CN109446625B - Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method - Google Patents

Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method Download PDF

Info

Publication number
CN109446625B
CN109446625B CN201811227688.1A CN201811227688A CN109446625B CN 109446625 B CN109446625 B CN 109446625B CN 201811227688 A CN201811227688 A CN 201811227688A CN 109446625 B CN109446625 B CN 109446625B
Authority
CN
China
Prior art keywords
data
dynamic threshold
fault data
calculating
probability density
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811227688.1A
Other languages
Chinese (zh)
Other versions
CN109446625A (en
Inventor
熊邦书
丛雷
李新民
雷鸰
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.)
Nanchang Hangkong University
Original Assignee
Nanchang Hangkong 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 Nanchang Hangkong University filed Critical Nanchang Hangkong University
Priority to CN201811227688.1A priority Critical patent/CN109446625B/en
Publication of CN109446625A publication Critical patent/CN109446625A/en
Application granted granted Critical
Publication of CN109446625B publication Critical patent/CN109446625B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • 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
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Complex Calculations (AREA)

Abstract

The invention discloses a Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method, which comprises the following steps: 1) preprocessing data, and filtering noise in the data; 2) extracting characteristics, namely extracting the characteristics of normal data and fault data respectively; 3) and calculating dynamic threshold values, namely calculating the dynamic threshold values of the normal data and the fault data by adopting a Bayesian inference method. The invention has the advantages that: in the feature extraction method, four methods of inner ring feature frequency energy extraction, outer ring feature frequency energy extraction, ball feature frequency energy extraction and M6A extraction are respectively adopted to extract features of normal data and fault data of different types, so that the vibration characteristics of the normal data and the fault data can be effectively reflected; in the dynamic threshold calculation, a Bayesian inference method is adopted, the probability density of normal data is considered, the probability density of different types of fault data is considered, and the calculation accuracy of the dynamic threshold is improved.

