CN114386700A - Power transmission system planning method and device considering wind-light correlation and storage medium - Google Patents

Power transmission system planning method and device considering wind-light correlation and storage medium Download PDF

Info

Publication number
CN114386700A
CN114386700A CN202210039670.9A CN202210039670A CN114386700A CN 114386700 A CN114386700 A CN 114386700A CN 202210039670 A CN202210039670 A CN 202210039670A CN 114386700 A CN114386700 A CN 114386700A
Authority
CN
China
Prior art keywords
rattan
wind
scene
copula
power
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210039670.9A
Other languages
Chinese (zh)
Inventor
彭穗
刘若平
余浩
左郑敏
李作红
龚贤夫
陈鸿琳
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.)
Guangdong Power Grid Co Ltd
Original Assignee
Guangdong Power Grid 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 Guangdong Power Grid Co Ltd filed Critical Guangdong Power Grid Co Ltd
Priority to CN202210039670.9A priority Critical patent/CN114386700A/en
Publication of CN114386700A publication Critical patent/CN114386700A/en
Pending legal-status Critical Current

Links

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
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • 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
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06313—Resource planning in a project environment
    • 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
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/381—Dispersed generators
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H02J3/466—Scheduling the operation of the generators, e.g. connecting or disconnecting generators to meet a given demand
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]

Abstract

The application discloses a power transmission system planning method, a device and a storage medium considering wind-solar correlation, wherein the method comprises the following steps: acquiring historical data sets of power system parameters and random variables; carrying out scene division on the historical data set according to a Mini Batch K-Means clustering method; determining parameters and types of Copula functions according to the data of each scene, and judging the optimal rattan structure in each scene by adopting the AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure in each scene to generate a wind-solar-lotus sample set; and inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model, and outputting a power grid planning result. The method and the device adopt a Mini Batch K-Means algorithm to divide the scenes, and reduce the operation time; and the AD distance is used as a rattan structure judgment standard, the modeling precision of the hybrid rattan Copula is improved, and the reliability of the power transmission system extension planning method is further improved.

Description

Power transmission system planning method and device considering wind-light correlation and storage medium
Technical Field
The present application relates to the field of power transmission system technologies, and in particular, to a method, an apparatus, a device, and a storage medium for planning a power transmission system in consideration of wind-light correlation.
Background
In recent years, with the increasing annual energy demand and the increasing environmental problems, the installed capacity of renewable energy sources has been increasing year by year. As is well known, wind power and photovoltaic output have strong randomness and intermittency, and these characteristics will bring a serious challenge to the expansion planning of the power system. Meanwhile, loads widely existing in the power system are influenced by differences and irregularities of power consumer behaviors, and strong uncertainty also exists. On the other hand, the randomness among wind power, photovoltaic output and load is not completely independent, but complex correlation exists between every two wind power, photovoltaic output and load. In order to deeply reveal the influence of the correlation among wind power, photovoltaic output and load (referred to as 'wind-solar load') on power grid planning, a power transmission system extension planning method considering the complex correlation of 'wind-solar load' needs to be researched.
A large-scale wind-solar complementary power grid expansion planning [ J ] power grid technology considering correlation, 2018,42(07) & 2120-2127 & gt establishes a wind-solar complementary output Copula model considering correlation of output between a wind power plant and a photovoltaic power station, and provides a wind-solar complementary multi-target power grid expansion planning model on the basis. However, this paper has two disadvantages: 1) only the correlation between wind speed and illumination intensity is considered. In an actual power system, wind speed, light irradiation and load have certain correlation. Considering only the correlation between wind speed and illumination will result in inaccurate planning results. 2) Only a single Copula function, i.e., Frank Copula function, is employed. The correlation between the "wind-solar-load" is different, and the dependency structure (correlation type) between the random variables described by each Copula function is different. Therefore, the correlation between different variables cannot be accurately described using only a single Copula function.
A wind-solar combined power generation correlation modeling based on a mixed rattan Copula model and application thereof in reactive power optimization [ J ] a power grid technology, 2017,41(03):791 and 798 ] proposes a mixed rattan Copula model built by combining a K-means clustering and rattan structure principle, and the mixed rattan Copula model is used for analyzing the correlation of wind speed, illumination and load combined output in a key point manner. However, there are three disadvantages to this paper:
1) the hybrid rattan Copula model is applied to the field of reactive power optimization of a power grid and is never applied to the field of extended planning of a power transmission network, and the technical difficulty of combination of the two technologies is unknown.
2) The data of the input random variable are clustered and divided by adopting a K-means clustering algorithm, and the method has higher calculation speed in a scene with smaller sample size. However, when the historical data of the random variables is huge, the calculation is long in time consumption and low in efficiency, the modeling precision based on the hybrid rattan Copula is seriously influenced, and the efficiency of the power transmission system expansion planning is further influenced.
3) The rattan structural rating index CvM distance measure used is the integral of the squared distance of the empirical cumulative distribution function and the target cumulative distribution function. If the data mostly fall on the tail positions of the left side and the right side, the difference of the tail parts is more important to be considered, and the total distribution deviation obtained by CvM distance does not consider the difference caused by the distribution position of large-scale data although all difference points are considered.
Disclosure of Invention
The application provides a power transmission system planning method, a device, equipment and a storage medium considering wind-light correlation, so as to solve the problems of unreliable and inaccurate power transmission system extension planning results in the prior art.
In order to solve the above technical problem, the present application provides a power transmission system planning method considering wind and light correlation, including: acquiring historical data sets of power system parameters and random variables; performing cluster classification and scene division on the historical data set according to a Mini Batch K-Means clustering method; respectively determining parameters and types of Copula functions according to each scene data after scene division, generating a sample corresponding to each scene data by using a rattan structure, and judging an optimal rattan structure under each scene by adopting an AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model; and inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result.
Optionally, the random variables include wind speed, light and load, and the clustering classification and scene division are performed on the historical data set according to a Mini Batch K-Means clustering method, including: inputting a historical data set D of continuous random variables of the power system about wind speed, illumination and load, clustering parameters K and sampling data N, and outputting 9 clustering schemes by using a Mini Batch K-Means clustering method; calculating to obtain DBI index value corresponding to each K value, and outputting the optimal clustering number KbestAnd its corresponding scene division set C ═ (C)1,C2,…,CK)。
Optionally, the method further comprises: input scene division set C ═ (C)1,C2,…,CK) Respectively fitting the probability density function of the multidimensional variable of each scene by a nonparametric kernel density estimation method to obtain an edge distribution function F (x)i) (ii) a According to the edge distribution function F (x)i) Selecting the optimal parameters of each alternative Copula function by using the Copula function carried by the MATLAB, and calculating an alternative Copula function CpAnd empirical Copula function CnThe Euclidean distance between the cumulative distribution functions is calculated, and each scene outputs three optimal Copula functions and parameters thereof; setting the total number O of samples, and generating an independent variable set L of each scene by using an LHS methodiWherein i ═ 1,2, …, K; set of independent variables LiInputting a rattan Copula model constructed in each scene, respectively sampling, and outputting a sampling sample Z; and according to the sampling sample Z, calculating the AD distance in each scene, and selecting a rattan Copula model with a small AD distance as a mixed rattan Copula model of the scene.
Optionally, the method further comprises: and obtaining a sample data set U according to the mixed rattan Copula model, obtaining an edge distribution function through a nonparametric kernel density method, and converting the sample data set U into a required wind-solar-load sample set X according to an inverse function of estimated edge distribution.
Optionally, set L of independent variablesiInputting the rattan Copula model built by each scene and sampling respectively, and outputting a sampling sample Z, wherein the method comprises the following steps: set of independent variables LiRandomly generating an n-dimensional sample set which is subjected to independent uniform distribution; according to the structural column writing equations of the C rattan and the D rattan, the first dimension sample points are uniformly distributed sampling points, sampling points corresponding to the second dimension are obtained by utilizing the property of condition distribution, the sample points of the first dimension are sequentially obtained one by one according to the previous dimension sample points until all the sample points of the n dimensions are obtained, and finally a sampling sample Z is generated.
Optionally, before inputting the wind-solar-load sample set and the power system parameters into the power transmission system planning model, the method further includes: establishing a power transmission system planning model according to the objective function and the constraint condition; the target function comprises a power grid construction cost value, a wind abandoning light abandoning economic loss value and the discharge capacity of a conventional power grid unit, and the constraint conditions comprise power balance constraint, load node new energy power generation penetration power constraint, branch power flow constraint, conventional power generator unit output upper and lower limit constraint, wind turbine unit operation condition constraint, photovoltaic unit operation condition constraint and equipment operation state 0-1 constraint.
Optionally, inputting the wind-solar-load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result, including: under the constraint condition, when the minimization is realized by the 3 objective functions, the optimal power grid planning result is output.
In order to solve the above technical problem, the present application provides a power transmission system planning apparatus considering wind and light correlation, including: the data acquisition module is used for acquiring historical data sets of the parameters and the random variables of the power system; the clustering module is used for carrying out clustering classification and scene division on the historical data set according to a Mini Batch K-Means clustering method; the mixed rattan Copula model module is used for respectively determining parameters and types of Copula functions according to each scene data after scene division, generating samples corresponding to each scene data by using rattan structures, and judging the optimal rattan structure in each scene by adopting AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model; (ii) a And the planning result module is used for inputting the wind-solar load sample set and the power system parameters into the power transmission system planning model and finally outputting a power grid planning result.
In order to solve the above technical problem, the present application provides an electronic device, which includes a memory and a processor, where the memory is connected to the processor, and the memory stores a computer program, and the computer program is executed by the processor to implement the above power transmission system planning method considering wind-light correlation.
In order to solve the above technical problem, the present application provides a computer-readable storage medium storing a computer program, which when executed implements the above power transmission system planning method considering wind-light correlation.
The application provides a power transmission system planning method, a device, equipment and a storage medium considering wind-solar correlation, wherein the method comprises the following steps: acquiring historical data sets of power system parameters and random variables; performing cluster classification and scene division on the historical data set according to a Mini Batch K-Means clustering method; respectively determining parameters and types of Copula functions according to each scene data after scene division, generating a sample corresponding to each scene data by using a rattan structure, and judging an optimal rattan structure under each scene by adopting an AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model; and inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result. By the mode, scene division is carried out on the multi-dimensional data by utilizing a Mini Batch K-Means algorithm, so that the operation time is greatly reduced; the AD distance is used as a judgment standard of the rattan structure, the modeling precision of the hybrid rattan Copula is improved, and the reliability of the power transmission system extension planning method is further improved.
Drawings
In order to more clearly illustrate the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a schematic flow diagram of an embodiment of a method for power transmission system planning that takes into account wind-solar correlations according to the present application;
FIG. 2 is a logical schematic of the rattan structure of the present application C;
FIG. 3 is a schematic logic diagram of a D vine structure of the present application;
FIG. 4 is a schematic block diagram of an embodiment of an improved IEEE-6 node system of the present application;
FIG. 5 is a schematic diagram of an embodiment of a power transmission system planning apparatus of the present application that considers wind-solar correlation;
FIG. 6 is a schematic structural diagram of an embodiment of an electronic device of the present application;
FIG. 7 is a schematic structural diagram of an embodiment of a computer-readable storage medium of the present application.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present application, the method, apparatus, device and storage medium for planning a power transmission system considering wind-solar correlation provided in the present application are described in further detail below with reference to the accompanying drawings and the detailed description.
Referring to fig. 1, fig. 1 is a schematic flow chart of an embodiment of a power transmission system planning method considering wind-light correlation in the present application, and in this embodiment, the power transmission system planning method considering wind-light correlation may include steps S110 to S140, where each step is specifically as follows:
s110: historical data sets of power system parameters and random variables are obtained.
Collecting power system parameters and a random variable D ═ x1,x2,...,xn) Wherein the random variables may include continuous random variables such as wind speed, light, and load.
S120: and performing cluster classification and scene division on the historical data set according to a Mini Batch K-Means clustering method.
When the correlation of the wind load and the solar load is considered, the analysis on the details of the data can enable the analysis structure to be more accurate due to the large amount of historical data. From this perspective, in this embodiment, the cluster division method is used to subdivide the historical data set to be analyzed into several categories, and then perform correlation analysis on each category of data to be analyzed. Wherein, the partitioning method selects a Mini Batch K-Means clustering method.
In the conventional K-Means algorithm, the distances of all sample points to all centroids are calculated. If the sample size is very large, for example, more than 10 ten thousand, and the feature is more than 100, the conventional K-Means algorithm is very time-consuming.
