CN114509158A - Acoustic-vibration fused blade crack fault detection method and application - Google Patents

Acoustic-vibration fused blade crack fault detection method and application Download PDF

Info

Publication number
CN114509158A
CN114509158A CN202210001190.3A CN202210001190A CN114509158A CN 114509158 A CN114509158 A CN 114509158A CN 202210001190 A CN202210001190 A CN 202210001190A CN 114509158 A CN114509158 A CN 114509158A
Authority
CN
China
Prior art keywords
vibration
samples
channel
sample
fusion
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.)
Granted
Application number
CN202210001190.3A
Other languages
Chinese (zh)
Other versions
CN114509158B (en
Inventor
宋狄
许飞云
胡建中
贾民平
黄鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southeast University
Original Assignee
Southeast University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Southeast University filed Critical Southeast University
Priority to CN202210001190.3A priority Critical patent/CN114509158B/en
Publication of CN114509158A publication Critical patent/CN114509158A/en
Application granted granted Critical
Publication of CN114509158B publication Critical patent/CN114509158B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
    • G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
    • G—PHYSICS
    • 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
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/04—Analysing solids
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G01N29/4481—Neural networks
    • 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
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/02—Indexing codes associated with the analysed material
    • G01N2291/023—Solids
    • G01N2291/0234—Metals, e.g. steel
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/72—Wind turbines with rotation axis in wind direction

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Immunology (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Data Mining & Analysis (AREA)
  • Acoustics & Sound (AREA)
  • Signal Processing (AREA)
  • Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention discloses a method for detecting blade crack faults through sound and vibration fusion, which is applied urgently and comprises the following steps: collecting two-channel sound signals and two-channel vibration signals, and dividing the signals into training and testing samples after standardization processing; respectively fusing two-channel vibration signal training and testing samples by using a data level fusion method, and generating vibration data fusion training and testing samples; inputting the two-channel acoustic signal, the two-channel vibration signal and the vibration data fusion training sample into a one-dimensional convolution neural network to obtain an initial detection result; and fusing the initial detection result by using a decision-level fusion method to obtain a final detection result, and realizing the crack fault detection of the compressor blade. The method is simple and easy to implement, and can fuse sound vibration signals at a data level and a decision level to realize the crack fault detection of the compressor blade compared with other existing crack detection technologies.

