CN110110907B - Method for extracting characteristic parameters of low-voltage transformer area - Google Patents

Method for extracting characteristic parameters of low-voltage transformer area Download PDF

Info

Publication number
CN110110907B
CN110110907B CN201910316214.2A CN201910316214A CN110110907B CN 110110907 B CN110110907 B CN 110110907B CN 201910316214 A CN201910316214 A CN 201910316214A CN 110110907 B CN110110907 B CN 110110907B
Authority
CN
China
Prior art keywords
line
voltage
item set
parameters
low
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
Application number
CN201910316214.2A
Other languages
Chinese (zh)
Other versions
CN110110907A (en
Inventor
滕永兴
曹国瑞
杨霖
孙淑娴
朱逸群
于学均
闵诚
钟睿君
王子南
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Corp of China SGCC
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by State Grid Corp of China SGCC, State Grid Tianjin Electric Power Co Ltd, Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN201910316214.2A priority Critical patent/CN110110907B/en
Publication of CN110110907A publication Critical patent/CN110110907A/en
Application granted granted Critical
Publication of CN110110907B publication Critical patent/CN110110907B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • 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
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply

Abstract

The invention relates to a method for extracting characteristic parameters of a low-voltage station area, which comprises the following steps: s1: establishing a station area characteristic index system; s2: adopting an association rule mining algorithm-Apriori to realize the selection of characteristic parameters of a low-voltage distribution area; s3: clustering sample data by adopting an improved K-Means algorithm aiming at the selected low-voltage distribution area characteristic parameters; s4: and extracting the line parameters of the distribution line by adopting a compact 2D FDFD algorithm. According to the invention, the circuit parameters in the distribution line are extracted, so that the line loss rate of the system can be reduced, the economic loss is reduced, and the energy-saving and emission-reducing work is further advanced.

