CN109472110B - Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model - Google Patents
Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model Download PDFInfo
- Publication number
- CN109472110B CN109472110B CN201811441584.0A CN201811441584A CN109472110B CN 109472110 B CN109472110 B CN 109472110B CN 201811441584 A CN201811441584 A CN 201811441584A CN 109472110 B CN109472110 B CN 109472110B
- Authority
- CN
- China
- Prior art keywords
- engine
- lstm
- data
- service life
- sensor
- 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
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
- 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)
- Combined Controls Of Internal Combustion Engines (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an aeroengine residual service life prediction method based on an LSTM network and an ARIMA model, which comprises the following steps: according to the historical degradation data of the engine, establishing n engine health index (LSTM-HI) models based on an LSTM deep neural network; training an ARIMA model and predicting engine sensor parameters in a backward multi-step manner according to the past sensor data of the engine; and evaluating whether the engine is degenerated to be invalid or not according to the LSTM-HI index by using the predicted sensor parameters to obtain the residual service life of the engine and the probability distribution thereof. The invention provides a novel method for predicting the residual service life of the aero-engine, which has higher accuracy and feasibility and has positive promotion effects on real-time health management and maintenance cost reduction of the aero-engine.
Description
Technical Field
The invention belongs to the technical field of residual service life of an aeroengine, and particularly relates to an aeroengine residual service life prediction method based on an LSTM (Long Short-Term Memory) network and a ARIMA (Autoregressive Integrated Moving Average Model) model.
Background
Aeroengines operate throughout the year under severe conditions of high temperature, high pressure, high load, and their operating conditions change often, which results in frequent engine failure. The aero-engine is used as a heart of an aircraft, and the health condition of the aero-engine directly determines the personal safety of passengers and the safety of civil aviation. To ensure its safety, the engine must be serviced in a timely manner. However, the premature maintenance tends to increase the operation cost of the airline company, so that the maintenance time must be mastered at a proper time, thereby realizing the use value of the engine to the maximum extent. Accurately predicting the Remaining Useful Life (RUL) of an aircraft engine is an important measure to ensure flight safety and reduce maintenance costs.
With the development of artificial intelligence technology, a data-based fault diagnosis method provides a new solution for solving the health prediction problem of a nonlinear complex system. For intelligent engine control today, data-driven sensor fault diagnosis and resolution redundancy techniques are also gaining importance. Deep learning networks have gained extensive attention and research due to their strong nonlinear mapping capabilities. Recurrent Neural Networks (RNNs) can model changes in data over time, with success in the fields of natural language processing, speech recognition, and the like. Long short term memory networks (LSTM) are a type of deep learning network developed on the basis of RNNs that specifically address the problem of long term dependency of data. The method has the advantages that the long-term information of the data can be remembered by default without complex parameter adjustment, so that the long-term dependence problem is solved.
In addition, the rapid development of sensor technology and on-line monitoring technology makes it possible to integrate, transmit, analyze and tie engine performance and state data during the running process of an aircraft, which provides beneficial technical support and sufficient data preparation for the prediction of the remaining service life of the engine.
Disclosure of Invention
Aiming at the technical problems, the invention provides an aeroengine residual service life prediction method based on an LSTM network and an ARIMA model, which has higher accuracy and feasibility and has positive promotion effects on real-time health management and maintenance cost reduction of the aeroengine.
The technical scheme is as follows: in order to achieve the above purpose, the invention adopts the following technical scheme:
an aeroengine residual service life prediction method based on an LSTM network and an ARIMA model comprises the following steps:
step 1), according to engine history degradation data, establishing n engine health index (LSTM-HI) models based on an LSTM deep neural network, and constructing a health index model library reflecting the residual service life;
step 2), training an ARIMA model according to a past sensor data set of the engine to be tested, designing a prediction window length l, and predicting engine sensor parameters of the window length l backwards;
step 3), according to the predicted sensor parameters and the LSTM network health index model library, resolving LSTM-HI indexes, evaluating whether the engine is degenerated to be invalid, if the failure threshold is exceeded, fitting probability distribution characteristic parameters, and obtaining the residual service life of the engine; otherwise, turning to the step 2), increasing the prediction window length l, and sequentially calculating.