Description

Acoustic-vibration fused blade crack fault detection method and application
Technical Field
The invention relates to the technical field of sound vibration signal analysis and fusion of rotating machinery, in particular to a method for detecting a crack fault of a blade through sound vibration fusion.
Background
The compressor is widely applied to the fields of petrochemical industry, electric power and the like, and the blade serving as a core component is easy to crack and break down under the action of centrifugal force, friction and unstable airflow load, so that the normal work of the whole compressor is influenced. Therefore, the method can detect the blade crack fault in time and has important significance for ensuring the safe and stable operation of the compressor. In addition, in actual engineering, blade failure detection is often performed using an acoustic signal, a vibration signal, or the like. However, a single acoustic signal or a single vibration signal is often doped with strong background noise, and fault analysis performed by the single signal results in low detection accuracy and unreliable detection results. Therefore, how to effectively utilize and fuse the acoustic signal and the vibration signal and realize reliable and accurate detection of the crack fault of the compressor blade is worthy of further research.
The traditional signal fusion comprises 3 fusion modes: data level, feature level, and decision level fusion. Generally, data level fusion is suitable for signals of the same type and different positions, and is the lowest level fusion mode; the feature level and decision level fusion can fuse different types of signals, wherein the decision level fusion is the highest level fusion mode and has the best fusion effect. Due to the existence of strong noise, a single fusion mode may not be capable of accurately detecting blade cracks, and different similar fusions often have certain conflicts. Therefore, a method for detecting crack faults of compressor blades, which can perform sound vibration signal fusion in different degrees according to different signal characteristics and realize reliability and accuracy, is urgently needed.
Disclosure of Invention
Aiming at the defects, the invention provides the sound-vibration fused blade crack fault detection method, which overcomes the current situations that the crack fault characteristics of the compressor blade are difficult to accurately reflect and the crack fault of the blade cannot be accurately detected by the conventional single signal and single fusion mode, and effectively realizes the sound-vibration signal fusion and the crack fault detection of the compressor blade.
In order to solve the technical problems, the invention adopts the following technical scheme:
a sound and vibration fused compressor blade crack fault detection method comprises the following steps:
a blade crack fault detection method based on sound and vibration fusion is characterized by comprising the following steps:
step 1: collecting at least two channel acoustic signals and at least two channel vibration signals;
step 2: standardizing the acquired at least two channel acoustic signals and at least two channel vibration signals, and dividing the standardized acoustic signals and vibration signals into training samples and testing samples;
and step 3: respectively fusing at least two channels of vibration signal training and testing samples by using a data level fusion method, and generating vibration data fusion training and testing samples;
and 4, step 4: fusing the acoustic signal and vibration signal training samples processed in the step (2) and the vibration data generated in the step (3) into a training sample, and inputting the training sample into a neural network to obtain an initial detection result;
and 5: and (4) fusing the initial detection result obtained in the step (4) by using a decision-level fusion method, and obtaining a final detection result to realize the crack fault detection of the blade.
In the step 1, a sound pressure sensor is adopted to obtain sound signals, the sound pressure sensor is respectively away from a set distance of 0.2-1.0 meter from the air inlet and the air outlet, a set included angle is formed by the sound pressure sensor and the sound pressure sensor in an upward inclined mode, and the set included angle can be 45 degrees.
In step 2, the normalization process is a 0-1 normalization process, and the expression of the 0-1 normalization process is as follows:
Figure BDA0003454154710000031
wherein x iss(t) represents the normalized signal from 0 to 1, x (t) represents the raw signal, min (x) and max (x) represent the minimum and maximum values of the raw signal, respectively.
Step 3, the data level fusion method comprises the following specific steps:
step 3.1, respectively calculating the Hoyer distances of the two channel samples, wherein the Hoyer distance expression is as follows:
Figure BDA0003454154710000032
wherein HxRepresenting the Hoyer distance of sample x and t representing the length of sample x.
Step 3.2, respectively calculating improved cosine similarity of the two-channel samples according to the historical samples and the Hoyer distance, wherein the expression of the improved cosine similarity is as follows:
Figure BDA0003454154710000033
Figure BDA0003454154710000034
wherein, thetaxRepresenting the cosine angle, x, of a sample xhRepresenting a history sample, G (x, x, H)x)、G(x,xh,Hx) And G (x)h,xh,Hx) Representing between samples x, sample x and historical sample xhAnd historical sample xhGaussian distance therebetween.