The main idea of Mini Batch K-Means is: and selecting part of representative samples in the massive sample points to perform K-Means clustering, so that the algorithm convergence time can be effectively reduced, and the clustering calculation efficiency is greatly improved. Especially when the 'wind-solar-load' correlation is considered, the historical data is multidimensional random variables. In such a scenario, the operation time is greatly reduced by using Mini Batch K-Means.
The Mini Batch K-Means algorithm has a degree of accuracy reduction compared to the traditional K-Means algorithm, but this reduction is within an acceptable range. Preferably, in order to improve the accuracy based on the Mini Batch K-Means algorithm, the embodiment may repeat the sampling for a plurality of times, and then calculate the mean value of the class coordinates obtained until the class center tends to be stable.
The number of divided scenes can influence the separation of different correlations of multidimensional variables, so that when historical data of random variables such as 'wind, light and load' in a power transmission system are clustered, the influence of cluster number change on data similarity classification needs to be analyzed, and the most appropriate classification number is selected.
The embodiment may determine the optimal cluster number by a density based cluster validity index (DBI). The DBI index can solve the defect that the clustering number of the DBI index needs to be specified according to an empirical value and the optimal clustering number which best accords with the clustering characteristic of the data is difficult to obtain, and the value is the optimal clustering number with the maximum value. For a historical data set D ═ x to be classified1,x2,...,xnAnd if each sample point is in the dimension P and the initially set clustering number is K, calculating the DBI index by the following steps:
1) calculating a radius Ri，RiIs represented as each data set CiTo class center ciAverage value of distance:
d(x,ci)＝|x-ci|……(1)
Figure BDA0003469387700000051
in the formula, CiRepresents the set of clusters output by the Mini Batch K-Means algorithm, where i ═ 1,2iIs of class CiCentral point of (1), N (C)i) Represents class CiThe number of history data contained in (1), d (x, c)i) Indicating the distance between samples.
2) Find the mid-class midpoint cij，cijRepresented as two class centers ciAnd cjA point on the connecting line of (1), the point having a radius of similarity R of the twoiAnd RjThe ratio of the center lines is equally divided:
Figure BDA0003469387700000052
in the formula (I), the compound is shown in the specification,
Figure BDA0003469387700000053
and
Figure BDA0003469387700000054
respectively represent ciAnd cjWherein n ═ R (1, 2.., P), λ ═ Ri/Rj。
3) Calculating the class center density Dinside()，Dinside() Is represented as class CiAll to class center point ciIs less than the radius of class RiThe number of data of (2):
Figure BDA0003469387700000063
4) calculating class edge density Dedge()，Dedge() Is represented as class CiAnd any class C in the data setjAll to two intermediate points cijDistance less than R ═ R (R)i+Rj) Average of the number of data samples of/2:
d(x,cij)＝|x-cij|……(5)
Figure BDA0003469387700000061
in the formula, K is the number of clusters.
5) Calculating DBI index value VDBI(i)，VDBI(i) Expressed as class center density Dinside(i) Sum and edge-like density Dedge() The ratio of the sums, which defines the formula:
Figure BDA0003469387700000062
in the formula, VDBI(i)，i＝(1,2,...,K)。
Kbest＝argmax{VDBI(i),i＝1,2…,K}……(8)
The best effect of considering clustering is that the tighter the data within a class is, the better the data between classes is, the better the dispersion is, so MThe best clustering number of ini Batch K-means is the maximum value V of the DBI indexDBI(i) Corresponding K value Kbest。
In some embodiments, the determining step of the scene division mode of the "wind-solar-load" correlation model is as follows:
step 1: given sample set D ═ x1,x2,...,xn) And n is the sample volume. Setting the clustering parameter K to be 2-10 types, namely 2 types, 3 types and 10 types, and totally 9 schemes;
it should be noted that the clustering parameter K may be selected from a suitable range according to actual situations, and in other embodiments, may also be set to 2-15, 4-12, and the like.
Step 2: selecting K data objects in D as initial clustering centers ci，i＝(1,2,...,K)；
And step 3: randomly extracting a sample subset S with the capacity of b from the D (S ═ S)1,…,sj,…,sb) To form a Batch;
and 4, step 4: for each sample point S in SjCalculating the similarity between the sample points and K clustering centers, and sampling the sample points sjDividing the class with the maximum similarity;
and 5: after all samples in the S pass through the step4, recalculating the clustering centers according to the class labels of the samples;
step 6: judging whether a clustering finishing condition is met (if the iteration times t are met), returning to the step3 if the clustering finishing condition is not met, or entering the step 7;
and 7: for each sample point in the D, dividing each sample point into a class with the maximum similarity according to the similarity of the sample point and the K clustering centers to obtain 9 clustering schemes;
and 8: determining the optimal clustering number K of the multidimensional data by comparing the DBI indexes corresponding to the results obtained by the clustering schemesbestAnd outputting the corresponding scene division result.
The main calculation amount of the Mini Batch K-Means algorithm is concentrated in the 4 th step, namely, the similarity of each sample point is calculated and the attribution is determined. According to the Mini Batch K-Means cluster analysis result, a proportion function Q (i) is defined in each scene, and is expressed as formula (9):
Figure BDA0003469387700000071
in the formula, N (C)i) Represents class CiThe number of data contained in (1), and N is the total number of data.
S130: and respectively determining parameters and types of Copula functions according to each scene data after scene division, generating a sample corresponding to each scene data by using the rattan structure, and judging the optimal rattan structure in each scene by adopting the AD distance.
S140: and establishing a mixed rattan Copula model according to the optimal rattan structure in each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model.
For an analysis method of random variable correlation, at present, main technical means include a spatial variation method and a Copula function method, wherein a spatial transformation method can only research one characteristic of the correlation, and the usage is limited, so that the Copula function method is used in the embodiment when historical data of wind speed, illumination and load are subjected to correlation modeling. Copula was derived from the Sklar theorem proposed by Sklar in 1959, which is defined as follows: let H be a joint distribution function, and F be an edge-cumulative distribution function1And F2Then there is a Copula, so that:
H(x1,x2)＝C(F1(x1),F2(x2))……(10)
if F1And F2If the operation is continuous, the Copula function C is uniquely determined; otherwise at F1And F2Is unique within the range of values of (A). Conversely, if C is a Copula function, F1And F2If it is an edge cumulative distribution function, the function H calculated by the formula (9) is a joint distribution function, and the edge cumulative distribution function is F1And F2. The Copula function defines the union distribution function of the multidimensional random variable and the connection function of the respective edge distribution functions, i.e. "join (Couple)" the union scoreA distribution function and a respective edge distribution function.
The class of Copula functions can be classified into elliptic Copula and archimedean elliptic Copula including Gaussian Copula, t-Copula, etc., and archimedean including Gumbel Copula, Frank Copula, and Clayton Copula, etc. Different Copula functions can characterize different dependency structures and have different advantages.
The probability distribution function of 5 Copula functions of Gaussian Copula, t-Copula and Gumbel Copula, Clayton Copula, Frank Copula is shown in table 1 as candidate functions:
TABLE 1 alternative Copula function
Figure BDA0003469387700000081
And each Copula function has a corresponding parameter to describe the correlation degree of the random variable. Therefore, the optimal parameters of the Copula function are required to be obtained by analyzing the correlation of the wind-light load. Before that, each edge distribution function of the multidimensional variables needs to be obtained, the non-parameter kernel density estimation is selected in the embodiment, the method does not need to predict the distribution characteristics of the data to be analyzed, the edge distribution function can be solved according to the characteristics of the data, and the influence of the parameter estimation method caused by improper selection of the distribution function is avoided.
If the theoretical probability density of the variable X is set to be fh(x) Then, the calculation mode of the kernel density estimation value is as the following formula (11):
Figure BDA0003469387700000082
where n is the number of samples, h is the kernel window width, and K (.) is the selected kernel function.
In this embodiment, a gaussian kernel is used, and the optimal window width thereof is selected with reference to empirical parameters, and the specific selection mode is as follows:
Figure BDA0003469387700000091
obtaining the edge distribution function F (x) of the variable to be analyzed according to the stepsi) Then, the corresponding optimal parameters of each alternative Copula function can be selected according to a maximum likelihood method on the basis, and the specific implementation of the step can be completed by using a MATLAB with Copula function.
The nature of the random variables described by each alternative Copula function is different. When the correlation between the "wind load and the solar load" is analyzed, the most suitable three Copula functions are selected for each scene in a Copula function library (refer to table 1) to be fitted.
In this embodiment, the euclidean distances between the candidate Copula function types and the cumulative distribution function curve of the empirical Copula function are respectively calculated by using the euclidean distances as indexes, and the Copula function with the minimum calculation result is the optimal Copula function type. Wherein, the cumulative distribution function C of the empirical Copula functionnThe calculation mode is as formula (13):
Figure BDA0003469387700000092
in the formula: wherein u, v ∈ [0,1]](ii) a The value of I is 0, 1; f (x)i) For joint cumulative distribution function, when F (x)i) When the content is less than or equal to u,
Figure BDA0003469387700000093
otherwise
Figure BDA0003469387700000094
The calculation mode of the Euclidean distance between the cumulative distribution functions of different Copula functions and the cumulative distribution function of the empirical Copula function is shown as the formula (14):
Figure BDA0003469387700000095
wherein C isnAnd CpRespectively representing empirical Copula functions and alternativesCumulative distribution function values of Copula function.
When the relevance analysis is carried out on the multidimensional variables, the dependency relationship of the original data structure can be better simulated by connecting the Copula functions by means of the rattan structure model. Meanwhile, the rattan Copula model can be used for converting the complex correlation between the wind, light and load into the correlation between every two simpler variables for analysis. In practical applications, there are two common rattan models, namely a rattan C and a rattan D, and the following descriptions are provided:
1) c rattan structure
The logic diagram of the C-rattan structure is shown in fig. 2, and it can be seen from the diagram that each dimension random variable represented by the rattan structure is connected with one dimension random variable, which indicates that the C-rattan structure is suitable for the situation that the correlation between one dimension variable and the remaining variables is strong, and the correlation between the remaining variables is weak.
When the C vine structure is used for analyzing the multi-dimensional related variable correlation, the joint probability density distribution expression mode is as shown in formula (15):
Figure BDA0003469387700000101
in the formula (f)i(xi) And Fi(xi) Respectively represent XiA probability density function and a cumulative distribution function.
2) D rattan structure
The D rattan model has obvious parallel structure, has better precision when the correlation degree between every two multivariable is close, and the logic diagram is shown in figure 3.
When the relevance of the multidimensional related variables is analyzed by using the D vine structure, the expression mode of the joint probability density distribution is shown as the formula (16)
Figure BDA0003469387700000102
In the formula (f)i(xi) And Fi(xi) Respectively represent XiProbability density function and accumulation ofA distribution function.
In addition, the step of establishing the mixed rattan Copula model further comprises the following steps:
i) optimal vine structure selection
1) Brief introduction to the principles
For variable X, if its cumulative distribution function is denoted by U ═ f (X), U ∈ [0,1], such as if the cumulative distribution function is invertible within the defined domain, then the cumulative distribution function may be transformed as in equation (17):
P(U<u)＝P(F(X)<u)＝P(X<F-1(u))＝F(F-1(u))＝u……(17)
in the formula, F-1(u) is the inverse of the cumulative distribution function. As can be seen from the above equation, the cumulative distribution function of the variable X is a function of [0,1]]A uniformly distributed function within the range. If order U1＝F1(x1),2＝F2(x2) Then, for the joint cumulative distribution function of the two, it is obtained as the following equation (18):
Figure BDA0003469387700000104
from the above equation, the Copula function can be approximately regarded as a joint cumulative distribution function of a uniform distribution function in the range of [0,1 ].
According to the property of the Copula function, the joint cumulative distribution function of the multidimensional variables can be decomposed by using the property of the conditional distribution, and then the variables corresponding to the obtained cumulative distribution function are independent from each other. The mode of obtaining the compound is as shown in formula (19)
Figure BDA0003469387700000103
In the formula vjRepresenting a certain vector v, v-jIndicates the removal of v from vjThe vectors remaining thereafter.
Based on this theoretical basis, for multidimensional variable X1,X2,…,XnA cumulative distribution function ofWith the following assumptions:
Figure BDA0003469387700000111
in the formula ZnRepresents [0,1]]Uniformly distributed data within the range, F (X)n|X1,X2,...,Xn-1) Is the inverse function of equation (20).
