CN117786583A - Fault diagnosis method and device based on variational modal decomposition and computer device - Google Patents
Fault diagnosis method and device based on variational modal decomposition and computer device Download PDFInfo
- Publication number
- CN117786583A CN117786583A CN202410207668.7A CN202410207668A CN117786583A CN 117786583 A CN117786583 A CN 117786583A CN 202410207668 A CN202410207668 A CN 202410207668A CN 117786583 A CN117786583 A CN 117786583A
- Authority
- CN
- China
- Prior art keywords
- fault
- signal
- time window
- signals
- clustering
- 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 41
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 40
- 238000003745 diagnosis Methods 0.000 title claims abstract description 32
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000004364 calculation method Methods 0.000 claims description 20
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000007621 cluster analysis Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000489 vacuum metal deposition Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method and device based on variation modal decomposition and a computer device, which realize accurate diagnosis of high-speed aircraft sensor faults in a strong noise environment. The fault diagnosis method based on the variation mode decomposition comprises the steps of collecting normal data signals, decomposing the collected normal data signals into a plurality of sub-modes through the variation mode decomposition, selecting an effective sub-mode for reconstruction, wherein the effective sub-mode refers to that the energy of an effective signal of the sub-mode is larger than that of noise, decomposing a reconstructed characteristic signal through a time window, training a self-encoder through the obtained time window data, identifying fault signals through the trained self-encoder, and clustering the identified fault signals through a self-adaptive density peak clustering algorithm. The method is suitable for fault diagnosis of the high-speed aircraft sensor.
Description
Technical Field
The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method and device based on variation modal decomposition and a computer device.
Background
The monitoring of various indexes and the control data of the performance of the high-speed aircraft in the flight process thereof all need the sensors to provide data. In the hypersonic flight process of an aircraft, physical phenomena such as a thin shock wave layer, an entropy layer, high-temperature gas and the like are often generated, and factors such as environment variation, electromagnetic communication, vibration, overweight and weightlessness exist, so that the fault rate of a sensor can be increased. The sensor fault of the aircraft has light influence, the deviation is measured, and the safety accident is caused by heavy influence. Therefore, the fault diagnosis of the sensor has important significance in the aspect of avoiding accidents. However, a strong noise environment tends to present a significant challenge to this.
The working environment of the sensor of the high-speed aircraft is a strong noise environment, the needed key signals are usually covered by noise, and common signal processing modes generally include wavelet transformation, empirical mode decomposition and the like. However, the wavelet transform requires pre-assignment of an optimal wavelet base, which may affect the decomposition result, and it is also limited by a hard band limitation, a poor time-frequency resolution, or the presence of cross terms. The lack of mathematical rigor of empirical mode decomposition has the inevitable drawbacks of end effects, mode mixing and limitation of noise sensitivity, while the method is recursive, which can lead to error propagation from one mode to another and eventually distortion of the decomposed signal.
At present, the fault diagnosis method mainly comprises two modes. One model-driven way is to analyze the environment of the diagnostic object, consider the influence of various stresses, and build a mathematical interpretation model, which often has a strict mathematical interpretation, but some scenes are not easy to detect, and building the model according to the environment is difficult. The other mode is a data driving mode, analysis is usually carried out according to collected data by using a machine learning mode, the mode focuses more on the collected data, a specific mathematical model is not required to be built according to the environment, the method is flexible, the method can be used in various situations, and the method is also used for researching the relatively popular direction of fault diagnosis.
The common data-driven fault diagnosis method adopts supervised learning, which often requires collecting a large amount of fault sample data for marking, which is often difficult for high-speed aircrafts, and marking the fault sample data also requires a certain priori knowledge, which increases the labor cost. At the same time, there is a certain calculation force requirement for processing data by building a depth network, and the internal calculation force of the high-speed aircraft is often tense, which is generally not preferable.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a fault diagnosis method and device based on variation modal decomposition and a computer device, which realize accurate diagnosis of high-speed aircraft sensor faults in a strong noise environment.