Description

Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method
Technical Field
The invention relates to the field of helicopter health monitoring and fault diagnosis, in particular to a Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method.
Background
Compared with a fixed wing airplane, the helicopter can take off and land vertically, has the characteristics of better maneuverability, quick response capability and no limitation by landforms and landforms, and can be widely applied to the civil and military fields.
With the wider application of helicopters, the safety of helicopters is more and more emphasized by people. Helicopter Health and Usage Monitoring System (HUMS) is an important technology and system for guaranteeing safe operation of helicopters in the world today. In order to reduce accidents, monitoring the working state of critical components, especially moving components, of a helicopter in real time is an important content of the HUMS. In HUMS, normal and fault data are collected through a fault implantation test under a test bed environment, a dynamic threshold value is obtained through calculation, and the dynamic threshold value is used for a real environment of a helicopter to monitor the health state of a moving part in real time.
At present, a weighted average method is adopted for calculating the dynamic threshold of a helicopter maneuvering part, the method only utilizes the characteristic mean value and the variance of normal data to calculate and obtain the dynamic threshold, and the method has the advantage of simple calculation, but does not utilize the characteristics of fault data and has the problem of low accuracy of the dynamic threshold.
Disclosure of Invention
The invention aims to provide a Bayesian inference-based dynamic threshold calculation method for helicopter maneuvering components, and aims to solve the problem that the dynamic threshold calculated in the prior art is low in accuracy when different types of fault data are targeted. The method of the invention comprises the following main steps:
1) the data preprocessing specifically comprises the following steps:
(1.1) collecting normal data, inner ring fault data, outer ring fault data and ball fault data;
and (1.2) preprocessing all the data acquired in the step (1.1) by using a data preprocessing method, and filtering noise in the data to obtain the denoised data.
The data preprocessing method is a singular value decomposition method, a wavelet de-noising method, a wavelet packet de-noising method, a sliding smoothing processing method, a basic form filtering method or a difference form filtering method.
2) The feature extraction specifically comprises the following steps:
(2.1) grouping the denoised normal data, the inner ring fault data, the outer ring fault data and the ball fault data obtained in the step 1);
and (2.2) extracting the features of all the data grouped in the step (2.1) by using a feature extraction method.
The characteristic extraction method comprises the steps of inner ring characteristic frequency energy extraction, outer ring characteristic frequency energy extraction, ball characteristic frequency energy extraction and M6A extraction.
3) And (3) calculating the dynamic threshold, namely calculating the probability density of the features extracted in the step 2), drawing a probability density curve according to the probability density, and finally calculating to obtain the dynamic threshold through a Bayesian inference method. The method specifically comprises the following steps:
(3.1) respectively calculating the probability densities of the normal data feature, the inner ring fault data feature, the outer ring fault data feature and the ball fault data feature by using the features extracted in the step 2);
(3.2) respectively drawing probability density curves of normal data features and inner ring fault data features on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is a dynamic threshold of the normal data and the inner ring fault data, and calculating the accuracy and the false alarm rate to judge the effectiveness of the calculation of the dynamic threshold;
(3.3) respectively drawing probability density curves of normal data characteristics and outer ring fault data characteristics on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is a dynamic threshold of the normal data and the outer ring fault data, and calculating the accuracy and the false alarm rate to judge the effectiveness of the calculation of the dynamic threshold;
(3.4) respectively drawing probability density curves of normal data features and ball fault data features on the same coordinate system, and calculating by a Bayesian reasoning method to obtain the probability density curves, wherein the intersection point of the two curves is the dynamic threshold of the normal data and the ball fault data, and calculating the accuracy and the false alarm rate to judge the effectiveness of the dynamic threshold calculation.
To verify the validity of the dynamic threshold calculation, the dynamic threshold test may be performed as follows.
4) The dynamic threshold test specifically comprises the following steps:
(4.1) collecting normal data, inner ring fault data, outer ring fault data and ball fault data for testing;
(4.2) preprocessing the data by adopting the same preprocessing method in the step (1.2), and filtering noise in the data to obtain denoised data;
(4.3) grouping the denoised normal data, the inner ring fault data, the outer ring fault data and the ball fault data obtained in the step (4.2);
(4.4) performing feature extraction on all the grouped data by adopting the same feature extraction method in the step (2.2);
(4.5) respectively calculating the probability densities of the normal data feature, the inner ring fault data feature, the outer ring fault data feature and the ball fault data feature by using the features extracted in the step (4.4);
(4.6) respectively drawing probability density curves of normal data characteristics and inner ring fault data characteristics on the same coordinate system, displaying the dynamic threshold calculated in the step (3.2) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of dynamic threshold calculation;
(4.7) respectively drawing probability density curves of the normal data characteristic and the outer ring fault data characteristic on the same coordinate system, displaying the dynamic threshold calculated in the step (3.3) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of calculation of the dynamic threshold;
(4.8) respectively drawing probability density curves of normal data characteristics and ball fault data characteristics on the same coordinate system, displaying the dynamic threshold calculated in the step (3.4) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of dynamic threshold calculation.
The invention has the advantages that: in the feature extraction method, four methods of inner ring feature frequency energy extraction, outer ring feature frequency energy extraction, ball feature frequency energy extraction and M6A extraction are respectively adopted to extract features of normal data and fault data of different types, so that the vibration characteristics of the normal data and the fault data can be effectively reflected; in the dynamic threshold calculation, a Bayesian inference method is adopted, the probability density of normal data is considered, the probability densities of different types of fault data are considered, and the calculation accuracy of the dynamic threshold is improved.
