CN109543357A - Fault degree quantitative evaluation method for multivariate regression model optimization - Google Patents
Fault degree quantitative evaluation method for multivariate regression model optimization Download PDFInfo
- Publication number
- CN109543357A CN109543357A CN201910024090.0A CN201910024090A CN109543357A CN 109543357 A CN109543357 A CN 109543357A CN 201910024090 A CN201910024090 A CN 201910024090A CN 109543357 A CN109543357 A CN 109543357A
- Authority
- CN
- China
- Prior art keywords
- fault degree
- index
- fault
- formula
- degree
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000005457 optimization Methods 0.000 title claims abstract description 24
- 238000011158 quantitative evaluation Methods 0.000 title abstract description 3
- 238000012360 testing method Methods 0.000 claims abstract description 60
- 239000011159 matrix material Substances 0.000 claims abstract description 36
- 238000011156 evaluation Methods 0.000 claims abstract description 6
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 29
- 230000008569 process Effects 0.000 claims description 10
- 238000004088 simulation Methods 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000002347 injection Methods 0.000 claims description 4
- 239000007924 injection Substances 0.000 claims description 4
- 238000013210 evaluation model Methods 0.000 abstract description 4
- 238000000605 extraction Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000007257 malfunction Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 208000037656 Respiratory Sounds Diseases 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009760 electrical discharge machining Methods 0.000 description 1
- 238000005538 encapsulation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a fault degree quantitative evaluation method for multivariate regression model optimization, which comprises fault degree evaluation modeling and fault degree calculation by using optimized measuring points, a fault evaluation model and fault degree indexes. The method calculates the fault degree evaluation result of the test sample; calculating a fault degree evaluation root-mean-square matrix of the test sample; and carrying out fault degree evaluation by using the optimized evaluation model and the fault degree index, wherein the optimally selected fault degree evaluation model can synthesize the relationship among the source of fault data acquisition, the extraction of the fault degree index and the fault degree to ensure that the root mean square error is evaluated to be minimum.
Description
Technical field
The invention belongs to fault degree quantitative evaluation fields, and in particular to a kind of fault degree of multivariate regression models optimization
Quantitative estimation method.It is mainly used in typical mechanical parts in Mechatronic Systems, such as bearing, gear, axis.
Background technique
Fault degree assessment refers to information according to the collected data, and analysis is extracted fault signature, assessed using intelligent algorithm
The severity or severity level of failure.Assessment result timely and accurately can effectively trigger maintenance decision mechanism, avoid event
The aggravation of barrier degree or secondary failure occur, and are of great significance for reducing functions of the equipments failure probability of happening.In reality
In engineering, for the good system of encapsulation, it is difficult directly to measure severity (such as crackle of failure by detecting instrument or equipment
Length, spot corrosion width), it can only be acquired by the detection device being deployed in equipment or sensor etc. and be made an uproar as caused by failure
The abnormal signals such as sound, vibration, temperature, output shift, oil liquid, in order to accurately obtain the true fault severity level of current device,
It is badly in need of the quantization function relationship established between a kind of connector arrangement monitoring data source, fault degree index and fault severity level.
Currently, well known method has the following problems:
First, the prior art not from fault degree signal source analysis data to the validity of assessment so that useful failure
On signal is useless, garbage signal strong jamming directly reduces assessment accuracy and efficiency.
Second, the influence that the non-comprehensive analysis data source of the prior art, index and model assess fault degree causes simple
Assessment accuracy by index and model is low.
Summary of the invention
In order to solve the above problems existing in the present technology, it is an object of that present invention to provide a kind of optimizations of multivariate regression models
Fault degree quantitative estimation method.
The technical scheme adopted by the invention is as follows:
A kind of fault degree quantitative estimation method of multivariate regression models optimization, including fault degree assessment modeling, and
Fault degree calculating, fault degree assessment modeling packet are carried out using the measuring point of optimization, assessment of failure model and fault degree index
Include following steps:
S1.1, fault degree simulation and injection: the component of the different severity failures of simulation carrying is simultaneously installed into system, fortune
Row system generates fault-signal;
S1.2, acquisition multivariate data: multiple measuring points are arranged in system inside and outside, monitor abnormal signal caused by fault propagation,
Obtain the fault-signal collection of each measuring point monitoring under different faults degree;
S1.3, building fault degree indicator vector;
S1.4, setting fault degree assessment models;
S1.5, the corresponding fault degree assessment models coefficient matrix of training sample is calculated, then calculates the event of test sample
Hinder scale evaluation result;
S1.6, the fault degree for calculating test sample assess root-mean-square error;
S1.7, optimum choice measuring point, assessment of failure model and fault degree index.
Preferably, the realization process of the S1.3 are as follows: use time-frequency statistical indicator extracting method, use formula (1) structure
Build fault degree indicator vector:
In formula, FnFor n-th of fault degree indicator vector, 1≤n≤NF, NFIt is total for alternative fault degree index
Number, fiFor the calculating function of i-th of fault degree index, N is total sample number, NTFor the measuring point sum of deployment, ojkFor at j-th
The data that k-th of measuring point acquires in sample, 1≤j≤N, 1≤k≤NT。
Preferably, the time-frequency statistical indicator extracting method include mean value, root-mean-square value, root amplitude, absolutely
It is value, degree of skewness, kurtosis, variance, peak value, standard deviation, peak-to-peak value, mean power, waveform index, peak index, pulse index, abundant
Spend index, degree of skewness index, kurtosis index or means frequency.
Preferably, the realization process of the S1.4 are as follows: alternative using formula (2) setting according to the number of measuring point
Multiple regression assessment models:
In formula, fiFor the calculating function of i-th of fault degree index, N is total sample number, NTIt is total for the measuring point of deployment,
ojkFor the data that k-th of measuring point acquires in j-th of sample, 1≤j≤N, 1≤k≤NT,For assessment models system
Number.
Preferably, in the S1.5, the meter of the corresponding fault degree assessment models coefficient matrix of training sample is calculated
Calculate formula are as follows:
In formula, (Bij)trainFor the corresponding assessment models coefficient of training sample for using i-th of model, j-th of index to obtain
Matrix, (Xij)trainFor the training sample observation matrix for using j-th of index to calculate for i-th of model, YtrainFor training
The true fault degree vector of sampleNtrainFor the sum of training sample.
Preferably, in the S1.5, the calculation formula of the fault degree assessment result of test sample is calculated are as follows:
In formula,To use fault degree model MiWith fault degree index FjAssess the failure journey of obtained test sample
Spend vector, (Bij)trainFor the assessment models coefficient matrix for using training sample to obtain, calculated by formula (3), (Xij)trainFor needle
To fault degree model MiWith fault degree index FjThe test sample observation matrix of calculating.
Preferably, the realization process of the S1.6 are as follows:
S1.6.1, N is calculated using formula (5)MN is used under a modelFA index calculates fault degree and assesses root-mean-square error square
Battle array:
In formula, R is that the fault degree of test sample assesses root mean square matrix, the corresponding fault degree model of the row of R, and column correspond to
Fault degree index, NMFor fault degree assessment models sum, rijDefinition uses fault degree model MiWith fault degree index Fj
The root-mean-square error of lower fault degree assessment;
S1.6.2, r is calculated using formula (6)ij:
In formula, NtestIt is total for test sample,To use fault degree model MiWith fault degree index FjFailure journey
Assessment result is spent, is calculated by formula (4), YtestFor the true fault degree vector of test sample,
Preferably, the realization process of the S1.7 are as follows: most using formula (7) optimum choice assessment of failure root-mean-square error
Small corresponding assessment models and fault degree index:
In formula, M*、F*The respectively minimum corresponding model of fault degree assessment root-mean-square error and fault degree index, root
According to model M*Expression formula determine optimal test points set T*With model parameter B*。
Preferably, described to carry out fault degree using the measuring point, assessment of failure model and fault degree index of optimization
Calculating process are as follows:
S2.1, the test points set T optimized using formula (7)*Acquire the data O under current statec;
S2.2, the fault degree index F optimized using formula (7)*The fault degree index of the data under current state is extracted,
Construct observation vector matrix Xc;
S2.3, the fault degree under current state is calculated using formula (8):
In formula, XcFor observation vector matrix, B*Optimal model parameter.
