CN115828611A - Method and device for evaluating health state of electromechanical component of spacecraft - Google Patents

Method and device for evaluating health state of electromechanical component of spacecraft Download PDF

Info

Publication number
CN115828611A
CN115828611A CN202211616097.XA CN202211616097A CN115828611A CN 115828611 A CN115828611 A CN 115828611A CN 202211616097 A CN202211616097 A CN 202211616097A CN 115828611 A CN115828611 A CN 115828611A
Authority
CN
China
Prior art keywords
data
sample
parameters
health
parameter
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
CN202211616097.XA
Other languages
Chinese (zh)
Other versions
CN115828611B (en
Inventor
刘成瑞
刘磊
刘文静
王淑一
徐赫屿
梁寒玉
李文博
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.)
Beijing Institute of Control Engineering
Original Assignee
Beijing Institute of Control Engineering
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 Beijing Institute of Control Engineering filed Critical Beijing Institute of Control Engineering
Priority to CN202211616097.XA priority Critical patent/CN115828611B/en
Publication of CN115828611A publication Critical patent/CN115828611A/en
Application granted granted Critical
Publication of CN115828611B publication Critical patent/CN115828611B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • YGENERAL 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
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02PCLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30Computing systems specially adapted for manufacturing

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention relates to a method and a device for evaluating the health state of an electromechanical component of a spacecraft, wherein the method comprises the following steps: acquiring data to be detected of an electromechanical component of the spacecraft, wherein the data to be detected comprises data of parameters of the electromechanical component and overall parameters of the spacecraft; sequentially performing outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain a target feature of each parameter; the method comprises the steps of inputting target characteristics of all parameters into a trained health state evaluation model to obtain health factors corresponding to data to be tested, wherein the health factors are used for representing the health state of an electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking sample characteristics of the electromechanical component as input and sample health factors corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially carrying out outlier processing, characteristic extraction processing and standardization processing on sample data of the electromechanical component. The method can effectively evaluate the health state of the electromechanical component of the spacecraft.

Description

Method and device for evaluating health state of electromechanical component of spacecraft
Technical Field
The invention relates to the technical field of aerospace, in particular to a method and a device for evaluating the health state of an electromechanical component of a spacecraft.
Background
Electromechanical components play an important role in spacecrafts, and are key components for determining the functions and the performances of the spacecrafts. Due to the complex composition of the electromechanical components, the on-track failure rate is high, and especially due to the existence of the rotating assembly, a performance degradation process is generated. The health state of the electromechanical components has an important influence on the reliability of the spacecraft, and determines the service life of the spacecraft to a certain extent. Therefore, the health state of the electromechanical parts is evaluated based on the on-orbit data of the spacecraft, which is of great significance for ensuring the safe and stable operation of the spacecraft. However, it is difficult for the related art to effectively assess the health of the electromechanical components of the spacecraft.
Disclosure of Invention
In order to effectively evaluate the health state of the electromechanical component of the spacecraft, the embodiment of the specification provides a health state evaluation method and a health state evaluation device for the electromechanical component of the spacecraft.
In a first aspect, embodiments of the present specification provide a method for evaluating a state of health of an electromechanical component of a spacecraft, including:
acquiring data to be detected of spacecraft electromechanical components; the data to be detected comprises the data of the self parameters of the electromechanical parts and the overall parameters of the spacecraft;
sequentially performing outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain a target feature of each parameter;
inputting the target characteristics of all parameters into a trained health state evaluation model to obtain health factors corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and taking the sample health factor corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially performing outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
In one possible design, the electromechanical component includes at least one of a momentum wheel, a control moment gyro, an inertial attitude sensor, and a windsurfing board drive mechanism;
and/or the presence of a gas in the atmosphere,
the self-parameter comprises at least one of a current parameter, a temperature parameter and a rotating speed parameter;
and/or the presence of a gas in the gas,
the overall parameters include at least one of attitude parameters, orbit parameters, environmental parameters, and command parameters.
In one possible design, the features extracted by the feature extraction process include at least one of peak, mean, standard deviation, absolute mean, square root magnitude, and kurtosis.
In one possible design, the sample health factor is obtained by:
carrying out dimension reduction processing on the sample characteristics of each parameter to obtain one-dimensional fusion characteristics of each parameter;
performing dimension reduction processing on the one-dimensional fusion characteristics of all the parameters to obtain one-dimensional target characteristics corresponding to the sample data;
performing curve fitting on the one-dimensional target characteristics to obtain a sample fitting curve;
and taking the fitting value of the sample fitting curve as a sample health factor.
In one possible design, the dimensionality reduction process employs principal component analysis.
In one possible design, the curve fitting the one-dimensional target feature to obtain a sample fitting curve includes:
based on the variation trend of the one-dimensional target features, carrying out standardization processing on the one-dimensional target features to obtain one-dimensional standard features; wherein the trend of change comprises an ascending trend and a descending trend;
and performing curve fitting on the one-dimensional standard characteristics by adopting a least square method to obtain a sample fitting curve.
In one possible design, the curve fitting the one-dimensional standard features using a least squares method includes:
and performing piecewise fitting on the one-dimensional standard features by adopting the following formula:
Figure BDA0004001747220000021
in the formula (I), the compound is shown in the specification,
Figure BDA0004001747220000022
for the one-dimensional standard features, i belongs to (1,2, …, p), p is the length of each feature in the feature extraction processing process, p = floor (n/m), floor () is a downward integer function, n is the total number of sample data, m is the length of a set sliding window, a, b and c are fitting parameters, t is t i The on-orbit running time corresponding to the ith parameter in the one-dimensional standard characteristic is obtained;
assuming that the one-dimensional standard features are divided into s sections for fitting, a of the j (j is less than or equal to s) th section j Is determined by the following formula:
Figure BDA0004001747220000031
in the formula, a j-1 、b j-1 、c j-1 Respectively, the fitting parameter, t, of the previous segment of the curve stage_j Is the starting time corresponding to the j section;
fitting parameters b of each segment j 、c j Is calculated by the following formula:
Figure BDA0004001747220000032
in the formula, n j And the number of data contained in the j section for the one-dimensional standard feature.
In a second aspect, embodiments of the present specification further provide a state of health assessment apparatus for an electromechanical component of a spacecraft, including:
the acquisition module is used for acquiring data to be detected of the electromechanical component of the spacecraft; the data to be detected comprises the data of the self parameters of the electromechanical parts and the overall parameters of the spacecraft;
the processing module is used for sequentially carrying out outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain a target feature of each parameter;
the evaluation module is used for inputting the target characteristics of all parameters into a trained health state evaluation model to obtain health factors corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and taking the sample health factor corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially performing outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
In a third aspect, an embodiment of the present specification further provides an electronic device, which includes a memory and a processor, where the memory stores a computer program, and the processor executes the computer program to implement the method according to any embodiment of the present specification.
In a fourth aspect, the embodiments of the present specification further provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed in a computer, the computer program causes the computer to execute the method according to any one of the embodiments of the present specification.
The embodiment of the specification provides a method and a device for evaluating the health state of an electromechanical component of a spacecraft, and the method and the device consider that data influencing the electromechanical component of the spacecraft are not only parameters of the electromechanical component but also possibly related to overall parameters of the spacecraft, so that the data to be tested comprise the parameters of the electromechanical component and the data of the overall parameters of the spacecraft, and the accuracy of the evaluation of the health state of the electromechanical component is ensured; the data to be measured of each parameter are sequentially subjected to outlier processing, feature extraction processing and standardization processing, so that more accurate and easier-to-calculate target features can be obtained; the health state evaluation model is set to establish the mapping relation between the data and the health factors, so that the health factors corresponding to the data to be tested can be obtained. Therefore, the scheme can effectively evaluate the health state of the electromechanical components of the spacecraft.
