CN111880117B - Fault diagnosis method and device for energy-fed power supply device and storage medium - Google Patents
Fault diagnosis method and device for energy-fed power supply device and storage medium Download PDFInfo
- Publication number
- CN111880117B CN111880117B CN202010738590.3A CN202010738590A CN111880117B CN 111880117 B CN111880117 B CN 111880117B CN 202010738590 A CN202010738590 A CN 202010738590A CN 111880117 B CN111880117 B CN 111880117B
- Authority
- CN
- China
- Prior art keywords
- cuckoo
- neural network
- network model
- individual
- fault
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/54—Testing for continuity
Abstract
The application provides an energy feedback device open-circuit fault diagnosis method, an energy feedback device open-circuit fault diagnosis device and a storage medium. The method comprises the following steps: acquiring a three-phase current waveform of a Pulse Width Modulation (PWM) rectifier of the energy-fed power supply device; determining fault characteristic quantities of the three-phase current waveforms; and inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier. The method can diagnose the open circuit fault of the switching element, effectively shorten the time of the energy-feedback power supply device in fault operation, and reduce the operation risk of the energy-feedback power supply device.
Description
Technical Field
The present disclosure relates to a fault diagnosis technique for an open circuit of a rectifier switching device, and more particularly, to a fault diagnosis method and apparatus for a power supply apparatus, and a storage medium.
Background
Urban rail transit is a preferred scheme for solving the increasingly serious urban congestion problem due to the advantages of safety, comfort, large passenger capacity, high running speed, energy conservation, environmental protection and the like. The direct current side voltage of the energy feedback power supply device is adjustable, the feedback reutilization of the regenerative braking energy of the train can be realized, the energy-saving effect is better, and meanwhile, the traction energy can be provided for the train, and the direct current power supply quality is provided. In addition, the energy-feeding power supply device also has the functions of reactive compensation, direct current voltage stabilization, harmonic compensation and the like, so that the energy-feeding power supply device can be used as a power supply device in an urban rail transit system. The advantage of being able to feed power is that it has a much better performance because it has more precise switching devices than conventional diode rectifiers. However, these precision switching devices are prone to failure during operation. Statistically, 38% of converter faults are caused by open and short circuit faults of the power switching devices. Because short-circuit faults of the switching devices are seriously damaged in a short time and the current rise is obvious, the current can be effectively diagnosed and the faults can be removed at the first time when the faults occur; and the open-circuit fault of the switching device is less harmful in a short time, and the fault current change is not obvious, so that the effective monitoring is difficult at present.
If the switching device is in an open-circuit fault state for a long time, not only the service life of the equipment is reduced, but also secondary faults can be caused, and more serious results are caused. Therefore, how to diagnose the open-circuit fault of the switching device and reduce the operation risk of the feedable power supply device is an urgent problem to be solved.
Disclosure of Invention
The application provides a fault diagnosis method and device for an energy-fed power supply device and a storage medium, which are used for diagnosing an open-circuit fault of a switching device and reducing the operation risk of the energy-fed power supply device.
In one aspect, the present application provides a method for diagnosing a fault of a feedable power supply apparatus, including:
acquiring a three-phase current waveform of a Pulse Width Modulation (PWM) rectifier of the energy-fed power supply device;
determining fault characteristic quantity of the three-phase current waveform, wherein the fault characteristic quantity is norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method;
and inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
In one embodiment, before inputting the fault feature quantity into the target neural network model and obtaining the real-time fault diagnosis result of the PWM rectifier, the method further includes:
determining structural data of an initial neural network model, wherein the structural data comprises the number of input layer nodes, the number of hidden layer nodes and the number of output layer nodes;
creating an initial neural network model according to the structural data of the initial neural network model;
and training the initial neural network model, and taking the trained initial neural network model as the target neural network model.
In one embodiment, the training the initial neural network model includes:
acquiring a cuckoo algorithm;
according to the structural data of the initial neural network model, encoding the parameters of the initial neural network model by using a cuckoo algorithm to obtain encoded data of the initial neural network;
and updating the initial neural network model according to the coded data to obtain the target neural network model.
In one embodiment, the encoding the parameters of the initial neural network model by using a cuckoo algorithm according to the structural data of the initial neural network model includes:
acquiring a weight to be encoded and a threshold to be encoded, which correspond to the initial neural network model, of each cuckoo individual of a cuckoo population in the cuckoo algorithm to obtain data to be encoded, which correspond to each cuckoo individual, wherein the data to be encoded comprises the weight to be encoded and the threshold to be encoded;
determining each fitness value corresponding to each cuckoo individual in the cuckoo population to obtain a fitness value data set, wherein the number of the fitness values in the fitness value data set is equal to the preset number;
obtaining target cuckoo individuals according to the average value of the fitness values in the fitness value data group;
and determining a weight value to be coded and a threshold value to be coded, which correspond to the initial neural network model, of the target cuckoo individual to obtain the coded data.
In one embodiment, the obtaining the target cuckoo individual according to the average value of the fitness values in the fitness value data set includes:
and if the average value of the fitness values in the fitness value data group is greater than or equal to a preset average value, acquiring the maximum fitness value in the fitness value data group, and determining the cuckoo individuals corresponding to the maximum fitness value to obtain target cuckoo individuals.
In one embodiment, the method further comprises:
and if the average value of the fitness values is smaller than the preset average value, iteratively updating the cuckoo population until the average value of the fitness values corresponding to the updated cuckoo population is larger than or equal to the preset average value or the iterative updating times reach preset updating times.
In one embodiment, the iteratively updating the cuckoo population until the updated fitness value average corresponding to the cuckoo population is greater than or equal to the preset average or the number of iterative updates reaches a preset update number includes:
A. obtaining a first cuckoo individual and a second cuckoo individual in the cuckoo population;
B. respectively obtaining the fitness value corresponding to the first cuckoo individual and the fitness value corresponding to the second cuckoo individual to obtain a first fitness value and a second fitness value;
C. if the second fitness value is larger than the first fitness value, replacing the first cuckoo individual with the second cuckoo individual;
D. rejecting nxpaThe individual with the worst fitness is selected,randomly generating n individuals to be supplemented into the cuckoo population to form the updated cuckoo population; wherein n is the number of nodes of the input layer, paThe value range is 0-1.
