CN111767675A - Transformer vibration fault monitoring method and device, electronic equipment and storage medium - Google Patents
Transformer vibration fault monitoring method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN111767675A CN111767675A CN202010591946.5A CN202010591946A CN111767675A CN 111767675 A CN111767675 A CN 111767675A CN 202010591946 A CN202010591946 A CN 202010591946A CN 111767675 A CN111767675 A CN 111767675A
- Authority
- CN
- China
- Prior art keywords
- characteristic information
- transformer
- vibration
- determining
- vibration signal
- 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
- 238000000034 method Methods 0.000 title claims abstract description 79
- 238000012544 monitoring process Methods 0.000 title claims abstract description 42
- 230000002159 abnormal effect Effects 0.000 claims abstract description 117
- 238000012545 processing Methods 0.000 claims abstract description 44
- 230000008569 process Effects 0.000 claims description 42
- 210000002569 neuron Anatomy 0.000 claims description 35
- 238000012549 training Methods 0.000 claims description 24
- 238000004458 analytical method Methods 0.000 claims description 22
- 238000004422 calculation algorithm Methods 0.000 claims description 22
- 238000013528 artificial neural network Methods 0.000 claims description 16
- 238000012806 monitoring device Methods 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 12
- 238000005192 partition Methods 0.000 claims description 5
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 10
- 230000005540 biological transmission Effects 0.000 description 9
- 238000004804 winding Methods 0.000 description 6
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 5
- 230000005611 electricity Effects 0.000 description 4
- 238000001816 cooling Methods 0.000 description 3
- 229910000976 Electrical steel Inorganic materials 0.000 description 2
- 230000002860 competitive effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 238000013316 zoning Methods 0.000 description 2
- 241001370313 Alepes vari Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 150000003376 silicon Chemical class 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2148—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the process organisation or structure, e.g. boosting cascade
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)
Abstract
The invention discloses a transformer vibration fault monitoring method and device, electronic equipment and a storage medium. The transformer vibration fault monitoring method comprises the following steps: and acquiring an abnormal vibration signal of the transformer, and processing the abnormal vibration signal to acquire corresponding characteristic information, wherein the characteristic information is used for reflecting the time frequency of the abnormal vibration signal. And determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model. And forming fault information of the transformer based on the vibration fault type and carrying out fault prompt. The invention realizes the real-time monitoring and prompting of the vibration fault of the transformer.
Description
Technical Field
The embodiment of the invention relates to a transformer monitoring technology, in particular to a transformer vibration fault monitoring method and device, electronic equipment and a storage medium.
Background
In the process of power development, a transformer is one of the most important devices in power equipment, and the role of the transformer in the power grid is very important. For example, the transformer not only needs to boost the voltage at the power generation side in the power grid to realize the long-distance transmission of electricity, but also transforms the transmitted voltage according to the electricity consumption requirements of users at different voltage levels to realize the power supply of the users. Therefore, if the fault type of the transformer can be accurately judged when the transformer has a fault, the fault treatment can be timely and effectively carried out on the transformer by workers, the normal operation of the transformer is ensured, the normal operation of a power grid is further ensured, and the protection and navigation of the first-class power grid in the world are built. The vibration of the transformer is the most main reason for the easy occurrence of faults in the operation of the transformer, and therefore, a transformer vibration fault monitoring method is urgently needed to realize the real-time monitoring and prompting of the transformer vibration fault.
Disclosure of Invention
The invention provides a method and a device for monitoring a vibration fault of a transformer, electronic equipment and a storage medium, which are used for realizing real-time monitoring and prompting of the vibration fault of the transformer.
In a first aspect, an embodiment of the present invention provides a method for monitoring a vibration fault of a transformer. The method comprises the following steps:
an abnormal vibration signal of the transformer is obtained,
processing the abnormal vibration signal to obtain corresponding characteristic information, wherein the characteristic information is used for reflecting the time frequency of the abnormal vibration signal;
determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model;
and forming fault information of the transformer based on the vibration fault type and carrying out fault prompt.
In a second aspect, an embodiment of the present invention further provides a transformer vibration fault monitoring apparatus, where the apparatus includes:
the acquisition module is used for acquiring an abnormal vibration signal of the transformer;
the processing module is used for processing the abnormal vibration signal to obtain corresponding characteristic information, and the characteristic information is used for reflecting the time frequency of the abnormal vibration signal;
the determining module is used for determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model;
and the prompting module is used for forming fault information of the transformer based on the vibration fault type and performing fault prompting.
In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the transformer vibration fault monitoring method according to any one of the first aspect.
In a fourth aspect, an embodiment of the present invention further provides a storage medium, where instructions are stored, and when the instructions are executed by the processor, the storage medium is configured to perform the transformer vibration fault monitoring method according to any one of the first aspect.
According to the method, the abnormal vibration signal of the transformer is acquired in real time, the corresponding characteristic information is acquired by processing the abnormal vibration signal, and the vibration fault type of the abnormal vibration signal is determined according to the characteristic information and a pre-trained learning model. And then fault information of the transformer is formed based on the vibration fault type and fault prompt is carried out, so that real-time monitoring and prompt of the vibration fault of the transformer are realized.