The invention adopts the following technical scheme to achieve the aim, and in a first aspect, the invention provides a fault diagnosis method based on variational modal decomposition, which comprises the following steps:
s1, collecting normal data signals;
s2, decomposing the collected normal data signals into a plurality of sub-modes through variation mode decomposition, and selecting an effective sub-mode for reconstruction, wherein the effective sub-mode means that the energy of the effective signal of the sub-mode is larger than that of noise;
s3, decomposing the reconstructed characteristic signals by using a time window;
s4, training the self-encoder through the acquired time window data, and identifying fault signals through the trained self-encoder;
s5, clustering the identified fault signals through a self-adaptive density peak clustering algorithm.
Further, the decomposing the reconstructed characteristic signal by using the time window specifically includes:
setting the length of a time window as L, setting the moving step length of the time window as S, moving the time window from the beginning of a signal to the end of the signal according to the designated step length, detecting the data of each window in the moving process, and storing the data of the time window obtained by detection.
Further, training the self-encoder by the acquired time window data specifically includes:
the acquired time window data is input from the encoder for encoding to obtain an intermediate code, and then the intermediate code is input from a decoder of the encoder, which decodes the output signal into an input signal.
Further, clustering the identified fault signals by the adaptive density peak clustering algorithm specifically comprises:
when the fault signals are identified, extracting time period signals with faults, extracting the time period signals with all the faults, and carrying out cluster analysis on the fault signals through a self-adaptive density peak value clustering algorithm, wherein the cluster calculation process is as follows:
inputting fault signal data setsThe number of samples isn;
Initializing a cutoff distanceCalculating local Density +.>And relative distance->;
Changing initialization cut-off distanceCalculate the correspondingGObtaining the minimumGCorresponding->And calculates +.>And,
;
calculation acquisitionAnd sort in descending order,/->;
Calculating the maximumTWhereincFor real numbers greater than 0, ensuring that the denominator is not 0, selectingAnd ratio->Large points as cluster center points, +.>;
According to the calculation resultClustering is carried out with a clustering center point to obtain a clustering result +.>The number of clusters isk。
In a second aspect, the present invention provides a fault diagnosis device based on decomposition of a variation mode, for implementing the fault diagnosis method based on decomposition of a variation mode, where the device includes:
the data acquisition module acquires normal data signals;
the system comprises a variation mode decomposition module, a conversion mode analysis module and a data processing module, wherein the variation mode decomposition module decomposes the collected normal data signals into a plurality of sub-modes, and selects an effective sub-mode for reconstruction, wherein the effective sub-mode refers to the sub-mode that the energy of an effective signal is larger than that of noise;
the decomposition module is used for decomposing the reconstructed characteristic signals by using a time window;
the self-encoder module trains the self-encoder through the acquired time window data and recognizes fault signals through the trained self-encoder;
and the clustering module clusters the identified fault signals through a self-adaptive density peak clustering algorithm.
Further, the decomposition module is specifically configured to set the length of the time window to L, and the moving step length of the time window to S, where the time window moves from the beginning of the signal to the ending of the signal according to the specified step length, and in the moving process, the data of each window needs to be detected, and the data of the time window obtained by detection is stored.
Further, the self-encoder module is specifically configured to input the obtained time window data from the encoder for encoding, obtain an intermediate code, and then input the intermediate code to a decoder of the encoder, where the decoder decodes the output signal into the input signal.
Further, the clustering module is specifically configured to extract a time period signal when a fault signal is identified, extract time period signals of all faults, and perform cluster analysis on the fault signal through an adaptive density peak clustering algorithm, where a cluster calculation process is as follows:
inputting fault signal data setsThe number of samples isn;
Initializing a cutoff distanceCalculating local Density +.>And relative distance->;
Changing initialization cut-off distanceCalculate the correspondingGObtaining the minimumGCorresponding->And calculates +.>And,
;
calculation acquisitionAnd sort in descending order,/->;
Calculating the maximumTWhereincFor real numbers greater than 0, ensuring that the denominator is not 0, selectingAnd ratio->Large points as cluster center points, +.>;
According to the calculation resultClustering is carried out with a clustering center point to obtain a clustering result +.>The number of clusters isk。
In a third aspect, the present invention provides a computer apparatus comprising a memory storing program instructions that, when executed, perform a method of fault diagnosis based on a variant mode decomposition as described above.