And 3.3, fusing the two-channel samples according to the improved cosine similarity of the two-channel samples, wherein the fused expression is as follows:
Figure BDA0003454154710000041
Figure BDA0003454154710000042
wherein x isdfDenotes the fused sample, ωiRepresents the ithThe weight of the sample.
Preferably, in step 4, the one-dimensional convolutional neural network includes 2 convolutional layers, 2 pooling layers, 1 full-link layer, and 1 Softmax layer, the size of the convolutional layer filter is 6 × 1, the step size of the convolutional layer is 1, the number of channels of the first convolutional layer is 4, the number of channels of the second convolutional layer is 8, the size of the pooling layer filter is 6 × 1, the step size of the pooling layer is 6, the number of channels of the first pooling layer is 4, the number of channels of the second pooling layer is 8, and the number of outputs of the Softmax layer is 3.
Preferably, the decision-level fusion method in step 5 specifically includes the steps of:
step 5.1, respectively calculating the precision and the accuracy of different types of different samples in the initial detection result, wherein the expressions of the precision and the accuracy are as follows:
Figure BDA0003454154710000043
Figure BDA0003454154710000044
wherein, PijIndicates the accuracy of the jth class in the ith sample, AiRepresents the accuracy of the ith sample, TPij、TNij、FPijAnd FNijRespectively representing the number of the true positive samples, the true negative samples, the false positive samples and the false negative samples.
Step 5.2, calculating the reliability of the sample, wherein the reliability expression is as follows:
Figure BDA0003454154710000045
wherein, CPijIndicating the confidence level of the jth class in the ith sample.
Step 5.3, calculating an initial decision-level fusion result, wherein the expression of the initial decision-level fusion is as follows:
Figure BDA0003454154710000051
wherein, FPTjAnd (5) representing the initial decision level fusion result of the jth category, and m representing the number of the initial detection results.
And 5.4, verifying the initial decision-level fusion result. When the probability of the initial decision-level fusion result exceeds 60%, the result is considered to be a valid result, and the initial decision-level fusion result is the final decision-level fusion result; otherwise, removing the initial detection result with the lowest accuracy, and repeating the steps 5.1-5.3, and verifying until the requirements are met.
The invention has the following beneficial effects:
1) according to the method for detecting the crack fault of the compressor blade with the sound-vibration fusion, two channels of vibration signals are fused in a data level mode, initial detection of the sound signals, the vibration signals and the vibration data level fusion signals is achieved by combining a one-dimensional convolution neural network, and finally decision level fusion is conducted on initial detection results to achieve the crack fault detection of the compressor blade. The defect that the crack fault of the blade cannot be accurately detected by a single signal or a single fusion mode is overcome, and the accuracy of the crack fault detection of the compressor blade is improved by fusing the acoustic vibration signals at a data level and a decision level.
2) The vibration signal coefficient degree is calculated through the Hoyer distance, and the cosine similarity is improved, so that the data level fusion method provided by the invention can be used for analyzing the vibration signal sparsity degree and the similarity with the historical signal, further accurately reflecting the blade crack fault characteristics, and improving the reliability of the vibration signal.
3) The decision-level fusion method provided by the invention carries out credibility distribution by calculating the precision and accuracy of the initial detection result, improves the reliability of sound vibration fusion on the basis of verifying the initial decision-level fusion result, and realizes high-accuracy detection of the blade crack fault.
Drawings
FIG. 1 is a block flow diagram of an embodiment of the present invention;
FIG. 2 is a flow chart of a data level fusion method in an embodiment of the invention;
FIG. 3 is a flow chart of a decision-level fusion method according to an embodiment of the present invention.
Detailed Description
The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses. It should be understood that throughout the drawings, corresponding reference numerals indicate like or corresponding parts and features.
As shown in FIG. 1, a method for detecting a crack fault of a compressor blade with sound and vibration fusion comprises the following steps:
step 1: two sound pressure sensors are arranged at the air inlet and the air outlet of the compressor to acquire two-channel sound signals, and two vibration sensors are arranged near the motor spindle to acquire two-channel vibration signals.
Step 2: and carrying out standardization processing on the four-channel acoustic signal and the vibration signal, and dividing the standardized acoustic signal and vibration signal into a training sample and a test sample.
And step 3: and respectively fusing two-channel vibration signal training and testing samples by using a data-level fusion method, and generating vibration data fusion training and testing samples.
And 4, step 4: and inputting the two-channel acoustic signal, the two-channel vibration signal and the vibration data fusion training sample into a one-dimensional convolution neural network to obtain an initial detection result.