And A-D is executed in an iterative mode until the average value of the fitness values corresponding to the updated cuckoo population is larger than or equal to a preset average value or the iterative updating times reach preset updating times.
In one embodiment, the determining the fault characteristic quantity of the three-phase current waveform includes:
and determining the fault characteristic quantity of the three-phase current waveform based on a wavelet packet algorithm.
In one embodiment, the method further comprises:
obtaining a three-phase current waveform corresponding to a known fault diagnosis result to obtain a test waveform;
inputting the test waveform into the target neural network model to obtain a fault diagnosis result;
comparing the fault diagnosis result with the known fault diagnosis result, and determining the diagnosis accuracy of the target neural network model;
if the diagnosis accuracy rate is greater than or equal to a preset accuracy rate, acquiring and storing a neural network structure and model parameters of the target neural network model to obtain a diagnosis neural network model, wherein the diagnosis neural network model is not updated any more;
and inputting the fault characteristic quantity into the diagnosis neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
On the other hand, this application still provides a device is presented to ability, and open circuit fault diagnosis device includes:
the acquisition module is used for modulating the three-phase current waveform of the PWM rectifier of the energy-fed power supply device;
the processing module is used for determining fault characteristic quantity of the three-phase current waveform, and the fault characteristic quantity is norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method;
and the processing module is also used for inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
In another aspect, the present application further provides a computer device, including a memory, a processor and a transceiver, where the memory is used for storing instructions, the transceiver is used for communicating with other devices, and the processor is used for executing the instructions stored in the memory, so that the computer device executes the open-circuit fault diagnosis method of the energy feeding apparatus according to the above embodiment.
In another aspect, the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the instructions are executed, the instructions cause a computer to execute the method for diagnosing the open circuit fault of the energy feeding device according to the above embodiment.
The application provides a fault diagnosis method of an energy-feedback power supply device, which comprises the steps of obtaining a three-phase current waveform of a PWM rectifier of the energy-feedback power supply device; determining fault characteristic quantities of the three-phase current waveforms; and inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier. The fault diagnosis method of the energy-feedback power supply device can realize online monitoring and real-time diagnosis of the faults of the PWM rectifier, can give out warning in the first time of the faults and help workers to take corresponding measures to solve the faults of the PWM rectifier in time. The fault diagnosis method for the energy-feed power supply device can effectively shorten the time of the energy-feed power supply device in fault operation, thereby reducing the operation risk of the energy-feed power supply device and prolonging the service life of the energy-feed power supply device.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
Fig. 1 is a schematic application scenario diagram of a fault diagnosis method for an energy feed power supply device according to an embodiment of the present application.
Fig. 2 is a schematic flowchart of a fault diagnosis method of an energy feeding apparatus according to an embodiment of the present application.
Fig. 3 is a schematic flowchart of a fault diagnosis method of an energy feeding apparatus according to another embodiment of the present application.
Fig. 4 is a schematic flowchart of a fault diagnosis method for an energy feeding apparatus according to another embodiment of the present application.
Fig. 5 is a schematic flowchart of a fault diagnosis method of an energy feeding apparatus according to another embodiment of the present application.
Fig. 6 is a schematic flowchart of a fault diagnosis method for an energy feeding apparatus according to another embodiment of the present application.
Fig. 7 is a schematic flowchart of a fault diagnosis method of an energy feeding apparatus according to another embodiment of the present application.
Fig. 8 is a schematic diagram of a fault diagnosis device capable of feeding a power supply device according to an embodiment of the present application.
FIG. 9 is a schematic diagram of a computer device provided by an embodiment of the present application.
With the foregoing drawings in mind, certain embodiments of the disclosure have been shown and described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
The terms referred to in this application are explained first:
pulse Width Modulation (PWM) rectifier: the dominant technology employed for medium capacity unity power factor generally requires the use of self-turn-off devices.
Urban rail transit is a preferred scheme for solving the increasingly serious urban congestion problem due to the advantages of safety, comfort, large passenger capacity, high running speed, energy conservation, environmental protection and the like. The direct current side voltage of the energy feedback power supply device is adjustable, the feedback reutilization of the regenerative braking energy of the train can be realized, the energy-saving effect is better, and meanwhile, the traction energy can be provided for the train, and the direct current power supply quality is provided. In addition, the energy-feeding power supply device also has the functions of reactive compensation, direct current voltage stabilization, harmonic compensation and the like, so that the energy-feeding power supply device can be used as a power supply device in an urban rail transit system. The advantage of being able to feed power is that it has a much better performance because it has more precise switching devices than conventional diode rectifiers. However, these precision switching devices are prone to failure during operation. Statistically, 38% of converter faults are caused by open and short circuit faults of the power switching devices. Because short-circuit faults of the switching devices are seriously damaged in a short time and the current rise is obvious, the current can be effectively diagnosed and the faults can be removed at the first time when the faults occur; and the open-circuit fault of the switching device is less harmful in a short time, and the fault current change is not obvious, so that the effective monitoring is difficult at present.
If the switching device is in an open-circuit fault state for a long time, not only the service life of the equipment is reduced, but also secondary faults can be caused, and more serious results are caused. Therefore, how to diagnose the open-circuit fault of the switching device and reduce the operation risk of the feedable power supply device is an urgent problem to be solved.
The present application provides a method, an apparatus, and a storage medium for fault diagnosis of an energy feeding apparatus, which are intended to solve the above technical problems in the prior art.
