CN109472110A - A kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model - Google Patents
A kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model Download PDFInfo
- Publication number
- CN109472110A CN109472110A CN201811441584.0A CN201811441584A CN109472110A CN 109472110 A CN109472110 A CN 109472110A CN 201811441584 A CN201811441584 A CN 201811441584A CN 109472110 A CN109472110 A CN 109472110A
- Authority
- CN
- China
- Prior art keywords
- engine
- lstm
- data
- indicate
- aero
- 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
Links
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
- 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
- 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"
-
- 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)
- Geometry (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Development Economics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Hardware Design (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The aero-engine remaining life prediction technique based on LSTM network and ARIMA model that the invention discloses a kind of, this method comprises: establishing n engine health indicator (LSTM-HI) models based on LSTM deep neural network according to engine history degraded data;According to the passing sensing data of engine, training ARIMA model and backward multi-step prediction engine sensor parameter;By the sensor parameters predicted, whether failure is degenerated to according to LSTM-HI metrics evaluation engine, obtains engine residual service life and its probability distribution.The invention proposes a kind of novel aero-engine remaining life prediction technique, accuracy rate and feasibility with higher play the role of actively promoting for the management of aero-engine real time health, reduction maintenance cost.
Description
Technical field
The invention belongs to aero-engine remaining life technical fields, more particularly to one kind to be based on LSTM (Long
Short-Term Memory) network and ARIMA (Autoregressive Integrated Moving Average Model)
The aero-engine remaining life prediction technique of model.
Background technique
Aero-engine is worked throughout the year under high temperature, high pressure, the harsh conditions of high load, and its working condition often changes
Become, which results in engine failure to take place frequently." heart " of the aero-engine as aircraft, health status directly determines
Passenger personal safety and civil aviaton flight safety.To guarantee its safety, engine must repair in time.
However premature maintenance will certainly cause the increase of airline operation cost, therefore, it is necessary to maintenance time is grasped in due course, from
And the use value of engine is realized to greatest extent.Accurately prediction aero-engine remaining life (RUL) is to ensure
Flight safety and the important measures for reducing maintenance cost.
With the development of artificial intelligence technology, the method for diagnosing faults based on data is to solve being good for for complicated nonlinear system
Health forecasting problem provides a kind of new solution route.For intelligent engine of today control, the sensor of data-driven
Fault diagnosis and analytic redundancy technology also have attracted much attention.Deep learning network is obtained because of its powerful non-linear mapping capability
Extensive concern and research.Recurrent neural network (RNN) can model variation of the data in time series, certainly
The fields such as right Language Processing, speech recognition achieve success.Shot and long term memory network (LSTM) is sent out on the basis of RNN
A kind of deep learning network of the special long-rang dependence for solving the problems, such as data of Zhan Erlai.Its advantage is that not needing complexity
Ginseng is adjusted, default can remember the long-term information of data, to solve long-term Dependence Problem.
In addition, the rapid development of sensor technology and on-line monitoring technique, so as to engine in aircraft operational process
Performance and status data are integrated, are transmitted, being analyzed and draw becomes possibility, this is engine residual life forecast
Provide advantageous technical support and sufficient data preparation.
Summary of the invention
In view of the above technical problems, it is surplus to provide a kind of aero-engine based on LSTM network and ARIMA model by the present invention
Remaining life forecast method, this method can accuracy rate and feasibility with higher, for aero-engine real time health pipe
Reason, reduction maintenance cost play the role of actively promoting.
Technical solution: to achieve the above object, the technical solution adopted by the present invention are as follows:
A kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model, including it is following
Step:
It is strong to establish the n engines based on LSTM deep neural network according to engine history degraded data for step 1)
Kang Zhibiao (LSTM-HI) model, the health indicator model library of building reflection remaining life;
Step 2), according to the passing sensor data set of spreadsheet engine to be measured, training ARIMA model, design prediction window is long
Spend l, and the engine sensor parameter of prediction window length l backward;
Step 3) resolves LSTM-HI and refers to according to the sensor parameters of prediction and LSTM network health index model library
Whether mark, engine evaluated degenerate to failure, if being more than failure threshold, Fitted probability distribution characteristics parameter obtains engine
Remaining life;Otherwise, step 2) is gone to, prediction window length l is further added by, is successively calculated.
It is preferred: according to engine history degraded data in the step 1), to establish n based on LSTM depth nerve net
Specific step is as follows for engine health indicator (LSTM-HI) model of network:
Step 1.1) is degenerated to the historical data of failure according to aero-engine by health, and suitable sensor is selected to join
Number, and noise reduction, smoothing processing are carried out, form n training dataset;
Step 1.2) is referred to based on deep learning network LSTM and training dataset building aero-engine health status evaluation
LSTM-HI is marked, expression is as follows:
Wherein,Indicate that the sensor parameters of selection, k indicate the quantity of the sensor parameters of selection, t indicates the moment;
Step 1.3), training deep learning network LSTM, obtains n LSTM-HI model, establishes aero-engine health
State evaluation index model library.
