CN114970645B - Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation - Google Patents

Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation Download PDF

Info

Publication number
CN114970645B
CN114970645B CN202210889066.5A CN202210889066A CN114970645B CN 114970645 B CN114970645 B CN 114970645B CN 202210889066 A CN202210889066 A CN 202210889066A CN 114970645 B CN114970645 B CN 114970645B
Authority
CN
China
Prior art keywords
data
vibration data
fault diagnosis
vibration
sampling
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
CN202210889066.5A
Other languages
Chinese (zh)
Other versions
CN114970645A (en
Inventor
孙黎
何慧虹
尚舵
景浩
高宇阳
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.)
China Industrial Internet Research Institute
Original Assignee
China Industrial Internet Research Institute
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 China Industrial Internet Research Institute filed Critical China Industrial Internet Research Institute
Priority to CN202210889066.5A priority Critical patent/CN114970645B/en
Publication of CN114970645A publication Critical patent/CN114970645A/en
Application granted granted Critical
Publication of CN114970645B publication Critical patent/CN114970645B/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
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • 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
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
    • 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
    • G06N3/045—Combinations of 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
    • G06N3/047—Probabilistic or stochastic 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
    • G06N3/048—Activation functions
    • 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
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C5/00—Registering or indicating the working of vehicles
    • G07C5/004—Indicating the operating range of the engine
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C5/00—Registering or indicating the working of vehicles
    • G07C5/008—Registering or indicating the working of vehicles communicating information to a remotely located station
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
    • H04W4/42—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for mass transport vehicles, e.g. buses, trains or aircraft
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
    • H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching
    • 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
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Signal Processing (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

The invention provides a civil aircraft engine fault diagnosis system and method based on 5G Bian Yun cooperation, which relates to the technical field of civil aircraft engines, and comprises the following steps: the edge end is used for carrying out full sampling on the vibration signal of the aircraft engine and carrying out undersampling on the basis of the observation matrix to respectively obtain full-sampling vibration data and compressed sampling vibration data; the cloud platform is used for receiving the full-sampling vibration data and the compression sampling vibration data, obtaining the observation matrix based on the full-sampling vibration data, selecting a dictionary, outputting the observation matrix to the edge end, reconstructing the compression sampling vibration data through the dictionary to obtain reconstructed data, and performing fault diagnosis based on the reconstructed data; and the 5G ATG base station is used for signal transmission between the edge end and the cloud platform. The fault diagnosis system and the fault diagnosis method can effectively solve the problem of signal transmission.