Drawings
Fig. 1 is a schematic structural diagram of a transformer fault monitoring system according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a transformer vibration fault monitoring method according to an embodiment of the present invention;
fig. 3 is a schematic flowchart of a learning model training process according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an SOM neural network according to an embodiment of the present invention;
fig. 5 is a schematic flowchart of a learning model training process according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a transformer vibration fault monitoring apparatus according to a second embodiment of the present invention;
fig. 7 is a schematic structural diagram of another transformer vibration fault monitoring apparatus according to a second embodiment of the present invention;
fig. 8 is a schematic structural diagram of another transformer vibration fault monitoring apparatus according to a second embodiment of the present invention;
fig. 9 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
In the process of power development, a transformer is one of the most important devices in power equipment, and the role of the transformer in the power grid is very important. For example, the transformer not only needs to boost the voltage at the power generation side in the power grid to realize the long-distance transmission of electricity, but also transforms the transmitted voltage according to the electricity consumption requirements of users at different voltage levels to realize the power supply of the users.
Different types of faults often occur when transformers are operated in environments with different voltage levels. Among them, the vibration of the transformer (also called as the vibration of the transformer body) is the most main cause of the transformer being prone to malfunction during operation. Therefore, the vibration of the transformer is monitored on line, so that the vibration fault of the transformer can be monitored in real time, and the method is an important monitoring means for ensuring the stable and healthy operation of the transformer and further ensuring the normal operation of a power grid.
The vibration of the transformer mainly comes from the vibration caused by the iron core, the winding and the cooling device which are included in the transformer and the foundation of the transformer. The iron core of the transformer is composed of laminated silicon steel sheets. The magnetic hysteresis expansion of the silicon steel sheet under the strong magnetic field causes the vibration of the iron core. Wherein the amplitude A1 of the vibration is proportional to the square of the excitation voltage U1 (i.e., A1 ℃. varies.U 1)2) And since the variation period T of the hysteresis is exactly half of the mains frequency power supply period Tu-f (i.e., T is 1/2Tu-f), the fundamental frequency f1 of the vibration of the transformer caused by the hysteresis is twice the power supply frequency fi (i.e., f1 is 2 × fi), i.e., 100Hz is the fundamental frequency of the vibration.
The transformer includes a winding that vibrates due to electromagnetic oscillations generated by current flowing in the winding. The amplitude A2 of the vibration is proportional to the square of the winding current IN (i.e., A ^ IN)2) And the fundamental frequency f2 of the vibration is twice the current frequency fi (i.e. f2 ═ 2 × fi), i.e. 100Hz is the fundamental frequency of the vibration.
Because the vibration that the base that the iron core, winding and the cooling device that the transformer included to and the transformer arouses all can transmit to the tank wall of transformer through different routes, so can through install vibration sensor additional on the tank wall to gather the vibration signal realization monitoring transformer inner iron core, winding and cooling device etc. state change of tank wall.
Fig. 1 is a schematic structural diagram of a transformer fault monitoring system according to an embodiment of the present invention. As shown in fig. 1, the transformer fault monitoring system may include: transformer 101, vibration sensor 102, data transmission device 103 and transformer vibration fault monitoring device 104. The vibration sensor 102 may be disposed on the transformer 101 (e.g., disposed on a tank wall of the transformer 101). The data transmission device 103 is electrically connected to the vibration sensor 102 and the transformer vibration failure monitoring device 104, respectively.
The vibration sensor 102 may be configured to collect a vibration signal of the transformer, where the vibration signal may be a current signal or a voltage signal. The data transmission device 103 may be configured to determine that the vibration signal of the transformer collected by the vibration sensor is a normal vibration signal or an abnormal vibration signal, and transmit the vibration signal to the transformer vibration fault monitoring device. The normal vibration information refers to a vibration signal acquired when the transformer fails; the abnormal vibration signal refers to a vibration signal acquired when the transformer fails. The transformer vibration fault monitoring device 104 may be configured to determine a vibration fault type of the transformer according to the abnormal vibration signal of the transformer transmitted by the data transmission device, and form fault information of the transformer based on the vibration fault type and perform fault prompt.
Example one
Fig. 2 is a schematic flow chart of a transformer vibration fault monitoring method according to an embodiment of the present invention, where this embodiment is applicable to real-time monitoring of a transformer vibration fault, and the method may be executed by the transformer vibration fault monitoring device in the transformer fault monitoring system shown in fig. 1, and specifically includes the following steps:
The transformer vibration fault monitoring device adopts a transmission protocol negotiated with the data transmission device to obtain and analyze an abnormal vibration signal sent by the data transmission device.
And step 202, processing the abnormal vibration signal to obtain corresponding characteristic information.
The characteristic information of the abnormal vibration signal can be used for reflecting the time frequency of the abnormal vibration signal. Optionally, the feature information may include M features, each feature may be used to reflect partial time-frequency information of the abnormal vibration signal, M is a positive integer, and M > 1.
Optionally, the process of processing the abnormal vibration signal to obtain corresponding characteristic information may include: and performing target processing on the abnormal vibration signal to determine the characteristic information of the abnormal vibration signal. The target process may include at least one of a wavelet packet analysis process, a time domain analysis process, or a frequency domain analysis process.