The invention has the benefit that
The fault degree assessment result of present invention calculating test sample;The fault degree for calculating test sample assesses root mean square
Matrix;Carry out fault degree using the assessment models of optimization, fault degree index to assess, the fault degree of optimum choice assesses mould
The source of type energy resultant fault data acquisition, fault degree index extraction, three aspect relationship of fault degree ensure to assess root mean square
Error is minimum.
Detailed description of the invention
Fig. 1 is the present invention-embodiment method flow diagram.
Fig. 2 is the original waveform figure under the present invention-four kinds of severity of embodiment inner ring failure.
Fig. 3 is the present invention-embodiment T*In two measuring points acquisition current state original waveform figure.
Specific embodiment
With reference to the accompanying drawing and specific embodiment the present invention is further elaborated.
Embodiment:
As shown in Figure 1, the fault degree quantitative estimation method that a kind of multivariate regression models of the present embodiment optimizes, first will
Polynary measuring point data is divided into training sample and test sample, is described using training sample building different faults degree and index polynary
Fault degree assessment models are returned, compare the root-mean-square error that each assessment models are assessed under different indexs using test sample
Size, and then select optimal fault degree assessment models and fault degree index.
Based on this, comprehensive optimal data source, fault degree index and assessment models realize accurate fault degree
Assessment finally ensures to improve the precision and efficiency of assessment.
A kind of fault degree quantitative estimation method of multivariate regression models optimization, is realized especially by following steps:
The first step, fault degree assessment modeling.
S1.1, fault degree simulation and injection.It is different tight that carrying is simulated by the modes such as software emulation or hardware processing
Tape jam component is installed into system by the component of weight degree failure, and operating system generates fault-signal.
S1.2, acquisition multivariate data.It is required according to system monitoring, multiple measuring points is set in system inside and outside, monitor failure
Abnormal signal caused by propagating obtains the fault-signal collection of each measuring point monitoring under different faults degree.
S1.3, building fault degree indicator vector.Using time-frequency statistical indicator extracting method common in engineering, such as
Value, root-mean-square value, root amplitude, absolute mean, degree of skewness, kurtosis, variance, peak value, standard deviation, peak-to-peak value, mean power, wave
Shape index, peak index, pulse index, margin index, degree of skewness index, kurtosis index, means frequency etc. use formula (1) structure
Build fault degree indicator vector:
In formula, FnFor n-th of fault degree indicator vector, 1≤n≤NF, NFIt is total for alternative fault degree index
Number, fiFor the calculating function of i-th of fault degree index, N is total sample number, NTFor the measuring point sum of deployment, ojkFor at j-th
The data that k-th of measuring point acquires in sample, 1≤j≤N, 1≤k≤NT。
S1.4, setting fault degree assessment models.It is set alternative polynary time according to the number of measuring point using formula (2)
Return assessment models.
In formula, fiFor the calculating function of i-th of fault degree index, N is total sample number, NTIt is total for the measuring point of deployment,
ojkFor the data that k-th of measuring point acquires in j-th of sample, 1≤j≤N, 1≤k≤NT,For assessment models system
Number.
S1.5, the fault degree assessment result for calculating test sample.