The fault diagnosis method of the energy-feeding power supply device is applied to computer equipment such as a computer, a server, a tablet computer, a mobile phone and the like. The fault diagnosis method of the energy-feeding power supply device is applied to a computer. Fig. 1 is a schematic diagram of an application of the fault diagnosis method of the energy feeding device provided by the present application, wherein the computer device receives three-phase current waveforms from a Pulse Width Modulation (PWM) rectifier of the energy feeding device.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Referring to fig. 2, the present application provides a method for diagnosing a fault of a power feeding apparatus, including:
s201, obtaining a three-phase current waveform of a PWM rectifier of the energy-feeding power supply device.
The three-phase current waveform of the PWM rectifier can also be understood as a three-phase current waveform on the rectified ac side of the PWM rectifier. The three-phase current waveforms are collected in real time during operation of the PWM rectifier.
S202, determining fault characteristic quantity of the three-phase current waveform, wherein the fault characteristic quantity is norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method;
the fault characteristic quantity refers to norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method, and for example, the fault characteristic quantity refers to characteristic quantity capable of representing fault types of the three-phase current waveform after the three-phase current waveform is decomposed by a wavelet packet and obtained by extracting the wavelet packet norm entropy. In one embodiment, the fault characteristic quantity of the three-phase current waveform can be determined based on a wavelet packet algorithm, and the corresponding fault characteristic quantity is extracted from the three-phase current waveform by adopting a wavelet packet decomposition method, so that the real-time performance and the rapidity of the operation are ensured.
As shown in fig. 1, the fault diagnosis method of the energy-feeding power supply device provided by the present application is directed to diagnosing five operation conditions of a main loop of a PWM rectifier, where the five operation conditions include a single-tube open-circuit fault, an open-circuit fault occurring simultaneously in two Insulated Gate Bipolar Transistors (IGBTs) on a cross half-bridge, an open-circuit fault occurring simultaneously in two IGBTs on a single-phase bridge arm, an open-circuit fault occurring simultaneously in two IGBTs of the same half-bridge, and a normal operation.
And S203, inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
And inputting the fault characteristic quantity into a target neural network, and outputting a real-time fault diagnosis result of the PWM rectifier main loop switching device by the target neural network. The diagnosis result may include an operation state of the PWM rectifier, such as rectification, a specific fault, such as a G5G6 fault, and a fault code. In one embodiment, the computer device can display the operation state, specific faults and fault codes of the PWM rectifier, and can also display specific faults on the PWM rectifier, such as fault positions, fault degrees and the like on the basis of a graph for displaying the PWM rectifier. The computer equipment can also display the three-phase current waveform of the PWM rectifier acquired in real time and display the characteristic value of the three-phase current waveform, namely the fault characteristic quantity of the three-phase current waveform. In one embodiment, the target neural network model is loaded into a monitoring program of the energy-feeding power supply device through programming of an experimental Virtual Instrument Engineering platform (Labview), so that online fault diagnosis of the PWM rectifier is realized.
In the embodiment, after the three-phase current waveforms of the PWM rectifier of the energy-feedback power supply device are obtained, the fault characteristic quantity of the waveforms is determined, and the fault characteristic quantity is input into the target neural network model, so that the real-time fault diagnosis result of the PWM rectifier can be obtained, the online monitoring and the real-time diagnosis of the faults of the PWM rectifier can be realized, the warning can be sent out at the first time of the occurrence of the faults, and the working personnel can be helped to take corresponding measures to solve the faults of the PWM rectifier in time. The fault diagnosis method for the energy-feed power supply device can effectively shorten the time of the energy-feed power supply device in fault operation, thereby reducing the operation risk of the energy-feed power supply device and prolonging the service life of the energy-feed power supply device.
Referring to fig. 3, in an embodiment of the present application, before step S203, the method further includes:
s301, determining structural data of the initial neural network model, wherein the structural data comprises the number of input layer nodes, the number of hidden layer nodes and the number of output layer nodes.
The initial neural network model comprises an input layer, a hidden layer and an output layer, wherein the number of nodes of the input layer is equal to the number of fault characteristic values, and as shown in fig. 1, three-phase current waveforms are decomposed by three-layer wavelet packets and then subjected to norm entropy calculation to obtain 24 fault characteristic values which are 3 multiplied by 8. The number of the output layer nodes is determined according to the codes of the fault types corresponding to the fault characteristic quantities, in one embodiment, the fault types are coded by using binary systems, for example, 22 fault types need 5-bit binary numbers to represent, and in addition, discrimination bits representing rectification or inversion states need 6-bit binary numbers in total, so that the number of the output layer nodes is 6. The number of the hidden layer nodes is usually obtained by trial and error based on an empirical formula, and the common empirical formula is as follows:
or m is log2n; or
Or m is 2n + 1; wherein, l represents the number of the output layer nodes; m represents the number of the nodes of the input layer; a is a constant and has a value range of 1-10.
S302, according to the structural data of the initial neural network model, the initial neural network model is created.
The structural data includes the number of input layer nodes, the number of hidden layer nodes, and the number of output layer nodes of the initial neural network model. It should be noted that the initial neural network model belongs to a multi-layer feedforward (BP) neural network trained according to an error back propagation algorithm. The BP neural network belongs to supervised learning, and the learning process mainly comprises forward transmission of signals and backward transmission of errors. The forward transmission of signals refers to that input signals serving as training data are transmitted to a hidden layer from an input layer and then transmitted to an output layer after being processed by the hidden layer. And the reverse transmission of the error means that if the output layer fails to obtain the expected output, the error between the actual output and the expected output is returned to the input layer through the hidden layer in a certain form, and the error is dispersed in all units among the layers, so that the reference error of each unit is obtained and used as the basis for modifying the weight among the neurons. And the forward transmission and the reverse transmission are performed once, and the system cannot stop until the output error is smaller than a set value or the maximum training times is reached.
S303, training the initial neural network model, and taking the trained initial neural network model as the target neural network model.
Referring to fig. 4, in an embodiment, an alternative implementation of step S302 includes:
s401, obtaining a cuckoo algorithm.