When the target processing includes wavelet packet analysis processing, the characteristic information of the abnormal vibration signal may include energy information of a frequency band of the abnormal vibration signal. For example, the abnormal vibration signal may be divided into 8 frequency bands, and the energy information of each frequency band may be a percentage of energy corresponding to the frequency band to the total energy value. When the target process includes a time-domain analysis process, the characteristic information of the abnormal vibration signal may include at least one of a peak-to-peak value, a standard deviation, a root-mean-square, and a skewness of the abnormal vibration signal. Illustratively, the characteristic information of the abnormal vibration signal includes a peak-to-peak value, a standard deviation, a root-mean-square, and a skewness of the abnormal vibration signal. When the target process includes a frequency domain analysis process, the characteristic information of the abnormal vibration signal includes at least one of a center-of-gravity frequency, a frequency variance, and a mean-square frequency of the abnormal vibration signal. Illustratively, the characteristic information of the abnormal vibration signal includes a center-of-gravity frequency, a frequency variance, and a mean square frequency of the abnormal vibration signal.
For example, the transformer vibration fault monitoring device may perform wavelet packet analysis processing, time domain analysis processing, and frequency domain analysis processing on the abnormal vibration signal, respectively, and the characteristic information determined by the device includes energy information of a frequency band of the abnormal vibration signal, a peak-to-peak value, a standard deviation, a root-mean-square, a skewness, a center-of-gravity frequency, a frequency variance, and a mean-square frequency. The process of performing wavelet packet analysis processing on the abnormal vibration signal may specifically be: and performing 4-layer wavelet packet transformation processing on the abnormal vibration signal by adopting a db6 wavelet of a multi-Behcet (Daubechies, db) wavelet series. Therefore, the transformer vibration fault monitoring device processes the abnormal vibration signal for multiple times, so that the multi-dimensional (multidimensional) characteristic information corresponding to the abnormal vibration signal can be acquired from different aspects, and the comprehensiveness and the universality of the characteristic information of the abnormal vibration signal are improved. And a foundation is laid for accurately determining the vibration fault type according to the characteristic information in the follow-up process.
And 203, screening the characteristic information by adopting an AdaBoost algorithm, and determining first important characteristic information with strong partition capacity.
An Adaptive Boosting (AdaBoost) algorithm is an iterative algorithm, and the core idea is to train different weak classifiers aiming at the same training set and then combine the weak classifiers to form a strong classifier.
In the embodiment of the invention, the AdaBoost algorithm is utilized to evaluate and screen the distinguishing capability of the characteristic information corresponding to the abnormal vibration signal, so that part of the screened characteristic information with higher distinguishing capability is determined as the first important characteristic information, unnecessary characteristics in the characteristic information are eliminated, and the dimension reduction characteristic information is optimized. Wherein the first significant characteristic information may include one or more characteristics. The distinguishing capability of the characteristic information refers to the capability of distinguishing whether the vibration signal corresponding to the characteristic information belongs to a normal vibration signal or an abnormal vibration signal according to the characteristic information, and the stronger the distinguishing capability is, the higher the accuracy of determining whether the vibration signal is normal or abnormal according to the characteristic information is; on the contrary, the weaker the distinguishing capability is, the lower the accuracy of determining whether the vibration signal is normal or abnormal according to the characteristic information is.
Optionally, the method for screening the feature information by using the AdaBoost algorithm to optimize the feature information may include: and constructing a weak classifier aiming at each feature in the feature information so that each weak classifier corresponds to one feature, and carrying out classification judgment by the weak classifier according to the numerical value of the feature. From the constructed weak classifiers, strong classifiers are formed (the process of forming strong classifiers can be considered as the process of picking features). In the process of forming the strong distributor, each round of circulation picks out the characteristics with the best classification performance under the current weight distribution.
In an example, taking the example that the acquired feature information of the abnormal vibration signal includes M features, an example of a process of screening the feature information by using an AdaBoost algorithm and determining first important feature information with strong zoning capability is described. The process may include the following steps a1 through E1. Wherein, a corresponding weak classifier is constructed for each of the M features included in the feature information, and then the strong classifier is composed of the M weak classifiers. The AdaBoost algorithm may be an AdaBoost-based samme.r algorithm.
And A1, constructing training data according to the abnormal vibration signal and the normal vibration signal.
Optionally, the transformer vibration fault monitoring device may store the normal vibration information of the transformer in advance. The transformer vibration fault monitoring device can construct a plurality of acquired abnormal vibration signals within a period of time and stored normal vibration signals into training data T.
T＝{(x1，y1)，(x2，y2)，...，(xi，yi)}，xi∈X，yi∈(+1，-1}，
xi is the vibration signal (sample) and yi is the sample class of xi. That is, yi ═ 1, xi is the normal vibration signal; yi ═ 1, xi is an abnormal vibration signal.
And step B1, initializing a weight distribution value (weight for short) of the training data.
Initializing weights of training data
And step C1, determining a strong classifier according to the weak classifier corresponding to each feature in the M features included in the feature information.
And each weak classifier corresponds to one feature, and M features included in the feature information respectively correspond to M classifiers. The process of determining a strong classifier according to a weak classifier corresponding to each of the M features included in the feature information may include:
first, the following steps (a) to (g) are performed for M weak classifiers, for the M-th weak classifier gm(xi)，m＝1,2,…,M：
Step (a) according to the weight
The m-th weak classifier learns the training data, and the output of the weak classifier is { +1, -1 }.
And (b) carrying out sample class identification on each datum in the training data.
Step (c) repeatedly performs the following step K times, for the K-th time, K is 1,2, …, K, and K is a positive integer.
