CN110414726A - A kind of power quality method for early warning based on Analysis on monitoring data - Google Patents
A kind of power quality method for early warning based on Analysis on monitoring data Download PDFInfo
- Publication number
- CN110414726A CN110414726A CN201910639454.6A CN201910639454A CN110414726A CN 110414726 A CN110414726 A CN 110414726A CN 201910639454 A CN201910639454 A CN 201910639454A CN 110414726 A CN110414726 A CN 110414726A
- Authority
- CN
- China
- Prior art keywords
- trend
- mode
- sequence
- power quality
- early warning
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000004458 analytical method Methods 0.000 title claims abstract description 16
- 238000012544 monitoring process Methods 0.000 title claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 21
- 230000006835 compression Effects 0.000 claims description 19
- 238000007906 compression Methods 0.000 claims description 19
- 239000008186 active pharmaceutical agent Substances 0.000 claims description 6
- 230000007423 decrease Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of power quality method for early warning based on Analysis on monitoring data, the following steps are included: step S1: the reading and pretreatment of power quality data use the Piecewise Linear Representation of Time Series method for extracting marginal point based on slope to carry out piecewise linearity processing the data read；Step S2: mode indicates, establishes trend sequence, and the trend sequence established is model, calculates the pattern distance between every two sequence；Step S3: calculating the assemble mode distance of each power quality time series Yu other times sequence, chooses assemble mode apart from the smallest trend sequence as normal model；Step S4: the pattern distance and percent similarity between other each trend sequences and the sequence are calculated；Step S5: the selected percent similarity critical value for needing to carry out early warning carries out early warning to the underproof time series of percent similarity.The present invention writes program and realizes algorithm, and realization efficiently analyzes power quality data, to judge whether power quality is qualified.
Description
Technical field
The present invention relates to a kind of power quality method for early warning based on Analysis on monitoring data, belongs to field of power system.
Background technique
With the development of science and technology, requirement of the power equipment for power quality is higher and higher, once power quality can not
It meets the requirements, will lead to equipment operation irregularity, or even damage equipment, cause economic loss.In view of the requirement to power quality
Higher and higher, the accuracy requirement now for data analysis is also higher and higher, the power quality based on Analysis on monitoring data
This project of early warning is efficiently handled data with algorithm, is write program and is realized algorithm, is expected to be able to use more simple
Clean method goes accurately to carry out electricity quality evaluation, to the stable operation of electric system and all kinds of more demanding to power quality
Instrument is of great significance.
The country pays more attention to the research of power quality at present, and relevant research achievement is more, but it is most of all
It is to be analyzed for data itself, to judge whether the power quality at a certain moment has reached relevant standard, and passes through
Time series is directly researched to judge that the research of power quality is less, but relevant mathematical method have been relatively mature, only
Fewer to use this field of power quality, can find suitable mathematical algorithm and apply to power quality field will make electricity
Whether qualified the judgement of energy quality be more simple, improves the operation efficiency of computer.
Therefore, on Practical Project, there is an urgent need to it is a kind of using program realize algorithm to power quality data carry out processing with
Analysis, to realize the early warning to power quality.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of power quality method for early warning based on Analysis on monitoring data,
It can be used for solving the problems, such as that current power quality data processing is more inefficient.
To achieve the above object, the present invention adopts the following technical scheme: a kind of electric energy matter suitable for Analysis on monitoring data
Measure method for early warning, comprising the following steps:
Step S1: the reading and pretreatment of power quality data use the data read and extract edge based on slope
The Piecewise Linear Representation of Time Series method of point carries out piecewise linearity processing；
Step S2: mode indicates, establishes trend sequence, and the trend sequence established is model, calculates every two sequence
Between pattern distance；
Step S3: calculating the assemble mode distance of each power quality time series Yu other times sequence, chooses assemble mode
Apart from the smallest trend sequence as normal model；
Step S4: the pattern distance and percent similarity between other each trend sequences and the sequence are calculated；
Step S5: the selected percent similarity critical value for needing to carry out early warning, to the underproof time sequence of percent similarity
Column carry out early warning.