Rattan C and rattan D have different choices for j, for rattan C:
Figure BDA0003469387700000112
for D vines:
Figure BDA0003469387700000113
2) multidimensional variable sampling point obtaining step
Combining the rattan structure, the mutual constraint between the dimensions of the sampling sample can be obtained. Therefore, the sampling point acquisition mode based on the rattan structure model is that firstly, an n-dimensional sample set which obeys independent uniform distribution is randomly generated, equations are written according to the structure columns of the C rattan and the D rattan respectively, the first-dimensional sample points are uniformly distributed sampling points, sampling points corresponding to the second dimension can be obtained by using the property of condition distribution, and the sample points of the first dimension are acquired one by one according to the previous sample points in sequence until the sample points of all dimensions are acquired. The specific process can be summarized as follows:
step 1: simulating and generating a group of num multiplied by N (data number x dimension) uniform variables z by using a MATLAB self-contained uniform function;
step 2: make the first set of variables to be solved equal to the first column in step1, i.e. u1＝z1；
And step 3: using the formula (19), the second dimension to be solved for u2Can utilize
Figure BDA0003469387700000114
To calculate, when z is calculated2Generated for step1, C (u)1,u2) To represent
Figure BDA0003469387700000115
u1,u2The function value at the corresponding Copula function is calculated by the Copula cdf function. Solving the right side, adding a while loop, and solving u by adopting a dichotomy in the loop2Until the difference between the calculated data on the right side of the equation and the data on the left side is very small, at which time u is randomly derived2The simulation data is the simulation data;
and 4, step 4: by using
Figure BDA0003469387700000116
And carrying out iterative solution on subsequent dimension simulation data.
Through the steps, the acquisition of the multi-dimensional variable sampling points can be completed. It should be noted that, because the generated Z ═ Z (Z)1,z2,…,zn) The interval is still [0,1]]It is necessary to convert these data into the desired samples according to the inverse function of the estimated edge distribution function.
ii) optimal vine structural index
1) Anderson-Darling (AD) distance Profile
The best rattan structure mode under each scene is selected to connect each Copula function, so that the 'wind-solar-load' correlation analysis is more accurate. From the angle, the embodiment selects an Anderson-Darling (AD) distance index, judges whether the simulation effect of different rattan structure models is good or not, and effectively improves the modeling precision of the mixed rattan Copula model.
The description indexes for the accuracy of the correlation modeling result mainly include the distance between Cramer Von Mise (CvM) and Kolmogorov-Smirnov (K-S), the distance between Anderson-Darling (AD), and the like. CvM the distance measure is the integral of the squared distance of the empirical cumulative distribution function and the target cumulative distribution function. If the data mostly fall on the tail positions of the left side and the right side, the difference of the tail parts is more important to be considered, and the total distribution deviation obtained by CvM distance does not consider the difference caused by the distribution position of large-scale data although all difference points are considered.
The AD distance is actually an improvement over the CvM distance. This weighting function, which is mainly used to limit the weight of the data location, is shown in equation (23) by adding a weighting function. Equivalently, higher weight is given when the square distance of the tail data is calculated; when the square distance of the intermediate data is calculated, smaller weight is given to the intermediate data, and the modeling precision of the rattan structure model can be tested more effectively.
Figure BDA0003469387700000121
2) Optimal vine structure selection based on AD distance
In the process of combining the traditional method with a rattan structure model, the rattan Copula model only relates to a certain rattan structure, so that the method has a disadvantage that the type of a dependent structure which can be analyzed by a certain rattan structure is fixed, and in the case of historical data of 'wind and light load', the single rattan structure connection mode cannot be used for describing the correlation of all 'wind and light load' data in detail and comprehensively due to the fact that the single rattan structure connection mode is too complicated and variable.
For the problem, in this embodiment, the AD distance is selected to check the quality degree of each scene that is simulated by the C rattan structure and the D rattan structure, respectively, and the optimal rattan structure in each scene is selected, so as to improve the accuracy degree of the correlation modeling result.
The AD distance d between the established model and the original data is based on the empirical distribution function expression mentioned aboveADIs represented by formula (24):
Figure BDA0003469387700000122
wherein d represents a random variable dimension, CnAnd CpThe empirical Copula function and the cumulative distribution function value of each alternative Copula function are respectively represented.
When the AD distances of different rattan structure modeling and original sample data are solved, the Copula function is usedThe form is complex, the corresponding multi-dimensional variable rattan structure expression is huge, and after each scene is modeled, the whole model of the original sample has no specific function form, so when the AD distance is calculated, the model established in each scene is sampled by equal sample number respectively, the sampling points with the same number of the original sample points are obtained, the experience distribution function of the newly sampled sample points is calculated, and the corresponding AD distance is solved. dADThe smaller the distance, the higher the modeling accuracy.