And 5: and fusing the initial detection result by using a decision-level fusion method, obtaining a final detection result, and realizing crack fault detection of the compressor blade.
In the embodiment, the two sound pressure sensors collect two-channel sound signals at the air inlet and the air outlet of the compressor, the two vibration sensors collect two-channel vibration signals near the main shaft of the motor, the two-channel vibration signals are divided into training and testing samples after standardization processing, the two-channel vibration signals are fused by using a data-level fusion method and input into a one-dimensional convolutional neural network to obtain an initial detection result, and finally the initial detection result is fused by using a decision-level fusion method to realize crack fault detection of the compressor blade.
It should be noted that, the sampling frequency of the acoustic signal and the vibration signal is 5120Hz, the sampling time is 1s, and the training sample and the test sample are as follows: 3, the training sample and the test sample may have a certain difference due to the randomness of the division, but the subsequent processing is not affected.
In the step 1, the sound pressure sensor is respectively 0.5 meter away from the air inlet and the air outlet, and the sound pressure sensor inclines upwards to form an included angle of 45 degrees.
It should be noted that the distance is referred to the center of the air inlet and the air outlet, and the 45 ° is referred to the horizontal direction.
In step 2, the normalization process is a 0-1 normalization process, and the expression of the 0-1 normalization process is as follows:
Figure BDA0003454154710000071
wherein x iss(t) represents the normalized signal from 0 to 1, x (t) represents the raw signal, min (x) and max (x) represent the minimum and maximum values of the raw signal, respectively.
In step 3, the data level fusion method comprises the following specific steps:
step 3.1, respectively calculating the Hoyer distances of the two channel samples, wherein the Hoyer distance expression is as follows:
Figure BDA0003454154710000081
wherein HxRepresenting the Hoyer distance of sample x and t representing the length of sample x.
Step 3.2, respectively calculating the improved cosine similarity of the two-channel samples according to the historical samples and the Hoyer distance, wherein the expression of the improved cosine similarity is as follows:
Figure BDA0003454154710000082
Figure BDA0003454154710000083
wherein, thetaxRepresenting the cosine angle, x, of a sample xhRepresenting a history sample, G (x, x, H)x)、G(x,xh,Hx) And G (x)h,xh,Hx) Representing between samples x, sample x and historical sample xhAnd historical sample xhGaussian distance therebetween.
And 3.3, fusing the two-channel samples according to the improved cosine similarity of the two-channel samples, wherein the fused expression is as follows:
Figure BDA0003454154710000084
Figure BDA0003454154710000085
wherein x isdfDenotes the fused sample, ωiRepresenting the weight of the ith sample.
Note that the history sample is the previous sample of the sample x, and the value of the history sample of the first sample is set to 0.
In step 4, the one-dimensional convolutional neural network comprises 2 convolutional layers, 2 pooling layers, 1 full-connection layer and 1 Softmax layer, the size of the convolutional layer filter is 6 × 1, the step length of the convolutional layers is 1, the number of channels of the first convolutional layer is 4, the number of channels of the second convolutional layer is 8, the size of the pooling layer filter is 6 × 1, the step length of the pooling layer is 6, the number of channels of the first pooling layer is 4, the number of channels of the second pooling layer is 8, and the output number of the Softmax layer is 3.
It should be noted that the one-dimensional convolutional neural network is constructed in the order of convolutional layer, pooling layer, full-link, and Softmax layer.
In step 5, the decision-level fusion method comprises the following specific steps:
step 5.1, respectively calculating the precision of different samples and the accuracy of different types in the initial detection result, wherein the expressions of the precision and the accuracy are as follows:
Figure BDA0003454154710000091
Figure BDA0003454154710000092
wherein, PijIndicates the accuracy of the jth class in the ith sample, AiRepresents the accuracy of the ith sample, TPij、TNij、FPijAnd FNijRespectively representing the number of the true positive samples, the true negative samples, the false positive samples and the false negative samples.
Step 5.2, calculating the reliability of the sample, wherein the reliability expression is as follows:
Figure BDA0003454154710000093
wherein, CPijIndicating the confidence level of the jth class in the ith sample.
Step 5.3, calculating an initial decision-level fusion result, wherein the expression of the initial decision-level fusion is as follows:
Figure BDA0003454154710000094
wherein, FPTjAnd (5) representing the initial decision level fusion result of the jth category, and m representing the number of the initial detection results.
And 5.4, verifying the initial decision-level fusion result. When the probability of the initial decision-level fusion result exceeds 60%, the result is regarded as a valid result, and the initial decision-level fusion result is the final decision-level fusion result; otherwise, removing the initial detection result with the lowest accuracy, and repeating the steps 5.1-5.3, and verifying until the requirements are met.