The technical solution that the present invention further limits is, the step S1 comprising the following steps:
Step S11: the time series data of related power quality index is read, is deposited into structural array, form is
(vi, ti), and input the initial value and step value of desired compression ratio and iteration；
Step S12: piecewise linearity processing judges the slope at each both ends, and the difference of calculating between the two,
It is compared with critical slope variation rate, judges whether it will be deposited into the amplitude of waypoint and time as waypoint
In another structural array；Critical slope variation rate is a definite value, is closed by the slope at certain point both ends and the size of the value
To determine whether regarding the certain point in time series as waypoint, its value is defining its initial value and journey after stepping for system
Sequence is selected automatically according to the compression ratio of selection.
Step S13: piecewise linearity calculates the length of array, the i.e. length of time series after piecewise linearity after completing
Degree, then calculates compression ratio, judge compression ratio whether meet demand, that is, the compression ratio being calculated is more than or equal to default
Compression ratio value, the value is selected according to the amount of data, and the data volume the big eligible bigger, is typically chosen the compression lower than 75%
Rate error can be smaller.Following step S2 is carried out if meeting the requirements, if being unsatisfactory for increasing critical slope variation rate, because
For the smaller of initial value setting, it is however generally that be that compression ratio is smaller, therefore the step value to critical slope variation rate plus setting is
Can, repetition step S12 both may be used after increasing slope variation rate；Wherein compression ratio are as follows:
The length of its Central Plains time series is n, and the length of time series after piecewise linearity is n '.
Waypoint it is selected it is specific there are two types of situations, meeting in following two condition condition can be used as being segmented
Point:
(1) slope at left and right sides of the point is positive and negative identical, and the absolute value of the two slope difference is greater than the critical oblique of setting
The value of rate change rate.
(2) slope at left and right sides of the point is positive and negative on the contrary, wherein the slope of certain side is greater than the critical slope variation of setting
The value of rate.
Further, step S2 specifically:
Step S21: mode expression is carried out to the data after the resulting piecewise linearity of step S1, obtains trend sequence；
Step S22: the pattern distance between gained trend sequence every two is calculated；
Mode indicates and trend sequence: mode expression is exactly to be partitioned into time series one by one according to variation tendency
Subsequence, each subsequence represent one mode, and essence is exactly that figure line is converted to a set of the trend of containing only.Gained
To sequence be referred to as trend sequence；
The representation method of trend sequence is as follows: will rise the mode 1 that is used as, decline is used as mode 1, keeps being used as mode 0, then
Any one time series can be expressed as: S={ (m1, t1), (m2, t2), (m3, t3) ..., (mn, tn), the table of the sequence
Show 0-t1Mode in period is m1, t1-t2Mode in period is m2, wherein miThe value of (1≤i≤n) be 0, -1 or
Person 1, ti(1≤i≤n) is miThe end time of this mode and mi+1The initial time of mode；
Pattern distance: pattern distance is the distance between identical mode of two time spans, it is assumed that some mode is
si=(mi, ti), another mode is sj=(mj, tj), ti=tjAnd ti-ti-1=tj-tj-1, then siWith sjBetween pattern distance
D(si, sj)=| mi-mj|, the value range of the pattern distance is { 0,1,2 }, the difference between distance two modes of bigger expression
It is bigger, indicate mode between the two completely on the contrary, only in m when distance is 2i=mjWhen pattern distance be just 0.Calculate two
Isotype processing must be first done when pattern distance between time series；
Isotype processing is for example to SA={ (1, t11), (- 1, t12), (1, t3), SB={ (1, t21), (- 1, t22), (1,
t3) (wherein (0 < t11< t21< t12< t22< t3)) two sequences carry out isotype processing, after carrying out isotype processing, two
A sequence becomes S ' respectivelyA={ (1, t11), (- 1, t21), (- 1, t12), (1, t22), (1, t3), S 'B={ (1, t11), (1,
t21), (- 1, t12), (- 1, t22), (1, t3), when becoming two after isotype processing has trend initial time and terminates
Between identical trend sequence；Pattern distance between two time serieses is defined as:
Wherein tnFor total time length, ti-1And tiThe time is originated and terminated for mode, and (isotype handles latter two sequence
It is completely the same), m1iAnd m2iFor mode, (value is 1,0 and 1).DSValue range be [0,2]；DSValue two times of bigger explanation
Similarity degree between sequence is lower, and the similarity degree between two time serieses of smaller explanation is higher.
Further, determine assemble mode apart from the smallest trend sequence as normal model, by other trend sequences with
Matching, calculate pattern distance, to obtain the trend percent similarity between other trend sequences and normal model；Trend phase
Like percentage specifically:
Trend phase recency between the value and two sequences is directly proportional, the bigger trend more phase proved between two trend sequences
Closely.