Cuckoo algorithm is also called cuckoo search algorithm or cuckoo search.
S402, according to the structural data of the initial neural network model, encoding the parameters of the initial neural network model by using a cuckoo algorithm to obtain the encoded data of the initial neural network.
And determining a cuckoo algorithm according to the structural data of the initial neural network model, and encoding the parameter according to the determined cuckoo algorithm. The parameters refer to an initial weight and an initial threshold of the initial neural network model, the initial weight and the initial threshold are coded by using a cuckoo algorithm, and the obtained coded data also comprises the weight and the threshold. The initial neural network model is optimized through a cuckoo algorithm, and the defect that the training of the neural network model is easy to fall into local optimization is overcome.
And S403, updating the initial neural network model according to the coded data to obtain the target neural network model.
Namely, the initial neural network model is updated according to the weight and the threshold value included in the data to be coded, and the target neural network model is obtained. And inputting the fault characteristic quantity into the target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
Referring to fig. 5, in an embodiment of the present application, an alternative implementation of S402 includes:
s501, obtaining a weight to be encoded and a threshold to be encoded, corresponding to the initial neural network model, of each cuckoo individual of the cuckoo population in the cuckoo algorithm, and obtaining data to be encoded corresponding to each cuckoo individual, wherein the data to be encoded comprises the weight to be encoded and the threshold to be encoded.
The weight value to be coded and the threshold value to be coded of each cuckoo individual are randomly generated, but the length of the data to be coded of each cuckoo individual is determined by the number of the weight values and the number of the threshold values contained in the structural data of the initial neural network model, and the length of the data to be coded of each cuckoo individual is equal. Each cuckoo individual code adopts a real number coding method, each cuckoo individual is represented by a real number string, the real number string is composed of weight data and threshold data among an input layer, a hidden layer and an output layer of the initial neural network model, and the weight data and the threshold data among the input layer, the hidden layer and the output layer are randomly generated and distributed.
And S502, determining each fitness value corresponding to each cuckoo individual in the cuckoo population to obtain a fitness value data set, wherein the number of the fitness values in the fitness value data set is equal to the preset number.
In one embodiment, according to a formula
Determining a fitness value, wherein S represents the fitness value, n represents the number of output layer nodes of the initial neural network model, k represents a coefficient, yiRepresenting the actual output, o, of the initial neural network model when the fault feature is at node iiAnd representing the predicted output of the initial neural network model when the fault characteristic quantity is at a node i, wherein i represents the node.
It is understood that each cuckoo individual has different n and i, and the fitness value, i.e., S, corresponding to each cuckoo individual can be determined according to the n and i corresponding to each cuckoo individual. And the fitness values of all cuckoo individuals in the cuckoo population form the fitness value data set.
And S503, obtaining the target cuckoo individual according to the average value of the fitness values in the fitness value data group.
In one embodiment, an optional implementation of step S503 includes:
and if the average value of the fitness values in the fitness value data group is greater than or equal to the preset average value, acquiring the maximum fitness value in the fitness value data group, and determining the cuckoo individuals corresponding to the maximum fitness value to obtain the target cuckoo individuals.
The average value of the fitness values refers to the average value of the fitness values obtained by dividing the sum of the fitness values of all cuckoo individuals in the cuckoo population by the number of the cuckoo individuals in the cuckoo population. The preset average value is set by a worker according to actual needs, and the application is not limited. And if the average value of the fitness values is larger than or equal to the preset average value, acquiring the maximum fitness value in the fitness value data set, and determining the cuckoo individual corresponding to the maximum fitness value to obtain the target cuckoo individual. That is, when the average value of the fitness values is greater than or equal to the preset average value, the cuckoo individual with the largest fitness value is found from the cuckoo population and is used as the target cuckoo individual.
Referring to fig. 6, in an embodiment of the present application, the method for diagnosing a fault of a power feeding apparatus further includes:
and S601, if the average value of the fitness values is smaller than the preset average value, iteratively updating the cuckoo population until the average value of the fitness values corresponding to the updated cuckoo population is larger than or equal to the preset average value or the iterative updating times reach preset updating times.
Wherein, an optional implementation manner of S601 includes:
A. and acquiring a first cuckoo individual and a second cuckoo individual in the cuckoo population.
B. And respectively obtaining the fitness value corresponding to the first cuckoo individual and the fitness value corresponding to the second cuckoo individual to obtain a first fitness value and a second fitness value.
C. If the second fitness value is larger than the first fitness value, replacing the first cuckoo individual with a second cuckoo individual;
D. rejecting nxpaRandomly generating n individuals with the worst fitness and filling the individuals into the cuckoo population to form the updated cuckoo population; wherein n is the number of nodes of the input layer, paThe value range is 0-1.
And E, iteratively executing the steps A-D until the average value of the fitness values corresponding to the updated cuckoo population is greater than or equal to a preset average value or the iterative updating times reach preset updating times.
The first cuckoo individual and the second cuckoo individual may be two cuckoo individuals sequentially selected in sequence, or may be two cuckoo individuals randomly selected in the cuckoo population, and the specific selection mode may be selected according to actual needs, which is not limited in the present application. In one embodiment, the first cuckoo individual and the second cuckoo individual may be two cuckoo individuals sequentially selected in the order of arrangement of the cuckoo individuals.
If the average value of the fitness values corresponding to the cuckoo population is smaller than the preset average value, selecting a larger fitness value by comparing the first fitness value with the second fitness value, and if the second fitness value is larger than the first fitness value, replacing the first cuckoo individual with the second cuckoo individual and rejecting the first cuckoo individual. Namely, the original first cuckoo individual and the original second cuckoo individual become two cuckoo individuals with equal fitness values, and the fitness values of the two cuckoo individuals are both equal to the second fitness value. And after the cuckoo individuals in the cuckoo population are replaced and removed for the first time, updating the cuckoo population for the first time. In addition, it is necessary to eliminate n × paRandomly generating n individuals with the worst fitness and filling the individuals into the cuckoo population to form the updated cuckoo population;wherein n is the number of nodes of the input layer, paThe value range is 0-1. The purpose of randomly generating n individuals to supplement the population is to keep the population size of the cuckoo unchanged. When the average value of the fitness values corresponding to the cuckoo population is still smaller than the preset average value, the cuckoo population needs to be continuously replaced, eliminated and updated by the cuckoo individuals until the average value of the fitness values corresponding to the cuckoo population is larger than or equal to the preset average value.