S1.5.1, the corresponding fault degree assessment models coefficient matrix of training sample is calculated using formula (3):
In formula, (Bij)trainFor the corresponding assessment models coefficient of training sample for using i-th of model, j-th of index to obtain
Matrix, (Xij)trainFor the training sample observation matrix for using j-th of index to calculate for i-th of model, YtrainFor training
The true fault degree vector of sampleNtrainFor the sum of training sample.
S1.5.2, the fault degree assessment result that test sample is calculated using formula (4):
In formula,To use fault degree model MiWith fault degree index FjAssess the failure journey of obtained test sample
Spend vector, (Bij)trainFor the assessment models coefficient matrix for using training sample to obtain, calculated by formula (3), (Xij)trainFor needle
To fault degree model MiWith fault degree index FjThe test sample observation matrix of calculating.
S1.6, the fault degree for calculating test sample assess root-mean-square error.N is calculated using formula (5)MIt is used under a model
NFA index calculates fault degree and assesses root-mean-square error matrix.
In formula, the row outside braces indicates feature vector, and the list representation model number outside braces, R is test sample
Fault degree assesses root mean square matrix, and the corresponding fault degree model of the row of R arranges corresponding fault degree index, NMFor fault degree
Assessment models sum, rijDefinition uses fault degree model MiWith fault degree index FjThe root mean square of lower fault degree assessment misses
R is calculated using formula (6) in differenceij:
In formula, NtestIt is total for test sample,To use fault degree model MiWith fault degree index FjFailure journey
Assessment result is spent, is calculated by formula (4), YtestFor the true fault degree vector of test sample,
S1.7, optimum choice measuring point, assessment of failure model and fault degree index.It is commented using formula (7) optimum choice failure
Estimate the minimum corresponding assessment models of root-mean-square error and fault degree index:
In formula, M*、F*The respectively minimum corresponding model of fault degree assessment root-mean-square error and fault degree index, root
According to model M*Expression formula determine optimal test points set T*With model parameter B*。
Second step, fault degree assessment.
S2.1, the test points set T optimized using formula (7)*Acquire the data O under current statec。
S2.2, the fault degree index F optimized using formula (7)*The fault degree index of the data under current state is extracted,
Construct observation vector matrix Xc。
S2.3, the fault degree under current state is calculated using formula (8):
In formula, XcFor observation vector matrix, B*Optimal model parameter.
Main thought of the invention is elaborated by taking rolling bearing fault simulated experiment platform as an example:
The first step, fault degree assessment modeling.
S1.1, fault degree simulation and injection.
Rolling bearing fault simulated experiment platform includes the motor (1hp=746w) of a 2hp, torque sensor, shows function
Device and electrical control gear.Experiment bearing is 6205-2RS JEMSKF type deep groove ball bearing, and the power of motor is 746W, defeated
Entering axis revolving speed is 1772r/min.Using the bearing inner ring failure of 4 kinds of severity of electrical discharge machining, fault severity level is divided into
0, the bearing for carrying 4 kinds of severity failures, is mounted on testing stand by 0.007 inch, 0.014 inch, 0.021 inch respectively
Operation.
S1.2, acquisition multivariate data.
One measuring point, i.e. measuring point collection { t are respectively set above the bearing block of testing stand motor drive end and fan end1,
t2, measuring point sum NT=2, the corresponding vibration number of 4 kinds of severity failures is acquired using the accelerometer being mounted on 2 measuring points
According to every kind of malfunction acquires 3 groups of data, amounts to 12 groups of data, the original being illustrated in figure 2 under 4 kinds of severity of inner ring failure
Beginning waveform.Each malfunction selects 2 groups of data as training sample, i.e. training sample sum Ntrain=8, training sample is true
Fault degree vector Ytrain=(0 0 0.007 0.007 0.014 0.014 0.021 0.021);Every kind of state remaining 1
Group data amount to 4 groups of data as test sample, i.e. test sample sum Ntest=4, the true fault degree of test sample to
Measure Ytest=(0 0.007 0.014 0.021).
S1.3, building fault degree indicator vector.