Description

Method for extracting characteristic parameters of low-voltage transformer area
Technical Field
The invention belongs to the field of electric power metering, relates to an electric power data mining and characteristic parameter extraction technology, and particularly relates to a method for extracting characteristic parameters of a low-voltage station area.
Background
The line loss rate is an important technical and economic index of the power system, and is a comprehensive technical and economic index for measuring the operation level and the management level of the power enterprises. In an effort to reduce the power consumption of the power grid, to minimize the management line loss, controlling the technical line loss within a reasonable range is an indispensable task for power enterprises at all levels.
The low-voltage transformer area is an important component of electricity consumption, low-voltage electricity consumption residents have large electricity consumption quantity and wide distribution, the structure of the power distribution network is complex, a big data analysis technology is applied, massive user electricity consumption data are mined, typical characteristic parameters are extracted, and the operation state of the low-voltage transformer area is analyzed in a multi-dimensional and comprehensive mode.
With the development of social economy and the increase of electricity load, the line loss electric quantity of the low-voltage distribution network accounts for about 40% of the loss of the whole power network, is the voltage layer with the largest loss, and has the maximum loss reduction and energy saving potential. The low-voltage transformer area is used as the end link of the power system, the root of the 'quarter' management, and the line loss rate is one of important assessment indexes for the transformer area management. The method accurately and rapidly calculates the line loss rate of the transformer area, and provides basis for formulating reasonable loss reduction measures, which is an important task of power supply enterprises. The original data required by line loss calculation is element parameters and wiring diagrams related to the power grid structure; and secondly, the operation parameters (current, voltage, power factor, active and reactive power, etc.) of the power network. However, due to the uneven construction and management conditions of the low-voltage transformer areas, huge number of transformer areas and terminal users, incomplete transformer account management, complex and various line distribution, and large difference of acquisition success rates of the electricity acquisition systems, a great amount of manpower and material resources are required to collect necessary operation data and data no matter the theoretical line loss rate is calculated or the statistical line loss rate is estimated, the workload is very large, and the power supply department is difficult to perform calculation work once a month. In addition, the line loss calculation value is inaccurate due to the common management reasons such as unclear user change relation, poor meter reading quantity, electricity stealing, metering faults and the like in the conventional line loss management of the transformer area. Based on the above-mentioned current situation, how to calculate the line loss rate of the area rapidly and accurately is a problem to be solved.
The loss in the line is not only related to the magnitude of the power consumption load, but also has a close relation with the magnitude of the impedance of the line and the time, but the power consumption load is determined by a user and is an uncontrollable factor, and a method for reducing the line loss by extracting the impedance of the line to perform some compensation is attractive. However, in the existing analysis method of the electric power system in the world, the method is mainly used for extracting parameters of the transmission line, but the method for extracting parameters of the distribution line in the low-voltage distribution network is few, so that whether the parameters of the distribution line can be extracted from the collected data based on a large amount of electricity consumption data is a problem to be solved.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a method for extracting characteristic parameters of a low-voltage transformer area.
The invention solves the technical problems by adopting the following technical scheme:
the extraction method of the characteristic parameters of the low-voltage station area is characterized by comprising the following steps of: the method comprises the following steps:
s1: establishing a station area characteristic index system;
s2: adopting an association rule mining algorithm-Apriori to realize the selection of characteristic parameters of a low-voltage distribution area;
s3: clustering sample data by adopting an improved K-Means algorithm aiming at the selected low-voltage distribution area characteristic parameters;
s4: and extracting the line parameters of the distribution line by adopting a compact 2D FDFD algorithm.
And, the step S1 of establishing a characteristic index system of the station area includes:
(1) Screening out electric characteristic parameters related to the grid structure and the load of the transformer area: comprising a power supply radius X 1 (m) total length X of low-voltage line 2 (m) load factor X 3 (%) and its electric property and proportion X 4 (%) and line loss rate X 5 (％)；
(2) The method comprises the following steps of: the number of the characteristic parameters is m, the number of the samples is N, and the standardized processing method is specifically as follows:
Figure BDA0002033172540000021
Figure BDA0002033172540000022
Figure BDA0002033172540000023
z in ij Is x ij Normalizing the amount after treatment;
Figure BDA0002033172540000024
is x ij Average value of (2); s is(s) j Is x ij Is a variance of (c).
In addition, in the step S2, the association rule mining algorithm-Apriori is adopted to implement the step of selecting the characteristic parameters of the low-voltage distribution area:
1) Setting a minimum support s and a minimum confidence C 0 ；
2) The Apriori algorithm uses a set of candidates. Firstly, generating a candidate item set, wherein when the support degree of the candidate item set is greater than or equal to the minimum support degree, the candidate item set is a frequent item set;
3) Firstly, reading things in a database, and calculating the support degree of each item (candidate 1-item set), wherein the candidate 2-item set is generated by frequent 1-item set;
4) And scanning the database to obtain a candidate 2-item set. Then obtaining candidate 3-item sets through the obtained frequent 2-item set sets;