Preferred: training deep learning network LSTM in the step 1.3) obtains n LSTM-HI model, establishes boat
Specific step is as follows for empty engine health status evaluation index model library:
Step 1.3.1), data are observed for one group of engine life-cycleWherein,
xtIndicate the sensor parameters of selection,Indicate cycle to failure i.e. its service life of engine, HItIt is the strong of arteface
Kang ZhibiaoAssuming that the input data at current time is xt, the output h of last moment LSTM networkt-1, then when current
The output calculating for carving LSTM network is as follows:
ft=σ (wf·[ht-1,xt]+bf)
it=σ (wi·[ht-1,xt]+bi)
ot=σ (wo·[ht-1,xt]+bo)
ht=ot·Φ(St)
Wherein, σ (), g () indicate that activation primitive, w indicate weight, and b indicates biasing, StIndicate the memory of LSTM network
Quantity of state, itIndicate the output of input gate, ftIndicate the output of forgetting door, otIndicate the output of out gate;
Step 1.3.2), by the output of the corresponding LSTM network of all input samples of step 1.3.1 forward calculation, and remember
Record the error of each training sample;
Step 1.3.3), the error obtained according to forward calculation, by error gradient descent method training LSTM network until
Precision is met the requirements;
Step 1.3.4), the n group engine life-cycle is observed into data, n LSTM network is respectively trained, obtains n LSTM-
HI model, establishes model library.
It is preferred: according to the passing sensing data of engine, training ARIMA model and backward multistep described in step 2)
Predict that specific step is as follows for engine sensor parameter:
Step 2.1), if the sensing data of the passing degeneration of known engine isWherein,
xtIndicate that the sensor measures parameters of selection, r indicate engine previous cycle number;For the sensor parameters of every one-dimensional selection,
Its ARIMA prediction model is established, expression formula is as follows:
wt=φ1·wt-1+φ2·wt-2+…+φp·wp-2
+δ+μt+θ1·μt-1+θ2·μt-2+…+θq·μq-2
Wherein, μtIt is white noise sequence, δ is to indicate sequence data without the constant of 0 equalization, and φ and respectively θ indicate poor
The coefficient of subitem and rolling average item, wtIndicate that data in the Difference Terms of t moment, indicate are as follows:
Wherein, d indicates to carry out the order of difference, and L is delay factor;
Step 2.2), after the order for determining ARIMA model, according to passing data, by Least Square Method
The parameter of ARIMA model;
Step 2.3) carries out multi-step prediction to each sensor parameters respectively according to obtained ARIMA model, obtains each
The prediction data of a sensor.
It is preferred: by the sensor parameters predicted described in step 3), according to LSTM-HI metrics evaluation engine whether
Failure is degenerated to, obtaining engine residual service life and its probability distribution, specific step is as follows:
Step 3.1), calculates the corresponding health indicator HI sequence of sensor prediction data, and definition failure threshold is θf, when
HI judges that engine degenerates to failure when being more than threshold value, and it is engine residual that corresponding recurring number, which subtracts previous cycle number,
Service life RUL;
Step 3.2), the sensor parameters prediction data obtained by LSTM-HI model library and step 2), calculates separately out n
A HI sequence obtains n engine RUL predicted value according to step 3.1);
Step 3.3) is fitted n engine RUL predicted value using gauss of distribution function, obtains engine RUL
Probability distribution.
The utility model has the advantages that the invention proposes a kind of novel aero-engine remaining life prediction technique, have compared with
High accuracy rate and feasibility has the work actively promoted for the management of aero-engine real time health, reduction maintenance cost
With.
Detailed description of the invention
Fig. 1 is aero-engine remaining life prediction technique schematic diagram.
Fig. 2 is C-MAPSS system structure diagram.
Fig. 3 is the curve after characteristic parameter real change curve and smoothing processing.
Fig. 4 is trend chart of the characteristic parameter after a window is predicted.
Fig. 5 is #1 test engine remaining life prediction result.
Fig. 6 is #2 test engine remaining life prediction result.
Fig. 7 is #3 test engine remaining life prediction result.
Fig. 8 is 20 engine residual life prediction results of test set.
Specific embodiment
A specific embodiment of the invention is further described with reference to the accompanying drawing.