Using 10 kinds of time-frequencies statistical indicator extracting method, that is, fault degree index sum N common in engineeringF=10, respectively
Corresponding mean value (F1), root-mean-square value (F2), root amplitude (F3), absolute mean (F4), degree of skewness (F5), kurtosis (F6), variance
(F7), peak value (F8), standard deviation (F9), peak-to-peak value (F10), fault degree indicator vector is constructed, following formula is to refer to using root amplitude
Mark (F3) establish training sample fault degree indicator vector:
S1.4, setting fault degree assessment models.Due to measuring point sum NT=2, then following 6 kinds of multiple regressions can be set
Model:
M1:y=b0+b1x1;
M2:y=b0+b1x2;
M6:y=b0+b1x1+b2x2。
Based on above-mentioned model, the corresponding observed value vector of 8 training samples is set:
It is similar using the training sample observation matrix method for solving of other 9 indexs, it does not repeat one by one herein.
S1.5, the fault degree assessment result for calculating test sample.The corresponding event of 8 training samples is calculated using formula (3)
Hinder scale evaluation model coefficient B:
Using fault degree index F3(i.e. root amplitude) establishes sight of 4 test samples under above-mentioned 6 kinds of assessment models
Examine value vector:
Service index F3The observation value matrix of 4 test samples is established with above-mentioned 6 assessment models:
The observation matrix method for solving of test sample is similar under other 9 indexs, does not repeat one by one herein.
It can be calculated using formula (4) in index F3The lower and upper fault degree for stating test sample under 6 kinds of assessment models is commented
Estimate result:
The fault degree assessment result of test sample is similar under other 9 indexs, does not repeat one by one herein.
S1.6, the fault degree for calculating 4 test samples assess root-mean-square error.It is calculated using formula (5) and is adopted under 6 models
The root-mean-square error matrix of fault degree assessment is carried out with 10 fault degree indexs.
As described above, the true fault degree vector of 4 test samples,
Ytest=(0 0.007 0.014 0.021).
Root-mean-square error value of 6 assessment models under index 1 is calculated using formula (6).
r13=0.3508 × 10-4,r23=0.5902 × 10-4,r33=0.3074 × 10-4,
r43=0.0617 × 10-4,r53=0.0383 × 10-4,r63=0.1524 × 10-4
It is compared from the root-mean-square error of above-mentioned assessment as can be seen that in fault degree index F3Under, it is commented using model 5
The root-mean-square error estimated minimum 0.0383 × 10-4.Similarly, other 9 indexs are calculated in 6 kinds of assessment moulds using formula (5) and (6)
The fault degree of test sample assesses root mean square under type, does not repeat one by one herein.
S1.7, according to above-mentioned solving result, using the assessment of failure model and fault degree index of formula (7) selection optimization,
It is compared and analyzed by above-mentioned calculated result.Assessment root mean square under 6 assessment models and 10 fault degree indexs misses
In poor matrix, the minimum corresponding model of assessment root-mean-square error is M5It is F with index2(i.e. root mean square), i.e. (M*,F*)=(M5,
F2)。
Pass through comparative analysis, 5 expression formula of model M are as follows: y=b0+b1f1+b2f1 2+b3f2+b4f2 2, that is, indicate the serious journey of failure
Degree and measuring point t1And t2In binary quadratic polynomial relationship, corresponding Optimized model coefficient B *=B52=(0.0035 0.2617-
0.2999 -0.2834 0.3436)T.According to model M5It is found that the test points set T of optimization*={ t1, t2}。
Second step, fault degree assessment.
S2.1, optimization measuring point T is used*In two measuring point t1And t2Acquire the data O under current statec,
It is illustrated in figure 3 T*In two measuring points acquisition current state original waveform.
S2.2, optimizing index F is used*={ F2(corresponding to root mean square) extract the index of current unknown state, building is seen
Measured value vector matrix Xc=(1 0.0005 0.00052 0.0002 0.00022)。
S2.3, the fault degree under current state is assessed using formula (8)(inch).