5) Repeatedly scanning the database, the method is the same as above until no new candidate item set is generated, namely, starting a loop from the 2-item set and generating a k-item frequent set from the frequent (k-1) -item set; the loop is ended when there is only one item set to loop into the k-item set.
Moreover, the step S3 is specifically as follows:
1) Initializing, determining the number k of categories and the initial clustering center point, and calculating the total contour coefficient S of the clustering result t Selecting an optimal k value;
for any sample point i, the calculation method is as follows:
Figure BDA0002033172540000031
wherein: q (i) is the average distance from point i to other points in the class to which it belongs; p (i) is the minimum value of the average distance from point i to all points in the non-belonging class;
the total profile coefficient of the clustering result is the average value of the profile coefficients of all sample points, and is calculated as follows:
Figure BDA0002033172540000032
2) Class division: calculating the distance between N samples and k initial center points according to the following formula, distributing the N samples to the nearest center point according to the distance to form k clusters
Figure BDA0002033172540000033
3) Solving a clustering center point: according to P of sample E The values are sequenced in an ascending order, samples are equally divided into k classes, a central sample of each class is selected as an initial clustering center of the class, and P is selected E Euclidean distance representing electrical characteristic parameter vector of sample and minimum electrical characteristic parameter vector
Figure BDA0002033172540000034
Wherein:
Figure BDA0002033172540000035
ω j the weight of the j-th electrical characteristic parameter is taken as 1;
4) Convergence criteria: judging whether to converge by using a formula
Figure BDA0002033172540000036
5) Repeating the step 2) and the step 3) until E reaches a limiting condition or the swing is small, which indicates that the algorithm tends to be stable and the clustering is finished.
The step S4 is specifically as follows:
1) Constructing a model;
2) A compact 2D FDFD method of four field components is employed to characterize the distribution line;
3) Introducing boundary conditions, and finally obtaining a characteristic value problem;
4) Solving a large sparse eigenvalue problem by adopting an implicit restarting Arnoldi method;
5) And extracting parameters of the line by adopting a 1D transmission line theory.
Moreover, the compact 2D FDFD equation in step 2) is given by the formula that only relates to the four lateral components of the electromagnetic field:
Figure BDA0002033172540000041
Figure BDA0002033172540000042
Figure BDA0002033172540000043
/>
Figure BDA0002033172540000044
epsilon in the formula r Is the relative dielectric constant. k (k) 0 Is the free space wave number. Δx and Δy are the FDFD unit sizes in the x and y directions, respectively. Kappa (kappa) a ，κ x And kappa (kappa) y Is defined as follows:
Figure BDA0002033172540000045
Figure BDA0002033172540000046
Figure BDA0002033172540000047
moreover, the step 3) is specifically: after introducing boundary conditions PMC and/or PEC into (1) - (4), the eigenvalue problem can be finally obtained:
[A]{x}＝λ{x} (13)
to facilitate implementation of the boundary conditions of the distribution line metal strip, (5) a slight modification to a new generalized eigenvalue problem:
[A]{x}＝λ[B]·{x} (14)
where λ is the characteristic value to be determined, λ=β/k 0 X represents an unknown eigenvector, i.e., x= { E x ,E y ,H x ,H y } T ，[B]Is an identity matrix [ A ]]Is a sparse matrix comprising the coefficients listed in (1) - (4), and there are at most seven non-zero entries in each row of the matrix.
Moreover, said step 4) solves the problem of generalized eigenvalues by implicitly restarting the Arnoldi method, which is implemented in a software package named ARPACK [15], which can be used in the public domain.
Moreover, the step 5) adopts a 1D transmission line theory to extract the parameters of the line, and adopts a one-dimensional telegraph equation to characterize the voltage and current changes along the line:
Figure BDA0002033172540000051
Figure BDA0002033172540000052
v (z, ω) and I (z, ω) are vectors of voltage and current in the formula. R (ω), L (ω), C (ω) and G (ω) are unknown matrices containing the frequency dependent resistance, inductance, capacitance and conductance per unit length, respectively;
for a lossless distribution line structure, (14) and (15) can be reduced to the following form by applying exp (-jβz) dependencies [9,14 ]:
βV(ω)＝ωL(ω)I(ω) (17)
βI(ω)＝ωC(ω)V(ω) (18)
wherein the eigenvectors of voltage (V (ω)) and current (I (ω)) can be easily derived from the corresponding eigenvectors solved in (11), ultimately obtaining frequency dependent inductances and capacitances:
Figure BDA0002033172540000053
Figure BDA0002033172540000054
in the formula beta i (i= 1,2N) is a propagation constant associated with the N fundamental mode of the transmission line structure at frequency ω; for N-conductor distribution lines in generalThe size of the matrices L (ω) and C (ω) are both nxn; r (omega) is obtained from the obtained L (omega) and C (omega) and from the eigenvectors of the voltage and current.
The utility model provides an extraction element of low pressure district characteristic parameter which characterized in that: comprises a platform region characteristic index system establishment module, a platform region characteristic parameter selection module, a calculation module and a parameter extraction module which are sequentially connected in sequence,
the station characteristic index system establishment module is used for establishing a station characteristic index system;
the station characteristic parameter selection module is used for selecting the characteristic parameters of the low-voltage distribution station by adopting an association rule mining algorithm-Apriori;
the computing module is used for clustering sample data by adopting an improved K-Means algorithm aiming at the selected low-voltage distribution area characteristic parameters;
the parameter extraction module is used for extracting line parameters of the distribution line by adopting a compact 2D FDFD algorithm.
The invention has the advantages and positive effects that:
1. according to the invention, the original data is classified by adopting an improved K-Means clustering algorithm according to the electrical characteristic parameters of the sample, so that two main defects in the traditional clustering are overcome: firstly, before clustering starts, the number k of categories needs to be given in advance; and secondly, the algorithm does not provide a selection principle of an initial clustering center.