Description

Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation
Technical Field
The invention relates to the technical field of civil aircraft engines, in particular to a civil aircraft engine fault diagnosis system and method based on 5G Bian Yun cooperation.
Background
An aircraft engine is the heart of an aircraft. Reliable and stable operation of the engine is directly related to flight safety, and serious engine failure can cause serious flight safety accidents. The aero-engine is used as an efficient thermodynamic machine, is complex in structural form, high in working temperature and prone to generating various faults. An Engine Health Management (EHM) system supports reliable and stable operation of an engine as an effective means for engine state monitoring and fault diagnosis, which are core contents thereof.
The existing EHM is composed of an airborne subsystem and a ground subsystem. For vibration state monitoring, an airborne subsystem is responsible for real-time operation and judgment in an air flight state, the airborne computational power is limited, the general monitoring criterion is monotonous vibration total threshold judgment, and the specific cause of the fault cannot be diagnosed in detail (for example, when the vibration value of a low-pressure rotor reaches 6 units or the vibration value of a high-pressure rotor reaches 4.2 units, a high-vibration alarm of a cockpit is triggered); the ground subsystem is responsible for offline analysis of the aircraft engine in a ground state, and vibration fault diagnosis of the engine is completed by means of big data analysis and the like under the support of sufficient computing resources. Thus, the latter is a functional complement to the former. Under the background that the flight safety is gradually improved, the interconnection, intercommunication and fusion development between the two is a necessary technical trend.
However, vibration monitoring of the engine faces the following problems in terms of communication between the onboard subsystem and the ground subsystem: due to the high rotational speed of the engine, the vibration sensor often needs a high sampling frequency to capture the fault information based on the nyquist sampling theorem, which results in a large amount of vibration data. At present, the networking communication mode of the airplane mainly adopts very high frequency, high frequency and GEO means, has high communication time delay, low speed and high cost, and is mainly used for key information transmission. Therefore, the vibration data is difficult to be effectively transmitted back to the ground subsystem from the machine-mounted subsystem due to the existing acquisition mode and communication mode, so that the possibility of online diagnosis of the vibration fault is limited.
Therefore, there is a need to provide a signal diagnosis system to solve the problem of signal transmission between vibration monitoring and fault diagnosis.
Disclosure of Invention
Solves the technical problem
Aiming at the defects in the prior art, the invention provides a civil aircraft engine fault diagnosis system based on 5G Bian Yun cooperation, and the fault diagnosis system and method can effectively solve the problem of signal transmission.
Technical scheme
In order to achieve the purpose, the invention is realized by the following technical scheme:
the invention provides a civil aviation engine fault diagnosis system based on 5G Bian Yun cooperation, which comprises:
the edge end is arranged on the aircraft engine and used for carrying out full sampling on the vibration signal of the aircraft engine and carrying out undersampling on the basis of the observation matrix to respectively obtain full-sampling vibration data and compression sampling vibration data;
the cloud platform is arranged on the ground, is used for receiving the full-sampling vibration data and the compression sampling vibration data, obtaining the observation matrix based on the full-sampling vibration data, selecting a dictionary, outputting the observation matrix to the edge end, and is also used for reconstructing signals of the compression sampling vibration data through the dictionary to obtain reconstruction data and performing fault diagnosis based on the reconstruction data;
and the 5G ATG base station is arranged on the ground and is used for signal transmission between the edge end and the cloud platform.
Optionally, the edge end at least includes a vibration acceleration sensor, a data acquisition device, a storage unit and a communication management module, the vibration acceleration sensor is used for measuring the vibration of the aircraft engine, the data acquisition device is used for obtaining the full-sampling vibration data and the compression sampling vibration data according to the measurement of the vibration acceleration sensor, the storage unit is used for storing the full-sampling vibration data and the compression sampling vibration data, and the communication management module is used for transmitting the full-sampling vibration data and the compression sampling vibration data to the 5G ATG base station through a 5G ATG transmission technology.
The application also provides an aeroengine fault diagnosis method based on 5G Bian Yun cooperation, and the aeroengine fault diagnosis system based on 5G Bian Yun cooperation, which is provided by the application, comprises the following steps:
