CN113591792B - Transformer fault identification method based on self-organizing competitive neural network algorithm - Google Patents
Transformer fault identification method based on self-organizing competitive neural network algorithm Download PDFInfo
- Publication number
- CN113591792B CN113591792B CN202110952066.0A CN202110952066A CN113591792B CN 113591792 B CN113591792 B CN 113591792B CN 202110952066 A CN202110952066 A CN 202110952066A CN 113591792 B CN113591792 B CN 113591792B
- Authority
- CN
- China
- Prior art keywords
- transformer
- vibration
- neural network
- signals
- self
- 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
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 24
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 23
- 230000002860 competitive effect Effects 0.000 title claims abstract description 16
- 238000005070 sampling Methods 0.000 claims abstract description 16
- 239000013598 vector Substances 0.000 claims abstract description 8
- 238000006243 chemical reaction Methods 0.000 claims abstract description 7
- 230000001133 acceleration Effects 0.000 claims abstract description 5
- 210000002569 neuron Anatomy 0.000 claims description 25
- 238000004804 winding Methods 0.000 claims description 18
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 4
- 230000002238 attenuated effect Effects 0.000 claims 1
- 238000012549 training Methods 0.000 abstract description 18
- 238000004364 calculation method Methods 0.000 abstract description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 15
- 238000012795 verification Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000005284 excitation Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 229910052742 iron Inorganic materials 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Signal Processing (AREA)
- Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)
Abstract
A transformer fault identification method based on self-organizing competitive neural network algorithm belongs to the technical field of transformers, and comprises the steps of performing time-frequency domain conversion on sampling signals on the basis of on-site sampling of the transformer vibration signals, extracting individual component amplitudes in the frequency domain as vibration acceleration feature vectors, and performing training and identification on the basis through SOM algorithm to achieve the purpose of transformer running state and fault identification based on the vibration signals. The invention has the advantages of science, rationality, reality, effectiveness, accurate calculation, high practical value and the like.
Description
Technical Field
The invention belongs to the technical field of transformers, and particularly relates to a transformer fault identification method based on a self-organizing competitive neural network algorithm, which is applied to power transformer fault running state identification and safety assessment.
Background
Power transformers are extremely important devices in the power grid, playing an important role in the interconnection of various levels of power grids and in the transmission of power. With the continuous construction and perfection of the power grid in China, the current power grid transformation capacity in China reaches 49.4 hundred million kilowatts, the power grid is influenced by the operation working condition and the environment, the fault of the transformer occurs at time, and the disassembly analysis of the transformer after the accident shows that the damage of the iron core and the winding structure caused by the electromagnetic force of the transformer is relatively high. At present, a diagnosis method based on electrical parameters is mostly adopted in a transformer fault identification method, but the electrical parameters are influenced by the running state of the transformer and the external environment, so that the electrical parameters are difficult to effectively and comprehensively reflect the internal faults of the transformer, and the vibration signals contain more internal component change information, so that the method for researching the vibration mechanism of the transformer and combining the vibration signals to realize the transformer running state identification has important significance.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: according to the vibration acquisition parameters of the on-site real-operation transformer, the frequency domain transformation result of the time domain acquisition signals of the transformer is obtained through fast Fourier decomposition, the time-frequency domain characteristics are combined on the basis to form a sample characteristic vector set, the sample characteristic training is carried out by adopting the self-organizing map neural network algorithm, and finally the purpose of transformer fault identification is achieved.