Further, the carry out early warning of threshold value is lower than to trend percent similarity.
Compared with the prior art, the invention has the following beneficial effects: the present invention can be used for efficient process power quality number
According to, power quality early warning is carried out, it is examined by example, realization efficiently analyzes power quality data, from
And judge whether power quality is qualified；Its result is objective credible, more engineering practicability, have stronger application value.
Detailed description of the invention
Fig. 1 is the method program flow chart in the embodiment of the present invention 1.
Fig. 2 is the comparison diagram of first day in the embodiment of the present invention 1 linear front and back of time series segmentation.
Fig. 3 is the comparison diagram of selected data and normal data in the embodiment of the present invention 1.
Fig. 4 is the comparison diagram of selected data and abnormal data in the embodiment of the present invention 1.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
Fig. 1-Fig. 4 is please referred to, the present invention provides a kind of power quality method for early warning suitable for Analysis on monitoring data, including
Following steps:
Step S1: the reading and pretreatment of power quality data use the data read and extract edge based on slope
The Piecewise Linear Representation of Time Series method of point carries out piecewise linearity processing；
Step S2: mode indicates, establishes trend sequence, and the trend sequence established is model, calculates every two sequence
Between pattern distance；
Step S3: calculating the assemble mode distance of each power quality time series Yu other times sequence, chooses assemble mode
Apart from the smallest trend sequence as normal model；
Step S4: the pattern distance and percent similarity between other each trend sequences and the sequence are calculated；
Step S5: the selected percent similarity critical value for needing to carry out early warning, to the underproof time sequence of percent similarity
Column carry out early warning.
In the present embodiment, preferably, the step S1 is specifically included:
Step S11: having chosen 13 day datas of the voltage total harmonic distortion factor in the monitoring data on certain 10kv bus,
The time series data for reading related power quality index, is deposited into structural array, and form is (vi, ti), and input and think
The initial value and step value of the compression ratio and iteration wanted, since existing data volume is smaller, by discovery choosing is run multiple times
Level pressure shrinkage can eliminate minor swing and retain all features substantially for 20%, therefore the compression ratio that the present embodiment is chosen is 20%,
Step value and initial value selection should be smaller, and the initial value and step value of the embodiment are 0.01.
Step S12: piecewise linearity processing judges the slope at each both ends, and the difference of calculating between the two,
It is compared with critical slope variation rate, judges whether it will be deposited into the amplitude of waypoint and time as waypoint
In another structural array.
Step S13: piecewise linearity calculates the length of array, the i.e. length of time series after piecewise linearity after completing
Degree, then calculate compression ratio, judge compression ratio whether meet demand, following step S2 is carried out if meeting, if not being inconsistent
Conjunction then increases critical slope variation rate, because initial value setting is smaller, it is however generally that is that compression ratio is smaller, therefore gives critical slope
Step value of the change rate plus setting.Increase slope variation rate after repeat step S12 both can, for first day, line
Property segmentation front and back comparison diagram such as Fig. 2.Wherein compression ratio are as follows:
The length of its Central Plains time series is n, and the length of time series after piecewise linearity is n '.
The selected of waypoint specifically includes two kinds of situations:
(1) slope at left and right sides of the point is positive and negative identical, and the absolute value of the two slope difference is greater than the critical oblique of setting
The value of rate change rate.
(2) slope at left and right sides of the point is positive and negative on the contrary, wherein the slope of certain side is greater than the critical slope variation of setting
The value of rate.
