algorithm and a kernel principal component analysis method, and generate low-dimensional irrelevant sample data;
and a fault classification unit 930 configured to classify the fault of the sample data subjected to the dimensionality reduction processing by the particle swarm algorithm and the kernel principal component analysis method by using the learning vector quantization neural network.
Based on the above embodiment, as an optional embodiment, the data acquired by the data acquiring unit 910 includes:
the original fault data of the pilot, the membrane, the input end pressure and the valve port valve position of the gas pressure regulator.
Based on the foregoing embodiment, as an optional embodiment, the preprocessing unit 920 is specifically configured to:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
Based on the foregoing embodiment, as an optional embodiment, the fault classification unit 930 specifically includes:
the data set setting unit is configured to randomly distribute the sample data into a training data set and a testing data set;
the model generation unit is configured for establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set and generating a trained gas pressure regulator fault diagnosis model;
and the diagnosis testing unit is configured to utilize the test sample set to test the trained learning vector quantization neural network and analyze a fault diagnosis result.
Based on the foregoing embodiment, as an optional embodiment, the model generating unit includes:
the initialization unit is configured for initializing the learning vector quantization neural network weight, and comprises the steps of determining the number of nodes of an input layer, a hidden layer and an output layer of the learning vector quantization neural network, connecting the weight and a neuron excitation function and the like;
the error calculation unit is configured for giving a sample vector, calculating the output of each layer of neuron and calculating the total error of an output layer;
the error judgment unit is configured for judging whether the total error of the iteration ending condition output layer meets the requirement or not;
and the weight value adjusting unit is configured and used for recalculating the error and adjusting the weight value if the requirement is not met, returning to the step of executing the output of calculating each layer of neurons, and sequentially performing each step downwards until the end of meeting the requirement is finally reached, thereby completing the fault matching training of the learning vector quantization neural network.
Referring to fig. 10, fig. 10 is a schematic structural diagram of a terminal 1000 according to an embodiment of the present disclosure, where the terminal 1000 may be used to execute the intelligent fault diagnosis method for a gas pressure regulator according to the embodiment of the present disclosure.
The terminal system 1000 may include: a processor 1001, a memory 1002, and a communication unit 1003. The components communicate over one or more buses, and those skilled in the art will appreciate that the architecture of the servers shown in the figures is not intended to be limiting, and may be a bus architecture, a star architecture, a combination of more or fewer components than those shown, or a different arrangement of components.
The memory 1002 may be used for storing instructions executed by the processor 1001, and the memory 1002 may be implemented by any type of volatile or non-volatile storage terminal or combination thereof, such as a Static Random Access Memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk. The executable instructions in memory 1002, when executed by processor 1001, enable terminal 1000 to perform some or all of the steps in the method embodiments described below.
The processor 1001 is a control center of the storage terminal, connects various parts of the entire electronic terminal using various interfaces and lines, and performs various functions of the electronic terminal and/or processes data by operating or executing software programs and/or modules stored in the memory 1002 and calling data stored in the memory. The processor may be composed of an Integrated Circuit (IC), for example, a single packaged IC, or may be composed of multiple packaged ICs with the same or different functions. For example, the processor 1001 may include only a Central Processing Unit (CPU). In the embodiment of the present invention, the CPU may be a single operation core, or may include multiple operation cores.
A communication unit 1003 for establishing a communication channel so that the storage terminal can communicate with other terminals. And receiving user data sent by other terminals or sending the user data to other terminals.
The present application also provides a computer storage medium, wherein the computer storage medium may store a program, and the program may include some or all of the steps in the embodiments provided by the present invention when executed. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a Random Access Memory (RAM).
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system provided by the embodiment, the description is relatively simple because the system corresponds to the method provided by the embodiment, and the relevant points can be obtained by referring to the description of the method part.
The principles and embodiments of the present application are explained herein using specific examples, which are provided only to help understand the method and the core idea of the present application. It should be noted that, for those skilled in the art, it is possible to make several improvements and modifications to the present application without departing from the principle of the present application, and such improvements and modifications also fall within the scope of the claims of the present application.