s1, first Bian Yun in synergy: the edge end carries out full sampling on vibration signals of an aircraft engine to obtain full sampling vibration data, the full sampling vibration data are output to the 5G ATG base station, the 5G ATG base station outputs the full sampling vibration data to the cloud end platform, the cloud end platform obtains an observation matrix based on the full sampling vibration data and selects a dictionary, the observation matrix is output to the edge end through the 5G ATG base station, the dictionary is stored, and the first-time edge cloud is finished in a coordinated mode;
s2, second Bian Yun synergy: the edge terminal is according to the vibration signal of observing the matrix to aeroengine is undersampled, obtains compression sampling vibration data, and will compression sampling vibration data pass through 5G ATG basic station is exported for the high in the clouds platform, the high in the clouds platform is according the dictionary is right compression sampling vibration data carries out signal reconstruction, obtains the reconsitution data, and based on the reconsitution data carries out fault diagnosis, passes through the diagnostic result 5G ATG basic station is exported for the edge terminal, and the second time cloud is in coordination over.
Optionally, step S1 is preceded by setting a criterion for determining system start, so that when the civil aircraft engine fault diagnosis system based on 5G Bian Yun cooperation meets the criterion, the first Bian Yun is cooperatively started.
Optionally, the judgment criterion is based on a vibration value and time.
Optionally, the selecting dictionary refers to a selecting sparse matrix, and is used for sparsely representing signals in a certain domain space of the sparse matrix;
wherein, the scope of selecting is: the common transformation orthogonal basis is stored in the cloud platform; the selection criterion is as follows: and selecting the dictionary with the maximum signal compression rate as the orthogonal base dictionary in the updating period under the constraint of setting the reconstruction precision signal.
Optionally, the cloud platform performs signal reconstruction on the compression sampling vibration data according to the dictionary, and specifically includes:
an orthogonal matching pursuit algorithm is adopted as a reconstruction algorithm, signal reconstruction refers to a process of reconstructing a sparse signal x with the length of N by a measurement vector y with the length of M, and a signal reconstruction result is as follows:
Figure 363647DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 223631DEST_PATH_IMAGE002
in order to obtain the sparse coefficient, the method comprises the following steps,
Figure 651201DEST_PATH_IMAGE003
is an estimate of the original signal.
Optionally, the cloud platform performs fault diagnosis based on the reconstruction data, wherein a network framework of the fault diagnosis is a two-dimensional convolutional neural network adopting supervised learning, and a data set used for training the two-dimensional convolutional neural network is vibration data of the aircraft engine of the same model during flight and fault diagnosis result data of flight maintenance.
Optionally, the input layer of the two-dimensional convolutional neural network is a two-dimensional data square matrix formed by drawing a vibration time sequence waveform signal, the two-dimensional data square matrix is expressed as a single-channel gray image, the network middle layer is formed by alternately stacking a convolutional layer, a pooling layer and an activation function layer, and the output layer is formed by a full connection layer and a Soft max layer.
Advantageous effects
1) According to the technical scheme provided by the invention, a set of edge end-cloud platform cooperative fault diagnosis system environment is established, the fault diagnosis of the civil aircraft engine under a side cloud cooperative mechanism is realized, and the traditional civil aircraft engine fault diagnosis system and a network transmission architecture thereof are updated;
2) The scheme can better guide the civil aviation engine to carry out on-line diagnosis on the mechanical system fault which needs to be emergently disposed and has abnormal vibration without exceeding the total vibration amount, and supports the maintenance of the engine according to the condition;
3) In the aspect of transmission, after a compressed sensing technology is used and data compression based on signal sparse representation is introduced, the pressure on acquisition equipment can be reduced, and the memory space and energy consumption required in the storage and space transmission process are reduced. Meanwhile, a 5G ATG transmission technology is introduced into an air-ground data transmission chain, air-ground network full coverage is realized by means of infrastructures such as a ground special base station and the like, and necessary conditions are provided for the transmission of a large amount of real-time sensing data of a civil aircraft engine;
4) In the aspect of fault diagnosis, the scheme provides an end-to-end algorithm utilizing an original time domain vibration signal, and the intelligent diagnosis algorithm based on a convolutional neural network is used for automatically completing feature extraction and fault recognition, so that the traditional intelligent diagnosis algorithm of 'feature extraction + classifier' based on signal processing is avoided, and the problems of high requirement on expert experience, time-consuming algorithm design and poor universality are solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
Fig. 1 is a schematic diagram of a civil aircraft engine fault diagnosis system based on 5G Bian Yun cooperation according to an embodiment of the present invention;
fig. 2 is a flowchart of a civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a two-dimensional convolutional neural network in a civil aircraft engine fault diagnosis method based on 5G Bian Yun synergy according to an embodiment of the present invention;
fig. 4 is a schematic diagram of two-dimensional convolution operation in a civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation according to an embodiment of the present invention;
fig. 5 is a schematic diagram of maximum pooling of a two-dimensional convolutional neural network in a civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The airborne vibration monitoring system of the existing civil aircraft engine defaults to use a low-pressure or high-pressure rotor vibration value as a fault judgment standard, and takes a CFM56-5B engine as an example, and a vibration total quantity limit value set by a manufacturer is that the low-pressure rotor vibration value reaches 6 units (1 unit is mils) or the high-pressure rotor vibration value reaches 4.2 units. After the value is exceeded, the airborne system gives an alarm to display that the vibration condition is abnormal, and if the vibration condition is serious, the pilot needs to close the engine to ensure the safety of the airplane and land and overhaul as soon as possible.
Referring to fig. 1, an embodiment of the present invention provides a civil aircraft engine fault diagnosis system based on 5G Bian Yun cooperation, including: the edge end is arranged on the aircraft engine and used for carrying out full sampling on the vibration signal of the aircraft engine and carrying out undersampling on the basis of the observation matrix to respectively obtain full-sampling vibration data and compression sampling vibration data; the cloud platform is arranged on the ground, is used for receiving the full-sampling vibration data and the compression-sampling vibration data, obtaining an observation matrix based on the full-sampling vibration data, selecting a dictionary, outputting the observation matrix to an edge end, and is also used for performing signal reconstruction on the compression-sampling vibration data through the dictionary to obtain reconstructed data and performing fault diagnosis based on the reconstructed data; and the 5G ATG base station is arranged on the ground and used for signal transmission between the edge end and the cloud platform.
In this embodiment, the edge terminal at least includes a vibration acceleration sensor, a data acquisition device, a storage unit and a communication management module. The 5G ATG base station should at least comprise an active antenna unit. The cloud platform at least comprises an application system server, a database server and a data processing server. The vibration acceleration sensor is used for measuring vibration of the aircraft engine, the data acquisition device is used for obtaining full-sampling vibration data and compression sampling vibration data according to measurement of the vibration acceleration sensor, the storage unit is used for storing the full-sampling vibration data and the compression sampling vibration data, and the communication management module is used for transmitting the full-sampling vibration data and the compression sampling vibration data to the 5G ATG base station through a 5G ATG transmission technology.
Based on the same invention concept, the invention also provides an aircraft engine fault diagnosis method based on 5G Bian Yun cooperation, and the aircraft engine fault diagnosis system based on 5G Bian Yun cooperation, which uses any one of the methods, comprises the following steps:
s1, first Bian Yun synergy: the method comprises the steps that the edge end carries out full sampling on vibration signals of the aero-engine to obtain full sampling vibration data, the full sampling vibration data are output to a 5G ATG base station, the 5G ATG base station outputs the full sampling vibration data to a cloud end platform, the cloud end platform obtains an observation matrix and selects a dictionary based on the full sampling vibration data, the observation matrix is output to the edge end through the 5G ATG base station, the dictionary is stored, and the first-time edge cloud is finished in a coordinated mode;
s2, second Bian Yun synergy: the method comprises the steps that the edge end carries out undersampling on vibration signals of the aircraft engine according to an observation matrix to obtain compression sampling vibration data, the compression sampling vibration data are output to a cloud platform through a 5G ATG base station, the cloud platform carries out signal reconstruction on the compression sampling vibration data according to a dictionary to obtain reconstruction data, fault diagnosis is carried out on the basis of the reconstruction data, a diagnosis result is output to the edge end through the 5G ATG base station, and secondary side cloud cooperation is finished.
In specific implementation, after the system is started, the edge end of the aircraft engine is deployed through the 5G ATG base station on the ground, and the cloud end platform of the ground subsystem is deployed between the edge end of the aircraft engine and the cloud end platform of the ground subsystem to perform secondary side cloud cooperation through a 5G ATG transmission technology.