The present invention is not limited to above-mentioned optional embodiment, anyone can show that other are various under the inspiration of the present invention
The product of form, however, make any variation in its shape or structure, it is all to fall into the claims in the present invention confining spectrum
Technical solution, be within the scope of the present invention.
Claims (9)
1. a kind of fault degree quantitative estimation method of multivariate regression models optimization, it is characterised in that: assessed including fault degree
Modeling, and fault degree calculating is carried out using the measuring point, assessment of failure model and fault degree index of optimization, fault degree is commented
Estimate modeling to include the following steps:
S1.1, fault degree simulation and injection: the component of the different severity failures of simulation carryings simultaneously is installed into system, runs and is
System generates fault-signal;
S1.2, acquisition multivariate data: multiple measuring points are arranged in system inside and outside, monitor abnormal signal caused by fault propagation, obtain
The fault-signal collection of each measuring point monitoring under different faults degree;
S1.3, building fault degree indicator vector;
S1.4, setting fault degree assessment models;
S1.5, the corresponding fault degree assessment models coefficient matrix of training sample is calculated, then calculates the failure journey of test sample
Spend assessment result;
S1.6, the fault degree for calculating test sample assess root-mean-square error;
S1.7, optimum choice measuring point, assessment of failure model and fault degree index.
2. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 1, feature exist
In: the realization process of the S1.3 are as follows: use time-frequency statistical indicator extracting method, construct fault degree mark sense using formula (1)
Amount:
In formula, FnFor n-th of fault degree indicator vector, 1≤n≤NF, NFFor alternative fault degree index sum, fiFor
The calculating function of i-th of fault degree index, N are total sample number, NTFor the measuring point sum of deployment, ojkFor in j-th of sample
The data of k-th of measuring point acquisition, 1≤j≤N, 1≤k≤NT。
3. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 2, feature exist
In: the time-frequency statistical indicator extracting method includes mean value, root-mean-square value, root amplitude, absolute mean, degree of skewness, kurtosis, side
Difference, peak value, standard deviation, peak-to-peak value, mean power, waveform index, peak index, pulse index, margin index, degree of skewness refer to
Mark, kurtosis index or means frequency.
4. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 2, feature exist
In: the realization process of the S1.4 are as follows: alternative multiple regression is set using formula (2) according to the number of measuring point and assesses mould
Type:
In formula, fiFor the calculating function of i-th of fault degree index, N is total sample number, NTFor the measuring point sum of deployment, ojkFor
The data that k-th of measuring point acquires in j-th of sample, 1≤j≤N, 1≤k≤NT,For assessment models coefficient.
5. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 4, feature exist
In: in the S1.5, calculate the calculation formula of the corresponding fault degree assessment models coefficient matrix of training sample are as follows:
In formula, (Bij)trainFor use i-th of model, j-th of index obtain the corresponding assessment models coefficient matrix of training sample,
(Xij)trainFor the training sample observation matrix for using j-th of index to calculate for i-th of model, YtrainIt is true for training sample
Real fault degree vectorNtrainFor the sum of training sample.
6. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 5, feature exist
In: in the S1.5, calculate the calculation formula of the fault degree assessment result of test sample are as follows:
In formula,To use fault degree model MiWith fault degree index FjAssess the obtained fault degree of test sample to
Amount, (Bij)trainFor the assessment models coefficient matrix for using training sample to obtain, calculated by formula (3), (Xij)trainFor for event
Hinder Degree Model MiWith fault degree index FjThe test sample observation matrix of calculating.
7. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 6, feature exist
In: the realization process of the S1.6 are as follows:
S1.6.1, N is calculated using formula (5)MN is used under a modelFA index calculates fault degree and assesses root-mean-square error matrix:
In formula, R is that the fault degree of test sample assesses root mean square matrix, and the corresponding fault degree model of the row of R arranges corresponding failure
Level index, NMFor fault degree assessment models sum, rijDefinition uses fault degree model MiWith fault degree index FjLower event
Hinder the root-mean-square error of scale evaluation;
S1.6.2, r is calculated using formula (6)ij:
In formula, NtestIt is total for test sample,To use fault degree model MiWith fault degree index FjFault degree comment
Estimate as a result, being calculated by formula (4), YtestFor the true fault degree vector of test sample,
8. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 7, feature exist
In: the realization process of the S1.7 are as follows: use the minimum corresponding assessment models of formula (7) optimum choice assessment of failure root-mean-square error
With fault degree index:
In formula, M*、F*The respectively minimum corresponding model of fault degree assessment root-mean-square error and fault degree index, according to mould
Type M*Expression formula determine optimal test points set T*With model parameter B*。
9. a kind of fault degree quantitative estimation method of multivariate regression models optimization according to claim 8, feature exist
In: it is described to carry out fault degree calculating process using the measuring point, assessment of failure model and fault degree index of optimization are as follows:
S2.1, the test points set T optimized using formula (7)*Acquire the data O under current statec;
S2.2, the fault degree index F optimized using formula (7)*Extract the fault degree index of the data under current state, building
Observation vector matrix Xc;
S2.3, the fault degree under current state is calculated using formula (8):
In formula, XcFor observation vector matrix, B*Optimal model parameter.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811215790 | 2018-10-18 | ||
CN201811215790X | 2018-10-18 |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109543357A true CN109543357A (en) | 2019-03-29 |
CN109543357B CN109543357B (en) | 2021-01-05 |
Family
ID=65834724
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910024090.0A Expired - Fee Related CN109543357B (en) | 2018-10-18 | 2019-01-10 | Fault degree quantitative evaluation method for multivariate regression model optimization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109543357B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110301230A (en) * | 2019-07-29 | 2019-10-08 | 中国农业大学 | A kind of combined harvester threshing cylinder fault simulation monitoring system and method |
CN110674891A (en) * | 2019-10-16 | 2020-01-10 | 北京天泽智云科技有限公司 | Data quality abnormity detection method for monitoring system |
CN112883569A (en) * | 2021-02-05 | 2021-06-01 | 吉林大学 | Method for analyzing fault propagation diffusion behavior of numerical control machine tool |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5487591B2 (en) * | 2008-10-28 | 2014-05-07 | Jfeスチール株式会社 | Diagnosis method of rotating equipment |
CN107101813A (en) * | 2017-04-26 | 2017-08-29 | 河北工业大学 | A kind of frame-type circuit breaker mechanical breakdown degree assessment method based on vibration signal |
CN107782551A (en) * | 2017-10-30 | 2018-03-09 | 电子科技大学 | Method for evaluating damage degree of mechanical part |
CN108092622A (en) * | 2017-12-14 | 2018-05-29 | 中国计量大学 | A kind of photovoltaic string formation method for diagnosing faults based on resistance calculations |
CN108596518A (en) * | 2018-05-14 | 2018-09-28 | 中国路桥工程有限责任公司 | A kind of Highway Geological Disaster risk assessment method |
-
2019
- 2019-01-10 CN CN201910024090.0A patent/CN109543357B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5487591B2 (en) * | 2008-10-28 | 2014-05-07 | Jfeスチール株式会社 | Diagnosis method of rotating equipment |
CN107101813A (en) * | 2017-04-26 | 2017-08-29 | 河北工业大学 | A kind of frame-type circuit breaker mechanical breakdown degree assessment method based on vibration signal |
CN107782551A (en) * | 2017-10-30 | 2018-03-09 | 电子科技大学 | Method for evaluating damage degree of mechanical part |