The present invention illustrates a kind of based on the prediction of the aero-engine remaining life of LSTM network and ARIMA model
Method, specifically includes the following steps:
It is strong to establish the n engines based on LSTM deep neural network according to engine history degraded data for step 1)
Kang Zhibiao (LSTM-HI) model, the health indicator model library of building reflection remaining life;
Step 1.1) is degenerated to the historical data of failure according to aero-engine by health, and suitable sensor is selected to join
Number, and noise reduction, smoothing processing are carried out, form n training dataset;
Step 1.2), needle are based on deep learning network LSTM and training dataset building aero-engine health status evaluation
Index LSTM-HI, expression are as follows:
Wherein,Indicate that the sensor parameters of selection, k indicate the quantity of the sensor parameters of selection, t indicates the moment.
Step 1.3), training deep learning network LSTM, obtains n LSTM-HI model, establishes aero-engine health
State evaluation index model library.
Step 1.3.1), data are observed for one group of engine life-cycleWherein,
xtIndicate the sensor parameters of selection,Indicate cycle to failure i.e. its service life of engine, HItIt is the strong of arteface
Kang ZhibiaoAssuming that the input data at current time is xt, the output h of last moment LSTM networkt-1, then currently
The output of moment LSTM network calculates as follows:
ft=σ (wf·[ht-1,xt]+bf)
it=σ (wi·[ht-1,xt]+bi)
ot=σ (wo·[ht-1,xt]+bo)
ht=ot·Φ(St)
Wherein, σ (), g () indicate that activation primitive, w indicate weight, and b indicates biasing, StIndicate the memory of LSTM network
Quantity of state, itIndicate the output of input gate, ftIndicate the output of forgetting door, otIndicate the output of out gate.
Step 1.3.2), by the output of the corresponding LSTM network of all input samples of step 1.3.1 forward calculation, and remember
Record the error of each training sample;
Step 1.3.3), the error obtained according to forward calculation, by error gradient descent method training LSTM network until
Precision is met the requirements;
Step 1.3.4), the n group engine life-cycle is observed into data, n LSTM network is respectively trained, obtains n LSTM-
HI model, establishes model library.
Step 2), according to the passing sensor data set of spreadsheet engine to be measured, training ARIMA model, design prediction window is long
Spend l, and the engine sensor parameter of prediction window length l backward;
Step 2.1), if the sensing data of the passing degeneration of known engine isWherein,
xtIndicate that the sensor measures parameters of selection, r indicate engine previous cycle number.For the sensor parameters of every one-dimensional selection,
Its ARIMA prediction model is established, expression formula is as follows:
wt=φ1·wt-1+φ2·wt-2+…+φp·wp-2
+δ+μt+θ1·μt-1+θ2·μt-2+…+θq·μq-2
Wherein, μtIt is white noise sequence, δ is to indicate sequence data without the constant of 0 equalization, and φ and respectively θ indicate poor
The coefficient of subitem and rolling average item, wtIndicate that data in the Difference Terms of t moment, indicate are as follows:
Wherein, d indicates to carry out the order of difference, and L is delay factor.
Step 2.2), after the order for determining ARIMA model, according to passing data, by Least Square Method
The parameter of ARIMA model;
Step 2.3) carries out multi-step prediction to each sensor parameters respectively according to obtained ARIMA model, obtains each
The prediction data of a sensor.
Step 3) resolves LSTM-HI and refers to according to the sensor parameters of prediction and LSTM network health index model library
Whether mark, engine evaluated degenerate to failure, if being more than failure threshold, Fitted probability distribution characteristics parameter obtains engine
Remaining life;Otherwise, step 2) is gone to, prediction window length l is further added by, is successively calculated;
Step 3.1) calculates the corresponding health indicator HI sequence of sensor prediction data.Definition failure threshold is θf, work as HI
Judge that engine degenerates to failure when more than threshold value, it is that engine residual makes that corresponding recurring number, which subtracts previous cycle number,
With service life RUL;
Step 3.2), the sensor parameters prediction data obtained by LSTM-HI model library and step 2), calculates separately out n
A HI sequence obtains n engine RUL predicted value according to step 3.1);
Step 3.3) is fitted n engine RUL predicted value using gauss of distribution function, obtains engine RUL
Probability distribution.
In order to verify the aero-engine sensor intelligent analytic redundancy designed by the present invention based on KEOS-ELM algorithm
The validity of design method has carried out relevant Digital Simulation under MATLAB environment.In this emulation, C-MAPSS is used
Data in data set in " train_FD001.txt " and " test_FD001.txt " both of these documents as training data and
Test data, the link of the modules of C-MAPSS and layout are as shown in Figure 2.Wherein, it includes 100 groups that training data is concentrated altogether
Different engines observe data from the life-cycle that health status degenerates to failure, and therefore, the present invention trains 100 LSTM-HI altogether
Model, and establish LSTM-HI model library.Different degrees of observation number is degenerated to comprising 100 groups of difference engines in test set
According to taking the prediction of 20 groups of carry out remaining lifes at random therefrom in order to verify the performance of the proposed method of the present invention.By right
The analysis of 21 sensing datas, choosing wherein effective 5 parameters includes that T24, Nf, Ps30, NRf and BPR carry out engine
Performance parameter prediction, the explanation of each parameter are as shown in table 1.