It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Claims (10)
1. A gas pressure regulator fault diagnosis method based on particle swarm, kernel principal component analysis and learning vector quantization neural network is characterized by comprising the following steps:
acquiring original fault data of a gas pressure regulator with a known fault type;
performing dimensionality reduction on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data;
and carrying out fault classification on the sample data subjected to dimensionality reduction by using a particle swarm algorithm and a kernel principal component analysis method by using a learning vector quantization neural network.
2. The intelligent fault diagnosis method for a gas pressure regulator according to claim 1, wherein the raw fault data of the gas pressure regulator comprises:
the original fault data of the pilot, the membrane, the input end pressure and the valve port valve position of the gas pressure regulator.
3. The intelligent fault diagnosis method for the gas pressure regulator according to claim 1, wherein the dimension reduction processing is performed on the original fault data to generate low-dimensional irrelevant sample data, and the method comprises the following steps:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
4. The intelligent fault diagnosis method for the gas pressure regulator according to claim 1, wherein the fault classification of the sample data subjected to the dimensionality reduction processing by the particle swarm algorithm and the kernel principal component analysis method by using the learning vector quantization neural network comprises the following steps:
randomly distributing the sample data into a training data set and a testing data set;
establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set, and generating a trained gas pressure regulator fault diagnosis model;
and (4) utilizing the learning vector quantization neural network finished by the test sample set test training to analyze the fault diagnosis result.
5. The intelligent fault diagnosis method for the gas pressure regulator according to claim 4, wherein the establishing of the learning vector quantization neural network, the training of the learning vector quantization neural network by using the training data set, and the generation of the trained fault diagnosis model for the gas pressure regulator comprises:
initializing a learning vector quantization neural network weight, including determining the number of nodes of an input layer, a hidden layer and an output layer of the learning vector quantization neural network, connecting the weight and a neuron excitation function, and the like;
giving a sample vector, calculating the output of each layer of neuron, and calculating the total error of an output layer;
and judging whether the total error of the output layer of the iteration end condition meets the requirement, if not, recalculating the error and adjusting the weight, returning to the step of executing and calculating the output of each layer of neurons, and sequentially performing the steps downwards until the end of meeting the requirement is finished, thereby finishing the fault matching training of the learning vector quantization neural network.
6. A gas pressure regulator fault diagnosis system based on particle swarm, kernel principal component analysis and learning vector quantization neural network is characterized by comprising:
the data acquisition unit is configured for acquiring original fault data of the gas pressure regulator with known fault type;
the preprocessing unit is configured to perform dimension reduction processing on the original fault data by adopting a particle swarm algorithm and a kernel principal component analysis method to generate low-dimensional irrelevant sample data;
and the fault classification unit is configured and used for classifying the faults of the sample data subjected to dimensionality reduction processing by the particle swarm algorithm and the kernel principal component analysis method by utilizing the learning vector quantization neural network.
7. The gas pressure regulator intelligent fault diagnosis system of claim 6, wherein the preprocessing unit is specifically configured to:
preprocessing the acquired original fault data to obtain a standard data matrix;
calculating the mean value and covariance matrix of each variable of the standard data matrix;
calculating an eigenvalue matrix and an eigenvector matrix of the covariance matrix;
and calculating the accumulated variance contribution rate of the principal elements by using the eigenvalue matrix and the eigenvector matrix, determining the number of the principal elements and further determining the principal components.
8. The gas pressure regulator intelligent fault diagnosis system of claim 6, wherein the fault classification list comprises:
the data set setting unit is configured to randomly distribute the sample data into a training data set and a testing data set;
the model generation unit is configured for establishing a learning vector quantization neural network, training the learning vector quantization neural network by using the training data set and generating a trained gas pressure regulator fault diagnosis model;
and the diagnosis testing unit is configured to utilize the test sample set to test the trained learning vector quantization neural network and analyze a fault diagnosis result.
