CN114034332B - Fault monitoring method for weapon equipment system - Google Patents
Fault monitoring method for weapon equipment system Download PDFInfo
- Publication number
- CN114034332B CN114034332B CN202111394862.3A CN202111394862A CN114034332B CN 114034332 B CN114034332 B CN 114034332B CN 202111394862 A CN202111394862 A CN 202111394862A CN 114034332 B CN114034332 B CN 114034332B
- Authority
- CN
- China
- Prior art keywords
- matrix
- output
- input
- sensing data
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 72
- 238000012544 monitoring process Methods 0.000 title claims abstract description 34
- 239000011159 matrix material Substances 0.000 claims abstract description 164
- 238000012545 processing Methods 0.000 claims abstract description 23
- 238000013507 mapping Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims description 38
- 239000013598 vector Substances 0.000 claims description 25
- 238000000605 extraction Methods 0.000 claims description 13
- 230000009467 reduction Effects 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 238000009826 distribution Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000000470 constituent Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 16
- 238000012360 testing method Methods 0.000 description 11
- 230000002159 abnormal effect Effects 0.000 description 9
- 230000009286 beneficial effect Effects 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000011895 specific detection Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 235000015842 Hesperis Nutrition 0.000 description 1
- 235000012633 Iberis amara Nutrition 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000007795 chemical reaction product Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000012367 process mapping Methods 0.000 description 1
- 239000000376 reactant Substances 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D18/00—Testing or calibrating apparatus or arrangements provided for in groups G01D1/00 - G01D15/00
Abstract
The invention discloses a fault monitoring method for a weapon equipment system, which comprises the following steps: s1, acquiring sensing data of a weapon equipment system, and performing linear standardization processing on the sensing data to obtain an input standard sensing data set and an output standard sensing data set; s2, performing high-dimensional mapping and nonlinear processing on the input standard sensing data set and the output standard sensing data set to obtain a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix; s3, calculating an input score matrix, an output score matrix, an input load matrix and an output load matrix; s4, collecting sensing data of a weapon equipment system to be detected, obtaining sensing data to be detected, and performing fault monitoring on the sensing data to be detected according to an input score matrix, an output score matrix, an input load matrix and an output load matrix; the invention solves the problem of inaccurate fault judgment in the existing fault monitoring method of the weapon equipment system.
Description
Technical Field
The invention relates to the field of quality monitoring of weapon equipment systems, in particular to a fault monitoring method for a weapon equipment system.
Background
With the development of modern technology, the system structure of large complex equipment such as rockets and missiles is more complex, and the system structure cannot be described by a simple physical model and a mechanism model, so that a series of fault diagnosis methods based on fault trees and physical models are difficult to apply. With the development of integrated circuits, a large number of sensors are placed in many critical locations for equipment testing, while obtaining a vast amount of data. Therefore, the data-driven modeling-based method is gradually becoming a research hotspot and is greatly developed and applied. In the data obtained by the industrial process, the characteristics of high coupling, non-Gaussian, non-linearity and the like are usually adopted, and how to propose a characteristic modeling of the process from the data is a difficult point of a data driving method. Meanwhile, when the performance of equipment is degraded, parameters in the testing process drift or even a certain link fails, abnormal changes can occur to process data monitored by a corresponding sensor, and how to find out the reasons of the faults from the changes is the key point and the difficulty of the current research.
In a nonlinear process, the measured values of the process variables typically vary non-linearly, and the process variables also tend to be non-linearly related. Aiming at the process, the traditional multivariate statistical analysis method applied to the stable process is difficult to effectively monitor the process, and can have larger false alarm and missing alarm situations. Meanwhile, in nonlinear process monitoring considering key performance indexes, key index variables also change in a nonlinear manner, and the traditional nonlinear partial least squares method is difficult to be applied to an input-output nonlinear process because only the process variables are considered to change in a nonlinear manner. To study the relationship between nonlinear input and nonlinear output, nonlinear features are extracted, and an input-output nonlinear model is proposed. Based on the constructed model, nonlinear quality-related and quality-independent subspaces are partitioned and statistics are established in the respective subspaces, respectively. An online nonlinear process monitoring strategy is also designed, comprising offline construction control limits, online process mapping projection construction statistics, and real-time monitoring of online samples.
Disclosure of Invention
Aiming at the defects in the prior art, the fault monitoring method for the weapon equipment system solves the problem that the existing fault monitoring method for the weapon equipment system is inaccurate in fault judgment.
In order to achieve the aim of the invention, the invention adopts the following technical scheme: a fault monitoring method for a weapon equipment system, comprising the steps of:
s1, acquiring sensing data of a weapon equipment system, and performing linear standardization processing on the sensing data to obtain an input standard sensing data set and an output standard sensing data set;
s2, performing high-dimensional mapping and nonlinear processing on the input standard sensing data set and the output standard sensing data set to obtain a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix;
s3, calculating an input score matrix, an output score matrix, an input load matrix and an output load matrix according to the nonlinear input matrix, the nonlinear output matrix, the input parameter matrix and the output parameter matrix;
s4, collecting sensing data of the weapon equipment system to be detected, obtaining sensing data to be detected, and performing fault monitoring on the sensing data to be detected according to the input score matrix, the output score matrix, the input load matrix and the output load matrix.
