CN113449472A - Turbofan engine residual life prediction method - Google Patents
Turbofan engine residual life prediction method Download PDFInfo
- Publication number
- CN113449472A CN113449472A CN202110720208.0A CN202110720208A CN113449472A CN 113449472 A CN113449472 A CN 113449472A CN 202110720208 A CN202110720208 A CN 202110720208A CN 113449472 A CN113449472 A CN 113449472A
- Authority
- CN
- China
- Prior art keywords
- stage
- model
- performance degradation
- data
- turbofan engine
- 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.)
- Pending
Links
Images
Classifications
-
- 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
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Combined Controls Of Internal Combustion Engines (AREA)
Abstract
The invention provides a method for predicting the residual life of a turbofan engine, which divides the performance degradation process of the turbofan engine into the following steps: an effective working stage, a slow degradation stage and a rapid degradation stage; processing the state monitoring data obtained from the sensor, carrying out normalization processing, and analyzing and reducing the dimension by using a principal component; calculating permutation entropy of the dimensionality reduction signal, extracting performance degradation characteristics by using the permutation entropy, and identifying a degradation stage; selecting different prediction models according to the identified different performance degradation stages for sectional prediction; compared with the prior art, the method can effectively improve the accuracy of predicting the residual life of the turbofan engine, reduce the error of life prediction and provide a basis for the optional maintenance of the engine.
Description
Technical Field
The invention belongs to the technical field of fault prediction and health management, and particularly relates to a method for predicting the residual life of a turbofan engine.
Background
The maintenance mode of the fan engine is mainly a timing maintenance mode and maintenance is carried out according to the maintenance time determined by a manufacturer. The timing maintenance mode does not consider the condition that the failure time of the turbofan engine is different due to different operating environments, so that corresponding excessive maintenance is caused, and the waste of component resources and the increase of maintenance cost are brought. The mode of maintenance according to the situation can avoid excessive maintenance, and when the residual service life of turbofan engine can not guarantee the normal operation of the turbofan engine, the maintenance work of the turbofan engine is carried out, the maintenance efficiency of the turbofan engine is improved, and the maintenance cost is reduced. The premise of the optional maintenance is to know the remaining life of the engine, so that a good method for predicting the remaining life of the turbofan engine is needed.
The performance degradation process of a turbofan engine typically goes through various stages, such as an active operating stage, a slow degradation stage, a fast degradation stage, etc. When the turbofan engine is just started to operate, the performance degradation of the turbofan engine is not obvious, the damage growth speed is slow in a slow degradation stage, the damage growth speed is fast in a fast degradation stage, and if the performance degradation stages are not distinguished, a single life prediction model is used for predicting the residual lives of different degradation stages, large life prediction errors can be caused.
Disclosure of Invention
The invention provides a method for predicting the residual life of a turbofan engine, aiming at the problem that the prior art does not distinguish performance degradation stages and can cause larger life prediction errors.
The invention is realized by the following scheme:
a method for predicting the residual life of a turbofan engine comprises the following steps:
a method for predicting the residual life of a turbofan engine comprises the following steps:
the performance degradation process of the turbofan engine is divided into 3 stages: the stage 1 is an effective working stage, the stage 2 is a slow degradation stage, and the stage 3 is a rapid degradation stage;
the prediction method comprises the following steps:
step 1, processing state monitoring data obtained from a sensor, carrying out normalization processing, and analyzing and reducing dimensions by using principal components;
step 2, calculating the permutation entropy of the dimensionality reduction signal, extracting performance degradation characteristics by using the permutation entropy, and identifying a degradation stage;
and 3, selecting different prediction models according to the identified different performance degradation stages, and performing segmented prediction.
Further: in the first step:
step 1.1, generating turbofan engine state monitoring data by using a C-MAPSS simulator; carrying out non-dimensionalization processing on the turbofan engine state monitoring data by adopting a maximum and minimum normalization method, and limiting the data size of each state monitoring parameter between [0 and 1], wherein the maximum and minimum normalization formula is as follows:
in the formula (I), the compound is shown in the specification,for the parameters of the non-dimensionalization,is the smallest value in the column and,is the largest value in the column;
step 1.2, reducing the dimension of the data by adopting a principal component analysis method;
let x1,x2,...,xnEach vector of the n times of data collected by the sensors representing the turbofan engine, wherein each vector has m variables, corresponding to the operating parameters and the condition monitoring parameters of the turbofan engine, the variables form an initial data set matrix X, namely:
in the formula, Xi=(x1i,x2i,...,xni)T,i=1,2,...,m;
And (3) calculating a covariance matrix of X of the data matrix, namely:
for covariance matrix CXAnd (3) carrying out eigenvalue decomposition to obtain an eigenvector U, namely:
CX=UDUT (4)
wherein D ═ diag (λ)1,λ2,...,λm),λ1>λ2>...>λmRepresenting a feature value diagonal matrix arranged in descending order of numerical value; u ═ U (U)1,U2,...,Um) Representing a characteristic value λiCorresponding feature vector UiU is a standard orthogonal matrix;
carrying out linear transformation on the coherent accumulation signal data matrix X by using the eigenvector matrix U to obtain each principal component vector Y, namely:
Y=UTX (5)
and then calculating to obtain the cumulative contribution rate corresponding to the first K principal components, namely:
calculating the reverse value of K of the accumulated contribution rate of each dimension according to the threshold limit, wherein K is the dimension after the dimension reduction of the principal component analysis;
and after Y dimension reduction is set, taking the first K principal component matrixes, and then:
Y=XUY (7)
in the formula of UY=(U1,U2,...,UK)
And according to the accumulated contribution rates corresponding to the first K principal components, retaining the data of the vectors of the first K principal components for data reconstruction, and realizing the mapping from high-dimensional data to low-dimensional data.
Further, in step two:
step 2.1, the selected permutation entropy parameter is embedded into the dimension d reduced in the step one, and the time delay τ is 1, and phase space reconstruction is performed on the original time sequence to obtain a corresponding d-dimension reconstruction vector: x (i) { x (i), x (i + τ), …, x (i + (d-1) τ) };
step 2.2, according to formula Hp=Hp(d) Calculating a normalized permutation entropy value by the ln (d!), and obtaining the permutation entropy value with the interval of 0 to 1 as an evaluation index of the degradation stage;
step 2.3, if the permutation entropies are all smaller than the threshold value, the performance degradation stage of the engine is considered to be in stage 1;
if 1 permutation entropy exceeds the threshold value, the engine performance degradation stage is considered to be in stage 2;
if the permutation entropies all exceed the threshold, the engine performance degradation phase is considered to be in phase 3.
Further, in step three:
if the performance degradation stage 1 is in an effective working stage, the model does not need to be updated, and the XGboost model is selected for service life prediction;
if the performance degradation stage 2 is in a slow degradation stage, the model needs to be updated, and an LSTM model is selected for service life prediction;
and if the performance degradation stage 3 is in a rapid degradation stage, the model needs to be updated, and the CNN-JANET model is selected for service life prediction.
Further, the air conditioner is provided with a fan,
the model training parameters were set as follows:
the hyper-parameters of the XGboost model corresponding to the phase 1 are set as follows: the learning rate is 0.00521, the maximum tree depth is 7, the total number of trees is 4000, the minimum sample number of trees is 5, and the sample sampling proportion is 0.9;
the LSTM model hyper-parameters corresponding to phase 2 are set as: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the activation function is tanh, and the others are default values;
the hyper-parameter setting of the CNN-JANET model corresponding to the stage 3 is as follows: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the number of iterations is 20, and the time window length is 100.
Further, the air conditioner is provided with a fan,
the three performance degradation stages can use the same model for life prediction, the model is trained by using data of different performance degradation stages, and the model can be selected from an XGboost model, an LSTM model and a CNN-JANET model.
The invention has the beneficial effects
(1) The method divides the performance degradation stage of the engine into 3 different stages, and in the life prediction process, a model with short training time and high prediction precision is selected for updating according to the characteristics of each stage;
(2) compared with the prior art, the method can effectively improve the accuracy of predicting the residual life of the turbofan engine, reduce the error of life prediction and provide a basis for the optional maintenance of the engine.
Drawings
Fig. 1 is a schematic block diagram of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments; all other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A method for predicting the residual life of a turbofan engine,
the performance degradation process of the turbofan engine is divided into 3 stages: the stage 1 is an effective working stage, the stage 2 is a slow degradation stage, and the stage 3 is a rapid degradation stage;
the prediction method comprises the following steps:
step 1, processing state monitoring data obtained from a sensor, carrying out normalization processing, and analyzing and reducing dimensions by using principal components;
step 2, calculating the permutation entropy of the dimensionality reduction signal, extracting performance degradation characteristics by using the permutation entropy, and identifying a degradation stage;
and 3, selecting different prediction models according to the identified different performance degradation stages, and performing segmented prediction.
The first embodiment is as follows:
the C-MAPSS simulator is used for generating turbofan engine state monitoring data, and a turbofan engine degradation process of a turbofan engine system level engine with 40k of thrust after a series of flight tasks is simulated, wherein each flight is about 90 minutes.
In the first step:
step 1.1, generating turbofan engine state monitoring data by using a C-MAPSS simulator; carrying out non-dimensionalization processing on the turbofan engine state monitoring data by adopting a maximum and minimum normalization method, and limiting the data size of each state monitoring parameter between [0 and 1], wherein the maximum and minimum normalization formula is as follows:
in the formula (I), the compound is shown in the specification,for the parameters of the non-dimensionalization,is the smallest value in the column and,is the largest value in the column;
step 1.2, reducing the dimension of the data by adopting a principal component analysis method;
let x1,x2,...,xnEach vector in n times of data collected by a sensor used for representing the turbofan engine, wherein each vector has m variables corresponding to operating parameters and state monitoring parameters of the turbofan engine, and the variables form an initial data set matrix XNamely:
in the formula, Xi=(x1i,x2i,...,xni)T,i=1,2,...,m;
And (3) calculating a covariance matrix of X of the data matrix, namely:
for covariance matrix CXAnd (3) carrying out eigenvalue decomposition to obtain an eigenvector U, namely:
CX=UDUT (4)
wherein D ═ diag (λ)1,λ2,...,λm),λ1>λ2>...>λmRepresenting a feature value diagonal matrix arranged in descending order of numerical value; u ═ U (U)1,U2,...,Um) Representing a characteristic value λiCorresponding feature vector UiU is a standard orthogonal matrix;
carrying out linear transformation on the coherent accumulation signal data matrix X by using the eigenvector matrix U to obtain each principal component vector Y, namely:
Y=UTX (5)
and then calculating to obtain the cumulative contribution rate corresponding to the first K principal components, namely:
calculating the reverse value of K of the accumulated contribution rate of each dimension according to the threshold limit, wherein K is the dimension after the dimension reduction of the principal component analysis;
and after Y dimension reduction is set, taking the first K principal component matrixes, and then:
Y=XUY (7)
in the formula of UY=(U1,U2,...,UK)
And according to the accumulated contribution rates corresponding to the first K principal components, retaining the data of the vectors of the first K principal components for data reconstruction, and realizing the mapping from high-dimensional data to low-dimensional data.
In this embodiment, the threshold is set to 0.94, and the sensor data of 21 dimensions is reduced to 6 dimensions.
In the second step:
step 2.1, the dimension d reduced in the selected permutation entropy parameter embedding step one is 6, the time delay τ is 1, and phase space reconstruction is performed on the original time series to obtain a corresponding d-dimension reconstruction vector: x (i) { x (i), x (i + τ), …, x (i + (d-1) τ) };
step 2.2, according to formula Hp=Hp(d) Calculating a normalized permutation entropy value by the ln (d!), and obtaining the permutation entropy value with the interval of 0 to 1 as an evaluation index of the degradation stage;
step 2.3, if the permutation entropies are all smaller than the threshold value 0.7, the performance degradation stage of the engine is considered to be in stage 1;
if 1 permutation entropy exceeds the threshold value of 0.7, the engine performance degradation stage is considered to be in stage 2;
if the permutation entropies all exceed the threshold value of 0.85, the engine performance degradation stage is considered to be in stage 3.
