CN108763729A - Process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy - Google Patents
Process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy Download PDFInfo
- Publication number
- CN108763729A CN108763729A CN201810508428.5A CN201810508428A CN108763729A CN 108763729 A CN108763729 A CN 108763729A CN 201810508428 A CN201810508428 A CN 201810508428A CN 108763729 A CN108763729 A CN 108763729A
- Authority
- CN
- China
- Prior art keywords
- variable
- network structure
- dcca
- coupling
- entropy
- 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
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)
- Complex Calculations (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses the process industry Mechatronic Systems couple state appraisal procedures based on network structure entropy,First pass through the quasi-periodicity that FFT methods seek sequence,So that it is determined that the time window width of coupling analysis,The correlation of multivariable between any two is calculated using DCCA algorithms,The weighted network model of structure reflection Multivariable Coupling relationship,Can the coupled relation between acquisition process monitored parameters in time variation,The fast accurate of realization system upstream and downstream is dispatched,The fining management and control of realization system,NSEn methods calculate the entropy of monitored parameters coupled relation network model in each time window,Utilize the coupling incidence relation between monitoring data,It can intuitively reflect the Dynamic Coupling process of system different parts by the dynamic change of network topology structure,The state evolution process for carrying out quantitative characterization system of network structure entropy,Comprehensive scheduling and maintenance decision information are provided for system manager,Improve science and intelligent level to process industry complex electromechanical systems safe and reliable operation decision under complex working condition.
Description
Technical field
The present invention relates to complex electromechanical systems military service security state evaluation fields, and in particular to one kind being based on DCCA-NSEn
Process industry Mechatronic Systems coupling network modeling and appraisal procedure.
Background technology
Process industry production system production equipment is various, and needs various auxiliary systems, between each structural unit constantly
The exchange of substance, information, energy is carried out, internal system conjunction coupling degree is high, is a distributed complex electromechanical systems.If
Standby failure and technique adjustment frequently result in systematic fluctuation, the operation troubles found in industrial process promptly and accurately and rationally
The recovery extent for assessing failure process, is particularly important the Reasonable Regulation And Control of flow system upstream and downstream.Dispatcher according to
Past scheduling experience dispatches the upstream and downstream of system, usually for the sake of security, easy tos produce scheduling or scheduling not in time etc.
Situation causes the interruption of production or produce load to decline, thus may cause huge economic loss.Therefore, it finds effective
State evaluating method makes assessment promptly and accurately to system running state, and it is real-time to provide reliable system as dispatcher
State, the urgent need of reduction economic loss caused by due to scheduling discomfort.In terms of complex electromechanical systems synthetical condition assessment,
Li Li etc. is according to the variable synthesis concept of information data correlation rule and theory of factors space, it is proposed that a kind of to power transformer
The method for carrying out status assessment.Cheng Yingying, Yang Huaxiao etc. propose a kind of based on monitoring data progress for electric energy metering device
State evaluating method.Yao Yunfeng etc. is established on the basis of parameter health status is equipped in assessment based on improvement evidence theory
Equipment health state evaluation model, to the synthesis of the health status of all parameters and decision.These researchs are directed to transformer etc.
Different specific objects are studied, and achieve preferable achievement in research, but for the such operating mode of distributed complex Mechatronic Systems
Complexity, monitored parameters are numerous, and coupled relation is complicated, exists simultaneously monitoring information redundancy and insufficient object, research method is still
Compare deficient.
Invention content
The purpose of the present invention is to provide the process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy,
With overcome the deficiencies in the prior art.
In order to achieve the above objectives, the present invention adopts the following technical scheme that：
Process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy, specifically includes following steps：
Step 1), selection need the variables set of the monitoring objective for the complex electromechanical systems analyzed, monitored parameters collection to be supervised from DCS
The monitoring data collection of certain time course is obtained in examining system, the monitoring time sequence data collection obtained is that n ties up monitoring time
Sequence matrix；
Step 2) pre-processes the monitoring data collection obtained, specifically includes time series noise reduction, deletes information content
Few variable, the adaptive fusion and the determination of sliding window of redundancy feature；
Step 3), to pretreated monitoring data sequent data using going trend alternate analysis, determining between each pair of variable is
No there are coupled relations, if there are coupled relation calculating to remove trend interaction coefficent；
Step 4), using monitored parameters as node, coupled relation is side, and it is complicated that the size of the coefficient of coup is that the weight on side is established
Mechatronic Systems coupling network model；
Step 5), by setting sliding step, coupling network is established to each sliding window, it is fixed using network structure entropy
The couple state of analysis system is measured, system military service performance state evolution curve is formed, is commented to complete industrial Mechatronic Systems coupling
Estimate.
Further, monitoring data sequent sample frequency needs to be set according to sampling cost and monitoring accuracy, and sets sample
This length obtains monitoring data collection from system operation historical data.
Further, to the preprocess method of monitoring data collection, following steps are specifically included：
(1) heterologous data are subjected to data normalization, the data for normalizing later is dropped using method of wavelet packet
It makes an uproar；
(2) comentropy of time series is utilized to calculate the information content of each variable, given threshold R, removal information content is less than R
Variable；
(3) redundant variables are merged using adaptive weighted fusion method, reduces follow-up computation complexity；
(4) selection of time window width goes out the quasi-periodicity of time series to represent time series by Algorithm Analysis
When period of change, i.e. sequence length are more than the quasi-periodicity of variable, then this section of time series can preferably reflect the feature of variable.
Further, the quasi-periodicity of multiple variables with chaotic characteristic is calculated using fft algorithm, obtains n
The quasi-periodicity T of time series1, T2..., Tn, to enable sequence length as much as possible to reflect the feature of each variable, this
It is reference to sentence quasi-periodicity longest variable in n variable, chooses most macrocyclic 2 times of T=2max (T1,T2,…,Tn) conduct
The time window width of sequence determines the length of sequence.
Further, using going coupled relation of trend alternate analysis (DCCA) method between variable to carry out qualitative point
Analysis calculates the DCCA coefficients of its coupling variable pair, in this, as the stiffness of coupling between variable if there is coupling.
Further, the structure of DCCA coefficient networks is utilized：
For the time series x of each variables of n1,x2,x3,…,xn, calculate separately its DCCA coefficient between any two, DCCA
(x1,x1), DCCA (x1,x2) ..., DCCA (xn,xn), the DCCA square formations of a n × n are formed, it is as follows：
D in formula (11)11To dnnDCCA coefficients between variable, x1To xnFor the selected n variables to be assessed, wherein
dijDCCA coefficients between variable i and variable j, that is, represent the correlation between two variables；Between any two by n variable
The degree of coupling constitutes the degree of coupling network (DCCAnet) of such a n × n；Since DCCA methods are symmetrical, so dij's
Value and djiEqual, DCCAnet matrixes are a symmetrical matrixes.
Further, advantage of the application network structure entropy in terms of characterizing network heterogeneity, quantitative analysis monitoring data exist
The network structure entropy of system in each sliding window；According to different required precisions, sliding step-length STEP is determined and is adjusted
It is whole.
Network knot when system normal operation is calculated in the data set obtained when further, using system normal operation
The reasonable threshold value of structure entropy.
Further, the conjunction determined when normal beyond system according to the network structure changes of entropy curve being calculated in real time
The magnitude of threshold value is managed to quantitatively judge complex electromechanical systems operating status intensity of anomaly.
Compared with prior art, the present invention has technique effect beneficial below：
The present invention is based on the process industry Mechatronic Systems couple state appraisal procedures of network structure entropy, first pass through and seek sequence
Quasi-periodicity calculate the correlation of multivariable between any two, structure reflection multivariable so that it is determined that the time window width of coupling analysis
The weighted network model of coupled relation, monitoring data sequent time window are slid with certain step-length, and it is dynamic to obtain system coupled relation network
State evolutionary model, can the coupled relation between acquisition process monitored parameters in time variation, realize the fast accurate of system upstream and downstream
Scheduling, and then realize the fining management and control of system, it will help the safety military service for promoting production procedure industrial system is horizontal；Pass through
The entropy for calculating monitored parameters coupled relation network model in each time window changes with time trend to complicated machine according to entropy
The military service Evolution States of electric system carry out comprehensive assessment, using the coupling incidence relation between monitoring data, establish reflection flow work
The coupling network model of industry complex electromechanical systems operating mechanism, system will generate different coupling networks in the different phase of military service
Topological structure can intuitively reflect the Dynamic Coupling process of system different parts by the dynamic change of network topology structure,
Using the state evolution process come quantitative characterization system of network structure entropy, to provide comprehensive tune for system manager
Degree and maintenance decision information, to improve under complex working condition to the section of process industry complex electromechanical systems safe and reliable operation decision
The property learned and intelligent level；This method not only realizes the comprehensive characterization of system service state, but also can pass through network topology
The Dynamic Evolution of structure determines system exception position, realizes the complete perception of system service state, and then implement to system
Effective military service security management and control.
Further, this method can not only provide the overall target change curve of system, can also be by constructed
The intrinsic coupled relation network model of system and system part are found in the Dynamic Evolution of system coupling network topological structure
The evolution process of network structure provides comprehensive information for complex electromechanical systems operational decisions, pacifies to be on active service for realization system
Complete accurate management and control provides reliable basis.
Description of the drawings
Fig. 1 is that the present invention is based on the modeling of the complex electromechanical systems coupling network of DCCA-NSEn and estimation flow figures.
Fig. 2 is process industry complex electromechanical systems network modelling and evaluation process schematic diagram.
Fig. 3 is normal service state variations per hour 1-8 monitoring time Sequence Trend figures.
Fig. 4 is design sketch before and after system monitoring variable G_AVIR_0401 noise reductions, and (a) is system monitoring variable G_AVIR_
Design sketch before 0401 noise reduction is (b) design sketch after system monitoring variable G_AVIR_0401 noise reductions.
Fig. 5 is process industry complex electromechanical systems military service safe condition curve graph.
Fig. 6 is Coupled Variable network under process industry complex electromechanical systems different conditions.
Fig. 7 is to extract process industry complex electromechanical systems coupling feature figure using CDFA methods.
Specific implementation mode
The present invention is described in further detail below in conjunction with the accompanying drawings：
As shown in Figures 1 to 7, the present invention is a kind of is based on going answering for trend alternate analysis-network structure entropy (DCCA-NSEn)
Miscellaneous Mechatronic Systems Multivariable Coupling network modelling and comprehensive assessment, this method first pass through the quasi-periodicity that FFT methods seek sequence, from
And determine the time window width of coupling analysis, the correlation of multivariable between any two is calculated using DCCA algorithms, structure reflection is changeable
The weighted network model of coupled relation is measured, monitoring data sequent time window is slid with certain step-length, obtains system coupled relation network
Dynamic Evolution Model；The entropy that monitored parameters coupled relation network model in each time window is calculated using NSEn methods, according to
The entropy trend that changes with time carries out comprehensive assessment to the military service Evolution States of complex electromechanical systems, and this method, which not only realizes, is
The comprehensive characterization of system service state, and can determine system exception position by the Dynamic Evolution of network topology structure,
The complete perception of realization system service state, and then effective military service security management and control is implemented to system.
Process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy, specifically includes following steps：
Step 1), selection need the variables set of the monitoring objective for the complex electromechanical systems analyzed, monitored parameters collection to be supervised from DCS
The monitoring data collection of certain time course is obtained in examining system, the monitoring time sequence data collection obtained is that n ties up monitoring time
Sequence matrix；
Step 2) pre-processes the monitoring data collection obtained, specifically includes time series noise reduction, deletes information content
Few variable sets variable threshold L, deletes the variable of the few remaining variable threshold L of information content, redundancy feature it is adaptive merge and
The determination of sliding window；
Step 3), to pretreated monitoring data sequent data using going trend alternate analysis, determining between each pair of variable is
No there are coupled relations, if there are coupled relation calculating to remove trend interaction coefficent；
Step 4), using monitored parameters as node, coupled relation is side, and it is complicated that the size of the coefficient of coup is that the weight on side is established
Mechatronic Systems coupling network model；
Step 5), by setting sliding step, coupling network is established to each sliding window, it is fixed using network structure entropy
The couple state of analysis system is measured, system military service performance state evolution curve is formed, is commented to complete industrial Mechatronic Systems coupling
Estimate.
Monitoring data sequent sample frequency needs to be set according to sampling cost and monitoring accuracy, and sets the length of sample,
Monitoring data collection is obtained from system operation historical data；
To the preprocess method of monitoring data collection, following steps are specifically included：
(1) heterologous data are subjected to data normalization, the data for normalizing later is dropped using method of wavelet packet
It makes an uproar；
(2) comentropy of time series is utilized to calculate the information content of each variable, given threshold R, removal information content is less than R
Variable；
(3) redundant variables are merged using adaptive weighted fusion method, reduces follow-up computation complexity；
(4) selection of time window width goes out the quasi-periodicity of time series to represent time series by Algorithm Analysis
When period of change, i.e. sequence length are more than the quasi-periodicity of variable, then this section of time series can preferably reflect the feature of variable.
The quasi-periodicity of multiple variables with chaotic characteristic is calculated using fft algorithm, obtains n time series
Quasi-periodicity T1, T2..., Tn, to enable sequence length as much as possible to reflect the feature of each variable, this sentences n variable
Middle quasi-periodicity longest variable is reference, chooses most macrocyclic 2 times of T=2max (T1,T2,…,Tn) time window as sequence
Width determines the length of sequence.
Using going coupled relation of trend alternate analysis (DCCA) method between variable to carry out qualitative analysis, if there are couplings
It closes, then the DCCA coefficients of its coupling variable pair is calculated, using this as the stiffness of coupling between variable.
Utilize the structure of DCCA coefficient networks
For the time series x of each variables of n1,x2,x3,…,xn, calculate separately its DCCA coefficient between any two, DCCA
(x1,x1), DCCA (x1,x2) ..., DCCA (xn,xn), form the DCCA square formations of a n × n.It is as follows：
D in formula (11)11To dnnDCCA coefficients between variable, x1To xnFor the selected n variables to be assessed.Wherein,
dijDCCA coefficients between variable i and variable j, that is, represent the correlation between two variables.Between any two by n variable
The degree of coupling constitutes the degree of coupling network (DCCAnet) of such a n × n；Since DCCA methods are symmetrical, so dij's
Value and djiIt is equal, so DCCAnet matrixes are a symmetrical matrixes.
Advantage of the application network structure entropy in terms of characterizing network heterogeneity, quantitative analysis monitoring data are in each sliding window
The network structure entropy of system in mouthful.Here, according to different required precisions, sliding step-length STEP is determined and is adjusted.
The conjunction of network structure entropy when the data set obtained when using system normal operation is calculated system normal operation
Manage threshold value.
When system breaks down, there is abnormal, to be consequently formed grid topology in the coupled relation between each variable
Structure changes, this can lead to the variation that network structure entropy will appear, and exceed reasonable threshold value.
The reasonable threshold value determined when normal beyond system according to the network structure changes of entropy curve that is calculated in real time
Magnitude quantitatively judges complex electromechanical systems operating status intensity of anomaly.
Embodiment：
One complex electromechanical systems for containing n element carries out wanted production factors according to the specific time cycle
Sampling, is recorded in system variable, then after m sampling period, system shares m × n system variable data：
Step 1：Monitoring data collection pre-processes
It first has to the data of heterologous isomery carrying out data normalization, for each value in variable X, carries out such as following formula
Processing：
X (i)=X (i)/mean (X)
Noise reduction process, the method for using wavelet-packet noise reduction herein are carried out to normalizing later data.
Step 2：The selection of time window width
Go out the quasi-periodicity of time series by Algorithm Analysis to represent the period of change of time series.I.e. sequence length is more than
When the quasi-periodicity of variable, then this section of time series can preferably reflect the feature of variable.
We calculate the quasi-periodicity of multiple variables with chaotic characteristic using fft algorithm here, obtain n
The quasi-periodicity T of time series1, T2..., Tn.To enable sequence length as much as possible to reflect the feature of each variable, this
It is reference to sentence quasi-periodicity longest variable in n variable, chooses most macrocyclic 2 times of T=2max (T1,T2,…,Tn) conduct
The time window width of sequence determines the length of sequence；
Step 3：The structure of DCCA coefficient networks
For the time series x of each variables of n1,x2,x3,…,xn, calculate separately its DCCA coefficient between any two, DCCA
(x1,x1), DCCA (x1,x2) ..., DCCA (xn,xn), form the DCCA square formations of a n × n.It is as follows：
D in formula (11)11To dnnDCCA coefficients between variable, x1To xnFor the selected n variables to be assessed；Wherein,
dijDCCA coefficients between variable i and variable j, that is, represent the correlation between two variables；Between any two by n variable
The degree of coupling constitutes the degree of coupling network (DCCAnet) of such a n × n.Since DCCA methods are nondirectional so that dij
Value and djiIt is equal, so DCCAnet matrixes are a symmetrical matrixes；
Step 4：System military service comprehensive state characterization based on NSEn
Using the algorithm of network structure entropy, the network structure entropy of DCCAnet is solved, here, according to different monitoring essences
Degree requires, and can be adjusted to sliding step-length STEP.
The operation situation that complex electromechanical systems entirety is judged according to network structure changes of entropy curve, two below Main Basiss
Point：
1. when system is in normal operating condition, network structure entropy stabilization fluctuates in certain section；
2. when system breaks down, there is exception in each correlation of variables, and the network structure entropy of DCCAnet will appear substantially
The variation of degree, and exceed reasonable threshold value.
Monitoring time sequence variables select
Compressor set is as using chemical enterprise as the typical unit of the process industry of representative, and safe operation is for entirely flowing
Journey industrial processes stable operation is most important.When raw material is stablized, the monitoring of equipment data in usual production process can
Indirectly reflection industrial process operating status, the fluctuation of monitored parameters can embody industrial process failure indirectly, it is possible to answer
Restore evaluation studies for procedure fault early warning and process with typical process fault data.
The malfunction monitoring number of compressor set before and after an equipment component scram of the application application coal chemical industry enterprises
According to failure can be described as：System is just initially located in normal operating condition, and rear unusual service condition persistently occurs, operating personnel's
Scheduling is under manipulation, and system service state has certain improvement, but abnormality several times deteriorated again later, caused repeatedly
Most equipment must be forced scram, be overhauled.In the process, 8 prisons relatively high with the failure degree of correlation are selected
Measuring point position carries out accident analysis, and monitoring site details are as shown in table 1.The compressor set DCS monitoring data collection sampling intervals are
1min verifies the validity that this method assesses system mode based on the fault data of 8 monitored parameters.These variables
The trend of monitoring data sequent is as shown in Figure 3.
1 compressor set of table monitors argument table
2 data predictions
Carry out data analysis before, need first by primary monitoring data be converted into can united analysis time series, this punishment
It is handled for two steps, i.e. data normalization processing and data noise reduction process.
Monitoring data collection is normalized, heterologous isomeric data unit disunity can be eliminated, data scale is not united
The influence of a pair of of analysis result.
It is monitored after data set normalized, for the influence of noise for eliminating among production process, also needs to be dropped
It makes an uproar processing, the method for using wavelet-packet noise reduction herein.Noise reducing of data process is divided into two steps of decomposition and reconstruct：1. to difference
After variable takes suitable wavelet basis function and Decomposition order, with fixed late value method to each wavelet details coefficient after decomposition into
Row is soft to close value processing；2. reconstructing last layer of approximation coefficient and all layers of detail coefficients, the variable sequence diagram after noise reduction is obtained.
Its noise reduction is as shown in Figure 4.
3 seeking time sequences slide window width T
To make data as much as possible reflect the feature of data, improves DCCA algorithms and calculate the accurate of the degree of coupling between variable
Property, the quasi-periodicity for using fft algorithm to carry out each section of sequence herein is calculated, and takes the week of quasi-periodicity maximum time series
Phase is as slip time window T.I.e. T=max (T (1), T (2) ..., T (n)).8 variables have been selected to carry out comprehensive point herein
Analysis, so n=8.The quasi-periodicity solving result of each variable is as shown in the table：Then window width is 1667, i.e. time window width
It is 1667 minutes.
The quasi-periodicity of 2 variable 1-8 of table
Coupling Analysis between 4 monitored parameters and structure
By the coupling mutation analysis between variable two-by-two it is found that in failure whole process, the coupling between variable is closed
Different degrees of variation has occurred in system.
The abnormal state information that can partly reflect system that DCCA index variation trend between each variable has.But if only
Reflect the state of system using the degree of coupling of certain two variable, then system evaluation can be brought excessively unilateral, it is inaccurate to obtain
Really, even wrong conclusion.And selection variable work is complicated, when variable is more, when coupled relation is inenarrable between variable,
Need the method that status assessment can be made to system comprehensive information.
The correlation between variable is calculated with DCCA algorithms, and builds monitored parameters degree of coupling network.Following formula is normal shape
The degree of coupling network of state curve paragraph 1 window.
During Coupling Degrees, the degree of coupling of two variables can be judged with DCCA coefficients.Criterion such as following table.
Coupling description of the table 3 based on DCCA coefficients
Here, to make analysis result tend towards stability, identification is high, removes weak coherent element, removal rule such as following formula
It is shown.
The 5 system service state Evolution analyses based on NSEn
The normal and abnormal data of 8 variables has been selected to compare and analyze herein, the time window of window width T=1666
With certain step-length STEP in the enterprising line slip of time series, STEP values are 200 herein, acquire the degree of coupling net of each period
Network, and solve its network structure entropy.It calculates the variation of its network structure entropy and to draw curve.And choose represent system normally and
The point of abnormality carries out coupling network model construction, and state change of the analysis system in different phase causes between each variable
The variation of coupled relation network.The military service security state evaluation curve of system different phase is as shown in Figure 5.
By analyzing, network structure entropy when system is in normal operating condition should fluctuate between 0.92 to 0.94.
Before scram, system is just initially located in abnormality, corresponding, and the network structure entropy of abnormality is bent
Line then significantly deviates from normal region.There is primary artificial intervention adjustment halfway, the military service quality state of system is made to have
The improvement of transience, later failure further deteriorate, NSEn values have deviated significantly from normal region again, even up to reach
0.98, and last from days, eventually lead to scram.
In conclusion DCCA-NSEn methods disclosed in this invention perceive sensitivity to the abnormality of system, with Fig. 7's
The multiple features curve of CDFA is compared, and this method, which not only assesses the normal condition of system, to be stablized, and can pass through network structure
Variation clearly find system exception node, the variation of network structure is as shown in fig. 6, so as to the abnormal shape to system
State makes comprehensive timely early warning, and reference is provided for the maintenance maintenance decision of dispatcher.
The present invention is directed to the coupled relations between the fully monitoring data sequent of application reflection system running state, by research and application
Time series coupling goes trend fluction analysis method with the method for network structure entropy with being combined, it is proposed that one kind is based on more
Complex electromechanical systems modeling and the comprehensive assessment new method of variable, specific effect are as follows：
Compared to single arguments, two time series variation analysis methods such as DFA, DCCA, MFDFA and MFDCCA, this method
Advantage with multiple variable synthetical assessment.
Compared to Multivariate Time Series analysis methods such as PCA, KPCA and CDFA, has to assess normal condition and imitate
Fruit is stablized, the advantage sensitive to abnormality perception.
DCCA-NSEn methods can not only provide the overall target change curve of system, can also pass through constructed system
System intrinsic coupled relation network model and system local area network are found in the Dynamic Evolution of coupling network topological structure
The evolution process of network structure provides comprehensive information for complex electromechanical systems operational decisions, safe to be on active service for realization system
Accurate management and control provide reliable basis.
Claims (9)
1. the process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy, which is characterized in that specifically include with
Lower step：
Step 1), selection need the variables set of the monitoring objective for the complex electromechanical systems analyzed, and monitored parameters collection is monitored from DCS is
The monitoring data collection of certain time course is obtained in system, the monitoring time sequence data collection obtained is that n ties up monitoring time sequence
Matrix；
Step 2) pre-processes the monitoring data collection obtained, specifically includes time series noise reduction, and it is few to delete information content
Variable, the adaptive fusion and the determination of sliding window of redundancy feature；
Step 3), to pretreated monitoring data sequent data using going trend alternate analysis, determine whether deposit between each pair of variable
In coupled relation, if there are coupled relation calculating to remove trend interaction coefficent；
Step 4), using monitored parameters as node, coupled relation is side, and it is complicated electromechanical that the size of the coefficient of coup is that the weight on side is established
System coupling network model；
Step 5), by set sliding step, coupling network is established to each sliding window, is quantitatively divided using network structure entropy
The couple state of analysis system forms system military service performance state evolution curve, to complete industrial Mechatronic Systems coupling assessment.
2. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It being characterized in that, monitoring data sequent sample frequency needs to be set according to sampling cost and monitoring accuracy, and sets the length of sample,
Monitoring data collection is obtained from system operation historical data.
3. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, to the preprocess method of monitoring data collection, specifically includes following steps：
(1) heterologous data are subjected to data normalization, noise reduction is carried out using method of wavelet packet to normalizing later data；
(2) comentropy of time series is utilized to calculate the information content of each variable, given threshold R, removal information content is less than the change of R
Amount；
(3) redundant variables are merged using adaptive weighted fusion method, reduces follow-up computation complexity；
(4) selection of time window width goes out the quasi-periodicity of time series to represent the variation of time series by Algorithm Analysis
When period, i.e. sequence length are more than the quasi-periodicity of variable, then this section of time series can preferably reflect the feature of variable.
4. the process industry Mechatronic Systems couple state appraisal procedure according to claim 3 based on network structure entropy,
It is characterized in that, the quasi-periodicity of multiple variables with chaotic characteristic is calculated using fft algorithm, obtains n time series
Quasi-periodicity T1, T2..., Tn, to enable sequence length as much as possible to reflect the feature of each variable, this sentences n change
Quasi-periodicity longest variable is reference in amount, chooses most macrocyclic 2 times of T=2max (T1,T2,…,Tn) time as sequence
Window width determines the length of sequence.
5. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, using going coupled relation of the trend alternate analysis method between variable to carry out qualitative analysis, if there is coupling,
The DCCA coefficients for calculating its coupling variable pair, in this, as the stiffness of coupling between variable.
6. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, utilizes the structure of DCCA coefficient networks：
For the time series x of each variables of n1,x2,x3,…,xn, calculate separately its DCCA coefficient between any two, DCCA (x1,
x1), DCCA (x1,x2) ..., DCCA (xn,xn), the DCCA square formations of a n × n are formed, it is as follows：
D in formula (11)11To dnnDCCA coefficients between variable, x1To xnFor the selected n variables to be assessed, wherein dijFor
DCCA coefficients between variable i and variable j, that is, represent the correlation between two variables；By the coupling of n variable between any two
Degree, constitutes the degree of coupling network of such a n × n；Since DCCA methods are symmetrical, so dijValue and djiIt is equal,
DCCAnet matrixes are a symmetrical matrixes.
7. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, advantage of the application network structure entropy in terms of characterizing network heterogeneity, quantitative analysis monitoring data are in each sliding
The network structure entropy of system in window；According to different required precisions, sliding step-length STEP is determined and is adjusted.
8. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, the conjunction of the network structure entropy when data set obtained when using system normal operation is calculated system normal operation
Manage threshold value.