1 engine sensor measurement parameter of table
Since there are biggish noises for original data, standardization first is carried out to it and smoothing processing obtains more
True engine data.Fig. 3 is that each sensor parameters of a certain group of training data after noise reduction and smoothing processing are bent
Line.Fig. 4 is the result for carrying out parameter multi-step prediction by ARIMA model with first group of Engine test data (#1).Fig. 5,
Fig. 6 and Fig. 7 is to choose 1-3 group Engine test data (#1, #2 and #3), and HI change curve and obtained residue use the longevity
Order prediction result and its probability distribution.Table 2 is whole 20 groups of obtained prediction results of Engine test data.
The whole 20 groups of Engine test data remaining life prediction results of table 2
In conjunction with Fig. 5, Fig. 6 and Fig. 7 and table 2 as can be seen that engine residual life forecast side proposed by the present invention
Method has higher precision of prediction, and Relative Error is 10% or so, and absolute error is largely all located at 10 circulations
Within.Fig. 8 is the result curve of all 20 groups of Engine test datas predictions.But for individual engine such as #4, #10
It is larger with the prediction error of #16 engine, this may be since its current degradation recurring number is smaller, passing sensing data compared with
Few, ARIMA model cannot effectively model its degenerative process, therefore cause to predict that error is larger.But generally speaking,
A kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model proposed by the present invention can be with
Effectively predict the remaining life of aero-engine, accuracy rate and feasibility with higher.
It should be pointed out that the above description is merely a specific embodiment, but protection scope of the present invention is not
Be confined to this, anyone skilled in the art in the technical scope disclosed by the present invention, the change that can be readily occurred in
Change and replace, should be covered by the scope of protection of the present invention.Therefore, protection scope of the present invention should be with the claim
Protection scope subject to.
Claims (5)
1. a kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model, it is characterised in that:
The following steps are included:
Step 1) establishes the n engine health indicators based on LSTM deep neural network according to engine history degraded data
(LSTM-HI) model, the health indicator model library of building reflection remaining life;
Step 2), according to the passing sensor data set of spreadsheet engine to be measured, training ARIMA model designs prediction window length l,
And the engine sensor parameter of prediction window length l backward;
Step 3) resolves LSTM-HI index, comments according to the sensor parameters of prediction and LSTM network health index model library
Whether valence engine degenerates to failure, if being more than failure threshold, Fitted probability distribution characteristics parameter, obtaining engine residual makes
Use the service life;Otherwise, step 2) is gone to, prediction window length l is further added by, is successively calculated.
2. a kind of aero-engine remaining life based on LSTM network and ARIMA model according to claim 1
Prediction technique, it is characterised in that: according to engine history degraded data in the step 1), establish n based on LSTM depth mind
Specific step is as follows for engine health indicator (LSTM-HI) model through network:
Step 1.1) is degenerated to the historical data of failure by health according to aero-engine, selects suitable sensor parameters, and
Noise reduction, smoothing processing are carried out, n training dataset is formed;
Step 1.2) constructs aero-engine health status evaluation index based on deep learning network LSTM and training dataset
LSTM-HI, expression are as follows:
Wherein,Indicate that the sensor parameters of selection, k indicate the quantity of the sensor parameters of selection, t indicates the moment;
Step 1.3), training deep learning network LSTM, obtains n LSTM-HI model, establishes aero-engine health status and comment
Valence index model library.
3. a kind of aero-engine remaining life based on LSTM network and ARIMA model according to claim 2
Prediction technique, it is characterised in that: training deep learning network LSTM in the step 1.3) obtains n LSTM-HI model, builds
Specific step is as follows for vertical aero-engine health status evaluation index model library:
Step 1.3.1), data are observed for one group of engine life-cycleWherein, xtTable
Show the sensor parameters of selection,Indicate cycle to failure i.e. its service life of engine, HItIt is that the health of arteface refers to
MarkAssuming that the input data at current time is xt, the output h of last moment LSTM networkt-1, then current time
The output of LSTM network calculates as follows:
ft=σ (wf·[ht-1,xt]+bf)
it=σ (wi·[ht-1,xt]+bi)
ot=σ (wo·[ht-1,xt]+bo)
ht=ot·Φ(St)
Wherein, σ (), g () indicate that activation primitive, w indicate weight, and b indicates biasing, StIndicate the memory state of LSTM network
Amount, itIndicate the output of input gate, ftIndicate the output of forgetting door, otIndicate the output of out gate;
Step 1.3.2), by the output of the corresponding LSTM network of all input samples of step 1.3.1 forward calculation, and record each
The error of training sample;
Step 1.3.3), the error obtained according to forward calculation, by error gradient descent method training LSTM network until precision
It meets the requirements;
Step 1.3.4), the n group engine life-cycle is observed into data, n LSTM network is respectively trained, obtains n LSTM-HI mould
Type establishes model library.