In step three:
if the performance degradation stage 1 is in an effective working stage, the model does not need to be updated, and the XGboost model is selected for service life prediction;
if the performance degradation stage 2 is in a slow degradation stage, the model needs to be updated, and an LSTM model is selected for service life prediction;
and if the performance degradation stage 3 is in a rapid degradation stage, the model needs to be updated, and the CNN-JANET model is selected for service life prediction.
The model training parameters were set as follows:
the hyper-parameters of the XGboost model corresponding to the phase 1 are set as follows: the learning rate is 0.00521, the maximum tree depth is 7, the total number of trees is 4000, the minimum sample number of trees is 5, and the sample sampling proportion is 0.9;
the LSTM model hyper-parameters corresponding to phase 2 are set as: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the activation function is tanh, and the others are default values;
the hyper-parameter setting of the CNN-JANET model corresponding to the stage 3 is as follows: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the number of iterations is 20, and the time window length is 100.
Example two:
the three performance degradation stages can use the same model for life prediction, the model is trained by using data of different performance degradation stages, and the model can be selected from an XGboost model, an LSTM model and a CNN-JANET model.
The prediction models of different performance degradation stages can also be the same model, and the models trained by the data of different performance degradation stages are used for prediction.
The present embodiment is further described with reference to the method for predicting the remaining life of a turbofan engine described in embodiment 1. In this embodiment, in the prediction model selection in step 3, a model trained by using different performance degradation stage data in the same model may be selected. Taking the XGBoost model as an example, the details are as follows:
the performance degradation stage 1 is in an effective working stage, and an XGboost model 1 trained by using the data in the stage 1 is selected for service life prediction; the performance degradation stage 2 is in a slow degradation stage, and an XGboost model 2 trained by using data in the stage 2 is selected for service life prediction; and the performance degradation stage 3 is in a rapid degradation stage, and the XGboost model 3 trained by using the data in the stage 3 is selected for service life prediction.
The method for predicting the remaining life of the turbofan engine, which is provided by the invention, is described in detail, and the principle and the implementation mode of the invention are explained, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (6)
1. A method for predicting the residual life of a turbofan engine is characterized by comprising the following steps:
the performance degradation process of the turbofan engine is divided into 3 stages: the stage 1 is an effective working stage, the stage 2 is a slow degradation stage, and the stage 3 is a rapid degradation stage;
the prediction method comprises the following steps:
step 1, processing state monitoring data obtained from a sensor, carrying out normalization processing, and analyzing and reducing dimensions by using principal components;
step 2, calculating the permutation entropy of the dimensionality reduction signal, extracting performance degradation characteristics by using the permutation entropy, and identifying a degradation stage;
and 3, selecting different prediction models according to the identified different performance degradation stages, and performing segmented prediction.