In this embodiment, under the first secondary side cloud coordination, the edge end performs full sampling on a vibration signal of the aircraft engine to obtain full-sampling vibration data, and outputs the full-sampling vibration data to the 5G ATG base station. And after receiving the full sampling vibration data, the 5G ATG base station forwards the full sampling vibration data to the cloud platform. The cloud platform designs an observation matrix and selects a dictionary according to the full sampling vibration data, wherein the selected observation matrix is generally a Gaussian random matrix and needs to meet RIP characteristics. The selection dictionary is a selection sparse matrix, and aims to enable signals to be sparsely represented in a certain domain space of the sparse matrix, wherein the selection range is as follows: performing domain transformation on the full sampling signal by common transformation orthogonal bases stored in the cloud platform, such as a Fourier transformation base, a discrete cosine transformation base, a wavelet base and the like to obtain a sparse signal; the selection criterion is as follows: and selecting the dictionary with the maximum signal compression rate as the orthogonal base dictionary in the updating period under the constraint of better signal reconstruction performance (for example, the reconstruction accuracy is 95%).
In the present embodiment, the signal compression ratio is defined as:
Figure 978277DEST_PATH_IMAGE004
wherein, CR is the signal compression ratio, num represents the number of points of the original signal before compression, num' is the number of points of the signal after compression sampling, and in addition, the selected dictionary must satisfy irrelevant to the observation matrix.
The mathematical expression of the signal compression sampling process is as follows:
Figure 465891DEST_PATH_IMAGE005
wherein y ∈ R M In order to compress the sampled signal(s),
Figure 19363DEST_PATH_IMAGE006
in order to observe the matrix, the system,
Figure 985045DEST_PATH_IMAGE007
the segment is the fully sampled signal.
Figure 369890DEST_PATH_IMAGE008
Wherein the content of the first and second substances,
Figure 497246DEST_PATH_IMAGE009
in the form of a matrix of dictionaries,
Figure 662648DEST_PATH_IMAGE010
the sparse coefficients under the dictionary.
The compressed sensing model of the segment of the fully sampled signal is as follows:
Figure 432021DEST_PATH_IMAGE011
wherein the content of the first and second substances,
Figure 871705DEST_PATH_IMAGE012
is a perceptual matrix.
The signal reconstruction is to solve θ, which can be converted into an optimization problem for solving the norm, namely:
Figure 169962DEST_PATH_IMAGE013
the reconstruction accuracy is expressed by using normalized root mean square error, and the formula is as follows:
Figure 494764DEST_PATH_IMAGE014
where NMSE represents the reconstruction accuracy, representing the difference between the reconstructed signal and the original signal.
In this embodiment, the cloud platform stores the selected dictionary, and sends the observation matrix to the 5G ATG base station located on the ground. And the 5G ATG base station transmits the received observation matrix to a communication management module of the edge end by utilizing a 5G ATG technology, and the first cooperation is finished.
In the embodiment, under the second minor edge cloud cooperation,
the marginal end is according to observing the matrix and is carrying out the undersampling to civil aviation engine's vibration signal, obtains compression sampling vibration data to export compression sampling vibration data for the high in the clouds platform through 5G ATG basic station, the high in the clouds platform carries out signal reconstruction to compression sampling vibration data according to the dictionary, specifically includes:
an orthogonal matching pursuit algorithm is adopted as a reconstruction algorithm, signal reconstruction refers to a process of reconstructing a sparse signal x with the length of N by a measurement vector y with the length of M, and a signal reconstruction result is as follows:
Figure 536669DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 364948DEST_PATH_IMAGE002
in order to obtain the sparse coefficient, the method comprises the following steps of,
Figure 302948DEST_PATH_IMAGE003
is an estimate of the original signal.
Specifically, an Orthogonal Matching Pursuit (OMP) algorithm is used as a reconstruction algorithm, the algorithm needs sparsity as a preset parameter, and the iteration number is determined according to the sparsity of the signal. And acquiring the reconstructed data, performing fault diagnosis based on the reconstructed data, outputting a diagnosis result to the edge end through the 5G ATG base station, and finishing the second-time edge cloud cooperation.
In this embodiment, referring to fig. 3, the cloud platform performs fault diagnosis based on the reconstructed data, where a network framework for fault diagnosis is a two-dimensional convolutional neural network for supervised learning, and a data set used for training the two-dimensional convolutional neural network is vibration data of an aircraft engine of the same model during flight and fault diagnosis result data of the flight maintenance. The input layer of the two-dimensional convolutional neural network is a two-dimensional data matrix formed by drawing vibration time sequence waveform signals and expressed as a single-channel gray image, the network middle layer is formed by alternately stacking a convolutional layer, a pooling layer and an activation function layer, and the output layer is formed by a full connection layer and a Softmax layer.
Specifically, the input layer functions to reduce the influence of dimension and singular samples, and is normalized by a z-score method (see the formula in particular) for the reconstructed signal.
Figure 115047DEST_PATH_IMAGE015
Wherein the content of the first and second substances,
Figure 488872DEST_PATH_IMAGE016
is a normalized signal that is a function of,
Figure 702815DEST_PATH_IMAGE017
is the original vibration signal, u is the mean value of the original vibration signal, and σ is the standard deviation of the original vibration signal. The normalized signals form a two-dimensional data matrix of size m (m ∈ N) in terms of timing.
Referring to fig. 4, the convolutional layer is used to extract data features, a convolutional kernel with a size of p × q (p, q < m & p, q ∈ N) slides on a two-dimensional data matrix by a certain step length(s), and performs dot product with an area covered by the convolutional kernel to complete feature extraction, and the specific calculation method is as follows.
Figure 77296DEST_PATH_IMAGE018
The convolution process represented by the above formula can represent the operation method that the layer I is the convolution layer. Wherein, the input of the jth neuron of the l + layer is represented; is an activation function; m represents the number of feature maps; is the output of the jth neuron at layer l; representing a convolution operation; a convolution kernel representing the ith neuron of the l-th layer and the jth neuron of the l + 1-th layer; b denotes a bias.
The function of the activation function is to introduce a nonlinear factor to enable the network to have complex logic judgment capability, and a ReLU function form is adopted here, and a specific calculation formula is as follows.
Figure 439007DEST_PATH_IMAGE019
Referring to fig. 5, the activation function is used to introduce a non-linear factor to make the network have a complex logic judgment capability, and here, the activation function is in the form of a ReLU function, and a specific calculation formula thereof is as follows.
Figure 353873DEST_PATH_IMAGE019
The fully-connected layer is used for converting the two-dimensional matrix obtained at the upstream into a one-dimensional matrix according to the row sequence. The role of the Softmax layer is to calculate the probability of the failure class occurring based on the data of the fully-connected layer. Here with a common rotor system vibration failure: and if the rotor is unbalanced, the rotor is not centered and the radial collision and abrasion faults are taken as examples, the Softmax layer outputs probability values of the three fault states and the normal state. The convolutional neural network approaches the optimal weight and bias of a convolutional kernel under the minimum cross entropy by a gradient descent method.
In this embodiment, referring to fig. 2, step S1 is preceded by setting a criterion for determining system start, so that Bian Yun is cooperatively turned on for the first time when the civil aircraft engine fault diagnosis system based on 5G Bian Yun cooperation meets the criterion. Wherein the criterion is based on the vibration value and the time. The system related by the invention is started under the condition that the vibration value is close to the fault standard threshold value, so that the judgment on potential faults is realized, and the pilot is assisted to carry out emergency treatment. The judgment criteria (or trigger logic) for system startup are as follows: and starting the edge cloud coordination mechanism when the low-pressure rotor vibration value is greater than 3 units for 60 seconds continuously or greater than 4 units for 10 seconds continuously. The combined time and vibration limit will effectively filter short term vibration rise conditions that occur when the engine is subjected to air flow fluctuations, icing, foreign object ingestion and acceleration and deceleration, and the like. To accommodate dynamic changes in the operating parameters and environment of the civil aircraft engine, the guidelines set the update interval to T0.
The method has the advantages that a set of edge end-cloud end platform cooperative fault diagnosis system environment is established, the fault diagnosis of the civil aircraft engine under the edge cloud cooperative mechanism is realized, and the traditional civil aircraft engine fault diagnosis system and the network transmission architecture thereof are updated; furthermore, the scheme can better guide the vibration of the civil aviation engine to be abnormal without exceeding the total vibration limit, and the mechanical system fault needing urgent treatment is diagnosed on line to support the visual maintenance of the engine; furthermore, in the aspect of transmission, after a compressed sensing technology is utilized and data compression based on signal sparse representation is introduced, the pressure on acquisition equipment can be reduced, and the memory space and energy consumption required in the storage and air-ground transmission processes are reduced. Meanwhile, a 5G ATG transmission technology is introduced into an air-ground data transmission chain, air-ground network full coverage is realized by means of infrastructures such as a ground special base station and the like, and necessary conditions are provided for the transmission of a large amount of real-time sensing data of a civil aircraft engine; finally, in the aspect of fault diagnosis, the scheme provides an end-to-end algorithm utilizing an original time domain vibration signal, and the intelligent diagnosis algorithm based on the convolutional neural network is used for automatically completing feature extraction and fault recognition, so that the traditional intelligent diagnosis algorithm of 'feature extraction plus classifier' based on signal processing is avoided, and the problems of high requirement on expert experience, time-consuming algorithm design and poor universality are solved.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and the modifications or the substitutions do not cause the essence of the corresponding technical solutions to depart from the scope of the technical solutions of the embodiments of the present invention.

Claims (4)

1. A civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation is characterized by comprising an aircraft engine fault diagnosis system based on 5G Bian Yun cooperation, wherein the diagnosis system comprises:
the edge end is arranged on the aircraft engine and used for carrying out full sampling on the vibration signal of the aircraft engine and carrying out undersampling on the basis of the observation matrix to respectively obtain full-sampling vibration data and compression sampling vibration data; the edge end at least comprises a vibration acceleration sensor, a data acquisition device, a storage unit and a communication management module, wherein the vibration acceleration sensor is arranged on the aircraft engine and used for measuring the vibration of the aircraft engine, the data acquisition device is used for obtaining full sampling vibration data and compression sampling vibration data according to the measurement of the vibration acceleration sensor, the storage unit is used for storing the full sampling vibration data and the compression sampling vibration data, and the communication management module is used for transmitting the full sampling vibration data and the compression sampling vibration data to the 5G ATG base station through a 5G ATG transmission technology;
the cloud platform is arranged on the ground, is used for receiving the full-sampling vibration data and the compression sampling vibration data, obtaining the observation matrix based on the full-sampling vibration data, selecting a dictionary, outputting the observation matrix to the edge end, reconstructing the compression sampling vibration data through the dictionary to obtain reconstructed data, and diagnosing faults based on the reconstructed data;
the 5G ATG base station is arranged on the ground and is used for signal transmission between the edge end and the cloud platform;
the diagnostic method comprises the following steps:
s1, setting a judgment criterion of system starting so that Bian Yun is started cooperatively for the first time when the aeroengine fault diagnosis system based on 5G Bian Yun cooperation meets the judgment criterion;
s2, first Bian Yun synergy: the edge end carries out full sampling on vibration signals of an aircraft engine to obtain full sampling vibration data, the full sampling vibration data are output to the 5G ATG base station, the 5G ATG base station outputs the full sampling vibration data to the cloud end platform, the cloud end platform obtains an observation matrix based on the full sampling vibration data and selects a dictionary, the observation matrix is output to the edge end through the 5G ATG base station, the dictionary is stored, and the first-time edge cloud is finished in a coordinated mode;
the selection dictionary refers to a selection sparse matrix and is used for enabling signals to be expressed sparsely in a certain domain space of the sparse matrix; wherein, the scope of choosing does: the common transformation orthogonal basis is stored by the cloud platform; the selection criterion is as follows: selecting a dictionary with the maximum signal compression ratio as an orthogonal base dictionary in an updating period under the constraint of setting a reconstruction precision signal;
s3, second Bian Yun synergy: the edge end performs undersampling on vibration signals of the aircraft engine according to the observation matrix to obtain compression sampling vibration data, the compression sampling vibration data are output to the cloud platform through the 5G ATG base station, the cloud platform performs signal reconstruction on the compression sampling vibration data according to the dictionary to obtain reconstruction data, fault diagnosis is performed on the basis of the reconstruction data, a diagnosis result is output to the edge end through the 5G ATG base station, and secondary edge cloud cooperation is finished;
s4, according to a side cloud cooperation judgment criterion, secondary side cloud cooperation is carried out every time an update period runs, the observation matrix and the dictionary are updated accordingly, a complete signal full sampling, undersampling and reconstruction process is executed, the process is repeated in a circulating mode, after the initial criterion condition is met and Bian Yun is cooperatively started, the update period can be adjusted according to the situation, the frequency of the side cloud cooperation is reduced or improved, and an adjustment strategy of the update period is not limited to the situation determination of pilots, the change situation of signal change characteristic quantity slipping along with time, fault diagnosis results, the use situation of transmission bandwidth and the like;
the judgment criterion of the system starting in the S1 is as follows: starting a side cloud coordination mechanism when the vibration value of the low-pressure rotor is greater than 3 units for 60 seconds continuously or greater than 4 units for 10 seconds continuously;
and the cloud platform carries out fault diagnosis based on the reconstruction data, wherein a network framework of the fault diagnosis is a two-dimensional convolutional neural network adopting supervised learning, and a data set used for training the two-dimensional convolutional neural network is vibration data of the same model of aircraft engine during flight and fault diagnosis result data of flight maintenance.
2. The civil aircraft engine fault diagnosis method based on 5G Bian Yun synergy of claim 1, wherein the judgment criterion is based on vibration value and time.
3. The civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation according to claim 1, wherein the cloud platform performs signal reconstruction on the compression sampling vibration data according to the dictionary, and specifically comprises:
an orthogonal matching pursuit algorithm is adopted as a reconstruction algorithm, signal reconstruction refers to a process of reconstructing a sparse signal x with the length of N by a measurement vector y with the length of M, and a signal reconstruction result is as follows:
Figure 490399DEST_PATH_IMAGE001
wherein, the first and the second end of the pipe are connected with each other,
Figure 835929DEST_PATH_IMAGE002
for the determined sparse coefficient>
Figure 799337DEST_PATH_IMAGE003
Is an estimate of the original signal.
4. The civil aircraft engine fault diagnosis method based on 5G Bian Yun cooperation according to claim 1, wherein an input layer of the two-dimensional convolution neural network is a two-dimensional data matrix formed by drawing vibration time sequence waveform signals and expressed as a single-channel gray image, a network middle layer is formed by alternately stacking a convolution layer, a pooling layer and an activation function layer, and an output layer is formed by a full connection layer and a Soft max layer.
CN202210889066.5A 2022-07-27 2022-07-27 Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation Active CN114970645B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210889066.5A CN114970645B (en) 2022-07-27 2022-07-27 Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210889066.5A CN114970645B (en) 2022-07-27 2022-07-27 Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation

Publications (2)

Publication Number Publication Date
CN114970645A CN114970645A (en) 2022-08-30
CN114970645B true CN114970645B (en) 2023-04-07

Family

ID=82970181

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210889066.5A Active CN114970645B (en) 2022-07-27 2022-07-27 Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation

Country Status (1)

Country Link
CN (1) CN114970645B (en)

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170178311A1 (en) * 2015-12-20 2017-06-22 Prophecy Sensors, Llc Machine fault detection based on a combination of sound capture and on spot feedback
CN106289508A (en) * 2016-07-19 2017-01-04 西南交通大学 A kind of vibration signal reconstruction method towards mechanical fault diagnosis
CN115280123A (en) * 2020-04-27 2022-11-01 西门子股份公司 Fault diagnosis method and device
CN112287889A (en) * 2020-11-22 2021-01-29 西北工业大学 Bearing fault intelligent diagnosis method based on compressed sensing and related vector machine
CN113033309B (en) * 2021-02-25 2023-12-19 北京化工大学 Fault diagnosis method based on signal downsampling and one-dimensional convolutional neural network
CN112989712B (en) * 2021-04-27 2021-08-03 浙大城市学院 Aeroengine fault diagnosis method based on 5G edge calculation and deep learning

Also Published As

Publication number Publication date
CN114970645A (en) 2022-08-30

Similar Documents

Publication Publication Date Title
Gou et al. Aeroengine control system sensor fault diagnosis based on CWT and CNN
Pan et al. A novel deep learning network via multiscale inner product with locally connected feature extraction for intelligent fault detection
WO2023123593A1 (en) Variational mode decomposition and residual network-based aviation bearing fault diagnosis method
Joly et al. Gas-turbine diagnostics using artificial neural-networks for a high bypass ratio military turbofan engine
Gao et al. A fault diagnosis method of rolling bearing based on complex morlet CWT and CNN
CN113469060A (en) Multi-sensor fusion convolution neural network aeroengine bearing fault diagnosis method
CN106197999B (en) A kind of planetary gear method for diagnosing faults
CN110866448A (en) Flutter signal analysis method based on convolutional neural network and short-time Fourier transform
CN110987439B (en) Aeroengine fault prediction method based on Logitics regression and Xgboost model
CN111680875A (en) Unmanned aerial vehicle state risk fuzzy comprehensive evaluation method based on probability baseline model
CN113514109A (en) Unmanned aerial vehicle flight fault detection method and system
CN113168313A (en) Computing environment system for monitoring aircraft engines
Deng et al. Application of BP neural network and convolutional neural network (CNN) in bearing fault diagnosis
CN109655266A (en) A kind of Wind turbines Method for Bearing Fault Diagnosis based on AVMD and spectral coherence analysis
Li et al. Civil aircraft health management research based on big data and deep learning technologies
Li et al. Intelligent fault diagnosis of aeroengine sensors using improved pattern gradient spectrum entropy
CN114970645B (en) Civil aviation engine fault diagnosis system and method based on 5G Bian Yun cooperation
Bi et al. Fault diagnosis of valve clearance in diesel engine based on BP neural network and support vector machine
Zhong et al. Fault diagnosis of Marine diesel engine based on deep belief network
CN116624343B (en) Wind turbine generator tower abnormal vibration monitoring and health degree evaluation method and system
CN112069738A (en) Method and system for predicting residual life of electric steering engine based on DBN and multi-layer fuzzy LSTM
Zurita et al. A review of vibration machine diagnostics by using artificial intelligence methods
CN115903760A (en) Fault diagnosis and performance evaluation method for unmanned aerial vehicle system
CN111766067A (en) Aircraft outfield aircraft engine fault prediction method based on deep learning
Shang et al. Varying Speed Bearing Fault Diagnosis Based on Synchroextracting Transform and Deep Residual Network

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