Further, the types of the sensing data in step S1 include: voltage current data, barometric pressure data, vibration data, temperature data, servo speed data, servo rotational speed data, and servo feedback voltage data, wherein the voltage current data, barometric pressure data, vibration data, and temperature data are used as input data set X, and the servo speed data, the servo rotational speed data, and the servo feedback voltage data are used as output data set Y.
Further, the formula for performing the linear normalization processing on the sensing data in the step S1 is as follows:
wherein ,standard sensed data for j-th sampled sensed data for a corresponding i-th variable in input dataset X, X i，j For the jth sample of the sensing data for the ith variable in the input dataset X,y i，j sampling the sensor data for the j-th time for the i-th variable in the output data set Y, +.>For outputting standard sensing data of the j-th sampling sensing data of the corresponding i-th variable in the data set Y, n is the sampling frequency of each variable, < +.>The set is input standard sensing data set, < ->The set of constituents is an output standard sensing dataset.
The beneficial effects of the above-mentioned further scheme are: unifying the dimension and the scale of the different scale data obtained by measuring by different sensors, avoiding the loss of information in the data dimension reduction process, and further ensuring the abnormal representation of fault data statistics in the fault detection of the complex equipment system (the abnormal representation is the abnormal change of the fault data dimension reduction construction statistics compared with the normal data statistics).
Further, step S2 includes the following sub-steps:
s21, mapping the input standard sensing data set and the output standard sensing data set to a high-dimensional space respectively to obtain a high-dimensional input matrix phi x And a high-dimensional output matrix phi y ；
S22, for high-dimensional input matrix phi x And a high-dimensional output matrix phi y Nonlinear processing is carried out to obtain a nonlinear input matrixAnd a nonlinear output matrix->
S23, according to the nonlinear input matrixAnd a nonlinear output matrix->Calculating input parameter matrix +.>And output parameter matrix->
Further, the nonlinear input matrix in step S22And a nonlinear output matrix->The calculation formula of (2) is as follows:
the input parameter matrix is calculated in the step S23And output parameter matrix->The formula of (2) is:
wherein ,1n Is thatIs the full 1 column vector of>Is real space.
The beneficial effects of the above-mentioned further scheme are: by mapping the original data monitored in the nonlinear process of the complex equipment to high dimensions, the nonlinear data can construct a linear relation in a high-dimensional space, and meanwhile, a kernel function is constructed to avoid the specific construction of the nonlinear function, so that the computational complexity of the nonlinear process is reduced, and the modeling efficiency of fault detection is improved.
Further, step S3 includes the following sub-steps:
s31, constructing a nonlinear feature extraction objective function according to a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix, and solving the nonlinear feature extraction objective function by adopting a Lagrangian theorem to obtain an input-output parameter relation equation;
s32, calculating initial output latent variable component information according to an input-output parameter relation equation;
s33, extracting an input score matrix, an output score matrix, an input load matrix and an output load matrix by adopting a nonlinear regression iteration method according to the initial output latent variable component information.
Further, in step S31, the nonlinear feature extraction objective function is:
the input-output parameter relation equation is as follows:
wherein w is a high-dimensional input matrix phi x C is the high-dimensional output matrix phi y Is defined by the projection vector of (a),is a nonlinear input matrix +.>Is a nonlinear output matrix +.>For inputting parameter matrix>For outputting the parameter matrix, u is the initial output latent variable component information, and θ is the characteristic value corresponding to u after characteristic decomposition.
The beneficial effects of the above-mentioned further scheme are: the modeling process of the equipment test system with nonlinear variation of input and output is generally provided, an objective function of input and output correlation extraction is established, and dimension reduction of equipment test data is realized by extracting a projection direction.
Further, step S33 includes the following sub-steps:
s331, initializing the iteration number i=1, outputting the latent variable component information u i Giving an initial value u;
s332, calculating the input latent variable component information t of the ith iteration i And update and output latent variable component information u i ；
wherein ,u′i For the updated output latent variable component signal Si i ；
S333, according to the output latent variable component information u' i And transportLatent variable component information t i Calculating an input load vector and an output load vector, and for a nonlinear input matrixNonlinear output matrix->Input parameter matrix->And output parameter matrix->Updating;
wherein ,py，i To output the load vector, p x，i For the input load vector, E is the identity matrix,for an updated non-linear input matrix +.>For an updated nonlinear output matrix +.>For updated input parameter matrix +.>For updated output parameter matrix +.>
S334, pairMiddle->Performing feature decomposition to obtain output latent variable component information u i+1 ；
S335, judging and outputting latent variable component information u i+1 If the iteration number i is not equal to 1, the step S332 is skipped;
s336, extracting A pieces of output latent variable component information u i A pieces of input latent variable component information t i A input load vectors p x，i And A output load vectors p y，i Obtain an input score matrix m= (t) 1 ，...，t A ) Output score matrix u= (U) 1 ，...，u A ) Input load matrix P x ＝(p x，1 ，...，p x，A ) And output load matrix P y ＝(p y，1 ，...，p y，A )。
The beneficial effects of the above-mentioned further scheme are: the nonlinear iteration specifically constructs a nonlinear projection direction, and the dimension reduction is realized by projecting along the direction after nonlinear mapping of test data in online monitoring.