First, the weights η of the samples classified into each class are calculated in a loopmkj：
Secondly, determine the sample weight sum η for each class to be classified correctlymkj＝[gm(xi)＝j]Whether it is greater than the sum of weights assigned to other types of samples ηmkj＝[gm(xi)≠j]And if the weight of the sample classified correctly in each class is greater than the sum of the weights of the samples classified into other classes, performing next cycle. And if the weight of the sample classified correctly in each class is not more than the sum of the weights classified into other samples, returning to the step to recalculate the weight sum of each sample.
Step (d) calculating weak classifiers gm(xi) E classification error rate ofm:
Step (e) calculating weak classifiers gm(xi) Occupied ratio weight α in strong classifierm:
Wherein the error rate e of classificationmWhen reduced, the weight of the proportion αmThe duty ratio weight α is increasedmCan be used to reflect weak classifier gm(xi) Importance in composing the Strong classifier the scaling weight αmCan also be used to reflect weak classifiers gm(xi) The partition capability of the corresponding feature. Weak classifier gm(xi) Is weighted αmThe larger the weak classifier g is, the weaker classifier g is indicatedm(xi) The stronger the zoning ability of the corresponding feature, the more important the feature is; otherwise, the weak classifier gm(xi) Is weighted αmThe smaller the weak classifier g is, the less the weak classifier g ism(xi) The weaker the partitioning capability of the corresponding feature, the less necessary the feature.
Step (f) updating the weight vector:
step (g) normalizing the updated weights
Secondly, obtain the strong classifier gf(x)：
And D1, determining all weak classifiers in the strong classifier determining process, wherein the first n weak classifiers are important weak classifiers according to the sequence of the proportion weight values corresponding to the weak classifiers in all the weak classifiers from large to small, n is a positive integer, and M is more than n and is more than or equal to 1.
Optionally, the ratio weight of each weak classifier in the strong classifier obtained by calculation in the step C1 may be recorded, the ratio weight corresponding to each weak classifier is compared, and the first n weak classifiers are determined to be important weak classifiers according to the sequence of decreasing the value of the ratio weight. For example, n may be the number of weak classifiers with a duty ratio weight not equal to 0.
And E1, determining the characteristics corresponding to the important weak classifier as the first important characteristic information. The first significant feature information includes n features.
For example, assuming that n is 3, the features corresponding to the 3 important weak classifiers determined in step D are the peak-to-peak value, the skewness and the center-of-gravity frequency, and the peak-to-peak value, the skewness and the center-of-gravity frequency are determined as the first important feature information, that is, the first important feature information includes the peak-to-peak value, the skewness and the center-of-gravity frequency.
And 204, updating the feature information, wherein the updated feature information is the first important feature information.
For example, assuming that the first important feature information includes a peak-to-peak value, a skewness, and a center-of-gravity frequency, after the feature information is updated, the updated feature information includes a peak-to-peak value, a skewness, and a center-of-gravity frequency.
And step 205, determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model.
Optionally, the feature information may be the updated feature information in step 204. The learning model may be a Self Organizing Map (SOM) neural network. In the embodiment of the invention, the learning model is taken as an SOM neural network as an example, and when the training processes of the learning models are different, the process of determining the vibration fault type of the abnormal vibration signal according to the characteristic information and the pre-trained learning model is also different.
The embodiment of the invention is schematically illustrated by the following two implementation modes aiming at the process of determining the vibration fault type of the abnormal vibration signal under the learning modules trained in different training processes.
A first alternative implementation: fig. 3 is a schematic flowchart illustrating a process of training a learning model according to an embodiment of the present invention. As shown in fig. 3, the learning model training step includes:
and 301, acquiring abnormal vibration signals of the transformer under different vibration fault types.
By way of example, assuming that the transformer has 4 vibration fault types, an abnormal vibration signal corresponding to each of the 4 vibration fault types is acquired.
And step 302, determining third important characteristic information of the abnormal vibration signal of the transformer under different vibration fault types.
The third important characteristic information of the abnormal vibration signals of the transformer under different vibration fault types is determined after the abnormal vibration signals of the transformer under different vibration fault types are processed to obtain corresponding characteristic information and the processed characteristic information is screened by adopting an AdaBoost algorithm. In addition, the process of determining the third important characteristic information may refer to step 202 and step 203, which is not described in detail in this embodiment of the present invention.
And step 303, inputting the third important characteristic information into an input layer of the learning model.
Fig. 4 is a schematic diagram of an SOM neural network according to an embodiment of the present invention. When the learning model is an SOM neural network, the SOM neural network is a self-organizing (competitive) neural network, and is composed of an input layer and an output layer (also called competitive layer), wherein the input layer and the output layer both include a plurality of neurons. The SOM neural network can map high-dimensional similar samples into neighboring neurons in the output layer for classification and clustering of samples, i.e., clustering similar samples together, without separation of samples that are dissimilar or less similar. For example, in the present invention, each neuron of the input layer of the SOM neural network may correspond to each type of feature in the third important feature information, and then the SOM neural network may map the neuron of the input layer into a neighboring neuron of the output layer.
And step 304, determining the neurons of the output layer of the learning model mapped by the third important feature information.
And after the third important characteristic information is input into an input layer of the SOM neural network, determining neurons of an output layer mapped by the third important characteristic information of the abnormal vibration signal of each vibration fault type of the transformer.