Drawings
FIG. 1 is a flow chart of the operation of the present invention.
Fig. 2 is a schematic view of a rolling bearing structure.
FIG. 3 is a graph of probability density for normal and fault data.
Detailed Description
The invention adopts a working flow chart of a dynamic threshold value calculation method of a helicopter maneuvering component based on Bayesian inference as shown in figure 1 to realize the calculation of the dynamic threshold value, and the specific implementation steps are as follows:
1) data pre-processing
(1.1) collecting normal data, inner ring fault data, outer ring fault data and ball fault data.
In the embodiment of the invention, the experimental data is normal and fault data acquired in a fault implantation mode.
And (1.2) preprocessing all the data acquired in the step (1.1) by using a data preprocessing method, and filtering noise in the data to obtain the denoised data.
The data preprocessing method is a singular value decomposition method, a wavelet de-noising method, a wavelet packet de-noising method, a sliding smoothing processing method, a basic form filtering method or a difference form filtering method.
2) Feature extraction
And (2.1) grouping the denoised normal data, the inner ring fault data, the outer ring fault data and the ball fault data obtained in the step 1).
In the embodiment of the invention, de-noised data are grouped according to the 1369 data points of one circle of rotation of the bearing.
And (2.2) performing feature extraction on all the data grouped in the step (2.1) by using a feature extraction method.
The characteristic extraction method comprises the steps of inner ring characteristic frequency energy extraction, outer ring characteristic frequency energy extraction, ball characteristic frequency energy extraction and M6A extraction.
In the embodiment of the invention, the specific implementation steps of inner ring characteristic frequency energy extraction, outer ring characteristic frequency energy extraction and ball characteristic frequency energy extraction are as follows:
(2.2.1) respectively calculating according to the formula (1), the formula (2) and the formula (3) to obtain the characteristic frequency of the inner ring, the characteristic frequency of the outer ring and the characteristic frequency of the ball of the bearing;
the rolling bearing structure is shown in FIG. 2, d is the diameter of the ball; alpha is a contact angle; d is the pitch diameter of the bearing; n is the rotational speed in revolutions per minute. According to different bearing fault parts, common fault frequencies include an inner ring fault frequency, an outer ring fault frequency and a ball fault frequency, and the corresponding characteristic frequency calculation formulas are as follows:
inner ring failure frequency:
Figure BDA0001836388790000041
outer ring fault frequency:
Figure BDA0001836388790000042
frequency of ball failure:
Figure BDA0001836388790000043
wherein N is the number of balls, f r The calculation formula is as follows for the fundamental frequency of the rotating shaft of the rolling bearing:
f r ＝n/60 (4)
in the embodiment of the invention, the value of D is 9.525mm, the value of alpha is 30 degrees, the value of D is 374.326mm, the value of N is 209rpm, and the value of N is 95.
(2.2.2) respectively finding the maximum values of the amplitudes of the inner ring characteristic frequency, the outer ring characteristic frequency and the ball characteristic frequency corresponding to the range of frequency multiplication (-1 to +1) in the spectrogram through Fast Fourier Transform (FFT) of all the data grouped in the step (2.1), and obtaining energy as the characteristic.
In the embodiment of the invention, the specific implementation steps of M6A extraction are as follows:
and (4) calculating all the data grouped in the step (2.1) according to the formula (5) to obtain the characteristics.
Let the vibration signal be { x i Calculation of 1,2,3, …, N, M6AThe formula is as follows:
Figure BDA0001836388790000051
in the formula,
Figure BDA0001836388790000052
is the average value of the vibration signal.
3) Dynamic threshold calculation
And (3.1) respectively calculating the probability densities of the normal data characteristic, the inner ring fault data characteristic, the outer ring fault data characteristic and the ball fault data characteristic by using the characteristics extracted in the step 2).
In the embodiment of the invention, the specific implementation steps of the probability density calculation are as follows:
(3.1.1) defining the length of a minimum interval, and dividing the difference between the characteristic average value and the characteristic minimum value into m parts as the minimum interval;
in the embodiment of the invention, the effect is best when m is 20-30;
and (3.1.2) calculating the number of features contained in each minimum interval in all the data.
And (3.1.3) calculating to obtain the probability density of each interval according to the probability of each interval.
And (3.2) respectively drawing probability density curves of the normal data characteristics and the inner ring fault data characteristics on the same coordinate system, and calculating by a Bayesian inference method, wherein the intersection point of the two curves is a dynamic threshold of the normal data and the inner ring fault data, and calculating the accuracy and the false alarm rate to judge the effectiveness of the dynamic threshold calculation.
In the embodiment of the invention, the Bayesian inference method describes that the prior information is combined with the sample information in a probability mode. Z represents characteristics including inner ring characteristic frequency energy, outer ring characteristic frequency energy, ball characteristic frequency energy and M6A, and X represents categories including normal and fault. When given Z, the posterior probability density function of X can be calculated by bayesian inference:
Figure BDA0001836388790000053
in the formula (6), p (x) represents the probability of normal data and fault data, both of which are 0.5; p (z) represents the probability of normal data or failure data under the feature; p (X | Z) is the conditional probability of X given Z, also called posterior probability, and represents the probability of normal or fault occurrence when a certain characteristic is given; p (Z | X) is the conditional probability of Z given X, also known as the likelihood estimate, which indicates the probability of the feature occurring at a certain normal or fault class. The basic concepts in bayesian inference statistics are prior distributions, likelihood functions and posterior distributions. The maximum posterior probability density shows that the intersection point of the two probability density curves of p (Z | X) is the optimal dynamic threshold t for obtaining the maximum accuracy.