9. A terminal, comprising:
a processor;
a memory for storing instructions for execution by the processor;
wherein the processor is configured to perform the method of any one of claims 1-5.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911420812.0A CN111191727A (en) | 2019-12-31 | 2019-12-31 | Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911420812.0A CN111191727A (en) | 2019-12-31 | 2019-12-31 | Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111191727A true CN111191727A (en) | 2020-05-22 |
Family
ID=70710586
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911420812.0A Pending CN111191727A (en) | 2019-12-31 | 2019-12-31 | Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111191727A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111695465A (en) * | 2020-06-01 | 2020-09-22 | 杭州英集动力科技有限公司 | Pipe network fault diagnosis and positioning method and system based on pressure wave mode identification |
| CN111929591A (en) * | 2020-08-21 | 2020-11-13 | 彩虹无线（北京）新技术有限公司 | Fault battery detection method, device, equipment and computer storage medium |
| CN112381141A (en) * | 2020-11-13 | 2021-02-19 | 西安建筑科技大学 | Air conditioner sensor fault detection method and system |
| CN112819107A (en) * | 2021-04-16 | 2021-05-18 | 四川九门科技股份有限公司 | Artificial intelligence-based fault prediction method for gas pressure regulating equipment |
| CN113743593A (en) * | 2021-09-27 | 2021-12-03 | 上海齐感电子信息科技有限公司 | Neural network quantization method, system, storage medium and terminal |
| CN113923104A (en) * | 2021-12-07 | 2022-01-11 | 南京信息工程大学 | Network fault diagnosis method, equipment and storage medium based on wavelet neural network |
| CN117668528A (en) * | 2024-02-01 | 2024-03-08 | 成都华泰数智科技有限公司 | Natural gas voltage regulator fault detection method and system based on Internet of things |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600140A (en) * | 2016-12-07 | 2017-04-26 | 河海大学常州校区 | Improved support vector machine-based gas pipeline fault prediction and early warning system and method |
| CN108921230A (en) * | 2018-07-25 | 2018-11-30 | 浙江浙能嘉华发电有限公司 | Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network |
| CN109344517A (en) * | 2018-10-14 | 2019-02-15 | 毛述春 | A kind of high-voltage isulation method for diagnosing faults of new-energy automobile |
-
2019
- 2019-12-31 CN CN201911420812.0A patent/CN111191727A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600140A (en) * | 2016-12-07 | 2017-04-26 | 河海大学常州校区 | Improved support vector machine-based gas pipeline fault prediction and early warning system and method |
| CN108921230A (en) * | 2018-07-25 | 2018-11-30 | 浙江浙能嘉华发电有限公司 | Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network |
| CN109344517A (en) * | 2018-10-14 | 2019-02-15 | 毛述春 | A kind of high-voltage isulation method for diagnosing faults of new-energy automobile |
Non-Patent Citations (3)
| Title |
|---|
| 宋乐见;张晓龙;陈同兴;薛宇涛;: "基于KPCA与PSO-WLSSVM的顶吹熔炼系统喷枪寿命预测研究", 计算机与应用化学 * |
| 杨东民;陈敏;吴庆朝;: "融合KPCA与PSO-RBF的数控机床故障诊断研究", 机械设计与制造 * |
| 郭小康 等: "基于 PSO-KPCA-LVQ 神经网络的雷达一维距离像识别", 电光与控制 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111695465A (en) * | 2020-06-01 | 2020-09-22 | 杭州英集动力科技有限公司 | Pipe network fault diagnosis and positioning method and system based on pressure wave mode identification |
| CN111695465B (en) * | 2020-06-01 | 2024-03-05 | 浙江英集动力科技有限公司 | Pipe network fault diagnosis and positioning method and system based on pressure wave mode identification |
| CN111929591A (en) * | 2020-08-21 | 2020-11-13 | 彩虹无线（北京）新技术有限公司 | Fault battery detection method, device, equipment and computer storage medium |
| CN112381141A (en) * | 2020-11-13 | 2021-02-19 | 西安建筑科技大学 | Air conditioner sensor fault detection method and system |
| CN112819107A (en) * | 2021-04-16 | 2021-05-18 | 四川九门科技股份有限公司 | Artificial intelligence-based fault prediction method for gas pressure regulating equipment |