2. The invention adopts a compact 2D FDFD extraction method to extract the parameters of the distribution line, and overcomes the defects of the 2D FDTD method: the propagation constants should be considered as input parameters and signal processing algorithms need to be used to extract the eigenfrequencies. Compared with other extraction algorithms, the algorithm has simpler programming and smaller memory occupation.
Drawings
FIG. 1 is a flow chart of extracting characteristic parameters of a low-voltage area provided by the invention;
FIG. 2 is a flow chart of the improved K-Means algorithm provided by the present invention;
FIG. 3 is a flow chart of a compact 2D FDFD algorithm provided by the present invention;
FIG. 4 is a modified model of a distribution line provided by the present invention;
fig. 5 shows the field components involved in equation (6) provided by the present invention.
Detailed Description
The invention will now be described in further detail by way of specific examples, which are given by way of illustration only and not by way of limitation, with reference to the accompanying drawings.
The invention is based on the collected electricity consumption data of a large number of users in the low-voltage area, so that the analysis is needed by a big data processing technology, and the big data processing is usually carried out by a cloud computing platform. The cloud computing of the computer is adopted, a large number of common computers are combined into a network system similar to a super computer, the system is meshed into a cloud, and the processing requirement of massive big data can be met. The terminal-intelligent ammeter facing the platform region collects a large amount of user electricity consumption data, not all electricity consumption data can be collected, and not all collected data have significance, so that data loss, data pollution and data omission can occur.
The embodiment adopts three layers of hierarchical clouds to process the acquired information:
the outer layer of the cloud platform is preprocessed through data fusion, and the fusion universality intelligent power grid is obtained. Each cloud platform is faced with massive data of a certain action, but the data may have abnormal and polluted dirty data, and of course, sometimes the abnormal data is true valuable data and needs to be reserved for further analysis. In this way, the cloud platform eliminates invalid data based on massive big data, simultaneously sorts the data, fuses sparse data based on local learning and model analysis, screens uncertain data, supplements incomplete data, and primarily mines input electric big data.
The middle layer of the cloud platform plays a role in connection, so that information sharing and privacy protection are realized. The middle layer contacts the outer layer and the inner layer of the cloud platform, so that output data of the outer layer is backed up and stored, and big data primary mining information is shared. In addition, based on privacy protection policies, the data is privacy marked for use by the inner layer reference.
The inner layer of the cloud platform serves as a big data mining platform, and the core of the cloud platform is calculation and data access. The method is a key of big data processing, and needs to integrate current data based on historical data, give out effective results and valuable conclusions of big data calculation processing, and complete data input of applications or provide the results to a next-stage cloud platform.
Referring to fig. 1, the method for extracting the characteristic parameters of the low-voltage area according to the embodiment includes the following steps:
s1: the building of the characteristic index system of the platform area specifically comprises the following steps:
(1) The electrical characteristic parameters are mainly studied from aspects of power supply, transmission, power distribution and the like. The factors influencing the line loss of the station area are many, and mainly comprise the aspects of distribution network structure, equipment state, management level and the like. Reflecting the power supply radius, load distribution level, power supply mode and the like of the distribution network structure; the state of the equipment is reflected by the performance of the transformer, the type of the line, the total length of the low-voltage line and the like. The management levels of the areas in different areas are greatly different, the management loss is unknown loss, and no exact physical factors exist. Screening out main electric characteristic parameters (characteristic parameters of the transformer area) related to the grid structure and load of the transformer area by considering the importance degree and the obtained difficulty degree of the index on the line loss of the transformer area, wherein the main electric characteristic parameters comprise a power supply radius X 1 (m) total length X of low-voltage line 2 (m) load factor X 3 (%) and its electric property and proportion X 4 (%) and line loss rate X 5 (%) and the like.
Meaning of characteristic parameters of the station area:
(1) radius of power supply X 1 (m)。
X 1 The distance between the farthest load point of the station area and the line of the distribution transformer is often used as an important parameter for judging whether the grid structure is reasonable.
(2) Total length X of low-voltage line 2 (m)。
X 2 Is the sum of all low-voltage line lengths in the bay.
(3) Load factor X 3 (％)。
X3 refers to the ratio of the power supply amount to the rated capacity of the distribution transformer, and this parameter reflects the average load level of the station area.
(4) Electrical properties and ratio X 4 (％)
The electricity property includes household electricity, industrial electricity, unusual electricity and other electricity. X is X 4 The ratio of the electricity consumption to the electricity supply for a specific electricity consumption property is the ratio of the electricity consumption to the electricity supply for a resident, and the ratio of the electricity consumption to the electricity supply for an industry is the ratio of the electricity consumption to the electricity supply for an industry.
(5) Line loss rate X 5 (％)
The line loss rate refers to the ratio of the loss electric quantity to the power supply quantity of the power grid, and correspondingly comprises a statistical line loss rate and a theoretical line loss rate.