Further, step S4 includes the following sub-steps:
s41, calculating latent variables after dimension reduction projection of single historical sensing data:
wherein t is a latent variable after dimension reduction projection of single historical sensing data, M is an input score matrix,a Gaussian kernel function for single historical sensing data and all the historical sensing data is standardized;
s42, establishing quality irrelevant space statistics of the historical sensing data according to the latent variable after the dimension reduction projection of the single historical sensing data:
Q＝||φ(x i )-P x t|| 2
where Q is a quality independent spatial statistic of the historical sensed data, φ (x i ) To x i Function, x, of high-dimensional mapping i ＝{x i1 ，...，x ij ，...，x im }，|||| 2 Is a two-norm operation;
s43, calculating a quality-related spatial control limit of the historical sensing data:
wherein ,for quality-related spatial control limit of historical sensing data, F A，n-A Is F distribution;
s44, calculating a quality irrelevant space control limit of the historical sensing data according to the quality irrelevant space statistic of the historical sensing data:
where g=ζ/2 μ, h=2μ 2 /ξ，J th，SPE For the quality independent spatial control limit of the historical sensing data, g is a control limit coefficient, ζ is the variance after constructing statistics Q for all the historical sensing data, μ is the mean after constructing statistics Q for all the historical sensing data,the degree of freedom of chi-square distribution is h;
s45, collecting sensing data of a weapon equipment system to be detected to obtain sensing data to be detected;
s46, mapping the sensor data to be measured to a high-dimensional space after standardized processing to obtain nonlinear sensor data to be measured;
s47, carrying out standardization processing on the Gaussian kernel function of the nonlinear sensing data to be detected to obtain the standardized Gaussian kernel function:
wherein ,is a normalized Gaussian kernel function, k new The method is characterized in that the method is a Gaussian kernel function of nonlinear sensing data to be detected, and K is an input data set X and a Gaussian kernel function of the input data set X;
s48, projecting the nonlinear sensing data to be detected along the quality-related space according to the Gaussian kernel function after the standardized processing to obtain a latent variable t new ：
S49, calculating quality related space statistics and quality unrelated space statistics of the sensing data to be detected:
wherein ,T′2 For the quality-related spatial statistics of the sensed data to be measured, Q' is the quality-independent spatial statistics of the sensed data to be measured,is nonlinear sensing data to be measured, +.>
S50, judging quality-related spatial statistics T 'of the sensing data to be detected' 2 Whether or not to be greater than or equal to a quality-related spatial control limit of the historical sensed dataIf yes, the fault that the existence of the weapon equipment system to be tested causes the system operation error, namely the quality related fault exists, if not, the step S51 is skipped;
s51, judging whether the quality irrelevant space statistic Q' of the sensing data to be detected is more than or equal to the quality irrelevant space control limit J of the historical sensing data th，SPE If so, the existence of the weapon equipment system to be tested does not cause disturbance of system operation errors, namely, quality irrelevant faults exist, and if not, the sensing data to be tested of the weapon equipment system to be tested is normal data, and the weapon equipment system to be tested operates normally.
The beneficial effects of the above-mentioned further scheme are: the control limit of fault detection in the testing process of the equipment system is determined by the historical sensing data, and real-time monitoring of the testing process is realized through statistic construction of the sensing data to be tested.
In summary, the invention has the following beneficial effects:
the existing method only considers the process of nonlinear change of input, so that the complex nonlinear process aimed by the method cannot be accurately described in modeling and feature extraction stages, and is finally reflected in higher fault detection rate and lower false alarm rate. The invention constructs the objective function of the correlation extraction of nonlinear input and output, effectively extracts the nonlinear characteristics of input and output data by constructing the projection direction of quality correlation through nonlinear iteration, has good process monitoring performance, and has more excellent quality-related fault detection rate and lower false alarm rate. And the type and severity of the fault can be judged according to the sensing data in the running process of the weapon equipment system.
Drawings
FIG. 1 is a flow chart of a fault monitoring method for a weapon equipment system;
FIG. 2 is a fault detection result of a quality-related fault IDV (14);
fig. 3 shows the fault detection result of the quality independent fault IDV (04).
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, a fault monitoring method for a weapon equipment system includes the steps of:
s1, acquiring sensing data of a weapon equipment system, and performing linear standardization processing on the sensing data to obtain an input standard sensing data set and an output standard sensing data set;
the types of the sensing data in step S1 include: voltage current data, barometric pressure data, vibration data, temperature data, servo speed data, servo rotational speed data, and servo feedback voltage data, wherein the voltage current data, barometric pressure data, vibration data, and temperature data are used as input data set X, and the servo speed data, the servo rotational speed data, and the servo feedback voltage data are used as output data set Y.
The corresponding type faults of the weapon equipment system are monitored, process data of the corresponding type faults can be acquired, an input data set X is obtained, output data which can respond to the corresponding faults and is output by the weapon equipment system is taken as an output data set Y, for example, when voltage and current are increased, the rotating speed of the servo mechanism can be increased, then the voltage and current data are process data, namely input data, the rotating speed of the servo mechanism is taken as index data, namely output data, and a corresponding relation between the input data and the output data under normal conditions is established and can be used for monitoring the running condition of the system.
The formula for performing linear normalization processing on the sensing data in the step S1 is as follows:
wherein ,standard sensed data for j-th sampled sensed data for a corresponding i-th variable in input dataset X, X i，j Sampling the sensed data for the jth time of the ith variable in the input dataset X, y i，j Sampling the sensor data for the j-th time for the i-th variable in the output data set Y, +.>For outputting standard sensed data of the j-th sampled sensed data of the corresponding i-th variable in the data set Y, n is the sum of the values of each variableSampling times->The set is input standard sensing data set, < ->The set of constituents is an output standard sensing dataset.
The variables mentioned above can be understood as some type of sensed data, e.g. x i，j Which can be interpreted as the j-th sampled sense data of the voltage data i in the input data set X.
S2, performing high-dimensional mapping and nonlinear processing on the input standard sensing data set and the output standard sensing data set to obtain a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix;
step S2 comprises the following sub-steps:
s21, mapping the input standard sensing data set and the output standard sensing data set to a high-dimensional space respectively to obtain a high-dimensional input matrix phi x And a high-dimensional output matrix phi y ；
wherein ,Φx ＝[φ(x 1 )，φ(x 2 )，...，φ(x n )] T ，Φ y ＝[φ(y 1 )，φ(y 2 )，...，φ(y n )] T ，，φ(x i ) To x i Function, x, of high-dimensional mapping i ＝{x i1 ，...，x ij ，...，x im }，y i ＝{y i1 ，...，y ij ，...，y im M is the number of variables.
S22, for high-dimensional input matrix phi x And a high-dimensional output matrix phi y Nonlinear processing is carried out to obtain a nonlinear input matrixAnd a nonlinear output matrix->
Nonlinear input matrix in step S22And a nonlinear output matrix->The calculation formula of (2) is as follows:
s23, according to the nonlinear input matrixAnd a nonlinear output matrix->Calculating input parameter matrix +.>And output parameter matrix->
The input parameter matrix is calculated in the step S23And output parameter matrix->The formula of (2) is:
wherein ,1n Is thatIs the full 1 column vector of>Is real space.
S3, calculating an input score matrix, an output score matrix, an input load matrix and an output load matrix according to the nonlinear input matrix, the nonlinear output matrix, the input parameter matrix and the output parameter matrix;
step S3 comprises the following sub-steps:
s31, constructing a nonlinear feature extraction objective function according to a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix, and solving the nonlinear feature extraction objective function by adopting a Lagrangian theorem to obtain an input-output parameter relation equation;
in step S31, the nonlinear feature extraction objective function is:
the input-output parameter relation equation is as follows:
wherein w is a high-dimensional input matrix phi x C is the high-dimensional output matrix phi y Is defined by the projection vector of (a),is a nonlinear input matrix +.>Is a nonlinear output matrix +.>For inputting parameter matrix>For outputting parameter matrix, u is the initial output latent variable component information, θ 2 For the eigenvalue of the input-output parameter relation equation, θ=2λ=2β.
The detailed process of solving the nonlinear feature extraction objective function by using the Lagrangian theorem is as follows:
a1, establishing Lagrange theorem to solve:
where λ, β are parameters of the lagrangian multiplier method (fixed expression).
A2, obtaining partial derivatives of w and c in A1:
a3, according to constraint: the expression in A2 can be reduced to:
a4, introducing the formula in the step A3 into the step A2:
a5, simplifying the formula in the step A4:
a6, multiplying the two sides of the formula in the step A5 by
u＝Φ y c is->Latent variables of (dimension reduction data, comprising +.>Most information of (b)>Can be rewritten as: />
S32, calculating initial output latent variable component information according to an input-output parameter relation equation;
in the step S32, the input/output parameter relation equationAnd performing characteristic decomposition to obtain u.
S33, extracting an input score matrix, an output score matrix, an input load matrix and an output load matrix by adopting a nonlinear regression iteration method according to the initial output latent variable component information.
Step S33 includes the following sub-steps:
s331, initializing the iteration number i=1, outputting the latent variable component information u i Giving an initial value u;
s332, calculating the input latent variable component information t of the ith iteration i And update and output latent variable component information u i ；
wherein ,U′i For updated output latent variable component information u i ；
S333, according to the output latent variable component information u' i And inputting latent variable component information t i Calculating an input load vector and an output load vector, and for a nonlinear input matrixNonlinear output matrix->Input parameter matrix->And output parameter matrix->Updating;
wherein ,py，i To output the load vector, p x，i For the input load vector, E is the identity matrix,for an updated non-linear input matrix +.>For an updated nonlinear output matrix +.>For updated input parameter matrix +.>For updated output parameter matrix +.>
S334, pairMiddle->Performing feature decomposition to obtain output latent variable component information u i+1 ；
S335, judging and outputting latent variable component information u i+1 Whether or not to convergeIf yes, go to step S336, if no, add 1 to the iteration number i, go to step S332;
s336, extracting A pieces of output latent variable component information u i A pieces of input latent variable component information t i A input load vectors p x，i And A output load vectors p y，i Obtain an input score matrix m= (t) 1 ，...，t A ) Output score matrix u= (U) 1 ，...，u A ) Input load matrix P x ＝(p x，1 ，...，p x，A ) And output load matrix P y ＝(p y，1 ，...，p y，A )。
S4, collecting sensing data of the weapon equipment system to be detected, obtaining sensing data to be detected, and performing fault monitoring on the sensing data to be detected according to the input score matrix, the output score matrix, the input load matrix and the output load matrix.