In order to further verify the effectiveness of the inventive method, the adopted compressor blade crack test bed further explains the scheme of the application:
the compressor blade crack test bench has 3 kinds of crack length blades of 0, 10mm and 20mm and operates at four rotation speeds of 1200rpm, 1500rpm, 1800rpm and 2100 rpm. And collecting two-channel acoustic signals and vibration signals, and dividing the signals into training samples and testing samples after standardization processing. And fusing the two-channel vibration samples by adopting a data-level fusion method, wherein the flow of the data-level fusion method is shown in figure 2. And inputting the acoustic signal, the vibration signal and the vibration data fusion sample into a one-dimensional convolution neural network for detection to obtain an initial detection result. And finally fusing the initial detection result by using a decision-level fusion method, wherein the flow is shown in fig. 3. And classifying and detecting crack faults according to the decision-level fusion result, wherein the crack fault detection results of the compressor blades at 4 rotating speeds are shown in the table 1. It has been found that the accuracy of fault detection of a single acoustic or vibration signal sample is low due to noise interference. After data level fusion is carried out, the accuracy of crack fault detection of the compressor blade is greatly improved and is all over 90%. Finally, after decision-making level fusion, the accuracy of blade crack fault detection exceeds 98% at 4 rotating speeds, and reaches 100% at 1500rpm, so that the accuracy and reliability of the method are verified.
TABLE 1 compressor blade crack Fault detection results
Rotational speed 1200rpm 1500rpm 1800rpm 2100rpm
Acoustic signal sample 1 83.2％ 88.8％ 79.6％ 72.9％
Acoustic signal sample 2 92.3％ 76.3％ 84.6％ 77.4％
Vibration signal sample 1 95.4％ 87.0％ 84.3％ 85.7％
Vibration signal sample 2 95.2％ 88.4％ 85.1％ 87.1％
Data level fusion 97.1％ 93.2％ 90.2％ 95.6％
Decision level fusion 99.3％ 100％ 98.8％ 99.3％
To further illustrate the advantages of the method of the present invention in the detection of cracks in compressor blades, a comparison was made with other fusion methods. The method comprises a feature level fusion method based on a one-dimensional convolutional neural network, a decision level fusion method based on voting, a fusion method based on a D-S evidence theory and a decision level fusion method based on a convolutional neural network. Based on the original acoustic vibration signals and the initial results of the experiment, four comparison methods are respectively used for detecting the cracks of the compressor blade, and the results are shown in table 2.
TABLE 2 comparison of different fusion methods
Figure BDA0003454154710000111
It can be found that the feature level fusion method based on the one-dimensional convolutional neural network is to perform fusion on a fully connected layer, and the average precision is 97.82%. However, the accuracy of this method decreases with increasing rotational speed, and is easily disturbed by strong noise at high rotational speeds, which is mainly that this method is too dependent on the features extracted by the deep network, and the accuracy of this method under all conditions is not as good as the method of the present invention.
For the voting-based decision-level fusion method, the accuracies at 1800rpm and 2100rpm were 94.92% and 95.74%, respectively. This is because the blindly obtained voting result is not accurate enough because the authenticity of the result is not analyzed. In addition, if the voting results of the two tags are the same, randomness and confusion are easily generated.
For the fusion method based on the D-S evidence theory, the average precision of the fusion method is 97.76% as a classical decision-level fusion method, but the method is lower than that of the method of the invention.
For the decision-level fusion method based on the convolutional neural network, the average accuracy is only 96.81%, which is the worst of all the methods. By directly fusing through CNN, acoustic vibration overfitting is easily generated, resulting in poor accuracy.
For the method of the present invention, it has the highest accuracy under four operating conditions and is superior to other comparative methods. The result shows that the method has good anti-noise performance and can accurately detect the cracks.
Compared with the prior art, the method for detecting the crack fault of the compressor blade through sound-vibration fusion has the advantages that two channels of vibration signals are fused through data level, the initial detection of the sound signals, the vibration signals and the vibration data level fusion signals is realized by combining a one-dimensional convolution neural network, and finally the crack fault of the compressor blade is detected by performing decision level fusion on the initial detection result. The defect that the crack fault of the blade cannot be accurately detected by a single signal or a single fusion mode is overcome, and the accuracy of the crack fault detection of the compressor blade is improved.
The embodiments of the present invention are merely preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. That is, all equivalent changes and modifications made according to the content of the claims of the present invention should be regarded as the technical scope of the present invention.