A transformer fault identification method based on a self-organizing competitive neural network algorithm is characterized by comprising the following steps: comprising the following steps, which are sequentially carried out,
step one, fault signal acquisition and preprocessing
Arranging vibration measuring points of a transformer in a corresponding box body in the middle of a transformer winding, collecting original vibration signals of the transformer, preprocessing data of the original vibration signals through fast Fourier decomposition (FFT), and assuming that sampling signals are obtained by linearly superposing a plurality of groups of signals in the FFT principle, wherein the original vibration sampling signals can be expressed as
g(t)＝a 0 +a 1 *t+a 2 *t 2 +...+a n-1 *t n-1
In which a is 0 、a 1 …a n-1 In order for the coefficients of decomposition to be chosen,
by parity arrangement and introduction of twiddle factorsConverting the time domain result of the transformer vibration sampling signal into a frequency domain result through FFT (fast Fourier transform), extracting vibration acceleration amplitude values under each frequency domain component, and forming a vibration signal feature vector;
step two, establishing self-organizing competitive neural network identification
Establishing SOM neural network structure comprising input layer and output layer, wherein the output layer can be expressed as two-dimensional plane array formed by a×b neurons assuming that the number of neurons in the input layer is m, and the Euclidean distance of the jth neuron is calculated as
Wherein w is ij Is the weight between the i-th neuron of the input layer and the j-th neuron of the mapping layer. The method for correcting the weight of the winning neuron in the weight learning process comprises the following steps of
Δw ij ＝w ij (t+1)-w ij (t)＝η(t)(x i (t)-w ij (t))
Wherein η is a constant of 0 to 1 and decays to 0 continuously with time.
In the first step, the original vibration sampling signal is obtained by parity arrangement
In the first step, a twiddle factor is introducedThe converted frequency domain result is
Through the design scheme, the invention has the following beneficial effects: a transformer fault identification method based on self-organizing competitive neural network algorithm is characterized in that on the basis of on-site sampling of a transformer vibration signal, time-frequency domain conversion is carried out on the sampled signal, individual component amplitude values under the frequency domain are extracted to serve as vibration acceleration characteristic vectors, on the basis, training and identification are carried out through SOM algorithm, the purpose of identifying the running state and faults of the transformer based on the vibration signal is achieved, and the method has the advantages of being scientific, reasonable, real, effective, accurate in calculation, high in practical value and the like.
Drawings
The invention is further described with reference to the drawings and detailed description which follow:
fig. 1 is an experimental wiring diagram of a transformer fault identification method based on an ad hoc competitive neural network algorithm.
Fig. 2 is a flowchart of a transformer fault identification method based on an ad hoc competitive neural network algorithm.
Fig. 3 is a topology diagram of a transformer fault identification method SOM algorithm based on a self-organizing competitive neural network algorithm.
Fig. 4 is a diagram of a current characteristic value recognition result of a transformer fault recognition method based on a self-organizing competitive neural network algorithm.
Fig. 5 is a diagram of a vibration characteristic value recognition result of a transformer fault recognition method based on a self-organizing competitive neural network algorithm.
Detailed Description
A transformer fault identification method based on self-organizing competitive neural network algorithm comprises the following steps of 1. Fault signal sampling and preprocessing
According to the existing specifications, the vibration measuring points of the transformer are arranged in the middle of the transformer winding corresponding to the box body, original vibration signals of the transformer are collected, and data preprocessing is carried out on the original vibration signals through fast Fourier decomposition FFT. Assuming that the sampling signal is linearly superimposed by a plurality of groups of signals in the FFT principle, the original vibration sampling signal can be expressed as a form shown in (1)
g(t)＝a 0 +a 1 *t+a 2 *t 2 +...+a n-1 *t n-1 (1)
The following formula is obtained by arranging the formula (1) according to parity
g(t)＝(a 0 +a 2 *t 2 +…+a n-2 *t n-2 )
+(a 1 +a 3 *t 3 +…+a n-1 *t n-1 ) (2)
＝g 1 (t)+g 2 (t)＝g 1 (t 2 )+tg 2 (t 2 )
Introduction of twiddle factorsIs brought into the above way to obtain
And converting the time domain result of the transformer vibration sampling signal into a frequency domain result through FFT (fast Fourier transform), and further extracting the vibration acceleration amplitude under each frequency domain component to form a vibration signal feature vector.