(2) The invention only researches the line loss rate in the distribution line. The line loss rate refers to the ratio of the loss electric quantity to the power supply quantity of the power grid, and correspondingly comprises a statistical line loss rate and a theoretical line loss rate. The power loss of the power grid line refers to the electric energy consumed by each element in the power transmission and distribution process of the power grid in a given time. In the line loss management work, the line loss electric quantity can be classified into a statistical line loss electric quantity, a theoretical (technical) line loss electric quantity, and a management line loss electric quantity.
The line loss electricity quantity statistics refers to the difference between the electricity quantity and the electricity selling quantity counted by the electricity meter.
The theoretical (technical) line loss electric quantity refers to the power grid loss calculated by a theoretical line loss calculation method according to the operation parameters of the power distribution equipment and the load condition of the power grid. Including transformer windings and power losses in the distribution lines that are proportional to the square of the current; transformer core, capacitor and cable insulation dielectric losses, corona losses, etc. related to operating voltage.
The management line loss electric quantity is the difference value between the statistical line loss electric quantity and the theoretical line loss electric quantity. The method comprises the steps of inaccurate statistics caused by comprehensive errors, meter reading different times, missing meter reading and wrong meter reading and calculation of various electric meters; leakage caused by poor insulation of electrified equipment, no electricity consumption and electricity loss caused by electricity stealing.
(3) The invention mainly researches line loss on a distribution network line, namely resistance loss in the line, wherein the loss generated by the resistance is mainly generated by current flowing in the line due to transmission power (including active power and reactive power), and the loss is mainly caused by the current. Under the condition that the transmission power is unchanged, the higher the voltage level is, the smaller the current in the transmission line is, so that the generated loss is smaller, which is the reason that the loss can be reduced by high-voltage transmission, but the voltage in the distribution network is almost unchanged, so that the method for reducing the network loss by adopting the voltage rise is not applicable; but the power distribution can be optimized by obtaining the parameters of the line, so that the network loss is reduced as much as possible under the condition of ensuring stable power supply. If the active power required on the load side is certain, the larger the reactive power flowing on the line is, the more the loss is, so the reactive power required to be transmitted on the line is reduced as much as possible.
(4) Because the value ranges and units of the various characteristic parameters are different, in order to calculate without being influenced by dimension, the original data needs to be standardized. The number of the characteristic parameters is m, the number of the samples is N, and the normalization method is as follows:
Figure BDA0002033172540000081
Figure BDA0002033172540000082
Figure BDA0002033172540000083
z in ij ——x ij Normalizing the amount after treatment;
Figure BDA0002033172540000084
——x ij average value of (2); s is(s) j ——x ij Is a variance of (c).
S2: the method comprises the following steps of mining the characteristic parameter types of the power distribution network by adopting an association rule mining algorithm, namely Apriori, and adopting the Apriori algorithm:
1) Setting a minimum support s and a minimum confidence C 0 ；
2) The Apriori algorithm uses a set of candidates. Firstly, generating a candidate item set, wherein when the support degree of the candidate item set is greater than or equal to the minimum support degree, the candidate item set is a frequent item set;
3) Firstly, reading things in a database, and calculating the support degree of each item (candidate 1-item set), wherein the candidate 2-item set is generated by frequent 1-item set;
4) And scanning the database to obtain a candidate 2-item set. Then obtaining candidate 3-item sets through the obtained frequent 2-item set sets;
5) Repeatedly scanning the database, the method is the same as above until no new candidate item set is generated, namely, starting a loop from the 2-item set and generating a k-item frequent set from the frequent (k-1) -item set; the loop is ended when there is only one item set to loop into the k-item set.
S3: the improved K-Means algorithm clusters the sample data, and the specific steps are as follows:
s31: and carrying out an initialization process, and determining the number k of categories and an initial clustering center point. By calculating the total profile coefficient S of the clustering result t To select the optimal k value. The contour coefficient is an evaluation of good or bad clustering effect, and the larger the total contour coefficient of the clustering result is, the better the clustering effect is.
For any sample point i, the calculation method is as follows:
Figure BDA0002033172540000091
wherein: q (i) is the average distance from point i to other points in the class to which it belongs; p (i) is the minimum of the average distance of point i to all points in the non-belonging class.
The total profile coefficient of the clustering result is the average value of the profile coefficients of all sample points, and is calculated as follows:
Figure BDA0002033172540000092
s32: class division: and (3) calculating the distance between the N samples and k initial center points according to the formula, and distributing the N samples to the nearest center points according to the distance to form k clusters.
Figure BDA0002033172540000093
S33: solving a clustering center point: according to P of sample E And (3) carrying out ascending sorting on the values, equally dividing the samples into k classes, and selecting a central sample of each class as an initial clustering center of the class. P (P) E Euclidean distance representing electrical characteristic parameter vector of sample and minimum electrical characteristic parameter vector
Figure BDA0002033172540000101
Wherein:
Figure BDA0002033172540000102
ω j the j-th electrical characteristic parameter is weighted, here 1.
S34: convergence criteria: judging whether to converge by using a formula
Figure BDA0002033172540000103
S35: repeating the step 2) and the step 3) until E reaches a limiting condition or the swing is small, which indicates that the algorithm tends to be stable and the clustering is finished.