Preferably: in the step 1), according to the historical degradation data of the engine, the specific steps of establishing n engine health index (LSTM-HI) models based on the LSTM deep neural network are as follows:
step 1.1), selecting proper sensor parameters according to historical data of the aero-engine from health degradation to failure, and performing noise reduction and smoothing treatment to form n training data sets;
step 1.2), constructing an aeroengine health state evaluation index LSTM-HI based on a deep learning network LSTM and a training data set, wherein the specific expression is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
representing selected sensor parameters, k representing the number of selected sensor parameters, t representing the time of day;
step 1.3), training a deep learning network LSTM to obtain n LSTM-HI models, and establishing an aircraft engine health state evaluation index model library.
Preferably: training a deep learning network LSTM in the step 1.3) to obtain n LSTM-HI models, and establishing an aeroengine health state evaluation index model library specifically comprises the following steps:
step 1.3.1), pairIn a group of engine life-span observation data
Wherein x is t Representing the selected sensor parameter, +.>
Indicating the number of cycles to failure of the engine, i.e. its service life, HI t Is an artificial health index +.>
Let the input data at the current time be x t Output h of last time LSTM network t-1 The output of the LSTM network at the current time is calculated as follows:
f t ＝σ(w f ·[h t-1 ,x t ]+b f )
i t ＝σ(w i ·[h t-1 ,x t ]+b i )
o t ＝σ(w o ·[h t-1 ,x t ]+b o )
h t ＝o t ·Φ(S t )
wherein σ (·), g (·) represents the activation function, w represents the weight, b represents the bias, S t Representing the memory state quantity, i, of an LSTM network t Representing the output of the input gate, f t Output of forgetting gate, o t Representing the output of the output gate;
step 1.3.2), forward calculating the outputs of the LSTM network corresponding to all input samples by the step 1.3.1, and recording the errors of all training samples;
step 1.3.3), training an LSTM network by an error gradient descent method according to the error obtained by forward calculation until the precision meets the requirement;
step 1.3.4), training n LSTM networks respectively according to n groups of engine life-span observation data to obtain n LSTM-HI models, and establishing a model library.
Preferably: the specific steps of training the ARIMA model and predicting engine sensor parameters in a backward multi-step mode according to the past sensor data of the engine in the step 2) are as follows:
step 2.1) if the sensor data of the past degradation of the engine is known to be
Wherein x is t Representing selected sensor measurement parameters, r representing the current number of engine cycles; for each one of the one-dimensional selected sensor parameters, an ARIMA prediction model is established, and the expression is as follows:
w t ＝φ 1 ·w t-1 +φ 2 ·w t-2 +…+φ p ·w p-2
+δ+μ t +θ 1 ·μ t-1 +θ 2 ·μ t-2 +…+θ q ·μ q-2
wherein mu t Is a white noise sequence, delta is a constant representing that the sequence data is not 0-averaged, phi and theta represent coefficients of a difference term and a moving average term, respectively, w t A differential term representing the data at time t, which is expressed as:
where d represents the order of difference and L is the delay factor;
step 2.2), after determining the order of the ARIMA model, estimating parameters of the ARIMA model by a least square method according to past data;
and 2.3) respectively carrying out multi-step prediction on each sensor parameter according to the obtained ARIMA model to obtain the prediction data of each sensor.
Preferably: and 3) evaluating whether the engine is degraded to be invalid or not according to the LSTM-HI index by using the predicted sensor parameters in the step 3), and obtaining the residual service life of the engine and the probability distribution thereof, wherein the specific steps are as follows:
step 3.1), calculating a health index HI sequence corresponding to sensor prediction data, and defining a failure threshold value as theta f Judging that the engine is degraded to be invalid when HI exceeds a threshold value, and subtracting the current cycle number from the corresponding cycle number to obtain the residual service life RUL of the engine;
step 3.2), respectively calculating n HI sequences according to the LSTM-HI model library and the sensor parameter prediction data obtained in the step 2), and obtaining n engine RUL prediction values according to the step 3.1);
and 3.3) fitting n predicted values of the RUL of the engine by adopting a Gaussian distribution function to obtain probability distribution of the RUL of the engine.
The beneficial effects are that: the invention provides a novel method for predicting the residual service life of the aero-engine, which has higher accuracy and feasibility and has positive promotion effects on real-time health management and maintenance cost reduction of the aero-engine.
Drawings
FIG. 1 is a schematic illustration of a method for predicting remaining useful life of an aircraft engine.
FIG. 2 is a schematic diagram of a C-MAPSS system architecture.
Fig. 3 is a graph of the actual variation curve of the characteristic parameter and the curve after the smoothing process.
Fig. 4 is a graph showing the trend of the characteristic parameters after prediction by a window.
Fig. 5 is a predicted result of the remaining life of the test engine # 1.
FIG. 6 is a predicted remaining life of the test engine of # 2.
Fig. 7 is a result of predicting remaining life of the test engine # 3.
Fig. 8 is a test set of 20 engine residual life predictions.
Detailed Description
The following describes the embodiments of the present invention further with reference to the drawings.
The invention discloses an aeroengine residual service life prediction method based on an LSTM network and an ARIMA model, which specifically comprises the following steps:
step 1), according to engine history degradation data, establishing n engine health index (LSTM-HI) models based on an LSTM deep neural network, and constructing a health index model library reflecting the residual service life;
step 1.1), selecting proper sensor parameters according to historical data of the aero-engine from health degradation to failure, and performing noise reduction and smoothing treatment to form n training data sets;
step 1.2), constructing an aeroengine health state evaluation index LSTM-HI based on a deep learning network LSTM and a training data set, wherein the specific expression is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
represents the selected sensor parameters, k represents the number of selected sensor parameters, and t represents the time of day.
Step 1.3), training a deep learning network LSTM to obtain n LSTM-HI models, and establishing an aircraft engine health state evaluation index model library.
Step 1.3.1) for a set of engine life observations
Wherein x is t Representing the selected sensor parameter, +.>
Indicating the number of cycles to failure of the engine, i.e. its service life, HI t Is an artificial health index +.>
Let the input data at the current time be x t Output h of last time LSTM network t-1 The output of the LSTM network at the current time is calculated as follows:
f t ＝σ(w f ·[h t-1 ,x t ]+b f )
i t ＝σ(w i ·[h t-1 ,x t ]+b i )
o t ＝σ(w o ·[h t-1 ,x t ]+b o )
h t ＝o t ·Φ(S t )
wherein σ (·), g (·) represents the activation function, w represents the weight, b represents the bias, S t Representing the memory state quantity, i, of an LSTM network t Representing the output of the input gate, f t Output of forgetting gate, o t Representing the output of the output gate.
Step 1.3.2), forward calculating the outputs of the LSTM network corresponding to all input samples by the step 1.3.1, and recording the errors of all training samples;
step 1.3.3), training an LSTM network by an error gradient descent method according to the error obtained by forward calculation until the precision meets the requirement;
step 1.3.4), training n LSTM networks respectively according to n groups of engine life-span observation data to obtain n LSTM-HI models, and establishing a model library.
Step 2), training an ARIMA model according to a past sensor data set of the engine to be tested, designing a prediction window length l, and predicting engine sensor parameters of the window length l backwards;
step 2.1) if the sensor data of the past degradation of the engine is known to be
Wherein x is t Indicating the selected sensor measurement parameter, r indicating the current number of engine cycles. For each one of the one-dimensional selected sensor parameters, an ARIMA prediction model is established, and the expression is as follows:
w t ＝φ 1 ·w t-1 +φ 2 ·w t-2 +…+φ p ·w p-2
+δ+μ t +θ 1 ·μ t-1 +θ 2 ·μ t-2 +…+θ q ·μ q-2
wherein mu t Is a white noise sequence, delta is a constant representing that the sequence data is not 0-averaged, phi and theta represent coefficients of a difference term and a moving average term, respectively, w t A differential term representing the data at time t, which is expressed as:
where d represents the order of difference and L is the delay factor.
Step 2.2), after determining the order of the ARIMA model, estimating parameters of the ARIMA model by a least square method according to past data;
and 2.3) respectively carrying out multi-step prediction on each sensor parameter according to the obtained ARIMA model to obtain the prediction data of each sensor.
Step 3), according to the predicted sensor parameters and the LSTM network health index model library, resolving LSTM-HI indexes, evaluating whether the engine is degenerated to be invalid, if the failure threshold is exceeded, fitting probability distribution characteristic parameters, and obtaining the residual service life of the engine; otherwise, turning to the step 2), increasing the length l of the prediction window, and sequentially calculating;
step 3.1), calculating a health index HI sequence corresponding to the sensor prediction data. Defining a failure threshold value as θ f Judging the degradation of the engine to failure when HI exceeds a threshold value, which corresponds toSubtracting the current cycle number from the cycle number of the engine to obtain the residual service life RUL of the engine;
step 3.2), respectively calculating n HI sequences according to the LSTM-HI model library and the sensor parameter prediction data obtained in the step 2), and obtaining n engine RUL prediction values according to the step 3.1);
and 3.3) fitting n predicted values of the RUL of the engine by adopting a Gaussian distribution function to obtain probability distribution of the RUL of the engine.