2. Building self-organizing competing neural network identification
Self-organizing feature mapping network (SOM) is a non-guided, self-organizing, self-learning network formed by an array of fully connected neural network elements. The fundamental theory of SOM considers that neurons at different positions in space have different division of labor, when a neural network receives an external input, different reaction regions will react to the external input, and the response characteristics of different reaction regions to the external input are different. Unlike traditional self-organizing networks, SOM networks can learn the distribution characteristics of training input parameters, and can learn the parameter topological structure of training input, so that visual cluster analysis of data is realized.
The SOM neural network structure includes an input layer and a competing layer (output layer), and the competing layer can be expressed as a two-dimensional planar array formed by a×b neurons assuming that the number of neurons in the input layer is m. The basic flow of the SOM algorithm is shown in FIG. 2.
The SOM neural network structure comprises an input layer and a competition layer (output layer), the competition layer can be expressed as a two-dimensional plane array formed by a×b neurons assuming that the number of neurons of the input layer is m, wherein the Euclidean distance of the jth neuron is calculated as follows
Wherein w is ij Is the weight between the i-th neuron of the input layer and the j-th neuron of the mapping layer. The method for correcting the weight of the winning neuron in the weight learning process comprises the following steps of
Δw ij ＝w ij (t+1)-w ij (t)＝η(t)(x i (t)-w ij (t)) (6)
Wherein eta is a constant of 0 to 1 and continuously decays to 0 with time.
3. Experimental analysis
Aiming at a customized three-phase dry-type double-winding experimental transformer (model: SG-10kVA 1.1kV/0.38 kV), a transformer movable mould experimental platform is built, and winding and iron core vibration original data of the transformer in different running states are collected. The transformer parameters are shown in table 1. The experimental platform is shown in fig. 1.
Table 1 experimental transformer parameters
The basic steps of the transformer vibration data acquisition experiment are as follows
1) And (3) equipment connection: the device is connected with an experimental transformer, a current data acquisition module (oscilloscope: lecroy WaveSurfer 4000 HD), a vibration data acquisition module (magnetic vibration pickup, YD-104/10 kHZ) and a voltage control module (voltage regulator), and the connected equipment respectively realizes vibration signal output, current data acquisition, vibration data acquisition and transformer load control.
2) Vibration measuring point arrangement: the vibration measuring points are respectively arranged for the transformer winding and the iron core, the winding vibration measuring points are arranged at the front center positions of the windings of each phase as shown in fig. 4 in consideration of the accuracy of the vibration signal acquisition of the vibration measuring points, and the iron core vibration measuring points are arranged at the center positions of the phases of the upper iron jaw.
3) Setting an operation state: in combination with the actual field operation condition and common faults of the transformer (the applied excitation is the power frequency excitation), the running state of the transformer is set to be 75% of normal load running, three-phase unbalanced running, direct current magnetic bias faults and winding/iron core loosening faults, wherein the running state experiments of the faults are set as follows:
a. three-phase unbalanced operation: the secondary side resistance is adjusted so that the a-phase of the transformer is in an operation state with unbalance rates of alpha=5%, 10% and 15%.
b. Direct current magnetic bias: firstly, measuring the no-load current of the transformer, and setting the direct current injection quantity to be beta=0.5, 1.0 and 2.0 times of the no-load current on the basis of the no-load current, wherein the direct current injection point is the neutral point of the primary side of the transformer.
c. The components are loosened, and the upper and lower fastening bolts and the iron jaw fastening bolts of the transformer winding are respectively adjusted, so that the transformer winding and the iron core are loosened.
4) Data acquisition and processing: winding currents and winding and iron core vibration data in different running states are respectively collected, and vibration signals are decomposed and reconstructed through WPT, so that vibration signal feature vectors are further obtained.
The original vibration signal and FFT result of the transformer are obtained, the original sampling signal of the transformer has certain periodicity, but the original waveform is complex, the effective identification of the running state of the transformer is difficult to be carried out directly through the original vibration signal, the state characteristics of the time domain result after FFT conversion are obvious, the vibration signal is concentrated at 100Hz and frequency multiplication thereof and is influenced by magnetic leakage, the FFT frequency domain result contains certain 50Hz frequency multiplication component, the FFT conversion result is extracted as the vibration characteristics, a vibration characteristic sample set is formed, and SOM training and identification are carried out.