S602, determining a weight value to be coded and a threshold value to be coded, which correspond to the initial neural network model, of the target cuckoo individual to obtain the coded data.
The average value of the fitness values corresponding to the cuckoo population to which the target cuckoo individual belongs is larger than or equal to the preset average value, and the target cuckoo individual is the cuckoo individual with the largest fitness value in the cuckoo population to which the target cuckoo individual belongs. The encoding data comprises the weight to be encoded and the threshold to be encoded of the target cuckoo individual, and the initial neural network model is updated according to the encoding data to obtain the target neural network model.
Optionally, the above steps S601 to S602 may be executed after the foregoing step S502, that is, after the fitness value data group is acquired, the processes in steps S601 to S602 may be used to update the cuckoo population according to the fitness value average value corresponding to the fitness value data group until the fitness value average value corresponding to the updated cuckoo population is greater than or equal to the preset average value.
Referring to fig. 7, in an embodiment of the present application, the method for diagnosing a fault of a power feeding apparatus further includes:
s701, obtaining three-phase current waveforms corresponding to known fault diagnosis results to obtain test waveforms.
The method comprises the steps of obtaining a fault diagnosis result of an open circuit corresponding to a certain three-phase current waveform of a PWM rectifier, and verifying the diagnosis accuracy of the open circuit fault of the PWM rectifier by using the certain three-phase current waveform as a test waveform.
S702, inputting the test waveform into the target neural network model to obtain a fault diagnosis result.
Specifically, the fault characteristic quantity of the test waveform is determined, and then the fault characteristic quantity of the test waveform is input into the target neural network model to obtain a fault diagnosis result corresponding to the test waveform.
And S703, comparing the fault diagnosis result with the known fault diagnosis result, and determining the diagnosis accuracy of the target neural network model.
The diagnosis accuracy is the overall diagnosis accuracy of 110 groups of data for testing, and the higher the diagnosis accuracy is, the better the overall diagnosis effect is.
S704, if the diagnosis accuracy is greater than or equal to a preset accuracy, acquiring and storing a neural network structure and model parameters of the target neural network model to obtain a diagnosis neural network model, wherein the diagnosis neural network model is not updated any more;
that is, if the diagnosis accuracy is greater than or equal to the preset accuracy, the structure and parameters of the neural network at the moment are saved and determined as the structure and parameters of the neural network used for diagnosis after training, and the parameters of the neural network are not updated any more. The preset accuracy is set by a worker according to actual needs, and the method is not limited in the application. If the diagnosis accuracy is greater than or equal to the preset accuracy, the fault diagnosis capability of the diagnosis neural network model is proved to meet the requirement, and the diagnosis neural network model can be put into use. The target neural network model is tested before being put into use, so that the risk of fault diagnosis errors can be reduced, and the accuracy of the real-time fault diagnosis result of the PWM rectifier is guaranteed.
And S705, inputting the fault characteristic quantity into the diagnosis neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
When the method is applied, the fault characteristic quantity is input into the trained neural network, namely the diagnosis neural network model, and a real-time fault diagnosis result of the PWM rectifier is obtained. It should be noted that the diagnostic neural network model refers to the target neural network model for which the neural network structure and model parameters have been determined, and the diagnostic neural network model is not updated. Therefore, it can be said that the fault feature quantity is input to the target neural network model, but at this time, the neural network structure and the model parameters of the target neural network model are already determined, the target neural network model is not updated any more, and the diagnosis accuracy of the target neural network model is greater than or equal to the preset accuracy.
Optionally, the steps S701 to S704 may be executed before the step S203, that is, after the fault feature quantity is obtained, the target neural network model may be verified and tested by using the processes of the steps S701 to S704, and the fault feature quantity is not input into the target neural network model until the diagnosis accuracy of the target neural network model reaches the preset accuracy.
Referring to fig. 8, the present application further provides an energy feeding apparatus open-circuit fault diagnosis apparatus 10, including:
the acquisition module 11 is used for acquiring a three-phase current waveform of a PWM rectifier of the energy-fed power supply device;
and the processing module 12 is used for determining a fault characteristic quantity of the three-phase current waveform, wherein the fault characteristic quantity is norm entropy extracted from the three-phase current waveform signal based on a wavelet packet decomposition method.
The processing module 12 is further configured to input the fault feature quantity into a target neural network model, so as to obtain a real-time fault diagnosis result of the PWM rectifier.
The processing module 12 is specifically configured to determine structural data of the initial neural network model, where the structural data includes the number of input layer nodes, the number of hidden layer nodes, and the number of output layer nodes; creating an initial neural network model according to the structural data of the initial neural network model; and training the initial neural network model, and taking the trained initial neural network model as the target neural network model.
The processing module 12 is specifically configured to obtain a cuckoo algorithm; according to the structural data of the initial neural network model, encoding the parameters of the initial neural network model by using a cuckoo algorithm to obtain encoded data of the initial neural network; and updating the initial neural network model according to the coded data to obtain the target neural network model.
The processing module 12 is specifically configured to obtain a weight to be encoded and a threshold to be encoded, which correspond to the initial neural network model, of each cuckoo individual of the cuckoo population in the cuckoo algorithm, and obtain data to be encoded corresponding to each cuckoo individual, where the data to be encoded includes the weight to be encoded and the threshold to be encoded; determining a plurality of fitness values corresponding to each cuckoo individual in the cuckoo population to obtain a fitness value data set, wherein the number of the fitness values in the fitness value data set is equal to the preset number; obtaining target cuckoo individuals according to the average value of the fitness values in the fitness value data set; and determining a weight value to be coded and a threshold value to be coded, which correspond to the initial neural network model, of the target cuckoo individual to obtain the coded data.
The processing module 12 is specifically configured to, if the average value of the fitness values in the fitness value data set is greater than or equal to the preset average value, obtain a maximum fitness value in the fitness value data set, and determine a cuckoo individual corresponding to the maximum fitness value, so as to obtain a target cuckoo individual.
The processing module 12 is specifically configured to iteratively update the cuckoo population if the fitness value average is smaller than the preset average until the fitness value average corresponding to the updated cuckoo population is greater than or equal to the preset average or the iterative update frequency reaches a preset update frequency.
The processing module 12 is specifically configured to a obtain a first cuckoo individual and a second cuckoo individual in the cuckoo population; B. respectively obtaining the fitness value corresponding to the first cuckoo individual and the fitness value corresponding to the second cuckoo individual to obtain a first fitness value and a second fitness value; C. if the second fitness value is larger than the first fitness value, replacing the first cuckoo individual with the second cuckoo individual, and D, rejecting nxpaRandomly generating n individuals with the worst fitness and filling the individuals into the cuckoo population to form the updated cuckoo population; wherein n is the number of nodes of the input layer, paThe value range is 0-1; iteratively executing steps A-D until after the updateThe average value of the fitness values corresponding to the cuckoo population is greater than or equal to a preset average value or the iterative updating times reach preset updating times.
The processing module 12 is specifically configured to determine a fault characteristic quantity of the three-phase current waveform based on a wavelet packet algorithm.
The test module 13 is used for obtaining a three-phase current waveform corresponding to a known fault diagnosis result to obtain a test waveform; inputting the test waveform into the target neural network model to obtain a fault diagnosis result; comparing the fault diagnosis result with the known fault diagnosis result, and determining the diagnosis accuracy of the target neural network model; if the diagnosis accuracy rate is greater than or equal to the preset accuracy rate, acquiring and storing a neural network structure and model parameters of the target neural network model to obtain a diagnosis neural network model, wherein the diagnosis neural network model is not updated any more; and inputting the fault characteristic quantity into the diagnosis neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
Referring to fig. 9, the present application further provides a computer device 20, including a memory 21, a processor 22 and a transceiver 23, where the memory 31 is used for storing instructions, the transceiver 23 is used for communicating with other devices, and the processor 22 is used for executing the instructions stored in the memory 21, so that the computer device executes the open-circuit fault diagnosis method of the feeding apparatus according to any one of the above methods.
The present application further provides a computer-readable storage medium having stored therein computer-executable instructions, which when executed, cause a processor to execute the computer-executable instructions for implementing the method for diagnosing an open circuit fault of an energy feeding apparatus as provided in any one of the above embodiments. The present application also provides another computer-readable storage medium, in which computer-executable instructions are stored, and when executed, the instructions cause a computer to perform the method for diagnosing the open circuit fault of the energy feeding device provided in any one of the above embodiments.
The computer-readable storage medium may be a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read Only Memory (EPROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a magnetic Random Access Memory (FRAM), a Flash Memory (Flash Memory), a magnetic surface Memory, an optical Disc, or a Compact Disc Read-Only Memory (CD-ROM). And may be various electronic devices such as mobile phones, computers, tablet devices, personal digital assistants, etc., including one or any combination of the above-mentioned memories.
It should be noted that, in this document, 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 like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method described in the embodiments of the present application.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (8)
1. A method of diagnosing a fault in an feedable power supply, comprising:
acquiring a three-phase current waveform of a Pulse Width Modulation (PWM) rectifier of the energy-fed power supply device;
determining fault characteristic quantity of the three-phase current waveform, wherein the fault characteristic quantity is norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method;
inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier;
before inputting the fault characteristic quantity into the target neural network model, the method further includes: creating an initial neural network model; training the initial neural network model, and taking the trained initial neural network model as the target neural network model;
wherein the training the initial neural network model comprises:
acquiring a cuckoo algorithm, and acquiring a weight to be encoded and a threshold to be encoded, which correspond to the initial neural network model, of each cuckoo individual of a cuckoo population in the cuckoo algorithm to obtain data to be encoded corresponding to each cuckoo individual, wherein the data to be encoded comprises the weight to be encoded and the threshold to be encoded;
determining each fitness value corresponding to each cuckoo individual in the cuckoo population to obtain a fitness value data set, wherein the number of the fitness values in the fitness value data set is equal to a preset number;
if the average value of the fitness values in the fitness value data group is larger than or equal to a preset average value, acquiring the maximum fitness value in the fitness value data group, and determining the cuckoo individuals corresponding to the maximum fitness value to obtain target cuckoo individuals;
determining a weight value to be coded and a threshold value to be coded, which correspond to the initial neural network model, of the target cuckoo individual to obtain coded data;
and updating the initial neural network model according to the coded data to obtain the target neural network model.
2. The method of claim 1, wherein creating the initial neural network model comprises:
determining structural data of an initial neural network model, wherein the structural data comprises the number of input layer nodes, the number of hidden layer nodes and the number of output layer nodes;
and creating an initial neural network model according to the structural data of the initial neural network model.
3. The method of claim 1, further comprising:
and if the average value of the fitness values is smaller than a preset average value, iteratively updating the cuckoo population until the average value of the fitness values corresponding to the updated cuckoo population is larger than or equal to the preset average value or the iterative updating times reach preset updating times.