The beneficial effects of the invention are as follows:
the invention uses VMD (Variational Mode Decomposition, variational modal decomposition) to process the signal aiming at fault signal diagnosis in a strong noise environment, thereby improving the accuracy of effective signal extraction.
According to the invention, aiming at the situation that fault data are difficult to acquire, only the normal signal is used for training the encoder, so that the fault signal can be accurately identified, and the cost of data acquisition and artificial labeling is saved.
According to the method, aiming at the acquisition of the fault signal, the time period of the fault signal is acquired by using a sliding time window mode, so that the information generated by the fault signal is acquired, and data support is provided for fault analysis.
According to the invention, ADPC (Adaptive Density Peak Clustering, self-adaptive density peak clustering algorithm) is used for analyzing fault signals, so that the fault signals of the same type can be clustered, and a reference is provided for generating and judging the fault type.
The invention abandons the scheme of the deep neural network, uses a machine learning algorithm with relatively small calculation force requirement, and improves the practicability.
Drawings
FIG. 1 is a flow chart of a fault diagnosis method based on variation modal decomposition provided by an embodiment of the invention;
FIG. 2 is a flow chart of decomposition and reconstruction of a variation modality provided by an embodiment of the present invention;
FIG. 3 is an exploded view of a reconstructed signature signal using a time window provided by an embodiment of the present invention;
FIG. 4 is a self-encoder training flowchart provided by an embodiment of the present invention;
FIG. 5 is a schematic diagram of clustering fault signals using an ADPC algorithm provided by an embodiment of the present invention;
fig. 6 is a schematic diagram of a clustering result provided by an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Aiming at the problem that the common signal processing method has poor processing effect on strong noise, the invention adopts the variation mode decomposition to reduce the end effect, so that the modes are mixed, the related frequency band is determined in a self-adaptive way, and the corresponding modes are estimated at the same time, thereby properly balancing the error between the modes. VMDs, on the other hand, are generalizing classical wiener filters to several adaptive bands. Thus, the VMD is particularly capable of solving noise present in an input signal, and has a certain degree of superiority in processing noise.
Aiming at the problem that the fault sample data is difficult to collect in a large quantity, the collection of the normal data is relatively easier, the invention uses the large quantity of the collected normal data, trains the normal characteristic signals to obtain the self-encoder, and obtains a large difference from the normal signal coding result after the fault signals enter the self-encoder so as to identify the fault signals.
Considering the situation that the calculation force of the aircraft is limited, fault classification is carried out in a frame mode without introducing a deep neural network, and an adaptive density peak clustering algorithm in a machine learning model is used, wherein the algorithm can adaptively select a clustering center and a cutting distance, can well obtain a clustering effect under the condition that noise exists, and has low requirement on the calculation force. The clustering result is obtained through an algorithm, so that the type of the fault can be obtained, the existence of the fault can be judged, and a basis is provided for analyzing the cause of the fault.
Based on the above principle, the present invention provides a fault diagnosis method based on decomposition of variation modes, the overall flow of which is shown in fig. 1, wherein a large amount of normal data is collected first to prepare a training model, the normal data contains a large amount of noise, the collected signal data is decomposed into a plurality of sub-modes by using VMD, then effective sub-modes are selected, and the sub-modes are reconstructed by using energy ratio criteria. Since it is necessary to analyze the cause of the occurrence of the fault, it is necessary to detect the time when the fault exists. Then, data is acquired through a sliding window, the acquired data is input into a self-encoder fault detector, and the existence time of the fault is detected. Finally, different types of faults are extracted from the signal by ADPC for subsequent analysis.
Specifically, as shown in fig. 2, processing the signal with noise using the VMD specifically includes: the VMD may adaptively determine the relevant frequency bands and estimate the corresponding modes, so that multiple sub-modes may be automatically decomposed for different input signals. In selecting an active sub-mode, the energy of the active signal is typically greater than the energy of the noise. Therefore, higher energy subcomponents are more likely to have significant signal components. To this end, the sub-mode with the three highest energy ratios will be used as the active sub-mode to reconstruct the filtered signal.