100 groups of samples of vibration data under 5 running states (normal running, three-phase imbalance, direct current magnetic bias, winding loosening and iron core loosening) are selected, wherein the total number of the samples is 500, 90 groups of samples are randomly selected as training sample sets (total 450 groups) for each of the 5 running states, and the rest 50 groups of samples are verification sample sets for each of the 5 running states.
In the aspect of selecting vibration characteristic parameters, the three-phase unbalance and winding loosening state select winding vibration characteristic parameters as identification parameters; selecting the vibration characteristic parameters of the iron core as identification parameters by DC magnetic bias and iron core looseness; and selecting the vibration characteristic parameters of the winding and the iron core in the normal running state as identification parameters, and forming a current-vibration characteristic parameter set on the basis.
The FFT result extracts 16 groups of characteristic parameters, and the number of neurons of a mapping layer is set to be 20 in consideration of the fact that the number of the neurons of the mapping layer of the SOM algorithm is required to be larger than the input characteristic dimension. Considering the uniformity of the training process of the characteristic parameter sets of different running states, the training times are respectively set to be 10 times, 20 times, 30 times, 50 times, 200 times, 500 times and 1000 times, and the proper training times are determined by taking the DC magnetic bias state (63 # sample) as a verification sample. As shown in table 2, the results are different training times.
After 1000 training, each state was divided into 5 regions, indicating that the training set was correctly divided.
The verification set samples are effectively identified, and the dead node problem does not exist.
TABLE 2SOM identification results
As can be seen from table 2, the overlapping condition of the classification results of each running state appears when the training times are 10, 20, 30 and 50, which indicates that the training times are insufficient, and when the training times are 200, the overlapping problem does not occur in each running state classification, which indicates that the training times are sufficient. When the training time is 1000 times, the number of the neuron of the verification vector output layer is 40, which is consistent with the number of the neuron of the output layer direct current magnetic bias state, so that the SOM algorithm can identify the running state of the transformer.
Further, the conventional current characteristic parameter recognition accuracy and vibration characteristic parameter recognition accuracy are compared and analyzed, wherein the recognition accuracy is the ratio of the number of the recognition correct samples to the number of the total verification samples, and the results shown in fig. 4 and 5 are obtained
According to the SOM algorithm current characteristic value recognition result, as can be seen from FIG. 4, 17 groups of verification samples only adopt 50 groups of current characteristic parameters, the recognition accuracy is 66%, the recognition accuracy of normal operation and three-phase unbalanced operation states is higher, the recognition accuracy of direct current magnetic bias and component loosening states is lower, and the main reasons of the recognition accuracy are analyzed, so that the port current change is smaller in the direct current magnetic bias and component loosening states, and therefore, the effective recognition of the internal fault problem of the transformer is difficult to realize through the traditional electrical parameter recognition method.
As can be seen from the SOM algorithm vibration characteristic value recognition result figure 5, 2 groups of recognition errors in 50 groups of verification samples adopting global characteristic parameters have the recognition accuracy of 96 percent, and the recognition accuracy is far higher than that of the detection samples adopting only current characteristic parameters, so that the detection samples adopting the global characteristic parameters can be used for effectively recognizing various faults of the transformer.
According to the transformer fault identification method based on the self-organizing competitive neural network algorithm, experimental analysis results show that various running states and faults of the transformer can be accurately identified by using the method, and the purposes of the invention are achieved and the effects are achieved.
The computing conditions, illustrations, etc. in the embodiments of the invention are provided for further illustration and are not intended to be exhaustive, and do not limit the scope of the claims, and other substantially equivalent substitutions will occur to those skilled in the art without inventive labor from the teachings of the examples of the invention, and are within the scope of the invention.
