CN109034461B - Voltage sag random estimation method based on actual power grid monitoring information - Google Patents
Voltage sag random estimation method based on actual power grid monitoring information Download PDFInfo
- Publication number
- CN109034461B CN109034461B CN201810738766.8A CN201810738766A CN109034461B CN 109034461 B CN109034461 B CN 109034461B CN 201810738766 A CN201810738766 A CN 201810738766A CN 109034461 B CN109034461 B CN 109034461B
- Authority
- CN
- China
- Prior art keywords
- fault
- sag
- node
- line
- index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000012544 monitoring process Methods 0.000 title claims abstract description 28
- 238000012937 correction Methods 0.000 claims abstract description 78
- 238000005259 measurement Methods 0.000 claims abstract description 44
- 238000005070 sampling Methods 0.000 claims abstract description 31
- 238000004088 simulation Methods 0.000 claims abstract description 25
- 238000012545 processing Methods 0.000 claims abstract description 18
- 239000011159 matrix material Substances 0.000 claims abstract description 11
- 238000012806 monitoring device Methods 0.000 claims description 35
- 238000004364 calculation method Methods 0.000 claims description 30
- 238000010606 normalization Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 9
- 238000010220 Pearson correlation analysis Methods 0.000 claims description 8
- 230000001172 regenerating effect Effects 0.000 claims description 7
- 230000002596 correlated effect Effects 0.000 claims description 6
- 230000000875 corresponding effect Effects 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000007619 statistical method Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000013178 mathematical model Methods 0.000 claims description 5
- 230000016507 interphase Effects 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 150000001875 compounds Chemical class 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 239000006185 dispersion Substances 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 abstract description 8
- 238000000342 Monte Carlo simulation Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005315 distribution function Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001122767 Theaceae Species 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000919 ceramic Substances 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
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
-
- 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)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (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)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a random voltage sag estimation method based on actual power grid monitoring information, and belongs to the technical field of power quality analysis. The method comprises the following steps: reading actual power grid parameters to obtain a node admittance matrix; processing the actually measured sag data of each node to obtain a probability model of the fault type and the fault duration; establishing a system fault information model; utilizing Latin hypercube sampling to generate fault information and forming a fault information original database; carrying out fault simulation, calculating the error between the estimated result and the actual measurement result, and correcting the fault information model according to whether the error meets the requirement or not until the error meets the requirement or reaches the preset correction times, so as to obtain the final estimated result; the invention can effectively avoid the problems of poor stability, slow convergence, long time consumption, large error of the estimated result and the like in the existing estimation method, and the estimated result is more accurate.
Description
Technical Field
The invention belongs to the technical field of power quality analysis, and particularly relates to a voltage sag random estimation method based on actual power grid monitoring information.
Background
With the wide application of sensitive devices such as power electronics in power grids, voltage sag has become one of the most important power quality problems. Typical sensitive equipment, such as a computer, an ac governor, an ac contactor, etc., may cause problems such as equipment shutdown, unstable operation or error, efficiency reduction or life shortening after being subjected to voltage sag disturbance, and further cause the yield and quality of the industrial production line to be reduced or even completely interrupt the production process or service activities, thereby causing huge economic loss. Therefore, the method has important significance in predicting the occurrence condition of the voltage sag of the power grid, finding out weak links of the power grid, pertinently taking measures to inhibit the voltage sag and avoiding the adverse effect of the voltage sag on sensitive equipment as much as possible.
The current voltage sag estimation method mainly comprises an actual measurement statistical method and a simulation estimation method. Considering the randomness of the grid fault, the evaluation result of the actual measurement statistical method can be accurate enough only under the condition that the monitoring period is long enough. Therefore, in general, the voltage sag is estimated by using a simulation estimation method. The simulation estimation method is mainly classified into a fault point method, a critical distance method, and a monte carlo method. The Monte Carlo method is a common method for randomly estimating the voltage sag at present, and is characterized in that a power grid fault random model is firstly established, voltage sag information of a power grid is obtained by simulating a short-circuit fault, and when simulation reaches a certain number of times and a certain time limit, an evaluation result can accurately reflect the voltage sag condition of the power grid. However, the Monte Carlo method has the defects of poor stability, slow convergence and long time. In addition, with the wide application of the electric energy quality on-line monitoring system in regional power grids of China, monitoring devices are installed on partial nodes in an actual system, enough sag events can be obtained, therefore, a voltage sag random estimation method suitable for the actual power grid can be provided, the defects of poor stability, slow convergence and long time use of a Monte Carlo method can be overcome, estimation results can be made to be more practical, sag information is provided for power grid companies and users, weak links of the power grid are discovered, measures are taken in advance to relieve voltage sag, and economic loss is reduced.
Disclosure of Invention
In order to solve the problems, the invention provides a random voltage sag estimation method based on actual power grid monitoring information, which is characterized by comprising the following steps of:
A. reading actual power grid parameters, and processing to obtain a node admittance matrix;
B. acquiring actually measured sag data of each node of an actual power grid monitoring system, and processing the actually measured sag data to obtain a fault information model based on the actually measured data, wherein the fault information model comprises a fault type information model PtypeAnd a fault duration information model Pdur；
C. Establishing a system fault information model including a fault line information model PlineFault location information model PspotAnd a fault impedance information model Pres；
D. Adopting Latin hypercube sampling to generate fault information, and establishing a fault information original database which comprises a fault line original database, a fault position original database, a fault type original database, a duration original database and a fault impedance original database;
E. carrying out fault simulation to obtain sag characteristic values of each sag event, wherein the sag characteristic values comprise sag amplitude values, duration times and sag types; calculating sag indexes of all nodes, comparing estimated sag indexes of nodes with installed monitoring devices with actually measured sag indexes of a power grid, and calculating to obtain errors of estimated results and actually measured results; and E, judging whether the error is less than 20%, if the error does not meet the requirements, correcting the fault information model, and repeating the step E until the error is less than 20% or reaches the preset correction times to obtain a final estimated result.
The actual grid parameters include: the number of system nodes, line parameters, transformer parameters, system power flow parameters and generator parameters.
The step B specifically comprises the following substeps:
b1, exporting actually measured sag data of each node;
b2, screening the actually measured sag data to obtain the total fault frequency of the actually measured system, wherein the specific method comprises the following steps:
b2-1, carrying out normalization processing on sag events with short sag occurrence time intervals and the same sag types;
b2-2, carrying out statistics again on the sag events of each node to obtain the fault frequency in the actual measurement period, and further obtaining the total fault frequency F of the actual measurement systemnumThe calculation formula is as follows:
wherein T represents an actual measurement period in years, and NtRepresenting the failure frequency in the actual measurement period;
b3, carrying out statistical analysis on the fault occurrence types, and establishing a fault type information model Ptype：
In the formula, PtypeRepresenting the probability of occurrence of each type of fault based on the measured data; pLG、P2LG、P2L、P3LGRespectively representing the fault probabilities of single-phase grounding, two-phase interphase and three-phase grounding;
b4, establishing a fault duration information model PdurThe establishing method comprises the following steps:
b4-1, extracting the duration of each sag when the actual measurement period reaches 3-5 years or the actual measurement sag data reaches more than 300 groups, fitting a mathematical model of the duration by using MATLAB, and establishing a fault duration information model Pdur；
B4-2, when the actual measurement period is less than 3 years or the actual measurement sag data is less than 300 groups, adopting a universal fault duration information model P of which the fault duration follows the standard normal distribution with the expectation of 0.06s and the standard deviation of 0.01sdurThe calculation formula is as follows:
Pdur＝N(0.06,0.01)
b5 selection of SARFI90The index is taken as each nodeReducing evaluation index, and calculating actual measurement temporary reduction frequency of each node, wherein SARFI90The calculation formula of the index is as follows:
in the formula, DTRepresenting the total days in the actual measurement period, D representing the number of days in the index calculation period, and taking 365 and NTIndicating that the node has the temporary descending frequency with the temporary descending amplitude lower than 90 percent in the monitoring period T.
The method for establishing the system fault information model in the step C comprises the following steps:
c1, for the fault line, assuming that the line fault probability is in direct proportion to the line length, counting the length of each line, further obtaining the probability of each line fault, and establishing a fault line information model Pline：
In the formula, K is the total number of lines; plineRepresenting the probability of a line fault; pj(j ═ 1,2, …, K) is the probability of failure of the jth line, and
Ljrepresents the length of the j-th line;
c2, for the fault position, assuming that the probability of the fault occurring at each point on the line is the same, the fault position obeys [0,1]]Is uniformly distributed, and a fault position information model P is establishedspotThe calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
representing the fault probability of the N position intervals of the jth line;
c3, for the fault impedance, assuming that the fault impedance follows a standard normal distribution with the expectation of 5 omega and the standard deviation of 1 omega, establishing a fault impedance information model PresThe calculation formula is as follows:
Pres＝N(5,1)
step D, generating fault information based on Latin hypercube sampling, and establishing a fault information original database further comprises the following substeps:
d1, establishing a fault line information model P according to the step ClineAdopting Latin hypercube sampling to obtain a failure line original database;
d2, establishing fault location information model P according to step CspotAdopting Latin hypercube sampling to obtain a failure position original database;
d3, establishing a fault type information model P according to the step BtypeAdopting Latin hypercube sampling to obtain a failure type original database;
d4, establishing a fault duration information model P according to the step BdurAdopting Latin hypercube sampling to obtain a duration original database;
d5, establishing a fault impedance information model P according to the step CresAnd adopting Latin hypercube sampling to obtain a fault impedance original database.
The step E further comprises the sub-steps of:
e1, inputting system parameters, and performing fault simulation according to the fault information original database obtained in the step D to obtain sag characteristic values of each sag event, including sag amplitude values, duration and sag types;
e2, calculating the error value of the estimated sag index and the power grid actual measurement sag index of each installed monitoring device node respectively, and assuming that the estimated sag index and the power grid actual measurement sag index of the ith node are I respectivelyes,iAnd Irel,iError value εiThe calculation formula of (2) is as follows:
e3, sequencing the nodes according to the error from large to small, setting the error allowable value of each installed monitoring device node, comparing the error of each installed monitoring device node with the error allowable value according to the node sequence with the error from large to small, if the error is not in the range of the error allowable value, obtaining the relevance between the node sag index and the fault information model by using a Pearson correlation analysis method, selecting the fault information with the maximum relevance with each node, and correcting the fault information model;
e4, regenerating a fault information original database by using the corrected fault information model, carrying out fault simulation, counting each sag event, calculating sag indexes of each node, and giving sag indexes of all nodes without the nodes of the monitoring device and sag characteristic values of each sag event of each node, wherein the sag characteristic values comprise sag types, sag amplitudes and sag durations.
The method for correcting the fault information model in the step E3 includes:
e3-1 data processing
Screening out sag indexes of each node and fault information causing sag events of each time aiming at each node, wherein the fault information comprises a fault line, a fault position, fault impedance and fault duration;
e3-2 normalization of data
And E3-1, performing normalization processing on the data obtained in the step E3-1 by using a dispersion normalization method, namely performing linear transformation on the original data to map a transformation result between [0 and 1], wherein the transformation function is as follows:
in the formula, c*Normalizing the data obtained in step E3-1C is original data before mapping, max is the maximum value of the sample data, and min is the minimum value of the sample data;
e3-3 model correction
And (4) obtaining the relevance between the sag indexes of each node and each fault information by using a Pearson correlation analysis method, sequencing the relevance of each node, selecting the fault information with the maximum relevance with the sag indexes, and correcting the fault information model established in the step B, C.
In the step E3-3, the specific method for model modification is as follows:
e3-3-1, selecting the fault information with the maximum relevance with the node, and adding a correction coefficient alpha to the corresponding fault information model to correct the fault information model;
e3-3-2, assuming that the failure information having the greatest relevance to node 1 is a failed line, the line connected to node 1 is corrected, and the correction coefficient α is added to each of the line failure rates connected to node 11And the corrected line fault rate of the line e connected with the node 1 is as follows:
Pe,re＝Pe+α1
in the formula, Pe,reFor corrected line fault rate, P, of line eeIs the original line fault rate of line e;
e3-3-3, subtracting correction coefficient alpha from the line with longer electrical distance from node 11If the corrected line fault rate of the line f connected to the node 1 is:
Pf,re＝Pf-α1
in the formula, Pf,reFor corrected line fault rate, P, of line ffIs the original line fault rate of line f;
e3-3-4, obtaining a new line fault information model after the correction, regenerating a fault information original database, carrying out fault simulation, counting each sag event, calculating each node sag index, comparing the estimated sag index of the node 1 with the power grid actual measurement sag index, carrying out error calculation after correction, if the error is more than 20%, continuing the correction until the error is less than 20% or reaches the preset correction times, and finally obtaining the line fault rate of the line connected with the node 1;
e3-3-5, correcting the h nodes one by one according to the correction method, if the nodes which are not in the error range still exist after correction, simultaneously correcting the nodes which are not in the error range according to the correction method, and finally obtaining the fault rate of the line connected with the h nodes provided with the monitoring devices.
The calculation method of the correction coefficient alpha is as follows:
1) assuming that the system has H nodes, wherein H nodes are provided with monitoring devices, the H nodes are respectively numbered as 1,2,3, …, H, H +1, H +2, …, H, namely the former H nodes are nodes provided with monitoring devices, and the latter H-H nodes are nodes not provided with monitoring devices;
2) the errors of the estimated sag indexes of the h nodes and the actually measured sag indexes of the power grid are assumed to be epsilon respectively1,ε2,ε3,…,εhSequentially correcting h nodes, and recording the correction coefficient of each node as alpha1,α2,α3,…,αh；
3) First, for node 1, set α1The initial value is 0.01, and each correction amount is 0.01, the correction coefficient of the k-th correction is alpha1＝|0.01+0.01*k|；
Wherein alpha is1The positive and negative of (A) are selected according to the following rules:
a) when the sag index of the node 1 is positively correlated with the fault information, the sag index is increased along with the increase of the fault information, and if epsilon is1If the estimated sag index is positive, the estimated sag index is larger than the actually measured sag index, and the fault information correction coefficient alpha is1Is negative; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is positive;
b) when the sag index of the node 1 is negatively correlated with the fault information, the sag index decreases with the increase of the fault information, if epsilon1If the estimated sag index is positive, the estimated sag index is larger than the actually measured sag index, and the fault information correction coefficient alpha is1Is positive; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is negative.
The invention has the beneficial effects that:
according to the voltage sag random estimation method based on the actual power grid monitoring information, provided by the invention, a fault random information model is established according to the actual power grid sag event, so that the fault information of the actual power grid can be more accurately reflected; the Latin hypercube sampling method is adopted to replace the traditional Monte Carlo method, and the sag condition of each node of the actual power grid can be estimated more quickly and stably on the premise of ensuring the accuracy; error calculation is carried out on the sag estimation result and the actual power grid sag estimation result, if the error does not meet the precision requirement, sag information with strong relevance with each node is selected by utilizing a Pearson correlation analysis method, a sag information model is corrected until the error meets the requirement or reaches the preset correction times, the estimation result is more accurate, the problems of poor stability, slow convergence, long time consumption, large estimation result error and the like in the existing estimation method can be effectively avoided, and the significance of timely taking measures to users and power grid companies to relieve the economic loss brought by voltage sag is obvious.
Drawings
FIG. 1 is a flow chart of a voltage sag random estimation method based on actual power grid monitoring information;
FIG. 2 is a schematic diagram of Latin hypercube sampling in an embodiment of the present invention;
FIG. 3 is a topological diagram of an actual system of a power grid in a certain city in China according to a specific embodiment of the present invention;
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and examples.
Detailed inferential analysis methods and exemplary analysis examples are disclosed below. However, the specific reasoning and analysis process details disclosed herein are for purposes of describing example analysis examples only.
It should be understood, however, that the intention is not to limit the invention to the particular exemplary embodiments disclosed, but to cover all modifications, equivalents, and alternatives falling within the scope of the disclosure. Like reference numerals refer to like elements throughout the description of the figures.
First, the following describes several indexes and methods used in the present invention:
(1) SARFI index
The SARFI index is a system mean square root value fluctuation frequency index and is used for describing the square root value fluctuation condition of a single measurement point in a specific time. The SARFI index includes two forms: one is a statistical index SARFI based on a certain threshold voltagexThe other is a statistical index SARFI based on the sensitive equipment curvecurve。
For SARFIxThe index, x is the root mean square value voltage threshold, expressed in percentage form, and the possible values are 180, 140, 120, 110, 90, 80, 70, 50 or 10, etc.; for SARFIcurveIndicators representing the frequency of voltage sag events outside the reference curve of a certain type of sensitive equipment, with different reference curves mapping different SARFIscurveAnd (4) indexes. Because the specific sensitive equipment of each node in the actual system is difficult to obtain and the tolerance curve of the sensitive equipment cannot be determined, the voltage sag condition of each node is mainly estimated in the invention, so that the value x in the invention is 90, namely SARFI is adopted90The index is used as the sag evaluation index of each node, and the SARFI of a certain node90The index calculation formula is as follows:
in the formula, DTRepresenting the total number of days in the actual measurement period, D representing the number of days in the index calculation period, where the value 365 is taken in years, NTIndicating that the node has the temporary descending frequency with the temporary descending amplitude lower than 90 percent in the monitoring period T.
(2) Latin hypercube sampling
In conjunction with the Latin hypercube sampling diagram shown in FIG. 2, assume a total of M random variables X in a probability problem1、X2、…、XM，XmIs any one of the random variables, and XmThe cumulative distribution function of (d) is: y ism＝Fm(Xm)。
Assuming the sampling times are N, the cumulative distribution function Y is obtainedmThe vertical axis of (1) is divided into N equal intervals, and the width of each interval is 1/N. Let each random variable be independent of each other, xmnIs the nth sample of the mth variable.
The Latin hypercube sampling method comprises the following basic steps:
1) generating an M x N dimensional matrix LM×NEach row of the matrix is a random sequence of (1, N) integers, amnIs m rows and n columns of elements;
2) generating an MxN dimensional matrix UM×NEach element of the matrix obeys [0,1]]Uniformly distributed, umnIs m rows and n columns of elements;
3) calculating to obtain an M multiplied by N dimensional sampling matrix XM×N，xmnFor its m rows and n columns of elements, then:
wherein M is 1,2, …, M; n is 1,2, …, N.
(3) Pearson correlation coefficient analysis method
The basic principle of the pearson correlation analysis is that assuming that there are two variables a, b, the pearson correlation coefficient of the two variables can be calculated by the following formula:
in the formula, Ra,bThe correlation of the variables a and b is represented,
representing the mathematical expectation of variables a and b, respectively.
The range of the Pearson correlation coefficient is [ -1,1], and the closer the absolute value is to 1, the stronger the correlation is; the closer the absolute value is to 0, the weaker the correlation is, the less the correlation coefficient is than 0 means that 2 variables are in negative correlation, and the more the correlation coefficient is than 0 means that 2 variables are in positive correlation. The correspondence between the absolute value of the pearson correlation coefficient and the strength of the correlation is shown in table 1.
TABLE 1 correlation between absolute value of Pearson correlation coefficient and correlation intensity
| | Relevance | |
| 0≤|R|<0.2 | Very weak | |
| 0.2≤|R|<0.4 | Weak (weak) | |
| 0.4≤|R|<0.6 | Medium and high grade | |
| 0.6≤|R|<0.8 | High strength | |
| 0.8≤|R|≤1.0 | Is very strong |
Fig. 1 is a flowchart of a voltage sag random estimation method based on actual grid monitoring information, as shown in fig. 1, the method includes the following steps:
A. reading actual power grid parameters, and processing to obtain a node admittance matrix;
B. acquiring actually measured sag data of each node of an actual power grid monitoring system, and processing the dataObtaining a fault information model based on measured data, including a fault type information model PtypeAnd a fault duration information model Pdur；
C. Establishing a system fault information model including a fault line information model PlineFault location information model PspotAnd a fault impedance information model Pres；
D. Adopting Latin hypercube sampling to generate fault information, and establishing a fault information original database which comprises a fault line original database, a fault position original database, a fault type original database, a duration original database and a fault impedance original database;
E. carrying out fault simulation to obtain sag characteristic values of each sag event, wherein the sag characteristic values comprise sag amplitude values, duration times and sag types; calculating sag indexes of all nodes, comparing estimated sag indexes of nodes with installed monitoring devices with actually measured sag indexes of a power grid, and calculating to obtain errors of estimated results and actually measured results; and E, judging whether the error is less than 20%, if the error does not meet the requirements, correcting the fault information model, and repeating the step E until the error is less than 20% or reaches the preset correction times to obtain a final estimated result.
Specifically, in the step a, reading an actual power grid parameter, and processing the parameter to obtain a node admittance matrix specifically includes:
a1, recording the number of system nodes, line parameters, transformer parameters, system power flow parameters and generator parameters according to an actual power grid;
and A2, processing the parameters to obtain a node admittance matrix.