Specifically, as shown in fig. 3, the decomposing the reconstructed characteristic signal using the time window specifically includes: setting the length of a time window as L, setting the moving step length of the time window as S, and moving the time window from the beginning of a signal to the end of the signal according to the designated step length, wherein the data of each window need to be detected in the moving process. The acquired data of the time window is reserved, and data preparation is performed for the training of AE (auto encoder).
Specifically, as shown in fig. 4, the time window data of the normal signal data acquired through the self-encoder training specifically includes: AE is implemented by a neural network by encoding an input signal to obtain an intermediate code, and then decoding the output signal as much as possible by a decoder into the input signal, where the number of layers of the network is designed to be relatively simple in view of the computational problem.
In the invention, after the AE is trained by using the obtained normal signal time window data, the detection encoder which obtains the correct data can identify the fault signal by using the AE. The specific identification process is as follows: if a fault signal is input, the residual error between the decoding result of AE and the input signal is greatly different from the preset residual error value in training, so that the signal can be judged as the fault signal. The identification step may be performed while the data is acquired in real time on the aircraft, and may be immediately identified upon failure.
Specifically, as shown in fig. 5 and 6, clustering fault signals using ADPC specifically includes: in the process of identifying fault signals by using an encoder, the occurrence time period of the fault signals can be obtained, when the fault signals are identified, the occurrence time period signals of the faults are extracted, after all the fault signals are extracted, the fault signals are subjected to cluster analysis by using an ADPC, so that the fault signals are classified, and the classification result can be expressed as: fault 1, fault 2, fault 3.
Calculation flow of ADPC: inputting a failure time period signal datasetThe number of samples isn;
Initializing a cutoff distanceCalculating local Density +.>And relative distance->;
Changing initialization cut-off distanceCalculate the correspondingGObtaining the minimumGCorresponding->And calculates +.>And,
;
calculation acquisitionAnd sort in descending order,/->;
Calculating the maximumTWhereincFor real numbers greater than 0, ensuring that the denominator is not 0, selectingAnd ratio->Large points as cluster center points, +.>;
According to the calculation resultClustering is carried out with a clustering center point to obtain a clustering result +.>The number of clusters isk。
The density peak clustering algorithm is a density-based clustering algorithm, the basic idea of which is that the cluster center has the highest density value in its neighborhood and is farther from other cluster centers. However, the method relies on manual selection of the clustering center and the cut-off distance, which is time-consuming, and the clustering effect is greatly reduced due to improper selection. Therefore, the invention uses the ADPC algorithm, and the algorithm can adaptively select the clustering center and the cut-off distance according to the actual condition of the data, thereby improving the efficiency and the accuracy.
The fault information can be initially classified after being analyzed. If a certain type of signal is classified as belonging to a certain fault, the fault type can be immediately judged, if the signal is not classified as belonging to the fault, the fault type can be further judged and analyzed according to the time and environment information of the fault after the flight is finished.
The foregoing is merely a preferred embodiment of the invention, and it is to be understood that the invention is not limited to the form disclosed herein but is not to be construed as excluding other embodiments, but is capable of numerous other combinations, modifications and environments and is capable of modifications within the scope of the inventive concept, either as taught or as a matter of routine skill or knowledge in the relevant art. And that modifications and variations which do not depart from the spirit and scope of the invention are intended to be within the scope of the appended claims.
Claims (9)
1. A method for fault diagnosis based on decomposition of a variant mode, the method comprising:
s1, collecting normal data signals;
s2, decomposing the collected normal data signals into a plurality of sub-modes through variation mode decomposition, and selecting an effective sub-mode for reconstruction, wherein the effective sub-mode means that the energy of the effective signal of the sub-mode is larger than that of noise;
s3, decomposing the reconstructed characteristic signals by using a time window;
s4, training the self-encoder through the acquired time window data, and identifying fault signals through the trained self-encoder;
s5, clustering the identified fault signals through a self-adaptive density peak clustering algorithm.