Claims (1)
1. A transformer fault identification method based on a self-organizing competitive neural network algorithm is characterized by comprising the following steps: comprising the following steps, which are sequentially carried out,
step one, fault signal acquisition and preprocessing
Arranging vibration measuring points of a transformer in a corresponding box body in the middle of a transformer winding, collecting original vibration signals of the transformer, preprocessing data of the original vibration signals through fast Fourier decomposition (FFT), and assuming that sampling signals are obtained by linearly superposing a plurality of groups of signals in the FFT principle, wherein the original vibration sampling signals can be expressed as
g(t)＝a 0 +a 1 *t+a 2 *t 2 +…+a n-1 *t n-1
In which a is 0 、a 1 …a n-1 In order for the coefficients of decomposition to be chosen,
by parity arrangement and introduction of twiddle factorsk is smaller than 0.5n, converting the time domain result of the transformer vibration sampling signal into a frequency domain result through FFT conversion, extracting vibration acceleration amplitude under each frequency domain component, and forming a vibration signal feature vector;
step two, establishing self-organizing competitive neural network identification
Establishing SOM neural network structure comprising input layer and output layer, wherein the output layer can be expressed as two-dimensional plane array formed by a×b neurons assuming that the number of neurons in the input layer is m, and the Euclidean distance of the jth neuron is calculated as
Wherein w is ij To input the weight between the ith neuron of the layer and the jth neuron of the mapping layer, the weight of the winning neuron is corrected in the weight learning process,the method is as follows
Δw ij ＝w ij (t+1)-w ij (t)＝η(t)(x i (t)-w ij (t))
Wherein eta is a constant of 0 to 1 and is continuously attenuated to 0 along with time;
in the first step, the original vibration sampling signal is obtained by parity arrangement
g(t)＝(a 0 +a 2 *t 2 +…+a n-2 *t n-2 )
+(a 1 +a 3 *t 3 +…+a n-1 *t n-1 )
＝g 1 (t)+g 2 (t)＝g 1 (t 2 )+tg 2 (t 2 )
In the first step, a twiddle factor is introducedk is less than 0.5n, and the converted frequency domain result is
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110952066.0A CN113591792B (en) | 2021-08-19 | 2021-08-19 | Transformer fault identification method based on self-organizing competitive neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110952066.0A CN113591792B (en) | 2021-08-19 | 2021-08-19 | Transformer fault identification method based on self-organizing competitive neural network algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113591792A CN113591792A (en) | 2021-11-02 |
| CN113591792B true CN113591792B (en) | 2023-11-28 |
Family
ID=78238498
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110952066.0A Active CN113591792B (en) | 2021-08-19 | 2021-08-19 | Transformer fault identification method based on self-organizing competitive neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113591792B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115222111A (en) * | 2022-06-29 | 2022-10-21 | 南方电网科学研究院有限责任公司 | Transformer health state evaluation method, device, equipment and readable storage medium |
| CN116092524A (en) * | 2022-12-27 | 2023-05-09 | 南方电网调峰调频发电有限公司储能科研院 | Electrical equipment sound identification method based on characteristic frequency band selection |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008071255A (en) * | 2006-09-15 | 2008-03-27 | Saga Univ | Pattern generating method using self-organizing map, its program and device |
| CN104330255A (en) * | 2014-11-04 | 2015-02-04 | 徐州隆安光电科技有限公司 | Gear fault diagnosing method based on multi-sensor information fusion |
| CN109829538A (en) * | 2019-02-28 | 2019-05-31 | 苏州热工研究院有限公司 | A kind of equipment health Evaluation method and apparatus based on deep neural network |
| CN110443302A (en) * | 2019-08-02 | 2019-11-12 | 天津相和电气科技有限公司 | Load discrimination method and its application based on Fusion Features and deep learning |
| CN112345252A (en) * | 2020-11-22 | 2021-02-09 | 国家电网有限公司 | Rolling bearing fault diagnosis method based on EEMD and improved GSA-SOM neural network |