When the training step of the learning model is as in steps 301 to 304 described above, the process of determining the vibration fault type of the abnormal vibration signal in step 205 according to the feature information and the pre-trained learning model may include:
and step A2, inputting the characteristic information into the learning model, and determining the target neuron of the output layer corresponding to the characteristic information.
Inputting the updated feature information in step 204 into the learning model, and determining the target neuron of the output layer corresponding to the updated feature information.
And step B2, inquiring the incidence relation between the neurons and the vibration fault types, and determining the vibration fault types matched with the target neurons.
The transformer vibration fault monitoring device can record and store the incidence relation between the neuron and the transformer vibration fault type corresponding to the third important characteristic information corresponding to the input layer after the learning model training. After determining the target neuron of the output layer corresponding to the updated characteristic information, the device can query the association relationship between the neuron and the vibration fault type, and determine the vibration fault type matched with the target neuron.
And step C2, determining the vibration fault type as the vibration fault type of the abnormal vibration signal.
And determining the vibration fault type matched with the target neuron as the vibration fault type of the abnormal vibration signal.
A second alternative implementation: fig. 5 is a schematic flowchart illustrating a process of training a learning model according to an embodiment of the present invention. As shown in fig. 5, the learning model training step includes:
and 501, acquiring abnormal vibration signals of the transformer under different vibration fault types.
And 502, determining third important characteristic information of the abnormal vibration signal of the transformer under different vibration fault types.
And step 503, inputting the third important characteristic information into an input layer of the learning model.
And step 504, determining the neurons of the output layer of the learning model mapped by the third important feature information.
The implementation manners of step 501 to step 504 may refer to step 301 to step 304, which are not described in detail in the embodiment of the present invention.
And 505, setting a vibration fault type identifier for the neuron of the output layer according to the vibration fault type corresponding to the third important characteristic information.
Alternatively, the vibration failure type identifier may be a numerical identifier or an alphabetical identifier.
When the training step of the learning model is as the above step 501 to step 505, the process of determining the vibration fault type of the abnormal vibration signal according to the feature information and the pre-trained learning model in step 205 may include:
and step A3, inputting the characteristic information into the learning model, and determining the target neuron of the output layer corresponding to the characteristic information.
Inputting the updated feature information in step 204 into the learning model, and determining the target neuron of the output layer corresponding to the updated feature information.
And step B3, identifying the vibration fault type of the target neuron as the identified vibration fault type, and determining the vibration fault type as the vibration fault type of the abnormal vibration signal.
And step 206, forming fault information of the transformer based on the vibration fault type and carrying out fault prompt.
Optionally, the fault information may include a vibration fault type of the transformer and a fault prediction processing method thereof. The fault prompt can be a text display prompt, a voice prompt, a sound effect prompt and the like.
According to the technical scheme of the embodiment, the abnormal vibration signal of the transformer can be obtained in real time, the corresponding characteristic information can be obtained by processing the abnormal vibration signal, and the vibration fault type of the abnormal vibration signal can be determined according to the characteristic information and a pre-trained learning model. And then fault information of the transformer is formed based on the vibration fault type and fault prompt is carried out, so that real-time monitoring and prompt of the vibration fault of the transformer are realized.
And when the vibration fault type of the obtained abnormal vibration signal is determined by using the SOM neural network, the sample trained by the SOM neural network (the third important feature information of the transformer under different vibration fault types) is determined by processing the abnormal vibration signal of the transformer under different vibration fault types to obtain corresponding feature information, and screening the obtained corresponding feature information by using an AdaBoost algorithm, that is, the sample trained by the SOM neural network is determined by optimizing the AdaBoost algorithm. The AdaBoost algorithm screens the characteristic information of the abnormal vibration signal according to the proportion weight, so that the number of the characteristics included in the characteristic information is reduced, and the accuracy of reflecting the abnormal vibration signal by the characteristic information is improved. Therefore, the speed and the accuracy of the determined vibration fault type are improved, and the accuracy and the real-time performance of the transformer vibration fault monitoring are effectively improved.
Furthermore, because the AdaBoost algorithm is a framework and does not limit the types of weak learners, the embodiment of the invention can use different learning algorithms to construct weak classifiers so as to implement the construction of the weak classifiers corresponding to any features without specific screening of the features, thereby avoiding the over-fitting phenomenon. In addition, when the AdaBoost algorithm is used for screening the characteristic information, the proportion weight of each weak classifier is fully considered, and compared with the method for screening the characteristic information by adopting the bagging algorithm and the RandomForest algorithm, the method has higher accuracy.
Example two
The transformer vibration fault monitoring device provided by the embodiment of the invention can execute the transformer vibration fault monitoring method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Fig. 6 is a schematic structural diagram of a transformer vibration fault monitoring apparatus according to a second embodiment of the present invention. The device includes:
the obtaining module 601 is configured to obtain an abnormal vibration signal of the transformer.
The processing module 602 is configured to process the abnormal vibration signal to obtain corresponding feature information, where the feature information is used to reflect a time frequency of the abnormal vibration signal.
The determining module 603 is configured to determine a vibration fault type of the abnormal vibration signal according to the feature information and a pre-trained learning model.
And the prompting module 604 is used for forming fault information of the transformer based on the vibration fault type and performing fault prompting.