Specifically, in the step B, the actually measured sag data of each node of the actual power grid monitoring system is obtained, and a fault information model based on the actually measured data is obtained after processing, including the fault type information model PtypeAnd a fault duration information model PdurThe method comprises the following substeps:
b1, exporting actually measured sag data of each node;
b2, screening the actually measured sag data to obtain the total fault frequency of the actually measured system, and specifically comprising the following substeps:
b2-1, normalizing sag events with short sag occurrence time intervals (within a few seconds) and the same sag types, and considering the sag events to be caused by the same fault event;
b2-2, carrying out statistics again on the sag events of each node to obtain the fault frequency in the actual measurement period, and further obtaining the annual total fault frequency F of the actual measurement systemnum：
Wherein T represents an actual measurement period in years, and NtRepresenting the failure frequency in the actual measurement period;
b3, because the single-phase, two-phase and three-phase sag are respectively mainly caused by single-phase, two-phase and three-phase faults, the sag type probability model obtained by statistics of the measured sag data of the power grid can be used as the probability model of each fault type. In the existing sag estimation method, a probability model which is relatively universal at home and abroad is generally adopted for the occurrence of fault types, but when an actual power grid is estimated, the probability of the power grid for the occurrence of various types of faults may have the characteristics of the power grid and is not necessarily consistent with the universal model, so that the sag estimation method is used for modeling the actual power grid again so as to meet the actual condition of the power grid. Carrying out statistical analysis on the type of the fault occurrence, and establishing a fault type information model P based on the measured datatype：
In the formula, PtypeRepresenting the probability of occurrence of each type of fault based on the measured data; pLG、P2LG、P2L、P3LGRespectively showing the fault probability of single-phase grounding, two-phase interphase and three-phase grounding.
B4, establishing a fault duration information moduleType PdurThe establishing method comprises the following steps:
b4-1, under the condition of more actual measurement data, when the actual measurement period can reach 3-5 years or the actual measurement sag data reach more than 300 groups, extracting the duration of each sag, fitting a mathematical model of the duration by adopting a function (such as a Gaussian function) in MATLAB, and establishing a fault duration information model Pdur；
B4-2, under the condition of less measured data, such as short measurement period or measured sag data less than 300 groups, the current universal fault duration information model P is still adopteddurThat is, assuming that the fault duration follows a standard normal distribution with an expected 0.06s and a standard deviation of 0.01s, the calculation formula is:
Pdur＝N(0.06,0.01)
b5 selection of SARFI90Index, calculating the actual measurement temporary reduction frequency of each node, wherein SARFI90The calculation formula of the index is as follows:
in the formula, DTRepresenting the total days in the actual measurement period, D representing the number of days in the index calculation period, and taking 365 and NTIndicating that the node has the temporary descending frequency with the temporary descending amplitude lower than 90 percent in the monitoring period T.
Specifically, in the step C, establishing a system fault information model includes the following substeps:
c1, for a fault line, generally assuming that the line fault probability is in direct proportion to the line length, counting the length of each line, further obtaining the fault probability of each line, and establishing a fault line information model Pline:
in the formula, K is the total number of lines; pline represents the probability of line fault; pj(j ═ 1,2, …, K) is the probability of failure of the jth line, and
Ljindicating the length of the j-th line.
C2, for the fault position, it is generally assumed that the probability of the fault occurring at each point on the line is the same, so the fault position follows [0,1]]Is uniformly distributed, and a fault position information model P is establishedspot：
In the formula (I), the compound is shown in the specification,
representing the fault probability of the N position intervals of the jth line;
c3, establishing a fault impedance information model P for the fault impedance, which is difficult to represent with precise numbers, assuming that the fault impedance follows a standard normal distribution with a standard deviation of 1 Ω and is expected to be 5 ΩresThe calculation formula is as follows:
Pres＝N(5,1)
specifically, in the step D, the fault information is generated based on latin hypercube sampling, and the establishing of the fault information original database further includes the following substeps:
d1, establishing a fault line information model P according to the step C1lineAdopting Latin hypercube sampling to obtain a failure line original database;
for a faulty line, assume that the random number y obeys 0,1]Uniformly distributed, adopting Latin hypercube sampling to generate random number y, and corresponding fault line FlineExpressed as:
in the formula, K is the total number of lines; pj(j ═ 1,2, …, K) is the probability of failure of the jth line, and
d2, establishing fault location information model P according to the step C2spotBecause the information model is continuous probability distribution, adopting Latin hypercube sampling to obtain a failure position original database;
d3, establishing a fault type information model P according to the step B3typeAdopting Latin hypercube sampling to obtain a failure type original database;
for the fault type, assume that the random number z obeys 0,1]Uniformly distributed, adopting Latin hypercube sampling to generate random number z, and corresponding fault type FtypeIs shown as
In the formula, PLG、P2LG、P2L、P3LRespectively showing the fault probability of single-phase grounding, two-phase interphase and three-phase grounding.
D4, establishing a fault duration information model P according to the step B4durBecause the probability distribution is continuous, adopting Latin hypercube sampling to obtain a duration original database;
d5, establishing fault impedance information model P according to the step C3resAnd because the probability distribution is continuous, the original fault impedance database is obtained by adopting Latin hypercube sampling.
Specifically, the step E specifically includes the following steps:
e1, inputting system parameters, and performing fault simulation according to the fault information original database obtained in the step D to obtain sag characteristic values of each sag event, including sag amplitude values, duration and sag types;
e2, statistically analyzing the sag event obtained by E1, calculating sag indexes of each node, respectively calculating error values of the estimated sag indexes and the actually measured sag indexes of the power grid of each node provided with the monitoring device, and assuming that the estimated sag indexes and the actually measured sag indexes of the power grid of the ith node are I respectivelyes,iAnd Irel,iError value εiThe calculation formula of (2) is as follows:
e3, sequencing the nodes according to the error from large to small, setting the error allowable value of each installed monitoring device node, comparing the error of each installed monitoring device node with the error allowable value according to the node sequence with the error from large to small, if the error is not in the range of the error allowable value, obtaining the relevance between the node sag index and the fault information model (such as a fault line, a fault position, fault impedance and fault duration) by using a Pearson correlation analysis method, selecting the fault information with the maximum relevance with each node, and correcting the fault information model established in the step B, C.
E4, regenerating a fault information original database by using the corrected fault information model obtained by E3, carrying out fault simulation, counting each sag event, calculating sag indexes of each node to obtain a final estimation result, and providing sag indexes of all nodes without the nodes of the monitoring device and sag characteristic values of each sag event of each node, wherein the sag characteristic values comprise sag types, sag amplitudes and sag durations.
The method for correcting the fault information model in the step E3 includes:
e3-1 data processing
Screening out sag indexes of each node and fault information causing sag events of each time aiming at each node, wherein the fault information comprises a fault line, a fault position, fault impedance and fault duration;
e3-2, carrying out normalization processing on the obtained data, and converting the data into a value between [0 and 1 ];
the normalization process is performed by min-max normalization (also called dispersion normalization), i.e. the raw data is linearly transformed, so that the result value is mapped between [0,1], and the conversion function is:
in the formula, c*Normalizing the mapped result of the data obtained in the step E3-1, wherein c is the original data before mapping, max is the maximum value of the sample data, and min is the minimum value of the sample data;
e3-3 model correction
Obtaining and sequencing the relevance between each node sag index and each fault information by adopting a Pearson correlation analysis method, selecting the fault information with the maximum relevance with the sag index, and correcting the corresponding fault mathematical model established in the step B, C, wherein the correction method comprises the following steps:
e3-3-1, selecting the fault information with the maximum relevance with the node, and adding a correction coefficient alpha to the corresponding fault information model to correct the fault information model;
e3-3-2, assuming that the failure information having the greatest relevance to node 1 is a failed line, the line connected to node 1 is corrected, and the correction coefficient α is added to each of the line failure rates connected to node 11And the corrected line fault rate of the line e connected with the node 1 is as follows:
Pe,re＝Pe+α1
in the formula, Pe,reFor corrected line fault rate, P, of line eeIs the original line fault rate of line e;
e3-3-3, subtracting correction coefficient alpha from the line with longer electrical distance from node 11If the corrected line fault rate of the line f connected to the node 1 is:
Pf,re＝Pf-α1
in the formula, Pf,reFor corrected line fault rate, P, of line ffAs origin of line fInitial line failure rate;
e3-3-4, obtaining a new line fault information model after the correction, regenerating a fault information original database, carrying out fault simulation, counting each sag event, calculating each node sag index, comparing the estimated sag index of the node 1 with the power grid actual measurement sag index, carrying out error calculation after correction, if the error is more than 20%, continuing the correction until the error is less than 20% or reaches the preset correction times, and finally obtaining the line fault rate of the line connected with the node 1;
e3-3-5, correcting the h nodes one by one according to the correction method, if the nodes which are not in the error range still exist after correction, simultaneously correcting the nodes which are not in the error range according to the correction method, and finally obtaining the fault rate of the line connected with the h nodes provided with the monitoring devices.
The calculation method of the correction coefficient alpha is as follows:
1) assuming that the system has H nodes, wherein H nodes are provided with monitoring devices, the H nodes are respectively numbered as 1,2,3, …, H, H +1, H +2, …, H, namely the former H nodes are nodes provided with monitoring devices, and the latter H-H nodes are nodes not provided with monitoring devices;
2) the errors of the estimated sag indexes of the h nodes and the actually measured sag indexes of the power grid are assumed to be epsilon respectively1,ε2,ε3,…,εhSequentially correcting the h nodes by adopting a one-by-one correction method, and recording the correction coefficient of each node as alpha1,α2,α3,…,αh；
3) First, for node 1, set α1The initial value is 0.01, and each correction amount is 0.01, the correction coefficient of the k-th correction is alpha1＝|0.01+0.01*k|；
Wherein alpha is1The positive and negative of (A) are selected according to the following rules:
a) when the sag index of the node 1 is positively correlated with the fault information, the sag index is increased along with the increase of the fault information, and if epsilon is1Positive, it shows that the estimated sag index is larger than the measured sag index, and a smaller estimation is neededTemporarily reducing the size of the index, thereby correcting the fault information by the factor alpha1Is negative; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is positive;
b) when the sag index of the node 1 is negatively correlated with the fault information, the sag index decreases with the increase of the fault information, if epsilon1If yes, the estimated sag index is larger than the actual measured sag index, and the size of the estimated sag index needs to be reduced, so that the fault information correction coefficient alpha1Is positive; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is negative.
Here, assuming that the failure information having the greatest relevance to the node 1 is a failed line, the line connected to the node 1 is corrected, and a positive correction coefficient α is added to each of the line failure rates connected to the node 11In summary, the corrected line fault rate of the line e connected to the node 1 is:
Pe,re＝Pe+α1
in the formula, Pe,reCorrected line fault rate, P, for line eeIs the original line fault rate for line e.
Since the sum of all line fault rates is 1, the correction factor α is subtracted from each line that is electrically far from the node 11That is, the corrected line fault rate of the line f connected to the node 1 is:
Pf,re＝Pf-α1
in the formula, Pf,reCorrected line fault rate, P, for line ffIs the original line fault rate for line f.
And obtaining a new line fault information model after the correction, then regenerating a fault information original database, carrying out fault simulation, counting each sag event, calculating sag indexes of each node, comparing the estimated sag indexes of the node 1 with the actually measured sag indexes of the power grid, carrying out error calculation after the correction, and if the error is not in an allowable value range, continuing the correction until the error meets the requirement or reaches the correction times, and finally obtaining the line fault rate of the line connected with the node 1.
Similarly, for the node 2, according to the same correction method, the line fault rate connected with the node 2 is obtained; by analogy, the fault rate of the lines connected with the h nodes provided with the monitoring devices can be finally obtained.
If the node errors are still not in the error range after the 'one-by-one correction', the nodes which are not in the error range are corrected integrally according to the method, namely the nodes are corrected simultaneously, and finally the fault rate of the line connected with the h nodes provided with the monitoring devices is obtained.
If the fault information with the strongest association with a certain node is not the line fault rate, the correction method is similar to the line fault rate correction method.
By the method, the corrected fault information model can be obtained.
Example 1
The technical effects of the present invention will be described below with reference to a specific embodiment.
FIG. 3 is a topological diagram of an actual system of a power grid in a certain city in China. As shown in fig. 3, the urban power grid has 2 500kV substations (nodes 1 and 2) and 19 220kV substations (nodes 3 to 21), and is formed by connecting 48 lines.
The method takes sag actual measurement data acquired by monitoring nodes of the urban power grid electric energy quality online monitoring system from 2017 month 1 to 2017 month 6 as research objects, screens sag events with short sag occurrence time intervals (within a few seconds) and the same sag types, considers the sag events to be caused by the same fault event, and carries out statistics again on the urban power grid to obtain the following results: within half a year, the total number of sag events is 38 and the number of fault events incurred is 26. Table 2 shows how each monitoring node of the urban power grid generates sag frequency, selects safi 90 as sag index, and reports sag evaluation results of each node to table 2. Statistical analysis is performed on each fault, and the occurrence probability of each type of fault is shown in table 3.
Table 2 sag estimation results of each node
| Numbering | Monitoring node | Frequency conversion time of temporary clock | SARFI90 index |
| 5 | Pu county | 4 | 8 |
| 3 | |
3 | 6 |
| 21 | Xingtang food | 2 | 4 |
| 19 | Yongle | 8 | 16 |
| 18 | Qiao Bei | 6 | 12 |
| 10 | |
3 | 6 |
| 16 | |
9 | 18 |
| 14 | |
3 | 6 |
| Total of | 38 | 76 |
TABLE 3 probability of occurrence of each type of failure
| Type of failure | Failure rate |
| LG | 53％ |
| 2L | 14 |
| 2LG | |
| 3％ | |
| 3LG | 30％ |
The simulation results and the actual measurement results of the nodes of the installed node monitoring equipment are shown in table 4 by performing actual system sag simulation analysis through two steps of E1 and E2. As can be seen from table 4, except for the node 3 and the node 18, the sag frequency simulation values of other nodes have larger errors with the measured value, and the sag estimation error of the node 10 reaches 80.00%, so that the original sag information model needs to be corrected to obtain a sag estimation random model more suitable for the actual system.
TABLE 4 comparison of simulation and actual measurement values of nodes of the installation node monitoring device
| Numbering | SARFI90Measured value | SARFI90Simulation value | Error (%) |
| 3 | 6 | 5.50 | -8.33 |
| 5 | 8 | 13.97 | 74.62 |
| 10 | 6 | 10.80 | 80.00 |
| 14 | 6 | 8.40 | 40.00 |
| 16 | 18 | 8.83 | -50.94 |
| 18 | 12 | 10.58 | -11.83 |
| 19 | 16 | 12.04 | -24.75 |
| 21 | 4 | 7.18 | 79.50 |
The mathematical model of the fault type based on the measured data is obtained in consideration of the establishment of the aforementioned fault information model, and therefore, it is not corrected. For each node SARFI in the area90The index and the fault information model are subjected to correlation coefficient calculation, and table 5 shows the correlation coefficient between the index and the fault information model of each node safi 90. As can be seen from table 5, the association between the index of each node SARFI90 and the sag route is the largest, so the method proposed by the E2.2 step will be adopted in the subsequent model modificationThe line fault is firstly corrected one by one and then corrected integrally, and the allowable error value is 20%.
TABLE 5 correlation coefficient between node SARFI90 indices and fault information model
| Node number | Fault line | Location of failure | Impedance to fault | Duration of fault |
| 1 | -0.5521 | 0.1081 | 0.1434 | 0.0018 |
| 2 | -0.4923 | 0.0636 | 0.2392 | -0.0087 |
| 3 | 0.5048 | -0.0049 | 0.1122 | 0.0066 |
| 4 | -0.1339 | -0.0876 | -0.0489 | -0.0396 |
| 5 | -0.1942 | -0.0099 | 0.0986 | 0.0141 |
| 6 | 0.3419 | 0.1756 | 0.1214 | -0.0044 |
| 7 | -0.2147 | -0.1613 | 0.0464 | 0.0160 |
| 8 | -0.4379 | -0.2388 | 0.0359 | 0.0128 |
| 9 | -0.4602 | -0.0147 | -0.1257 | -0.0216 |
| 10 | -0.5275 | 0.3374 | 0.1058 | -0.0015 |
| 11 | -0.3345 | 0.0191 | 0.1702 | -0.0113 |
| 12 | -0.3490 | -0.0154 | 0.1163 | 0.0028 |
| 13 | -0.2648 | -0.0530 | 0.0642 | 0.0031 |
| 14 | -0.2416 | 0.0027 | 0.1568 | -0.0004 |
| 15 | -0.3345 | -0.0027 | 0.1444 | -0.0107 |
| 16 | -0.1592 | -0.0288 | 0.0853 | 0.0185 |
| 17 | -0.1508 | -0.0802 | 0.0122 | 0.0110 |
| 18 | -0.3496 | -0.0738 | 0.0656 | 0.0002 |
| 19 | -0.3744 | -0.3169 | -0.0596 | -0.0092 |
| 20 | 0.3576 | -0.2897 | -0.0241 | 0.0057 |
| 21 | -0.5096 | -0.3445 | 0.0592 | -0.0038 |
Table 6 shows the comparison result between the node simulation value and the actual measurement value of the installed monitoring device after the correction. Comparing table 4 with table 6, it can be seen that the error between the corrected sag estimation result and the measured value is reduced, most nodes meet the error within 20%, and the influence of the monitoring period and some accidental factors (weather, human misoperation, etc.) is considered, so although the corrected estimated value still has a certain error with the value, the estimated value can basically meet the estimated requirement.
After the fault information model is corrected, sag random estimation is carried out again, and sag estimation results of other nodes without monitoring devices shown in table 7 are obtained.
TABLE 6 comparison of simulated values and measured values of nodes of installed monitoring devices after correction
| Node number | SARFI90Measured value | SARFI90Correcting simulation value | Error (%) |
| 3 | 6 | 5.06 | -15.67 |
| 5 | 8 | 10.57 | 32.12 |
| 10 | 6 | 4.65 | -22.50 |
| 14 | 6 | 6.60 | 10.00 |
| 16 | 18 | 7.54 | -58.11 |
| 18 | 12 | 12.68 | 5.67 |
| 19 | 16 | 13.42 | -16.12 |
| 21 | 4 | 4.75 | 18.75 |
TABLE 7 sag estimation results for nodes without installed monitoring devices
| Node number | Predictive SARFI90Value of | Node number | Predictive SARFI90Value of |
| 1 | 5.50 | 11 | 6.59 |
| 2 | 6.60 | 12 | 3.67 |
| 4 | 1.88 | 13 | 1.50 |
| 6 | 9.62 | 15 | 6.60 |
| 7 | 9.57 | 17 | 9.39 |
| 8 | 9.25 | 20 | 13.42 |
| 9 | 0.64 | - | - |
The present invention is not limited to the above embodiments, and any changes or substitutions that can be easily made by those skilled in the art within the technical scope of the present invention are also within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (5)
1. A random voltage sag estimation method based on actual power grid monitoring information is characterized by comprising the following steps:
A. reading actual power grid parameters, and processing to obtain a node admittance matrix;
B. acquiring actually measured sag data of each node of an actual power grid monitoring system, and processing the actually measured sag data to obtain a fault information model based on the actually measured data, wherein the fault information model comprises a fault type information model PtypeAnd a fault duration information model Pdur；
C. Establishing a system fault information model including a fault line information model PlineFault location information model PspotAnd a fault impedance information model Pres；
D. Adopting Latin hypercube sampling to generate fault information, and establishing a fault information original database which comprises a fault line original database, a fault position original database, a fault type original database, a duration original database and a fault impedance original database;
E. carrying out fault simulation to obtain sag characteristic values of each sag event, wherein the sag characteristic values comprise sag amplitude values, duration times and sag types; calculating sag indexes of all nodes, comparing estimated sag indexes of nodes with installed monitoring devices with actually measured sag indexes of a power grid, and calculating to obtain errors of estimated results and actually measured results; judging whether the error is less than 20%, if the error does not meet the requirements, correcting the fault information model, and repeating the step E until the error is less than 20% or reaches the preset correction times to obtain a final estimated result;
the step E further comprises the sub-steps of:
e1, inputting system parameters, and performing fault simulation according to the fault information original database obtained in the step D to obtain sag characteristic values of each sag event, including sag amplitude values, duration and sag types;
e2, calculating the error value of the estimated sag index and the power grid actual measurement sag index of each installed monitoring device node respectively, and assuming that the estimated sag index and the power grid actual measurement sag index of the ith node are I respectivelyes,iAnd Irel,iError value εiThe calculation formula of (2) is as follows:
e3, sequencing the nodes according to the error from large to small, setting the error allowable value of each installed monitoring device node, comparing the error of each installed monitoring device node with the error allowable value according to the node sequence with the error from large to small, if the error is not in the range of the error allowable value, obtaining the relevance between the node sag index and the fault information model by using a Pearson correlation analysis method, selecting the fault information with the maximum relevance with each node, and correcting the fault information model;
the method for correcting the fault information model comprises the following steps:
e3-1 data processing
Screening out sag indexes of each node and fault information causing sag events of each time aiming at each node, wherein the fault information comprises a fault line, a fault position, fault impedance and fault duration;
e3-2 normalization of data
And E3-1, performing normalization processing on the data obtained in the step E3-1 by using a dispersion normalization method, namely performing linear transformation on the original data to map a transformation result between [0 and 1], wherein the transformation function is as follows:
in the formula, c*Normalizing the mapped result of the data obtained in the step E3-1, wherein c is the original data before mapping, max is the maximum value of the sample data, and min is the minimum value of the sample data;
e3-3 model correction
Obtaining the relevance between the sag indexes of each node and each fault information by using a Pearson correlation analysis method, sequencing the relevance of each node, selecting the fault information with the maximum relevance with the sag indexes, and correcting the fault information model established in the step B, C;
the specific method for correcting the model comprises the following steps:
e3-3-1, selecting the fault information with the maximum relevance with the node, and adding a correction coefficient alpha to the corresponding fault information model to correct the fault information model;
e3-3-2, assuming that the failure information having the greatest relevance to node 1 is a failed line, the line connected to node 1 is corrected, and the correction coefficient α is added to each of the line failure rates connected to node 11And the corrected line fault rate of the line e connected with the node 1 is as follows:
Pe,re＝Pe+α1
in the formula, Pe,reFor corrected line fault rate, P, of line eeIs the original line fault rate of line e;
e3-3-3, subtracting correction coefficient alpha from the line with longer electrical distance from node 11If the corrected line fault rate of the line f connected to the node 1 is:
Pf,re＝Pf-α1
in the formula, Pf,reFor corrected line fault rate, P, of line ffIs the original line fault rate of line f;
e3-3-4, obtaining a new line fault information model after the correction, regenerating a fault information original database, carrying out fault simulation, counting each sag event, calculating each node sag index, comparing the estimated sag index of the node 1 with the power grid actual measurement sag index, carrying out error calculation after correction, if the error is more than 20%, continuing the correction until the error is less than 20% or reaches the preset correction times, and finally obtaining the line fault rate of the line connected with the node 1;
e3-3-5, correcting the h nodes one by one according to the correction method, if the nodes which are not in the error range still exist after correction, simultaneously correcting the nodes which are not in the error range according to the correction method, and finally obtaining the fault rate of the line connected with the h nodes provided with the monitoring devices;
e4, regenerating a fault information original database by using the corrected fault information model, carrying out fault simulation, counting each sag event, calculating sag indexes of each node, and giving sag indexes of all nodes without the nodes of the monitoring device and sag characteristic values of each sag event of each node, wherein the sag characteristic values comprise sag types, sag amplitudes and sag durations;
the calculation method of the correction coefficient alpha is as follows:
1) assuming that the system has H nodes, wherein H nodes are provided with monitoring devices, the H nodes are respectively numbered as 1,2,3, …, H, H +1, H +2, …, H, namely the former H nodes are nodes provided with monitoring devices, and the latter H-H nodes are nodes not provided with monitoring devices;
2) the errors of the estimated sag indexes of the h nodes and the actually measured sag indexes of the power grid are assumed to be epsilon respectively1,ε2,ε3,…,εhSequentially correcting h nodes, and recording the correction coefficient of each node as alpha1,α2,α3,…,αh；
3) First, for node 1, set α1The initial value is 0.01, and each correction amount is 0.01, the correction coefficient of the k-th correction is alpha1＝|0.01+0.01*k|；
Wherein alpha is1The positive and negative of (A) are selected according to the following rules:
a) when the sag index of the node 1 is positively correlated with the fault information, the sag index is increased along with the increase of the fault information, and if epsilon is1If the estimated sag index is positive, the estimated sag index is larger than the actually measured sag index, and the fault information correction coefficient alpha is1Is negative; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is positive;
b) when the sag index of the node 1 is negatively correlated with the fault information, the sag index decreases with the increase of the fault information, if epsilon1If the estimated sag index is positive, the estimated sag index is larger than the actually measured sag index, and the fault information correction coefficient alpha is1Is positive; if epsilon1If the estimated sag index is negative, the estimated sag index is smaller than the actually measured sag index, and the fault information correction coefficient alpha is1Is negative.
2. The method according to claim 1, wherein the actual grid parameters include: the number of system nodes, line parameters, transformer parameters, system power flow parameters and generator parameters.
3. The voltage sag random estimation method based on actual power grid monitoring information according to claim 1, wherein the step B specifically comprises the following substeps:
b1, exporting actually measured sag data of each node;
b2, screening the actually measured sag data to obtain the total fault frequency of the actually measured system, wherein the specific method comprises the following steps:
b2-1, carrying out normalization processing on sag events with short sag occurrence time intervals and the same sag types;
b2-2, carrying out statistics again on the sag events of each node to obtain the fault frequency in the actual measurement period, and further obtaining the total fault frequency F of the actual measurement systemnumThe calculation formula is as follows:
wherein T represents an actual measurement period in years, and NtShowing the measured weekFrequency of faults within a period;
b3, carrying out statistical analysis on the fault occurrence types, and establishing a fault type information model Ptype：
In the formula, PtypeRepresenting the probability of occurrence of each type of fault based on the measured data; pLG、P2LG、P2L、P3LGRespectively representing the fault probabilities of single-phase grounding, two-phase interphase and three-phase grounding;
b4, establishing a fault duration information model PdurThe establishing method comprises the following steps:
b4-1, extracting the duration of each sag when the actual measurement period reaches 3-5 years or the actual measurement sag data reaches more than 300 groups, fitting a mathematical model of the duration by using MATLAB, and establishing a fault duration information model Pdur；
B4-2, when the actual measurement period is less than 3 years or the actual measurement sag data is less than 300 groups, adopting a universal fault duration information model P of which the fault duration follows the standard normal distribution with the expectation of 0.06s and the standard deviation of 0.01sdurThe calculation formula is as follows:
Pdur＝N(0.06,0.01)
b5 selection of SARFI90The index is used as the sag evaluation index of each node, and the actually measured sag frequency of each node is calculated, wherein SARFI90The calculation formula of the index is as follows:
in the formula, DTRepresenting the total days in the actual measurement period, D representing the number of days in the index calculation period, and taking 365 and NTAnd indicating that the node has the temporary reduction frequency with the temporary reduction amplitude lower than 90% in the monitoring period T.