In order to verify the effectiveness of the intelligent analysis redundancy design method of the aeroengine sensor based on the KEOS-ELM algorithm, the invention carries out relevant digital simulation in an MATLAB environment. In this simulation, the data in the two files "train_fd001.Txt" and "test_fd001.Txt" in the C-MAPSS dataset are used as training data and test data, and the links and layout of the respective modules of the C-MAPSS are shown in fig. 2. Wherein the training data set contains 100 groups of full life observation data of different engines from health state degradation to failure, so the invention trains 100 LSTM-HI models and builds an LSTM-HI model library. The test set contains 100 groups of observed data of different engine degradation to different degrees, and 20 groups are randomly taken from the test set to predict the residual service life in order to verify the performance of the method. By analyzing the 21 sensor data, 5 parameters including T24, nf, ps30, NRf and BPR were selected for engine performance parameter prediction, and the description of each parameter is shown in table 1.
Table 1 engine sensor measurement parameters
Because the original data has larger noise, the original data is normalized and smoothed to obtain more real engine data. FIG. 3 is a graph of various sensor parameters for a set of training data after noise reduction and smoothing. Fig. 4 is a result of a multi-step prediction of parameters via ARIMA model with a first set of engine test data (# 1). Fig. 5, 6 and 7 are graphs of the HI variation curves and resulting residual life predictions and their probability distributions for 1-3 sets of engine test data (# 1, #2 and # 3) were selected. Table 2 shows the predicted results obtained for all 20 sets of engine test data.
Table 2 results of residual life predictions for all 20 sets of engine test data
As can be seen from fig. 5, 6 and 7 and table 2, the prediction method of the remaining service life of the engine provided by the invention has higher prediction accuracy, the relative prediction error is about 10%, and the absolute error is mostly within 10 cycles. Fig. 8 is a graph of predicted results for all 20 sets of engine test data. However, the prediction error is large for individual engines such as #4, #10 and #16, which may be due to the small number of current degradation cycles, less past sensor data, and the ARIMA model is not able to model its degradation process effectively, thus resulting in large prediction errors. But in general, the method for predicting the residual service life of the aero-engine based on the LSTM network and the ARIMA model can effectively predict the residual service life of the aero-engine, and has higher accuracy and feasibility.
It should be noted that the foregoing is only specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any changes and substitutions that can be easily contemplated by those skilled in the art within the technical scope of the present invention should be covered by the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (4)
1. An aeroengine residual service life prediction method based on an LSTM network and an ARIMA model is characterized by comprising the following steps of: the method comprises the following steps:
step 1), according to engine history degradation data, establishing n engine health index LSTM-HI models based on an LSTM deep neural network, and constructing a health index model library reflecting the residual service life;
step 2), training an ARIMA model according to a past sensor data set of the engine to be tested, designing an engine sensor parameter with a prediction window length of l and a backward prediction window length of l;
step 3), according to the predicted sensor parameters and the LSTM network health index model library, resolving LSTM-HI indexes, evaluating whether the engine is degenerated to be invalid, if the failure threshold is exceeded, fitting probability distribution characteristic parameters, and obtaining the residual service life of the engine; otherwise, turning to the step 2), increasing the length l of the predicted window on the original predicted window length, and sequentially calculating;
in the step 1), according to the historical degradation data of the engine, the specific steps of establishing n LSTM-HI models of the engine health index based on the LSTM deep neural network are as follows:
step 1.1), selecting proper sensor parameters according to historical data of the aero-engine from health degradation to failure, and performing noise reduction and smoothing treatment to form n training data sets;
step 1.2), constructing an engine health index LSTM-HI based on an LSTM deep learning network and a training data set, wherein the specific expression is as follows:
wherein HI t Is a health index of an artificial structure,
representing selected sensor parameters, k representing the number of selected sensor parameters, t representing the time of day;
step 1.3), training a deep learning network LSTM to obtain n LSTM-HI models, and establishing an aircraft engine health state evaluation index model library.