| CN112734001A (en) * | 2020-12-09 | 2021-04-30 | 东南大学 | Wind power transmission chain intelligent fault diagnosis method based on order spectrum migration |
-
2021
- 2021-08-19 CN CN202110952066.0A patent/CN113591792B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008071255A (en) * | 2006-09-15 | 2008-03-27 | Saga Univ | Pattern generating method using self-organizing map, its program and device |
| CN104330255A (en) * | 2014-11-04 | 2015-02-04 | 徐州隆安光电科技有限公司 | Gear fault diagnosing method based on multi-sensor information fusion |
| CN109829538A (en) * | 2019-02-28 | 2019-05-31 | 苏州热工研究院有限公司 | A kind of equipment health Evaluation method and apparatus based on deep neural network |
| CN110443302A (en) * | 2019-08-02 | 2019-11-12 | 天津相和电气科技有限公司 | Load discrimination method and its application based on Fusion Features and deep learning |
| CN112345252A (en) * | 2020-11-22 | 2021-02-09 | 国家电网有限公司 | Rolling bearing fault diagnosis method based on EEMD and improved GSA-SOM neural network |
| CN112734001A (en) * | 2020-12-09 | 2021-04-30 | 东南大学 | Wind power transmission chain intelligent fault diagnosis method based on order spectrum migration |
Non-Patent Citations (1)
| Title |
|---|
| 基于自组织特征映射网络的配电网故障类型识别;赵智;任桂山;陈津刚;陈学梅;;自动化技术与应用(第09期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113591792A (en) | 2021-11-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113591792B (en) | Transformer fault identification method based on self-organizing competitive neural network algorithm | |
| CN109948597B (en) | High-voltage circuit breaker mechanical fault diagnosis method | |
| CN106443316B (en) | Multi-information detection method and device for deformation state of power transformer winding | |
| CN112578312B (en) | Micro-grid fault diagnosis method based on whale algorithm optimization extreme learning machine | |
| CN109344918B (en) | Big data power distribution network fault line selection analysis method based on improved particle swarm optimization | |
| CN109142851A (en) | A kind of novel power distribution network internal overvoltage recognition methods | |
| CN110703151A (en) | Transformer fault diagnosis method based on vibration blind source separation and Bayesian model | |
| Xiao et al. | Multifeature extraction and semi-supervised deep learning scheme for state diagnosis of converter transformer | |
| CN112114232A (en) | Monopole earth fault distance measuring method based on wavelet decomposition and DBN | |
| CN115267428B (en) | LCC-MMC monopole ground fault positioning method based on VMD-ET feature selection | |
| CN116089857A (en) | Transformer fault identification method based on CEEMDAN-DBN | |
| CN111723684A (en) | Method for identifying transient overvoltage type in offshore wind farm | |
| CN112083328A (en) | Fault diagnosis method, system and device for high-voltage circuit breaker | |
| CN110174255A (en) | A kind of transformer vibration signal separation method based on radial base neural net | |
| CN110991481A (en) | High-voltage shunt reactor internal loosening fault diagnosis method based on cross wavelet transformation | |
| CN112130026A (en) | MMC-HVDC single-pole grounding fault positioning method based on wavelet packet energy entropy and DBN | |
| CN111157843A (en) | Power distribution network line selection method based on time-frequency domain traveling wave information | |
| CN116992326A (en) | Transformer fault identification method based on wavelet packet-mapping network algorithm | |
| CN112100923A (en) | State evaluation method for frequency converter IGBT of full-power generation system | |
| CN113991652B (en) | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG | |
| CN113092934B (en) | Single-phase earth fault judgment method and system based on clustering and LSTM | |
| CN115577584A (en) | Dry-type transformer winding state identification method based on time-frequency analysis and multi-field coupling | |
| CN114091534A (en) | Winding vibration characteristic identification method based on wavelet packet scale-energy ratio | |
| CN114186590A (en) | Power distribution network single-phase earth fault identification method based on wavelet and deep learning | |
| CN114814391A (en) | Charging pile fault identification method and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |