CN117093938A - Fan bearing fault detection method and system based on deep learning - Google Patents
Fan bearing fault detection method and system based on deep learning Download PDFInfo
- Publication number
- CN117093938A CN117093938A CN202310975564.6A CN202310975564A CN117093938A CN 117093938 A CN117093938 A CN 117093938A CN 202310975564 A CN202310975564 A CN 202310975564A CN 117093938 A CN117093938 A CN 117093938A
- Authority
- CN
- China
- Prior art keywords
- deep learning
- fan bearing
- learning model
- fault
- data
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000013135 deep learning Methods 0.000 title claims abstract description 30
- 238000012544 monitoring process Methods 0.000 claims abstract description 49
- 238000013136 deep learning model Methods 0.000 claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims description 28
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 15
- 238000013527 convolutional neural network Methods 0.000 claims description 14
- 230000015654 memory Effects 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000003745 diagnosis Methods 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
- G01M13/045—Acoustic or vibration analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Acoustics & Sound (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a fault detection method and a fault detection system for a fan bearing based on deep learning, wherein the fault detection method comprises the steps of collecting real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data; constructing a deep learning model, extracting first features and second features of the real-time monitoring signal, and training the deep learning model; and inputting the signal to be detected into a deep learning model, carrying out fault identification on the fan bearing by using a classifier, and outputting a fault detection result. Compared with the prior art, the technical scheme of the invention can effectively solve the problems of low recognition precision, large calculated amount, poor robustness and the like, has higher recognition accuracy under the condition of small sample learning, and has strong application prospect.
Description
Technical Field
The invention relates to the technical field of new energy, in particular to a fan bearing fault detection method and system based on deep learning.
Background
In recent years, new energy development in China is rapid, wherein the development of wind power is particularly outstanding, the installed capacity of wind power is rapidly increased, and the wind power has become the global maximum wind power market.
The fan bearing is an important component part in the wind generating set, and the quality and performance of the fan bearing directly influence the generating efficiency and the operation safety of the wind generating set. Due to the long-term operation and the influence of external environmental factors, the fan bearing is easy to generate faults such as abrasion, corrosion, cracks and the like, so that the performance of the fan bearing is reduced, the service life of the fan bearing is shortened, and the damage of the fan bearing and the complete machine fault can be caused when the service life of the fan bearing is serious. Therefore, the method and the device can timely and accurately detect and diagnose the faults of the fan bearing, and have important significance for guaranteeing the safe operation of the wind generating set and improving the generating efficiency.
At present, the fault detection of the fan bearing mainly adopts methods such as vibration signal analysis, temperature signal analysis, sound signal analysis and the like, and the running state and the fault condition of the fan bearing are judged by analyzing the frequency spectrum characteristics, the time domain characteristics, the amplitude characteristics and the like of the signals. The traditional sound signal analysis method mainly adopts mathematical methods such as Fourier transform, wavelet transform, short-time Fourier transform and the like to carry out signal processing and feature extraction, and then uses classifiers such as a support vector machine, an artificial neural network and the like to carry out classification and identification on the features. Although the method has a certain application value in the field of fault detection of the fan bearing, the problems of low recognition precision, large calculation amount, poor robustness and the like still exist due to the complexity of signal processing and feature extraction and the complexity of a classifier, and the requirements of practical application are difficult to meet.
In summary, it is difficult to accurately and reliably detect a fan fault in the current technology.
Disclosure of Invention
In view of the above, the method and the system for detecting the fault of the fan bearing based on deep learning provided by the embodiment of the invention can effectively solve the problems of low recognition precision, large calculation amount, poor robustness and the like, and also have higher recognition accuracy under the condition of small sample learning, and have strong application prospects.
The embodiment of the invention provides a fault detection method of a fan bearing based on deep learning, which comprises the following steps:
collecting real-time monitoring signals of a fan bearing, wherein the real-time monitoring signals are time sequence data;
constructing a deep learning model, extracting first features and second features of the real-time monitoring signals, and training the deep learning model;
and inputting a signal to be detected into the deep learning model, carrying out fault identification on the fan bearing by using a classifier, and outputting a fault detection result.
Illustratively, the real-time monitoring signal of the fan bearing comprises fan bearing vibration, bearing temperature and rotational speed information; the fault identification result comprises serious faults, general faults and early warning faults, and the fault detection result comprises fault diagnosis suggestions output according to the fault grade.
Illustratively, the acquiring the real-time monitoring signal of the fan bearing includes:
acquiring original sample data containing real-time monitoring signals;
the original sample data is subjected to labeling classification to generate a first data set, a second data set and a third data set, wherein the monitoring data acquisition success rate of the first data set is not less than 90%, the monitoring data acquisition success rate of the second data set is more than 50% and less than 90%, and the monitoring data acquisition success rate of the third data set is not more than 50%;
screening and checking the labeled classified data set to obtain optimized sample data so as to improve the feature extraction capability of the small sample data.
Illustratively, the screening the labeled classified data set for check includes:
fitting analysis is carried out on the first data set, the second data set and the third data set by adopting a local linear regression method, and local information is amplified by utilizing a Gaussian function;
and extracting a sample data interval with the goodness-of-fit meeting the condition to obtain the optimized sample data.
Illustratively, the constructing the deep learning model, performing the first feature extraction and the second feature extraction on the real-time monitoring signal includes:
constructing a training set, a verification set and a test set for deep learning by using the optimized sample data and adopting a sliding time window method, and generating a support set by using the training set to randomly select data pairs;
and constructing a one-dimensional convolutional neural network and a long-short-term memory artificial neural network, extracting the first characteristic and the second characteristic by using the one-dimensional convolutional neural network and the long-short-term memory artificial neural network, and training the deep learning model.
Illustratively, the one-dimensional convolutional neural network includes an input layer, a convolutional layer, and a pooling layer, the constructing a one-dimensional convolutional neural network and a long-short-term memory artificial neural network, and extracting the first feature and the second feature using the one-dimensional convolutional neural network and the long-short-term memory artificial neural network, and performing the deep learning model training includes:
inputting the support set to the input layer, and obtaining the first characteristic through the convolution layer and the pooling layer;
the first characteristic sequence data is directly input into the long-short-term memory artificial neural network to obtain the second characteristic, the training process is optimized by utilizing a self-adaptive moment estimation method, network parameters are updated by minimizing a loss function, and the parameters are fixed on the deep learning model;
and adjusting the parameters by using the verification set.
Illustratively, the inputting the signal to be detected into the deep learning model, performing fault identification on the fan bearing by using a classifier, and outputting a fault detection result includes:
inputting the test set to the deep learning model;
and obtaining a fault probability vector by using a softmax function, carrying out fan fault identification according to the fault probability vector, and outputting the fault diagnosis suggestion according to an identification result.
Illustratively, the learning rate of the deep learning model is 0.015 and the number of iterations is 150.
Another embodiment of the present invention provides a fault detection system for a deep learning-based fan bearing, including:
the signal acquisition unit is used for acquiring real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data;
the learning training unit is used for constructing a deep learning model, extracting first features and second features of the real-time monitoring signals, and training the deep learning model;
the fault detection unit is used for inputting a signal to be detected into the deep learning model, carrying out fault identification on the fan bearing by using a classifier and outputting a fault detection result.
The invention provides a fault detection method and a fault detection system for a fan bearing based on deep learning, wherein the fault detection method comprises the steps of collecting real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data; constructing a deep learning model, extracting first features and second features of the real-time monitoring signal, and training the deep learning model; and inputting the signal to be detected into a deep learning model, carrying out fault identification on the fan bearing by using a classifier, and outputting a fault detection result. Compared with the prior art, the technical scheme of the invention can effectively solve the problems of low recognition precision, large calculated amount, poor robustness and the like, has higher recognition accuracy under the condition of small sample learning, and has strong application prospect.
Drawings
In order to more clearly illustrate the technical solutions of the present invention, the drawings that are required for the embodiments will be briefly described, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope of the present invention. Like elements are numbered alike in the various figures.
FIG. 1 is a schematic flow chart of a fault detection method of a fan bearing based on deep learning according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a method in step S101 according to an embodiment of the present invention;
fig. 3 is a schematic flow chart of a method in step S102 according to an embodiment of the present invention;
fig. 4 is a flowchart of a method in step S302 according to an embodiment of the present invention;
fig. 5 is a schematic flow chart of a method in step S103 according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a fault detection system of a blower bearing based on deep learning according to an embodiment of the present invention.
Description of main reference numerals:
10-a signal acquisition unit; 20-a learning training unit; 30-a fault detection unit.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
In the common fan power generation system faults, the fault occurrence probability of the transmission system is higher, and as the transmission system is an important component for generating energy conversion of the wind turbine, the influence caused by the traditional system faults is also larger, long-time shutdown and maintenance can be caused, and the safety reliability and the power generation economy of the wind power plant are seriously reduced. In this regard, the method and the system for detecting the fault of the fan bearing based on the deep learning provided by the embodiment of the invention realize the fault diagnosis of the deep neural network by combining the deep learning theory, and can be well applied to a transmission system of a wind turbine.
Example 1
Referring to fig. 1, the present embodiment provides a fault detection method for a fan bearing based on deep learning, including:
step S101, collecting real-time monitoring signals of a fan bearing, wherein the real-time monitoring signals are time sequence data;
specifically, the real-time monitoring signals of the fan bearing comprise information such as the fan bearing vibration exceeds standard, the bearing temperature is too high, and the moving blades are blocked and rotate to stall.
Step S102, a deep learning model is built, first feature extraction and second feature extraction are carried out on the real-time monitoring signals, and deep learning model training is carried out;
step S103, inputting a signal to be detected into a deep learning model, carrying out fault identification on a fan bearing by using a classifier, and outputting a fault detection result;
specifically, the fault recognition results include severe faults, general faults and early warning faults, and the fault detection results include fault diagnosis suggestions according to fault levels.
Referring to fig. 2, step S101 includes:
step S201, obtaining original sample data containing real-time monitoring signals;
step S202, labeling and classifying original sample data to generate a first data set, a second data set and a third data set, wherein the monitoring data acquisition success rate of the first data set is not less than 90%, the monitoring data acquisition success rate of the second data set is more than 50% and less than 90%, and the monitoring data acquisition success rate of the third data set is not more than 50%;
specifically, due to large change of field conditions, acquisition failure often occurs in monitoring data acquisition, and in this embodiment of the invention, before data is applied, original sample data is classified according to data acquisition quality to obtain a first data set, a second data set and a third data set, and the data quality of the three data sets is sequentially decreased. In the model training process, the first data set or the second data set is preferably selected for better feature extraction, and the third data set is selected according to the data application condition.
And step S203, screening and checking the labeled and classified data set to obtain optimized sample data so as to improve the feature extraction capability of the small sample data.
Specifically, the use of high quality data such as the first data set and the second data set is a simpler screening mode, and the embodiment of the invention provides another high quality data screening check method. Firstly, establishing a data quality screening check model, carrying out fitting analysis on the first data set, the second data set and the third data set by adopting a local linear regression method, amplifying local information by utilizing a Gaussian function, and carrying out weight assignment on a screened high-quality data set by adopting the Gaussian function as a weight value calculation function; and then extracting a sample data interval of the goodness-of-fit goodness load condition, and splicing to obtain optimized sample data.
Referring to fig. 3, step S102 includes:
step S301, constructing a training set, a verification set and a test set for deep learning by using optimized sample data and adopting a sliding time window method, and generating a support set by using a mode of randomly selecting data pairs by using the training set;
step S302, a one-dimensional convolutional neural network and a long-short-term memory artificial neural network are constructed, and the one-dimensional convolutional neural network and the long-short-term memory artificial neural network are utilized to extract first features and second features, so that deep learning model training is performed.
Specifically, since the real-time monitoring data belongs to one-dimensional data, the embodiment of the invention adopts a one-dimensional convolutional neural network for training, and mainly comprises an input layer, a convolutional layer and a pooling layer, wherein the input layer receives one-dimensional sequence data (namely optimized sample data) as the input of a model, a series of trainable convolutional kernels are used for sliding and extracting features on the input data, local feature information is extracted in the way, local distribution of the input sequence is extracted, and finally, the output result of the convolutional layer is subjected to dimension reduction processing in the pooling layer, so that the calculated amount is reduced, and the robustness of the model is improved. In the training process, the embodiment of the invention optimally calculates the super parameters required to be set by the network, including the aspects of the convolution kernel size, the number of convolution layers and the like, and performs optimization verification.
Referring to fig. 4, step S302 includes:
step S401, inputting a support set into an input layer, and processing in a convolution layer and a pooling layer to obtain a first characteristic;
step S402, the first characteristic sequence data is directly input into a long-short-term memory artificial neural network to obtain a second characteristic, a training process is optimized by using a self-adaptive moment estimation method, network parameters are updated by minimizing a loss function, and the parameters are fixed on a deep learning model;
step S403, adjusting the parameters by using the verification set.
In particular, the feature extraction in the prior art often stays in the aspect of local features, and long-term dependence problems existing in the recurrent neural network are not effectively considered, so that the diagnosis result may be biased. According to the embodiment of the invention, after the one-dimensional convolutional neural network is used for extracting the sequence data local time sequence characteristics of the wind turbine generator, the long-term and short-term memory artificial neural network is used for extracting the second characteristics, so that a more comprehensive wind turbine generator bearing fault judging result is formed.
Referring to fig. 5, step S103 includes:
step S501, inputting a test set into a deep learning model;
step S502, obtaining a fault probability vector by using a softmax function, carrying out fan fault identification according to the fault probability vector, and outputting fault diagnosis suggestions according to the identification result.
Specifically, the learning rate of the deep learning model is 0.015, and the number of iterations is 150. The fault detection method for the fan bearing based on deep learning provided by the embodiment of the invention can provide the diagnosis opinion of the fault degree on the basis of the fault diagnosis of the fan bearing such as excessive vibration, excessive bearing temperature, moving blade jamming, rotating stall information and the like. By taking the bearing vibration exceeding as an example, the embodiment of the invention can automatically judge the bearing vibration exceeding seriously, the bearing vibration exceeding generally or the bearing vibration exceeding early warning, so as to generate corresponding counter measure suggestions, and effectively help technicians to perform daily monitoring and maintenance work.
The invention provides a fault detection method of a fan bearing based on deep learning, which comprises the steps of collecting real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data; constructing a deep learning model, extracting first features and second features of the real-time monitoring signal, and training the deep learning model; and inputting the signal to be detected into a deep learning model, carrying out fault identification on the fan bearing by using a classifier, and outputting a fault detection result. Compared with the prior art, the technical scheme of the invention can effectively solve the problems of low recognition precision, large calculated amount, poor robustness and the like, has higher recognition accuracy under the condition of small sample learning, and has strong application prospect.
Example 2
Referring to fig. 6, the present embodiment further proposes a fault detection system of a fan bearing based on deep learning, including:
the signal acquisition unit 10 is used for acquiring real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data;
the learning training unit 20 is configured to construct a deep learning model, perform first feature extraction and second feature extraction on the real-time monitoring signal, and perform deep learning model training;
the fault detection unit 30 is configured to input a signal to be detected to the deep learning model, perform fault recognition on the fan bearing by using the classifier, and output a fault detection result.
It is to be understood that the above-described fault detection system of a deep learning-based fan bearing corresponds to the fault detection method of the deep learning-based fan bearing of embodiment 1. Any of the alternatives in embodiment 1 are also applicable to this embodiment and will not be described in detail here.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention.
Claims (9)
1. The fault detection method of the fan bearing based on deep learning is characterized by comprising the following steps of:
collecting real-time monitoring signals of a fan bearing, wherein the real-time monitoring signals are time sequence data;
constructing a deep learning model, extracting first features and second features of the real-time monitoring signals, and training the deep learning model;
and inputting a signal to be detected into the deep learning model, carrying out fault identification on the fan bearing by using a classifier, and outputting a fault detection result.
2. The method and system for detecting faults of a deep learning based fan bearing of claim 1 in which the real time monitoring signals of the fan bearing include fan bearing vibration, bearing temperature and rotational speed information; the fault identification result comprises serious faults, general faults and early warning faults, and the fault detection result comprises fault diagnosis suggestions output according to the fault grade.
3. The method and system for detecting a fault of a deep learning based fan bearing according to claim 2, wherein the collecting real-time monitoring signals of the fan bearing comprises:
acquiring original sample data, wherein the original sample data comprises a real-time monitoring signal;
the original sample data is subjected to labeling classification to generate a first data set, a second data set and a third data set, wherein the monitoring data acquisition success rate of the first data set is not less than 90%, the monitoring data acquisition success rate of the second data set is more than 50% and less than 90%, and the monitoring data acquisition success rate of the third data set is not more than 50%;
screening and checking the labeled classified data set to obtain optimized sample data so as to improve the feature extraction capability of the small sample data.
4. The method and system for detecting a failure of a blower bearing based on deep learning of claim 3, wherein the screening the labeled classified data set for check includes:
fitting analysis is carried out on the first data set, the second data set and the third data set by adopting a local linear regression method, and local information is amplified by utilizing a Gaussian function;
and extracting a sample data interval with the goodness-of-fit meeting the condition to obtain the optimized sample data.
5. The method and system for fault detection of a deep learning based fan bearing of claim 4, wherein constructing a deep learning model, performing a first feature extraction and a second feature extraction on the real-time monitoring signal comprises:
constructing a training set, a verification set and a test set for deep learning by using the optimized sample data and adopting a sliding time window method, and generating a support set by using the training set to randomly select data pairs;
and constructing a one-dimensional convolutional neural network and a long-short-term memory artificial neural network, extracting the first characteristic and the second characteristic by using the one-dimensional convolutional neural network and the long-short-term memory artificial neural network, and training the deep learning model.
6. The method and system for detecting a failure of a fan bearing based on deep learning according to claim 5, wherein the one-dimensional convolutional neural network includes an input layer, a convolutional layer, and a pooling layer, the constructing one-dimensional convolutional neural network and long-short-term memory artificial neural network, and extracting the first feature and the second feature by using the one-dimensional convolutional neural network and the long-short-term memory artificial neural network, and the training of the deep learning model includes:
inputting the support set to the input layer, and obtaining the first characteristic through the convolution layer and the pooling layer;
the first characteristic sequence data is directly input into the long-short-term memory artificial neural network to obtain the second characteristic, the training process is optimized by utilizing a self-adaptive moment estimation method, network parameters are updated by minimizing a loss function, and the parameters are fixed on the deep learning model;
and adjusting parameters by using the verification set.
7. The method and system for detecting a failure of a fan bearing based on deep learning according to claim 6, wherein inputting a signal to be detected to the deep learning model, performing failure recognition on the fan bearing by using a classifier, and outputting a failure detection result comprises:
inputting the test set to the deep learning model;
and obtaining a fault probability vector by using a softmax function, carrying out fan fault identification according to the fault probability vector, and outputting the fault diagnosis suggestion according to an identification result.
8. The method and system for detecting a failure of a deep learning based fan bearing according to claim 7, wherein the learning rate of the deep learning model is 0.015 and the number of iterations is 150.
9. A deep learning based fan bearing fault detection system, comprising:
the signal acquisition unit is used for acquiring real-time monitoring signals of the fan bearing, wherein the real-time monitoring signals are time sequence data;
the learning training unit is used for constructing a deep learning model, extracting first features and second features of the real-time monitoring signals, and training the deep learning model;
the fault detection unit is used for inputting a signal to be detected into the deep learning model, carrying out fault identification on the fan bearing by using a classifier and outputting a fault detection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310975564.6A CN117093938A (en) | 2023-08-03 | 2023-08-03 | Fan bearing fault detection method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310975564.6A CN117093938A (en) | 2023-08-03 | 2023-08-03 | Fan bearing fault detection method and system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117093938A true CN117093938A (en) | 2023-11-21 |
Family
ID=88782149
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310975564.6A Pending CN117093938A (en) | 2023-08-03 | 2023-08-03 | Fan bearing fault detection method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117093938A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117574327A (en) * | 2023-12-14 | 2024-02-20 | 盐城市崇达石化机械有限公司 | Fracturing pump fault detection method, system and storage medium |
| CN118656665A (en) * | 2024-08-19 | 2024-09-17 | 山东康吉诺技术有限公司 | Deep learning model-based wind turbine gearbox bearing temperature state detection method |
-
2023
- 2023-08-03 CN CN202310975564.6A patent/CN117093938A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117574327A (en) * | 2023-12-14 | 2024-02-20 | 盐城市崇达石化机械有限公司 | Fracturing pump fault detection method, system and storage medium |
| CN117574327B (en) * | 2023-12-14 | 2024-07-23 | 盐城市崇达石化机械有限公司 | Fracturing pump fault detection method, system and storage medium |
| CN118656665A (en) * | 2024-08-19 | 2024-09-17 | 山东康吉诺技术有限公司 | Deep learning model-based wind turbine gearbox bearing temperature state detection method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cui et al. | Quantitative trend fault diagnosis of a rolling bearing based on Sparsogram and Lempel-Ziv | |
| Saidi et al. | Wind turbine high-speed shaft bearings health prognosis through a spectral Kurtosis-derived indices and SVR | |
| CN103115789B (en) | Second generation small-wave support vector machine assessment method for damage and remaining life of metal structure | |
| CN117093938A (en) | Fan bearing fault detection method and system based on deep learning | |
| CN110259648B (en) | Fan blade fault diagnosis method based on optimized K-means clustering | |
| CN113738595B (en) | Wind generating set blade state monitoring method and system | |
| CN111209934A (en) | Fan fault prediction and alarm method and system | |
| CN112487890A (en) | Bearing acoustic signal fault diagnosis method based on parallel sparse filtering | |
| CN109255333A (en) | A kind of large-scale wind electricity unit rolling bearing fault Hybrid approaches of diagnosis | |
| CN114881087A (en) | Building robot bearing performance degradation assessment method | |
| CN116771610A (en) | Method for adjusting fault evaluation value of variable pitch system of wind turbine | |
| CN114463143A (en) | Method for enhancing SCADA fault data of offshore doubly-fed wind generator | |
| Zhang et al. | A novel hybrid compound fault pattern identification method for gearbox based on NIC, MFDFA and WOASVM | |
| CN117668646A (en) | Speed reducer running state detection method, system, equipment and medium based on multi-feature fusion | |
| Saha et al. | Investigating Rotor Conditions on Wind Turbines Using Integrating Tree Classifiers | |
| Sobha et al. | A comprehensive approach for gearbox fault detection and diagnosis using sequential neural networks | |
| Ahilan et al. | A Self-Improved Optimizer-Based CNN for Wind Turbine Fault Detection | |
| CN114962186A (en) | Method for predicting icing of wind turbine generator blade | |
| Zeng et al. | Data-Driven Anomaly Detection Method Based on Similarities of Multiple Wind Turbines | |
| Cheng et al. | Deep robust autoencoder based framework for bearing fault detection | |
| Wang et al. | Wind turbine spindle condition monitoring based on operational data | |
| CN117723303B (en) | Acoustic monitoring method for wind generating set bearing | |
| Saha et al. | Research Article Investigating Rotor Conditions on Wind Turbines Using Integrating Tree Classifiers | |
| Wang et al. | An AI-Driven Approach to Wind Turbine Bearing Fault Diagnosis from Acoustic Signals | |
| Owolabi et al. | Intermediate Stage Fault Classification for Wind Turbine Gearbox: A Comparative Analysis of Distance Metrics with k-NN Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |