CN108549346B - A kind of historical data section automatic searching method suitable for System Discrimination - Google Patents
A kind of historical data section automatic searching method suitable for System Discrimination Download PDFInfo
- Publication number
- CN108549346B CN108549346B CN201810458154.3A CN201810458154A CN108549346B CN 108549346 B CN108549346 B CN 108549346B CN 201810458154 A CN201810458154 A CN 201810458154A CN 108549346 B CN108549346 B CN 108549346B
- Authority
- CN
- China
- Prior art keywords
- data segment
- data
- system discrimination
- value
- searching method
- 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
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Manufacturing & Machinery (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
Abstract
The invention discloses a kind of historical data section automatic searching methods suitable for System Discrimination, it include: the different data section for finding probability distribution and there is variation, judge whether each data segment passes through an apparent amplitude variation or for a steady state value, it verifies between two adjacent data sections with the presence or absence of significant difference, obtains the data segment that can be used for System Discrimination.One aspect of the present invention overcomes existing system identification experiment because needing especially to apply pumping signal, and can interfere normal industrial production, and be difficult the problem carried out in practice；On the other hand, the present invention, which can be used in the data length that prolonged data sample is searched for, and searched for, can satisfy the requirement of user.It is indispensable a part that the present invention evaluates control loop, diagnoses and improve in performance in process industry.
Description
Technical field
The present invention relates to the fields of automation technology such as process control, more particularly to a kind of history number suitable for System Discrimination
According to section automatic searching method.
Background technique
In process industry, the mathematical model of dynamic process is frequently used for evaluation, diagnosis and the performance for improving control system.
The performance of closed-loop control system has very important work for the safe and highly efficient operation of the process industries such as petroleum, chemical industry and power generation
With.The census of manufacturing shows that many industrial control systems are faced with the phenomenon of performance difference, this make product quality, operating cost and
Process safety etc. produces negative effect very serious.
In thermal power plant, mistake that the primary process variables such as coal-supplying amount, quantity of steam and confluent change in active power
Cheng Zhonghui generation is widely varied, and the relevant devices such as coal pulverizer, blower and water pump can be run under a variety of operating conditions, therefore this
The controlled device of a little control systems has different kinetic characteristics.
The degradation that will lead to control system is mismatched between controller parameter and controlled device, therefore in a kind of operation
The controller parameter designed under operating condition may not be suitable for the dynamic process under another operating condition.In order to improve control system
Performance, need to readjust control parameter.System Discrimination is widely used in process industry, controlled for obtaining
The mathematical model of object, and then control parameter is readjusted according to mathematical model.
It can usually be tested by the identification of special designing, for recognizing the mathematical model of dynamic process.But this germline
System identification experiment needs especially to apply pumping signal, can interfere normal industrial production, so this experiment is difficult in practice
It carries out.
In conclusion distinguishing in the prior art to how finding suitable system from a large amount of history data of control system
The data segment of knowledge, still shortage effective solution scheme.
Summary of the invention
In order to solve the deficiencies in the prior art, the present invention provides a kind of historical data sections suitable for System Discrimination to look into automatically
Look for method.On the one hand, the present invention overcomes existing system identification experiments because needing especially to apply pumping signal, and can interfere normal
Industrial production, and in practice be difficult carry out problem；On the other hand, the present invention can be used in prolonged data sample
Search, and the data length searched for can satisfy the requirement of user.
A kind of historical data section automatic searching method suitable for System Discrimination, comprising the following steps:
The data sample for obtaining the time signal composition of series of discrete changes detection using the probability distribution of imparametrization
Method obtains the position of change point in data sample, finds the different data section that probability distribution has variation；
Judge whether each data segment passes through an apparent amplitude variation or for a steady state value, using the latter as
Null hypothesis, the former, it is assumed that form hypothesis testing, judges whether data segment can be used as system according to hypothesis testing as substitution
Identification Data section；
Based on the assumption that indicator sequence is all to verify in the case where zero between two adjacent data sections with the presence or absence of aobvious in examining
Write difference；
The setting value of comprehensive closed-loop control system obtains the data segment that can be used for System Discrimination.
Further preferred technical solution, the probability distribution change detecting method using imparametrization obtain data sample
The position of change point in this, specifically:
(1) current data segment x (t is initializedk:tk+ K-1)={ x (tk),x(tk+1),...,x(tk+ K-1) }, it includes
Entire data sample x (1), x (2) ..., x (N)；
(2) a current data section x (t is calculatedk:tk+ K-1) interior x (t) a relative position, be denoted as:
Wherein, t=tk,tk+1,…,tk+ K-1,
Provide statistic U (t)=U (t-1)+V (t), initial value U (a tk- 1)=0, it is specified that a hypothesis testing is yes
No change location is the maximum value of time point U (t) absolute value, is denoted as:The P value of hypothesis testing is remembered
Make:If the value of P is less than the probability value of Error type I α, tmaxExactly change point
Position, otherwise, tmaxIt is not just the position of change point；
(3) current data section x (tk:tk+ K-1) divide x (tk:tmax)={ x (tk,x(tk+1),...,x(tmax) and x
(tmax+1:tk+ K-1)={ x (tmax+1),x(tmax+2),...,x(tk+ K-1) } two data segments, tmaxIt is the position of variation；
(4) position of the step of the repeating (2) and (3) to each data segment until can not find variation, the length of data segment is not
Less than the parameter K of user's selection0, minimum length as data segment.
Further preferred technical solution a, if data segment, length is that K is greater than 2K0, then the data segment should
It is divided into K/K0Section, by operation, can obtain the position t of all changes pointi' s (i=1,2 ..., I), wherein I is change point
Number.
Further preferred technical solution, if a data segment x (ti:ti+1- 1) it is optional that range is greater than a user
The parameter d selected0, d0Change the smallest value for sample significance, null hypothesis is rejected, that is to say, that the previous case is received, then
x(ti:ti+1- 1) it can be used as System Discrimination data segment.
Further preferred technical solution, in a data segment x (ti:ti+1- 1) when whether being used as System Discrimination data segment
With index series Ix(t) indicate that can data segment be used to System Discrimination, if data segment can be used to System Discrimination, Ix(ti:
ti+1- 1)=1；Otherwise Ix(ti:ti+1- 1)=0.
Further preferred technical solution, when verifying between two adjacent data sections with the presence or absence of significant difference, with hypothesis
It examines to compare the average value of two adjacent data sections.
Further preferred technical solution, when comparing the average value of two adjacent data sections with hypothesis testing, for mark
Quasi- Gaussian ProfileGreater than a determining probability γ of Error type I type, here
γ is the positive number of a very little,And NiIt is the length of sample average, standard deviation and data segment respectively, then, they put down
Equal difference is greater than minimum conspicuousness changing value d0If second data segment have one relative to first apparent width of data segment
Degree variation, then the index I of the second segment dataxValue become 1 from 0.
The setting value of further preferred technical solution, comprehensive closed-loop control system obtains the data that can be used for System Discrimination
Duan Shi defines an indicator sequence I (t), works as IrOr I (t)=1y(t)=1 when, I (t)=1, it may be assumed that I (t)=Ir(t)∪Iy
(t), t is usedk,sAnd tk,eThe time of k-th of data segment beginning and end is respectively indicated, I (t)=1 is continuous, it may be assumed thatI(tk,s- 1)=0, I (tk,e+ 1)=0；Wherein Ir(t) and IyIt (t) is to use closed-loop control system
Setting value r (t) and controlled volume y (t) substitution x (t), and the System Discrimination data segment that respectively obtains is applicable in index series.
Further preferred technical solution, if r (t) does not change in k-th of data segment, this data segment cannot
For System Discrimination, corresponding index series becomes at this time: Ix(tk,s:tk,e)=0, the data segment value for System Discrimination are
Ix(t)=1, point I at other timesx(t)=0.
Further preferred technical solution, the above-mentioned historical data section automatic searching method suitable for System Discrimination of the application
System suitable for multiple input single output.
Compared with prior art, the beneficial effects of the present invention are:
The invention proposes a kind of to search for the new method for being suitable for System Discrimination data segment, search condition from historical data
It is that apparent variation occur in setting value and controlled volume in closed-loop control system.On the one hand, the present invention overcomes existing systems to distinguish
Experiment is known because needing especially to apply pumping signal, and can interfere normal industrial production, and be difficult the difficulty carried out in practice
Topic；On the other hand, the present invention can be used in prolonged data sample and search for, and the data length searched for can satisfy user
Requirement.It is indispensable a part that the present invention evaluates control loop, diagnoses and improve in performance in process industry.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 is closed-loop control system structural schematic diagram；
Fig. 2 is System Discrimination historical data section automatic searching method flow chart of the invention；
Fig. 3 (a) to Fig. 3 (c) is respectively the tendency chart of r (t), y (t), u (t) and d (t) in 24 hours；
Fig. 4 is the I in step 2y(t) it is represented by dashed line, the I in step 3y(t) it is indicated with point；
Fig. 5 (a) is the signal graph of y (t)；
Fig. 5 (b) is the signal graph of u (t), and Fig. 5 (a) and Fig. 5 (b) are used to look for the position of first change point；
Fig. 6 (a) is that y (t) ∈ [7405,8493] compare(solid line) and(dotted line)；
Fig. 6 (b) is that y (t) ∈ [8494,0721] compare(solid line) and(dotted line)；
R (t) is indicated by the solid line in Fig. 7 (a), and y (t) is indicated by a dotted line, Ir(t) it is indicated with dotted line；
Fig. 7 (b) is u (t) signal graph；
Fig. 7 (c) is d (t) signal graph.
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
In view of the closed-loop control system as described in Fig. 1, y (t) and u (t) indicates the output of dynamic process G and defeated here
Enter.R (t) is the setting value of closed-loop control system, and d (t) is the external disturbance for influencing closed-loop control system, v (t) be one can not
The noise of survey.Controller C changes the value of u (t) by different y (t) and r (t), and q is delay factor, such as q-1X (t)=x (t-
1), if dynamic process G and interference model H are linearly invariant, the system of a multiple input single output be may be expressed as:
Y (t)=G (q-1)u(t)+H(q-1)d(t)
The purpose of System Discrimination is to obtain the mathematical model for being equal or close to G from the historical data section of collection.Cause
This, problem to be solved is, how given historical data section r (t), y (t), u (t) wherein t=1,2 ..., N find out suitable
Close the data segment of identification G.The data sample collected in this embodiment example is the data of industry spot.
As shown in Fig. 2, System Discrimination historical data section automatic searching method operational flow diagram of the invention, comprising:
Step 1: find the different data section that probability distribution has variation: the data sample being collected into is (1) x, x
..., (2), the time signal x (t) of the series of discrete such as x (N).Existing change point method for detecting position is based on priori knowledge
It is assumed that these hypothesis are difficult to verify whether to set up in practice.Therefore the probability distribution for using an imparametrization changes inspection
Survey method obtains the position of change point.Data sample in the step is to refer to general signal, and wherein r (t) and y (t) is to be
Specific implementation example signal in system.
Specific step is as follows for this method:
(1) current data segment x (t is initializedk:tk+ K+1)={ x (tk),x(tk+1),...,x(tk+ K-1) }, it includes
Entire data sample x (1), x (2) ... x (N).
Wherein, tkIndicate that first sample moment of current data section in kth time circulation, K are the sample of current data section
Number, as k=1, tk=1, i.e. x (tk)=x (1), x (tk+ K-1)=x (N).
(2) a current data section x (t is calculatedk:tk+ K-1) interior x (t) a relative position, be denoted as:
Wherein, t=tk,tk+1,...,tk+K-1。
Provide statistic U (t)=U (t-1)+V (t), wherein t=tk,tk+1,...,tk+K-1.Initial value U (tk-
1)=0.Provide a hypothesis testing be whether change location be time point U (t) absolute value maximum value, be denoted as:The P value of hypothesis testing is denoted as:
If the value of P is less than the probability value of Error type I α, tmaxIt is exactly the position of change point, otherwise, tmaxJust not
It is the position of change point, it is assumed that the P value of inspection is the probability of Error type I.
(3) current data section x (tk:tk+ K-1) divide x (tk:tmax)={ x (tk,x(tk+1),...,x(tmax) and x
(tmax+1:tk+ K-1)={ x (tmax+1),x(tmax+2),...,x(tk+ K-1) } two data segments, tmaxIt is the position of variation.
(4) position of the step of the repeating (2) and (3) to each data segment until can not find variation, the length of data segment is not
Less than the parameter K of user's selection0, minimum length as data segment.On the other hand, in order to avoid too long data segment, if
One data segment, length is that K is greater than 2K0, then the data segment should be divided into K/K0Section.By the operation of step 1, can obtain
To the position t of all changes pointi' s (i=1,2 ..., I), I is the number of change point.
Step 2: judge whether each data segment passes through an apparent amplitude variation (i.e. value range for data segment
Greater than parameter d0) or be a steady state value.Using the latter as null hypothesis, i.e., data segment is a steady state value, the former is as replacing
In generation, is it is assumed that i.e. there are amplitude variations for data segment, to form hypothesis testing.If x (ti:ti+1- 1) range is greater than a use
The selectable parameter d in family0(sample significance changes the smallest value), null hypothesis is rejected, that is to say, that substitution is assumed to be received.
So, x (ti:ti+1- 1) it can be used as System Discrimination data segment.Herein, with indicator sequence Ix(t) indicate that can data segment be made
For the data segment of System Discrimination, if data segment can be used to System Discrimination, then Ix(ti:ti+1- 1)=1；Otherwise Ix(ti:ti+1-
1)=0.
In the null hypothesis of this paper, the probability of each samples span mean value is identical, therefore crosses over two phases of mean value
Adjacent time difference obeys exponential distribution.T is denoted as at the time of by samples span mean valuel' s, l=1,2 ..., L, L be that samples span is equal
It is worth the number at moment, these moment are defined by the formulaWherein, ti≤tl≤ti+1-1.Use Tl
=tl+1-tlIndicate poor across two adjacent moments of mean value.Define DTIt is the difference of exponential distribution function and estimation distribution function
Maximum value, such as following formula:
In formula,It is the exponential distribution function of stochastic variable T, expression formula is as follows:
Wherein,It is estimation distribution function, expression formula is as follows:
Wherein,
If DTGreater than corresponding threshold value, then the adjacent moment difference across mean value is not just exponential distribution, so zero is false
If being rejected.If x (ti:ti+1- 1) range is greater than an at user option parameter d0(sample significance variation is the smallest
Value).So, x (ti:ti+1- 1) it can be used as the candidate data section of System Discrimination.Herein, with index series Ix(t) it indicates, if
Data segment can be used to System Discrimination, then Ix(ti:ti+1- 1)=1；Otherwise Ix(ti:ti+1- 1)=0.
Step 3: verifying two adjacent data sections between whether there is significant difference: the step for be to refer in second step
It is carried out in the case that show sequence all be zero.Compare the average value of two adjacent data sections with hypothesis testing.For standard height
This distributionGreater than a determining probability γ of Error type I type, γ is here
The positive number of one very little.And NiIt is sample average, the length of standard deviation and data segment respectively.So, their mean difference
It is different to be greater than minimum conspicuousness changing value d0.If second data segment has one to become relative to the apparent amplitude of first data segment
Change, then the index I of the second segment dataxValue become 1 from 0.
Step 4: comprehensive r (t) and y (t) obtain the data segment that can be used for System Discrimination.An indicator sequence I (t) is defined,
Work as Ir(t) or Iy(t) when being equal to 1, I (t)=1, wherein Ir(t) and IyIt (t) is the setting value for using closed-loop control system
R (t) and controlled volume y (t) substitution x (t), and the System Discrimination data segment respectively obtained is applicable in index series；That is:
I (t)=Ir(t)∪Iy(t) (7)
Wherein, t is usedk,sAnd tk,eThe time for respectively indicating k-th of data segment beginning and end is that I (t)=1 is continuous,
That is:I(tk,s- 1)=0, I (tk,e+ 1)=0.But if r (t) does not become in k-th of data segment
Change, then this data segment cannot be used to System Discrimination, corresponding index series becomes at this time: I (tk,s:tk,e)=0.It can be used for
The data segment value of System Discrimination is I (t)=1, at other times point I (t)=0.
It should be noted that r (t) has higher priority here, i.e. r (t) is constant, then entirely in the examples of implementation
Data segment cannot use, and not need to consider further that y (t).
It is application of the method for the invention in specific example below.
Here an industrial application case is provided, y (t) is the water level of boiler, and r (t) is desired water level value, and u (t) is defeated
Controller out, d (t) are the main steam flows of boiler.The variation range of y (t) is -300mm to 300mm, and height warning level is arranged
Value is -100mm, 100mm, and u (t) is the inflow that boiler is flowed by two water pump controls.Fig. 3 (a) to Fig. 3 (c) gives r
(t), the tendency chart of y (t), u (t) and d (t) in 24 hours.
This method has found the data segment of three variations, [2990,9721], [16549,24553], [72296,77886].
Take the time trend of the y (t) of t ∈ [1,12000], in Fig. 4, y (t) takes 0 to 12000 solid lines, IyIt (t) is dotted line, vertical point
Setting-out is the position of change point.Y (t) and u (t) is set forth in the concrete outcome of first data segment in Fig. 5 (a)-Fig. 5 (b).
Step 1, find the position of change point in data segment: Fig. 4 is the time trend of the y (t) of t ∈ [1,12000] and leads to
It crosses above-mentioned steps one and calculates the position (vertical dotted line) for obtaining change point, the position of first change point is t=7404, p
=5.5887-181。DT=0.2688, threshold value takes 0.2313.
Step 2, Fig. 6 (a) compare in formula 9WithY (t) ∈ [7405,8493], DTIt is maximized
0.2688, threshold value 0.2313, indicator sequence Iy(t) 1 is taken in y (t) ∈ [7405,8493], comparison diagram 6 (b) is y (t) ∈
Other data of [8494,9721] interior y (t), DT=0.0880, it is less than threshold value 0.0919.Therefore, the I in this data segmenty
(t)=0.
Step 3, such as the I in Fig. 4 in step 2y(t) it is represented by dashed line, the I in step 3y(t) it is indicated with point, it is high respectively
In be lower than Iy(t) level.Data segment y (t) ∈ [2912,4782] is Amplitude Ratio former point time big data segment.So
Another data segment is [7405,8493] y (t) ∈.Therefore, the I in step 3y(t) than the I in step 2y(t) have and more widen
Section.
Step 4, Fig. 7 (a)-Fig. 7 (c) give two indicator sequence Ir(t) and Iy(t).It is obvious that Ir(t) it is centered around r
(t) value 1 near the variation of slope occurs.However, due to variable d (t), Iy(t) take 1 range ratio r (t) big.Therefore, having must
By Ir(t) and Iy(t) it carries out inclusive-OR operation and obtains indicator sequence I (t) to the end.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (9)
1. a kind of historical data section automatic searching method suitable for System Discrimination, characterized in that the following steps are included:
The data sample for obtaining the time signal composition of series of discrete, using the probability distribution change detecting method of imparametrization
The different data section that probability distribution has variation is found in the position for obtaining change point in data sample；
Judge whether each data segment passes through an apparent amplitude variation or for a steady state value, it is false using the latter as zero
If the former, it is assumed that form hypothesis testing, judges whether data segment can be used as System Discrimination according to hypothesis testing as substitution
Data segment；
Based on the assumption that indicator sequence is all to verify in the case where zero between two adjacent data sections with the presence or absence of significance difference in examining
It is different；
The setting value of comprehensive closed-loop control system obtains the data segment that can be used for System Discrimination；
The probability distribution change detecting method using imparametrization obtains the position of change point in data sample, specifically:
(1) current data segment x (t is initializedk:tk+ K-1)={ x (tk),x(tk+1),...,x(tk+ K-1) }, it includes entire
Data sample x (1), x (2) ..., x (N)；
(2) a current data section x (t is calculatedk:tk+ K-1) interior x (t) a relative position, be denoted as:
Wherein, t=tk,tk+1,…,tk+ K-1,
Provide statistic U (t)=U (t-1)+V (t), initial value U (a tk- 1)=0, it is specified that whether a hypothesis testing is to become
Change the maximum value that position is time point U (t) absolute value, be denoted as:The P value of hypothesis testing is denoted as:If the value of P is less than the probability value of Error type I α, tmaxIt is exactly the position of change point
It sets, otherwise, tmaxIt is not just the position of change point；
(3) current data section x (tk:tk+ K-1) divide x (tk:tmax)={ x (tk,x(tk+1),...,x(tmax) and x (tmax+1:tk
+ K-1)={ x (tmax+1),x(tmax+2),...,x(tk+ K-1) } two data segments, tmaxIt is the position of variation；
(4) position of the step of the repeating (2) and (3) to each data segment until can not find variation, the length of data segment are not less than
The parameter K of user's selection0, minimum length as data segment.
2. a kind of historical data section automatic searching method suitable for System Discrimination as described in claim 1, characterized in that if
One data segment, length is that K is greater than 2K0, then the data segment should be divided into K/K0Section, by operation, can obtain all changes
Change the position t of pointi' s (i=1,2 ..., I), wherein I is the number of change point.
3. a kind of historical data section automatic searching method suitable for System Discrimination as described in claim 1, characterized in that if
One data segment x (ti:ti+1- 1) range is greater than an at user option parameter d0, d0Change for sample significance the smallest
Value, null hypothesis are rejected, that is to say, that the previous case is received, then x (ti:ti+1- 1) it can be used as System Discrimination data segment.
4. a kind of historical data section automatic searching method suitable for System Discrimination as described in claim 1, characterized in that one
A data segment x (ti:ti+1- 1) with index series I when whether as System Discrimination data segmentx(t) indicate that can data segment be used to be
System identification, if data segment can be used to System Discrimination, Ix(ti:ti+1- 1)=1；Otherwise Ix(ti:ti+1- 1)=0.
5. a kind of historical data section automatic searching method suitable for System Discrimination as described in claim 1, characterized in that verifying
When whether there is significant difference between two adjacent data sections, the average value of two adjacent data sections is compared with hypothesis testing.
6. a kind of historical data section automatic searching method suitable for System Discrimination as claimed in claim 5, characterized in that with vacation
If when the average value examined to compare two adjacent data sections, being distributed for standard gaussianGreater than a determining probability γ of Error type I type,And NiIt is respectively
The length of sample average, standard deviation and data segment, then, their mean difference is greater than minimum conspicuousness changing value d0If
Second data segment has one relative to the apparent amplitude variation of first data segment, then the index I of the second segment dataxTake
Value becomes 1 from 0.
7. a kind of historical data section automatic searching method suitable for System Discrimination as described in claim 1, characterized in that comprehensive
When the setting value acquisition of closed-loop control system can be used for the data segment of System Discrimination, an indicator sequence I (t) is defined, I is worked asr(t)
=1 or Iy(t)=1 when, I (t)=1, it may be assumed that I (t)=Ir(t)∪Iy(t), t is usedk,sAnd tk,eK-th of data segment is respectively indicated to open
The time begun and terminated, I (t)=1 is continuous, it may be assumed thatI(tk,s- 1)=0, I (tk,e+ 1)=0.
8. a kind of historical data section automatic searching method suitable for System Discrimination as claimed in claim 4, characterized in that if
Historical data section r (t) does not change in k-th of data segment, then this data segment cannot be used to System Discrimination, it is corresponding at this time
Index series becomes: Ix(tk,s:tk,e)=0, wherein using tk,sAnd tk,eRespectively indicate k-th of data segment beginning and end when
Between, the data segment value for System Discrimination is Ix(t)=1, point I at other timesx(t)=0.
9. a kind of historical data section automatic searching method suitable for System Discrimination a method as claimed in any one of claims 1-8, feature
It is to be suitable for the system of multiple input single output suitable for the historical data section automatic searching method of System Discrimination.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810458154.3A CN108549346B (en) | 2018-05-14 | 2018-05-14 | A kind of historical data section automatic searching method suitable for System Discrimination |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810458154.3A CN108549346B (en) | 2018-05-14 | 2018-05-14 | A kind of historical data section automatic searching method suitable for System Discrimination |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108549346A CN108549346A (en) | 2018-09-18 |
| CN108549346B true CN108549346B (en) | 2019-04-02 |
Family
ID=63494846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810458154.3A Active CN108549346B (en) | 2018-05-14 | 2018-05-14 | A kind of historical data section automatic searching method suitable for System Discrimination |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108549346B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635431B (en) * | 2018-12-12 | 2019-10-29 | 山东科技大学 | Dynamical system static gain estimation method based on historical data ramp response |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2627844T3 (en) * | 2002-11-04 | 2017-07-31 | Ge Intelligent Platforms, Inc. | Monitoring the status of a system using a recurring local learning machine |
| CN102332011B (en) * | 2011-09-09 | 2012-12-26 | 北京空间飞行器总体设计部 | Method for selecting effective data of in-orbit spacecraft |
| CN104573850B (en) * | 2014-12-18 | 2017-08-11 | 湖南大唐先一科技有限公司 | A kind of Power Plant Equipment state evaluating method |
| CN105259756B (en) * | 2015-10-20 | 2018-05-04 | 广东电网有限责任公司电力科学研究院 | Power plant's control loop identification Method and system |
| CN106249728B (en) * | 2016-09-28 | 2019-02-22 | 清华大学 | A kind of thermal power generation unit on-line performance monitoring method based on characteristics of components |
| CN106683001B (en) * | 2016-12-06 | 2020-12-15 | 华北电力大学 | Selection method of thermal power generating unit identification data based on historical operating data |
| CN106778053B (en) * | 2017-03-31 | 2019-04-09 | 国网山东省电力公司电力科学研究院 | A kind of alert correlation variable detection method and system based on correlation |
-
2018
- 2018-05-14 CN CN201810458154.3A patent/CN108549346B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108549346A (en) | 2018-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106368813B (en) | A kind of abnormal alarm data detection method based on multivariate time series | |
| Li et al. | An improved fault detection method for incipient centrifugal chiller faults using the PCA-R-SVDD algorithm | |
| Shang et al. | A note on the extended Rosenbrock function | |
| US9146800B2 (en) | Method for detecting anomalies in a time series data with trajectory and stochastic components | |
| CN108335021A (en) | A kind of method and maintenance decision optimization of wind energy conversion system state Reliability assessment | |
| CN105205288A (en) | Mode evolution-based forecasting method for satellite long-term on-orbit running state | |
| CN104750973B (en) | Unit load (standard) steady state condition clustering algorithm based on data smoothness function | |
| CN109298225B (en) | Automatic identification model system and method for abnormal state of voltage measurement data | |
| CN108549346B (en) | A kind of historical data section automatic searching method suitable for System Discrimination | |
| CN106778053A (en) | A kind of alert correlation based on correlation becomes quantity measuring method and system | |
| CN108022058A (en) | A kind of wind energy conversion system state reliability estimation method | |
| CN110648480A (en) | Single variable alarm system and method based on change rate | |
| CN103488561A (en) | kNN (k-nearest neighbor) fault detection method for online upgrading master sample model | |
| CN111797889A (en) | Single variable alarm system performance evaluation method and system based on piecewise linear representation | |
| CN103246279A (en) | CPA method in chemical process with viscosity of actuated valve | |
| CN107205244A (en) | A kind of design method of the sensor network anomaly data detection based on temporal correlation | |
| CN113341087A (en) | Intelligent monitoring and analyzing method and system for river water pollution source and computer storage medium | |
| CN110298765B (en) | Power distribution network power consumption abnormality detection method based on objective correlation factors | |
| Wang et al. | Concurrent fault detection and anomaly location in closed-loop dynamic systems with measured disturbances | |
| CN108519760A (en) | A kind of Primary Processing stable state recognition methods based on detection of change-point theory | |
| CN108508860B (en) | Process industrial production system data monitoring method based on coupling relation | |
| Lucke et al. | Variable selection for fault detection and identification based on mutual information of alarm series | |
| Ji | Data-driven sensor fault diagnosis under closed-loop control with slow feature analysis | |
| CN108959739A (en) | A kind of analysis method and device of the pressure fluctuation of hydroenergy storage station transient process | |
| CN110298767B (en) | Thermal power plant time sequence variable abnormity monitoring method and system |
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 |