In the embodiment of the present invention, a probability density curve is shown in fig. 3, where the accuracy is a True Positive Rate (TPR), and the TPR is TP/(TP + FN), which represents a ratio of actual examples in the positive class predicted by the classifier to all positive examples; the false alarm rate is a negative positive class rate (FPR), which is FP/(FP + TN), and represents a ratio of actual negative examples to all negative examples in the positive class predicted by the classifier.
(3.3) respectively drawing probability density curves of normal data characteristics and outer ring fault data characteristics on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is the dynamic threshold of the normal data and the outer ring fault data, and calculating the accuracy and the false alarm rate to judge the effectiveness of the dynamic threshold calculation.
And (3.4) respectively drawing probability density curves of the normal data characteristics and the ball fault data characteristics on the same coordinate system, and calculating by a Bayesian inference method, wherein the intersection point of the two curves is the dynamic threshold of the normal data and the ball fault data, and the accuracy and the false alarm rate are calculated to judge the effectiveness of the dynamic threshold calculation.
4) Dynamic threshold testing
And (4.1) acquiring normal data, inner ring fault data, outer ring fault data and ball fault data for testing.
And (4.2) preprocessing the data by adopting the same preprocessing method in the step (1.2), and filtering noise in the data to obtain the denoised data.
And (4.3) grouping the denoised normal data, the inner ring fault data, the outer ring fault data and the ball fault data obtained in the step (4.2).
And (4.4) performing feature extraction on all the grouped data by adopting the same feature extraction method in the step (2.2).
And (4.5) respectively calculating the probability densities of the normal data feature, the inner ring fault data feature, the outer ring fault data feature and the ball fault data feature by using the features extracted in the step (4.4).
And (4.6) respectively drawing probability density curves of normal data characteristics and inner ring fault data characteristics on the same coordinate system, displaying the dynamic threshold calculated in the step (3.2) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of the calculation of the dynamic threshold.
(4.7) respectively drawing probability density curves of normal data characteristics and outer ring fault data characteristics on the same coordinate system, displaying the dynamic threshold calculated in the step (3.3) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of dynamic threshold calculation.
(4.8) respectively drawing probability density curves of normal data characteristics and ball fault data characteristics on the same coordinate system, displaying the dynamic threshold calculated in the step (3.4) on the coordinate system, and calculating the accuracy and the false alarm rate to judge the effectiveness of dynamic threshold calculation.
In the embodiment of the invention, the dynamic threshold calculation method and the weighted average method of the helicopter maneuvering components based on Bayesian inference are compared and researched. Normal data used for dynamic threshold calculation, inner ring fault data of scratch 1.5, outer ring fault data of scratch 1.2 and ball fault data of scratch 1.3 are 1200000 points respectively, and normal data used for threshold test, inner ring fault data of scratch 1.5, outer ring fault data of scratch 1.2 and ball fault data of scratch 1.3 are 300000 points respectively.
And calculating and comparing dynamic threshold values of normal data and inner ring fault data. (a) Noise is removed by using a singular value decomposition method, inner ring characteristic frequency energy is used for characteristic extraction, the dynamic threshold value obtained by adopting a weighted average method is 4.44, and TPR and FPR obtained by testing data are respectively 98.91% and 47.70%; the dynamic threshold value calculated by the Bayesian inference method is 3.10, and the TPR and the FPR obtained by measuring data are 90.04% and 22.35% respectively. (b) Removing noise by using a singular value decomposition method, extracting features by using an M6A method, calculating by using a weighted average method to obtain a dynamic threshold value of 11.92, and obtaining TPR and FPR respectively of 98.82% and 7.43% through test data; the dynamic threshold value calculated by adopting a Bayesian inference method is 10.84, and the TPR and the FPR obtained by testing data are 96.67 percent and 1.47 percent respectively.
And calculating and comparing the normal data and the outer ring fault data by using dynamic threshold values. (a) Removing noise by using a singular value decomposition method, performing feature extraction by using outer ring feature frequency energy, calculating by using a weighted average method to obtain a dynamic threshold value of 1.71, and obtaining TPR and FPR which are 99.01% and 0 respectively through test data; the dynamic threshold value calculated by adopting a Bayesian inference method is 11.42, and the TPR and the FPR obtained by testing data are 100% and 0 respectively. (b) Noise is removed by using a singular value decomposition method, feature extraction is carried out by using an M6A method, the size of a dynamic threshold value obtained by adopting a weighted average method is 11.92, the size of TPR and FPR obtained through test data is 98.82% and 0 respectively, the size of a dynamic threshold value obtained through adopting a Bayesian inference method is 34.54, and the size of TPR and FPR obtained through test data is 100% and 0 respectively.
And calculating and comparing the normal data and the ball fault data by using dynamic threshold values. (a) Noise is removed by using a singular value decomposition method, characteristic extraction is carried out by using ball characteristic frequency energy, the size of a dynamic threshold value obtained by using a weighted average method is 2.14, and TPR and FPR obtained by testing data are respectively 98.44% and 50.53%; the dynamic threshold value calculated by adopting a Bayesian inference method is 1.23; the TPR and FPR obtained by the test data were 83.39% and 16.12%, respectively. (b) Noise is removed by using a singular value decomposition method, a M6A method is adopted for feature extraction, the dynamic threshold value calculated by using a weighted average method is 11.92, the TPR and the FPR obtained through test data are 98.82% and 31.08% respectively, the dynamic threshold value calculated by using a Bayesian inference method is 9.58, and the TPR and the FPR obtained through test data are 90.29% and 3.86% respectively.
As can be seen from the comparison experiments, the dynamic threshold calculated by the Bayesian inference method is slightly lower than the dynamic threshold calculated by the weighted average method in accuracy as a whole, but the false alarm rate is much lower, so that the calculation accuracy of the dynamic threshold is improved.