The transformer vibration fault monitoring device provided by this embodiment can obtain the abnormal vibration signal of the transformer in real time through the obtaining module, the processing module processes the abnormal vibration signal to obtain the corresponding characteristic information, and the determining module determines the vibration fault type of the abnormal vibration signal according to the characteristic information and the pre-trained learning model. And then the prompt module can form fault information of the transformer based on the vibration fault type and prompt the fault, so that the real-time monitoring and prompting of the vibration fault of the transformer are realized.
Optionally, as shown in fig. 7, the transformer vibration fault monitoring apparatus may further include:
the screening module 605 is configured to screen the feature information by using an AdaBoost algorithm, and determine first important feature information with a strong partition capability.
And an updating module 606 for updating the feature information, wherein the updated feature information is the first important feature information.
Optionally, the feature information includes M features, where M is a positive integer and M > 1, and the filtering module 605 is configured to: and constructing training data according to the abnormal vibration signals and the normal vibration signals. A weight distribution of the training data is initialized. And determining a strong classifier according to the weak classifier corresponding to each feature in the M features included in the feature information. And determining all weak classifiers in the process of determining the strong classifiers according to the sequence of the proportion weight values corresponding to the weak classifiers in all the weak classifiers from large to small, wherein the first n weak classifiers are important weak classifiers, n is a positive integer, and M is more than n and is more than or equal to 1. And determining the characteristics corresponding to the important weak classifiers as first important characteristic information, wherein the first important characteristic information comprises n characteristics.
Optionally, the processing module 602 is configured to:
and performing target processing on the abnormal vibration signal, and determining the characteristic information of the abnormal vibration signal, wherein the target processing comprises at least one of wavelet packet analysis processing, time domain analysis processing or frequency domain analysis processing. When the target processing comprises wavelet packet analysis processing, the characteristic information comprises energy information of a frequency band of the abnormal vibration signal; when the target processing comprises time domain analysis processing, the characteristic information comprises a peak-to-peak value, a standard deviation, a root-mean-square and a skewness of the abnormal vibration signal; when the target process includes a frequency domain analysis process, the characteristic information includes a barycentric frequency, a frequency variance, and a mean square frequency of the abnormal vibration signal.
Optionally, the learning model comprises a SOM neural network.
Optionally, as shown in fig. 8, the transformer vibration fault monitoring apparatus may further include:
and the information acquisition module 607 is used for acquiring abnormal vibration signals of the transformer under different vibration fault types. And determining third important characteristic information of the abnormal vibration signals of the transformer under different vibration fault types, wherein the third important characteristic information is determined after processing the abnormal vibration signals of the transformer under different vibration fault types to obtain corresponding characteristic information and screening the obtained corresponding characteristic information by adopting an AdaBoost algorithm. Inputting the third important feature information into an input layer of the learning model. Neurons of an output layer of the learning model mapped by the third important feature information are determined.
A determining module 603, further configured to: and inputting the characteristic information into the learning model, and determining the target neuron of the output layer corresponding to the characteristic information. And inquiring the incidence relation between the neurons and the vibration fault types, and determining the vibration fault types matched with the target neurons. And determining the vibration fault type as the vibration fault type of the abnormal vibration signal.
Optionally, the information collecting module 607 is further configured to: and setting a vibration fault type identifier for the neuron of the output layer according to the vibration fault type corresponding to the third important characteristic information.
A determining module 603, further configured to: and inputting the characteristic information into the learning model, and determining the target neuron of the output layer corresponding to the characteristic information. And identifying the vibration fault type identified by the vibration fault type identification of the target neuron, and determining the vibration fault type as the vibration fault type of the abnormal vibration signal.
EXAMPLE III
Fig. 9 is a schematic structural diagram of an electronic apparatus according to a third embodiment of the present invention, as shown in fig. 9, the electronic apparatus includes a processor 901, a memory 902, an input device 903, and an output device 904; the number of the processors 901 in the electronic device may be one or more, and one processor 901 is taken as an example in fig. 9; the processor 901, the memory 902, the input device 903 and the output device 904 in the electronic apparatus may be connected by a bus or other means, and fig. 9 illustrates an example of connection by a bus.
The memory 902 is used as a computer-readable storage medium and can be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the transformer vibration fault monitoring method in the embodiment of the present invention (for example, the obtaining module 501, the processing module 502, the determining module 503, and the prompting module 504 in the transformer vibration fault monitoring method). The processor 901 executes various functional applications and data processing of the electronic device by running software programs, instructions and modules stored in the memory 902, so as to implement the transformer vibration fault monitoring method described above.
The memory 902 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 902 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory 902 may further include memory located remotely from the processor 901, which may be connected to an electronic device through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The output device 904 may include a display device such as a display screen.
Example four
The fourth embodiment of the invention also provides a storage medium. The storage medium stores instructions, and when the instructions are executed by the processor, the method operations as described above may be executed, and related operations in the transformer vibration fault monitoring method provided by any embodiment of the present invention may also be executed.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only MeMory (ROM), a Random Access MeMory (RAM), a FLASH MeMory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the above search apparatus, each included unit and module are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. A transformer vibration fault monitoring method is characterized by comprising the following steps:
acquiring an abnormal vibration signal of the transformer;
processing the abnormal vibration signal to obtain corresponding characteristic information, wherein the characteristic information is used for reflecting the time frequency of the abnormal vibration signal;
determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model;
and forming fault information of the transformer based on the vibration fault type and carrying out fault prompt.