2. The prediction method according to claim 1, characterized in that: in the first step:
step 1.1, generating turbofan engine state monitoring data by using a C-MAPSS simulator; carrying out non-dimensionalization processing on the turbofan engine state monitoring data by adopting a maximum and minimum normalization method, and limiting the data size of each state monitoring parameter between [0 and 1], wherein the maximum and minimum normalization formula is as follows:
in the formula (I), the compound is shown in the specification,for the parameters of the non-dimensionalization,is the smallest value in the column and,is the largest value in the column;
step 1.2, reducing the dimension of the data by adopting a principal component analysis method;
let x1,x2,...,xnEach vector of the n times of data collected by the sensors representing the turbofan engine, wherein each vector has m variables, corresponding to the operating parameters and the condition monitoring parameters of the turbofan engine, the variables form an initial data set matrix X, namely:
in the formula, Xi=(x1i,x2i,...,xni)T,i=1,2,...,m;
And (3) calculating a covariance matrix of X of the data matrix, namely:
for covariance matrix CXAnd (3) carrying out eigenvalue decomposition to obtain an eigenvector U, namely:
CX=UDUT (4)
wherein D ═ diag (λ)1,λ2,...,λm),λ1>λ2>...>λmRepresenting a feature value diagonal matrix arranged in descending order of numerical value; u ═ U (U)1,U2,...,Um) Representing a characteristic value λiCorresponding feature vector UiU is a standard orthogonal matrix;
carrying out linear transformation on the coherent accumulation signal data matrix X by using the eigenvector matrix U to obtain each principal component vector Y, namely:
Y=UTX (5)
and then calculating to obtain the cumulative contribution rate corresponding to the first K principal components, namely:
calculating the reverse value of K of the accumulated contribution rate of each dimension according to the threshold limit, wherein K is the dimension after the dimension reduction of the principal component analysis;
and after Y dimension reduction is set, taking the first K principal component matrixes, and then:
Y=XUY (7)
in the formula of UY=(U1,U2,...,UK)
And according to the accumulated contribution rates corresponding to the first K principal components, retaining the data of the vectors of the first K principal components for data reconstruction, and realizing the mapping from high-dimensional data to low-dimensional data.
3. The prediction method according to claim 2, characterized in that: in the second step:
step 2.1, the selected permutation entropy parameter is embedded into the dimension d reduced in the step one, and the time delay τ is 1, and phase space reconstruction is performed on the original time sequence to obtain a corresponding d-dimension reconstruction vector: x (i) { x (i), x (i + τ), …, x (i + (d-1) τ) };
step 2.2, according to formula Hp=Hp(d) Calculating a normalized permutation entropy value by the ln (d!), and obtaining the permutation entropy value with the interval of 0 to 1 as an evaluation index of the degradation stage;
step 2.3, if the permutation entropies are all smaller than the threshold value, the performance degradation stage of the engine is considered to be in stage 1;
if 1 permutation entropy exceeds the threshold value, the engine performance degradation stage is considered to be in stage 2;
if the permutation entropies all exceed the threshold, the engine performance degradation phase is considered to be in phase 3.
4. The prediction method according to claim 3, wherein: in step three:
if the performance degradation stage 1 is in an effective working stage, the model does not need to be updated, and the XGboost model is selected for service life prediction;
if the performance degradation stage 2 is in a slow degradation stage, the model needs to be updated, and an LSTM model is selected for service life prediction;
and if the performance degradation stage 3 is in a rapid degradation stage, the model needs to be updated, and the CNN-JANET model is selected for service life prediction.
5. The prediction method according to claim 4, wherein:
the model training parameters were set as follows:
the hyper-parameters of the XGboost model corresponding to the phase 1 are set as follows: the learning rate is 0.00521, the maximum tree depth is 7, the total number of trees is 4000, the minimum sample number of trees is 5, and the sample sampling proportion is 0.9;
the LSTM model hyper-parameters corresponding to phase 2 are set as: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the activation function is tanh, and the others are default values;
the hyper-parameter setting of the CNN-JANET model corresponding to the stage 3 is as follows: the learning rate is 0.001, the hidden layer is 100, the output layer is 1, the batch size is 1024, the number of iterations is 20, and the time window length is 100.