Claims (10)

1. A blade crack fault detection method based on sound and vibration fusion is characterized by comprising the following steps:
step 1: collecting at least two channel acoustic signals and at least two channel vibration signals;
step 2: standardizing the acquired at least two channel acoustic signals and at least two channel vibration signals, and dividing the standardized acoustic signals and vibration signals into training samples and testing samples;
and step 3: respectively fusing at least two channels of vibration signal training and testing samples by using a data level fusion method, and generating vibration data fusion training and testing samples;
and 4, step 4: fusing the acoustic signal and vibration signal training samples processed in the step (2) and the vibration data generated in the step (3) into a training sample, and inputting the training sample into a neural network to obtain an initial detection result;
and 5: and (5) fusing the initial detection result obtained in the step (4) by using a decision-level fusion method, obtaining a final detection result, and realizing the crack fault detection of the blade.
2. The method for detecting the crack fault of the acoustic-vibration fused machine blade according to claim 1, wherein in the step 1, two-channel acoustic signals and two-channel vibration signals are collected; the two-channel acoustic signals are signals at set distances from the air inlet and the air outlet respectively.
3. The method for detecting the crack fault of the acoustic-vibration fused blade as claimed in claim 2, wherein the set distance between the sound signals of the two channels and the air inlet and the set distance between the sound signals of the two channels and the set distance between the air inlet and the set distance between the air outlet are 0.2-1.0 meter.
4. The method for detecting the crack fault of the acoustic-vibration fused blade as claimed in claim 1, wherein in the step 2, the standardization process is a 0-1 standardization process, and the expression of the 0-1 standardization process is as follows:
Figure FDA0003454154700000021
wherein x iss(t) represents the normalized signal from 0 to 1, x (t) represents the raw signal, min (x) and max (x) represent the minimum and maximum values of the raw signal, respectively.
5. The method for detecting the blade crack fault through the acoustic-vibration fusion according to claim 1, wherein in the step 3, at least two channel vibration signal training and testing samples are respectively fused by using a data-level fusion method, and the method comprises the following steps:
step 3.1, respectively calculating the Hoyer distances of the two channel samples:
Figure FDA0003454154700000022
wherein HxTo representThe Hoyer distance of sample x, t represents the length of sample x;
step 3.2, respectively calculating the improved cosine similarity of the two channel samples according to the historical samples and the Hoyer distance:
Figure FDA0003454154700000023
wherein, thetaxRepresenting the cosine angle, x, of a sample xhRepresenting a history sample, G (x, x, H)x) Denotes the Gaussian distance between samples x, G (x, x)h,Hx) Representative sample x and historical sample xhGaussian distance between, G (x)h,xh,Hx) Representing historical samples xhThe gaussian distance between;
sample x and historical sample xhGaussian distance between G (x, x)h,Hx) Comprises the following steps:
Figure FDA0003454154700000024
and 3.3, fusing the two-channel samples according to the improved cosine similarity of the two-channel samples, wherein the fused expression is as follows:
Figure FDA0003454154700000031
wherein x isdfThe fused samples are represented as a result of the fusion,
Figure FDA0003454154700000032
represents a sample xiThe cosine angle of (c).
6. The method for detecting the crack fault of the acoustic-vibration fused blade as claimed in claim 5, wherein in the step 5, the initial detection result obtained in the step 4 is fused by using a decision-level fusion method, and the method comprises the following steps:
step 5.1, respectively calculating the precision of different samples and the accuracy of different types in the initial detection result:
Figure FDA0003454154700000033
Figure FDA0003454154700000034
wherein, PijIndicates the accuracy of the jth class in the ith sample, AiRepresents the accuracy of the ith sample, TPij、TNij、FPijAnd FNijRespectively representing the number of the true positive samples, the true negative samples, the false positive samples and the false negative samples;
step 5.2, calculating the sample credibility:
Figure FDA0003454154700000035
wherein, CPijRepresenting the credibility of the jth category in the ith sample;
Prijindicating the precision of the jth category in the initial detection result in the ith sample;
step 5.3, calculating an initial decision-level fusion result:
Figure FDA0003454154700000036
wherein, FPTjRepresenting the initial decision level fusion result of the jth category, and m representing the number of the initial detection results;
step 5.4, verifying the initial decision-level fusion result; when the probability of the initial decision-level fusion result exceeds a set value, the initial decision-level fusion result is the final decision-level fusion result; otherwise, removing the initial detection result with the lowest accuracy, and repeating the steps 5.1-5.3, and verifying until the requirements are met.
7. The method for detecting the blade crack fault through the acoustic-vibration fusion according to claim 6, wherein in the step 5.4, the set value for verifying the initial decision-level fusion result is 50% -70%.
8. The method for detecting the blade crack fault through the acoustic-vibration fusion as claimed in claim 1, wherein in the step 4, the neural network is a one-dimensional convolution neural network and comprises 2 convolution layers, 2 pooling layers, 1 full-link layer and 1 Softmax layer.
9. The method of claim 8, wherein the convolutional layer filter size of the one-dimensional convolutional neural network is 6 x 1, the convolutional layer step size is 1, the first layer convolutional layer channel number is 4, the second layer convolutional layer channel number is 8, the pooling layer filter size is 6 x 1, the pooling layer step size is 6, the first layer pooling layer channel number is 4, the second layer pooling layer channel number is 8, and the output number of the Softmax layers is 3.
10. The blade crack fault detection method based on the sound-vibration fusion of any one of claims 1 to 9 is applied to a compressor and used for carrying out crack fault identification on the blade of the compressor.
CN202210001190.3A 2022-01-04 2022-01-04 Acoustic-vibration-fused blade crack fault detection method and application Active CN114509158B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210001190.3A CN114509158B (en) 2022-01-04 2022-01-04 Acoustic-vibration-fused blade crack fault detection method and application