4. The voltage sag random estimation method based on actual power grid monitoring information according to claim 1, wherein the method for establishing the system fault information model in the step C comprises the following steps:
c1, for the fault line, assuming that the line fault probability is in direct proportion to the line length, counting the length of each line, further obtaining the probability of each line fault, and establishing a fault line information model Pline：
In the formula, K is the total number of lines; plineRepresenting the probability of a line fault; pj(j ═ 1,2, …, K) is the probability of failure of the jth line, and
Ljrepresents the length of the j-th line;
c2, for the fault position, assuming that the probability of the fault occurring at each point on the line is the same, the fault position obeys [0,1]]Is uniformly distributed, and a fault position information model P is establishedspotThe calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
representing the fault probability of the N position intervals of the jth line;
c3, for the fault impedance, assuming that the fault impedance follows a standard normal distribution with the expectation of 5 omega and the standard deviation of 1 omega, establishing a fault impedance information model PresThe calculation formula is as follows:
Pres＝N(5,1)。
5. the voltage sag random estimation method based on actual power grid monitoring information as claimed in claim 1, wherein the step D generates fault information based on latin hypercube sampling, and the establishing of the fault information original database further comprises the following sub-steps:
d1, establishing a fault line information model P according to the step ClineAdopting Latin hypercube sampling to obtain a failure line original database;
d2, establishing fault location information model P according to step CspotAdopting Latin hypercube sampling to obtain a failure position original database;
d3, establishing a fault type information model P according to the step BtypeAdopting Latin hypercube sampling to obtain a failure type original database;
d4, establishing a fault duration information model P according to the step BdurAdopting Latin hypercube sampling to obtain a duration original database;
d5, establishing a fault impedance information model P according to the step CresAnd adopting Latin hypercube sampling to obtain a fault impedance original database.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810738766.8A CN109034461B (en) | 2018-07-06 | 2018-07-06 | Voltage sag random estimation method based on actual power grid monitoring information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810738766.8A CN109034461B (en) | 2018-07-06 | 2018-07-06 | Voltage sag random estimation method based on actual power grid monitoring information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109034461A CN109034461A (en) | 2018-12-18 |
| CN109034461B true CN109034461B (en) | 2021-04-20 |
Family
ID=64640581
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810738766.8A Active CN109034461B (en) | 2018-07-06 | 2018-07-06 | Voltage sag random estimation method based on actual power grid monitoring information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109034461B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135659A (en) * | 2019-05-29 | 2019-08-16 | 南京亿顺弘信息技术有限公司 | A kind of prediction technique of the voltage dip mean intensity based on least square method |
| CN110442939B (en) * | 2019-07-25 | 2022-08-09 | 国网山西省电力公司电力科学研究院 | Direct-current power distribution and utilization system reliability evaluation method based on network equivalence |
| CN110795866B (en) * | 2019-11-18 | 2024-04-23 | 国网青海省电力公司 | Voltage sag area prediction method based on limit fault point |
| CN111007357B (en) * | 2019-12-18 | 2021-12-24 | 广东电网有限责任公司 | Sag domain identification method considering load voltage sag tolerance capacity curve |
| CN111506878B (en) * | 2020-04-08 | 2022-05-03 | 国网湖北省电力有限公司电力科学研究院 | Node number selection device and method based on sag system index random sample estimation |
| CN113011026B (en) * | 2021-03-19 | 2022-06-17 | 福州大学 | Power grid voltage sag simulation method |
| CN113281584B (en) * | 2021-04-21 | 2022-09-09 | 东南大学溧阳研究院 | Voltage sag source positioning method based on correction parameters |
| CN113609681B (en) * | 2021-08-06 | 2022-02-08 | 四川大学 | Voltage sag amplitude evaluation method based on electric energy quality measured data correction |
| CN113919162B (en) * | 2021-10-15 | 2022-08-09 | 福州大学 | Voltage sag risk early warning method based on simulation and multi-source measured data fusion |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101504437B (en) * | 2009-03-06 | 2011-06-15 | 华南理工大学 | Method for searching electric network weakness causing voltage sag |
| CN106093638A (en) * | 2016-06-06 | 2016-11-09 | 国网天津市电力公司 | A kind of Voltage Drop root-mean-square value and the mode identification method falling frequency estimation |
| CN106372812A (en) * | 2016-09-23 | 2017-02-01 | 广东工业大学 | Voltage sag risk assessment method and system |
| CN108181527A (en) * | 2017-12-30 | 2018-06-19 | 国网天津市电力公司电力科学研究院 | A kind of appraisal procedure of the voltage dip synthesis severity based on energy indexes |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104115077B (en) * | 2011-12-16 | 2018-12-07 | 施耐德电气美国股份有限公司 | Co-location electrical architecture |
-
2018
- 2018-07-06 CN CN201810738766.8A patent/CN109034461B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101504437B (en) * | 2009-03-06 | 2011-06-15 | 华南理工大学 | Method for searching electric network weakness causing voltage sag |
| CN106093638A (en) * | 2016-06-06 | 2016-11-09 | 国网天津市电力公司 | A kind of Voltage Drop root-mean-square value and the mode identification method falling frequency estimation |
| CN106372812A (en) * | 2016-09-23 | 2017-02-01 | 广东工业大学 | Voltage sag risk assessment method and system |
| CN108181527A (en) * | 2017-12-30 | 2018-06-19 | 国网天津市电力公司电力科学研究院 | A kind of appraisal procedure of the voltage dip synthesis severity based on energy indexes |
Non-Patent Citations (2)
| Title |
|---|
| "CLASSIFICATION OF VOLTAGE SAG DISTURBANCE SOURCES USING FUZZY COMPREHENSIVE EVALUATION METHOD";Chenyi LI，Yonghai XU，Jiali YANG，Yapen WU;《24th International Conference & Exhibition on Electricity Distribution (CIRED) 》;20171031;第1-5页 * |
| "电压暂降源的定位与暂降经济性研究";章雪萌;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20100915;第42-244页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109034461A (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109034461B (en) | Voltage sag random estimation method based on actual power grid monitoring information | |
| CN112699913B (en) | Method and device for diagnosing abnormal relationship of household transformer in transformer area | |
| CN109103874B (en) | Power distribution network reliability evaluation method considering load characteristics and distributed power access | |
| CN110516912B (en) | Method for identifying household transformer relation of distribution station | |
| Cheng et al. | Enhanced state estimation and bad data identification in active power distribution networks using photovoltaic power forecasting | |
| CN103699668A (en) | Power distribution network electric equipment combination state evaluation method based on data section consistency | |
| CN112068064A (en) | Running three-phase voltage transformer error calculation method with background influence quantity deducted | |
| CN113189418B (en) | Topological relation identification method based on voltage data | |
| CN117748507B (en) | Distribution network harmonic access uncertainty assessment method based on Gaussian regression model | |
| CN109657913B (en) | Transmission and distribution network joint risk assessment method considering distributed power supply | |
| CN107305651B (en) | Power transmission system reliability assessment method and system | |
| CN110705859A (en) | PCA-self-organizing neural network-based method for evaluating running state of medium and low voltage distribution network | |
| Emmanuel et al. | Net electricity clustering at different temporal resolutions using a SAX-based method for integrated distribution system planning | |
| CN110443481B (en) | Power distribution automation terminal state evaluation system and method based on hybrid K-nearest neighbor algorithm | |
| CN113466520A (en) | Method for on-line identifying misalignment electric energy meter | |
| Wei et al. | Online distribution system topology monitoring with limited smart meter communication | |
| CN108090237A (en) | A kind of modeling method of definite distribution transformer failure rate | |
| CN116029165A (en) | Power cable reliability analysis method and system considering lightning influence | |
| CN112001073A (en) | Reliability analysis research method of traction power supply system | |
| CN113780775A (en) | Method and system for evaluating theoretical line loss calculation result of power grid | |
| CN107305648B (en) | Power distribution network operation and power supply reliability assessment method and system | |
| Xie et al. | Energy System Time Series Data Quality Maintenance System Based on Data Mining Technology | |
| Lingang et al. | Research on integrated calculation method of theoretical line loss of MV and LV distribution Network based on Adaboost integrated learning | |
| CN106296461B (en) | Power grid parameter identification method based on particle swarm optimization algorithm and local state estimation | |
| Wang et al. | Study on Daily Line Loss Calculation Based on Operation Data and Resistance Characteristics of Low Voltage Distribution Transformer Area |
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 |