6. The prediction method according to claim 3, wherein:
the three performance degradation stages can use the same model for life prediction, the model is trained by using data of different performance degradation stages, and the model can be selected from an XGboost model, an LSTM model and a CNN-JANET model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110720208.0A CN113449472A (en) | 2021-06-28 | 2021-06-28 | Turbofan engine residual life prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110720208.0A CN113449472A (en) | 2021-06-28 | 2021-06-28 | Turbofan engine residual life prediction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113449472A true CN113449472A (en) | 2021-09-28 |
Family
ID=77813375
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110720208.0A Pending CN113449472A (en) | 2021-06-28 | 2021-06-28 | Turbofan engine residual life prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113449472A (en) |
Citations (5)
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 |
CN109033499A (en) * | 2018-06-07 | 2018-12-18 | 南京航空航天大学 | A kind of aero-engine method for predicting residual useful life of multistage consistency check |
CN109916634A (en) * | 2019-02-26 | 2019-06-21 | 武汉科技大学 | A kind of aviation turbofan engine remaining life prediction technique and system |
CN112287596A (en) * | 2020-09-22 | 2021-01-29 | 南京航空航天大学 | Engine residual life prediction method based on clustering and LSTM |
CN112683535A (en) * | 2021-01-14 | 2021-04-20 | 大连理工大学 | Bearing life prediction method based on multi-stage wiener process |
-
2021
- 2021-06-28 CN CN202110720208.0A patent/CN113449472A/en active Pending
Patent Citations (5)
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 |
CN109033499A (en) * | 2018-06-07 | 2018-12-18 | 南京航空航天大学 | A kind of aero-engine method for predicting residual useful life of multistage consistency check |
CN109916634A (en) * | 2019-02-26 | 2019-06-21 | 武汉科技大学 | A kind of aviation turbofan engine remaining life prediction technique and system |
CN112287596A (en) * | 2020-09-22 | 2021-01-29 | 南京航空航天大学 | Engine residual life prediction method based on clustering and LSTM |
CN112683535A (en) * | 2021-01-14 | 2021-04-20 | 大连理工大学 | Bearing life prediction method based on multi-stage wiener process |
Non-Patent Citations (2)
Title |
---|
沈泊言: "涡扇发动机剩余寿命预测算法研究及组件开发", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
车畅畅等: "基于多尺度排列熵和长短时记忆神经网络的航空发动机剩余寿命预测", 《交通运输工程学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110688288B (en) | Automatic test method, device, equipment and storage medium based on artificial intelligence | |
CN114282579A (en) | Aviation bearing fault diagnosis method based on variational modal decomposition and residual error network | |
CN112461537A (en) | Wind power gear box state monitoring method based on long-time neural network and automatic coding machine | |
CN112257263B (en) | Equipment residual life prediction system based on self-attention mechanism | |
CN113743016B (en) | Engine residual life prediction method based on self-encoder and echo state network | |
CN113642508A (en) | Bearing fault diagnosis method based on parameter self-adaptive VMD and optimized SVM | |
CN115204491A (en) | Production line working condition prediction method and system based on digital twinning and LSTM | |
CN110766215B (en) | Wind power climbing event prediction method based on feature adaptive selection and WDNN | |
CN114169091A (en) | Method for establishing prediction model of residual life of engineering mechanical part and prediction method | |
CN111766513B (en) | Capsule network-based variable-working-condition multi-fault diagnosis method for three-phase induction motor | |
CN116292130A (en) | Wind driven generator state monitoring system and method based on cooperative analysis | |
CN116738868A (en) | Rolling bearing residual life prediction method | |
CN114962390A (en) | Hydraulic system fault diagnosis method and system and working machine | |
CN114418146A (en) | Electromechanical equipment fault diagnosis method, system and terminal based on rule engine | |
CN113449472A (en) | Turbofan engine residual life prediction method | |
CN116908684A (en) | Motor fault prediction method and device, electronic equipment and storage medium | |
CN116861214A (en) | Health state identification method and system based on convolution long short-time memory network | |
CN117154680A (en) | Wind power prediction method based on non-stationary transducer model | |
CN114792026A (en) | Method and system for predicting residual life of aircraft engine equipment | |
CN115114770B (en) | Baseline self-adaptive auxiliary power device performance trend analysis method | |
CN116010896B (en) | Wind driven generator fault diagnosis method based on countermeasure training and transducer | |
CN116150666B (en) | Energy storage system fault detection method and device and intelligent terminal | |
CN117992863B (en) | Rotary machine fault diagnosis method based on interpretable stationary wavelet packet convolution network | |
CN110110814B (en) | Distributed parallel PCA process monitoring modeling method based on continuous MapReduce | |
Zhou et al. | Rudder Health Monitoring and Data Visualization Based on Feature Extraction |
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 |