4. a kind of aero-engine remaining life based on LSTM network and ARIMA model according to claim 1
Prediction technique, it is characterised in that: according to the passing sensing data of engine described in step 2), training ARIMA model and backward
Specific step is as follows for multi-step prediction engine sensor parameter:
Step 2.1), if the sensing data of the passing degeneration of known engine isWherein, xtIt indicates
The sensor measures parameters of selection, r indicate engine previous cycle number;For the sensor parameters of every one-dimensional selection, it is established
ARIMA prediction model, expression formula are as follows:
wt=φ1·wt-1+φ2·wt-2+…+φp·wp-2+δ+μt+θ1·μt-1+θ2·μt-2+…+θq·μq-2
Wherein, μtWhite noise sequence, δ be indicate sequence data without the constant of 0 equalization, φ and respectively θ indicate Difference Terms and
The coefficient of rolling average item, wtIndicate that data in the Difference Terms of t moment, indicate are as follows:
Wherein, d indicates to carry out the order of difference, and L is delay factor;
Step 2.2), after the order for determining ARIMA model, according to passing data, by Least Square Method ARIMA model
Parameter;
Step 2.3) carries out multi-step prediction to each sensor parameters respectively, obtains each biography according to obtained ARIMA model
The prediction data of sensor.
5. a kind of aero-engine remaining life based on LSTM network and ARIMA model according to claim 1
Prediction technique, it is characterised in that: by the sensor parameters predicted described in step 3), according to LSTM-HI metrics evaluation engine
Whether failure is degenerated to, and obtaining engine residual service life and its probability distribution, specific step is as follows:
Step 3.1), calculates the corresponding health indicator HI sequence of sensor prediction data, and definition failure threshold is θf, when HI is more than
Judge that engine degenerates to failure when threshold value, it is engine residual service life that corresponding recurring number, which subtracts previous cycle number,
RUL;
Step 3.2), the sensor parameters prediction data obtained by LSTM-HI model library and step 2), calculates separately out n HI
Sequence obtains n engine RUL predicted value according to step 3.1);
Step 3.3) is fitted n engine RUL predicted value using gauss of distribution function, obtains the probability of engine RUL
Distribution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811441584.0A CN109472110B (en) | 2018-11-29 | 2018-11-29 | Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811441584.0A CN109472110B (en) | 2018-11-29 | 2018-11-29 | Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109472110A true CN109472110A (en) | 2019-03-15 |
CN109472110B CN109472110B (en) | 2023-06-27 |
Family
ID=65673329
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811441584.0A Active CN109472110B (en) | 2018-11-29 | 2018-11-29 | Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109472110B (en) |
Cited By (30)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109991542A (en) * | 2019-03-27 | 2019-07-09 | 东北大学 | Lithium ion battery residual life prediction technique based on WDE optimization LSTM network |
CN110108957A (en) * | 2019-05-07 | 2019-08-09 | 武汉理工大学 | A kind of tractor electric fault diagnosis method based on structured analysis method |
CN110110804A (en) * | 2019-05-15 | 2019-08-09 | 上海飞机客户服务有限公司 | Flight control system method for predicting residual useful life based on CNN and LSTM |
CN110414033A (en) * | 2019-05-31 | 2019-11-05 | 太原理工大学 | The mechanical equipment approaches of predictive maintenance that jointing edge calculates and number is twin |
CN110737948A (en) * | 2019-10-15 | 2020-01-31 | 南京航空航天大学 | method for predicting residual life of aero-engine based on deep FNN-LSTM hybrid network |
CN110737952A (en) * | 2019-09-17 | 2020-01-31 | 太原理工大学 | prediction method for residual life of key parts of mechanical equipment by combining AE and bi-LSTM |
CN111008728A (en) * | 2019-11-01 | 2020-04-14 | 深圳供电局有限公司 | Method for predicting short-term output of distributed photovoltaic power generation system |
CN111178553A (en) * | 2019-12-16 | 2020-05-19 | 北京航天智造科技发展有限公司 | Industrial equipment health trend analysis method and system based on ARIMA and LSTM algorithms |
CN111709577A (en) * | 2020-06-17 | 2020-09-25 | 上海海事大学 | RUL prediction method based on long-range correlation GAN-LSTM |
CN112100865A (en) * | 2019-11-04 | 2020-12-18 | 中国人民解放军国防科技大学 | Method for predicting remaining life of aircraft engine based on parallel CNN model |
CN112232404A (en) * | 2020-10-13 | 2021-01-15 | 中国铁路上海局集团有限公司南京供电段 | Reliability calculation method and system based on historical abnormity and operation and maintenance information of railway power supply equipment |
CN112231980A (en) * | 2020-10-19 | 2021-01-15 | 西安建筑科技大学 | Engine life prediction method, storage medium and computing device |
CN112257935A (en) * | 2020-10-26 | 2021-01-22 | 中国人民解放军空军工程大学 | Aviation safety prediction method based on LSTM-RBF neural network model |
CN112312411A (en) * | 2019-07-29 | 2021-02-02 | 中国移动通信集团广东有限公司 | Traffic prediction method of VoLTE service and terminal equipment |
CN112560347A (en) * | 2020-12-17 | 2021-03-26 | 中国科学院空间应用工程与技术中心 | Method and system for predicting remaining service life of aero-engine and electronic equipment |
CN112668249A (en) * | 2020-07-17 | 2021-04-16 | 国网山东省电力公司电力科学研究院 | Online construction method and system for major repair technical modification scheme of primary equipment of power grid |
CN112685961A (en) * | 2021-01-12 | 2021-04-20 | 武汉大学 | Method and system for predicting remaining service life of analog circuit |
CN112926273A (en) * | 2021-04-13 | 2021-06-08 | 中国人民解放军火箭军工程大学 | Method for predicting residual life of multivariate degradation equipment |
CN112966441A (en) * | 2021-03-08 | 2021-06-15 | 中国人民解放军海军航空大学 | Equipment residual life evaluation method based on continuous Weiner process damage |
CN113076913A (en) * | 2021-04-16 | 2021-07-06 | 嘉兴毕格智能科技有限公司 | Aircraft engine fault prediction method |
CN113158348A (en) * | 2021-05-21 | 2021-07-23 | 上海交通大学 | Aircraft engine residual life prediction method based on deep learning coupling modeling |
CN113393034A (en) * | 2021-06-16 | 2021-09-14 | 国网山东省电力公司泰安供电公司 | Electric quantity prediction method of online self-adaptive OSELM-GARCH model |
CN113836822A (en) * | 2021-10-28 | 2021-12-24 | 重庆大学 | Aero-engine service life prediction method based on MCLSTM model |
CN114282443A (en) * | 2021-12-28 | 2022-04-05 | 浙江大学 | Residual service life prediction method based on MLP-LSTM supervised joint model |
WO2022112895A1 (en) * | 2020-11-30 | 2022-06-02 | International Business Machines Corporation | Automated deep learning architecture selection for time series prediction with user interaction |
CN114580705A (en) * | 2022-01-12 | 2022-06-03 | 中国电子科技集团公司第十研究所 | Method for predicting residual life of avionic product |
CN116147928A (en) * | 2023-04-20 | 2023-05-23 | 清华大学 | Method, device and equipment for determining health state of aeroengine thermal jet device |
WO2024020960A1 (en) * | 2022-07-28 | 2024-02-01 | 西门子股份公司 | Method for predicting remaining useful life of circuit breaker, and electronic device and storage medium |
CN117807718A (en) * | 2024-03-01 | 2024-04-02 | 南京工业大学 | Aero-engine maintenance decision design method based on failure probability prediction |
US11952149B2 (en) | 2020-09-17 | 2024-04-09 | General Electric Company | System and method of forecasting aircraft engine operational data for predictive analytics |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104166787A (en) * | 2014-07-17 | 2014-11-26 | 南京航空航天大学 | Aero-engine remaining life prediction method based on multi-stage information fusion |
CN105808957A (en) * | 2016-03-15 | 2016-07-27 | 清华大学 | Method for predicting residual life of aero-engine |
US20180165592A1 (en) * | 2016-12-13 | 2018-06-14 | Industrial Technology Research Institute | System and method for predicting remaining lifetime of a component of equipment |
-
2018
- 2018-11-29 CN CN201811441584.0A patent/CN109472110B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104166787A (en) * | 2014-07-17 | 2014-11-26 | 南京航空航天大学 | Aero-engine remaining life prediction method based on multi-stage information fusion |
CN105808957A (en) * | 2016-03-15 | 2016-07-27 | 清华大学 | Method for predicting residual life of aero-engine |
US20180165592A1 (en) * | 2016-12-13 | 2018-06-14 | Industrial Technology Research Institute | System and method for predicting remaining lifetime of a component of equipment |
Non-Patent Citations (1)
Title |
---|
王文庆等: "基于CNN与GRU的航空发动机剩余寿命预测", 《第37届中国控制会议论文集(F)》 * |
Cited By (45)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109991542A (en) * | 2019-03-27 | 2019-07-09 | 东北大学 | Lithium ion battery residual life prediction technique based on WDE optimization LSTM network |
WO2020191800A1 (en) * | 2019-03-27 | 2020-10-01 | 东北大学 | Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network |
CN110108957A (en) * | 2019-05-07 | 2019-08-09 | 武汉理工大学 | A kind of tractor electric fault diagnosis method based on structured analysis method |
CN110110804A (en) * | 2019-05-15 | 2019-08-09 | 上海飞机客户服务有限公司 | Flight control system method for predicting residual useful life based on CNN and LSTM |
CN110110804B (en) * | 2019-05-15 | 2021-06-25 | 上海飞机客户服务有限公司 | Flight control system residual life prediction method based on CNN and LSTM |
CN110414033A (en) * | 2019-05-31 | 2019-11-05 | 太原理工大学 | The mechanical equipment approaches of predictive maintenance that jointing edge calculates and number is twin |
CN112312411A (en) * | 2019-07-29 | 2021-02-02 | 中国移动通信集团广东有限公司 | Traffic prediction method of VoLTE service and terminal equipment |
CN110737952A (en) * | 2019-09-17 | 2020-01-31 | 太原理工大学 | prediction method for residual life of key parts of mechanical equipment by combining AE and bi-LSTM |
CN110737948A (en) * | 2019-10-15 | 2020-01-31 | 南京航空航天大学 | method for predicting residual life of aero-engine based on deep FNN-LSTM hybrid network |
CN111008728A (en) * | 2019-11-01 | 2020-04-14 | 深圳供电局有限公司 | Method for predicting short-term output of distributed photovoltaic power generation system |
CN111008728B (en) * | 2019-11-01 | 2023-06-09 | 深圳供电局有限公司 | Prediction method for short-term output of distributed photovoltaic power generation system |
CN112131760A (en) * | 2019-11-04 | 2020-12-25 | 中国人民解放军国防科技大学 | CBAM model-based prediction method for residual life of aircraft engine |
CN112100865A (en) * | 2019-11-04 | 2020-12-18 | 中国人民解放军国防科技大学 | Method for predicting remaining life of aircraft engine based on parallel CNN model |
CN112149316B (en) * | 2019-11-04 | 2022-04-19 | 中国人民解放军国防科技大学 | Aero-engine residual life prediction method based on improved CNN model |
CN112149316A (en) * | 2019-11-04 | 2020-12-29 | 中国人民解放军国防科技大学 | Aero-engine residual life prediction method based on improved CNN model |
CN111178553A (en) * | 2019-12-16 | 2020-05-19 | 北京航天智造科技发展有限公司 | Industrial equipment health trend analysis method and system based on ARIMA and LSTM algorithms |
CN111709577A (en) * | 2020-06-17 | 2020-09-25 | 上海海事大学 | RUL prediction method based on long-range correlation GAN-LSTM |
CN111709577B (en) * | 2020-06-17 | 2023-08-15 | 上海海事大学 | RUL prediction method based on long-range correlation GAN-LSTM |
CN112668249A (en) * | 2020-07-17 | 2021-04-16 | 国网山东省电力公司电力科学研究院 | Online construction method and system for major repair technical modification scheme of primary equipment of power grid |
US11952149B2 (en) | 2020-09-17 | 2024-04-09 | General Electric Company | System and method of forecasting aircraft engine operational data for predictive analytics |
CN112232404A (en) * | 2020-10-13 | 2021-01-15 | 中国铁路上海局集团有限公司南京供电段 | Reliability calculation method and system based on historical abnormity and operation and maintenance information of railway power supply equipment |
CN112231980B (en) * | 2020-10-19 | 2024-06-07 | 西安建筑科技大学 | Engine life prediction method, storage medium and computing device |
CN112231980A (en) * | 2020-10-19 | 2021-01-15 | 西安建筑科技大学 | Engine life prediction method, storage medium and computing device |
CN112257935A (en) * | 2020-10-26 | 2021-01-22 | 中国人民解放军空军工程大学 | Aviation safety prediction method based on LSTM-RBF neural network model |
CN112257935B (en) * | 2020-10-26 | 2023-05-09 | 中国人民解放军空军工程大学 | Aviation safety prediction method based on LSTM-RBF neural network model |
GB2616770A (en) * | 2020-11-30 | 2023-09-20 | Ibm | Automated deep learning architecture selection for time series prediction with user interaction |
WO2022112895A1 (en) * | 2020-11-30 | 2022-06-02 | International Business Machines Corporation | Automated deep learning architecture selection for time series prediction with user interaction |
CN112560347A (en) * | 2020-12-17 | 2021-03-26 | 中国科学院空间应用工程与技术中心 | Method and system for predicting remaining service life of aero-engine and electronic equipment |
CN112685961B (en) * | 2021-01-12 | 2022-06-21 | 武汉大学 | Method and system for predicting remaining service life of analog circuit |
CN112685961A (en) * | 2021-01-12 | 2021-04-20 | 武汉大学 | Method and system for predicting remaining service life of analog circuit |
CN112966441A (en) * | 2021-03-08 | 2021-06-15 | 中国人民解放军海军航空大学 | Equipment residual life evaluation method based on continuous Weiner process damage |
CN112966441B (en) * | 2021-03-08 | 2022-04-29 | 中国人民解放军海军航空大学 | Equipment residual life evaluation method based on continuous Weiner process damage |
CN112926273A (en) * | 2021-04-13 | 2021-06-08 | 中国人民解放军火箭军工程大学 | Method for predicting residual life of multivariate degradation equipment |
CN113076913A (en) * | 2021-04-16 | 2021-07-06 | 嘉兴毕格智能科技有限公司 | Aircraft engine fault prediction method |
CN113158348B (en) * | 2021-05-21 | 2023-10-03 | 上海交通大学 | Aircraft engine residual life prediction method based on deep learning coupling modeling |
CN113158348A (en) * | 2021-05-21 | 2021-07-23 | 上海交通大学 | Aircraft engine residual life prediction method based on deep learning coupling modeling |
CN113393034A (en) * | 2021-06-16 | 2021-09-14 | 国网山东省电力公司泰安供电公司 | Electric quantity prediction method of online self-adaptive OSELM-GARCH model |
CN113836822A (en) * | 2021-10-28 | 2021-12-24 | 重庆大学 | Aero-engine service life prediction method based on MCLSTM model |
CN114282443A (en) * | 2021-12-28 | 2022-04-05 | 浙江大学 | Residual service life prediction method based on MLP-LSTM supervised joint model |
CN114580705A (en) * | 2022-01-12 | 2022-06-03 | 中国电子科技集团公司第十研究所 | Method for predicting residual life of avionic product |
CN114580705B (en) * | 2022-01-12 | 2023-07-25 | 中国电子科技集团公司第十研究所 | Method for predicting residual life of avionics product |
WO2024020960A1 (en) * | 2022-07-28 | 2024-02-01 | 西门子股份公司 | Method for predicting remaining useful life of circuit breaker, and electronic device and storage medium |
CN116147928A (en) * | 2023-04-20 | 2023-05-23 | 清华大学 | Method, device and equipment for determining health state of aeroengine thermal jet device |
CN117807718A (en) * | 2024-03-01 | 2024-04-02 | 南京工业大学 | Aero-engine maintenance decision design method based on failure probability prediction |
CN117807718B (en) * | 2024-03-01 | 2024-05-10 | 南京工业大学 | Aero-engine maintenance decision design method based on failure probability prediction |
Also Published As
Publication number | Publication date |
---|---|
CN109472110B (en) | 2023-06-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109472110A (en) | A kind of aero-engine remaining life prediction technique based on LSTM network and ARIMA model | |
CN107941537B (en) | A kind of mechanical equipment health state evaluation method | |
CN107153874B (en) | Water quality prediction method and system | |
CN108960303B (en) | Unmanned aerial vehicle flight data anomaly detection method based on LSTM | |
Lindemann et al. | Anomaly detection and prediction in discrete manufacturing based on cooperative LSTM networks | |
CN110376003B (en) | Intelligent train service life prediction method and system based on BIM | |
CN111274737A (en) | Method and system for predicting remaining service life of mechanical equipment | |
CN110361180B (en) | Intelligent train pantograph service performance dynamic monitoring and evaluating method and system | |
CN110807257A (en) | Method for predicting residual life of aircraft engine | |
CN106529145A (en) | Bridge monitoring data prediction method based on ARIMA-BP neural network | |
CN108873859B (en) | Bridge type grab ship unloader fault prediction model method based on improved association rule | |
CN107704962B (en) | Steam flow interval prediction method based on incomplete training data set | |
CN113722985B (en) | Method and system for evaluating health state and predicting residual life of aero-engine | |
CN112785091A (en) | Method for performing fault prediction and health management on oil field electric submersible pump | |
Nguyen et al. | New methodology for improving the inspection policies for degradation model selection according to prognostic measures | |
CN114943179A (en) | Reliability evaluation and residual life prediction method based on multi-source degradation data fusion | |
Liao et al. | Nonparametric and semi-parametric sensor recovery in multichannel condition monitoring systems | |
Taha et al. | Aircraft engine remaining useful life prediction framework for industry 4.0 | |
CN113158448A (en) | Self-adaptive calculation method for equipment threshold of ship system | |
CN117521512A (en) | Bearing residual service life prediction method based on multi-scale Bayesian convolution transducer model | |
Xie et al. | Correlation analysis of aeroengine operation monitoring using deep learning | |
Wenqiang et al. | Remaining useful life prediction for mechanical equipment based on temporal convolutional network | |
CN114399024B (en) | Oil gas concentration big data intelligent detection system | |
CN116502123A (en) | Non-supervision cross-domain prediction method for residual service life of aero-engine | |
CN112560252B (en) | Method for predicting residual life of aeroengine |
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 |