2. The method according to claim 1, wherein before determining the vibration fault type of the abnormal vibration signal according to the feature information and a pre-trained learning model, the method further comprises:
screening the characteristic information by adopting a self-adaptive enhanced AdaBoost algorithm, and determining first important characteristic information with strong partition capacity;
and updating the feature information, wherein the updated feature information is the first important feature information.
3. The method according to claim 2, wherein the feature information includes M features, M is a positive integer and M > 1, and the screening the feature information by using the adaptive enhanced AdaBoost algorithm to determine the first important feature information with strong partition capability includes:
constructing training data according to the abnormal vibration signal and the normal vibration signal;
initializing a weight distribution of the training data;
determining a strong classifier according to a weak classifier corresponding to each of the M features included in the feature information;
determining all weak classifiers in the process of determining the strong classifiers, and determining the first n weak classifiers as important weak classifiers according to the sequence of the proportion weight values corresponding to the weak classifiers in all the weak classifiers from large to small, wherein n is a positive integer, and M is more than n and is more than or equal to 1;
and determining the features corresponding to the important weak classifiers as the first important feature information, wherein the first important feature information comprises n features.
4. The method of claim 3, wherein the processing the abnormal vibration signal to obtain corresponding characteristic information comprises:
performing target processing on the abnormal vibration signal, and determining characteristic information of the abnormal vibration signal, wherein the target processing comprises at least one of wavelet packet analysis processing, time domain analysis processing or frequency domain analysis processing;
wherein, when the target processing includes the wavelet packet analysis processing, the characteristic information includes energy information of a frequency band of the abnormal vibration signal; when the target process includes the time-domain analysis process, the characteristic information includes at least one of a peak-to-peak value, a standard deviation, a root-mean-square, and a skewness of the abnormal vibration signal; when the target process includes the frequency domain analysis process, the characteristic information includes at least one of a barycentric frequency, a frequency variance, and a mean square frequency of the abnormal vibration signal.
5. The method of any one of claims 1-4, wherein the learning model comprises a self-organizing map (SOM) neural network.
6. The method of claim 5, wherein the learning model training step comprises:
acquiring abnormal vibration signals of the transformer under different vibration fault types;
determining third important characteristic information of the abnormal vibration signals of the transformer under different vibration fault types, wherein the third important characteristic information is determined after processing the abnormal vibration signals of the transformer under different vibration fault types to obtain corresponding characteristic information and screening the obtained corresponding characteristic information by adopting a self-adaptive enhanced AdaBoost algorithm;
inputting the third important feature information to an input layer of the learning model;
determining neurons of an output layer of the learning model to which the third important feature information maps;
the determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model comprises the following steps:
inputting the characteristic information into the learning model, and determining a target neuron of an output layer corresponding to the characteristic information;
inquiring the incidence relation between the neurons and the vibration fault types, and determining the vibration fault types matched with the target neurons;
determining the vibration fault type as a vibration fault type of the abnormal vibration signal.
7. The method of claim 6, wherein the learning model training step further comprises: setting a vibration fault type identifier for the neuron of the output layer according to the vibration fault type corresponding to the third important characteristic information;
the determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model comprises the following steps:
inputting the characteristic information into the learning model, and determining a target neuron of an output layer corresponding to the characteristic information;
and identifying the vibration fault type identified by the vibration fault type identifier of the target neuron, and determining the vibration fault type as the vibration fault type of the abnormal vibration signal.
8. A transformer vibration fault monitoring device, comprising:
the acquisition module is used for acquiring an abnormal vibration signal of the transformer;
the processing module is used for processing the abnormal vibration signal to obtain corresponding characteristic information, and the characteristic information is used for reflecting the time frequency of the abnormal vibration signal;
the determining module is used for determining the vibration fault type of the abnormal vibration signal according to the characteristic information and a pre-trained learning model;
and the prompting module is used for forming fault information of the transformer based on the vibration fault type and performing fault prompting.
9. An electronic device, characterized in that the electronic device comprises:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the transformer vibration fault monitoring method of any one of claims 1-7.