2. The fault diagnosis method based on variation modal decomposition according to claim 1, wherein decomposing the reconstructed characteristic signal using a time window specifically includes:
setting the length of a time window as L, setting the moving step length of the time window as S, moving the time window from the beginning of a signal to the end of the signal according to the designated step length, detecting the data of each window in the moving process, and storing the data of the time window obtained by detection.
3. The method for fault diagnosis based on decomposition of variation modalities according to claim 1, wherein training the self-encoder by the acquired time window data specifically comprises:
the acquired time window data is input from the encoder for encoding to obtain an intermediate code, and then the intermediate code is input from a decoder of the encoder, which decodes the output signal into an input signal.
4. The fault diagnosis method based on variation modal decomposition according to claim 1, wherein clustering the identified fault signals by an adaptive density peak clustering algorithm specifically comprises:
when the fault signals are identified, extracting time period signals with faults, extracting the time period signals with all the faults, and carrying out cluster analysis on the fault signals through a self-adaptive density peak value clustering algorithm, wherein the cluster calculation process is as follows:
inputting fault signal data setsThe number of samples isn;
Initializing a cutoff distanceCalculating local Density +.>And relative distance->;
Changing initialization cut-off distanceCalculate the correspondingGObtaining the minimumGCorresponding->And calculates +.>And->,
;
Calculation acquisitionAnd sort in descending order,/->;
Calculating the maximumTWhereincFor real numbers greater than 0, ensuring that the denominator is not 0, selectingAnd ratio->Large points as cluster center points, +.>;
According to the calculation resultClustering is carried out with a clustering center point to obtain a clustering result +.>Clustering individualThe number isk。
5. A fault diagnosis device based on a variant mode decomposition for implementing the fault diagnosis method based on a variant mode decomposition according to any one of claims 1 to 4, characterized in that the device comprises:
the data acquisition module acquires normal data signals;
the system comprises a variation mode decomposition module, a conversion mode analysis module and a data processing module, wherein the variation mode decomposition module decomposes the collected normal data signals into a plurality of sub-modes, and selects an effective sub-mode for reconstruction, wherein the effective sub-mode refers to the sub-mode that the energy of an effective signal is larger than that of noise;
the decomposition module is used for decomposing the reconstructed characteristic signals by using a time window;
the self-encoder module trains the self-encoder through the acquired time window data and recognizes fault signals through the trained self-encoder;
and the clustering module clusters the identified fault signals through a self-adaptive density peak clustering algorithm.
6. The fault diagnosis device based on variation modal decomposition according to claim 5, wherein the decomposition module is specifically configured to set the length of the time window to L, the moving step length of the time window to S, and the time window moves from the beginning of the signal to the end of the signal according to the designated step length, and the data of each window need to be detected during the moving process, and save the data of the time window obtained by the detection.
7. The fault diagnosis device based on decomposition of variation modes according to claim 5, wherein the self-encoder module is specifically configured to input the obtained time window data from the encoder for encoding, obtain an intermediate code, and then input the intermediate code to a decoder of the encoder, and the decoder decodes the output signal into the input signal.