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210001190.3A CN114509158B (en) 2022-01-04 2022-01-04 Acoustic-vibration-fused blade crack fault detection method and application

Publications (2)

Publication Number Publication Date
CN114509158A true CN114509158A (en) 2022-05-17
CN114509158B CN114509158B (en) 2023-05-30

Family

ID=81549898

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210001190.3A Active CN114509158B (en) 2022-01-04 2022-01-04 Acoustic-vibration-fused blade crack fault detection method and application

Country Status (1)

Country Link
CN (1) CN114509158B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115166032A (en) * 2022-05-23 2022-10-11 东南大学 Device and method for detecting cracks of fan blade

Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS55162698A (en) * 1979-06-06 1980-12-18 Matsushita Electric Ind Co Ltd Acoustic diaphragm plate
US20020059831A1 (en) * 2000-10-10 2002-05-23 Snecma Moteurs Acoustic inspection of one-piece bladed wheels
JP2004317408A (en) * 2003-04-18 2004-11-11 Calsonic Kansei Corp System and method for surveying vibration source of rotor
CN107016404A (en) * 2017-02-24 2017-08-04 沈阳工业大学 Wind power generating set failure prediction method based on D S evidence fusions
US20180348303A1 (en) * 2017-06-01 2018-12-06 General Electric Company Wind Turbine Fault Detection Using Acoustic, Vibration, and Electrical Signals
CN110131109A (en) * 2019-04-25 2019-08-16 浙江大学 A kind of pneumatic equipment bladess unbalance detection based on convolutional neural networks
CN110376522A (en) * 2019-09-03 2019-10-25 宁夏西北骏马电机制造股份有限公司 A kind of Method of Motor Fault Diagnosis of the deep learning network of data fusion
CN111239249A (en) * 2020-03-18 2020-06-05 北京工业大学 Ventilator blade crack fault diagnosis method based on Hilbert-Huang transform
US20200219527A1 (en) * 2019-01-06 2020-07-09 3D Signals Ltd. Extracting Overall Equipment Effectiveness by Analysis of a Vibro-Acoustic Signal
CN111612030A (en) * 2020-03-30 2020-09-01 华电电力科学研究院有限公司 Wind turbine generator blade surface fault identification and classification method based on deep learning
CN111678698A (en) * 2020-06-17 2020-09-18 沈阳建筑大学 Rolling bearing fault detection method based on sound and vibration signal fusion
CN112255538A (en) * 2020-10-14 2021-01-22 国网青海省电力公司海西供电公司 Multi-feature optimization fusion high-voltage circuit breaker fault diagnosis method
CN112326210A (en) * 2019-07-17 2021-02-05 华北电力大学（保定） Large motor fault diagnosis method combining sound vibration signals with 1D-CNN
CN112393800A (en) * 2020-10-28 2021-02-23 浙江工业大学 Monitoring device that sound vibration signal is united
CN112684012A (en) * 2020-12-02 2021-04-20 青岛科技大学 Equipment key force-bearing structural part fault diagnosis method based on multi-parameter information fusion
CN113033309A (en) * 2021-02-25 2021-06-25 北京化工大学 Fault diagnosis method based on signal downsampling and one-dimensional convolution neural network
CN113255777A (en) * 2021-05-28 2021-08-13 郑州轻工业大学 Equipment fault early warning method and system based on multi-mode sensitive feature selection fusion
CN113469060A (en) * 2021-07-02 2021-10-01 浙大城市学院 Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method
CN113469281A (en) * 2021-07-22 2021-10-01 西北工业大学 Industrial gear box multi-source information fusion fault diagnosis method
CN113533962A (en) * 2021-07-29 2021-10-22 上海交通大学 Induction motor health diagnosis system based on decision fusion of multiple physical signal sensors

Patent Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS55162698A (en) * 1979-06-06 1980-12-18 Matsushita Electric Ind Co Ltd Acoustic diaphragm plate
US20020059831A1 (en) * 2000-10-10 2002-05-23 Snecma Moteurs Acoustic inspection of one-piece bladed wheels
JP2004317408A (en) * 2003-04-18 2004-11-11 Calsonic Kansei Corp System and method for surveying vibration source of rotor
CN107016404A (en) * 2017-02-24 2017-08-04 沈阳工业大学 Wind power generating set failure prediction method based on D S evidence fusions
US20180348303A1 (en) * 2017-06-01 2018-12-06 General Electric Company Wind Turbine Fault Detection Using Acoustic, Vibration, and Electrical Signals
US20200219527A1 (en) * 2019-01-06 2020-07-09 3D Signals Ltd. Extracting Overall Equipment Effectiveness by Analysis of a Vibro-Acoustic Signal
CN110131109A (en) * 2019-04-25 2019-08-16 浙江大学 A kind of pneumatic equipment bladess unbalance detection based on convolutional neural networks
CN112326210A (en) * 2019-07-17 2021-02-05 华北电力大学（保定） Large motor fault diagnosis method combining sound vibration signals with 1D-CNN
CN110376522A (en) * 2019-09-03 2019-10-25 宁夏西北骏马电机制造股份有限公司 A kind of Method of Motor Fault Diagnosis of the deep learning network of data fusion
CN111239249A (en) * 2020-03-18 2020-06-05 北京工业大学 Ventilator blade crack fault diagnosis method based on Hilbert-Huang transform
CN111612030A (en) * 2020-03-30 2020-09-01 华电电力科学研究院有限公司 Wind turbine generator blade surface fault identification and classification method based on deep learning
CN111678698A (en) * 2020-06-17 2020-09-18 沈阳建筑大学 Rolling bearing fault detection method based on sound and vibration signal fusion
CN112255538A (en) * 2020-10-14 2021-01-22 国网青海省电力公司海西供电公司 Multi-feature optimization fusion high-voltage circuit breaker fault diagnosis method
CN112393800A (en) * 2020-10-28 2021-02-23 浙江工业大学 Monitoring device that sound vibration signal is united
CN112684012A (en) * 2020-12-02 2021-04-20 青岛科技大学 Equipment key force-bearing structural part fault diagnosis method based on multi-parameter information fusion
CN113033309A (en) * 2021-02-25 2021-06-25 北京化工大学 Fault diagnosis method based on signal downsampling and one-dimensional convolution neural network
CN113255777A (en) * 2021-05-28 2021-08-13 郑州轻工业大学 Equipment fault early warning method and system based on multi-mode sensitive feature selection fusion
CN113469060A (en) * 2021-07-02 2021-10-01 浙大城市学院 Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method
CN113469281A (en) * 2021-07-22 2021-10-01 西北工业大学 Industrial gear box multi-source information fusion fault diagnosis method
CN113533962A (en) * 2021-07-29 2021-10-22 上海交通大学 Induction motor health diagnosis system based on decision fusion of multiple physical signal sensors