2. The method for predicting the remaining service life of an aeroengine based on an LSTM network and an ARIMA model according to claim 1, wherein the method comprises the following steps: training a deep learning network LSTM in the step 1.3) to obtain n LSTM-HI models, and establishing an aeroengine health state evaluation index model library specifically comprises the following steps:
step 1.3.1) for a set of engine life observations
Wherein x is t Representing the selected sensor parameter, +.>
Indicating the number of cycles to failure of the engine, i.e. its service life, HI t Is an artificial health index +.>
Let the input data at the current time be x t Output h of last time LSTM network t-1 The output of the LSTM network at the current time is calculated as follows:
f t ＝σ(w f ·[h t-1 ,x t ]+b f )
i t ＝σ(w i ·[h t-1 ,x t ]+b i )
o t ＝σ(w o ·[h t-1 ,x t ]+b o )
h t ＝o t ·Φ(S t )
wherein σ (·), Φ (·) represents the activation function, w represents the weight, b represents the bias, S t Representing the memory state quantity, i, of an LSTM network t Representing the output of the input gate, f t Output of forgetting gate, o t Representing the output of the output gate;
step 1.3.2), forward calculating the outputs of the LSTM network corresponding to all input samples by the step 1.3.1, and recording the errors of all training samples;
step 1.3.3), training an LSTM network by an error gradient descent method according to the error obtained by forward calculation until the precision meets the requirement;
step 1.3.4), training n LSTM networks respectively according to n groups of engine life-span observation data to obtain n LSTM-HI models, and establishing a model library.
3. The method for predicting the remaining service life of an aeroengine based on an LSTM network and an ARIMA model according to claim 1, wherein the method comprises the following steps: the specific steps of training the ARIMA model and predicting engine sensor parameters in a backward multi-step mode according to the past sensor data of the engine in the step 2) are as follows:
step 2.1) if the sensor data of the past degradation of the engine is known to be
Wherein x is t Representing selected sensor measurement parameters, r representing the current number of engine cycles; for each one of the one-dimensional selected sensor parameters, an ARIMA prediction model is established, and the expression is as follows:
w t ＝φ 1 ·w t-1 +φ 2 ·w t-2 +L+φ p ·w p-2 +δ+μ t +θ 1 ·μ t-1 +θ 2 ·μ t-2 +L+θ q ·μ q-2
wherein mu t Is a white noise sequence, delta is a constant representing that the sequence data is not 0-averaged, phi and theta represent coefficients of a difference term and a moving average term, respectively, w t Representing data inA differential term at time t, expressed as:
where d represents the order of difference and L is the delay factor;
step 2.2), after determining the order of the ARIMA model, estimating parameters of the ARIMA model by a least square method according to past data;
and 2.3) respectively carrying out multi-step prediction on each sensor parameter according to the obtained ARIMA model to obtain the prediction data of each sensor.
4. The method for predicting the remaining service life of an aeroengine based on an LSTM network and an ARIMA model according to claim 1, wherein the method comprises the following steps: and 3) evaluating whether the engine is degraded to be invalid or not according to the LSTM-HI index by using the predicted sensor parameters in the step 3), and obtaining the residual service life of the engine and the probability distribution thereof, wherein the specific steps are as follows:
step 3.1), calculating a health index HI sequence corresponding to sensor prediction data, and defining a failure threshold value as theta f Judging that the engine is degraded to be invalid when HI exceeds a threshold value, and subtracting the current cycle number from the corresponding cycle number to obtain the residual service life RUL of the engine;
step 3.2), respectively calculating n HI sequences according to the LSTM-HI model library and the sensor parameter prediction data obtained in the step 2), and obtaining n engine RUL prediction values according to the step 3.1);
and 3.3) fitting n predicted values of the RUL of the engine by adopting a Gaussian distribution function to obtain probability distribution of the RUL of the engine.