S4: the method adopts a compact 2D FDFD algorithm to extract the line parameters of the distribution line, and comprises the following specific steps:
s41: construction of a model
The distribution network has an accurate model and an improved model, the difference is that the distribution network has a pair of ground branches, the corona phenomenon represented by the pair of ground branches is used for researching a low-voltage station area, and the pair of ground branches is corresponding to a 380V system and can be ignored, so that the improved model is adopted, and the model is suitable for a shorter distribution overhead line. The parameters characterizing the low voltage distribution line thus include resistance and reactance. The model is shown in fig. 4:
s42: the transmission lines are characterized by a compact 2D FDFD method of four field components and their frequency dependent circuit parameters are extracted.
In this context, it is assumed that the three-phase loads of the distribution line are in a three-phase balanced state, which is a special case in the distribution network. A compact 2D FDFD method of four field components is used to characterize the transmission lines and extract their frequency dependent circuit parameters. The final compact 2D FDFD equation is given by the equation that relates only to the four lateral components of the electromagnetic field:
Figure BDA0002033172540000104
Figure BDA0002033172540000105
Figure BDA0002033172540000106
Figure BDA0002033172540000111
epsilon in the formula r Is the relative dielectric constant. k (k) 0 Is the free space wave number. Δx and Δy are the FDFD unit sizes in the x and y directions, respectively. Kappa (kappa) a ，κ x And kappa (kappa) y Is defined as follows:
Figure BDA0002033172540000112
Figure BDA0002033172540000113
Figure BDA0002033172540000114
fig. 5 shows the relative positions of the field components involved in (6).
S43: problem of generalized eigenvalue
After introducing boundary conditions (PMC and/or PEC) into (1) - (4), the eigenvalue problem can be finally obtained:
[A]{x}＝λ{x} (13)
to facilitate implementation of the boundary conditions of the distribution line metal strip, (5) a slight modification to a new generalized eigenvalue problem:
[A]{x}＝λ[B]·{x} (14)
where λ is the characteristic value to be determined, λ=β/k 0 . x represents an unknown eigenvector, i.e. x= { E x ,E y ,H x ,H y } T 。[B]Is an identity matrix. [ A ]]Is a sparse matrix comprising the coefficients listed in (1) - (4), and there are at most seven non-zero entries in each row of the matrix.
S44: solution of sparse eigenvalues
(14) The generalized eigenvalue problem in (2) can be solved by an implicit restart Arnoldi method, which is used for accurately and effectively solving the problem of large sparse eigenvalue. Since the Arnoldi method can find a small number of feature values of interest, CPU time can be saved. The Arnoldi method of implicit restart is implemented in a software package named ARPACK [15], which can be used in the public domain.
S45: extraction of characteristic parameters
The solution to the generalized eigenvalue problem in equation (14) produces eigenvalues and eigenvectors that correspond to the wave propagation constants and model field distribution in the distribution line. However, for circuit engineers, it is more preferable to characterize distribution lines in circuit simulations with circuit elements. Since three phases are assumed to be balanced, a 1D transmission line theory may be employed for extracting circuit parameters in the distribution network.
In the special case of three-phase balancing, the distribution lines are uniform, so it is assumed that they are uniform along the z-axis and that the electromagnetic waves in them travel in the positive z-direction. For the structure of the distribution line in this particular case and under the assumption of quasi-TEM mode propagation, the one-dimensional telegraph equation can characterize the voltage and current variations therealong:
Figure BDA0002033172540000121
Figure BDA0002033172540000122
v (z, ω) and I (z, ω) are vectors of voltage and current in the formula. R (ω), L (ω), C (ω) and G (ω) are unknown matrices containing the frequency dependent resistance, inductance, capacitance and conductance per unit length, respectively.
For a lossless distribution line structure, (14) and (15) can be reduced to the following form by applying exp (-jβz) dependencies [9,14 ]:
βV(ω)＝ωL(ω)I(ω) (17)
βI(ω)＝ωC(ω)V(ω) (18)
wherein the voltage (V (ω)) and current (I (ω)) eigenvectors can be easily derived from the corresponding eigenvectors solved in (11). The frequency dependent inductance and capacitance can now be finally obtained:
Figure BDA0002033172540000123
Figure BDA0002033172540000124
in the formula beta i (i= 1,2N) is the propagation constant associated with the N fundamental mode of the transmission line structure at frequency ω. For a general N-conductor distribution line structure, the matrices L (ω) and C (ω) are scaledThe dimensions are N. R (omega) can be obtained from the obtained L (omega) and C (omega) and from the eigenvectors of the voltage and current.
The utility model provides an extraction element of low pressure district characteristic parameter which characterized in that: comprises a platform region characteristic index system establishment module, a platform region characteristic parameter selection module, a calculation module and a parameter extraction module which are sequentially connected in sequence,
the station characteristic index system establishment module is used for establishing a station characteristic index system;
the station characteristic parameter selection module is used for selecting the characteristic parameters of the low-voltage distribution station by adopting an association rule mining algorithm-Apriori;
the computing module is used for clustering sample data by adopting an improved K-Means algorithm aiming at the selected low-voltage distribution area characteristic parameters;
the parameter extraction module is used for extracting line parameters of the distribution line by adopting a compact 2D FDFD algorithm.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Although the embodiments of the present invention and the accompanying drawings have been disclosed for illustrative purposes, those skilled in the art will appreciate that: various substitutions, changes and modifications are possible without departing from the spirit and scope of the invention and the appended claims, and therefore the scope of the invention is not limited to the embodiments and the disclosure of the drawings.