Step S4 comprises the following sub-steps:
s41, calculating latent variables after dimension reduction projection of single historical sensing data:
wherein t is a latent variable after dimension reduction projection of single historical sensing data, M is an input score matrix,a Gaussian kernel function for single historical sensing data and all the historical sensing data is standardized;
s42, establishing quality irrelevant space statistics of the historical sensing data according to the latent variable after the dimension reduction projection of the single historical sensing data:
Q＝||φ(x i )-P x t|| 2
where Q is a quality independent spatial statistic of the historical sensed data, φ (x i ) To x i Function, x, of high-dimensional mapping i ＝{x i1 ，...，x ij ，...，x im }，|||| 2 Is a two-norm operation;
s43, calculating a quality-related spatial control limit of the historical sensing data:
wherein ,for quality-related spatial control limit of historical sensing data, F A，n-A Is F distribution;
s44, calculating a quality irrelevant space control limit of the historical sensing data according to the quality irrelevant space statistic of the historical sensing data:
where g=ζ/2 μ, h=2μ 2 /ξ，J th，SPE For the quality independent spatial control limit of the historical sensing data, g is a control limit coefficient, ζ is the variance after constructing statistics Q for all the historical sensing data, μ is the mean after constructing statistics Q for all the historical sensing data,the degree of freedom of chi-square distribution is h;
s45, collecting sensing data of a weapon equipment system to be detected to obtain sensing data to be detected;
s46, mapping the sensor data to be measured to a high-dimensional space after standardized processing to obtain nonlinear sensor data to be measured;
s47, carrying out standardization processing on the Gaussian kernel function of the nonlinear sensing data to be detected to obtain the standardized Gaussian kernel function:
wherein ,is a normalized Gaussian kernel function, k new The method is characterized in that the method is a Gaussian kernel function of nonlinear sensing data to be detected, and K is an input data set X and a Gaussian kernel function of the input data set X;
s48, projecting the nonlinear sensing data to be detected along the quality-related space according to the Gaussian kernel function after the standardized processing to obtain a latent variable t new ：
S49, calculating quality related space statistics and quality unrelated space statistics of the sensing data to be detected:
wherein ,T′2 For the quality-related spatial statistics of the sensed data to be measured, Q' is the quality-independent spatial statistics of the sensed data to be measured,is nonlinear sensing data to be measured, +.>
S50, judging quality-related spatial statistics T 'of the sensing data to be detected' 2 Whether or not to be greater than or equal to a quality-related spatial control limit of the historical sensed dataIf yes, wait forDetecting faults of the weapon equipment system, which cause system operation errors, namely quality related faults, if not, jumping to the step S51;
s51, judging whether the quality irrelevant space statistic Q' of the sensing data to be detected is more than or equal to the quality irrelevant space control limit J of the historical sensing data th，SPE If so, the existence of the weapon equipment system to be tested does not cause disturbance of system operation errors, namely, quality irrelevant faults exist, and if not, the sensing data to be tested of the weapon equipment system to be tested is normal data, and the weapon equipment system to be tested operates normally.
Failure causing system operation errors: a large number of sensors for monitoring process variables are distributed in the equipment testing process, sensing data of the sensors for monitoring the process variables are collected to obtain an input data set X, and if the process has a fault and the process variables are abnormal, the key performance index is abnormal, namely the output data set Y is abnormal, the fault is caused to cause the system to operate in error;
no disturbance of system operation errors is caused: the process is also abnormal when the process fails, but the key performance index is not abnormal, and the disturbance of system operation errors is not caused.
Experiment:
considering that large complex equipment systems are multi-index, high-dimensional, large samples, presence of process variables and key performance indicators, like tennessee-eastmann (TEP), TEP is employed to verify the validity of the present invention. The method proposed by the present invention is verified by data collected in tenacissian-eastmann (TEP) experiments. TEP is a small industrial process developed by isman chemical company Downs and Vogel in 1993, the whole process consisting of five operating units including chemical reactors, condensers, compressors, vapor/liquid separators and separators.
TEP contains eight components: a, B, C, D, E, F, G and H, wherein gaseous species A, C, D and E and inert species B are reactants, G and H are reaction products, and F is a reaction byproduct.
TABLE 1 15 known faults (IDV)
TEP co-generation into 22 data sets for process monitoring and fault diagnosis, including 1 normal data and 8 quality-related fault training sets and 4 quality-independent fault training sets. In the training set, the normal data set contains 480 samples for building a nonlinear model (i.e., the process of steps S1-S3 of the present invention), and the fault data set contains 480 fault samples for building a fault library; in the test set, each test data set contains 960 samples, the first 160 being normal samples and the second 800 being faulty samples for experimental verification. Each input sample included 33 variables and the test sample was 3 variables. The fault type IDV (1, 2,6-8, 12, 13) is a quality-related fault dataset (corresponding to a fault causing a system operation error) and the IDV (3, 4,9, 15) is a quality-independent fault dataset (corresponding to a disturbance not causing a system operation error).