4. The method of claim 3, wherein the iteratively updating the cuckoo population until the average of the fitness values corresponding to the updated cuckoo population is greater than or equal to the preset average or the number of iterative updates reaches a preset number of updates comprises:
A. obtaining a first cuckoo individual and a second cuckoo individual in the cuckoo population;
B. respectively obtaining the fitness value corresponding to the first cuckoo individual and the fitness value corresponding to the second cuckoo individual to obtain a first fitness value and a second fitness value;
C. if the second fitness value is larger than the first fitness value, replacing the first cuckoo individual with the second cuckoo individual;
D. rejecting n
The individual with the worst fitness is randomly generated and n individuals are supplemented into the individualThe cuckoo population forms the updated cuckoo population; wherein n is the number of nodes of the input layer,
the value range is 0-1;
and A-D is executed in an iterative mode until the average value of the fitness values corresponding to the updated cuckoo population is larger than or equal to a preset average value or the iterative updating times reach preset updating times.
5. The method according to any one of claims 1 to 4, further comprising:
obtaining a three-phase current waveform corresponding to a known fault diagnosis result to obtain a test waveform;
inputting the test waveform into the target neural network model to obtain a fault diagnosis result;
comparing the fault diagnosis result with the known fault diagnosis result, and determining the diagnosis accuracy of the target neural network model;
if the diagnosis accuracy rate is greater than or equal to a preset accuracy rate, acquiring and storing a neural network structure and model parameters of the target neural network model to obtain a diagnosis neural network model, wherein the diagnosis neural network model is not updated any more;
and inputting the fault characteristic quantity into the diagnosis neural network model to obtain a real-time fault diagnosis result of the PWM rectifier.
6. An energy feed device open circuit fault diagnostic device, comprising:
the acquisition module is used for acquiring the three-phase current waveform of the Pulse Width Modulation (PWM) rectifier of the energy feed power supply device;
the processing module is used for determining fault characteristic quantity of the three-phase current waveform, and the fault characteristic quantity is norm entropy extracted from a three-phase current waveform signal based on a wavelet packet decomposition method;
the processing module is further used for inputting the fault characteristic quantity into a target neural network model to obtain a real-time fault diagnosis result of the PWM rectifier;
wherein, before inputting the fault characteristic quantity into the target neural network model, the processing module is further configured to: creating an initial neural network model; training the initial neural network model, and taking the trained initial neural network model as the target neural network model;
wherein, when training the initial neural network model, the processing module is specifically configured to:
acquiring a cuckoo algorithm, and acquiring a weight to be encoded and a threshold to be encoded, which correspond to the initial neural network model, of each cuckoo individual of a cuckoo population in the cuckoo algorithm to obtain data to be encoded corresponding to each cuckoo individual, wherein the data to be encoded comprises the weight to be encoded and the threshold to be encoded;
determining each fitness value corresponding to each cuckoo individual in the cuckoo population to obtain a fitness value data set, wherein the number of the fitness values in the fitness value data set is equal to a preset number;
if the average value of the fitness values in the fitness value data group is larger than or equal to a preset average value, acquiring the maximum fitness value in the fitness value data group, and determining the cuckoo individuals corresponding to the maximum fitness value to obtain target cuckoo individuals;
determining a weight value to be coded and a threshold value to be coded, which correspond to the initial neural network model, of the target cuckoo individual to obtain coded data;
and updating the initial neural network model according to the coded data to obtain the target neural network model.
7. A computer device comprising a memory for storing instructions, a processor for communicating with other devices, and a transceiver for executing the instructions stored in the memory to cause the computer device to perform the method of fault diagnosis of an energy feed apparatus according to any one of claims 1 to 5.
8. A computer-readable storage medium having stored therein computer-executable instructions that, when executed, cause a computer to perform the method of fault diagnosis of a feedable power unit according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010738590.3A CN111880117B (en) | 2020-07-28 | 2020-07-28 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010738590.3A CN111880117B (en) | 2020-07-28 | 2020-07-28 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111880117A CN111880117A (en) | 2020-11-03 |
| CN111880117B true CN111880117B (en) | 2022-04-08 |
Family
ID=73200888
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010738590.3A Active CN111880117B (en) | 2020-07-28 | 2020-07-28 | Fault diagnosis method and device for energy-fed power supply device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111880117B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103364683A (en) * | 2013-07-01 | 2013-10-23 | 东南大学 | Method for detecting open-circuit fault of inverter circuit |
| CN103837791A (en) * | 2014-03-20 | 2014-06-04 | 上海应用技术学院 | Three-level inverter multi-mode fault diagnosis circuit and diagnosis method thereof |
| CN104462770A (en) * | 2014-11-10 | 2015-03-25 | 福州大学 | Signal reconstruction method taking sensor performance degradation into consideration |
| CN106384122A (en) * | 2016-09-05 | 2017-02-08 | 江苏科技大学 | Device fault mode identification method based on improved CS-LSSVM |
| CN107632258A (en) * | 2017-09-12 | 2018-01-26 | 重庆大学 | A kind of fan converter method for diagnosing faults based on wavelet transformation and DBN |
| CN107766638A (en) * | 2017-10-13 | 2018-03-06 | 西南交通大学 | A kind of EMUs four-quadrant pulse rectifier IGBT open fault diagnostic methods |
| CN108596212A (en) * | 2018-03-29 | 2018-09-28 | 红河学院 | Based on the Diagnosis Method of Transformer Faults for improving cuckoo chess game optimization neural network |