Claims (3)

1. The extraction method of the characteristic parameters of the low-voltage station area is characterized by comprising the following steps of: the method comprises the following steps:
s1: the establishing of the characteristic index system of the platform area comprises the following steps:
(1) Screening out electric characteristic parameters related to the grid structure and the load of the transformer area: comprising a power supply radius X 1 Total length X of low-voltage line 2 Load factor X 3 Electric properties and ratio X 4 Line loss rate X 5 ；
(2) Carrying out standardization processing on the original data;
the standardized processing method comprises the following steps: the number of the characteristic parameters is m, the number of the samples is N, and the normalization method is as follows:
Figure FDA0004003618190000011
Figure FDA0004003618190000012
Figure FDA0004003618190000013
z in ij ——x ij Normalizing the amount after treatment;
Figure FDA0004003618190000014
——x ij average value of (2); s is(s) j ——x ij Is a variance of (2);
s2: adopting an association rule mining algorithm-Apriori to realize the selection of characteristic parameters of a low-voltage distribution area;
s3: clustering sample data by adopting an improved K-Means algorithm aiming at the selected low-voltage distribution area characteristic parameters; the method comprises the following specific steps:
1) Initializing, determining the number k of categories and the initial clustering center point, and calculating the total contour coefficient S of the clustering result t Selecting an optimal k value;
for any sample point i, the calculation method is as follows:
Figure FDA0004003618190000015
wherein: q (i) is the average distance from point i to other points in the class to which it belongs; p (i) is the minimum value of the average distance from point i to all points in the non-belonging class;
the total profile coefficient of the clustering result is the average value of the profile coefficients of all sample points, and is calculated as follows:
Figure FDA0004003618190000016
2) Class division: calculating the distance between N samples and k initial center points according to the following formula, distributing the N samples to the nearest center point according to the distance to form k clusters
Figure FDA0004003618190000017
3) Solving a clustering center point: according to P of sample E The values are sequenced in an ascending order, samples are equally divided into k classes, a central sample of each class is selected as an initial clustering center of the class, and P is selected E Euclidean distance representing electrical characteristic parameter vector of sample and minimum electrical characteristic parameter vector
Figure FDA0004003618190000021
Wherein:
Figure FDA0004003618190000022
ω j the weight of the j-th electrical characteristic parameter is taken as 1; />
4) Convergence criteria: judging whether to converge by using a formula
Figure FDA0004003618190000023
5) Repeating the step 2) and the step 3) until E reaches a limiting condition or the swing is small, which indicates that the algorithm tends to be stable and the clustering is finished;
s4: the method adopts a compact 2D FDFD algorithm to extract the line parameters of the distribution line, and comprises the following specific steps:
1) Constructing a model;
2) A compact 2D FDFD method of four field components is employed to characterize the distribution line; the compact 2D FDFD equation is given by the equation that relates only to the four lateral components of the electromagnetic field:
Figure FDA0004003618190000024
Figure FDA0004003618190000025
Figure FDA0004003618190000026
Figure FDA0004003618190000027
epsilon in the formula r Is the relative dielectric constant, k 0 Is the free space wavenumber, Δx and Δy are the FDFD unit sizes in the x and y directions, respectively, κ a ，κ x And kappa (kappa) y Is defined as follows:
Figure FDA0004003618190000031
Figure FDA0004003618190000032
Figure FDA0004003618190000033
3) Introducing boundary conditions, the problem of characteristic values can be finally obtained, specifically: after introducing the boundary conditions PMC and/or PEC into equations (1) - (4), the eigenvalue problem can be finally obtained:
[A]{x}＝λ{x} (13)
to facilitate implementation of the boundary conditions for the distribution line metal strip, equation (5) is modified to a new generalized eigenvalue problem:
[A]{x}＝λ[B]·{x} (14)
where λ is the characteristic value to be determined, λ=β/k 0 X represents an unknown eigenvector, i.e., x= { E x ,E y ,H x ,H y } T ，[B]Is an identity matrix [ A ]]Is a sparse matrix comprising coefficients listed in formulas (1) - (4), and there are at most seven non-zero terms in each row of the matrix;
4) Solving a large sparse eigenvalue problem by adopting an implicit restarting Arnoldi method, wherein the method comprises the following steps of: solving the generalized eigenvalue problem by implicit restarting Arnoldi method, which is implemented in a software package named ARPACK that can be used in the public domain
5) And extracting parameters of the line by adopting a 1D transmission line theory.
2. The method for extracting the characteristic parameters of the low-voltage area according to claim 1, wherein the method comprises the following steps: step S2, namely mining the characteristic parameter types of the power distribution network by adopting an Apriori algorithm:
1) Setting a minimum support s and a minimum confidence C 0 ；
2) The Apriori algorithm uses a set of candidates. Firstly, generating a candidate item set, wherein when the support degree of the candidate item set is greater than or equal to the minimum support degree, the candidate item set is a frequent item set;
3) Firstly, reading things in a database, and calculating the support degree of each candidate 1-item set, wherein a candidate 2-item set is generated by frequent 1-item sets;
4) Scanning a database to obtain a candidate 2-item set, and obtaining a candidate 3-item set through the obtained frequent 2-item set;
5) Repeatedly scanning the database, the method is the same as above until no new candidate item set is generated, namely, starting a loop from the 2-item set and generating a k-item frequent set from the frequent (k-1) -item set; the loop is ended when there is only one item set to loop into the k-item set.
3. The method for extracting the characteristic parameters of the low-voltage area according to claim 1, wherein the method comprises the following steps: the step 5) adopts a 1D transmission line theory to extract the parameters of the line, and adopts a one-dimensional telegraph equation to represent the voltage and current changes along the line:
Figure FDA0004003618190000041
Figure FDA0004003618190000042
v (z, ω) and I (z, ω) are vectors of voltage and current in the formula. R (ω), L (ω), C (ω) and G (ω) are unknown matrices containing the frequency dependent resistance, inductance, capacitance and conductance per unit length, respectively;
for a lossless distribution line structure, (14) and (15) can be reduced to the following form by applying exp (-jβz) dependencies [9,14 ]:
βV(ω)＝ωL(ω)I(ω) (17)
βI(ω)＝ωC(ω)V(ω) (18)
wherein the eigenvectors of voltage (V (ω)) and current (I (ω)) can be easily derived from the corresponding eigenvectors solved in (11), ultimately obtaining frequency dependent inductances and capacitances:
Figure FDA0004003618190000043
Figure FDA0004003618190000044
in the formula beta i (i=1, 2 … N) is the propagation constant associated with the N fundamental mode of the transmission line structure at frequency ω; for general purposesThe dimensions of the matrices L (ω) and C (ω) are n×n; r (omega) is obtained from the obtained L (omega) and C (omega) and from the eigenvectors of the voltage and current.
CN201910316214.2A 2019-04-19 2019-04-19 Method for extracting characteristic parameters of low-voltage transformer area Active CN110110907B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910316214.2A CN110110907B (en) 2019-04-19 2019-04-19 Method for extracting characteristic parameters of low-voltage transformer area

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910316214.2A CN110110907B (en) 2019-04-19 2019-04-19 Method for extracting characteristic parameters of low-voltage transformer area

Publications (2)

Publication Number Publication Date
CN110110907A CN110110907A (en) 2019-08-09
CN110110907B true CN110110907B (en) 2023-05-30

Family

ID=67485770

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910316214.2A Active CN110110907B (en) 2019-04-19 2019-04-19 Method for extracting characteristic parameters of low-voltage transformer area

Country Status (1)

Country Link
CN (1) CN110110907B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111091251A (en) * 2020-03-24 2020-05-01 汉谷云智(武汉)科技有限公司 Boiler operation optimization method and system based on big data technology
CN112801333B (en) * 2020-11-24 2023-12-08 国网福建省电力有限公司宁德供电公司 XGBoost-based power distribution network line summer peak load prediction method
CN113010917B (en) * 2021-03-16 2023-05-05 国网福建省电力有限公司 Loss reduction analysis processing method with privacy protection for contemporaneous line loss management system
CN116522181B (en) * 2023-06-28 2024-03-29 广东电网有限责任公司 Method and system for extracting daily line loss characteristics of distribution area
CN116680546A (en) * 2023-08-02 2023-09-01 北京前景无忧电子科技股份有限公司 Low-voltage distribution network station line loss rate identification method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106126577A (en) * 2016-06-17 2016-11-16 北京理工大学 A kind of weighted association rules method for digging based on data source Matrix dividing
CN106156792A (en) * 2016-06-24 2016-11-23 中国电力科学研究院 A kind of low-voltage platform area clustering method based on platform district electric characteristic parameter
CN109272176A (en) * 2018-12-10 2019-01-25 贵州电网有限责任公司 Calculation method is predicted to platform area line loss per unit using K-means clustering algorithm

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106126577A (en) * 2016-06-17 2016-11-16 北京理工大学 A kind of weighted association rules method for digging based on data source Matrix dividing
CN106156792A (en) * 2016-06-24 2016-11-23 中国电力科学研究院 A kind of low-voltage platform area clustering method based on platform district electric characteristic parameter
CN109272176A (en) * 2018-12-10 2019-01-25 贵州电网有限责任公司 Calculation method is predicted to platform area line loss per unit using K-means clustering algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"Efficient 2-D Finite-Difference Frequency Domain Method for Switching Noise Analysis in Multilayer Boards";Babak Alavikiad等;《IEEE TRANSACTIONS ON COMPONENTS》;20130531;第3卷(第5期);全文 *
"一种配电网理论线损实时计算分析系统";李云薪等;《电力需求侧管理》;20151130;第17卷(第6期);全文 *

Also Published As

Publication number Publication date
CN110110907A (en) 2019-08-09

Similar Documents

Publication Publication Date Title
CN110110907B (en) Method for extracting characteristic parameters of low-voltage transformer area
Zang et al. Adaptive method for harmonic contribution assessment based on hierarchical K‐means clustering and Bayesian partial least squares regression
CN106372747B (en) Random forest-based reasonable line loss rate estimation method for transformer area
CN111428355B (en) Modeling method for intelligent comprehensive statistics of power load numbers
CN101232180A (en) Power distribution system load obscurity model building device and method
CN112510707B (en) Power distribution station power impedance topological graph generation method and system
CN109286188B (en) 10kV power distribution network theoretical line loss calculation method based on multi-source data set
CN102982394A (en) Method and system for distribution network load parameter identification
CN108490288A (en) A kind of stealing detection method and system
CN112149873A (en) Low-voltage transformer area line loss reasonable interval prediction method based on deep learning
CN108268999B (en) Multi-principal-element modal analysis method for harmonic characteristic analysis of power network nodes
CN105406461B (en) A kind of load self-adapting dynamic monitoring method of power distribution network power-off event
CN103793556B (en) Comprehensively go out transformer station's load modeling method of specific electric load
CN113189418B (en) Topological relation identification method based on voltage data
CN110659693A (en) K-nearest neighbor classification-based rapid topology identification method and system for power distribution network and readable storage medium
CN106208043B (en) A kind of low and medium voltage distribution network key energy consumption link discrimination method
Han et al. Multi‐objective robust dynamic VAR planning in power transmission girds for improving short‐term voltage stability under uncertainties
CN104376207A (en) Power distribution network alternating current transmission loss computing and parameter estimation method
Montano-Martinez et al. Detailed primary and secondary distribution system model enhancement using AMI data
Lu et al. Adaptive weighted fuzzy clustering algorithm for load profiling of smart grid customers
CN104952002A (en) Urban low-voltage partitioning method based on spectral clustering algorithm
Grigoras et al. Processing of smart meters data for peak load estimation of consumers
CN107832959B (en) Voltage stability evaluation method considering load characteristics and power supply constraints
CN103972889B (en) A kind of distribution line impedance on-line identification method
Lin et al. Voltage sag severity analysis based on improved FP-Growth algorithm and AHP algorithm

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