Substituting the normal data set as a training sample into the step S1 of the invention, calculating to obtain parameters in the steps S2 and S3, and substituting the IDV (1-15) as sensing data to be detected into the step S4 of the invention to obtain a detection result shown in the table 2.
TABLE 2 quality related fault detection
Table 2 shows the detection conditions of quality-related faults in the fault IDV (8), and the detection rate can be seen that the invention has good monitoring performance aiming at the nonlinear chemical process, and the detection rate is more than 90%. The specific detection result of the fault IDV (14) is shown in fig. 2. In the IDV (14), the first 160 are normal samples and the last 800 are fault samples, and it can be seen that the normal samples are basically under the control limit, and the fault samples are basically and effectively alarmed.
TABLE 3 quality independent fault detection results
Table 3 shows the detection of class 4 quality independent faults. Of these faults, the process variable X has failed, but is quality independent and therefore does not cause a failure of the key performance indicators. Therefore, in the detection of key performance indexes, the faults are not warned, and if the faults are warned, the faults are false alarms. As can be seen from Table 3, the false positive rates of class 4 quality independent faults are all within 15%, wherein the faults 3,4 and 9 are all less than 10%. The specific detection result given for the faulty IDV (4) is shown in fig. 3. In FIG. 3, T' 2 The statistics only has a small amount of false alarms, and quality irrelevant faults are effectively monitored in the quality irrelevant space of the SPE, so that the faults are identified, and the detection performance is good.
Claims (8)
1. A fault monitoring method for a weapon equipment system, comprising the steps of:
s1, acquiring sensing data of a weapon equipment system, and performing linear standardization processing on the sensing data to obtain an input standard sensing data set and an output standard sensing data set;
s2, performing high-dimensional mapping and nonlinear processing on the input standard sensing data set and the output standard sensing data set to obtain a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix;
s3, calculating an input score matrix, an output score matrix, an input load matrix and an output load matrix according to the nonlinear input matrix, the nonlinear output matrix, the input parameter matrix and the output parameter matrix;
s4, collecting sensing data of a weapon equipment system to be detected, obtaining sensing data to be detected, and performing fault monitoring on the sensing data to be detected according to an input score matrix, an output score matrix, an input load matrix and an output load matrix;
the step S4 includes the following sub-steps:
s41, calculating latent variables after dimension reduction projection of single historical sensing data:
wherein ,for an input parameter matrix, U is an output score matrix, t is a latent variable obtained by performing dimension reduction projection on single historical sensing data, M is an input score matrix, and +.>A Gaussian kernel function for single historical sensing data and all the historical sensing data is standardized;
s42, establishing quality irrelevant space statistics of the historical sensing data according to the latent variable after the dimension reduction projection of the single historical sensing data:
Q＝‖φ(x i )-P x t‖ 2
where Q is a quality independent spatial statistic of the historical sensed data, φ (x i ) To x i Function, x, of high-dimensional mapping i ＝{x i1 ,…,x ij ,…,x im }，P x For inputting a load matrix II 2 Is a two-norm operation;
s43, calculating a quality-related spatial control limit of the historical sensing data:
wherein ,for quality-related spatial control limit of historical sensing data, F A,n-A For F distribution, A is the number of each extracted variable, and n is the sampling times of each variable; the variables comprise output latent variable component information, input load vectors and output load vectors;
s44, calculating a quality irrelevant space control limit of the historical sensing data according to the quality irrelevant space statistic of the historical sensing data:
where g=ζ/2 μ, h=2μ 2 /ξ，J th，SPE For the quality independent spatial control limit of the historical sensing data, g is a control limit coefficient, ζ is the variance after constructing statistics Q for all the historical sensing data, μ is the mean after constructing statistics Q for all the historical sensing data,the degree of freedom of chi-square distribution is h; s45, collecting sensing data of a weapon equipment system to be detected to obtain sensing data to be detected;
s46, mapping the sensor data to be measured to a high-dimensional space after standardized processing to obtain nonlinear sensor data to be measured;
s47, carrying out standardization processing on the Gaussian kernel function of the nonlinear sensing data to be detected to obtain the standardized Gaussian kernel function:
wherein ,is a normalized Gaussian kernel function, k new Is Gaussian kernel function of nonlinear sensing data to be detected, K is input data set X and input data setGaussian kernel function of data set X, 1 n Is->Is the full 1 column vector of>Is real space;
s48, projecting the nonlinear sensing data to be detected along the quality-related space according to the Gaussian kernel function after the standardized processing to obtain a latent variable t new ：
S49, calculating quality related space statistics and quality unrelated space statistics of the sensing data to be detected:
wherein ,T′2 For the quality-related spatial statistics of the sensed data to be measured, Q' is the quality-independent spatial statistics of the sensed data to be measured,is nonlinear sensing data to be measured, +.>
S50, judging quality-related spatial statistics T 'of the sensing data to be detected' 2 Whether or not to be greater than or equal to a quality-related spatial control limit of the historical sensed dataIf yes, the fault that the existence of the weapon equipment system to be tested causes the system operation error, namely the quality related fault exists, if not, the step S51 is skipped;
s51, judging whether the quality irrelevant space statistic Q' of the sensing data to be detected is more than or equal to the quality irrelevant space control limit J of the historical sensing data th，SPE If so, the existence of the weapon equipment system to be tested does not cause disturbance of system operation errors, namely, quality irrelevant faults exist, and if not, the sensing data to be tested of the weapon equipment system to be tested is normal data, and the weapon equipment system to be tested operates normally.