| CN108805256A (en) * | 2018-08-07 | 2018-11-13 | 南京工业大学 | Photovoltaic module method for diagnosing faults based on cuckoo algorithm and BP neural network |
| CN108872751A (en) * | 2018-07-05 | 2018-11-23 | 西南交通大学 | A kind of method for diagnosing faults of three level Cascade H-Bridge Inverter neural network based |
| CN110061494A (en) * | 2019-04-09 | 2019-07-26 | 上海电力学院 | A kind of probabilistic three-phase imbalance reconstruction method of power distribution network of consideration DG power output |
| CN110068776A (en) * | 2019-05-16 | 2019-07-30 | 合肥工业大学 | Three-level inverter open-circuit fault diagnostic method based on Support Vector Machines Optimized |
| CN110766137A (en) * | 2019-10-18 | 2020-02-07 | 武汉大学 | Power electronic circuit fault diagnosis method based on longicorn whisker optimized deep confidence network algorithm |
| US10678233B2 (en) * | 2017-08-02 | 2020-06-09 | Strong Force Iot Portfolio 2016, Llc | Systems and methods for data collection and data sharing in an industrial environment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107153735A (en) * | 2017-04-28 | 2017-09-12 | 无锡开放大学 | Motor driven systems PWM inverter method for diagnosing faults |
| CN108416103A (en) * | 2018-02-05 | 2018-08-17 | 武汉大学 | A kind of method for diagnosing faults of electric automobile of series hybrid powder AC/DC convertor |
-
2020
- 2020-07-28 CN CN202010738590.3A patent/CN111880117B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103364683A (en) * | 2013-07-01 | 2013-10-23 | 东南大学 | Method for detecting open-circuit fault of inverter circuit |
| CN103837791A (en) * | 2014-03-20 | 2014-06-04 | 上海应用技术学院 | Three-level inverter multi-mode fault diagnosis circuit and diagnosis method thereof |
| CN104462770A (en) * | 2014-11-10 | 2015-03-25 | 福州大学 | Signal reconstruction method taking sensor performance degradation into consideration |
| CN106384122A (en) * | 2016-09-05 | 2017-02-08 | 江苏科技大学 | Device fault mode identification method based on improved CS-LSSVM |
| US10678233B2 (en) * | 2017-08-02 | 2020-06-09 | Strong Force Iot Portfolio 2016, Llc | Systems and methods for data collection and data sharing in an industrial environment |
| CN107632258A (en) * | 2017-09-12 | 2018-01-26 | 重庆大学 | A kind of fan converter method for diagnosing faults based on wavelet transformation and DBN |
| CN107766638A (en) * | 2017-10-13 | 2018-03-06 | 西南交通大学 | A kind of EMUs four-quadrant pulse rectifier IGBT open fault diagnostic methods |
| CN108596212A (en) * | 2018-03-29 | 2018-09-28 | 红河学院 | Based on the Diagnosis Method of Transformer Faults for improving cuckoo chess game optimization neural network |
| CN108872751A (en) * | 2018-07-05 | 2018-11-23 | 西南交通大学 | A kind of method for diagnosing faults of three level Cascade H-Bridge Inverter neural network based |
| CN108805256A (en) * | 2018-08-07 | 2018-11-13 | 南京工业大学 | Photovoltaic module method for diagnosing faults based on cuckoo algorithm and BP neural network |
| CN110061494A (en) * | 2019-04-09 | 2019-07-26 | 上海电力学院 | A kind of probabilistic three-phase imbalance reconstruction method of power distribution network of consideration DG power output |
| CN110068776A (en) * | 2019-05-16 | 2019-07-30 | 合肥工业大学 | Three-level inverter open-circuit fault diagnostic method based on Support Vector Machines Optimized |
| CN110766137A (en) * | 2019-10-18 | 2020-02-07 | 武汉大学 | Power electronic circuit fault diagnosis method based on longicorn whisker optimized deep confidence network algorithm |
Non-Patent Citations (3)
| Title |
|---|
| Open-circuit fault diagnosis of power rectifier using sparse autoencoder based deep neural network;Lin Xu 等;《Neurocomputing》;20180519;第1-10页 * |
| 基于布谷鸟算法的光伏组件故障诊断模型优化;张杰 等;《电源技术》;20200131;第76-80页 * |
| 小波范数熵特征提取的模拟电路故障诊断方法;肖迎群 等;《机械与电子》;20151231;第3-9页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111880117A (en) | 2020-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109501834B (en) | Method and device for predicting fault of turnout switch machine | |
| Sobajic et al. | An artificial intelligence system for power system contingency screening | |
| CN107622360A (en) | A kind of critical circuits recognition methods for considering subjective and objective factor | |
| CN104217104A (en) | Power transformer service life analysis method and system based on risk evaluation | |
| Feng et al. | A technical framework of PHM and active maintenance for modern high-speed railway traction power supply systems | |
| CN106982144B (en) | Power communication network fault recovery method | |
| CN111516539A (en) | Power distribution method and device, and computer readable storage medium | |
| Lin et al. | Adaptive multiple fault detection and alarm processing for loop system with probabilistic network | |
| CN109543814A (en) | A kind of each equipment fault prediction technique of subway signal system | |
| CN110689186A (en) | Power distribution network power supply reliability assessment method and system based on wind power output randomness | |
| CN111880117B (en) | Fault diagnosis method and device for energy-fed power supply device and storage medium | |
| CN104037760A (en) | Anticipated accident selection method for electric power system with uncertain injection power | |
| CN116523294A (en) | Knowledge graph-based power grid risk early warning and decision system, method and equipment | |
| CN105896545A (en) | Accident chain mode based distributed power flow controller location method | |
| CN116151562A (en) | Mobile emergency vehicle scheduling and power distribution network toughness improving method based on graphic neural network reinforcement learning | |
| CN111210092A (en) | Stacking machine predictive maintenance method and system based on deep learning | |
| Zhao et al. | Method to decide a multi-fault rush repair robust strategy in power distribution networks | |
| CN116132350A (en) | Path selection method, device and storage medium | |
| CN113344238A (en) | Power distribution network security risk assessment and reduction method, device and storage medium | |
| CN113537703B (en) | Power distribution network pre-disaster emergency resource deployment method and device and computer equipment | |
| CN114330884B (en) | Multi-time scale coordinated AC/DC power grid dynamic safety early warning method and system | |
| CN110783911A (en) | Soft intelligent switch configuration method and system for medium and low voltage distribution network | |
| CN112508313B (en) | Method, device and system for evaluating operation sequence of recoverable equipment after fault | |
| Zhang et al. | Vulnerability analysis of link-weighted Shanghai Metrorail transit network | |
| Gromov et al. | Synthesis of Data Transmission Networks with Specified Survivability under Negative External Influences |
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 |