| CN112819107B (en) * | 2021-04-16 | 2021-07-02 | 四川九门科技股份有限公司 | Artificial intelligence-based fault prediction method for gas pressure regulating equipment |
| CN113743593A (en) * | 2021-09-27 | 2021-12-03 | 上海齐感电子信息科技有限公司 | Neural network quantization method, system, storage medium and terminal |
| CN113743593B (en) * | 2021-09-27 | 2023-08-22 | 上海齐感电子信息科技有限公司 | Neural network quantization method, system, storage medium and terminal |
| CN113923104A (en) * | 2021-12-07 | 2022-01-11 | 南京信息工程大学 | Network fault diagnosis method, equipment and storage medium based on wavelet neural network |
| CN113923104B (en) * | 2021-12-07 | 2022-03-11 | 南京信息工程大学 | Network fault diagnosis method, equipment and storage medium based on wavelet neural network |
| CN117668528A (en) * | 2024-02-01 | 2024-03-08 | 成都华泰数智科技有限公司 | Natural gas voltage regulator fault detection method and system based on Internet of things |
| CN117668528B (en) * | 2024-02-01 | 2024-04-12 | 成都华泰数智科技有限公司 | Natural gas voltage regulator fault detection method and system based on Internet of things |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111191727A (en) | Gas pressure regulator fault diagnosis method, system, terminal and computer storage medium based on PSO-KPCA-LVQ | |
| CN111191725A (en) | RBF fault diagnosis method, system, terminal and computer storage medium based on PCA data processing | |
| CN105117602B (en) | A kind of metering device running status method for early warning | |
| CN110097297A (en) | A kind of various dimensions stealing situation Intellisense method, system, equipment and medium | |
| CN102183621B (en) | Aquaculture dissolved oxygen concentration online forecasting method and system | |
| CN107949812A (en) | For detecting the abnormal combined method in water distribution system | |
| US11840998B2 (en) | Hydraulic turbine cavitation acoustic signal identification method based on big data machine learning | |
| CN112819107B (en) | Artificial intelligence-based fault prediction method for gas pressure regulating equipment | |
| CN116757534A (en) | Intelligent refrigerator reliability analysis method based on neural training network | |
| CN115270965A (en) | Power distribution network line fault prediction method and device | |
| CN111291783A (en) | Intelligent fault diagnosis method, system, terminal and storage medium for gas pressure regulating equipment | |
| CN108921230A (en) | Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network | |
| CN110879377B (en) | Metering device fault tracing method based on deep belief network | |
| CN113221441A (en) | Method and device for health assessment of power plant equipment | |
| CN114969990A (en) | Multi-model fused avionic product health assessment method | |
| CN111697560B (en) | Method and system for predicting load of power system based on LSTM | |
| CN113537645B (en) | Soybean yield prediction method based on machine learning fusion satellite and weather data | |
| CN115688581A (en) | Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium | |
| CN114999527A (en) | Transformer anomaly detection model training and deploying method and device | |
| CN114116832A (en) | Power distribution network abnormity identification method based on data driving | |
| CN114358116A (en) | Oil-immersed transformer fault diagnosis method and system and readable storage medium | |
| CN115858606A (en) | Method, device and equipment for detecting abnormity of time series data and storage medium | |
| CN117674119A (en) | Power grid operation risk assessment method, device, computer equipment and storage medium | |
| CN116776209A (en) | Method, system, equipment and medium for identifying operation state of gateway metering device | |
| CN111476400A (en) | Circuit fault prediction method, device, equipment and computer readable medium |
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 |