8. The fault diagnosis device based on variation modal decomposition according to claim 5, wherein the clustering module is specifically configured to extract a time period signal when a fault occurs when the fault signal is identified, and perform a cluster analysis on the fault signal through an adaptive density peak clustering algorithm after extracting all the time period signals of the fault, where the cluster calculation process is as follows:
inputting fault signal data setsThe number of samples isn;
Initializing a cutoff distanceCalculating local Density +.>And relative distance->;
Changing initialization cut-off distanceCalculate the correspondingGObtaining the minimumGCorrespondingly, and calculate +.>And->,
;
Calculation acquisitionAnd sort in descending order,/->;
Calculating the maximumTWhereincFor real numbers greater than 0, ensuring that the denominator is not 0, selectingAnd ratio->Large points as cluster center points, +.>;
According to the calculation resultClustering is carried out with a clustering center point to obtain a clustering result +.>The number of clusters isk。
9. A computer device comprising a memory storing program instructions that, when executed, perform the method of fault diagnosis based on decomposition of a variant mode as claimed in any one of claims 1 to 4.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410207668.7A CN117786583A (en) | 2024-02-26 | 2024-02-26 | Fault diagnosis method and device based on variational modal decomposition and computer device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410207668.7A CN117786583A (en) | 2024-02-26 | 2024-02-26 | Fault diagnosis method and device based on variational modal decomposition and computer device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117786583A true CN117786583A (en) | 2024-03-29 |
Family
ID=90382015
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410207668.7A Pending CN117786583A (en) | 2024-02-26 | 2024-02-26 | Fault diagnosis method and device based on variational modal decomposition and computer device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117786583A (en) |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109409400A (en) * | 2018-08-28 | 2019-03-01 | 西安电子科技大学 | Merge density peaks clustering method, image segmentation system based on k nearest neighbor and multiclass |
CN113177537A (en) * | 2021-06-29 | 2021-07-27 | 湖北博华自动化***工程有限公司 | Fault diagnosis method and system for rotary mechanical equipment |
CN113255437A (en) * | 2021-04-12 | 2021-08-13 | 中国民航大学 | Fault diagnosis method for deep convolution sparse automatic encoder of rolling bearing |
CN113670605A (en) * | 2021-07-02 | 2021-11-19 | 中国地质大学(武汉) | VGSET time-frequency analysis method, device, equipment and storage medium for bearing fault signals |
CN113887398A (en) * | 2021-09-29 | 2022-01-04 | 中南大学 | GPR signal denoising method based on variational modal decomposition and singular spectrum analysis |
CN114964778A (en) * | 2022-05-13 | 2022-08-30 | 重庆理工大学 | Bearing fault diagnosis method based on wavelet time-frequency graph and deep learning |
CN115659255A (en) * | 2022-10-14 | 2023-01-31 | 桂林海威科技股份有限公司 | Street lamp fault diagnosis method based on time-frequency domain modeling and feature selection |
CN115905832A (en) * | 2022-10-24 | 2023-04-04 | 广西电网有限责任公司河池供电局 | Transformer fault diagnosis method based on vibration signal clustering identification |
CN116912522A (en) * | 2023-07-24 | 2023-10-20 | 吉林大学 | Self-adaptive density peak clustering band selection method |
CN116976857A (en) * | 2023-08-08 | 2023-10-31 | 中国铁建电气化局集团北方工程有限公司 | Predictive maintenance method and system for railway power equipment based on deep learning |
CN117172589A (en) * | 2023-08-30 | 2023-12-05 | 浙江大学 | Industry typical user power demand response potential evaluation method |
CN117574187A (en) * | 2023-11-09 | 2024-02-20 | 西安电子科技大学 | Multi-radiation-source signal sorting time delay estimation method and system |
-
2024
- 2024-02-26 CN CN202410207668.7A patent/CN117786583A/en active Pending
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109409400A (en) * | 2018-08-28 | 2019-03-01 | 西安电子科技大学 | Merge density peaks clustering method, image segmentation system based on k nearest neighbor and multiclass |
CN113255437A (en) * | 2021-04-12 | 2021-08-13 | 中国民航大学 | Fault diagnosis method for deep convolution sparse automatic encoder of rolling bearing |
CN113177537A (en) * | 2021-06-29 | 2021-07-27 | 湖北博华自动化***工程有限公司 | Fault diagnosis method and system for rotary mechanical equipment |
