CN112733880B - Aircraft engine fault diagnosis method and system and electronic equipment - Google Patents
Aircraft engine fault diagnosis method and system and electronic equipment Download PDFInfo
- Publication number
- CN112733880B CN112733880B CN202011495593.5A CN202011495593A CN112733880B CN 112733880 B CN112733880 B CN 112733880B CN 202011495593 A CN202011495593 A CN 202011495593A CN 112733880 B CN112733880 B CN 112733880B
- Authority
- CN
- China
- Prior art keywords
- parameter
- sensitive
- parameters
- fault diagnosis
- correlation coefficient
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M15/00—Testing of engines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Medical Informatics (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Computational Linguistics (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to a method, a system and electronic equipment for diagnosing faults of an aero-engine, wherein parameters of the aero-engine to be diagnosed are analyzed twice, specifically, the parameters of the aero-engine to be diagnosed are analyzed by adopting at least three analysis methods, and every two sensitive parameters in a first sensitive parameter union set are analyzed by adopting at least three correlation coefficient algorithms; through the fusion of various methods, the association information between the parameters can be sufficiently mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the operation safety of the airplane is ensured, and the investment of maintenance cost can be reduced to a great extent.
Description
Technical Field
The invention relates to the technical field of aircraft engines, in particular to a method and a system for diagnosing faults of an aircraft engine and electronic equipment.
Background
Aircraft engines act as aircraft "hearts" and therefore their importance is self-evident. Aircraft engines have a very complex structure, and are often in very harsh operating conditions of high temperature, high pressure, high rotational speed, and the like, so that various unknown faults are easily generated. And as the core equipment of the airplane, the generated faults can cause immeasurable loss. Through carrying out fault diagnosis to aeroengine key parts, can be convenient get rid of the trouble that appears, can avoid like this because major failure that the trouble arouses and then produce the loss that is difficult to estimate, reduce cost of maintenance, ensure aeroengine's healthy steady operation.
The traditional aircraft engine fault diagnosis method mainly comprises a model-based method and a data-driven method, and specifically comprises the following steps:
1) the method for diagnosing the faults of the model-based engine needs to establish a complex aircraft engine model, depends on expert knowledge, and has the diagnosis result greatly depending on the accuracy of establishing a mathematical model, but still has the problems of low training efficiency, poor accuracy and the like.
2) Based on a data driving method, a mode of artificial feature extraction and classification model is mainly used, a large amount of priori knowledge needs to be combined, appropriate feature parameters are selected and designed from a large amount of monitoring data, the implementation cost is high, and the diagnosis precision depends on the accuracy and effectiveness of the artificially extracted features; when different key components of the aircraft engine are diagnosed, models need to be established or characteristics need to be extracted, the process is complex, and the efficiency is low.
Disclosure of Invention
The invention aims to solve the technical problem of the prior art and provides a method and a system for diagnosing faults of an aircraft engine and electronic equipment.
The technical scheme of the aircraft engine fault diagnosis method is as follows:
analyzing parameters of the aircraft engine to be diagnosed by adopting at least three analysis methods to obtain a first sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method, and collecting a union set of all the first sensitive parameter sets to obtain a first sensitive parameter union set;
analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms to obtain a first strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm, and taking intersection of all the first strongly correlated parameter pair sets to obtain a first strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-related parameter pair in the first strongly-related parameter pair intersection set from the first sensitive parameter union set to obtain a second sensitive parameter union set;
and carrying out fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result.
The aircraft engine fault diagnosis method has the beneficial effects that:
analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of multiple methods, the association information among the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the healthy and stable operation of the aero-engine is guaranteed, the operation safety of the aircraft is guaranteed, and the investment of maintenance cost can be reduced to a great extent, namely the maintenance and guarantee cost of the aero-engine is reduced.
On the basis of the scheme, the aircraft engine fault diagnosis method can be further improved as follows.
Further, still include:
acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until obtaining a second sensitive parameter set which corresponds to each analysis method of all fault modes;
a union set is acquired for all the second sensitive parameter sets to obtain a third sensitive parameter union set;
analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm;
taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the second strongly-relevant parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and training according to the plurality of fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
Further, still include:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
The beneficial effect of adopting the further scheme is that: random noise is added into an input layer of an automatic encoder model, and the robustness and the generalization capability of the deep classification network are improved.
Further, the at least three analysis methods at least comprise a single factor sensitivity analysis method, a principal component analysis method and a random forest algorithm;
the at least three correlation coefficient algorithms include at least a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm, and a kender correlation coefficient algorithm.
The technical scheme of the aircraft engine fault diagnosis system is as follows:
the system comprises a first analysis module, a second analysis module, a deletion module and a fault diagnosis module;
the first analysis module is used for analyzing parameters of the aircraft engine to be diagnosed by adopting at least three analysis methods to obtain a first sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method, and performing union on all the first sensitive parameter sets to obtain a first sensitive parameter union set;
the second analysis module is used for analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms to obtain a first strong correlation parameter pair set containing strong correlation parameter pairs corresponding to each correlation coefficient algorithm, and taking intersection of all the first strong correlation parameter pair sets to obtain a first strong correlation parameter pair intersection set;
the deleting module is used for deleting any sensitive parameter in each strongly-related parameter pair in the first strongly-related parameter pair intersection set from the first sensitive parameter union set to obtain a second sensitive parameter union set;
and the fault diagnosis module is used for carrying out fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result.
The aircraft engine fault diagnosis system has the following beneficial effects:
analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of various methods, the association information between the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the loss which is difficult to estimate and is generated by further serious faults is avoided, the healthy and stable operation of the aero-engine is guaranteed, the operation safety of the airplane is guaranteed, and the investment of maintenance cost can be reduced to a great extent, namely the maintenance and guarantee cost of the aero-engine is reduced.
On the basis of the scheme, the aircraft engine fault diagnosis system can be further improved as follows.
Further, the system further comprises a model training module, wherein the model training module is used for:
acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until obtaining a second sensitive parameter set which corresponds to each analysis method of all fault modes;
a union set is acquired for all the second sensitive parameter sets to obtain a third sensitive parameter union set;
analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm;
taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the second strongly-relevant parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and training according to the plurality of fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
Further, the system also comprises a deep classification network building module, wherein the deep classification network building module is used for:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
The beneficial effect of adopting the further scheme is that: random noise is added into an input layer of an automatic encoder model, and the robustness and the generalization capability of the deep classification network are improved.
Further, the at least three analysis methods at least comprise a single factor sensitivity analysis method, a principal component analysis method and a random forest algorithm;
the at least three correlation coefficient algorithms include at least a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm, and a kender correlation coefficient algorithm.
The technical scheme of the electronic equipment is as follows:
comprising a memory, a processor and a program stored on the memory and running on the processor, wherein the processor implements the steps of an aircraft engine fault diagnosis method as described in any one of the above when executing the program.
The electronic equipment has the following beneficial effects:
analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of various methods, the association information between the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the healthy and stable operation of the aero-engine is guaranteed, the operation safety of the aircraft is guaranteed, and the investment of maintenance cost can be reduced to a great extent, namely the maintenance and guarantee cost of the aero-engine is reduced.
Drawings
FIG. 1 is a schematic flow chart of a method for diagnosing faults of an aircraft engine according to an embodiment of the invention;
FIG. 2 is a schematic flow chart of analyzing all parameters corresponding to any failure mode by adopting a single-factor sensitivity analysis method, a principal component analysis method and a random forest algorithm;
FIG. 3 is a schematic flow chart of analyzing every two sensitive parameters in the third sensitive parameter union set by using a Pearson correlation coefficient algorithm, a Spierman correlation coefficient algorithm and a Kendell correlation coefficient algorithm respectively;
FIG. 4 is a schematic structural diagram of an aircraft engine fault diagnosis system according to an embodiment of the invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
As shown in fig. 1, an aircraft engine fault diagnosis method according to an embodiment of the present invention includes the following steps:
s1, analyzing parameters of the aero-engine to be diagnosed by adopting at least three analysis methods to obtain a first sensitive parameter set corresponding to each analysis method and containing sensitive parameters, and merging all the first sensitive parameter sets to obtain a first sensitive parameter merged set;
s2, analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms to obtain a first strong correlation parameter pair set containing strong correlation parameter pairs corresponding to each correlation coefficient algorithm, and taking intersection of all the first strong correlation parameter pair sets to obtain a first strong correlation parameter pair intersection set;
s3, deleting any sensitive parameter in each strong correlation parameter pair in the first strong correlation parameter pair intersection set from the first sensitive parameter union set to obtain a second sensitive parameter union set;
and S4, performing fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result.
Analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of a plurality of methods, the association information among the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the operation safety of the airplane is ensured, and the investment of maintenance cost can be reduced to a great extent.
Wherein the at least three analysis methods at least comprise a single factor sensitivity analysis method, a principal component analysis method and a random forest algorithm; the at least three correlation coefficient algorithms at least comprise a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm and a kendell correlation coefficient algorithm, and are explained by taking three analysis methods, namely a single-factor sensitivity analysis method, a principal component analysis method and a random forest algorithm, and three correlation coefficient algorithms, namely the pearson correlation coefficient algorithm, the spearman correlation coefficient algorithm and the kendell correlation coefficient algorithm, as examples:
s10, analyzing the parameters of the aero-engine to be diagnosed by adopting a single-factor sensitivity analysis method, a principal component analysis method and a random forest algorithm respectively to obtain a first sensitive parameter set containing sensitive parameters corresponding to each analysis method, and specifically:
the parameters of the aircraft engine to be diagnosed can be understood as: the method comprises the steps that parameters monitored on all parts of the aero-engine to be diagnosed or parameters of key parts of the aero-engine to be diagnosed are set manually according to experience, for example, parts which are easy to damage in the aero-engine are used as preset key parts, parts which are easy to cause airplane crash after being damaged in the aero-engine are used as preset key parts, and the like, for example, the preset key parts are turbines, journals, drums and the like of the aero-engine;
the method comprises the following steps that a plurality of sensors arranged on the aero-engine are used for detecting data of all parts of the aero-engine, so that parameters of a turbine, a shaft neck and a drum barrel can be directly inquired and obtained from the detected data, wherein the parameters comprise temperature, vibration coefficient and the like, and the temperature and the vibration coefficient of the turbine, the temperature and the vibration coefficient of the shaft neck and the temperature and the vibration coefficient of the drum barrel are obtained at the moment; then:
1) the process of analyzing the parameters of the aircraft engine to be diagnosed by taking the single-factor sensitivity analysis method as an example is as follows:
sensitivity excavation is carried out based on the parameter time domain signal variation amplitude as an analysis factor, and sensitive parameters are excavated from multi-dimensional parameters, wherein the multi-dimensional parameters can be understood as follows: the temperature and the vibration coefficient of the turbine, the temperature and the vibration coefficient of the shaft neck and the temperature and the vibration coefficient of the drum barrel are multidimensional, the engine monitoring parameters at the same acquisition time of the high-pressure rotor rotating speed, the low-pressure rotor rotating speed, the exhaust temperature after the turbine, the outlet pressure of the gas compressor, the outlet pressure after the turbine, the angle of the fan guide vane, the angle of the gas compressor guide vane, the diameter of the nozzle throat and the like are multidimensional, and whether the parameters are sensitive parameters is judged through a first preset condition, specifically:
the first preset condition is that the variation amplitude gradient is larger than a first preset threshold value, and when the variation amplitude of the mean value of the parameter is larger than a second preset threshold value, the parameter is judged to be a sensitive parameter, otherwise, the parameter is not the sensitive parameter; wherein the preset first threshold may be set to 25%, 27%, etc., and the second preset threshold may be set to 30%, 35%, etc.; thus, a first sensitive parameter set S containing sensitive parameters corresponding to the single-factor sensitivity analysis method is obtained 1 。
2) The process of analyzing the parameters of the aircraft engine to be diagnosed by taking the principal component analysis method as an example is as follows:
sensitivity mining is carried out on multi-dimensional parameters, the multi-dimensional parameters are used as input, each principal component is obtained through a principal component analysis method, arrangement is carried out according to the contribution rate, for example, the principal component with the accumulated contribution rate larger than 80% is extracted in the application, weighting calculation is carried out based on the weight of the principal component occupied by each parameter, the statistic final result is the sensitivity, and each parameter is arranged according to the sensitivity rate to obtain the sensitivityQueues, in particular: the sensitivity of any parameter is the product of the sum of the contribution rates of each principal component and the weight of that parameter; then according to the formula:
obtaining a sensitivity threshold of the sensitivity of any two adjacent parameters in the queue, wherein the parameter a and the parameter b are any two parameters in the queue, and the sensitivity of the parameter a is greater than that of the parameter b;
when the sensitivity gradient between the sensitivity of the parameter a and the sensitivity of the parameter b is greater than a preset gradient threshold, the parameter a is a sensitive parameter, and thus a first sensitive parameter set S containing the sensitive parameter corresponding to the principal component analysis method is obtained 2 。
3) Analyzing the parameters of the aircraft engine to be diagnosed by taking the random forest algorithm as an example, and judging whether the parameters meet a second preset condition, wherein the second preset condition can be manually set according to experience, so that a first sensitive parameter set S which corresponds to the random forest algorithm and contains sensitive parameters is obtained 3 。
It is understood that since the specific calculation processes of the single-factor sensitivity analysis method, the principal component analysis method, and the random forest algorithm are well known to those skilled in the art, they will not be described in detail.
S11, obtaining a first sensitive parameter union set, specifically:
first set S of sensitivity parameters including sensitivity parameters corresponding to a one-factor sensitivity analysis 1 A first sensitive parameter set S containing sensitive parameters corresponding to the principal component analysis method 2 A first set S of sensitive parameters corresponding to a random forest algorithm and comprising sensitive parameters 3 Taking a union set to obtain a first sensitive parameter union set;
s12, obtaining first strong correlation parameter pair sets respectively corresponding to the pearson correlation coefficient algorithm, the spearman correlation coefficient algorithm, and the kendell correlation coefficient algorithm, specifically:
1) the process of analyzing every two sensitive parameters in the first sensitive parameter union set by the pearson correlation coefficient algorithm is as follows:
the correlation coefficient between each two sensitive parameters is calculated according to the following formula:
x and Y represent any two sensitive parameters in the first sensitive parameter union set, (X, Y) is a two-dimensional random variable vector formed by the any two sensitive parameters, cov (X, Y) is the covariance of the two-dimensional random variable vector, and sigma is X Is the standard deviation of X, σ Y Is the standard deviation of Y. When the correlation coefficient between any two sensitive parameters is greater than a first preset correlation coefficient, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and therefore a first strong correlation parameter pair set C corresponding to the Pearson correlation coefficient algorithm is obtained 1 (ii) a It can be understood that any two sensitive parameters are a pair of sensitive parameters;
2) the process of analyzing every two sensitive parameters in the first sensitive parameter union set by the spearman correlation coefficient algorithm is as follows:
the correlation coefficient between each two sensitive parameters is calculated according to the following formula:
d i the difference value of the levels of any two sensitive parameters in the first sensitive parameter union set is obtained, N is the number of all sensitive parameters in the first sensitive parameter union set, and N is a positive integer. When the correlation coefficient between any two sensitive parameters is larger than a second preset correlation coefficient, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and therefore a first strong correlation parameter pair set C corresponding to the spearman correlation coefficient algorithm is obtained 2 (ii) a It can be understood that any two sensitive parameters are a pair of sensitive parameters;
3) sensitivity to first with Kendel correlation coefficient algorithmAnalyzing every two sensitive parameters in the parameter union set, wherein when the correlation coefficient between any two sensitive parameters is greater than a third preset correlation coefficient, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and thus obtaining a first strong correlation parameter pair set C corresponding to the Kendel correlation coefficient algorithm 3 ；
It is understood that since the specific calculation processes of the pearson correlation coefficient algorithm, the spearman correlation coefficient algorithm, and the kender correlation coefficient algorithm are well known to those skilled in the art, they will not be described in detail.
S13, obtaining a first strongly correlated parameter pair intersection set, specifically:
first strong correlation parameter pair set C corresponding to Pearson correlation coefficient algorithm 1 A first strong correlation parameter pair set C corresponding to the spearman correlation coefficient algorithm 2 A first strong correlation parameter pair set C corresponding to Kendel correlation coefficient algorithm 3 Taking an intersection to obtain a first strong correlation parameter pair intersection set;
s14, obtaining a second sensitive parameter union set, specifically:
deleting any sensitive parameter in each strongly correlated parameter pair in the first set of strongly correlated parameter pair intersections from the first set of sensitive parameter union.
And S15, performing fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result.
Preferably, in the above technical solution, the method further comprises:
s20, acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
s21, analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until a second sensitive parameter set which corresponds to each analysis method of all fault modes is obtained;
s22, taking a union set of all the second sensitive parameter sets to obtain a third sensitive parameter union set;
s23, analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strong correlation parameter pair set containing strong correlation parameter pairs corresponding to each correlation coefficient algorithm;
s24, taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
s25, deleting any sensitive parameter in each strongly-related parameter pair in the second strongly-related parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and S26, training according to the multiple fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
Wherein, a plurality of fault modes such as noise fault mode, vibration and fault mode can be artificially defined according to experience, and three analysis methods, namely a single-factor sensitivity analysis method, a principal component analysis method and a random forest algorithm, and three correlation coefficient algorithms, namely a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm and a kendell correlation coefficient algorithm, are used as examples for explanation:
s30, analyzing all parameters corresponding to any fault mode by using a single-factor sensitivity analysis method, a principal component analysis method, and a random forest algorithm, respectively, to obtain a second sensitive parameter set including sensitive parameters corresponding to each analysis method, as shown in fig. 2, specifically:
1) the process of analyzing all parameters of any fault mode by taking the single-factor sensitivity analysis method as an example is as follows:
sensitivity excavation is carried out based on parameter time domain signal variation amplitude serving as an analysis factor, sensitive parameters are excavated from multi-dimensional parameters, whether each parameter is a sensitive parameter or not is judged through a first preset condition, and specifically:
the first preset condition is that the variation amplitude gradient is larger than a first preset threshold value and the variation amplitude of the mean value of the parameter is larger than a second preset threshold valueWhen a threshold value is preset, judging the parameter as a sensitive parameter, otherwise, judging the parameter as a non-sensitive parameter; wherein the preset first threshold may be set to 25%, 27%, etc., and the second preset threshold may be set to 30%, 35%, etc.; thereby obtaining a second sensitive parameter set S 'containing sensitive parameters corresponding to the single-factor sensitivity analysis method' 1 。
2) The process of analyzing all parameters of any fault mode by taking the principal component analysis method as an example is as follows:
sensitivity mining is carried out on multidimensional parameters, the multidimensional parameters are used as input, each principal component is obtained through a principal component analysis method, and arrangement is carried out according to the contribution rate, for example, the principal component with the accumulated contribution rate larger than 80% is extracted in the application, weighting calculation is carried out based on the weight of the principal component occupied by each parameter, the statistic final result is sensitivity, each parameter is arranged according to the sensitivity rate to obtain a queue, specifically: the sensitivity of any parameter is the product of the sum of the contribution rates of each principal component and the weight of that parameter; then according to the formula:
obtaining a sensitivity threshold of the sensitivity of any two adjacent parameters in the queue, wherein the parameter a and the parameter b are any two parameters in the queue, and the sensitivity of the parameter a is greater than that of the parameter b;
when the sensitivity gradient between the sensitivity of the parameter a and the sensitivity of the parameter b is greater than a preset gradient threshold value, the parameter a is a sensitive parameter, and a second sensitive parameter set S 'containing the sensitive parameter and corresponding to the principal component analysis method is obtained' 2 。
3) Analyzing all parameters of any fault mode by taking a random forest algorithm as an example, and judging whether second preset conditions are met, wherein the second preset conditions can be manually set according to experience, so that a second sensitive parameter set S 'containing sensitive parameters and corresponding to the random forest algorithm is obtained' 3 . Until obtaining a second sensitive parameter set corresponding to each analysis method of all fault modes;
it is understood that since the specific calculation processes of the single-factor sensitivity analysis method, the principal component analysis method, and the random forest algorithm are well known to those skilled in the art, they will not be described in detail.
S31, obtaining a third sensitive parameter union set, specifically:
second set of sensitivity parameters S 'corresponding to all failure modes' 1 Second set of sensitivity parameters S' 2 And a second set of sensitivity parameters S' 3 And taking the union set to obtain a third sensitive parameter union set.
S32, obtaining second strong correlation parameter pair sets respectively corresponding to the pearson correlation coefficient algorithm, the spearman correlation coefficient algorithm, and the kender correlation coefficient algorithm, as shown in fig. 3, specifically:
1) the process of analyzing every two sensitive parameters in the third sensitive parameter union set by the pearson correlation coefficient algorithm is as follows:
the correlation coefficient between each two sensitive parameters is calculated according to the following formula:
when the correlation coefficient between any two sensitive parameters is greater than a first preset correlation coefficient, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and therefore a second strong correlation parameter pair set C 'corresponding to the Pearson correlation coefficient algorithm is obtained' 1 (ii) a It can be understood that any two sensitive parameters are a pair of sensitive parameters;
2) the process of analyzing every two sensitive parameters in the third sensitive parameter union set by the spearman correlation coefficient algorithm is as follows:
the correlation coefficient between each two sensitive parameters is calculated according to the following formula:
when the correlation coefficient between any two sensitive parameters is largeWhen a second preset correlation coefficient is obtained, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and therefore a second strong correlation parameter pair set C 'corresponding to the spearman correlation coefficient algorithm is obtained' 2 (ii) a It can be understood that any two sensitive parameters are a pair of sensitive parameters;
3) analyzing every two sensitive parameters in the third sensitive parameter union set by using a Kendel correlation coefficient algorithm, wherein when the correlation coefficient between any two sensitive parameters is greater than a third preset correlation coefficient, the two sensitive parameters are strong correlation parameter pairs, otherwise, the two sensitive parameters are not strong correlation parameter pairs, and thus obtaining a second strong correlation parameter pair set C 'corresponding to the Kendel correlation coefficient algorithm' 3 ；
It is understood that since the specific calculation processes of the pearson correlation coefficient algorithm, the spearman correlation coefficient algorithm, and the kender correlation coefficient algorithm are well known to those skilled in the art, they will not be described in detail.
S33, obtaining a second strong correlation parameter pair intersection set, specifically:
second strong correlation parameter pair set C 'corresponding to Pearson correlation coefficient algorithm' 1 And a second strong correlation parameter pair set C 'corresponding to the spearman correlation coefficient algorithm' 2 And a second strong correlation parameter pair set C 'corresponding to the Kendel correlation coefficient algorithm' 3 Taking the intersection to obtain a second strongly correlated parameter pair intersection set;
s14, obtaining a fourth sensitive parameter union set, specifically:
deleting any sensitive parameter in each strongly related parameter pair in the second strongly related parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set; redundant data are removed, the redundancy of the data is reduced, a fourth sensitive parameter union set is obtained, and the machine learning efficiency, namely the model training speed, is improved.
S15, model training, specifically:
training according to the multiple fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model, dividing multiple groups of data into a training set and a verification set, wherein one group of data represents one fault mode and all parameters corresponding to the fault mode in the fourth sensitive parameter union set, and then training on the basis of the deep classification network, wherein when the preset precision is achieved, the corresponding model is the fault diagnosis model, and generally, the more the data, the higher the precision of the trained fault diagnosis model is.
Preferably, in the above technical solution, the method further comprises:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
Random noise is added into an input layer of an automatic encoder model, and robustness and generalization capability of the deep classification network are improved.
In order to verify the effectiveness of the aircraft engine fault diagnosis method, the following experiments are carried out, specifically:
the failure modes are set as surge failure, throttle lever failure and electronic control failure, wherein the surge failure refers to pneumatic mismatching, the compressor is unstable to work, and the flow and the pressure of airflow show low-frequency large-amplitude axial oscillation; the failure of the throttle lever refers to the clamping stagnation of the fuel supply valve of the afterburner, and the failure of the electronic control refers to the logic problem of software fault reporting mostly, so that:
firstly, according to parameters respectively corresponding to a surge fault, a throttle lever fault and an electronic control fault, specifically including 306-dimensional parameters, and according to S21-S25, obtaining 17 sensitive parameters in a fourth sensitive parameter union set, and then training according to the surge fault, the throttle lever fault, the electronic control fault, the 17 sensitive parameters in the fourth sensitive parameter union set and a depth classification network to obtain a fault diagnosis model;
secondly, acquiring parameters of the aero-engine to be diagnosed, acquiring a second sensitive parameter union set according to S1-S3, inputting the sensitive parameters in the second sensitive parameter union set into a fault diagnosis model to diagnose the fault of the aero-engine to be diagnosed, and acquiring a fault diagnosis result; the experimental results show that:
1) when the single fault mode is adopted, namely the to-be-diagnosed aviation engine only has surge fault, throttle lever fault or electronic control fault, the accuracy rate of the fault diagnosis result is 83.33%;
2) when the aircraft engine to be diagnosed has at least two of surge faults, throttle lever faults or electronic control faults in a multi-fault mode, the accuracy rate of the fault diagnosis result is 82.30%;
therefore, the accuracy of the fault diagnosis result obtained by the aircraft engine fault diagnosis method is enough to provide data support for fault diagnosis; therefore, fault removal is conveniently carried out on the basis of the fault diagnosis result, the loss which is difficult to measure and is generated by further serious faults is avoided, the operation safety of the airplane is ensured, and the investment of maintenance cost can be reduced to a great extent.
In the foregoing embodiments, although the steps are numbered as S1, S2, etc., but only the specific embodiments are given in this application, and those skilled in the art may adjust the execution order of S1, S2, etc. according to the actual situation, which is also within the protection scope of the present invention, and it is understood that some embodiments may include some or all of the above embodiments.
As shown in fig. 4, an aircraft engine fault diagnosis system 200 according to an embodiment of the present invention includes a first analysis module 210, a second analysis module 220, a deletion module 230, and a fault diagnosis module 240;
the first analysis module 210 is configured to analyze parameters of the aircraft engine to be diagnosed by using at least three analysis methods to obtain a first sensitive parameter set including sensitive parameters corresponding to each analysis method, and obtain a union set of all the first sensitive parameter sets to obtain a first sensitive parameter union set;
the second analysis module 220 is configured to analyze every two sensitive parameters in the first sensitive parameter union set by using at least three correlation coefficient algorithms to obtain a first strongly correlated parameter pair set including strongly correlated parameter pairs corresponding to each correlation coefficient algorithm, and obtain an intersection of all the first strongly correlated parameter pair sets to obtain a first strongly correlated parameter pair intersection set;
the deleting module 230 is configured to delete any sensitive parameter in each strongly related parameter pair in the first strongly related parameter pair intersection set from the first sensitive parameter union set, so as to obtain a second sensitive parameter union set;
the fault diagnosis module 240 is configured to perform fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model, so as to obtain a fault diagnosis result.
Analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of various methods, the association information between the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the healthy and stable operation of the aero-engine is guaranteed, the operation safety of the aircraft is guaranteed, and the investment of maintenance cost can be reduced to a great extent, namely the maintenance and guarantee cost of the aero-engine is reduced.
Preferably, in the above technical solution, the method further includes a model training module, where the model training module is configured to:
acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until obtaining a second sensitive parameter set which corresponds to each analysis method of all fault modes;
a union set is acquired for all the second sensitive parameter sets to obtain a third sensitive parameter union set;
analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm;
taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the second strongly-relevant parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and training according to the plurality of fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
Preferably, in the above technical solution, the method further includes constructing a deep classification network module, where the deep classification network module is configured to:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
Random noise is added into an input layer of an automatic encoder model, and the robustness and the generalization capability of the deep classification network are improved.
Preferably, in the above technical solution, the at least three analysis methods at least include a single-factor sensitivity analysis method, a principal component analysis method, and a random forest algorithm;
the at least three correlation coefficient algorithms include at least a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm, and a kender correlation coefficient algorithm.
The above steps for realizing the corresponding functions of each parameter and each unit module in the aircraft engine fault diagnosis system 200 according to the present invention may refer to each parameter and step in the above embodiment of an aircraft engine fault diagnosis method, which are not described herein again.
As shown in fig. 5, an electronic device 300 according to an embodiment of the present invention includes a memory 310, a processor 320, and a program 330 stored in the memory 310 and running on the processor 320, where the processor 320 executes the program 330 to implement any of the steps of the method for diagnosing an aircraft engine fault as described above.
Analyzing the parameters of the aero-engine to be diagnosed twice, specifically, analyzing the parameters of the aero-engine to be diagnosed by adopting at least three analysis methods, and analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms; through the fusion of various methods, the association information between the parameters can be fully mined, the association relation between the fault mode and the parameters is established, and information support is provided for fault diagnosis; on the basis, the fault diagnosis is carried out on the aero-engine to be diagnosed through the fault diagnosis model to obtain an accurate fault diagnosis result, so that fault elimination is carried out on the basis of the fault diagnosis result, the inestimable loss caused by further major faults is avoided, the healthy and stable operation of the aero-engine is guaranteed, the operation safety of the aircraft is guaranteed, and the investment of maintenance cost can be reduced to a great extent, namely the maintenance and guarantee cost of the aero-engine is reduced.
The electronic device 300 may be a computer, a mobile phone, or the like, and correspondingly, the program 330 is computer software or a mobile phone APP, and the parameters and steps in the electronic device 300 according to the present invention may refer to the parameters and steps in the above embodiment of the aircraft engine fault diagnosis method, which are not described herein again.
As will be appreciated by one skilled in the art, the present invention may be embodied as a system, method or computer program product.
Accordingly, the present disclosure may be embodied in the form of: the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a "circuit," module "or" system. Furthermore, in some embodiments, the invention may also be embodied in the form of a computer program product in one or more computer-readable media having computer-readable program code embodied in the medium.
Any combination of one or more computer-readable media may be employed. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (5)
1. An aircraft engine fault diagnosis method, characterized by comprising:
analyzing parameters of the aircraft engine to be diagnosed by adopting at least three analysis methods to obtain a first sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method, and collecting a union set of all the first sensitive parameter sets to obtain a first sensitive parameter union set;
the parameter of the aero-engine to be diagnosed is to be diagnosed as the parameter monitored on all components of the aero-engine to be diagnosed, or the parameter of the key component of the aero-engine to be diagnosed is set manually according to experience;
analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms to obtain a first strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm, and taking intersection of all the first strongly correlated parameter pair sets to obtain a first strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the first strongly-relevant parameter pair intersection set from the first sensitive parameter union set to obtain a second sensitive parameter union set;
performing fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result;
the at least three analysis methods at least comprise a single factor sensitivity analysis method, a principal component analysis method and a random forest algorithm;
the at least three correlation coefficient algorithms at least comprise a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm and a kender correlation coefficient algorithm;
further comprising:
acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until obtaining a second sensitive parameter set which corresponds to each analysis method of all fault modes;
a union set is acquired for all the second sensitive parameter sets to obtain a third sensitive parameter union set;
analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm;
taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the second strongly-relevant parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and training according to the plurality of fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
2. The aircraft engine fault diagnosis method according to claim 1, characterized by further comprising:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
3. The aircraft engine fault diagnosis system is characterized by comprising a first analysis module, a second analysis module, a deletion module and a fault diagnosis module;
the first analysis module is used for analyzing parameters of the aircraft engine to be diagnosed by adopting at least three analysis methods to obtain a first sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method, and performing union on all the first sensitive parameter sets to obtain a first sensitive parameter union set;
the parameter of the aero-engine to be diagnosed is to be diagnosed as the parameter monitored on all components of the aero-engine to be diagnosed, or the parameter of the key component of the aero-engine to be diagnosed is set manually according to experience;
the second analysis module is used for analyzing every two sensitive parameters in the first sensitive parameter union set by adopting at least three correlation coefficient algorithms to obtain a first strong correlation parameter pair set containing strong correlation parameter pairs corresponding to each correlation coefficient algorithm, and taking intersection of all the first strong correlation parameter pair sets to obtain a first strong correlation parameter pair intersection set;
the deleting module is used for deleting any sensitive parameter in each strongly-related parameter pair in the first strongly-related parameter pair intersection set from the first sensitive parameter union set to obtain a second sensitive parameter union set;
the fault diagnosis module is used for carrying out fault diagnosis on the aviation engine to be diagnosed according to the second sensitive parameter union set and the fault diagnosis model to obtain a fault diagnosis result;
the at least three analysis methods at least comprise a single factor sensitivity analysis method, a principal component analysis method and a random forest algorithm;
the at least three correlation coefficient algorithms at least comprise a pearson correlation coefficient algorithm, a spearman correlation coefficient algorithm and a kender correlation coefficient algorithm;
still include the model training module, the model training module is used for:
acquiring parameters corresponding to a plurality of failure modes of the aircraft engine respectively;
analyzing all parameters corresponding to any fault mode by adopting the at least three analysis methods to obtain a second sensitive parameter set which comprises sensitive parameters and corresponds to each analysis method of the fault mode until obtaining a second sensitive parameter set which corresponds to each analysis method of all fault modes;
a union set is acquired for all the second sensitive parameter sets to obtain a third sensitive parameter union set;
analyzing every two sensitive parameters in the third sensitive parameter union set by adopting the at least three correlation coefficient algorithms to obtain a second strongly correlated parameter pair set containing strongly correlated parameter pairs corresponding to each correlation coefficient algorithm;
taking intersection of all the second strongly correlated parameter pair sets to obtain a second strongly correlated parameter pair intersection set;
deleting any sensitive parameter in each strongly-relevant parameter pair in the second strongly-relevant parameter pair intersection set from the third sensitive parameter union set to obtain a fourth sensitive parameter union set;
and training according to the plurality of fault modes, the fourth sensitive parameter union set and the deep classification network to obtain a fault diagnosis model.
4. The aircraft engine fault diagnosis system according to claim 3, further comprising a build-depth classification network module configured to:
adding random noise into an input layer of an automatic encoder model to obtain a depth noise reduction laminated automatic encoder;
and combining the depth denoising stacking automatic encoder with a Softmax classifier to obtain the depth classification network.
5. An electronic device comprising a memory, a processor and a program stored on the memory and run on the processor, characterized in that the steps of a method of diagnosing a malfunction of an aircraft engine according to claim 1 or 2 are implemented when the program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011495593.5A CN112733880B (en) | 2020-12-17 | 2020-12-17 | Aircraft engine fault diagnosis method and system and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011495593.5A CN112733880B (en) | 2020-12-17 | 2020-12-17 | Aircraft engine fault diagnosis method and system and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112733880A CN112733880A (en) | 2021-04-30 |
| CN112733880B true CN112733880B (en) | 2022-09-20 |
Family
ID=75603916
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011495593.5A Active CN112733880B (en) | 2020-12-17 | 2020-12-17 | Aircraft engine fault diagnosis method and system and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112733880B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343364A (en) * | 2021-07-05 | 2021-09-03 | 中国国际航空股份有限公司 | Aviation big data analysis method and tool |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3019248A1 (en) * | 2016-04-05 | 2017-10-12 | Taleris Global Llp | Systems and methods to determine maintenance-focused aircraft data recording frame configurations |
| CN109618361A (en) * | 2019-01-22 | 2019-04-12 | 北京市天元网络技术股份有限公司 | A kind of base station 4G hidden failure checks method and device |
| CN109993232A (en) * | 2019-04-04 | 2019-07-09 | 浙江大学 | Turbine Vibrationfault Diagnosis method based on deep learning artificial neural network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI549007B (en) * | 2013-02-07 | 2016-09-11 | 先知科技股份有限公司 | Method for searching and analyzing process parameters and computer program product thereof |
-
2020
- 2020-12-17 CN CN202011495593.5A patent/CN112733880B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3019248A1 (en) * | 2016-04-05 | 2017-10-12 | Taleris Global Llp | Systems and methods to determine maintenance-focused aircraft data recording frame configurations |
| CN109618361A (en) * | 2019-01-22 | 2019-04-12 | 北京市天元网络技术股份有限公司 | A kind of base station 4G hidden failure checks method and device |
| CN109993232A (en) * | 2019-04-04 | 2019-07-09 | 浙江大学 | Turbine Vibrationfault Diagnosis method based on deep learning artificial neural network |
Non-Patent Citations (2)
| Title |
|---|
| 一种基于自适应神经网络的航空发动机故障诊断方法;艾剑良等;《中国科学:技术科学》;20180319(第03期);第94-103页 * |
| 基于知识库关联的航空发动机气路故障模型设计;郭慧玲等;《计算机测量与控制》;20140225(第02期);第70-72+80页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112733880A (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108334907B (en) | A kind of complex equipment point method for detecting abnormality and system based on deep learning | |
| US9818242B2 (en) | Gas turbine engine anomaly detections and fault identifications | |
| CN106094570B (en) | A kind of aero-engine complete machine health evaluating method under variable working condition based on this distance of operating mode's switch and paddy | |
| JP2016224913A (en) | Artificial intelligence-based health management of host system | |
| EP2388672A1 (en) | Identifying of turbomachine faults | |
| CN111581763B (en) | Method for evaluating air path fault diagnosis result of aero-engine | |
| CN110987439B (en) | Aeroengine fault prediction method based on Logitics regression and Xgboost model | |
| CN106873568A (en) | Sensor fault diagnosis method based on H infinity robust Unknown Input Observer | |
| CN112052871B (en) | Rocket engine health diagnosis method and system based on support vector machine | |
| CN112733880B (en) | Aircraft engine fault diagnosis method and system and electronic equipment | |
| EP2354876B1 (en) | Engine Monitoring | |
| CN103914386B (en) | Software defect positioning method based on input parameter characteristic spectrum | |
| Stenfelt et al. | Automatic gas turbine matching scheme adaptation for robust GPA diagnostics | |
| US11054340B2 (en) | Parametric trending architecture concept and design | |
| CN115859791A (en) | Method and device for estimating residual life of aero-engine | |
| CN110186684A (en) | A kind of aero-engine mechanical oscillation fault-signal feature extracting method | |
| Huang et al. | Gas path deterioration observation based on stochastic dynamics for reliability assessment of aeroengines | |
| CN109711000A (en) | A kind of Aero-Engine Start method for diagnosing faults based on firing test data | |
| CN110287594B (en) | Aero-engine state diagnosis method based on neural network algorithm | |
| CN112749789A (en) | Aero-engine multiple fault diagnosis device based on self-association neural network | |
| CN114154238A (en) | Aeroengine performance monitoring method, system and medium for adaptive evaluation of section parameters | |
| CN111336003B (en) | Method for positioning key influence factors of turbocharging system of aircraft engine | |
| Shen et al. | Quantitative Detection of Blade Crack Damage Based on Vibro-Acoustic Information and Multi-Dimensional Feature Fusion CNN | |
| CN117272638B (en) | Method for positioning faults of aircraft in real time based on layered directed graph | |
| CN117454145A (en) | Aeroengine fault diagnosis method based on multi-mode sensor |
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 |