10. A storage medium having stored therein instructions for performing the transformer vibration fault monitoring method of any one of claims 1-7 when executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010591946.5A CN111767675A (en) | 2020-06-24 | 2020-06-24 | Transformer vibration fault monitoring method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010591946.5A CN111767675A (en) | 2020-06-24 | 2020-06-24 | Transformer vibration fault monitoring method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111767675A true CN111767675A (en) | 2020-10-13 |
Family
ID=72721789
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010591946.5A Pending CN111767675A (en) | 2020-06-24 | 2020-06-24 | Transformer vibration fault monitoring method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111767675A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112665710A (en) * | 2020-12-21 | 2021-04-16 | 陕西宝光集团有限公司 | Method and device for detecting running state of equipment, electronic equipment and storage medium |
| CN116277040A (en) * | 2023-05-23 | 2023-06-23 | 佛山隆深机器人有限公司 | Mechanical arm vibration suppression method, device, equipment and medium based on deep learning |
| WO2024000764A1 (en) * | 2022-06-29 | 2024-01-04 | 南方电网科学研究院有限责任公司 | Transformer health state assessment method, apparatus, and device, and readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404901A (en) * | 2015-12-24 | 2016-03-16 | 上海玮舟微电子科技有限公司 | Training method of classifier, image detection method and respective system |
| CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
| CN106441547A (en) * | 2016-08-31 | 2017-02-22 | 许继集团有限公司 | Transformer vibration monitoring method and apparatus |
| CN107560850A (en) * | 2017-08-26 | 2018-01-09 | 中南大学 | Shafting fault recognition method based on Threshold Denoising and AdaBoost |
| CN108229581A (en) * | 2018-01-31 | 2018-06-29 | 西安工程大学 | Based on the Diagnosis Method of Transformer Faults for improving more classification AdaBoost |
| CN109975634A (en) * | 2019-03-04 | 2019-07-05 | 河南理工大学 | A kind of fault diagnostic method for transformer winding based on atom sparse decomposition |
-
2020
- 2020-06-24 CN CN202010591946.5A patent/CN111767675A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404901A (en) * | 2015-12-24 | 2016-03-16 | 上海玮舟微电子科技有限公司 | Training method of classifier, image detection method and respective system |
| CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
| CN106441547A (en) * | 2016-08-31 | 2017-02-22 | 许继集团有限公司 | Transformer vibration monitoring method and apparatus |
| CN107560850A (en) * | 2017-08-26 | 2018-01-09 | 中南大学 | Shafting fault recognition method based on Threshold Denoising and AdaBoost |
| CN108229581A (en) * | 2018-01-31 | 2018-06-29 | 西安工程大学 | Based on the Diagnosis Method of Transformer Faults for improving more classification AdaBoost |
| CN109975634A (en) * | 2019-03-04 | 2019-07-05 | 河南理工大学 | A kind of fault diagnostic method for transformer winding based on atom sparse decomposition |
Non-Patent Citations (3)
| Title |
|---|
| 夏玉剑;李敏;向天堂;秦少鹏;邓权伦;王昕;: "基于SOM的变压器绕组和铁芯故障诊断", 电力科学与技术学报, no. 02 * |
| 汪开正;黄亦翔;张旭东;李彦明;: "基于AdaBoost-SOM方法的电机故障诊断", 机械设计与研究, no. 02, 20 April 2019 (2019-04-20) * |
| 黄新波;王宁;朱永灿;马玉涛;吴明松;: "基于RST-SOM的高压断路器故障诊断", 高压电器, no. 03 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112665710A (en) * | 2020-12-21 | 2021-04-16 | 陕西宝光集团有限公司 | Method and device for detecting running state of equipment, electronic equipment and storage medium |
| WO2024000764A1 (en) * | 2022-06-29 | 2024-01-04 | 南方电网科学研究院有限责任公司 | Transformer health state assessment method, apparatus, and device, and readable storage medium |
| CN116277040A (en) * | 2023-05-23 | 2023-06-23 | 佛山隆深机器人有限公司 | Mechanical arm vibration suppression method, device, equipment and medium based on deep learning |
| CN116277040B (en) * | 2023-05-23 | 2023-07-18 | 佛山隆深机器人有限公司 | Mechanical arm vibration suppression method, device, equipment and medium based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | A novel data-driven approach for transient stability prediction of power systems considering the operational variability | |
| CN111767675A (en) | Transformer vibration fault monitoring method and device, electronic equipment and storage medium | |
| US11927609B2 (en) | Condition monitoring via energy consumption audit in electrical devices and electrical waveform audit in power networks | |
| CN112885372B (en) | Intelligent diagnosis method, system, terminal and medium for power equipment fault sound | |
| CN110213244A (en) | A kind of network inbreak detection method based on space-time characteristic fusion | |
| CN108062572A (en) | A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on DdAE deep learning models | |
| CN110879917A (en) | Electric power system transient stability self-adaptive evaluation method based on transfer learning | |
| Wang et al. | Classification of power quality events using optimal time-frequency representations-part 1: Theory | |
| CN106327357A (en) | Load identification method based on improved probabilistic neural network | |
| CN106628097A (en) | Ship equipment fault diagnosis method based on improved radial basis function neutral network | |
| Goodwin et al. | A pattern recognition approach for peak prediction of electrical consumption | |
| Qiu et al. | Detection of synchrophasor false data injection attack using feature interactive network | |
| CN113538037B (en) | Method, system, equipment and storage medium for monitoring charging event of battery car | |
| CN106650932A (en) | Intelligent fault classification method and device for data center monitoring system | |
| Promper et al. | Anomaly detection in smart grids with imbalanced data methods | |
| Kumar et al. | Cloud-based electricity consumption analysis using neural network | |
| CN114358092B (en) | Method and system for online diagnosis of internal insulation performance of capacitor voltage transformer | |
| Irfan et al. | Energy theft identification using AdaBoost Ensembler in the Smart Grids | |
| Zhou et al. | A method of CNN traffic classification based on sppnet | |
| Ullah et al. | Adaptive data balancing method using stacking ensemble model and its application to non-technical loss detection in smart grids | |
| CN113884807A (en) | Power distribution network fault prediction method based on random forest and multi-layer architecture clustering | |
| CN109711450A (en) | A kind of power grid forecast failure collection prediction technique, device, electronic equipment and storage medium | |
| CN113343123A (en) | Training method and detection method for generating confrontation multiple relation graph network | |
| Tangrand | Some new contributions to neural networks and wavelets with applications | |
| CN116523001A (en) | Method, device and computer equipment for constructing weak line identification model of power grid |
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 |