CN113670605A (en) * | 2021-07-02 | 2021-11-19 | 中国地质大学(武汉) | VGSET time-frequency analysis method, device, equipment and storage medium for bearing fault signals |
CN113887398A (en) * | 2021-09-29 | 2022-01-04 | 中南大学 | GPR signal denoising method based on variational modal decomposition and singular spectrum analysis |
CN114964778A (en) * | 2022-05-13 | 2022-08-30 | 重庆理工大学 | Bearing fault diagnosis method based on wavelet time-frequency graph and deep learning |
CN115659255A (en) * | 2022-10-14 | 2023-01-31 | 桂林海威科技股份有限公司 | Street lamp fault diagnosis method based on time-frequency domain modeling and feature selection |
CN115905832A (en) * | 2022-10-24 | 2023-04-04 | 广西电网有限责任公司河池供电局 | Transformer fault diagnosis method based on vibration signal clustering identification |
CN116912522A (en) * | 2023-07-24 | 2023-10-20 | 吉林大学 | Self-adaptive density peak clustering band selection method |
CN116976857A (en) * | 2023-08-08 | 2023-10-31 | 中国铁建电气化局集团北方工程有限公司 | Predictive maintenance method and system for railway power equipment based on deep learning |
CN117172589A (en) * | 2023-08-30 | 2023-12-05 | 浙江大学 | Industry typical user power demand response potential evaluation method |
CN117574187A (en) * | 2023-11-09 | 2024-02-20 | 西安电子科技大学 | Multi-radiation-source signal sorting time delay estimation method and system |
Non-Patent Citations (4)
Title |
---|
ZEXIAN SUN: "Hybrid model with secondary decomposition, randomforest algorithm, clustering analysis and long short memory network principal computing for short-term wind power forecasting on multiple scales", 《ENERGY》, 31 January 2021 (2021-01-31), pages 1 - 24 * |
吴斌 等: "自适应密度峰值聚类算法", 《计算机应用》, 21 January 2020 (2020-01-21), pages 1654 - 1661 * |
王哲川: "自动确定类数的密度峰值聚类算法研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》, 15 May 2021 (2021-05-15), pages 138 - 564 * |
马淑华 等: "一种自适应的密度峰值聚类算法", 《东北大学学报(自然科学版)》, vol. 43, no. 6, 15 June 2022 (2022-06-15), pages 761 - 768 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113868006B (en) | Time sequence detection method and device, electronic equipment and computer storage medium | |
CN108171119B (en) | SAR image change detection method based on residual error network | |
CN111189638B (en) | HMM and QPSO optimization algorithm-based bearing fault degree identification method | |
CN113532829A (en) | Reciprocating compressor fault diagnosis method based on improved RCMDE | |
CN113822201B (en) | Deep learning method for underwater object shape recognition based on flow field velocity component time course | |
CN111275198A (en) | Bearing abnormity detection method and system | |
CN111089856A (en) | Post-processing method for extracting Raman spectrum weak signal | |
CN116805061B (en) | Leakage event judging method based on optical fiber sensing | |
CN115953666B (en) | Substation site progress identification method based on improved Mask-RCNN | |
CN112380992A (en) | Method and device for evaluating and optimizing accuracy of monitoring data in machining process | |
CN115221973A (en) | Aviation bearing fault diagnosis method based on enhanced weighted heterogeneous ensemble learning | |
CN113837122B (en) | Wi-Fi channel state information-based contactless human body behavior recognition method and system | |
CN113758709A (en) | Rolling bearing fault diagnosis method and system combining edge calculation and deep learning | |
CN116738221B (en) | Pressurized pipeline gas analysis method and system | |
CN117786583A (en) | Fault diagnosis method and device based on variational modal decomposition and computer device | |
CN105513058B (en) | A kind of brain active region detection method and device | |
CN115310499B (en) | Industrial equipment fault diagnosis system and method based on data fusion | |
CN116227172A (en) | Rolling bearing performance degradation evaluation method based on convolutional neural network | |
CN113409213B (en) | Method and system for enhancing noise reduction of time-frequency diagram of fault signal of plunger pump | |
CN116992389B (en) | False data detection method and system for Internet of things | |
CN109446625A (en) | A kind of Helicopter Dynamic Components dynamic threshold calculation method based on Bayesian inference | |
CN112505779B (en) | Method for removing collected footprints based on feature decomposition | |
CN109036452A (en) | A kind of voice information processing method, device, electronic equipment and storage medium | |
He et al. | Nonlinear methods for rolling bearing fault diagnosis | |
CN112560674B (en) | Method and system for detecting sound signal quality |
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 |