2. The fault monitoring method for a weapon equipment system according to claim 1, wherein the type of sensed data in step S1 comprises: voltage current data, barometric pressure data, vibration data, temperature data, servo speed data, servo rotational speed data, and servo feedback voltage data, wherein the voltage current data, barometric pressure data, vibration data, and temperature data are used as input data set X, and the servo speed data, the servo rotational speed data, and the servo feedback voltage data are used as output data set Y.
3. The fault monitoring method for a weapon equipment system according to claim 2, wherein the formula for performing the linear normalization processing on the sensing data in step S1 is:
wherein ,standard sensed data for j-th sampled sensed data for a corresponding i-th variable in input dataset X, X i,j Sampling the sensed data for the jth time of the ith variable in the input dataset X, y i,j Sampling the sensor data for the j-th time for the i-th variable in the output data set Y, +.>For outputting standard sensing data of the j-th sampling sensing data of the corresponding i-th variable in the data set Y, n is the sampling frequency of each variable, < +.>The set is input standard sensing data set, < ->The set of constituents is an output standard sensing dataset.
4. A fault monitoring method for a weapon equipment system according to claim 3, characterized in that said step S2 comprises the following sub-steps:
s21, mapping the input standard sensing data set and the output standard sensing data set to a high-dimensional space respectively to obtain a high-dimensional input matrix phi x And a high-dimensional output matrix phi y ；
S22, for high-dimensional input matrix phi x And a high-dimensional output matrix phi y Nonlinear processing is carried out to obtain a nonlinear input matrixAnd a nonlinear output matrix->
S23, according to the nonlinear input matrixAnd a nonlinear output matrix->Calculating input parameter matrix +.>And output parameter matrix->
5. The fault monitoring method for a weapon equipment system according to claim 4, wherein said step S22 is a nonlinear input matrixAnd a nonlinear output matrix->The calculation formula of (2) is as follows:
the input parameter matrix is calculated in the step S23And output parameter matrix->The formula of (2) is:
wherein ,1n Is thatIs the full 1 column vector of>Is real space.
6. The fault monitoring method for a weapon equipment system according to claim 5, wherein said step S3 comprises the sub-steps of:
s31, constructing a nonlinear feature extraction objective function according to a nonlinear input matrix, a nonlinear output matrix, an input parameter matrix and an output parameter matrix, and solving the nonlinear feature extraction objective function by adopting a Lagrangian theorem to obtain an input-output parameter relation equation;
s32, calculating initial output latent variable component information according to an input-output parameter relation equation;
s33, extracting an input score matrix, an output score matrix, an input load matrix and an output load matrix by adopting a nonlinear regression iteration method according to the initial output latent variable component information.
7. The fault monitoring method for a weapon equipment system according to claim 6, wherein the nonlinear feature extraction objective function in step S31 is:
the input-output parameter relation equation is as follows:
wherein w is a high-dimensional input matrix phi x C is the high-dimensional output matrix phi y Is defined by the projection vector of (a),is a nonlinear input matrix +.>Is a nonlinear output matrix +.>For inputting parameter matrix>For outputting the parameter matrix, u is the initial output latent variable component information, and θ is the characteristic value corresponding to u after characteristic decomposition.