Claims (3)

1. A dynamic threshold calculation method for helicopter maneuvering components based on Bayesian inference comprises the following steps:
1) preprocessing data;
2) extracting characteristics;
3) calculating the probability density of the features extracted in the step 2), drawing a probability density curve according to the probability density, and finally calculating to obtain a dynamic threshold value through a Bayesian inference method;
wherein,
the step 1) of data preprocessing specifically comprises the following steps:
(1.1) collecting normal data, inner ring fault data, outer ring fault data and ball fault data;
(1.2) preprocessing all the data acquired in the step (1.1) by using a data preprocessing method, and filtering noise in the data to obtain denoised data;
step 2) the feature extraction specifically comprises the following steps:
(2.1) grouping the denoised normal data, inner ring fault data, outer ring fault data and ball fault data obtained in the step 1);
(2.2) extracting the features of all the data grouped in the step (2.1) by using a feature extraction method;
step 3) the dynamic threshold calculation specifically comprises the following steps:
(3.1) respectively calculating the probability densities of the normal data characteristic, the inner ring fault data characteristic, the outer ring fault data characteristic and the ball fault data characteristic by using the characteristics extracted in the step 2);
(3.2) respectively drawing probability density curves of the normal data characteristics and the inner ring fault data characteristics on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is the dynamic threshold of the normal data and the inner ring fault data;
(3.3) respectively drawing probability density curves of normal data characteristics and outer ring fault data characteristics on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is a dynamic threshold of the normal data and the outer ring fault data;
(3.4) respectively drawing probability density curves of normal data characteristics and ball fault data characteristics on the same coordinate system, and calculating by using a Bayesian inference method, wherein the intersection point of the two curves is a dynamic threshold of the normal data and the ball fault data;
the probability density calculation of step (3.1) comprises the following steps:
(3.1.1) defining the length of a minimum interval, dividing the difference between the characteristic average value and the characteristic minimum value into m parts as the minimum interval, wherein m is 20-30;
(3.1.2) calculating the number of the features contained in each minimum interval in all the data;
and (3.1.3) calculating the probability density of each interval according to the probability of each interval.
2. The method of claim 1, wherein the data preprocessing method is a singular value decomposition method, a wavelet de-noising method, a wavelet packet de-noising method, a sliding smoothing method, a fundamental morphological filtering method, or a differential morphological filtering method.
3. The method of claim 2, wherein the feature extraction methods are inner ring feature frequency energy extraction, outer ring feature frequency energy extraction, ball feature frequency energy extraction, and M6A extraction.
CN201811227688.1A 2018-10-22 2018-10-22 Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method Active CN109446625B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811227688.1A CN109446625B (en) 2018-10-22 2018-10-22 Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811227688.1A CN109446625B (en) 2018-10-22 2018-10-22 Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method