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 CN109472110A (en) | 2019-03-15 |
| CN109472110B true 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) |
Families Citing this family (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109991542B (en) * | 2019-03-27 | 2021-05-18 | 东北大学 | Lithium ion battery residual life prediction method based on WDE optimization LSTM network |
| CN110108957B (en) * | 2019-05-07 | 2020-06-02 | 武汉理工大学 | Tractor electrical fault diagnosis method based on structural analysis method |
| 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 |
| CN110414033B (en) * | 2019-05-31 | 2023-04-18 | 太原理工大学 | Mechanical equipment predictive maintenance method combining edge calculation and digital twinning |
| 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 |
| 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 |
| CN110807257A (en) * | 2019-11-04 | 2020-02-18 | 中国人民解放军国防科技大学 | Method for predicting residual life of aircraft engine |
| 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 |
| CN111709577B (en) * | 2020-06-17 | 2023-08-15 | 上海海事大学 | RUL prediction method based on long-range correlation GAN-LSTM |
| CN112668249B (en) * | 2020-07-17 | 2023-05-02 | 国网山东省电力公司电力科学研究院 | Online construction method and system for power grid primary equipment overhaul technical modification scheme |
| US11952149B2 (en) | 2020-09-17 | 2024-04-09 | General Electric Company | System and method of forecasting aircraft engine operational data for predictive analytics |
| CN112232404B (en) * | 2020-10-13 | 2023-12-22 | 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 abnormality and operation and maintenance information of railway power supply equipment |
| 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 |
| US20220172038A1 (en) * | 2020-11-30 | 2022-06-02 | International Business Machines Corporation | Automated deep learning architecture selection for time series prediction with user interaction |
| CN112560347B (en) * | 2020-12-17 | 2022-09-20 | 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 |
| 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 |
| CN112926273B (en) * | 2021-04-13 | 2023-04-18 | 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 |
| 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 |
| CN114282443B (en) * | 2021-12-28 | 2023-03-17 | 浙江大学 | Residual service life prediction method based on MLP-LSTM supervised joint model |
| 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 |
| CN116147928B (en) * | 2023-04-20 | 2023-07-14 | 清华大学 | Method, device and equipment for determining health state of aeroengine thermal jet device |
| 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 |
Citations (2)
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI603210B (en) * | 2016-12-13 | 2017-10-21 | 財團法人工業技術研究院 | 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 (2)
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| 基于CNN与GRU的航空发动机剩余寿命预测;王文庆等;《第37届中国控制会议论文集（F）》;20180725;第9118-9123页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109472110A (en) | 2019-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109472110B (en) | Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model | |
| Zheng et al. | Long short-term memory network for remaining useful life estimation | |
| Xiang et al. | LSTM networks based on attention ordered neurons for gear remaining life prediction | |
| CN109766583B (en) | Aircraft engine life prediction method based on unlabeled, unbalanced and initial value uncertain data | |
| CN110955963B (en) | Aviation cable residual life prediction method | |
| CN110807257A (en) | Method for predicting residual life of aircraft engine | |
| CN111274737A (en) | Method and system for predicting remaining service life of mechanical equipment | |
| CN110222371B (en) | Bayes and neural network-based engine residual life online prediction method | |
| CN110361180B (en) | Intelligent train pantograph service performance dynamic monitoring and evaluating method and system | |
| CN101436057A (en) | Numerical control machining tool heat error Bayes network compensation method | |
| Li et al. | Domain adaptation remaining useful life prediction method based on AdaBN-DCNN | |
| CN110824914B (en) | Intelligent wastewater treatment monitoring method based on PCA-LSTM network | |
| CN114297918A (en) | Aero-engine residual life prediction method based on full-attention depth network and dynamic ensemble learning | |
| Mo et al. | Few-shot RUL estimation based on model-agnostic meta-learning | |
| Wang et al. | A remaining useful life prediction model based on hybrid long-short sequences for engines | |
| CN114266201B (en) | Self-attention elevator trapping prediction method based on deep learning | |
| CN115062528A (en) | Prediction method for industrial process time sequence data | |
| Ruan et al. | Remaining useful life prediction for aero-engine based on lstm and cnn | |
| CN114239990A (en) | Time series data prediction method based on time series decomposition and LSTM | |
| Chen et al. | Remaining useful life prediction of turbofan engine using global health degradation representation in federated learning | |
| Zhao et al. | Rolling bearing remaining useful life prediction based on wiener process | |
| CN113158448A (en) | Self-adaptive calculation method for equipment threshold of ship system | |
| CN112257893A (en) | Complex electromechanical system health state prediction method considering monitoring error | |
| Kakati et al. | Remaining useful life predictions for turbofan engine degradation using online long short-term memory network | |
| CN115600695B (en) | Fault diagnosis method for metering equipment |
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 |