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
FARID K. MOGHADAM: "Theoretical and experimental study of wind turbine drivetrain fault diagnosis by using torsional vibrations and modal estimation", JOURNAL OF SOUND AND VIBRATION *
MARTIN D. ULRIKSEN: "Damage detection in an operating Vestas V27 wind turbine blade by use of outlier analysis", 2015 IEEE WORKSHOP ON ENVIRONMENTAL, ENERGY, AND STRUCTURAL MONITORING SYSTEMS (EESMS) PROCEEDINGS *
翟栋;: "基于多技术融合的风电大部件健康状态大数据分析", 风能产业（2017年第1期 总第90期） *
郭林: "基于声振特征融合的机车走行部轴承状态评估方法研究", 铁道技术监督 *
颉潭成;: "基于信息融合的滚动轴承故障诊断", 轴承 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115166032A (en) * 2022-05-23 2022-10-11 东南大学 Device and method for detecting cracks of fan blade
CN115166032B (en) * 2022-05-23 2024-04-19 东南大学 Device and method for detecting cracks of fan blades

Also Published As

Publication number Publication date
CN114509158B (en) 2023-05-30

Similar Documents

Publication Publication Date Title
CN111353482B (en) LSTM-based fatigue factor recessive anomaly detection and fault diagnosis method
CN109115491B (en) Evidence fusion method for mechanical fault diagnosis of shafting propulsion system of electrically-propelled ship
CN113375941B (en) Open-set fault diagnosis method for bearings of high-speed motor train unit
CN115166032B (en) Device and method for detecting cracks of fan blades
CN110259648B (en) Fan blade fault diagnosis method based on optimized K-means clustering
Mechefske et al. Fault detection and diagnosis in low speed rolling element bearings Part II: The use of nearest neighbour classification
Meng et al. Health indicator of bearing constructed by rms-CUMSUM and GRRMD-CUMSUM with multifeatures of envelope spectrum
CN114509158B (en) Acoustic-vibration-fused blade crack fault detection method and application
CN105866250A (en) Vibration-based identification method for cracks of ventilating vane
CN112487890B (en) Bearing acoustic signal fault diagnosis method based on parallel sparse filtering
Saimurugan et al. A comparative study of sound and vibration signals in detection of rotating machine faults using support vector machine and independent component analysis
Khan et al. System design for early fault diagnosis of machines using vibration features
Song et al. Data and decision level fusion-based crack detection for compressor blade using acoustic and vibration signal
CN110701087A (en) Axial flow compressor pneumatic instability detection method based on single-classification overrun learning machine
CN111062093B (en) Automobile tire service life prediction method based on image processing and machine learning
CN113314142B (en) Engine fault detection method, system, storage medium and equipment
CN108760268B (en) Step fault diagnosis method for vertical mill operation data based on information entropy
CN110749443B (en) Rolling bearing fault diagnosis method and system based on high-order origin moment
CN116307938A (en) Health state assessment method for feeding system of machining center
CN114528868A (en) Crack fault detection method for compressor blade
Zheng et al. Wavelet packet decomposition and neural network based fault diagnosis for elevator excessive vibration
CN116992389B (en) False data detection method and system for Internet of things
WO2024065777A1 (en) Method, apparatus, electronic device, and storage medium for diagnosing industrial fault
Zhao et al. Health indicator selection and health assessment of rolling element bearing
CN116819318B (en) Motor fault detection method and system

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