CN110807257A - Method for predicting residual life of aircraft engine - Google Patents

Method for predicting residual life of aircraft engine Download PDF

Info

Publication number
CN110807257A
CN110807257A CN201911064237.5A CN201911064237A CN110807257A CN 110807257 A CN110807257 A CN 110807257A CN 201911064237 A CN201911064237 A CN 201911064237A CN 110807257 A CN110807257 A CN 110807257A
Authority
CN
China
Prior art keywords
aircraft engine
variables
monitoring
residual life
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
Application number
CN201911064237.5A
Other languages
Chinese (zh)
Inventor
牛彩云
贾祥
柳冬林
陈英武
郭波
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National University of Defense Technology
Original Assignee
National University of Defense Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by National University of Defense Technology filed Critical National University of Defense Technology
Priority to CN201911064237.5A priority Critical patent/CN110807257A/en
Publication of CN110807257A publication Critical patent/CN110807257A/en
Priority to CN202011206202.3A priority patent/CN112131760B/en
Priority to CN202011205335.9A priority patent/CN112100865B/en
Priority to CN202011205350.3A priority patent/CN112149316B/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • 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
    • 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

Abstract

The invention provides a method for predicting the remaining life of an aircraft engine, which comprises the steps of firstly obtaining historical aircraft engine failure data, carrying out feature selection on aircraft engine monitoring variables, carrying out standardization processing on the monitoring variables selected by the features, then carrying out first-order difference operation to generate new variables, and obtaining a three-dimensional matrix form of an engine performance degradation data set. And constructing a bidirectional LSTM network with an attention mechanism, constructing input and output of a sample according to a mapping relation between a monitoring variable and the residual life, and training the bidirectional LSTM network with the attention mechanism to obtain a trained residual life prediction model of the aircraft engine. And constructing the input of a test sample for the monitoring data of the in-service aircraft engine to be subjected to residual life prediction, forming a test set and inputting the test set into the model for predicting the residual life of the aircraft engine to obtain the predicted value of the residual life of the in-service aircraft engine. The method provided by the invention is simple and effective in calculation process and high in prediction precision.

Description

Method for predicting residual life of aircraft engine
Technical Field
The invention relates to a prediction method of the residual life of an aircraft engine, in particular to a prediction method of the residual life of an aircraft engine of a bidirectional LSTM (Long Short-Term Memory) model with an attention mechanism.
Background
The aeroengine as a high-precision technical device needs to have the capability of adapting to extreme environments such as high temperature, high pressure, extreme cold and the like besides working in the conventional environment of sea level. Therefore, the degradation process of the engine is not solved by a simple single variable or single fault prediction model due to the complex and variable working environment and the influence of self-coupling action. In order to greatly improve the reliability and the safety of the aircraft engine and reduce the operation cost, the mass data generated by the sensor are mainly analyzed in real time or at regular time by a big data means, the working states of the engine and all parts are supervised and detected, and the problems are efficiently found or prevented, so that maintenance measures are made in advance to avoid catastrophic accidents, the economic benefit of engine operation is maximized, and the effect of safe operation is achieved while the maintenance cost is reduced.
In view of the advances in modern technology capabilities, having an integrated health management and diagnostic strategy becomes an important component of the system operational lifecycle. It is well known that complexity and noisy operating conditions make it difficult to build physical models from a priori knowledge, and most physical models cannot be updated with online measurement data, limiting their effectiveness and flexibility.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a method for predicting the residual life of an aircraft engine.
In order to realize the purpose of the invention, the following technical scheme is adopted for realizing the purpose:
a method for predicting the residual life of an aircraft engine comprises the following steps:
acquiring historical aeroengine failure data,form a training set XN×V×LWherein N represents the number of aero-engine samples, V represents the number of monitoring variables collected by each sensor in the aero-engine, and LnAnd (3) representing the monitoring track length of the nth aircraft engine sample N in the N aircraft engine samples.
Performing feature selection on the V monitoring variables to obtain F monitoring variables; standardizing the F monitoring variables, and performing first-order difference operation on the standardized variables to generate new variables to obtain a three-dimensional matrix form X of the engine performance degradation data setN×2F×L。
Constructing a bidirectional LSTM network with attention mechanism, and connecting XN×2F×LAnd constructing input and output of the sample according to the mapping relation between the monitoring variable X and the residual life RUL, and using the input and output to train a bidirectional LSTM network with an attention mechanism to obtain a trained prediction model of the residual life of the aircraft engine.
Establishing input of a test sample for monitoring data of an in-service aircraft engine to be subjected to residual life prediction to form a test set; and inputting the constructed test set into a trained prediction model of the residual life of the aircraft engine to obtain a predicted value of the residual life of the in-service aircraft engine.
Further, in the present invention, the feature selection on the V monitoring variables means that a constant variable in the V monitoring variables is deleted, and a trend variable is retained to obtain F monitoring variables.
Further, in the present invention, the normalization of the F monitoring variables means: the selected F monitoring variables are normalized to the range of [ -1,1] according to a 'min-max' normalization method.
Further, in the present invention, the input and output method for constructing the sample is: to XN×2F×LThe input of the sample is constructed by adopting a window sliding method, and the corresponding output label, namely the residual service life RUL, is corrected according to a step linear function, and finally the input and the output of the paired samples are obtained. The window sliding method is described as follows:
for XN×2F×LData of the middle engine sample n
Figure BDA0002258826160000031
The following form of samples is obtained for step d ═ 1:
wherein the content of the first and second substances,
Figure BDA0002258826160000033
wherein the order linear function expression is as follows:
Figure BDA0002258826160000034
wherein Label represents a Label for constructing sample data, RUL represents the actual residual life in the acquired historical aeroengine failure data, and RearlyIndicates a threshold value set according to the situation. Which may be taken as a default value of 125 as in the present invention.
Further, in the present invention, the mapping relationship between the monitoring variable X and the remaining lifetime RUL is expressed as follows:
f:X→RUL i.e.,RUL(t)＝f(Xt-s+1,Xt-s+2,…,Xt)
wherein t represents time, s represents time step, XiAnd i is t-s +1, …, and t represents the monitoring data corresponding to the time i and is in a vector form with the length of 2F.
Further, the invention also comprises a method for evaluating the prediction result of the test set, wherein the prediction result of the test set is evaluated by adopting two indexes of a mean square error root RMSE and a Score function Score:
Figure BDA0002258826160000041
Figure BDA0002258826160000042
wherein n represents the number of test samples in the test set, di＝RUL′i-RULiAnd the error between the predicted value and the true value of the ith test sample in the test set is represented.
In another aspect, the present invention further provides a system for predicting a remaining life of an aircraft engine, including:
the data acquisition module is used for acquiring historical aeroengine failure data to form a training set;
the characteristic selection module is used for selecting characteristics of the monitoring variables of the aero-engine, deleting constant variables and keeping trend variables;
the standardization processing module is used for standardizing the trend variables output by the characteristic selection module;
the first-order difference module is used for performing first-order difference operation on the standardized variables to generate new variables so as to obtain a three-dimensional matrix form of the engine performance degradation data set;
the training module is used for constructing a bidirectional LSTM network with an attention mechanism, constructing input and output of a sample by using an engine performance degradation data set in a three-dimensional matrix form output by the first-order difference module according to a mapping relation between a monitoring variable and the residual life, and training the bidirectional LSTM network with the attention mechanism to obtain a trained prediction model of the residual life of the aero-engine;
the prediction module is used for constructing the input of a test sample for the monitoring data of the in-service aircraft engine to be subjected to residual life prediction to form a test set; and inputting the constructed test set into a trained prediction model of the residual life of the aircraft engine to obtain a predicted value of the residual life of the in-service aircraft engine.
Further, the prediction system for the residual life of the aircraft engine further comprises a prediction result evaluation module, and the prediction result of the test set is evaluated by adopting two indexes, namely a mean square error Root (RMSE) and a Score function Score.
In another aspect, the present invention further provides an intelligent device, which includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of any one of the methods for predicting the remaining life of an aircraft engine when executing the computer program.
In another aspect, the present invention further provides a readable storage medium, on which a computer program is stored, the computer program, when being executed by a processor, implementing the steps of any one of the methods for predicting the remaining life of an aircraft engine.
The invention has the advantages that:
taking the residual service life RUL of the aircraft engine as a prediction target, firstly, establishing a mapping relation between a characteristic variable for representing a system degradation process and the prediction target; secondly, in order to learn a model capable of better describing a system degradation process, based on the angle of system degradation characteristic information extraction, an LSTM (Long Short-Term Memory) model is utilized to have the capability of learning a sequence data correlation relationship, an enhanced version of the bidirectional LSTM simultaneously utilizes past and future dependency information to acquire more information of a time sequence relationship, the introduction of an attention mechanism can help to highlight strong degradation information and weaken the influence of redundant or useless information, and the bidirectional LSTM model with the attention mechanism is provided to learn the mapping relationship.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments or the prior art of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for the ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of a method for predicting the remaining life of an aircraft engine according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a bi-directional LSTM network with attention mechanism;
FIG. 3 is a logical relationship diagram of an aircraft engine architecture and simulation module according to an embodiment of the present invention;
FIG. 4 is a scatter plot of 21 sensor signals from an aircraft engine in accordance with an embodiment of the present invention;
the diagram shows a comparison of the sequenced predicted results of 5100 test engines;
FIG. 6 tests engine #24 remaining life prediction results;
FIG. 7 test Engine #34 remaining Life prediction results;
FIG. 8 test Engine #76 remaining Life prediction results;
fig. 9 tests engine #100 remaining life prediction results.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments of the present invention, belong to the protection scope of the present invention.
Referring to fig. 1, fig. 1 is a flowchart of a method for predicting the remaining life of an aircraft engine according to an embodiment of the present invention, including the following steps:
(1) obtaining historical aircraft engine failure data, and carrying out sample reconstruction on the historical aircraft engine failure data to form a training set XN×V×L. Wherein N represents the number of aero-engine samples, V represents the number of monitoring variables collected by each sensor in the aero-engine, and LnAnd (3) representing the monitoring track length of the nth aircraft engine sample N in the N aircraft engine samples.
(2) And selecting the characteristics of the V monitoring variables, deleting constant variables in the V monitoring variables, keeping trend variables, and reducing the number of the monitoring variables in the engine sample from V to F to obtain F monitoring variables.
(3) Normalizing the selected F monitoring variables to the range of [ -1,1] according to a 'minimum-maximum' normalization method, wherein the calculation formula is as follows:
wherein x isi(n),j∈XN×F×LThe i (n) th raw data representing the sample engine signal j,
Figure BDA0002258826160000072
is xi(n),jNormalized value, and
Figure BDA0002258826160000073
and
Figure BDA0002258826160000074
representing the maximum and minimum values of signal j, respectively.
Then, the normalized variables are subjected to first-order difference operation to generate new variables, and finally, a three-dimensional matrix form X of the engine performance degradation data set is obtainedN×2F×L。
The calculation formula of the k-order difference operation is as follows:
Figure BDA0002258826160000081
in the invention, k is defaulted to be 1, namely, the first-order difference operation is carried out. The new variable generated by the first order difference operation can be understood to characterize the acceleration of the degeneration process.
(4) Constructing a network architecture of a bidirectional LSTM model with an attention mechanism; mixing XN×2F×LAnd constructing input and output of the sample according to the mapping relation between the monitoring variable X and the residual life RUL, and using the input and output to train a bidirectional LSTM network with an attention mechanism to obtain a trained prediction model of the residual life of the aircraft engine. The network architecture of the bidirectional LSTM model with attention mechanism employed by the present invention is shown in fig. 2. The proposed model has the benefits that: because the time sequence monitoring data has strong time dependence, the bidirectional LSTM can process sequence data through a forward hidden layer and a backward hidden layer respectively and then feed forward to the same output layer; the introduction of attention mechanism can play a role of 'highlighting strong degradation information and weakening redundant or useless information'.
Specifically, the activation mechanism of the bidirectional LSTM hidden layer function is as follows:
for the input data, the bidirectional LSTM network selectively masks some redundant information by utilizing a forgetting gate through forward calculation and backward calculation respectively, and utilizes an input gate and an input node to select information needing to be updated and record the new information. The basic formulas for training the predictive model based on the forward → and reverse ← time of bidirectional LSTM are as follows:
Figure BDA0002258826160000091
Figure BDA0002258826160000092
Figure BDA0002258826160000093
Figure BDA0002258826160000094
Figure BDA0002258826160000095
Figure BDA0002258826160000096
Figure BDA0002258826160000097
Figure BDA0002258826160000098
Figure BDA0002258826160000099
Figure BDA00022588261600000910
wherein the content of the first and second substances,
Figure BDA00022588261600000911
indicating the entry gate (what information needs to be updated),
Figure BDA00022588261600000912
indicating that the door was forgotten (what information was discarded),
Figure BDA00022588261600000913
indicating the output gate (determining what information to output),
Figure BDA00022588261600000914
Figure BDA00022588261600000915
an output representing a status update (history information accumulation),
Figure BDA00022588261600000916
the information representing the output is displayed on the display,
Figure BDA00022588261600000917
Figure BDA00022588261600000918
and
Figure BDA00022588261600000919
as a weight matrix,
Figure BDA00022588261600000920
Figure BDA00022588261600000921
and
Figure BDA00022588261600000922
is a bias vector.
Hidden unit h for bidirectional LSTM final outputtExpressed as the connected vectors output by the forward and backward processes, as follows:
Figure BDA00022588261600000923
as shown in FIG. 2, the nature of the attention mechanism is to compute weights, so the bidirectional LSTM with attention mechanism is actually a correction operation for hidden states, which can be expressed as a convex sum as follows:
ht＝(1-at)·ht-1+at·h′t
wherein h ist-1Indicates a hidden state of the previous time, h'tIs a correction based on attention mechanism made at the current time to the previous time hidden state, atRepresents the weight coefficient:
h′t＝g(W·ht-1+U·xt+b)
where W and U are linear transfer functions for the previous and current time instants, respectively, and b represents an offset, typically the activation function g uses a rectifying linear unit (ReLU). Performing three-dimensional matrix form engine performance degradation data set X obtained in the step (3)N×2F×LThe input and output of the sample are constructed according to the mapping relationship between the monitoring variable X and the remaining lifetime RUL.
Wherein the mapping relationship between the monitoring variable X and the remaining lifetime RUL is expressed as follows:
f:X→RUL i.e.,RUL(t)＝f(Xt-s+1,Xt-s+2,…,Xt)
wherein t represents time, s represents time step, XiAnd i is t-s +1, …, and t represents the monitoring data corresponding to the time i and is in a vector form with the length of 2F.
To XN×2F×LThe data in the step (1) is input by adopting a window sliding method to construct a sample, and a label (RUL) corresponding to the output of the sample is corrected according to a stepped linear function, so that input and output of paired sample data are finally obtained and are used for training a bidirectional LSTM network with an attention mechanism.
The window sliding method is described as follows:
for XN×2F×LOf medium engine sample nData of
Figure BDA0002258826160000101
The following form of samples is obtained for step d ═ 1:
Figure BDA0002258826160000102
wherein the content of the first and second substances,
Figure BDA0002258826160000111
wherein the order linear function expression is as follows:
Figure BDA0002258826160000112
wherein Label represents a Label for constructing sample data, RUL represents the actual residual life in the acquired historical aeroengine failure data, and RearlyIndicates a threshold value set according to the situation. Which may be taken as a default value of 125 as in the present invention.
(5) And for the monitoring data of the in-service aircraft engine to be subjected to residual life prediction, establishing the input of the test sample by adopting a window sliding method to form a test set. And inputting the constructed test set into a trained prediction model of the residual life of the aircraft engine to obtain a predicted value of the residual life of the in-service aircraft engine.
Further, the method is used for evaluating the prediction result of the method provided by the invention. The prediction result of the method provided by the invention is evaluated by adopting two indexes of a mean square error Root (RMSE) and a Score function (Score), and the expressions of the two indexes of the mean square error Root (RMSE) and the Score function (Score) are respectively:
Figure BDA0002258826160000113
wherein n represents the number of test samples in the test set, di＝RUL′i-RULiAnd the error between the predicted value and the true value of the ith test sample in the test set is represented.
The following describes the implementation and prediction effect of the present invention with reference to a specific application example:
in this implementation, NASA is used to provide a CMAPSS simulation data set for an aircraft engine. CMAPSS is a modularized aviation propulsion system simulation software developed by Green research center of NASA in the United states, and aims to simulate the whole degradation process of an airplane from normal to fault and provide a data base for a prediction model. Simulation experiments were created under the MatlabSimulink tool, simulating an engine model with 90000 pounds of thrust, and the program included an atmospheric model and an electrical management system involving five component modules for a fan, a Low Pressure Compressor (LPC), a High Pressure Compressor (HPC), a high pressure turbine (HPC), and a Low Pressure Turbine (LPT). The logical structural relationship of five modules in an aircraft engine simulation experiment is shown in fig. 3.
The open source data comprises four groups of simulation data in total, the specific implementation process of the invention selects 'train _ FD 001' and 'test _ FD 001' as a training set and a test set respectively, wherein each subdata set comprises 26 columns, namely, a number, an operation period, an environment setting 1, an environment setting 2, an environment setting 3 and 21 monitoring indexes, 21 monitoring data are used for outputting signal data in the engine degradation process in the simulation experiment, and the specific meaning represented by the data is described as shown in Table 1.
TABLE 1 Engine monitoring index description
Figure BDA0002258826160000121
Figure BDA0002258826160000131
Figure BDA0002258826160000141
The specific process of using the method of the invention to predict the service life is as follows:
step 1, multivariate degradation characteristic screening: first, raw monitoring data from 21 sensors in the data set train _ FD001 is visualized as shown in fig. 4. The trend of the different monitored variables over the life cycle can be found. It can be readily seen that sensor data can be broadly divided into two categories: constant and changing (i.e., increasing or decreasing trend). Thus, the constant signal is clearly not functional in characterizing the engine degradation process and is therefore not considered on the input variables of the later model. In addition, Sensor6 shows a binary variable that is also considered to not contribute to characterizing engine degradation. Thus, the remaining 14 sensor data are used as raw input features for the RUL prediction model, numbered 2, 3, 4, 7, 8, 9, 11, 12, 13, 14, 15, 17, 20 and 21, respectively.
Step 2, data standardization and difference technology: the 14 screened features were normalized according to the "max-min" normalization method. In addition, the 14 features are subjected to difference processing to generate new features.
Step 3, sample construction and label resetting: constructing the size of N for each engine unit in the training set according to a time window methodt×NfIs input of samples of (1), wherein Nt＝30,Nf28. When constructing the sample output, the real residual life is reset by adopting a step linear function method, and the assumption is that the sample output has a constant RUL value R in the initial stageearly＝125。
Step 4, network construction and training: the specific architecture of the bidirectional LSTM network with attention mechanism is as follows: InputLayer (30 × 28) → Dropout Layer (0.5) → Bidirectional LSTM Layer (28) → orientation → Dropout Layer (0.5) → Bidirectional LSTM Layer (14) → orientation → full connection Layer (280) → Output Layer (1). And inputting the samples constructed by the training set into the network architecture, and obtaining a trained prediction model of the residual life of the aircraft engine by taking the training round epoch as 30.
And 5, constructing a sample input of the network for the data in the test set according to the same method in the step 3. And inputting the samples of the test set into the trained prediction model of the residual life of the aircraft engine to obtain a prediction result, as shown in fig. 5.
The prediction performance index values of the prediction model of the residual life of the aircraft engine are respectively as follows: RMSE 12.58 and Score 240.21. In order to further observe the prediction effect of the model, samples constructed by four test engines with the numbers of #24, #34, #76 and #100 are respectively input into the trained prediction model of the residual life of the aircraft engine to obtain comparison effect graphs of a predicted value and a true value, which are respectively shown in fig. 6, fig. 7, fig. 8 and fig. 9, wherein fig. 6 is a prediction result graph of the residual life of the test engine # 24; FIG. 7 is a graph of the remaining life prediction results for test engine # 34; FIG. 8 is a graph of the remaining life prediction results for test engine # 76; fig. 9 is a remaining life prediction result map of the test engine # 100.
In conclusion, the invention provides a method for predicting the residual life of an aeroengine of a bidirectional LSTM model with an attention mechanism, which establishes a mapping relation between a characteristic variable for representing a system degradation process and a prediction target (RUL); according to the characteristic that the LSTM model has the capability of learning the sequence data correlation relationship, and the enhanced version of the bidirectional LSTM simultaneously utilizes the past and future dependency information to acquire more information of the time sequence relationship, and the bidirectional LSTM model with the attention mechanism is provided for learning the proposed mapping relationship by utilizing the attention mechanism to play a role of 'highlighting strong degradation information and weakening redundant or useless information'. Preprocessing original monitoring data, constructing a sample for inputting the proposed model, inputting the constructed sample into a set model, and training to obtain a final model; and finally, inputting the test sample into the trained model to obtain a final result and a predicted performance index. The method well solves the problem of predicting the residual service life of the data-driven aircraft engine through the steps, and is beneficial to building the bridge connected with big data and intelligent system health management. Compared with the existing method, the algorithm provided by the invention is simple and effective in calculation process. According to the specific embodiment, the algorithm provided by the invention has high prediction precision.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. A method for predicting the residual life of an aircraft engine is characterized by comprising the following steps:
obtaining historical aeroengine failure data to form a training set XN×V×LWherein N represents the number of aero-engine samples, V represents the number of monitoring variables collected by each sensor in the aero-engine, and LnRepresenting the monitoring track length of the nth aeroengine sample N in the N aeroengine samples;
performing feature selection on the V monitoring variables to obtain F monitoring variables; standardizing the F monitoring variables, and performing first-order difference operation on the standardized variables to generate new variables to obtain a three-dimensional matrix form X of the engine performance degradation data setN×2F×L；
Constructing a bidirectional LSTM network with attention mechanism, and connecting XN×2F×LConstructing input and output of a sample according to a mapping relation between a monitoring variable X and the residual life RUL, and training a bidirectional LSTM network with an attention mechanism to obtain a trained prediction model of the residual life of the aircraft engine;
establishing input of a test sample for monitoring data of an in-service aircraft engine to be subjected to residual life prediction to form a test set; and inputting the constructed test set into a trained prediction model of the residual life of the aircraft engine to obtain a predicted value of the residual life of the in-service aircraft engine.
2. The method for predicting the remaining life of the aircraft engine according to claim 1, wherein the characteristic selection of the V monitoring variables is to delete a constant variable from the V monitoring variables and retain a trend variable to obtain F monitoring variables.
3. The method for predicting the remaining life of an aircraft engine according to claim 1, wherein the normalization process of the F monitored variables is performed by: the selected F monitoring variables are normalized to the range of [ -1,1] according to a 'min-max' normalization method.
4. The method of predicting the remaining life of an aircraft engine according to claim 1, wherein the input and output methods of the construction samples are: to XN×2F×LThe input of the sample is constructed by adopting a window sliding method, and the corresponding output label, namely the residual service life RUL, is corrected according to a step linear function, and finally the input and the output of the paired samples are obtained.
5. The method of predicting the remaining life of an aircraft engine according to claim 2, wherein the mapping between the monitoring variable X and the remaining life RUL is represented as follows:
f:X→RULi.e.,RUL(t)＝f(Xt-s+1,Xt-s+2,…,Xt)
wherein t represents time, s represents time step, XiAnd i is t-s +1, …, and t represents the monitoring data corresponding to the time i and is in a vector form with the length of 2F.
6. The method for predicting the residual life of an aircraft engine according to claim 3, further comprising a method for evaluating the prediction results of the test set, wherein the prediction results of the test set are evaluated by two indexes, namely a mean square error Root (RMSE) and a Score function (Score):
Figure FDA0002258826150000021
Figure FDA0002258826150000022
wherein n represents the number of test samples in the test set, di＝RULi′-RULiAnd the error between the predicted value and the true value of the ith test sample in the test set is represented.
7. An aircraft engine remaining life prediction system, comprising:
the data acquisition module is used for acquiring historical aeroengine failure data to form a training set;
the characteristic selection module is used for selecting characteristics of the monitoring variables of the aero-engine, deleting constant variables and keeping trend variables;
the standardization processing module is used for standardizing the trend variables output by the characteristic selection module;
the first-order difference module is used for performing first-order difference operation on the standardized variables to generate new variables so as to obtain a three-dimensional matrix form of the engine performance degradation data set;
the training module is used for constructing a bidirectional LSTM network with an attention mechanism, constructing input and output of a sample by using an engine performance degradation data set in a three-dimensional matrix form output by the first-order difference module according to a mapping relation between a monitoring variable and the residual life, and training the bidirectional LSTM network with the attention mechanism to obtain a trained prediction model of the residual life of the aero-engine;
the prediction module is used for constructing the input of a test sample for the monitoring data of the in-service aircraft engine to be subjected to residual life prediction to form a test set; and inputting the constructed test set into a trained prediction model of the residual life of the aircraft engine to obtain a predicted value of the residual life of the in-service aircraft engine.
8. The system of claim 7, further comprising a prediction evaluation module for evaluating the prediction results of the test set using two criteria, Root of Mean Square Error (RMSE) and Score function (Score).
9. An intelligent device comprising a memory and a processor, the memory storing a computer program, wherein the processor when executing the computer program implements the steps of the method for predicting the remaining life of an aircraft engine as defined in any one of claims 1 to 6.
10. A readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method for predicting the remaining life of an aircraft engine according to any one of claims 1 to 6.
CN201911064237.5A 2019-11-04 2019-11-04 Method for predicting residual life of aircraft engine Pending CN110807257A (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
CN201911064237.5A CN110807257A (en) 2019-11-04 2019-11-04 Method for predicting residual life of aircraft engine
CN202011206202.3A CN112131760B (en) 2019-11-04 2020-11-02 CBAM model-based prediction method for residual life of aircraft engine
CN202011205335.9A CN112100865B (en) 2019-11-04 2020-11-02 Method for predicting remaining life of aircraft engine based on parallel CNN model
CN202011205350.3A CN112149316B (en) 2019-11-04 2020-11-02 Aero-engine residual life prediction method based on improved CNN model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911064237.5A CN110807257A (en) 2019-11-04 2019-11-04 Method for predicting residual life of aircraft engine

Publications (1)

Publication Number Publication Date
CN110807257A true CN110807257A (en) 2020-02-18

Family

ID=69500993

Family Applications (4)

Application Number Title Priority Date Filing Date
CN201911064237.5A Pending CN110807257A (en) 2019-11-04 2019-11-04 Method for predicting residual life of aircraft engine
CN202011205335.9A Active CN112100865B (en) 2019-11-04 2020-11-02 Method for predicting remaining life of aircraft engine based on parallel CNN model
CN202011205350.3A Active CN112149316B (en) 2019-11-04 2020-11-02 Aero-engine residual life prediction method based on improved CNN model
CN202011206202.3A Active CN112131760B (en) 2019-11-04 2020-11-02 CBAM model-based prediction method for residual life of aircraft engine

Family Applications After (3)

Application Number Title Priority Date Filing Date
CN202011205335.9A Active CN112100865B (en) 2019-11-04 2020-11-02 Method for predicting remaining life of aircraft engine based on parallel CNN model
CN202011205350.3A Active CN112149316B (en) 2019-11-04 2020-11-02 Aero-engine residual life prediction method based on improved CNN model
CN202011206202.3A Active CN112131760B (en) 2019-11-04 2020-11-02 CBAM model-based prediction method for residual life of aircraft engine

Country Status (1)

Country Link
CN (4) CN110807257A (en)

Cited By (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111361759A (en) * 2020-03-02 2020-07-03 哈尔滨工业大学 Airplane auxiliary power device on-wing residual life prediction method based on hybrid model
CN111460728A (en) * 2020-03-09 2020-07-28 华南理工大学 Method and device for predicting residual life of industrial equipment, storage medium and equipment
CN111507046A (en) * 2020-04-16 2020-08-07 哈尔滨工程大学 Method and system for predicting remaining service life of electric gate valve
CN111597663A (en) * 2020-07-27 2020-08-28 中国人民解放军国防科技大学 Momentum wheel residual life prediction method fusing residual life empirical data
CN111639467A (en) * 2020-06-08 2020-09-08 长安大学 Aero-engine service life prediction method based on long-term and short-term memory network
CN111695521A (en) * 2020-06-15 2020-09-22 哈尔滨理工大学 Attention-LSTM-based rolling bearing performance degradation prediction method
CN112115643A (en) * 2020-09-15 2020-12-22 中南大学 Smart train service life non-invasive prediction method
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
CN112560252A (en) * 2020-12-07 2021-03-26 厦门大学 Prediction method for residual life of aircraft engine
CN112613227A (en) * 2020-12-15 2021-04-06 大连理工大学 Model for predicting remaining service life of aero-engine based on hybrid machine learning
CN112613226A (en) * 2020-12-10 2021-04-06 大连理工大学 Feature enhancement method for residual life prediction
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
CN113139278A (en) * 2021-04-08 2021-07-20 上海工程技术大学 Deep learning-based bearing residual life prediction method and system
CN113158445A (en) * 2021-04-06 2021-07-23 中国人民解放军战略支援部队航天工程大学 Prediction algorithm for residual service life of aero-engine with convolution memory residual self-attention mechanism
CN113297787A (en) * 2021-05-12 2021-08-24 北京航空航天大学 Method for predicting remaining life of aircraft engine based on transfer learning
CN114266278A (en) * 2021-12-29 2022-04-01 合肥工业大学 Dual-attention-network-based method for predicting residual service life of equipment
CN114510870A (en) * 2022-01-07 2022-05-17 华东交通大学 Method and device for predicting residual life of underground structure of urban rail transit
CN114707257A (en) * 2022-02-18 2022-07-05 江苏赛德力制药机械制造有限公司 Mechanical residual service life prediction method based on all-state attention and BiLSTM
CN114880767A (en) * 2022-04-24 2022-08-09 西北工业大学 Aero-engine remaining service life prediction method based on attention mechanism Dense-GRU network
JP2023501030A (en) * 2020-09-30 2023-01-18 西南石油大学 Engine surge failure prediction system and method based on fusion neural network model
CN116502516A (en) * 2023-01-15 2023-07-28 北京控制工程研究所 Identification method and device for degradation stage of spacecraft component
WO2023160300A1 (en) * 2022-02-22 2023-08-31 上海三一重机股份有限公司 Structural part remaining service life prediction method and device, and operation machine
WO2024045377A1 (en) * 2022-09-02 2024-03-07 同济大学 Deep-learning-based method for predicting residual life of aero-engine

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112881518B (en) * 2021-01-08 2022-07-12 东冶及策河北能源技术有限公司 Method for predicting residual life of dynamic filter compensator
CN113344281B (en) * 2021-06-22 2022-02-11 太仓比泰科自动化设备有限公司 Probe performance prediction method, system, device and storage medium
CN113722989B (en) * 2021-08-23 2023-04-28 南京航空航天大学 CPS-DP model-based aeroengine service life prediction method
CN113602526B (en) * 2021-08-27 2023-05-23 中国航空工业集团公司上海航空测控技术研究所 Verification test method and system for electromechanical fault prediction and health management system of airplane
CN115982942B (en) * 2022-11-25 2023-07-14 浙江长龙航空有限公司 Method, device, equipment and storage medium for predicting residual life of aviation precooler
CN116561528B (en) * 2023-05-08 2024-03-01 重庆市荣冠科技有限公司 RUL prediction method of rotary machine
CN117556261A (en) * 2024-01-08 2024-02-13 浙江大学 MCNN-based diaphragm pump check valve service life prediction method and system

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170116552A1 (en) * 2010-06-04 2017-04-27 Sapience Analytics Private Limited System and Method to Measure, Aggregate and Analyze Exact Effort and Time Productivity
CN103336908B (en) * 2013-07-23 2016-01-20 哈尔滨工业大学 The method for predicting residual useful life of dull echo state network
CN106169001B (en) * 2016-07-11 2019-06-21 北京航空航天大学 A kind of aero-engine complete machine method for predicting residual useful life based on gas circuit performance parameter Graphic Pattern Matching
CN107153759B (en) * 2017-04-27 2019-10-15 西安交通大学 A kind of aviation turbofan engine method for predicting residual useful life of multisource data fusion
CN108959778B (en) * 2018-07-06 2020-09-15 南京航空航天大学 Method for predicting residual life of aircraft engine based on consistency of degradation modes
CN109272114A (en) * 2018-08-08 2019-01-25 西安交通大学 The deep learning method that mechanized equipment remaining life is quickly and effectively predicted
CN109460574A (en) * 2018-10-09 2019-03-12 南京航空航天大学 A kind of prediction technique of aero-engine remaining life
CN109472110B (en) * 2018-11-29 2023-06-27 南京航空航天大学 Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model
CN109766583B (en) * 2018-12-14 2023-06-20 南京航空航天大学 Aircraft engine life prediction method based on unlabeled, unbalanced and initial value uncertain data
CN109783906B (en) * 2018-12-29 2023-07-07 东北大学 Intelligent analysis system and method for detecting magnetic flux leakage data in pipeline
CN109829136B (en) * 2019-01-22 2022-10-04 中国人民解放军火箭军工程大学 Method and system for predicting residual life of degradation equipment with random jump
CN109992872B (en) * 2019-03-27 2020-07-28 西安交通大学 Mechanical equipment residual life prediction method based on stacked separation convolution module

Cited By (38)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111361759A (en) * 2020-03-02 2020-07-03 哈尔滨工业大学 Airplane auxiliary power device on-wing residual life prediction method based on hybrid model
CN111361759B (en) * 2020-03-02 2023-02-03 哈尔滨工业大学 Airplane auxiliary power device on-wing residual life prediction method based on hybrid model
CN111460728A (en) * 2020-03-09 2020-07-28 华南理工大学 Method and device for predicting residual life of industrial equipment, storage medium and equipment
CN111460728B (en) * 2020-03-09 2022-08-12 华南理工大学 Method and device for predicting residual life of industrial equipment, storage medium and equipment
CN111507046A (en) * 2020-04-16 2020-08-07 哈尔滨工程大学 Method and system for predicting remaining service life of electric gate valve
CN111507046B (en) * 2020-04-16 2022-09-06 哈尔滨工程大学 Method and system for predicting remaining service life of electric gate valve
CN111639467A (en) * 2020-06-08 2020-09-08 长安大学 Aero-engine service life prediction method based on long-term and short-term memory network
CN111639467B (en) * 2020-06-08 2024-04-16 长安大学 Aero-engine service life prediction method based on long-term and short-term memory network
CN111695521A (en) * 2020-06-15 2020-09-22 哈尔滨理工大学 Attention-LSTM-based rolling bearing performance degradation prediction method
CN111695521B (en) * 2020-06-15 2022-06-10 哈尔滨理工大学 Attention-LSTM-based rolling bearing performance degradation prediction method
CN111597663A (en) * 2020-07-27 2020-08-28 中国人民解放军国防科技大学 Momentum wheel residual life prediction method fusing residual life empirical data
CN112115643B (en) * 2020-09-15 2022-06-10 中南大学 Smart train service life non-invasive prediction method
CN112115643A (en) * 2020-09-15 2020-12-22 中南大学 Smart train service life non-invasive prediction method
JP7242101B2 (en) 2020-09-30 2023-03-20 西南石油大学 Engine surge failure prediction system and method based on fusion neural network model
JP2023501030A (en) * 2020-09-30 2023-01-18 西南石油大学 Engine surge failure prediction system and method based on fusion neural network model
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
CN112560252B (en) * 2020-12-07 2023-09-15 厦门大学 Method for predicting residual life of aeroengine
CN112560252A (en) * 2020-12-07 2021-03-26 厦门大学 Prediction method for residual life of aircraft engine
CN112613226A (en) * 2020-12-10 2021-04-06 大连理工大学 Feature enhancement method for residual life prediction
CN112613226B (en) * 2020-12-10 2022-11-18 大连理工大学 Feature enhancement method for residual life prediction
CN112613227B (en) * 2020-12-15 2022-09-30 大连理工大学 Model for predicting remaining service life of aero-engine based on hybrid machine learning
CN112613227A (en) * 2020-12-15 2021-04-06 大连理工大学 Model for predicting remaining service life of aero-engine based on hybrid machine learning
CN113158445A (en) * 2021-04-06 2021-07-23 中国人民解放军战略支援部队航天工程大学 Prediction algorithm for residual service life of aero-engine with convolution memory residual self-attention mechanism
CN113139278A (en) * 2021-04-08 2021-07-20 上海工程技术大学 Deep learning-based bearing residual life prediction method and system
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
CN113297787B (en) * 2021-05-12 2021-12-17 北京航空航天大学 Method for predicting remaining life of aircraft engine based on transfer learning
CN113297787A (en) * 2021-05-12 2021-08-24 北京航空航天大学 Method for predicting remaining life of aircraft engine based on transfer learning
CN114266278A (en) * 2021-12-29 2022-04-01 合肥工业大学 Dual-attention-network-based method for predicting residual service life of equipment
CN114266278B (en) * 2021-12-29 2024-02-20 合肥工业大学 Dual-attention network-based equipment residual service life prediction method
CN114510870A (en) * 2022-01-07 2022-05-17 华东交通大学 Method and device for predicting residual life of underground structure of urban rail transit
CN114510870B (en) * 2022-01-07 2024-04-16 华东交通大学 Method and device for predicting residual life of underground structure of urban rail transit
CN114707257A (en) * 2022-02-18 2022-07-05 江苏赛德力制药机械制造有限公司 Mechanical residual service life prediction method based on all-state attention and BiLSTM
WO2023160300A1 (en) * 2022-02-22 2023-08-31 上海三一重机股份有限公司 Structural part remaining service life prediction method and device, and operation machine
CN114880767A (en) * 2022-04-24 2022-08-09 西北工业大学 Aero-engine remaining service life prediction method based on attention mechanism Dense-GRU network
CN114880767B (en) * 2022-04-24 2024-03-08 西北工业大学 Aero-engine residual service life prediction method based on attention mechanism Dense-GRU network
WO2024045377A1 (en) * 2022-09-02 2024-03-07 同济大学 Deep-learning-based method for predicting residual life of aero-engine
CN116502516A (en) * 2023-01-15 2023-07-28 北京控制工程研究所 Identification method and device for degradation stage of spacecraft component
CN116502516B (en) * 2023-01-15 2023-12-01 北京控制工程研究所 Identification method and device for degradation stage of spacecraft component

Also Published As

Publication number Publication date
CN112131760A (en) 2020-12-25
CN112149316B (en) 2022-04-19
CN112131760B (en) 2022-08-23
CN112100865B (en) 2022-08-23
CN112149316A (en) 2020-12-29
CN112100865A (en) 2020-12-18

Similar Documents

Publication Publication Date Title
CN110807257A (en) Method for predicting residual life of aircraft engine
CN109766583B (en) Aircraft engine life prediction method based on unlabeled, unbalanced and initial value uncertain data
CN109472110B (en) Method for predicting residual service life of aeroengine based on LSTM network and ARIMA model
Xu et al. PHM-oriented integrated fusion prognostics for aircraft engines based on sensor data
Elattar et al. Prognostics: a literature review
CN108960303B (en) Unmanned aerial vehicle flight data anomaly detection method based on LSTM
CN111274737A (en) Method and system for predicting remaining service life of mechanical equipment
Liu et al. Integration of data fusion methodology and degradation modeling process to improve prognostics
CN108304661B (en) Diagnosis and prediction method based on TDP model
CN108780315B (en) Method and apparatus for optimizing diagnostics of rotating equipment
Phillips et al. A knowledge driven approach to aerospace condition monitoring
JP2009053938A (en) Equipment diagnosing system and equipment-diagnosing method on the basis of multiple model
CN108427400B (en) Aircraft airspeed head fault diagnosis method based on neural network analytic redundancy
Si et al. An integrated reliability estimation approach with stochastic filtering and degradation modeling for phased-mission systems
CN114266278B (en) Dual-attention network-based equipment residual service life prediction method
CN110309537B (en) Intelligent health prediction method and system for aircraft
CN103324155A (en) System monitoring
Liao et al. Nonparametric and semi-parametric sensor recovery in multichannel condition monitoring systems
CN114281054A (en) LSTM-CNN-based airplane remote fault analysis method and system
Atlas et al. An evolvable tri-reasoner ivhm system
CN116662743A (en) Engine residual life prediction method based on multi-mode deep learning
CN112560252B (en) Method for predicting residual life of aeroengine
Chen et al. Aircraft complex system diagnosis based on design knowledge and real-Time monitoring information
Reed Development of a parametric-based indirect aircraft structural usage monitoring system using artificial neural networks
Peng et al. Remaining useful life prediction for aircraft engines based on grey model

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20200218