Drawings
In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present specification, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for evaluating a state of health of an electromechanical component of a spacecraft, according to an embodiment of the present disclosure;
fig. 2 is a hardware architecture diagram of an electronic device provided in an embodiment of the present specification;
fig. 3 is a structural diagram of a health status evaluation apparatus for an electromechanical component of a spacecraft according to an embodiment of the present disclosure;
FIG. 4 is a flow diagram of outlier processing provided in an embodiment of the present description;
fig. 5 is a model diagram of a health status assessment model according to an embodiment of the present disclosure.
Detailed Description
In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure more clear, the technical solution in the embodiments of the present disclosure will be described clearly and completely with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are a part of the embodiments of the present disclosure, but not all embodiments, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts based on the embodiments of the present disclosure belong to the protection scope of the present disclosure.
Referring to fig. 1, an embodiment of the present specification provides a method for evaluating a state of health of an electromechanical component of a spacecraft, including:
step 100: acquiring data to be tested of spacecraft electromechanical components; the data to be detected comprise the self parameters of the electromechanical parts and the data of the overall parameters of the spacecraft;
step 102: sequentially carrying out outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain target features of each parameter;
step 104: inputting the target characteristics of all parameters into a trained health state evaluation model to obtain health factors corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and the sample health factor corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially carrying out outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
In the embodiment of the specification, in consideration of the fact that data influencing electromechanical components of the spacecraft are not only parameters of the electromechanical components, but also may be related to overall parameters of the spacecraft, the data to be measured comprise the parameters of the electromechanical components and the data of the overall parameters of the spacecraft, so that the accuracy of the health state evaluation of the electromechanical components is ensured; the data to be measured of each parameter are sequentially subjected to outlier processing, feature extraction processing and standardization processing, so that more accurate and easier-to-calculate target features can be obtained; the health state evaluation model is set to establish the mapping relation between the data and the health factors, so that the health factors corresponding to the data to be detected can be obtained. Therefore, the scheme can effectively evaluate the health state of the electromechanical components of the spacecraft.
The manner in which the various steps shown in fig. 1 are performed is described below.
With respect to step 100:
in some embodiments, the electromechanical component comprises at least one of a momentum wheel, a control moment gyro, an inertial attitude sensor, and a windsurfing board drive mechanism. The momentum wheel, the control moment gyroscope, the inertia attitude sensor and the like are core components of the attitude and orbit control system of the spacecraft, and the sailboard driving mechanism is a core component for energy supply of the spacecraft.
In some embodiments, the own parameter of the electromechanical component comprises at least one of a current parameter, a temperature parameter, and a rotational speed parameter.
In some embodiments, the overall parameter of the electromechanical component comprises at least one of an attitude parameter, a trajectory parameter, an environmental parameter, and a command parameter. The attitude parameters include three-axis attitude angles and three-axis attitude angular velocities, the orbit parameters include six orbit elements such as orbit inclination angles and eccentricity, the environment parameters include environment temperature, and the command parameters include three-axis control torque, engine jet duration and the like.
With respect to step 102:
the data of the spacecraft fluctuates in the acquisition and transmission processes, so that the data contains outliers, and the outliers are required to be processed firstly.
As shown in fig. 4, in some embodiments, outlier processing (i.e., culling) may be performed, for example, using a 3 σ criterion, specifically:
1) For the ith parameter x i (I is less than or equal to I), the length of a sliding window is set to be m, the starting point of a data window is set to be s, and the data in the sliding window can be expressed as
Figure BDA0004001747220000061
Averaging all data in the sliding window, i.e. the following formula:
Figure BDA0004001747220000062
in the formulaAnd I is the number of all the parameters,
Figure BDA0004001747220000063
as parameter x within the sliding window i Average value of (d);
2) And calculating the standard deviation of all data in the sliding window, namely the following formula:
Figure BDA0004001747220000064
in the formula (I), the compound is shown in the specification,
Figure BDA0004001747220000065
as parameter x within the sliding window i The standard deviation of (a);
3) If the following formula is satisfied:
Figure BDA0004001747220000066
then determine
Figure BDA0004001747220000067
The data are outliers and are removed; otherwise determining
Figure BDA0004001747220000068
Is a normal value; wherein j represents the jth data point in the window;
4) For data in the next sliding window
Figure BDA0004001747220000069
And repeating the steps 1) -3) until all data are processed.
It is understood that the data refers to data to be measured or sample data.
Of course, other methods may be adopted for outlier processing, which are not limited and described herein.
In some embodiments, the features extracted by the feature extraction process include at least one of a peak value, a mean value, a standard deviation, an absolute mean value, a square root magnitude, and a kurtosis. The characteristics are the characteristics which can reflect the degradation trend of the electromechanical components, so that the accurate health state evaluation model can be established in the follow-up process.
After outlier processing, a new sample X is obtained I×n =(x 1 ,x 2 ,…,x I ) (ii) a Wherein the content of the first and second substances,
Figure BDA00040017472200000610
(I is less than or equal to I). Respectively extracting the characteristics of I variables in the new sample, and for the ith parameter x i Setting the window length of feature extraction as m, wherein the calculation formula of the features is as follows:
peak value:
Figure BDA0004001747220000071
k=(j-1)*m+1,(j-1)*m+2,...,(j-1)*m+m
average value:
Figure BDA0004001747220000072
standard deviation:
Figure BDA0004001747220000073
absolute average value:
Figure BDA0004001747220000074
square root amplitude:
Figure BDA0004001747220000075
kurtosis:
Figure BDA0004001747220000076
in some embodiments, the normalization process may employ the (0,1) method (i.e., the Z-score method), specifically:
the features of each parameter are formed into a p x 6 matrix
Figure BDA0004001747220000077
Wherein
Figure BDA0004001747220000078
(j is less than or equal to 6) respectively corresponds to 6 feature quantities subjected to feature extraction, p is the length of each feature in the feature extraction processing process, p = floor (n/m), floor () is a downward rounding function, n is the total number of data, and m is the length of a set sliding window. The feature values of each column are normalized and converted into a matrix having a mean value of 0 and a standard deviation of 1
Figure BDA0004001747220000079
Namely:
Figure BDA00040017472200000710
wherein, the first and the second end of the pipe are connected with each other,
Figure BDA00040017472200000711
is the column average of the jth characteristic quantity of the parameter i,
Figure BDA00040017472200000712
column standard deviations.
Obtained as above
Figure BDA00040017472200000713
Is the target characteristic of parameter i, hence Y * (t)=[Y 1* (t),Y 2* (t),…,Y l* (t)]And t is the on-orbit running time length corresponding to the ith parameter.
With respect to step 104:
the following description focuses on the training process of the health factor assessment model.
Firstly, sample characteristics are obtained based on sample data, and the sample characteristics are obtained by sequentially performing outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component, which are not described herein again.
Secondly, a sample health factor is obtained based on the sample characteristics. In some embodiments, the sample health factor is obtained by:
carrying out dimension reduction processing on the sample characteristics of each parameter to obtain one-dimensional fusion characteristics of each parameter;
performing dimension reduction processing on the one-dimensional fusion characteristics of all the parameters to obtain one-dimensional target characteristics corresponding to the sample data;
performing curve fitting on the one-dimensional target characteristics to obtain a sample fitting curve;
and taking the fitting value of the sample fitting curve as a sample health factor.
In this embodiment, the one-dimensional target features corresponding to the sample data can be obtained by performing dimension reduction processing on the sample features of each parameter and the one-dimensional fusion features of all the parameters in sequence, so that data dimension reduction and data fusion are completed; and then, a sample fitting curve is obtained by performing curve fitting on the one-dimensional target characteristics, so that the health factor with better monotonicity and trend can be obtained.
In some embodiments, the dimension reduction process employs principal component analysis, specifically:
1) Computing
Figure BDA0004001747220000081
Covariance matrix of
Figure BDA0004001747220000082
c r,s =c s,r =cov(Dim r ,Dim s )
In the formula, dim r ,Dim s Respectively represent matrices
Figure BDA0004001747220000083
The element values of the r-th column and the s-th column are respectively used for calculating the covariance between every two rows to finally obtain a covariance matrix C i
2) To C i Performing characteristic decomposition to obtain a matrix characteristic value and a corresponding characteristic vector:
C i =U ii (U i ) -1
in the formula (E) i =diag(λ i 1 ,λ i 2 ,…,λ i 6 ) Expressed as a matrix of eigenvalues, U, obtained by decomposition i =(u i 1 ,u i 2 ,…,u i 6 ) Is a corresponding feature vector matrix;
3) The feature vector u j (j is less than or equal to 6) arranging the upper and lower characteristic values, and calculating the cumulative contribution ratio of the first q principal elements according to the following formula:
Figure BDA0004001747220000091
wherein q is selected to satisfy η i (q)>85%;
4) Forming a transformation matrix U by taking eigenvectors corresponding to the first q larger eigenvalues i* Calculating the characteristic quantity f after dimensionality reduction i =Y i* *U i* (ii) a Wherein, the weight vector w is designed according to the proportion of the first q characteristic values i =[w1,w2,w3,…wq],
Figure BDA0004001747220000092
5) Weight vector w according to design i Calculating the one-dimensional fusion characteristic after the fusion of the characteristic of the parameter i
Figure BDA0004001747220000093
6) Repeating the operation to obtain the one-dimensional fusion characteristic F of all the parameters i (I ≦ I), form the feature matrix F = (F) for all parameters 1 ,F 2 ,…,F I )。
7) For the characteristic matrix F obtained in the step 6), reducing the dimension of the I-dimension characteristic quantity by using a principal component analysis method again, repeating the operations of the steps 1) -5), reducing the dimension of the characteristic and fusing to finally obtain the one-dimensional target characteristic F *
Of course, the dimension reduction process may also adopt other methods, such as singular value decomposition, which are not specifically limited and described herein.
It should be noted that, because the on-orbit data of the spacecraft has a high-dimensional characteristic, for example, the data related to the health state of the electromechanical component not only includes parameters of the component itself, such as current, temperature, and rotation speed, but also includes parameters of the spacecraft, such as attitude, orbit, environment, and command. The high-dimensional data are mutually coupled, and any single data cannot completely reflect the health state of the electromechanical component, so that the high-dimensional data are fused to realize accurate evaluation of the health state.
However, the related art cannot mine the internal relationship between the high-dimensional data, and it is difficult to establish the mapping relationship between the high-dimensional on-rail data and the health state of the electromechanical component, so that a large error exists between the health state analysis result of the electromechanical component and the actual situation.
In order to solve the above problems, the inventors found in the development process that: the method can realize the fusion and association relationship mining of the on-orbit high-dimensional data and the establishment of the complex mapping relationship between the high-dimensional data and the health state of the electromechanical component, thereby realizing the accurate evaluation of the health state of the electromechanical component. The method comprises the steps that through the data dimension reduction capability of principal component analysis, the characteristics which can reflect the degradation trend of electromechanical parts in high-dimensional data are extracted, and an accurate health state evaluation model is established; by utilizing the nonlinear fitting capability of the deep neural network, the mapping relation from high-dimensional data to health factors is established, and the accurate evaluation of the health state of the electromechanical component is realized.
In some embodiments, the step of "performing curve fitting on the one-dimensional target feature to obtain a sample fitting curve" may specifically include:
based on the variation trend of the one-dimensional target features, carrying out standardization processing on the one-dimensional target features to obtain one-dimensional standard features; wherein the variation trend comprises an ascending trend and a descending trend;
and performing curve fitting on the one-dimensional standard characteristics by adopting a least square method to obtain a sample fitting curve.
In this embodiment, the one-dimensional standard features are subjected to the curve fitting by performing the normalization processing on the one-dimensional target features and adopting the least square method, which is beneficial to obtaining the health factor with better monotonicity and trend.
Specifically, for the acquired one-dimensional target feature F * The normalization process between (0,1) is performed to obtain the feature quantity
Figure BDA0004001747220000101
Wherein:
if the one-dimensional object feature F * In the ascending trend, then:
Figure BDA0004001747220000102
if the one-dimensional object feature F * In a downward trend, then:
Figure BDA0004001747220000103
in the formula (I), the compound is shown in the specification,
Figure BDA0004001747220000104
is F * The maximum value of (a) is,
Figure BDA0004001747220000105
is F * The minimum value of (d).
In some embodiments, the step of "curve fitting the one-dimensional standard feature by using the least square method" may specifically include:
and (3) performing piecewise fitting on the one-dimensional standard features by adopting the following formula:
Figure BDA0004001747220000106
in the formula (I), the compound is shown in the specification,
Figure BDA0004001747220000107
for one-dimensional standard features, i belongs to (1,2, …, p), p is the length of each feature in the feature extraction process, p = floor (n/m), floor () is a downward integer function, n is the total number of sample data, m is the set sliding window length, a, b and c are fitting parameters, t is t i The on-orbit operation time corresponding to the ith parameter in the one-dimensional standard characteristic is obtained;
assuming that the one-dimensional standard features are divided into s sections for fitting, a of the j (j is less than or equal to s) th section j Is determined by the following formula:
Figure BDA0004001747220000108
in the formula, a j-1 、b j-1 、c j-1 Respectively, the fitting parameter, t, of the previous segment of the curve stage_j Is the starting time corresponding to the j section;
fitting parameters b of each segment j 、c j Is calculated by the following formula:
Figure BDA0004001747220000111
in the formula, n j The number of data contained in the j-th segment for a one-dimensional standard feature.
In this embodiment, a health factor curve with good monotonicity and tendency can be obtained by performing least square fitting on the one-dimensional standard features.
Obtaining characteristic quantity
Figure BDA0004001747220000112
And after the curve is fitted, the fitting value of the sample fitting curve is used as a sample health factor (namely the construction of the input sample and the output result is completed) for the next neural network training.
In some embodiments, a fuzzy neural network may be used to construct a complex mapping from target features to health factors, and the specific implementation steps are as follows:
establishing a relationship between the health factor and the input variable using a fuzzy neural network may be expressed as:
HI(t)=f(Y * (t))
in the formula, HI (t) is a health factor at the time t; y is * (t) is a feature value after normalization at time t, Y * (t)=[Y 1* (t),Y 2* (t),…,y l* (t)],
Figure BDA0004001747220000113
f () is a non-linear function between the health factor and the input variable, approximated by a fuzzy neural network.
As shown in fig. 5, the high-dimensional data feature-health factor mapping network comprises four layers, the first layer is an input layer, and the number of hidden layers is determined by the sum of the number of input variables; the second layer is a radial basis function layer, the main function of the second layer is to fuzzify input variables, and the number of hidden layers is mostly designed based on artificial experience; the third layer is a regular layer, the main function of the third layer is defuzzification, and the number of hidden layers is the same as that of the radial basis function layers; the fourth layer is the output layer whose main function is to compute the health factor in the form of a linear sum.
The abnormal symptom evolution model based on the fuzzy neural network can be specifically described as follows:
HI(t)=ω(t)v T (t)
where HI (t) is the predicted output of the health factor at time t, ω (t) = [ ω = [ ω = ] 1 (t),ω 2 (t),…,ω 10 (t)]Is a weight vector between the regular layer and the output layer, v (t) = [ v = 1 (t),v 2 (t),…,ω 10 (t)]Is time tThe output of the third rule layer is,
Figure BDA0004001747220000114
wherein vl (t) is the ith neuron output of the regular layer at the time t, thetaj (t) is the jth neuron output of the radial basis function layer of the second layer at the time t,
Figure BDA0004001747220000121
Figure BDA0004001747220000122
wherein x (t) = [ x ] 1 (t),x 2 (t),…,x n (t)]Is the first input, where x (T) = Y * (t), n =6*l denotes the number of input layer neurons, c j (t)=[c 1j (t),c 2j (t),…,c nj (t)]Is the center of the jth neuron of the radial basis function layer at time t, σ j (t)=[σ 1j (t),σ 2j (t),…,σ nj (t)]T is the width of the jth neuron of the radial basis function layer at time T.
The output of the fourth output layer obtained by the weighting factor method is described as follows:
Figure BDA0004001747220000123
to realize the health factor prediction model parameters ω (t), c (t) (c (t) = [ c) based on the fuzzy neural network 1 (t),c 2 (t),…,c 10 (t)]) And σ (t) (σ (t) = [ σ) ] 1 (t),σ 2 (t),…,σ 10 (t)]) The self-adaptive adjustment of the method designs a second-order L-M-based parameter updating algorithm, and the algorithm can realize the simultaneous adjustment of model parameters, accelerate the calculation speed and ensure the prediction of the constructed abnormal symptom evolution model based on the health factorsAnd (4) precision. The update formula of the model parameters can be expressed as:
Ξ(t+1)=Ξ(t)+(H(t)+κ(t)I) -1 ·G(t)
wherein xi (t) is a parameter vector at time t, xi (t) = [ omega (t), c 1 (t),c 2 (t),…,c 10 (t),σ 1 (t),σ 2 (t),…,σ 10 (t)]The calculation process of H (t) is as follows:
H(t)=J T (t)J(t)
j (t) is the Jacobian vector at time t:
Figure BDA0004001747220000124
Figure BDA0004001747220000125
Figure BDA0004001747220000126
Figure BDA0004001747220000127
Figure BDA0004001747220000128
Figure BDA0004001747220000131
e (t) is the prediction error of HI at time t:
e(t)=HI′(t)-HI(t)
where HI' (t) is the actual output of HI at time t, I (t) is the identity matrix used to avoid ill-conditioned cases in the matrix inversion, and G (t) is the gradient vector:
G(t)=J T (t)e(t)
in the formula, κ (t) is an adaptive learning rate at time t, and is used to improve the convergence rate of the health factor prediction model, and is calculated as follows:
Figure BDA0004001747220000132
where ξ (t) is the adjustment coefficient of the learning rate at time t, ε min (t) is the minimum eigenvalue, ε, of H (t) max (t) is the maximum eigenvalue of H (t), 0<ε min (t)<ε max (t),0<κ(t)<1。
The above is the whole training process of the health factor evaluation model.
As shown in fig. 2 and 3, embodiments of the present specification provide a state of health assessment apparatus for electromechanical components of a spacecraft. The device embodiments may be implemented by software, or by hardware, or by a combination of hardware and software. From a hardware aspect, as shown in fig. 2, for a hardware architecture diagram of an electronic device in which a health status evaluation apparatus for an electromechanical component of a spacecraft is provided in an embodiment of the present disclosure, in addition to the processor, the memory, the network interface, and the nonvolatile memory shown in fig. 2, the electronic device in which the apparatus is located may generally include other hardware, such as a forwarding chip responsible for processing a message, and the like. Taking a software implementation as an example, as shown in fig. 3, as a logical device, a CPU of the electronic device reads a corresponding computer program in the non-volatile memory into the memory for running.
As shown in fig. 3, the present embodiment provides a health status evaluation apparatus for an electromechanical component of a spacecraft, including:
the acquisition module 300 is used for acquiring data to be detected of the electromechanical component of the spacecraft; the data to be detected comprises the data of the self parameters of the electromechanical parts and the overall parameters of the spacecraft;
the processing module 302 is configured to perform outlier processing, feature extraction processing, and normalization processing on the to-be-detected data of each parameter in sequence to obtain a target feature of each parameter;
the evaluation module 304 is configured to input the target features of all the parameters into the trained health state evaluation model to obtain a health factor corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and taking the sample health factors corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially carrying out outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
In this embodiment, the obtaining module 300 may be configured to perform the step 100 in the above method embodiment, the processing module 302 may be configured to perform the step 102 in the above method embodiment, and the evaluating module 304 may be configured to perform the step 104 in the above method embodiment.
In one embodiment of the present description, the electromechanical component comprises at least one of a momentum wheel, a control moment gyro, an inertial attitude sensor, and a windsurfing board drive mechanism;
and/or the presence of a gas in the gas,
the self-parameter comprises at least one of a current parameter, a temperature parameter and a rotating speed parameter;
and/or the presence of a gas in the gas,
the overall parameters include at least one of attitude parameters, orbit parameters, environmental parameters, and command parameters.
In one embodiment of the present specification, the feature extracted by the feature extraction process includes at least one of a peak value, a mean value, a standard deviation, an absolute mean value, a square root amplitude, and a kurtosis.
In one embodiment of the present description, the sample health factor is obtained by:
carrying out dimension reduction processing on the sample characteristics of each parameter to obtain one-dimensional fusion characteristics of each parameter;
performing dimension reduction processing on the one-dimensional fusion characteristics of all the parameters to obtain one-dimensional target characteristics corresponding to the sample data;
performing curve fitting on the one-dimensional target characteristics to obtain a sample fitting curve;
and taking the fitting value of the sample fitting curve as a sample health factor.
In one embodiment of the present specification, the dimensionality reduction process employs a principal component analysis.
In an embodiment of the present specification, the performing curve fitting on the one-dimensional target feature to obtain a sample fitting curve includes:
based on the variation trend of the one-dimensional target features, carrying out standardization processing on the one-dimensional target features to obtain one-dimensional standard features; wherein the trend of change comprises an ascending trend and a descending trend;
and performing curve fitting on the one-dimensional standard characteristics by adopting a least square method to obtain a sample fitting curve.
In an embodiment of the present specification, the curve fitting the one-dimensional standard feature by using a least square method includes:
and performing segmentation fitting on the one-dimensional standard features by adopting the following formula:
Figure BDA0004001747220000151
in the formula (I), the compound is shown in the specification,
Figure BDA0004001747220000152
for the one-dimensional standard features, i belongs to (1,2, …, p), p is the length of each feature in the feature extraction processing process, p = floor (n/m), floor () is a downward integer function, n is the total number of sample data, m is the length of a set sliding window, a, b and c are fitting parameters, t is t i The on-orbit running time corresponding to the ith parameter in the one-dimensional standard characteristic is obtained;
assuming that the one-dimensional standard features are divided into s sections for fitting, a of the j (j is less than or equal to s) th section j Is determined by the following formula:
Figure BDA0004001747220000153
in the formula, a j-1 、b j-1 、c j-1 Respectively, the fitting parameter, t, of the previous segment of the curve stage_j Is the starting time corresponding to the jth section;
fitting parameters b of each segment j 、c j Is calculated by the following formula:
Figure BDA0004001747220000154
in the formula, n j And the number of data contained in the j section for the one-dimensional standard feature.
It is to be understood that the illustrated structure of the embodiments of the present description does not constitute a specific limitation of a state of health assessment apparatus for spacecraft electromechanical components. In other embodiments of the present description, a state of health assessment apparatus for spacecraft electromechanical components may include more or fewer components than illustrated, or combine certain components, or split certain components, or a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.
For the information interaction, execution process, and other contents between the modules in the apparatus, the specific contents may refer to the description in the method embodiment of the present specification because the same concept is based on the method embodiment of the present specification, and are not described herein again.
An embodiment of the present specification further provides an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor, when executing the computer program, implements a method for evaluating a health state of an electromechanical component of a spacecraft in any embodiment of the present specification.
Embodiments of the present description further provide a computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, causes the processor to perform a method of health assessment of an electromechanical component of a spacecraft as in any of the embodiments of the present description.
Specifically, a system or an apparatus equipped with a storage medium on which software program codes that realize the functions of any of the above-described embodiments are stored may be provided, and a computer (or a CPU or MPU) of the system or the apparatus is caused to read out and execute the program codes stored in the storage medium.
In this case, the program code itself read from the storage medium can realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code constitute a part of this specification.
Examples of the storage medium for supplying the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD + RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer via a communications network.
Further, it should be clear that the functions of any one of the above-described embodiments may be implemented not only by executing the program code read out by the computer, but also by causing an operating system or the like operating on the computer to perform a part or all of the actual operations based on instructions of the program code.
Further, it is to be understood that the program code read out from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then a CPU or the like mounted on the expansion board or the expansion module is caused to perform part or all of the actual operations based on instructions of the program code, thereby realizing the functions of any of the embodiments described above.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solutions of the present specification, and not to limit them; although the present description has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the spirit and scope of the technical solutions of the embodiments of the present specification.

Claims (10)

1. A method of assessing the state of health of an electromechanical component of a spacecraft, comprising:
acquiring data to be tested of spacecraft electromechanical components; the data to be detected comprise the data of the self parameters of the electromechanical parts and the overall parameters of the spacecraft;
sequentially performing outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain a target feature of each parameter;
inputting the target characteristics of all parameters into a trained health state evaluation model to obtain health factors corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and taking the sample health factor corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially performing outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
2. The method of claim 1, wherein the electromechanical components include at least one of a momentum wheel, a control moment gyro, an inertial attitude sensor, and a windsurfing board drive mechanism;
and/or the presence of a gas in the gas,
the self-parameter comprises at least one of a current parameter, a temperature parameter and a rotating speed parameter;
and/or the presence of a gas in the gas,
the overall parameters include at least one of attitude parameters, orbit parameters, environmental parameters, and command parameters.
3. The method of claim 1, wherein the features extracted by the feature extraction process include at least one of peak, mean, standard deviation, absolute mean, square root magnitude, and kurtosis.
4. The method of any one of claims 1-3, wherein the sample health factor is obtained by:
carrying out dimension reduction processing on the sample characteristics of each parameter to obtain one-dimensional fusion characteristics of each parameter;
performing dimension reduction processing on the one-dimensional fusion characteristics of all the parameters to obtain one-dimensional target characteristics corresponding to the sample data;
performing curve fitting on the one-dimensional target characteristics to obtain a sample fitting curve;
and taking the fitting value of the sample fitting curve as a sample health factor.
5. The method of claim 4, wherein the dimension reduction process employs principal component analysis.
6. The method of claim 4, wherein said curve fitting said one-dimensional target feature to obtain a sample fitting curve comprises:
based on the variation trend of the one-dimensional target features, carrying out standardization processing on the one-dimensional target features to obtain one-dimensional standard features; wherein the trend of change comprises an ascending trend and a descending trend;
and performing curve fitting on the one-dimensional standard characteristics by adopting a least square method to obtain a sample fitting curve.
7. The method of claim 6, wherein said curve fitting said one-dimensional standard features using a least squares method comprises:
and performing piecewise fitting on the one-dimensional standard features by adopting the following formula:
Figure FDA0004001747210000021
in the formula (I), the compound is shown in the specification,
Figure FDA0004001747210000022
for the one-dimensional standard features, i belongs to (1,2, …, p), p is the length of each feature in the feature extraction processing process, p = floor (n/m), floor () is a downward integer function, n is the total number of sample data, m is the length of a set sliding window, a, b and c are fitting parameters, t is i The on-orbit running time corresponding to the ith parameter in the one-dimensional standard characteristic is obtained;
assuming that the one-dimensional standard features are divided into s sections for fitting, a of the j (j is less than or equal to s) th section j Is determined by the following formula:
Figure FDA0004001747210000023
in the formula, a j-1 、b j-1 、c j-1 Respectively, the fitting parameter, t, of the previous segment of the curve stage_j Is the starting time corresponding to the j section;
fitting parameters b of each segment j 、c j Is calculated by the following formula:
Figure FDA0004001747210000024
in the formula, n j And the number of data contained in the j section for the one-dimensional standard feature.
8. A state of health assessment device for an electromechanical component of a spacecraft, comprising:
the acquisition module is used for acquiring data to be detected of the spacecraft electromechanical component; the data to be detected comprise the data of the self parameters of the electromechanical parts and the overall parameters of the spacecraft;
the processing module is used for sequentially carrying out outlier processing, feature extraction processing and standardization processing on the data to be detected of each parameter to obtain the target feature of each parameter;
the evaluation module is used for inputting the target characteristics of all the parameters into a trained health state evaluation model to obtain the health factors corresponding to the data to be tested; the health factor is used for representing the health state of the electromechanical component, the health state evaluation model is obtained by training a preset neural network by taking the sample characteristics of the electromechanical component as input and taking the sample health factor corresponding to the sample characteristics as output, and the sample characteristics are obtained by sequentially performing outlier processing, characteristic extraction processing and standardization processing on the sample data of the electromechanical component.
9. An electronic device comprising a memory in which a computer program is stored and a processor which, when executing the computer program, carries out the method according to any one of claims 1-7.
10. A computer-readable storage medium, on which a computer program is stored which, when executed in a computer, causes the computer to carry out the method of any one of claims 1-7.
CN202211616097.XA 2022-12-15 2022-12-15 Health state evaluation method and device for spacecraft electromechanical component Active CN115828611B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211616097.XA CN115828611B (en) 2022-12-15 2022-12-15 Health state evaluation method and device for spacecraft electromechanical component

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211616097.XA CN115828611B (en) 2022-12-15 2022-12-15 Health state evaluation method and device for spacecraft electromechanical component

Publications (2)

Publication Number Publication Date
CN115828611A true CN115828611A (en) 2023-03-21
CN115828611B CN115828611B (en) 2023-08-01

Family

ID=85547476

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211616097.XA Active CN115828611B (en) 2022-12-15 2022-12-15 Health state evaluation method and device for spacecraft electromechanical component

Country Status (1)

Country Link
CN (1) CN115828611B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116401137A (en) * 2023-06-06 2023-07-07 中诚华隆计算机技术有限公司 Core particle health state prediction method and device, electronic equipment and storage medium
CN116449135A (en) * 2023-04-19 2023-07-18 北京航空航天大学 Method and system for determining health state of electromechanical system component and electronic equipment
CN117828517A (en) * 2024-03-06 2024-04-05 北京开运联合信息技术集团股份有限公司 Spacecraft on-orbit running state evaluation method based on data mining

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112034789A (en) * 2020-08-25 2020-12-04 国家机床质量监督检验中心 Health assessment method, system and assessment terminal for key parts and complete machine of numerical control machine tool
CN113722985A (en) * 2021-08-12 2021-11-30 武汉科技大学 Method and system for evaluating health state and predicting residual life of aircraft engine

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112034789A (en) * 2020-08-25 2020-12-04 国家机床质量监督检验中心 Health assessment method, system and assessment terminal for key parts and complete machine of numerical control machine tool
CN113722985A (en) * 2021-08-12 2021-11-30 武汉科技大学 Method and system for evaluating health state and predicting residual life of aircraft engine

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116449135A (en) * 2023-04-19 2023-07-18 北京航空航天大学 Method and system for determining health state of electromechanical system component and electronic equipment
CN116449135B (en) * 2023-04-19 2024-01-30 北京航空航天大学 Method and system for determining health state of electromechanical system component and electronic equipment
CN116401137A (en) * 2023-06-06 2023-07-07 中诚华隆计算机技术有限公司 Core particle health state prediction method and device, electronic equipment and storage medium
CN116401137B (en) * 2023-06-06 2023-09-26 中诚华隆计算机技术有限公司 Core particle health state prediction method and device, electronic equipment and storage medium
CN117828517A (en) * 2024-03-06 2024-04-05 北京开运联合信息技术集团股份有限公司 Spacecraft on-orbit running state evaluation method based on data mining

Also Published As

Publication number Publication date
CN115828611B (en) 2023-08-01

Similar Documents

Publication Publication Date Title
CN115828611A (en) Method and device for evaluating health state of electromechanical component of spacecraft
CN107944648B (en) Large ship speed and oil consumption rate prediction method
CN110046376B (en) Multi-working-condition health assessment method for satellite attitude control system based on Bayesian network
Spielberg et al. Neural network model predictive motion control applied to automated driving with unknown friction
US6954744B2 (en) Combinatorial approach for supervised neural network learning
Saviolo et al. Physics-inspired temporal learning of quadrotor dynamics for accurate model predictive trajectory tracking
Yoon et al. Optimal PID control for hovering stabilization of quadcopter using long short term memory
CN114357872A (en) Ship motion black box identification modeling and motion prediction method based on stacking model fusion
Løver et al. Explainable AI methods on a deep reinforcement learning agent for automatic docking
CN113485443B (en) Unmanned aerial vehicle control method based on deep learning, storage medium and equipment
Xin et al. Accelerated inverse reinforcement learning with randomly pre-sampled policies for autonomous driving reward design
Brüdigam et al. Structure-preserving learning using Gaussian processes and variational integrators
CN116011109B (en) Spacecraft service life prediction method and device, electronic equipment and storage medium
Pinguet et al. A Neural Autopilot Training Platform based on a Matlab and X-Plane co-simulation
Chen et al. Adaptive fuzzy PD+ control for attitude maneuver of rigid spacecraft
Wang et al. A data driven method of feedforward compensator optimization for autonomous vehicle control
CN114083543A (en) Active fault diagnosis method for space manipulator
CN115373411A (en) Decision-making method and system for airplane autopilot control strategy
CN115453880A (en) Training method of generative model for state prediction based on antagonistic neural network
CN115289917A (en) Rocket substage landing real-time optimal guidance method and system based on deep learning
CN114610039A (en) Robot control method, device, robot and storage medium
Li et al. Morphing Strategy Design for UAV based on Prioritized Sweeping Reinforcement Learning
Rahimi et al. Fault isolation of reaction wheels onboard 3-axis controlled in-orbit satellite using ensemble machine learning techniques
Xia et al. Data association-based fault diagnosis of IMUs: Optimized DBN design and wheeled robot evaluation
Gao Soft computing methods for control and instrumentation

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