8. The fault monitoring method for a weapon equipment system according to claim 7, wherein said step S33 comprises the sub-steps of:
s331, initializing iteration timesFor outputting latent variable component information->Giving an initial value u;
s332, calculate the firstInput latent variable component information of multiple iterations +.>And update the output latent variable component information +.>
wherein ,for updated output latent variable component information +.>
S333, according to the output latent variable component informationAnd input latent variable component information->Calculating an input load vector and an output load vector and applying to a nonlinear input matrix>Nonlinear output matrix->Input parameter matrix->And output parameter matrix->Updating;
wherein ,for outputting the load vector>For inputting load vector, E is identity matrix, < ->For updated non-linear inputMatrix->For an updated nonlinear output matrix +.> For updated input parameter matrix +.>For updated output parameter matrix +.>
S334, pairMiddle->Performing feature decomposition to obtain output latent variable component information +.>
S335, judging and outputting latent variable component informationIf the number of iterations is converged, if yes, the process goes to step S336, if no, the iteration number is increased>1 is added, and the process jumps to step S332;
s336, extracting A pieces of output latent variable component informationA input latent variable component information->A input load vectors->And A output load vectors->Obtain an input score matrix m= (t) 1 ,…,t A ) Output score matrix u= (U) 1 ,…,u A ) Input load matrix P x ＝(p x,1 ,…,p x,A ) And output load matrix P y ＝(p y,1 ,…,p y,A )。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111394862.3A CN114034332B (en) | 2021-11-23 | 2021-11-23 | Fault monitoring method for weapon equipment system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111394862.3A CN114034332B (en) | 2021-11-23 | 2021-11-23 | Fault monitoring method for weapon equipment system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114034332A CN114034332A (en) | 2022-02-11 |
| CN114034332B true CN114034332B (en) | 2023-09-22 |
Family
ID=80145222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111394862.3A Active CN114034332B (en) | 2021-11-23 | 2021-11-23 | Fault monitoring method for weapon equipment system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114034332B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2431877A2 (en) * | 2010-09-17 | 2012-03-21 | Siemens Corporation | System and method for modeling conditional dependence for anomaly detection in machine condition monitoring |
| CN107895224A (en) * | 2017-10-30 | 2018-04-10 | 北京工业大学 | A kind of MKECA fermentation process fault monitoring methods based on extension nuclear entropy load matrix |
| CN108664009A (en) * | 2017-08-03 | 2018-10-16 | 湖州师范学院 | Divided stages based on correlation analysis and fault detection method |
| CN110928262A (en) * | 2019-12-17 | 2020-03-27 | 中国人民解放军火箭军工程大学 | Quality-related fault online monitoring method for efficiently updating model in time-varying system |
| CN113420815A (en) * | 2021-06-24 | 2021-09-21 | 江苏师范大学 | Semi-supervised RSDAE nonlinear PLS intermittent process monitoring method |
-
2021
- 2021-11-23 CN CN202111394862.3A patent/CN114034332B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2431877A2 (en) * | 2010-09-17 | 2012-03-21 | Siemens Corporation | System and method for modeling conditional dependence for anomaly detection in machine condition monitoring |
| CN108664009A (en) * | 2017-08-03 | 2018-10-16 | 湖州师范学院 | Divided stages based on correlation analysis and fault detection method |
| CN107895224A (en) * | 2017-10-30 | 2018-04-10 | 北京工业大学 | A kind of MKECA fermentation process fault monitoring methods based on extension nuclear entropy load matrix |
| CN110928262A (en) * | 2019-12-17 | 2020-03-27 | 中国人民解放军火箭军工程大学 | Quality-related fault online monitoring method for efficiently updating model in time-varying system |
| CN113420815A (en) * | 2021-06-24 | 2021-09-21 | 江苏师范大学 | Semi-supervised RSDAE nonlinear PLS intermittent process monitoring method |
Non-Patent Citations (2)
| Title |
|---|
| The Multi-space Generalization of Total Projection to Latent Structures (MsT-PLS) and Its Application to Online Process Monitoring;Chunhui Zhao 等;2013 10th IEEE International Conference on Control and Automation (ICCA);1441-1446 * |
| 基于偏最小二乘得分重构的质量相关故障检测;孔祥玉 等;控制理论与应用;第37卷(第11期);2321-2332 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114034332A (en) | 2022-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Don et al. | Dynamic process fault detection and diagnosis based on a combined approach of hidden Markov and Bayesian network model | |
| Said et al. | Machine learning technique for data-driven fault detection of nonlinear processes | |
| CN111780800B (en) | Method and system for monitoring, isolating and reconstructing sensor fault | |
| CN114841396B (en) | Method for predicting metamorphic trend and warning catastrophe risk in petrochemical production process | |
| CN111639304B (en) | CSTR fault positioning method based on Xgboost regression model | |
| CN110942258B (en) | Performance-driven industrial process anomaly monitoring method | |
| Luo et al. | Key-performance-indicators-related fault subspace extraction for the reconstruction-based fault diagnosis | |
| CN113703422B (en) | Gas turbine pneumatic actuator fault diagnosis method based on feature analysis processing | |
| Namaki-Shoushtari et al. | Bayesian control loop diagnosis by combining historical data and process knowledge of fault signatures | |
| Ayech et al. | New adaptive moving window PCA for process monitoring | |
| CN114034332B (en) | Fault monitoring method for weapon equipment system | |
| Liu et al. | Nonstationary fault detection and diagnosis for multimode processes | |
| Wang et al. | A hybrid fault diagnosis strategy for chemical process startups | |
| Liu et al. | Structured sequential Gaussian graphical models for monitoring time-varying process | |
| CN115047853B (en) | Minor fault detection method based on recursion standard variable residual error and kernel principal component analysis | |
| Dong et al. | Dynamic-inner canonical correlation analysis based process monitoring | |
| Geng et al. | Fault detection of large-scale process control system with higher-order statistical and interpretative structural model | |
| Ma et al. | A novel dynamic nonlinear partial least squares based on the cascade structure | |
| CN113076211B (en) | Quality-related fault diagnosis and false alarm feedback method based on fault reconstruction | |
| Liu et al. | Graph attention network with Granger causality map for fault detection and root cause diagnosis | |
| CN114383648A (en) | Temperature instrument fault diagnosis method and device | |
| Guo et al. | Fault Diagnosis Combining Information Entropy with Transfer Entropy for Chemical Processes | |
| CN114112374A (en) | Fault diagnosis method of weapon equipment system based on orthogonal sub-block projection | |
| Said et al. | Improved dynamic optimized kernel partial least squares for nonlinear process fault detection | |
| Ma et al. | Process monitoring of the pneumatic control valve using canonical variate analysis |
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 |