Publications (2)

Publication Number Publication Date
CN109446625A CN109446625A (en) 2019-03-08
CN109446625B true CN109446625B (en) 2022-07-26

Family

ID=65547512

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811227688.1A Active CN109446625B (en) 2018-10-22 2018-10-22 Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method

Country Status (1)

Country Link
CN (1) CN109446625B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114104332B (en) * 2021-11-19 2023-09-22 中国直升机设计研究所 Method and device for acquiring state index dynamic threshold of helicopter moving part

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2344908A1 (en) * 2000-07-20 2002-01-20 Siemens Building Technologies, Inc. Model based fault detection and diagnosis methodology for hvac subsystems
CN102842105A (en) * 2012-07-09 2012-12-26 中国电力科学研究院 Online transient state stability risk evaluating method for metering wind power uncertainty
CN103684910A (en) * 2013-12-02 2014-03-26 北京工业大学 Abnormality detecting method based on industrial control system network traffic
CN103995985A (en) * 2014-06-10 2014-08-20 华东理工大学 Fault detection method based on Daubechies wavelet transform and elastic network
CN104677632A (en) * 2015-01-21 2015-06-03 大连理工大学 Rolling bearing fault diagnosis method using particle filtering and spectral kurtosis
CN104931255A (en) * 2015-06-04 2015-09-23 中国航空工业集团公司上海航空测控技术研究所 Method for evaluating whether fault feature parameter of gearbox good or bad
CN108304287A (en) * 2018-01-22 2018-07-20 腾讯科技（深圳）有限公司 A kind of disk failure detection method, device and relevant device

