CN113722833A - Turbofan engine residual service life prediction method based on dual-channel long-short time memory network - Google Patents

Turbofan engine residual service life prediction method based on dual-channel long-short time memory network Download PDF

Info

Publication number
CN113722833A
CN113722833A CN202111053571.8A CN202111053571A CN113722833A CN 113722833 A CN113722833 A CN 113722833A CN 202111053571 A CN202111053571 A CN 202111053571A CN 113722833 A CN113722833 A CN 113722833A
Authority
CN
China
Prior art keywords
service life
time
dual
memory network
residual service
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
CN202111053571.8A
Other languages
Chinese (zh)
Other versions
CN113722833B (en
Inventor
彭成
吴佳期
唐朝晖
陈青
张龙信
桂卫华
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.)
Hunan University of Technology
Original Assignee
Hunan University of Technology
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 Hunan University of Technology filed Critical Hunan University of Technology
Priority to CN202111053571.8A priority Critical patent/CN113722833B/en
Publication of CN113722833A publication Critical patent/CN113722833A/en
Application granted granted Critical
Publication of CN113722833B publication Critical patent/CN113722833B/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]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/04—Ageing analysis or optimisation against ageing
    • 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)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Operations Research (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Mathematical Physics (AREA)
  • Evolutionary Biology (AREA)
  • Computing Systems (AREA)
  • Development Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Biomedical Technology (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)

Abstract

The method for predicting the residual service life of the turbine engine based on the dual-channel long and short term memory neural network comprises the steps of firstly measuring the variability of a sensor data index for monitoring the state of the engine, obtaining the sensor data index with the variability, then processing a difference value between the data index and the data index by using the dual-channel long and short term memory neural network, designing a convolution neural network module to extract a local time characteristic of an output result of a long and short term memory neural network sequence, then using the local time characteristic for input of a two-layer fully-connected neural network to predict the residual service life of the engine, and finally using a predicted value at the previous moment as a buffer for a predicted result at the current moment to smoothly calibrate the current predicted value. The method effectively reduces the interference of fault noise, improves the capability of memorizing the neural network processing time sequence in long and short time, and finally improves the accuracy of fitting the residual service life.

Description

Turbofan engine residual service life prediction method based on dual-channel long-short time memory network
Technical Field
The invention belongs to the field of life prediction in equipment health management, and particularly relates to a turbine engine residual service life prediction method based on a dual-channel long-time and short-time memory network.
Background
The turbine engine is the heart of the aircraft and provides power for the flight of the aircraft. However, since the engine is often operated in a high temperature and high pressure environment, the problem of failure is difficult to avoid. If the engine fails, the airplane cannot take off, the passenger changes his label, the reputation of the airline company is damaged, and if the engine fails, the airplane is damaged, and personnel are injured. Therefore, it is very important to accurately predict the remaining service life of the engine before the fault occurs. At present, the traditional deep learning method is used to obtain good effect, but still faces the following problems:
(1) the problem to be solved is how to select useful time characteristics for the phenomenon that the influence of the time characteristics on the life prediction changes when the turbine engine runs under different environments, so as to avoid the occurrence of characteristic redundancy or invalid characteristics.
(2) For time characteristics, researchers often make a model pay attention to the size of time characteristics at a certain moment, and ignore the difference between the time characteristics at two different moments, for example, if a certain time characteristic value is continuously large, but the characteristic difference is small, how to use the two parameters to predict the service life reduces noise influence, and the model has robustness, which is also an urgent problem to be solved.
(3) Generally, the service life of a turbine engine is smooth and stable, that is, the length of the remaining service life of the engine in a certain period of time is not very different, the fluctuation times are relatively rare, but under the work of a severe environment, data transmitted back to a system by a sensor is often not clean, the traditional deep learning method learns and predicts according to the data, the remaining service life predicted by using a neural network fluctuates up and down, the curve of the remaining service life is jagged, and the deviation from the actual remaining service life is larger.
Disclosure of Invention
In order to solve the problems, the text provides a method for predicting the remaining service life of the turbofan engine through a dual-channel long-short time memory network. Firstly, selecting characteristics with predictability by using characteristic variability (prognosibility) by adopting a self-adaptive characteristic selection method aiming at different data sets, then, extracting output characteristics of each time after the processing of the long and short term memory network by using a dual-channel long and short term memory network, processing a time characteristic value and a characteristic difference value, then, extracting the output characteristics of each time after the processing of the long and short term memory network by using a convolutional neural network, then, predicting the residual service life by using a fully-connected neural network, and finally, providing a momentum smoothing method to process the actual residual service life curve aiming at the problem that the service life curve is jagged under the inspiration of gradient momentum.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the service life prediction method based on the dual-channel long-and-short time memory network comprises the following steps:
A. preprocessing data, selecting characteristics useful for model prediction, standardizing the data, accelerating the convergence of the model, and expanding the data in the following three aspects of characteristic selection, standardization processing and time window processing;
a. solving the probnosibility of the features, selecting the features which change greatly within a certain time range for training the model, wherein the probnosibility formula is as follows:
Figure BDA0003253810760000021
xja measurement vector representing a feature on the jth system, the variable M being the number of systems monitored, NjIs the number of measurements on the jth system, it can be observed that for some features, if prognosibility is equal to 0 or NaN, these features are removed, forming a new sample set;
b. data normalization was performed using z-score:
Figure BDA0003253810760000031
u represents the mean of all selected features, σ represents the standard deviation of all selected features, and x represents a value of a selected feature;
c. dividing a time window, the window width being denoted NtThe sliding step is denoted as s, and the first time window Input1 ═ x1,x2,...,xNf]，xiA certain time-characteristic vector is represented,
Figure BDA0003253810760000032
representing a temporal feature vector xiAt the first value of the time window, and so on;
Figure BDA0003253810760000033
second time window (time characteristic difference) Input2＝[d1,d2,...,dNf]，diA difference value representing a time feature vector;
Figure BDA0003253810760000034
B. dual-channel long-and-short-term memory network processing Input1And Input2Then, two Output outputs are obtained1＝[h1,h2,...,hhidden_size]And Output2＝[g1,g2,...,ghidden_size]，hiLength of the expressionShort-time memory network processing Input1Subsequent vector, giIndicating long-short time memory network processing Input2The subsequent vector;
Figure BDA0003253810760000041
Figure BDA0003253810760000042
output will1After (N)t-1) row vectors and Output2Adding directly to obtain Output which is o1,o2,...,ohidden_size]，oiFrom hiAnd giAdding to obtain;
Figure BDA0003253810760000043
C. dividing the life prediction into two parts, wherein one part is to extract the local time characteristics output by the dual-channel long-and-short memory network sequence by using a convolutional neural network, and the other part is to predict the residual service life by using a fully-connected neural network;
D. considering the relationship that the residual service life of the turbine engine in the early stage has cache to the residual service life at the current moment, and being inspired by momentum gradient reduction, the momentum smoothing residual service life method is adopted for a test set, and the formula is as follows:
predictt＝k×yt+(1-k)×predictt-1,0≤k≤1 (8)
ytthe method is to use a dual-channel long-time memory network to predict the residual service life at the time t, predcitt-1Is the prediction result, predict, after the last time smoothingtIs the remaining useful life after smoothing at the current time t, k denotes ytOn predicttThe ratio of (A) to (B).
The invention has the beneficial effects that:
the two-channel long-short term memory network model can learn a time value and a time difference value of a time characteristic, then is processed by a convolutional neural network to predict the residual service life, and finally, the predicted residual service life can be more fit with a real residual service life curve under the action of momentum smoothing.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a time window diagram for data preprocessing;
FIG. 3 is a diagram of a dual channel long and short term memory network;
FIG. 4 is a diagram of a remaining life prediction architecture;
FIG. 5 is a graph of the remaining useful life of a portion of the engine unit of the FD001 sub-data set;
fig. 6 is a momentum smoothness comparison graph.
Detailed Description
The present invention is described in further detail below with reference to the attached drawing figures.
Referring to fig. 1, the method for predicting the remaining service life based on the dual-channel long-and-short time memory network comprises the following steps:
A. preprocessing data, selecting characteristics useful for model prediction, standardizing the data, accelerating the convergence of the model, and expanding the data in the following three aspects of characteristic selection, standardization processing and time window processing;
a. table 1 lists the 21 sensor signatures monitored for each engine;
introduction to the Sensors of Table 1
Figure BDA0003253810760000051
Figure BDA0003253810760000061
Solving the probnosibility of the monitoring data, and selecting the characteristics with larger variation for training the model, wherein the prognosibility formula is as follows:
Figure BDA0003253810760000062
xj represents a measurement vector of a certain feature on the jth system, the variable M is the number of monitored systems, Nj is the number of measurements on the jth system, and for some features, if the prognosibility of the features is equal to 0 or NaN, the features are removed to form a new sample set;
b. data normalization was performed using z-score:
Figure BDA0003253810760000071
u represents the mean of all selected features, σ represents the standard deviation of all selected features, and x represents a value of a selected feature;
c. referring to FIG. 2, a time window is divided, with the window width denoted NtWith sliding step denoted s, first time window Input1＝[x1,x2,...,xNf]，xiA certain time-characteristic vector is represented,
Figure BDA0003253810760000072
representing a temporal feature vector xiAt the first value of the time window, and so on;
Figure BDA0003253810760000073
second time window (time characteristic difference) Input2＝[d1,d2,...,dNf]，diA difference value representing a time feature vector;
Figure BDA0003253810760000074
B. referring to FIG. 3, dual-channel long-short time memory network processing Input1And Input2Thereafter, two outputs Output can be obtained1＝[h1,h2,...,hhidden_size]And Output2＝[g1, g2,...,ghidden_size]，hiIndicating long-short time memory network processing Input1Subsequent vector, giIndicating long-short time memory network processing Input2The subsequent vector;
Figure BDA0003253810760000075
Figure BDA0003253810760000081
output will1After (N)t-1) row vectors and Output2Adding directly to obtain Output which is o1,o2,...,ohidden_size]，oiFrom hiAnd giAdding to obtain;
Figure BDA0003253810760000082
C. referring to fig. 4, the life prediction can be divided into two parts, one part is to extract the local time characteristics output by the dual-channel long-and-short memory network sequence by using the convolutional neural network, and the other part is to predict the remaining service life by using the fully-connected neural network;
D. considering the relationship that the residual service life of the turbine engine in the early stage has cache to the residual service life at the current moment, and being inspired by momentum gradient reduction, the method for smoothing the residual service life of momentum is adopted for a test set, and the formula is as follows:
predictt＝k×yt+(1-k)×predictt-1,0≤k≤1 (16)
ytthe method is to use a dual-channel long-time memory network to predict the residual service life at the time t, predcitt-1Is the prediction result, predict, after the last time smoothingtIs the remaining useful life after smoothing at the current time t, k denotes ytOn predicttThe ratio of the component (A) to (B);
E. to demonstrate the effectiveness of the method of the present invention, which is further described by comparing it to different methods, the commercial modular aviation propulsion system simulation (C-MAPSS) raw data set is detailed in table 2:
table 2 data set introduction
Figure BDA0003253810760000083
Figure BDA0003253810760000091
The C-MAPSS data set is composed of four different sub data sets FD001, FD002, FD003 and FD004, the number of engine units and the number of fault modes of each sub data set are different, each sub data set is divided into a training data set and a testing data set, the real remaining service life of the vortex engine from a healthy state to a degraded state at each moment is recorded, and aiming at the data sets, the method disclosed by the invention firstly preprocesses data, then memorizes network processing time characteristics and characteristic difference values by utilizing two channels in length, and finally smoothly predicts a remaining service life curve. The experimental results of the method of the invention and some advanced methods were analyzed by comparison and are detailed in table 3:
TABLE 3 comparison of the results of the methods
Figure BDA0003253810760000092
Figure BDA0003253810760000101
The score formula and the RMSE formula are as follows:
Figure BDA0003253810760000102
predictiindicates the predicted value, RULiRepresenting the true remaining useful life, N representing the amount of all sample data, the graph of the remaining useful life of a portion of the engine units is shown in fig. 5, table 4 depicts the effect of using momentum smoothing, and fig. 6 compares the remaining useful life curve after momentum smoothing.
TABLE 4 k-value influence Effect
Figure BDA0003253810760000103

Claims (3)

1. The method for predicting the remaining service life of the turbine engine based on the dual-channel long-short time memory network is characterized by comprising the following steps of:
A. preprocessing data, selecting characteristics useful for model prediction, standardizing the data, accelerating the convergence of the model, and expanding the data in the following three aspects of characteristic selection, standardization processing and time window processing;
a. solving the variability (prognosability) of the features, selecting the features which have larger variation within a certain time range for training the model, wherein the prognosability formula is as follows:
Figure FDA0003253810750000011
xja measurement vector representing a feature on the jth system, the variable M being the number of systems monitored, NjIs the number of measurements on the jth system, it can be observed that for certain features, if their prognosability is equal to 0 or NaN, these features are removed, resulting inA new sample set;
b. data normalization was performed using z-score:
Figure FDA0003253810750000012
u represents the mean of all selected features, σ represents the standard deviation of all selected features, and x represents a value of a selected feature;
c. dividing a time window, the window width being denoted NtThe sliding step is denoted as s, and the first time window Input1 ═ x1,x2,...,xNf]，xiA certain time-characteristic vector is represented,
Figure FDA0003253810750000013
representing a temporal feature vector xiAt the first value of the time window, and so on;
Figure FDA0003253810750000021
second time window (time characteristic difference) Input2＝[d1,d2,...,dNf]，diA difference value representing a time feature vector;
Figure FDA0003253810750000022
B. dual-channel long-and-short-term memory network processing Input1And Input2Then, two Output outputs are obtained1＝[h1,h2,...,hhidden_size]And Output2＝[g1,g2,...,ghidden_size]，hiIndicating long-short time memory network processing Input1Subsequent vector, giIndicating long-short time memory network processing Input2The subsequent vector;
Figure FDA0003253810750000023
Figure FDA0003253810750000024
output will1After (N)t-1) row vectors and Output2Adding directly to obtain Output which is o1,o2,...,ohidden_size]，oiFrom hiAnd giAdding to obtain;
Figure FDA0003253810750000025
C. dividing the life prediction into two parts, wherein one part is to extract the local time characteristics of the output result of the two-channel long-and-short time memory network sequence by using a convolutional neural network, and the other part is to predict the residual service life by using a fully-connected neural network;
D. considering the relationship that the residual service life of the turbine engine in the early stage has cache to the residual service life at the current moment, and being inspired by momentum gradient reduction, the momentum smoothing residual service life method is adopted for a test set, and the formula is as follows:
predictt＝k×yt+(1-k)×predictt-1,0≤k≤1 (8)
ytthe method is to use a dual-channel long-time memory network to predict the residual service life at the time t, predcitt-1Is the prediction result, predict, after the last time smoothingtIs the remaining useful life after smoothing at the current time t, k denotes ytOn predicttThe ratio of (A) to (B).
2. The method of claim 1 for turbine engine life prediction based on a dual channel long-short term memory network, wherein: the data set can be divided into four subdata sets of FD001, FD002, FD003 and FD004, each data set comprises a training set and a testing set, and the details are shown in Table 1:
table 1 data set introduction
Figure FDA0003253810750000031
Figure FDA0003253810750000041
3. The method for predicting the life of a turbine engine based on the dual-channel long-short time memory network as claimed in claim 1, wherein the detailed parameter settings are detailed in table 2:
table 2 parameter details
Figure FDA0003253810750000042
CN202111053571.8A 2021-09-09 2021-09-09 Turbofan engine residual service life prediction method based on double-channel long-short-term memory network Active CN113722833B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111053571.8A CN113722833B (en) 2021-09-09 2021-09-09 Turbofan engine residual service life prediction method based on double-channel long-short-term memory network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111053571.8A CN113722833B (en) 2021-09-09 2021-09-09 Turbofan engine residual service life prediction method based on double-channel long-short-term memory network

Publications (2)

Publication Number Publication Date
CN113722833A true CN113722833A (en) 2021-11-30
CN113722833B CN113722833B (en) 2023-06-06

Family

ID=78682835

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111053571.8A Active CN113722833B (en) 2021-09-09 2021-09-09 Turbofan engine residual service life prediction method based on double-channel long-short-term memory network

Country Status (1)

Country Link
CN (1) CN113722833B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114547803A (en) * 2022-02-28 2022-05-27 扬州中卓泵业有限公司 Ceramic pump turbine life detection system and method

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1878877A2 (en) * 2006-07-15 2008-01-16 Rolls-Royce plc Shape memory material actuator
CN102855349A (en) * 2012-08-06 2013-01-02 南京航空航天大学 Quick prototype design method and platform for gas path fault diagnosis for aeroengine
EP3051075A1 (en) * 2015-02-02 2016-08-03 General Electric Company Wash timing based on turbine operating parameters
CN106065830A (en) * 2016-06-01 2016-11-02 南京航空航天大学 A kind of pulse detonation combustor device based on rotary valve and pneumatic operated valve combination
CN106524223A (en) * 2016-12-15 2017-03-22 中国科学院工程热物理研究所 Combustion chamber with main nozzle module and micro nozzle module
CN110073301A (en) * 2017-08-02 2019-07-30 强力物联网投资组合2016有限公司 The detection method and system under data collection environment in industrial Internet of Things with large data sets
CN112295145A (en) * 2019-07-24 2021-02-02 上海长智系统集成有限公司 Artificial intelligent fire-fighting robot, disaster detection method, computer device and medium
CN112613226A (en) * 2020-12-10 2021-04-06 大连理工大学 Feature enhancement method for residual life prediction
CN112703457A (en) * 2018-05-07 2021-04-23 强力物联网投资组合2016有限公司 Method and system for data collection, learning and machine signal streaming for analysis and maintenance using industrial internet of things
CN112966355A (en) * 2021-03-30 2021-06-15 西安电子科技大学 Method for predicting residual service life of shield machine cutter based on deep learning
CN113094822A (en) * 2021-03-12 2021-07-09 华中科技大学 Method and system for predicting residual life of mechanical equipment
CN113158445A (en) * 2021-04-06 2021-07-23 中国人民解放军战略支援部队航天工程大学 Prediction algorithm for residual service life of aero-engine with convolution memory residual self-attention mechanism
CN113204921A (en) * 2021-05-13 2021-08-03 哈尔滨工业大学 Method and system for predicting remaining service life of airplane turbofan engine

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1878877A2 (en) * 2006-07-15 2008-01-16 Rolls-Royce plc Shape memory material actuator
CN102855349A (en) * 2012-08-06 2013-01-02 南京航空航天大学 Quick prototype design method and platform for gas path fault diagnosis for aeroengine
EP3051075A1 (en) * 2015-02-02 2016-08-03 General Electric Company Wash timing based on turbine operating parameters
CN106065830A (en) * 2016-06-01 2016-11-02 南京航空航天大学 A kind of pulse detonation combustor device based on rotary valve and pneumatic operated valve combination
CN106524223A (en) * 2016-12-15 2017-03-22 中国科学院工程热物理研究所 Combustion chamber with main nozzle module and micro nozzle module
CN110073301A (en) * 2017-08-02 2019-07-30 强力物联网投资组合2016有限公司 The detection method and system under data collection environment in industrial Internet of Things with large data sets
CN112703457A (en) * 2018-05-07 2021-04-23 强力物联网投资组合2016有限公司 Method and system for data collection, learning and machine signal streaming for analysis and maintenance using industrial internet of things
CN112295145A (en) * 2019-07-24 2021-02-02 上海长智系统集成有限公司 Artificial intelligent fire-fighting robot, disaster detection method, computer device and medium
CN112613226A (en) * 2020-12-10 2021-04-06 大连理工大学 Feature enhancement method for residual life prediction
CN113094822A (en) * 2021-03-12 2021-07-09 华中科技大学 Method and system for predicting residual life of mechanical equipment
CN112966355A (en) * 2021-03-30 2021-06-15 西安电子科技大学 Method for predicting residual service life of shield machine cutter based on deep learning
CN113158445A (en) * 2021-04-06 2021-07-23 中国人民解放军战略支援部队航天工程大学 Prediction algorithm for residual service life of aero-engine with convolution memory residual self-attention mechanism
CN113204921A (en) * 2021-05-13 2021-08-03 哈尔滨工业大学 Method and system for predicting remaining service life of airplane turbofan engine

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
鲍霁星: "一维CNN模型在车轴疲劳裂纹声发射信号中的应用研究" *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114547803A (en) * 2022-02-28 2022-05-27 扬州中卓泵业有限公司 Ceramic pump turbine life detection system and method

Also Published As

Publication number Publication date
CN113722833B (en) 2023-06-06

Similar Documents

Publication Publication Date Title
CN112149316B (en) Aero-engine residual life prediction method based on improved CNN model
CN109472110B (en) Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model
Li et al. A directed acyclic graph network combined with CNN and LSTM for remaining useful life prediction
Wu et al. Remaining useful life estimation of engineered systems using vanilla LSTM neural networks
CN110222371B (en) Bayes and neural network-based engine residual life online prediction method
CN110866314A (en) Method for predicting residual life of rotating machinery of multilayer bidirectional gate control circulation unit network
CN105606914A (en) IWO-ELM-based Aviation power converter fault diagnosis method
WO2024087128A1 (en) Multi-scale hybrid attention mechanism modeling method for predicting remaining useful life of aero engine
Wang et al. A remaining useful life prediction model based on hybrid long-short sequences for engines
Ruan et al. Remaining useful life prediction for aero-engine based on LSTM and CNN
CN114266201B (en) Self-attention elevator trapping prediction method based on deep learning
CN113722833A (en) Turbofan engine residual service life prediction method based on dual-channel long-short time memory network
CN116840764A (en) Method and system for evaluating comprehensive error state of capacitive voltage transformer
Li et al. A sequence-to-sequence remaining useful life prediction method combining unsupervised LSTM encoding-decoding and temporal convolutional network
CN115859777A (en) Method for predicting service life of product system in multiple fault modes
Yao et al. RUL prediction method for rolling bearing using convolutional denoising autoencoder and bidirectional LSTM
CN113673774A (en) Aero-engine remaining life prediction method based on self-encoder and time sequence convolution network
Morsalin et al. Induction motor inter-turn fault detection using heuristic noninvasive approach by artificial neural network with Levenberg Marquardt algorithm
CN116521406A (en) Method for detecting anomaly of non-overrun flight parameter data of aero-engine based on residual gate GRU-VAE model
Cao et al. Fuzzy Analytic Hierarchy Process Evaluation Method of Gas Turbine Based on Health Degree
Yan et al. Remaining Useful Life Interval Prediction for Complex System Based on BiGRU Optimized by Log-Norm
Cheng et al. A Spatio-Temporal Attention Mechanism Based Approach for Remaining Useful Life Prediction of Turbofan Engine
Peng et al. Research Article A Spatio-Temporal Attention Mechanism Based Approach for Remaining Useful Life Prediction of Turbofan Engine
Dang et al. Reliability prediction based on degradation measure distribution and wavelet neural network
Wang et al. Performance Optimization of Aero Turboshaft Engine Based on Bayesian 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