In the present embodiment, step S2 tool includes:
Step S21: carrying out mode expression to the data after the resulting piecewise linearity of step S1, obtain trend sequence, wherein
It is changed into trend sequence after first day time series segmentation is linear to obtain later:
SA=(- 1,8), (1,9), (0,10), (- 1,11), (- 1,12), (1,15), (0,19), (0,20), (- 1,
40), (- Isosorbide-5-Nitrae 1), (0,42), (Isosorbide-5-Nitrae 3), (- 1,51), (1,56), (- 1,57), (- 1,60), (1,71), (- 1,72), (1
77), (0,84), (- 1,85), (1,86), (- 1,94), (1,100), (- 1,101), (0,104), (- 1,105), (0,109)
(1,110), (0,111), (- 1,119), (1,120), (0,121), (- 1,122), (- 1,126), (1,131), (- 1,132),
(0,137), (1,138), (0,139), (- 1,140), (1,141), (- 1,151), (- 1,152), (0,154), (- 1,155),
(- 1,157), (0,158), (1,159), (- 1,161), (- 1,165), (0,168), (- 1,178), (- 1,179), (- 1,
201), (1,202), (- 1,203), (- 1,204), (- 1,215), (1,216), (- 1,222), (1,223), (1,226), (1,
227), (0,229), (1,231), (1,234), (1,236), (1,237), (1,238), (1,240), (1,241), (0,242)
(- 1,243), (0,249), (- 1,255), (- 1,261), (- 1,262), (- 1,264), (0,266), (- 1,267), (- 1,
269), (- 1,271), (1,272), (- 1,276), (1,287), (- 1,288), (- 1,293), (- 1,296), (1,297), (0,
299), (1,300), (0,305), (- 1,314), (1,317), (- 1,319), (- 1,322), (1,323), (- 1,326), (1,
336), (1,337), (- 1,339), (1,341), (1,342), (- 1,344), (1,347), (0,348), (1,350), (0,
354), (1,355), (0,358), (- 1,359), (1,375), (- 1,376), (1,377), (0,379), (1,380), (- 1,
386), (1,388), (- 1,391), (1,392), (0,393), (- 1,394), (1,395), (Isosorbide-5-Nitrae 02), (- Isosorbide-5-Nitrae 04), (- 1,
409), (Isosorbide-5-Nitrae 10), (- Isosorbide-5-Nitrae 15), (Isosorbide-5-Nitrae 26), (- Isosorbide-5-Nitrae 27), (0,428), (Isosorbide-5-Nitrae 29), (- Isosorbide-5-Nitrae 30), (0,432), (1,
434), (- Isosorbide-5-Nitrae 35), (0,437), (Isosorbide-5-Nitrae 38), (- Isosorbide-5-Nitrae 39), (Isosorbide-5-Nitrae 40), (0,441), (- Isosorbide-5-Nitrae 42), (Isosorbide-5-Nitrae 44), (1,
449), (- Isosorbide-5-Nitrae 50), (0,451), (Isosorbide-5-Nitrae 52), (- Isosorbide-5-Nitrae 55), (Isosorbide-5-Nitrae 56), (- Isosorbide-5-Nitrae 59), (Isosorbide-5-Nitrae 60), (0,463), (1,
465), (Isosorbide-5-Nitrae 66), (Isosorbide-5-Nitrae 70), (Isosorbide-5-Nitrae 71), (Isosorbide-5-Nitrae 72), (- Isosorbide-5-Nitrae 74), (Isosorbide-5-Nitrae 75), (- Isosorbide-5-Nitrae 76), (- Isosorbide-5-Nitrae 79) }.
Step S22: the pattern distance between gained trend sequence every two is calculated.
In the present embodiment, step S3 tool includes:
Step S31: the assemble mode distance between each trend sequence and other sequences is calculated, by assemble mode apart from the smallest
As normal model, the pattern distance summation that this 13 day every day and other days are obtained after program operation is respectively as follows: 7.91,
7.74、7.56、7.70、7.75、8.02、7.65、7.86、8.45、7.65、8.25、8.38、8.42。
Step S32: the trend sequence with the assemble mode of other trend sequences apart from the smallest third day is chosen as normal
Model.Mode indicates and trend sequence: mode expression is exactly that time series is partitioned into sub- sequence one by one according to variation tendency
Column, each subsequence represent one mode, and essence is exactly that figure line is converted to a set of the trend of containing only, obtained
Sequence is referred to as trend sequence.
The representation method of trend sequence is as follows: will rise the mode 1 that is used as, decline is used as mode 1, keeps being used as mode 0, then
Any one time series can be expressed as: S={ (m1, t1), (m2, t2), (m3, t3) ..., (mn, tn), the table of the sequence
Show 0-t1Mode in period is m1, t1-t2Mode in period is m2, wherein miThe value of (1≤i≤n) be 0, -1 or
Person 1, ti(1≤i≤n) is miThe end time of this mode and mi+1The initial time of mode.
Pattern distance: pattern distance is the distance between identical mode of two time spans, it is assumed that some mode is
si=(mi, ti), another mode is sj=(mj, tj), ti=tjAnd ti-ti-1=tj-tj-1, then siWith sjBetween pattern distance
D(si, sj)=| mi-mj|, the value range of the pattern distance is { 0,1,2 }, the difference between distance two modes of bigger expression
It is bigger, indicate mode between the two completely on the contrary, only in m when distance is 2i=mjWhen pattern distance be just 0.Calculate two
Isotype processing must be first done when pattern distance between time series.
Isotype processing is for example to SA={ (1, t11), (- 1, t12), (1, t3), SB={ (1, t21), (- 1, t22), (1,
t3) (wherein (0 < t11< t21< t12< t22< t3)) two sequences carry out isotype processing, after carrying out isotype processing, two
A sequence becomes S ' respectivelyA={ (1, t11), (- 1, t21), (- 1, t12), (1, t22), (1, t3), S 'B={ (1, t11), (1,
t21), (- 1, t12), (- 1, t22), (1, t3), when becoming two after isotype processing has trend initial time and terminates
Between identical trend sequence.
Pattern distance between two time serieses is defined as:
Wherein tnFor total time length, ti-1And tiIt is originated for mode and terminates the time, isotype handles latter two sequence
It is completely the same, m1iAnd m2iFor mode, being worth is -1,0 and 1.DSValue range be [0,2]；DSValue two times of bigger explanation
Similarity degree between sequence is lower, and the similarity degree between two time serieses of smaller explanation is higher.
In the present embodiment, step S4 tool includes:
Determine trend sequence of the assemble mode apart from the smallest third day as normal model, therewith by other trend sequences
Matching calculates pattern distance, to obtain the trend percent similarity between other trend sequences and normal model, obtained knot
Fruit are as follows: 68.95%, 70.10%, 69.15%, 70.40%, 69.50%, 70.10%, 68.85%, 66.35%, 70.40%,
65.30%, 65.75%, 66.85%.
Trend percent similarity specifically:
Trend phase recency between the value and two sequences is directly proportional, the bigger trend more phase proved between two trend sequences
Closely.
In the present embodiment, step S5 is specifically included:
The selected percent similarity critical value for needing to carry out early warning, in this embodiment after the verifying of multi-group data most
66.6% critical value as early warning at last, what Fig. 3 chose is and third day trend percent similarity the maximum ten day and the
Comparison in three days, Fig. 4 selection are and the smallest the 11st day for needing to carry out power quality early warning of third day percent similarity
It is compared with third day, from image, we can judge to compared with the 11st day, and the tenth day trend is obviously and third
It trend is more close, which demonstrates that this method is really feasible, can be by the close degree numerical value between trend
It shows to carry out early warning to power quality abnormal data, and calculating process is simple, accuracy is higher.Then to trend hundred
Divide the carry out early warning than being lower than threshold value.
The foregoing is merely presently preferred embodiments of the present invention, all equivalent changes done according to scope of the present invention patent with
Modification, is all covered by the present invention.
Claims (7)
1. a kind of power quality method for early warning based on Analysis on monitoring data, which comprises the following steps:
Step S1: the reading and pretreatment of power quality data use the data read and extract marginal point based on slope
Piecewise Linear Representation of Time Series method carries out piecewise linearity processing；
Step S2: mode indicates, establishes trend sequence, and the trend sequence established is model, calculates between every two sequence
Pattern distance；
Step S3: calculating the assemble mode distance of each power quality time series Yu other times sequence, chooses assemble mode distance
The smallest trend sequence is as normal model；
Step S4: the pattern distance and trend percent similarity between other each trend sequences and the sequence are calculated；
Step S5: the selected percent similarity critical value for needing to carry out early warning, to the underproof time series of percent similarity into
Row early warning.
2. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 1, it is characterised in that: institute
State step S1 specifically:
Step S11: reading the time series data of related power quality index, be deposited into structural array, and form is (vi,
ti), and input the initial value and step value of desired compression ratio and iteration；
Step S12: piecewise linearity processing judges the slope at each both ends, and calculates difference between the two, and faces
Boundary's slope variation rate compares, and judges in addition whether it will be deposited into the amplitude of waypoint and time as waypoint
In one structural array；
Step S13: piecewise linearity calculates the length of array, the i.e. length of time series after piecewise linearity after completing, with
After calculate compression ratio, judge compression ratio whether meet demand, following step S2 is carried out if meeting the requirements, if being unsatisfactory for
It is required that then increasing critical slope variation rate, repetition step S12 both may be used after increasing critical slope variation rate；Its compression ratio formula
Are as follows:
The length of its Central Plains time series is n, and the length of time series after piecewise linearity is n '.
3. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 2, it is characterised in that: institute
Stating the selected of waypoint includes two kinds of situations: the slope at left and right sides of the point is positive and negative identical, and the two slope difference is absolute
Value is greater than the value of the critical slope variation rate of setting；Or slope at left and right sides of the point it is positive and negative on the contrary, wherein certain side it is oblique
Rate is greater than the value of the critical slope variation rate of setting.
4. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 2, it is characterised in that: institute
State step S2 specifically:
Step S21: mode expression is carried out to the data after the resulting piecewise linearity of step S1, obtains trend sequence；
Step S22: the pattern distance between gained trend sequence every two is calculated.
5. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 4, it is characterised in that: institute
State mode expression, trend sequence and pattern distance specifically:
Mode indicates and trend sequence: by time series according to variation tendency, being partitioned into subsequence one by one, each subsequence
One mode is represented, essence is exactly that figure line is converted to a set of the trend of containing only；Obtained sequence is referred to as
Gesture sequence；
The representation method of trend sequence is as follows: will rise the mode 1 that is used as, decline is used as mode 1, keeps being used as mode 0, then can be with
Any one time series is expressed as: S={ (m1, t1), (m2, t2), (m3, t3) ..., (mn, tn), the expression 0- of the sequence
t1Mode in period is m1, t1-t2Mode in period is m2, wherein miThe value of (1≤i≤n) is 0,1 or 1, ti
(1≤i≤n) is miThe end time of this mode and mi+1The initial time of mode；
Pattern distance: pattern distance is the distance between identical mode of two time spans, it is assumed that some mode is si=
(mi, ti), another mode is sj=(mj, tj), ti=tjAnd ti-ti-1=tj-tj-1, then siWith sjBetween pattern distance D
(si, sj)=| mi-mj|, the value range of the pattern distance is { 0,1,2 }, the difference between distance two modes of bigger expression
It is bigger, indicate mode between the two completely on the contrary, only in m when distance is 2i=mjWhen pattern distance be just 0.
6. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 5, it is characterised in that: institute
State pattern distance between isotype processing and time series specifically:
Such as to SA={ (1, t11), (- 1, t12), (1, t3), SB={ (1, t21), (- 1, t22), (1, t3) (wherein (0 < t11
< t21< t12< t22< t3)) two sequences carry out isotype and handle two sequences becoming S ' respectivelyA={ (1, t11), (one 1,
t21), (- 1, t12), (1, t22), (1, t3), S 'B={ (1, t11), (1, t21), (- 1, t12), (- 1, t22), (1, t3), wait moulds
Becoming two after formula processing has trend initial time trend sequence identical with the time is terminated；
Pattern distance between two time serieses is defined as:
Wherein tnFor total time length, ti-1And tiTime, m are originated and terminated for mode1iAnd m2iFor mode；DSValue range be
[0,2], DSValue it is bigger explanation two time serieses between similarity degree it is lower, between two time serieses of smaller explanation
Similarity degree is higher.
7. a kind of power quality method for early warning based on Analysis on monitoring data according to claim 6, it is characterised in that: institute
State trend percent similarity specifically:
The close degree of trend between the value and two sequences is directly proportional, and the close degree of trend i.e. two time serieses are in same time
The similarity degree of section variation tendency is bigger to prove that the trend between two trend sequences is more close.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910639454.6A CN110414726A (en) | 2019-07-15 | 2019-07-15 | A kind of power quality method for early warning based on Analysis on monitoring data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910639454.6A CN110414726A (en) | 2019-07-15 | 2019-07-15 | A kind of power quality method for early warning based on Analysis on monitoring data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110414726A true CN110414726A (en) | 2019-11-05 |
Family
ID=68361616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910639454.6A Pending CN110414726A (en) | 2019-07-15 | 2019-07-15 | A kind of power quality method for early warning based on Analysis on monitoring data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110414726A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116990465A (en) * | 2023-09-25 | 2023-11-03 | 北京金水永利科技有限公司 | Air quality data abnormity early warning method and system thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682573A (en) * | 2012-03-31 | 2012-09-19 | 上海海洋大学 | Time sequence analysis-based storm surge disaster early warning system |
| CN103235981A (en) * | 2013-04-10 | 2013-08-07 | 东南大学 | Wind power quality trend predicting method |
| CN103577562A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Multi-measurement time series similarity analysis method |
| CN108171425A (en) * | 2017-12-28 | 2018-06-15 | 国网冀北电力有限公司秦皇岛供电公司 | Power quality partition method, device and storage medium |
| CN108238066A (en) * | 2017-12-22 | 2018-07-03 | 同济大学 | Track switch operation curve template choosing method and its application |
| CN109359135A (en) * | 2018-09-04 | 2019-02-19 | 河海大学 | A kind of Algorithm for Similarity Search in Time Series based on segmentation weight |
-
2019
- 2019-07-15 CN CN201910639454.6A patent/CN110414726A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682573A (en) * | 2012-03-31 | 2012-09-19 | 上海海洋大学 | Time sequence analysis-based storm surge disaster early warning system |
| CN103235981A (en) * | 2013-04-10 | 2013-08-07 | 东南大学 | Wind power quality trend predicting method |
| CN103577562A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Multi-measurement time series similarity analysis method |
| CN108238066A (en) * | 2017-12-22 | 2018-07-03 | 同济大学 | Track switch operation curve template choosing method and its application |
| CN108171425A (en) * | 2017-12-28 | 2018-06-15 | 国网冀北电力有限公司秦皇岛供电公司 | Power quality partition method, device and storage medium |
| CN109359135A (en) * | 2018-09-04 | 2019-02-19 | 河海大学 | A kind of Algorithm for Similarity Search in Time Series based on segmentation weight |
Non-Patent Citations (2)
| Title |
|---|
| 林意等: "基于一阶滤波的时间序列分段线性表示方法", 《计算机工程》 * |
| 王达等: "时间序列的模式距离", 《浙江大学学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116990465A (en) * | 2023-09-25 | 2023-11-03 | 北京金水永利科技有限公司 | Air quality data abnormity early warning method and system thereof |
| CN116990465B (en) * | 2023-09-25 | 2023-12-19 | 北京金水永利科技有限公司 | Air quality data abnormity early warning method and system thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110212592B (en) | Thermal power generating unit load regulation maximum rate estimation method and system based on piecewise linear expression | |
| CN109921426A (en) | Wind-electricity integration system probability load flow calculation method based on CV-KDE | |
| CN110414726A (en) | A kind of power quality method for early warning based on Analysis on monitoring data | |
| CN112289392A (en) | Method for predicting performance of anode aluminum foil for aluminum electrolytic capacitor | |
| CN113537245B (en) | Neural network pruning method based on feature map | |
| CN108415880B (en) | Line loss characteristic analysis method based on sample entropy and wavelet transformation | |
| CN109459609B (en) | Distributed power supply frequency detection method based on artificial neural network | |
| CN111861121A (en) | Abnormal electricity utilization judgment method based on user electricity fluctuation and line loss fluctuation analysis | |
| CN113484818B (en) | Sliding window-based accurate positioning method for high-frequency acquisition anomaly-resistant electric energy meter | |
| CN116298670A (en) | Intelligent fault positioning method and system suitable for multi-branch distribution line | |
| CN101917007A (en) | Method for determining inductance value at alternating side of shunt active power filter | |
| CN114139964A (en) | Reliability assessment method and device for electrical comprehensive energy system | |
| CN110619147B (en) | Second-order and multi-order battery equivalent circuit model construction method applied to constant-voltage working condition | |
| CN110896218B (en) | Harmonic modeling method and system for establishing collective residential load | |
| CN114818827A (en) | Non-invasive load decomposition method based on seq2point network | |
| CN110095743B (en) | Distribution network terminal wave recording performance test module and waveform fitting method thereof | |
| CN108614147B (en) | Voltage fluctuation detection method and gateway meter thereof | |
| CN109524983B (en) | Photovoltaic output modeling method based on typical state | |
| CN112230056A (en) | Multi-harmonic source contribution calculation method based on OFMMK-Means clustering and composite quantile regression | |
| CN113128574A (en) | Scene reduction method and device and terminal equipment | |
| CN110580494A (en) | Data analysis method based on quantile logistic regression | |
| CN117706390B (en) | Rolling optimization estimation method and device for battery state of charge | |
| Chen et al. | Implementation of Constant Temperature–Constant Voltage Charging Method with Energy Loss Minimization for Lithium-Ion Batteries | |
| CN114137323B (en) | Metering multiplying power file accuracy judging method and system | |
| CN117034206A (en) | Weight calculation method and device for C-type voltage qualification rate influence factors of power distribution network |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191105 |