CN108092622A (en) * | 2017-12-14 | 2018-05-29 | 中国计量大学 | A kind of photovoltaic string formation method for diagnosing faults based on resistance calculations |
CN108596518A (en) * | 2018-05-14 | 2018-09-28 | 中国路桥工程有限责任公司 | A kind of Highway Geological Disaster risk assessment method |
Non-Patent Citations (1)
Title |
---|
李莹等: "基于隐马尔科夫模型的风电齿轮箱故障程度评估", 《太阳能学报》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110301230A (en) * | 2019-07-29 | 2019-10-08 | 中国农业大学 | A kind of combined harvester threshing cylinder fault simulation monitoring system and method |
CN110301230B (en) * | 2019-07-29 | 2023-05-23 | 中国农业大学 | Threshing cylinder fault simulation monitoring system and method for combine harvester |
CN110674891A (en) * | 2019-10-16 | 2020-01-10 | 北京天泽智云科技有限公司 | Data quality abnormity detection method for monitoring system |
CN110674891B (en) * | 2019-10-16 | 2021-11-30 | 北京天泽智云科技有限公司 | Data quality abnormity detection method for monitoring system |
CN112883569A (en) * | 2021-02-05 | 2021-06-01 | 吉林大学 | Method for analyzing fault propagation diffusion behavior of numerical control machine tool |
Also Published As
Publication number | Publication date |
---|---|
CN109543357B (en) | 2021-01-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Feng et al. | Digital twin-driven intelligent assessment of gear surface degradation | |
CN103868692B (en) | Based on the rotary machinery fault diagnosis method of Density Estimator and K-L divergence | |
Cui et al. | Quantitative fault analysis of roller bearings based on a novel matching pursuit method with a new step-impulse dictionary | |
Ghate et al. | Cascade neural-network-based fault classifier for three-phase induction motor | |
CN103776654B (en) | The method for diagnosing faults of multi-sensor information fusion | |
CN105736140B (en) | A kind of diesel engine flash speed measures and cylinder stops working trouble-shooter and method | |
CN109543357A (en) | Fault degree quantitative evaluation method for multivariate regression model optimization | |
CN109613428A (en) | It is a kind of can be as system and its application in motor device fault detection method | |
CN110006645B (en) | Multi-source fusion high-voltage circuit breaker mechanical fault diagnosis method | |
CN104390657B (en) | A kind of Generator Unit Operating Parameters measurement sensor fault diagnosis method and system | |
CN105116872A (en) | Fault diagnosis method based on metric learning and time sequence during industrial process | |
KR100867938B1 (en) | Prediction method for watching performance of power plant measuring instrument by dependent variable similarity and kernel feedback | |
CN101272580B (en) | Self-adapting mobile base station system reliability estimation method based on feedback | |
CN105138770B (en) | Space product Reliablility simulation appraisal procedure based on indirect characteristic quantities | |
CN110399986A (en) | A kind of generation method of pumping plant unit fault diagnosis system | |
CN105041631A (en) | Method and system for detecting vibration signal of driving shaft of gas compressor | |
CN111504647A (en) | AR-MSET-based performance degradation evaluation method for rolling bearing | |
CN117195137B (en) | Rotor die casting error detecting system based on data analysis | |
CN108204895A (en) | A kind of Wind turbines change gear box fault diagnosis method and system | |
CN103193125A (en) | Elevator hydraulic buffer resetting performance dynamic measuring method and tester adopted by method | |
CN111122162A (en) | Industrial system fault detection method based on Euclidean distance multi-scale fuzzy sample entropy | |
Ulmer et al. | Early fault detection based on wind turbine scada data using convolutional neural networks | |
CN115542099A (en) | Online GIS partial discharge detection method and device | |
Cheng et al. | Quantitative damage detection for planetary gear sets based on physical models | |
CN110414086A (en) | A kind of combined stress accelerated factor calculation method based on sensitivity |
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 | ||
CB03 | Change of inventor or designer information | ||
CB03 | Change of inventor or designer information |
Inventor after: Tan Xiaodong Inventor before: Tan Xiaodong Inventor before: Zuo Mingjian |
|
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210105 |