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2344908A1 (en) * 2000-07-20 2002-01-20 Siemens Building Technologies, Inc. Model based fault detection and diagnosis methodology for hvac subsystems
CN102842105A (en) * 2012-07-09 2012-12-26 中国电力科学研究院 Online transient state stability risk evaluating method for metering wind power uncertainty
CN103684910A (en) * 2013-12-02 2014-03-26 北京工业大学 Abnormality detecting method based on industrial control system network traffic
CN103995985A (en) * 2014-06-10 2014-08-20 华东理工大学 Fault detection method based on Daubechies wavelet transform and elastic network
CN104677632A (en) * 2015-01-21 2015-06-03 大连理工大学 Rolling bearing fault diagnosis method using particle filtering and spectral kurtosis
CN104931255A (en) * 2015-06-04 2015-09-23 中国航空工业集团公司上海航空测控技术研究所 Method for evaluating whether fault feature parameter of gearbox good or bad
CN108304287A (en) * 2018-01-22 2018-07-20 腾讯科技（深圳）有限公司 A kind of disk failure detection method, device and relevant device

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
优化神经网络在矿用电机轴承故障诊断的应用;王煜 等;《安徽理工大学学报(自然科学版)》;20180715;30-35 *
周期能量与优化LMD结合的轴承故障诊断方法;熊邦书 等;《振动.测试与诊断》;20160227;第36卷(第02期);372-377+407 *
基于应力波分析的状态监控与故障预测研究;吴天舒 等;《仪器仪表学报》;20171215;第38卷(第12期);3061-3070 *

Also Published As

Publication number Publication date
CN109446625A (en) 2019-03-08

Similar Documents

Publication Publication Date Title
Sun et al. Bearing fault diagnosis based on EMD and improved Chebyshev distance in SDP image
CN107956708B (en) A kind of potential cavitation fault detection method of pump based on quick spectrum kurtosis analysis
Rai et al. A review on signal processing techniques utilized in the fault diagnosis of rolling element bearings
CN110135492B (en) Equipment fault diagnosis and abnormality detection method and system based on multiple Gaussian models
Liao et al. Bearing fault feature enhancement and diagnosis based on statistical filtering and 1.5-dimensional symmetric difference analytic energy spectrum
CN111120348A (en) Centrifugal pump fault early warning method based on support vector machine probability density estimation
CN110514441A (en) A kind of Fault Diagnosis of Roller Bearings based on vibration signal denoising and Envelope Analysis
CN108760327A (en) A kind of diagnostic method of aeroengine rotor failure
CN104677632A (en) Rolling bearing fault diagnosis method using particle filtering and spectral kurtosis
CN103902844A (en) Transformer vibration signal de-noising method based on EEMD kurtosis threshold value
CN108760316A (en) Information fusion method is joined in the change of variation mode decomposition
Han et al. Roller bearing fault diagnosis based on LMD and multi-scale symbolic dynamic information entropy
CN104198187A (en) Mechanical vibration fault characteristic time domain blind extraction method
CN111170103B (en) Equipment fault identification method
CN111238813B (en) Method for extracting fault features of rolling bearing under strong interference
Duan et al. Adaptive morphological analysis method and its application for bearing fault diagnosis
CN110987433B (en) Bearing fault early warning method based on high-frequency signal characteristic amplitude
CN104215456A (en) Plane clustering and frequency-domain compressed sensing reconstruction based mechanical fault diagnosis method
CN108444715A (en) Bearing state diagnostic method, device, storage medium and electronic equipment
CN109446625B (en) Bayesian inference-based helicopter maneuvering component dynamic threshold calculation method
CN108801634A (en) The method and its application of bearing fault characteristics frequency are extracted based on singular value decomposition and the frequency band entropy of optimization
CN109612726A (en) A kind of multiple superstage analysis method extracted for vibration signal characteristics
Yang et al. A new singular spectrum decomposition method based on cao algorithm and amplitude aware permutation entropy
Tan et al. A new method for machinery fault diagnoses based on an optimal multiscale morphological filter
Serbes et al. Embolic Doppler ultrasound signal detection using modified dual tree complex wavelet transform

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