9. the process industry Mechatronic Systems couple state appraisal procedure according to claim 1 based on network structure entropy,
It is characterized in that, the reasonable threshold value determined when normal beyond system according to the network structure changes of entropy curve that is calculated in real time
Magnitude quantitatively judges complex electromechanical systems operating status intensity of anomaly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810508428.5A CN108763729B (en) | 2018-05-24 | 2018-05-24 | Process industry electromechanical system coupling state evaluation method based on network structure entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810508428.5A CN108763729B (en) | 2018-05-24 | 2018-05-24 | Process industry electromechanical system coupling state evaluation method based on network structure entropy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108763729A true CN108763729A (en) | 2018-11-06 |
| CN108763729B CN108763729B (en) | 2020-07-28 |
Family
ID=64005597
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810508428.5A Active CN108763729B (en) | 2018-05-24 | 2018-05-24 | Process industry electromechanical system coupling state evaluation method based on network structure entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108763729B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109614451A (en) * | 2018-12-06 | 2019-04-12 | 南京邮电大学 | Industrial big data intellectual analysis decision making device |
| CN109634233A (en) * | 2018-12-06 | 2019-04-16 | 南京邮电大学 | Industrial big data intellectual analysis decision-making technique, readable storage medium storing program for executing and terminal |
| CN110032146A (en) * | 2019-04-24 | 2019-07-19 | 西安交通大学 | A kind of complicated processing process stability appraisal procedure based on the multi-machine collaborative factor |
| CN111008363A (en) * | 2019-11-21 | 2020-04-14 | 西安交通大学 | Multivariable causal-driven complex electromechanical system service safety situation evaluation method |
| CN111080074A (en) * | 2019-11-21 | 2020-04-28 | 西安交通大学 | System service security situation element obtaining method based on network multi-feature association |
| CN111582603A (en) * | 2020-05-19 | 2020-08-25 | 中煤科工集团重庆研究院有限公司 | Intelligent early warning method for coal and gas outburst based on multi-source information fusion |
| CN111861272A (en) * | 2020-07-31 | 2020-10-30 | 西安交通大学 | Multi-source data-based complex electromechanical system abnormal state detection method |
| CN112001295A (en) * | 2020-08-19 | 2020-11-27 | 北京航天飞行控制中心 | Performance evaluation method and device for high-speed rotor shafting, storage medium and processor |
| CN112116198A (en) * | 2020-08-04 | 2020-12-22 | 西安交通大学 | Data-driven process industrial state perception network key node screening method |
| CN113411821A (en) * | 2021-06-18 | 2021-09-17 | 北京航空航天大学 | System reconfiguration capability evaluation method and system for complex network |
| CN114637263A (en) * | 2022-03-15 | 2022-06-17 | 中国石油大学(北京) | Method, device and equipment for monitoring abnormal working conditions in real time and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8739081B2 (en) * | 2010-04-30 | 2014-05-27 | International Business Machines Corporation | Method and system for computing fourier series coefficients for mask layouts using FFT |
| CN106682835A (en) * | 2016-12-29 | 2017-05-17 | 西安交通大学 | Data-driven complex electromechanical system service quality state evaluation method |
| CN107248047A (en) * | 2017-08-01 | 2017-10-13 | 西安交通大学 | Process method for evaluating state based on process state entropy dynamic calculation |
| CN107292061A (en) * | 2017-07-28 | 2017-10-24 | 西安交通大学 | A kind of process industry complex electromechanical systems information modelling approach of data-driven |
-
2018
- 2018-05-24 CN CN201810508428.5A patent/CN108763729B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8739081B2 (en) * | 2010-04-30 | 2014-05-27 | International Business Machines Corporation | Method and system for computing fourier series coefficients for mask layouts using FFT |
| CN106682835A (en) * | 2016-12-29 | 2017-05-17 | 西安交通大学 | Data-driven complex electromechanical system service quality state evaluation method |
| CN107292061A (en) * | 2017-07-28 | 2017-10-24 | 西安交通大学 | A kind of process industry complex electromechanical systems information modelling approach of data-driven |
| CN107248047A (en) * | 2017-08-01 | 2017-10-13 | 西安交通大学 | Process method for evaluating state based on process state entropy dynamic calculation |
Non-Patent Citations (4)
| Title |
|---|
| BORIS PODOBNIK等: "《Detrended Cross-Correlation Analysis: A New Method for Analyzing Two Nonstationary Time Series》", 《PHYSICAL REVIEW LETTERS》 * |
| GAO ZHIYONG等: "《Coupling analysis-based false monitoring information identification of production system in process industry》", 《SCIENCE CHINA》 * |
| 蔡萌等: "《一种基于点和边差异性的网络结构熵》", 《物理学报》 * |
| 谭跃进等: "《网络结构熵及其在非标度网络中的应用》", 《***工程理论与实践》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109634233A (en) * | 2018-12-06 | 2019-04-16 | 南京邮电大学 | Industrial big data intellectual analysis decision-making technique, readable storage medium storing program for executing and terminal |
| CN109614451A (en) * | 2018-12-06 | 2019-04-12 | 南京邮电大学 | Industrial big data intellectual analysis decision making device |
| CN110032146A (en) * | 2019-04-24 | 2019-07-19 | 西安交通大学 | A kind of complicated processing process stability appraisal procedure based on the multi-machine collaborative factor |
| CN110032146B (en) * | 2019-04-24 | 2020-10-27 | 西安交通大学 | Complex machining process stability evaluation method based on multi-machine synergistic factors |
| CN111080074B (en) * | 2019-11-21 | 2022-07-12 | 西安交通大学 | System service security situation element obtaining method based on network multi-feature association |
| CN111008363A (en) * | 2019-11-21 | 2020-04-14 | 西安交通大学 | Multivariable causal-driven complex electromechanical system service safety situation evaluation method |
| CN111080074A (en) * | 2019-11-21 | 2020-04-28 | 西安交通大学 | System service security situation element obtaining method based on network multi-feature association |
| CN111582603A (en) * | 2020-05-19 | 2020-08-25 | 中煤科工集团重庆研究院有限公司 | Intelligent early warning method for coal and gas outburst based on multi-source information fusion |
| CN111861272A (en) * | 2020-07-31 | 2020-10-30 | 西安交通大学 | Multi-source data-based complex electromechanical system abnormal state detection method |
| CN112116198A (en) * | 2020-08-04 | 2020-12-22 | 西安交通大学 | Data-driven process industrial state perception network key node screening method |
| CN112116198B (en) * | 2020-08-04 | 2023-06-20 | 西安交通大学 | Data-driven process industrial state perception network key node screening method |
| CN112001295A (en) * | 2020-08-19 | 2020-11-27 | 北京航天飞行控制中心 | Performance evaluation method and device for high-speed rotor shafting, storage medium and processor |
| CN112001295B (en) * | 2020-08-19 | 2023-12-08 | 北京航天飞行控制中心 | Performance evaluation method and device of high-speed rotor shaft system, storage medium and processor |
| CN113411821B (en) * | 2021-06-18 | 2021-12-03 | 北京航空航天大学 | System reconfiguration capability evaluation method and system for complex network |
| CN113411821A (en) * | 2021-06-18 | 2021-09-17 | 北京航空航天大学 | System reconfiguration capability evaluation method and system for complex network |
| CN114637263A (en) * | 2022-03-15 | 2022-06-17 | 中国石油大学(北京) | Method, device and equipment for monitoring abnormal working conditions in real time and storage medium |
| CN114637263B (en) * | 2022-03-15 | 2024-01-12 | 中国石油大学(北京) | Abnormal working condition real-time monitoring method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108763729B (en) | 2020-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108763729A (en) | Process industry Mechatronic Systems couple state appraisal procedure based on network structure entropy | |
| CN110414155B (en) | Fan component temperature abnormity detection and alarm method with single measuring point | |
| CN109240244B (en) | Data-driven equipment running state health degree analysis method and system | |
| CN104732276B (en) | One kind metering production facility on-line fault diagnosis method | |
| CN107707431A (en) | The data safety monitoring method and system of a kind of facing cloud platform | |
| CN106407589B (en) | Fan state evaluation and prediction method and system | |
| CN104573850A (en) | Method for evaluating state of thermal power plant equipment | |
| CN105930629B (en) | A kind of on-line fault diagnosis method based on magnanimity service data | |
| CN113036913B (en) | Method and device for monitoring state of comprehensive energy equipment | |
| CN103512751A (en) | Bearing health state identification method based on probabilistic neural network | |
| WO2022171788A1 (en) | Prediction model for predicting product quality parameter values | |
| CN116050665B (en) | Heat supply equipment fault prediction method | |
| CN113554361B (en) | Comprehensive energy system data processing and calculating method and processing system | |
| CN114169254A (en) | Abnormal energy consumption diagnosis method and system based on short-term building energy consumption prediction model | |
| CN109491339B (en) | Big data-based substation equipment running state early warning system | |
| CN103440410A (en) | Main variable individual defect probability forecasting method | |
| CN114519923A (en) | Intelligent diagnosis and early warning method and system for power plant | |
| CN112417627A (en) | Power distribution network operation reliability analysis method based on four-dimensional index system | |
| CN116520739A (en) | Intelligent heating system self-healing control platform and method based on multiple intelligent agents | |
| CN104778340A (en) | Bearing life prediction method based on enhanced particle filter | |
| CN117494009A (en) | Electrical equipment state evaluation method based on insulating material pyrolysis analysis and cloud platform | |
| CN112418662A (en) | Power distribution network operation reliability analysis method using artificial neural network | |
| CN115936663A (en) | Maintenance method and device for power system | |
| CN111367255A (en) | Performance evaluation test system and method for multi-variable control system | |
| Aftabi et al. | A Variational Autoencoder Framework for Robust, Physics-Informed Cyberattack Recognition in Industrial Cyber-Physical Systems |
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 |