CN112613186A - Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics - Google Patents

Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics Download PDF

Info

Publication number
CN112613186A
CN112613186A CN202011608883.6A CN202011608883A CN112613186A CN 112613186 A CN112613186 A CN 112613186A CN 202011608883 A CN202011608883 A CN 202011608883A CN 112613186 A CN112613186 A CN 112613186A
Authority
CN
China
Prior art keywords
engine
fault
data
diagnosis
parameters
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011608883.6A
Other languages
Chinese (zh)
Other versions
CN112613186B (en
Inventor
范满意
孔祥兴
张瑞
杨博闻
李洋洋
梁宁宁
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.)
China Aero Engine Research Institute
Original Assignee
China Aero Engine Research Institute
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 China Aero Engine Research Institute filed Critical China Aero Engine Research Institute
Priority to CN202011608883.6A priority Critical patent/CN112613186B/en
Publication of CN112613186A publication Critical patent/CN112613186A/en
Application granted granted Critical
Publication of CN112613186B publication Critical patent/CN112613186B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • 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
    • G06F16/21—Design, administration or maintenance of databases
    • G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Analysis (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Evolutionary Computation (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • Quality & Reliability (AREA)
  • Algebra (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The invention provides an aeroengine gas circuit fault fusion diagnosis method based on statistical distribution characteristics, which constructs a matching diagnosis module based on parameter deviation statistical characteristics, fuses diagnosis results by adopting a weighted D-S evidence theory based on fault sensitivity, reduces and even eliminates the adverse effects of engine volume effect, metal heat transfer and measurement noise and offset on engine fault diagnosis, and improves the accuracy and reliability of engine fault diagnosis; in order to solve the problem of multi-measurement parameter fusion conflict, the reliability of different measurement parameters is obtained through fault sensitivity analysis, so that different weight coefficients are given to the fault diagnosis results of different measurement parameter statistical distribution characteristics in the fusion process, and the reliability of diagnosis is improved.

Description

Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics
Technical Field
The disclosure relates to the technical field of aeroengine fault diagnosis and health management, in particular to an aeroengine gas circuit fault fusion diagnosis method based on statistical distribution characteristics.
Background
The aeroengine is a large-scale complex electromechanical system integrating mechanical, electrical, hydraulic, pneumatic and various high and new technologies. The engine has a complex structure, high requirements on processing and mounting of parts and parts, a severe working environment, a maximum turbine front temperature of 2000K, a rotor rotating speed of more than 10000rpm, frequent take-off, climbing, cruising and other tasks in the use process, and inevitably fails to bear high centrifugal load, aerodynamic load, high-temperature and atmospheric temperature difference load, alternating load of vibration and the like for a long time. In order to ensure safe and reliable operation of the engine, the engine is subjected to fault diagnosis through measurement parameters acquired by an engine sensor, and the problem which occurs or is likely to occur is evaluated, so that the method has important value and significance.
The engine fault diagnosis, especially the gas circuit fault diagnosis, is greatly influenced by the engine state, because most methods in the current engine gas circuit fault diagnosis methods are used for diagnosing in a stable working state, but the engine is in a quasi-steady state process from one stable working state to another stable working state within a considerable period of time due to the influence of the engine volume effect and the metal heat transfer. Meanwhile, the sensor of the engine senses signals such as pressure, temperature, rotating speed and flow of each section of the engine and converts the signals into data used by the engine fault diagnosis system, and the four processes of sensing, transmitting, collecting and converting are approximately carried out, and the signals are inevitably interfered by the external electromagnetic environment in the process, so that the data analyzed by the engine is noisy and even has deviation. These factors clearly increase the difficulty of engine fault diagnosis.
In the current aeroengine fault diagnosis based on measurement parameter deviation, the fault diagnosis is mainly completed according to the magnitude of parameter offset, namely, when the engine offset exceeds a certain threshold value, the judgment result is that the fault is detected, otherwise, the fault is normal. The engine is influenced by the volume effect, the metal heat transfer and the measurement noise, the parameters have certain fluctuation, and most of the current diagnosis methods do not consider the fluctuation of the measurement parameters, so that the accuracy of fault diagnosis in real engine data is limited.
Disclosure of Invention
In order to solve at least one of the technical problems, the disclosure provides an aeroengine gas circuit fault fusion diagnosis method based on statistical distribution characteristics, which can improve the accuracy and reliability of engine fault diagnosis.
In order to achieve the purpose, the following technical scheme is adopted in the disclosure:
an aeroengine gas circuit fault fusion diagnosis method based on statistical distribution characteristics comprises the following steps:
data preprocessing: respectively carrying out data cleaning, similarity conversion, steady state interception and data intelligent tag adding operation on historical data of the engine to obtain steady state data with consistent import conditions and data intelligent tags;
constructing a reference model: taking data with a normal label in the steady-state data as input, and carrying out reference model modeling by establishing a mapping relation between main measurement parameters of the engine and controlled parameters of the engine to obtain an engine reference model so as to provide support for fault diagnosis;
extracting deviation statistical characteristics: based on the engine reference model, taking data with a fault label in the steady-state data as input, and acquiring a difference value between an engine measurement parameter and the engine reference model; then, sensitivity analysis is carried out according to the influence of the engine fault on the engine measurement parameters; extracting statistical characteristics based on deviation, converting engine measurement parameters into statistical distribution characteristics capable of supporting fault diagnosis, and constructing an engine statistical distribution characteristic library;
fusion diagnosis: carrying out statistical distribution characteristic matching of single engine parameters on an engine statistical distribution characteristic library, and carrying out fault diagnosis based on single measurement parameter deviation statistical characteristics; normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics; carrying out diagnosis result fusion on the fault diagnosis results of different measurement parameter statistical distribution characteristics by using a weighted D-S evidence theory method; optimizing and integrating the fusion diagnosis result to obtain a final diagnosis result.
According to at least one embodiment of the present disclosure, the data preprocessing comprises the following specific steps:
data cleaning: cleaning abnormal values in the historical data of the engine by adopting a mode based on median filtering and mean filtering;
and (4) similar conversion: by utilizing the similarity principle of the engine, the parameters of the sections of the runners of the engine are adjusted by aligning the temperature and the pressure of the inlet of the engine to the standard atmospheric condition, so that the influence of the inconsistency of the inlet height, the Mach number, the temperature and the pressure on the analysis result is eliminated;
and (3) steady state interception: setting a moving window with a certain length, intercepting steady-state section data according to the slope of a controlled parameter in the moving window, and extracting data of a period of time with a relatively stable engine state to obtain steady-state data;
data smart label: according to the statistical characteristics of the data of each steady-state section of the engine, classification of engine fault data and abnormal data is realized by adopting an unsupervised learning method, labels of the engine data are determined by adopting a C-means clustering algorithm, and corresponding labels are added to the clustered data according to the engine test run record, the hole detection and the issuing of the disassembled analysis result.
According to at least one embodiment of the present disclosure, the reference model is constructed by the following specific steps:
dividing the data with normal labels through the fan guide vane angle, the compressor guide vane angle and the tail nozzle area;
and solving the mean value of each steady-state section of different data sets to construct an interpolation table, and obtaining an engine reference model by establishing a mapping relation between main measurement parameters and controlled parameters of the engine.
According to at least one embodiment of the present disclosure, the extracting of the deviation statistical features specifically includes the following steps:
obtaining deviation: acquiring a difference value between engine measurement data and an engine reference model under the condition that parameters of a controlled rotating speed, a fan guide vane angle, a compressor guide vane angle and a tail nozzle area are aligned;
and (3) fault sensitivity analysis: for typical fault modes of m types of engines, the influence of each type of fault mode on n measurement parameters is different, and a sensitivity matrix SM of the engine measurement parameters on the fault modes is obtained by comparing the average relative change of the engine measurement parameters before and after the engine fault occurs;
Figure BDA0002874163040000031
in the formula, smijRepresenting the sensitivity of the jth parameter to the ith fault mode; i is 1,2, …, m; j is 1,2, …, n;
the confidence w of the diagnosis result of the pattern matching of the relative deviation statistic of different measurement parametersijComprises the following steps:
Figure BDA0002874163040000041
in the formula (I), the compound is shown in the specification,
Figure BDA0002874163040000042
the maximum value of the average relative change absolute value of the measurement parameter before and after the ith fault mode occurs is represented;
Figure BDA0002874163040000043
indicating the absolute value of the average relative change of the measured parameters before and after the occurrence of the ith fault modeMinimum value of (d); i is 1,2, …, m, epsilon1、ε2Represents two smaller numbers;
the confidence matrix for the engine measured parameters is then:
Figure BDA0002874163040000044
extracting statistical characteristics based on deviation: after the deviation of the engine measurement parameters is obtained, the characteristics of the mean value and the variance of the engine measurement parameters are counted, and an engine statistical distribution characteristic library is established.
According to at least one embodiment of the present disclosure, the fusion diagnosis comprises the following specific steps:
fault diagnosis based on single measurement parameter deviation statistical characteristics: matching the mean value and the variance of the deviation of the measured parameters to be diagnosed in the statistical distribution characteristic library of the engine with the mean value and the variance of the deviation of the measured parameters of the known fault mode;
let x be a measurement parameter, x ═ x1,x2,…xn]N represents the number of measured parameters, and the probability distribution function of the i-th fault mode is fi(xj) The probability distribution function of the data to be diagnosed is f (x)j) Wherein j is 1,2, …, n, xjIs in the range of [ -inf, inf]Inf represents infinity;
then the j-th measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijComprises the following steps:
Figure BDA0002874163040000045
normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics;
measuring the jth measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijAnd (3) carrying out normalization:
Figure BDA0002874163040000046
the failure diagnosis result after normalization is:
Figure BDA0002874163040000047
in the formula, m represents the number of failure modes;
performing fusion diagnosis on the fault diagnosis results of different measurement parameter statistical distribution characteristics: calculating the average matching degree of the measured parameters of each engine, namely the average value of the diagnosis results of the measured parameters of different engines, and fusing the diagnosis results by using a weighted D-S evidence theory method;
weighting the diagnosis results of the different engine measurement parameters by the following formula to obtain weighted fault diagnosis results and uncertainty based on single measurement parameter feature matching as follows:
Figure BDA0002874163040000051
in the formula (I), the compound is shown in the specification,
Figure BDA0002874163040000052
represents the uncertainty of the diagnostic result of the jth measured parameter,
Figure BDA0002874163040000053
representing the weighted fault diagnosis, wijRepresents a weighting coefficient;
the weighted fault diagnosis results based on different parameters of different fault feature libraries are represented in a matrix form as follows:
Figure BDA0002874163040000054
in the formula, the last line represents the uncertainty of different measurement parameters;
synthesis of rule pair matrices using Dempster
Figure BDA0002874163040000055
M-1 fusion from left to right: first pair matrix
Figure BDA0002874163040000056
The first column and the second column are fused, then the result after the first fusion is fused with the third column of the matrix, the result after the second fusion is fused with the fourth column, and so on until the matrix is fused
Figure BDA0002874163040000057
All columns are fused until the completion;
optimizing and integrating fusion diagnosis results: firstly, the average probability of the occurrence of the fault of the data to be diagnosed is calculated, and the fusion diagnosis result is optimized on the basis to obtain the fault diagnosis result.
Compared with the prior art, the present disclosure has the advantages that:
according to the aircraft engine gas circuit fault fusion diagnosis method based on the statistical distribution characteristics, the matching diagnosis module based on the parameter deviation statistical characteristics is constructed, the local diagnosis results are fused by adopting the weighted D-S evidence theory based on the fault sensitivity, the adverse effects of the engine volume effect, the metal heat transfer and the measurement noise and offset on the engine fault diagnosis are reduced and even eliminated, and the accuracy and the reliability of the engine fault diagnosis are improved; in order to solve the problem of multi-measurement parameter fusion conflict, the reliability of different measurement parameters is obtained through fault sensitivity analysis, so that different weight coefficients are given to the fault diagnosis results of different measurement parameter statistical distribution characteristics in the fusion process, and the reliability of diagnosis is improved; meanwhile, the average probability of fault occurrence is added to further optimize the diagnosis result, so that the diagnosis result not only contains the diagnosis information that the engine belongs to a certain fault mode, but also contains the current engine fault severity information, and the diagnosis result is more credible.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the disclosure and together with the description serve to explain the principles of the disclosure.
FIG. 1 is a flow chart of fusion diagnosis of gas path faults of an aircraft engine based on statistical distribution.
FIG. 2 is a reference model of the engine constructed in accordance with the present invention.
Fig. 3 is a schematic diagram of fault diagnosis based on probability distribution according to the present invention.
Fig. 4 is the diagnostic result of the present invention for three faults.
FIG. 5 is a diagnostic result of the present invention on the last three run data.
Detailed Description
The present disclosure will be described in further detail with reference to the drawings and embodiments. It is to be understood that the specific embodiments described herein are for purposes of illustration only and are not to be construed as limitations of the present disclosure. It should be further noted that, for the convenience of description, only the portions relevant to the present disclosure are shown in the drawings.
It should be noted that the embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict. The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
The current engine gas circuit fault diagnosis mainly comprises a fault diagnosis method based on a model, a fault diagnosis method based on a measurement parameter deviation threshold value and a fault diagnosis method based on artificial intelligence. The dependence of the model-based diagnosis on the aerodynamic thermodynamic model is high, and under the condition that the parameters deviate from normal values due to the factors, the aerodynamic thermodynamic model cannot be applied, and the flow and efficiency of iterative computation deviate from real values, so that the fault diagnosis result is inaccurate. The diagnosis based on the measured parameter deviation threshold value is to obtain a diagnosis result by comparing the measured parameter deviation with the threshold value, and a false alarm may exist in the threshold value-based method when the engine is in a quasi-steady state process. Artificial intelligence based methods require re-training of the model after the addition of a new fault.
Aiming at the problems in the prior art, the invention provides the fusion diagnosis method for the gas circuit fault of the aero-engine based on the statistical distribution characteristics, which reduces or even eliminates the influence of the fluctuation of the measured parameters on the fault diagnosis of the aero-engine by obtaining the statistical distribution characteristics of the deviation of the measured parameters, further completes the preliminary diagnosis by matching the statistical characteristics of the deviation of the single measured parameter, and finally fuses the diagnosis results of the deviation statistical characteristics of different measured parameters to realize the accurate diagnosis of the fault of the aero-engine.
The technical scheme adopted by the invention is as follows:
as shown in fig. 1, a statistical distribution feature-based fusion diagnosis method for an aircraft engine gas path fault includes the following steps:
s1, data preprocessing: respectively carrying out data cleaning, similarity conversion, steady state interception and data intelligent tag adding operation on historical data of the engine, and obtaining steady state data which has no large deviation point, consistent inlet conditions and data intelligent tags under the condition of replacing abnormal values; the data preprocessing comprises the following specific steps:
data cleaning: cleaning abnormal values in the historical data of the engine by adopting a mode based on median filtering and mean filtering; replacing the abnormal value by the data before or after the abnormal data at a certain time or the mean value of the data before and after the abnormal data;
and (4) similar conversion: by utilizing the similarity principle of the engine, the parameters of the sections of the runners of the engine are adjusted by aligning the temperature and the pressure of the inlet of the engine to the standard atmospheric condition, so that the influence of the inconsistency of the inlet height, the Mach number, the temperature and the pressure on the analysis result is eliminated;
and (3) steady state interception: setting a moving window with a certain length, intercepting steady-state section data according to the slope of a controlled parameter in the moving window, abandoning a data set with smaller steady-state section data points, cutting off the head and the tail of the remaining steady-state section data, and extracting data of a period of time with a relatively stable engine state to obtain steady-state data; the controlled parameters are determined by a control plan, wherein the low-pressure rotor conversion rotating speed or the high-pressure rotor conversion rotating speed is selected preferentially by the controlled parameters;
data smart label: according to the statistical characteristics of the data of each steady-state section of the engine, classification of engine fault data and abnormal data is realized by adopting an unsupervised learning method, labels of the engine data are determined by adopting a C-means clustering algorithm, and corresponding labels are added to the clustered data according to the engine test run record, the hole detection and the issuing of the disassembled analysis result.
S2, constructing a reference model: taking data with a normal label in the steady-state data as input, and carrying out reference model modeling by establishing a mapping relation between main measurement parameters of the engine and controlled parameters of the engine to obtain an engine reference model so as to provide support for fault diagnosis;
dividing the data with normal labels through the fan guide vane angle, the compressor guide vane angle and the tail nozzle area;
and solving the mean value of each steady-state section of different data sets to construct an interpolation table, and obtaining an engine reference model by establishing a mapping relation between main measurement parameters and controlled parameters of the engine.
S3, extracting deviation statistical characteristics: based on the engine reference model, taking data with a fault label in the steady-state data as input, and acquiring a difference value between an engine measurement parameter and the engine reference model; then, sensitivity analysis is carried out according to the influence of the engine fault on the engine measurement parameters; extracting statistical characteristics based on deviation, converting engine measurement parameters into statistical distribution characteristics capable of supporting fault diagnosis, and constructing an engine statistical distribution characteristic library; the specific steps of extracting the deviation statistical characteristics are as follows:
obtaining deviation: acquiring a difference value between engine measurement data and an engine reference model under the condition that parameters of a controlled rotating speed, a fan guide vane angle, a compressor guide vane angle and a tail nozzle area are aligned;
and (3) fault sensitivity analysis: for typical fault modes of m types of engines, the influence of each type of fault mode on n measurement parameters is different, and a sensitivity matrix SM of the engine measurement parameters on the fault modes is obtained by comparing the average relative change of the engine measurement parameters before and after the engine fault occurs;
Figure BDA0002874163040000081
in the formula, smijRepresenting the sensitivity of the jth parameter to the ith fault mode; i is 1,2, …, m; j is 1,2, …, n;
the confidence w of the diagnosis result of the pattern matching of the relative deviation statistic of different measurement parametersijComprises the following steps:
Figure BDA0002874163040000082
in the formula (I), the compound is shown in the specification,
Figure BDA0002874163040000083
the maximum value of the average relative change absolute value of the measurement parameter before and after the ith fault mode occurs is represented;
Figure BDA0002874163040000084
the minimum value of the average relative change absolute value of the measurement parameter before and after the ith fault mode occurs is represented; i is 1,2, …, m, epsilon1、ε2Represents two smaller numbers;
the confidence matrix for the engine measured parameters is then:
Figure BDA0002874163040000091
extracting statistical characteristics based on deviation: after the deviation of the engine measurement parameters is obtained, the characteristics of the mean value and the variance of the engine measurement parameters are counted, and an engine statistical distribution characteristic library is established.
S4, fusion diagnosis: carrying out statistical distribution characteristic matching of single engine parameters on an engine statistical distribution characteristic library, and carrying out fault diagnosis based on single measurement parameter deviation statistical characteristics; normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics; carrying out diagnosis result fusion on the fault diagnosis results of different measurement parameter statistical distribution characteristics by using a weighted D-S evidence theory method; optimizing and integrating the fusion diagnosis result to obtain a final diagnosis result; the fusion diagnosis comprises the following specific steps:
fault diagnosis based on single measurement parameter deviation statistical characteristics: matching the mean value and the variance of the deviation of the measured parameters to be diagnosed in the statistical distribution characteristic library of the engine with the mean value and the variance of the deviation of the measured parameters of the known fault mode;
let x be a measurement parameter, x ═ x1,x2,…xn]N represents the number of measured parameters, and the probability distribution function of the i-th fault mode is fi(xj) The probability distribution function of the data to be diagnosed is f (x)j) Wherein j is 1,2, …, n, xjIs in the range of [ -inf, inf]Inf represents infinity;
then the j-th measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijComprises the following steps:
Figure BDA0002874163040000092
normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics: measuring the jth measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijAnd (3) carrying out normalization:
Figure BDA0002874163040000093
the failure diagnosis result after normalization is:
Figure BDA0002874163040000094
in the formula, m represents the number of failure modes;
matching the measured parameters to be diagnosed with a plurality of fault modes to obtain a plurality of diagnosis results, and in order to realize the fusion of the diagnosis results of different measured parameters, carrying out the normalization of the fault diagnosis results based on the statistical distribution characteristics of single measured parameters;
performing fusion diagnosis on the fault diagnosis results of different measurement parameter statistical distribution characteristics: calculating the average matching degree of the measured parameters of each engine, namely the average value of the diagnosis results of the measured parameters of different engines, and fusing the diagnosis results by using a weighted D-S evidence theory method;
weighting the diagnosis results of the different engine measurement parameters by the following formula to obtain weighted fault diagnosis results and uncertainty based on single measurement parameter feature matching as follows:
Figure BDA0002874163040000101
in the formula (I), the compound is shown in the specification,
Figure BDA0002874163040000102
represents the uncertainty of the diagnostic result of the jth measured parameter,
Figure BDA0002874163040000103
represents the weighted fault diagnosis result, wijRepresents a weighting coefficient;
the weighted fault diagnosis results based on different parameters of different fault feature libraries are represented in a matrix form as follows:
Figure BDA0002874163040000104
in the formula, the last line represents the uncertainty of different measurement parameters;
synthesis of rule pair matrices using Dempster
Figure BDA0002874163040000105
M-1 fusion from left to right: first pair matrix
Figure BDA0002874163040000106
The first column and the second column are fused, then the result after the first fusion is fused with the third column of the matrix, the result after the second fusion is fused with the fourth column, and so on until the matrix is fused
Figure BDA0002874163040000107
All columns are fused until the completion;
optimizing and integrating fusion diagnosis results: firstly, the average probability of the occurrence of the fault of the data to be diagnosed is calculated, and the fusion diagnosis result is optimized on the basis to obtain the fault diagnosis result.
In the fusion diagnosis process, the normalization of the diagnosis results of the same measurement parameter causes the filtering of the severity of the engine fault, and only the type attribute of the fault is reserved, so that the average probability of the fault is introduced to correct the fusion diagnosis result in order to solve the problem.
Verification of the examples:
in order to verify the effectiveness of the method for diagnosing the gas circuit fault fusion of the aero-engine based on the statistical distribution characteristics, the extraction and fault diagnosis of the statistical distribution characteristics of the engine are carried out based on real engine measurement data.
The method comprises the steps of analyzing the high-pressure rotor reduced rotating speed, the fan inlet temperature, the throttle lever, the high-pressure rotor reduced rotating speed and the low-pressure rotor rotating speed N of three faults of engine high-pressure turbine block falling, high-pressure turbine fracture and compressor fracture1Reduced rotation speed of low-pressure rotor, inlet temperature of compressor, outlet temperature P3 of compressor, outlet pressure of compressor and outlet pressure P of low-pressure turbine6Low pressure turbine outlet temperature T6Area of throat of tail nozzle, fan outlet culvert pressure P13And (3) carrying out data preprocessing on the equal parameters to obtain more regular labeled data, and establishing an engine reference model as shown in figure 2.
And then, engine fault diagnosis algorithm research and verification are carried out by utilizing the historical fault data of the engine. In fig. 4, a, b, and c are a diagnosis result of a high-pressure turbine blade chipping of an engine, a diagnosis result of a high-pressure turbine blade fracture of an engine, and a diagnosis result of a compressor fracture of an engine, respectively, where an abscissa represents time, each segment represents a trial run number, an ordinate represents a probability of a fault occurrence, and points of different shapes represent different fault diagnosis results, and as can be seen from a in fig. 4, a probability that the engine is diagnosed as a high-pressure turbine fracture at the beginning is the largest, but does not exceed 40%, and as a trial run progresses, a probability of diagnosing a high-pressure turbine fracture gradually decreases, and is almost zero at last. Starting from the 4 th test run, the fault removing diagnosis method detects that the high-pressure turbine blade chipping fault possibly occurs, along with the 5 th, 6 th and 8 th test runs, the probability of diagnosing the high-pressure turbine blade chipping fault is gradually increased, the actual condition of engine fault development is met, and the fault removing diagnosis method can find the occurrence of the engine blade chipping fault 3-4 times in advance. As can be seen from b and c in fig. 4, the fault diagnosis results of the last two test runs of the engine are relatively accurate, but the engine may give wrong fault diagnosis results at a certain time point or within a certain period of time during the previous test runs.
The diagnosis results of the engine high-pressure turbine (HPT) blade chipping, the high-pressure turbine (HTP) blade cracking and the compressor (HPC) cracking obtained by the last three trials are shown in fig. 5, d, e and f in fig. 5 are respectively the diagnosis result of the third from last trial, the diagnosis result of the second from last trial and the diagnosis result of the first from last trial, and it can be found that the diagnosis accuracy of the high-pressure turbine (HPT) blade chipping fault is 100%, the diagnosis accuracy of the high-pressure turbine (HTP) blade cracking fault is 100%, 73.7% and 100%, the diagnosis accuracy of the second from last trial is 11.8% misdiagnosed as the compressor cracking, and the diagnosis accuracy of the compressor (HPC) cracking fault is also 100%. The final three diagnosis results are integrated and shown in table 1, and the average fault diagnosis accuracy rate of the final three-stage test run of the three faults is 95.57%.
TABLE 1 Fault diagnosis results of the last three engine test runs
Figure BDA0002874163040000121
According to the aeroengine gas circuit fault fusion diagnosis method based on the statistical distribution characteristics, the influences of the engine volume effect, the metal heat transfer and the measurement noise and offset can be reduced and even eliminated in the characteristic extraction process, and the fault diagnosis accuracy of the engine is improved.
With the continuous deepening and increasing use of the development of the engines, the obtained engine data are continuously increased, the advantage of fault diagnosis based on deviation statistical characteristics is more and more obvious, and the method becomes an important means in the gas circuit fault diagnosis of the aircraft engine.
In the description herein, reference to the description of the terms "one embodiment/mode," "some embodiments/modes," "example," "specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment/mode or example is included in at least one embodiment/mode or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to be the same embodiment/mode or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments/modes or examples. Furthermore, the various embodiments/aspects or examples and features of the various embodiments/aspects or examples described in this specification can be combined and combined by one skilled in the art without conflicting therewith.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
It will be understood by those skilled in the art that the foregoing embodiments are merely for clarity of illustration of the disclosure and are not intended to limit the scope of the disclosure. Other variations or modifications may occur to those skilled in the art, based on the foregoing disclosure, and are still within the scope of the present disclosure.

Claims (5)

1. An aeroengine gas circuit fault fusion diagnosis method based on statistical distribution characteristics is characterized by comprising the following steps:
data preprocessing: respectively carrying out data cleaning, similarity conversion, steady state interception and data intelligent tag adding operation on historical data of the engine to obtain steady state data with consistent import conditions and data intelligent tags;
constructing a reference model: taking data with a normal label in the steady-state data as input, and carrying out reference model modeling by establishing a mapping relation between main measurement parameters of the engine and controlled parameters of the engine to obtain an engine reference model so as to provide support for fault diagnosis;
extracting deviation statistical characteristics: based on the engine reference model, taking data with a fault label in the steady-state data as input, and acquiring a difference value between an engine measurement parameter and the engine reference model; then, sensitivity analysis is carried out according to the influence of the engine fault on the engine measurement parameters; extracting statistical characteristics based on deviation, converting engine measurement parameters into statistical distribution characteristics capable of supporting fault diagnosis, and constructing an engine statistical distribution characteristic library;
fusion diagnosis: carrying out statistical distribution characteristic matching of single engine parameters on an engine statistical distribution characteristic library, and carrying out fault diagnosis based on single measurement parameter deviation statistical characteristics; normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics; carrying out diagnosis result fusion on the fault diagnosis results of different measurement parameter statistical distribution characteristics by using a weighted D-S evidence theory method; optimizing and integrating the fusion diagnosis result to obtain a final diagnosis result.
2. The method for fusion diagnosis of the air circuit fault of the aircraft engine based on the statistical distribution characteristics as claimed in claim 1, wherein the data preprocessing comprises the following specific steps:
data cleaning: cleaning abnormal values in the historical data of the engine by adopting a mode based on median filtering and mean filtering;
and (4) similar conversion: by utilizing the similarity principle of the engine, the parameters of the sections of the runners of the engine are adjusted by aligning the temperature and the pressure of the inlet of the engine to the standard atmospheric condition, so that the influence of the inconsistency of the inlet height, the Mach number, the temperature and the pressure on the analysis result is eliminated;
and (3) steady state interception: setting a moving window with a certain length, intercepting steady-state section data according to the slope of a controlled parameter in the moving window, and extracting data of a period of time with a relatively stable engine state to obtain steady-state data;
data smart label: according to the statistical characteristics of the data of each steady-state section of the engine, classification of engine fault data and abnormal data is realized by adopting an unsupervised learning method, labels of the engine data are determined by adopting a C-means clustering algorithm, and corresponding labels are added to the clustered data according to the engine test run record, the hole detection and the issuing of the disassembled analysis result.
3. The method for fusion diagnosis of the air circuit fault of the aircraft engine based on the statistical distribution characteristics as claimed in claim 1, wherein the reference model is constructed by the following specific steps:
dividing the data with normal labels through the fan guide vane angle, the compressor guide vane angle and the tail nozzle area;
and solving the mean value of each steady-state section of different data sets to construct an interpolation table, and obtaining an engine reference model by establishing a mapping relation between main measurement parameters and controlled parameters of the engine.
4. The method for fusion diagnosis of the air circuit fault of the aircraft engine based on the statistical distribution characteristics as claimed in claim 1, wherein the specific steps of extracting the deviation statistical characteristics are as follows:
obtaining deviation: acquiring a difference value between engine measurement data and an engine reference model under the condition that parameters of a controlled rotating speed, a fan guide vane angle, a compressor guide vane angle and a tail nozzle area are aligned;
and (3) fault sensitivity analysis: for typical fault modes of m types of engines, the influence of each type of fault mode on n measurement parameters is different, and a sensitivity matrix SM of the engine measurement parameters on the fault modes is obtained by comparing the average relative change of the engine measurement parameters before and after the engine fault occurs;
Figure FDA0002874163030000021
in the formula, smijRepresenting the sensitivity of the jth parameter to the ith fault mode; i is 1,2, …, m; j is 1,2, …, n;
the confidence w of the diagnosis result of the pattern matching of the relative deviation statistic of different measurement parametersijComprises the following steps:
Figure FDA0002874163030000022
in the formula (I), the compound is shown in the specification,
Figure FDA0002874163030000023
the maximum value of the average relative change absolute value of the measurement parameter before and after the ith fault mode occurs is represented;
Figure FDA0002874163030000024
the minimum value of the average relative change absolute value of the measurement parameter before and after the ith fault mode occurs is represented; i is 1,2, …, m, epsilon1、ε2Represents two smaller numbers;
the confidence matrix for the engine measured parameters is then:
Figure FDA0002874163030000031
extracting statistical characteristics based on deviation: after the deviation of the engine measurement parameters is obtained, the characteristics of the mean value and the variance of the engine measurement parameters are counted, and an engine statistical distribution characteristic library is established.
5. The method for fusion diagnosis of the air circuit fault of the aircraft engine based on the statistical distribution characteristics as claimed in claim 1, wherein the fusion diagnosis comprises the following specific steps:
fault diagnosis based on single measurement parameter deviation statistical characteristics: matching the mean value and the variance of the deviation of the measured parameters to be diagnosed in the statistical distribution characteristic library of the engine with the mean value and the variance of the deviation of the measured parameters of the known fault mode;
let x be a measurement parameter, x ═ x1，x2，…xn]N represents the number of measured parameters, and the probability distribution function of the i-th fault mode is fi(xj) The probability distribution function of the data to be diagnosed is f (x)j) Wherein j is 1,2, …, n, xjIs in the range of [ -inf, inf]Inf represents infinity;
then the j-th measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijComprises the following steps:
Figure FDA0002874163030000032
normalizing the fault diagnosis result based on the single measurement parameter statistical distribution characteristics;
measuring the jth measurement parameter xjProbability p of determined data to be diagnosed belonging to class i fault modeijAnd (3) carrying out normalization:
Figure FDA0002874163030000033
the failure diagnosis result after normalization is:
Figure FDA0002874163030000034
in the formula, m represents the number of failure modes;
performing fusion diagnosis on the fault diagnosis results of different measurement parameter statistical distribution characteristics: calculating the average matching degree of the measured parameters of each engine, namely the average value of the diagnosis results of the measured parameters of different engines, and fusing the diagnosis results by using a weighted D-S evidence theory method;
weighting the diagnosis results of the different engine measurement parameters by the following formula to obtain weighted fault diagnosis results and uncertainty based on single measurement parameter feature matching as follows:
Figure FDA0002874163030000041
in the formula (I), the compound is shown in the specification,
Figure FDA0002874163030000042
represents the uncertainty of the diagnostic result of the jth measured parameter,
Figure FDA0002874163030000043
representing the weighted fault diagnosis, wijRepresents a weighting coefficient;
the weighted fault diagnosis results based on different parameters of different fault feature libraries are represented in a matrix form as follows:
Figure FDA0002874163030000044
in the formula, the last line represents the uncertainty of different measurement parameters;
synthesis of rule pairs Using DempsterMatrix array
Figure FDA0002874163030000045
M-1 fusion from left to right: first pair matrix
Figure FDA0002874163030000046
The first column and the second column are fused, then the result after the first fusion is fused with the third column of the matrix, the result after the second fusion is fused with the fourth column, and so on until the matrix is fused
Figure FDA0002874163030000047
All columns are fused until the completion;
optimizing and integrating fusion diagnosis results: firstly, the average probability of the occurrence of the fault of the data to be diagnosed is calculated, and the fusion diagnosis result is optimized on the basis to obtain the fault diagnosis result.
CN202011608883.6A 2020-12-30 2020-12-30 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics Active CN112613186B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011608883.6A CN112613186B (en) 2020-12-30 2020-12-30 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011608883.6A CN112613186B (en) 2020-12-30 2020-12-30 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics

Publications (2)

Publication Number Publication Date
CN112613186A true CN112613186A (en) 2021-04-06
CN112613186B CN112613186B (en) 2022-04-12

Family

ID=75249382

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011608883.6A Active CN112613186B (en) 2020-12-30 2020-12-30 Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics

Country Status (1)

Country Link
CN (1) CN112613186B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112989522A (en) * 2021-05-10 2021-06-18 创新奇智(成都)科技有限公司 Model training method, fault prediction method and device and electronic equipment
CN114676572A (en) * 2022-03-25 2022-06-28 中国航空发动机研究院 Parameter determination method and device and computer readable storage medium
CN114969645A (en) * 2022-05-24 2022-08-30 中国航空发动机研究院 Statistical distribution model-based aircraft engine state monitoring and evaluation method
CN115096595A (en) * 2022-06-30 2022-09-23 中国航空发动机研究院 Feature self-learning method, feature self-learning system and engine fault diagnosis method
CN115839848A (en) * 2023-02-22 2023-03-24 中国船舶集团有限公司第七一九研究所 Data acquisition and management method for ship diagnosis
CN116304538A (en) * 2023-05-19 2023-06-23 中国标准化研究院 Method for evaluating uncertainty of detection result by using big data

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040002810A1 (en) * 2002-07-01 2004-01-01 Syu Akuzawa Malfunction diagnosis system for engine
CN101206160A (en) * 2006-12-18 2008-06-25 联合工艺公司 Method and system for identifying gas turbine engine faults
KR101505975B1 (en) * 2014-05-30 2015-03-26 충남대학교산학협력단 Method and system for fault dignosis of engine
CN105911975A (en) * 2016-04-14 2016-08-31 南京航空航天大学 Integrated diagnosis method of gas path fault of aeroengine
CN108332970A (en) * 2017-11-17 2018-07-27 中国铁路总公司 A kind of Method for Bearing Fault Diagnosis based on LS-SVM and D-S evidence theory
CN108710756A (en) * 2018-05-18 2018-10-26 上海电力学院 The method for diagnosing faults of lower multicharacteristic information Weighted Fusion is analyzed based on spectral clustering
CN109857094A (en) * 2019-03-14 2019-06-07 杭州电子科技大学 Fault Diagnosis of Aeroengines method based on two-stage Kalman filter algorithm
CN110334395A (en) * 2019-05-28 2019-10-15 中国地质大学（武汉） The satellite momentum wheel fault diagnosis method and system of initialization EM algorithm based on JADE
CN111580506A (en) * 2020-06-03 2020-08-25 南京理工大学 Industrial process fault diagnosis method based on information fusion

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040002810A1 (en) * 2002-07-01 2004-01-01 Syu Akuzawa Malfunction diagnosis system for engine
CN101206160A (en) * 2006-12-18 2008-06-25 联合工艺公司 Method and system for identifying gas turbine engine faults
KR101505975B1 (en) * 2014-05-30 2015-03-26 충남대학교산학협력단 Method and system for fault dignosis of engine
CN105911975A (en) * 2016-04-14 2016-08-31 南京航空航天大学 Integrated diagnosis method of gas path fault of aeroengine
CN108332970A (en) * 2017-11-17 2018-07-27 中国铁路总公司 A kind of Method for Bearing Fault Diagnosis based on LS-SVM and D-S evidence theory
CN108710756A (en) * 2018-05-18 2018-10-26 上海电力学院 The method for diagnosing faults of lower multicharacteristic information Weighted Fusion is analyzed based on spectral clustering
CN109857094A (en) * 2019-03-14 2019-06-07 杭州电子科技大学 Fault Diagnosis of Aeroengines method based on two-stage Kalman filter algorithm
CN110334395A (en) * 2019-05-28 2019-10-15 中国地质大学（武汉） The satellite momentum wheel fault diagnosis method and system of initialization EM algorithm based on JADE
CN111580506A (en) * 2020-06-03 2020-08-25 南京理工大学 Industrial process fault diagnosis method based on information fusion

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张宗杰 等: "基于多分类器融合的航空发动机气路故障诊断", 《沈阳航空工业学院学报》 *
范满意: "基于数据调和的燃气轮机状态监测与故障诊断研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112989522A (en) * 2021-05-10 2021-06-18 创新奇智(成都)科技有限公司 Model training method, fault prediction method and device and electronic equipment
CN112989522B (en) * 2021-05-10 2021-07-30 创新奇智(成都)科技有限公司 Model training method, fault prediction method and device and electronic equipment
CN114676572A (en) * 2022-03-25 2022-06-28 中国航空发动机研究院 Parameter determination method and device and computer readable storage medium
CN114676572B (en) * 2022-03-25 2023-02-17 中国航空发动机研究院 Parameter determination method and device and computer readable storage medium
CN114969645A (en) * 2022-05-24 2022-08-30 中国航空发动机研究院 Statistical distribution model-based aircraft engine state monitoring and evaluation method
CN115096595A (en) * 2022-06-30 2022-09-23 中国航空发动机研究院 Feature self-learning method, feature self-learning system and engine fault diagnosis method
CN115839848A (en) * 2023-02-22 2023-03-24 中国船舶集团有限公司第七一九研究所 Data acquisition and management method for ship diagnosis
CN116304538A (en) * 2023-05-19 2023-06-23 中国标准化研究院 Method for evaluating uncertainty of detection result by using big data
CN116304538B (en) * 2023-05-19 2023-07-21 中国标准化研究院 Method for evaluating uncertainty of detection result by using big data

Also Published As

Publication number Publication date
CN112613186B (en) 2022-04-12

Similar Documents

Publication Publication Date Title
CN112613186B (en) Aero-engine gas circuit fault fusion diagnosis method based on statistical distribution characteristics
CN111175054B (en) Aeroengine fault diagnosis method based on data driving
CN108256173B (en) Gas circuit fault diagnosis method and system for dynamic process of aircraft engine
CN106021826B (en) One kind is based on aero-engine complete machine method for predicting residual useful life under operating mode's switch and the matched variable working condition of similitude
CN110018670B (en) Industrial process abnormal working condition prediction method based on dynamic association rule mining
US8682824B2 (en) Method and device for monitoring the state of a facility
CN110008253B (en) Industrial data association rule mining and abnormal working condition prediction method
CN112766303B (en) CNN-based aeroengine fault diagnosis method
CN114282579A (en) Aviation bearing fault diagnosis method based on variational modal decomposition and residual error network
CN109298633A (en) Chemical production process fault monitoring method based on adaptive piecemeal Non-negative Matrix Factorization
WO2024045246A1 (en) Spike echo state network model for aero engine fault prediction
CN115375026A (en) Method for predicting service life of aircraft engine in multiple fault modes
CN114757269A (en) Complex process refined fault detection method based on local subspace-neighborhood preserving embedding
CN112381965A (en) Aeroengine health state identification system and method based on data mining
CN115545092A (en) Aero-engine small sample fault diagnosis method based on deep twin self-attention network
WO2023097705A1 (en) Air compressor rotation stall prediction method based on multi-source data fusion
Viale et al. Least squares smoothed k-nearest neighbors online prediction of the remaining useful life of a NASA turbofan
CN110826600B (en) Engine surge prediction method based on adaptive resonance network online incremental learning
CN115618733B (en) Multi-scale hybrid attention mechanism modeling method for predicting remaining service life of aircraft engine
CN116028865A (en) Aeroengine fault diagnosis method based on feature amplification
CN111622815A (en) Blade crack online measurement method based on blade tip timing and naive Bayes optimization
Ren et al. A clustering-based method for health conditions evaluation of aero-engines
CN113836787A (en) Supersonic air inlet flow state monitoring method based on discriminative feature learning
CN110657031B (en) Method for identifying surge of aircraft engine
CN113465930B (en) Gas turbine multi-sensor fault detection method based on hybrid method

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