S150: and inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result.
Optionally, before inputting the wind-solar-load sample set and the power system parameters into the power transmission system planning model, the method further includes: and establishing a power transmission system planning model according to the objective function and the constraint condition.
The target function comprises a power grid construction cost value, a wind abandoning light abandoning economic loss value and the discharge capacity of a conventional power grid unit, and the constraint conditions comprise power balance constraint, load node new energy power generation penetration power constraint, branch power flow constraint, conventional power generator unit output upper and lower limit constraint, wind turbine unit operation condition constraint, photovoltaic unit operation condition constraint and equipment operation state 0-1 constraint. Under the constraint condition, when the minimization is realized by the 3 objective functions, the optimal power grid planning result is output by the power transmission system planning model.
The power transmission system planning model is further described below in terms of both objective functions and constraint conditions:
i) objective function
1) Cost of power grid construction
In consideration of the economy of power grid investment, the embodiment takes the power grid construction cost as one of the objective functions of power grid planning, including the investment cost, the loss cost and the power generation cost of the power grid.
f1＝CⅠ+CNL+CG……(25)
CⅠWhich is an investment cost.
Figure BDA0003469387700000131
In the formula: x is the number ofm、yn、zoThe decision variables are 0-1, and respectively represent the operation states of a conventional generator set, a wind power plant grid-connected line and a photovoltaic power plant grid-connected line, wherein 0 represents the outage, and 1 represents the operation; cIG、CIWTG、CIPVGEqual annual value for corresponding investment cost; i. j and k are the total number of the conventional generator sets to be increased, the total number of the grid-connected lines of the wind power plant to be increased and the total number of the grid-connected lines of the photovoltaic power station to be increased respectively.
Figure BDA0003469387700000132
Annual value of investment cost of conventional generating set to be added for mth station, IIG,mIs the initial investment cost of a conventional generator set, q is the annual discount rate of the investment, NiThe service life of the conventional generator set.
CIWTG,n、CIPVG,oThe calculation formula is the same as C for the annual value of the investment cost of the nth wind power plant grid-connected line to be increased and the annual value of the investment cost of the photovoltaic power plant grid-connected line to be increasedIG,m。
CNLThe loss cost of the network.
Figure BDA0003469387700000141
In the formula: delta is unit grid power loss price, ten thousand yuan/(kW.h); l is the total number of the original transmission lines of the system; i isu、In、IoThe current flowing on the line is the current flowing on the corresponding time period t; ru、Rn、RoResistance of the corresponding line; t is the total annual time period.
CGWhich is the cost of electricity generation.
Figure BDA0003469387700000142
In the formula: rhoGThe unit power generation cost of a conventional generator set; rhoHThe unit power generation cost of the wind-solar complementary output unit is obtained; pG,tThe active power output of the conventional generator set is the corresponding time period t; pH,tThe active power output of the wind-solar complementary power output unit is obtained in the corresponding time period t.
2) Economic loss of wind and light
The phenomenon of wind abandoning and light abandoning caused by new energy grid connection cannot be ignored, and the economic loss caused by wind abandoning and light abandoning is taken as one of the objective functions of power grid planning in the embodiment.
Figure BDA0003469387700000143
In the formula: c. CWTG、cPVGRespectively is the unit air loss and the economic loss caused by the unit light loss; pWTG0、PPVG0Respectively the planned active power output of the wind turbine generator and the planned active power output of the photovoltaic generator; pWTG、PPVGThe actual active power output of the wind turbine generator and the actual active power output of the photovoltaic generator are respectively.
3) Discharge capacity of system
The pollution discharged by the conventional unit can have a plurality of adverse effects on the environment, and the embodiment takes the discharge capacity of the conventional unit of the power grid as one of the objective functions of the power grid planning, and tries to reduce the discharge capacity to the minimum.
Figure BDA0003469387700000144
In the formula: n is a radical ofEIs the number of pollutants, and r is NO2Where r is 2 is CO2R is 3 SO2；β0、β1、β2Mu and epsilon are the pollution discharge coefficients of the conventional generator set.
ii) constraint conditions
1) And (4) power balance constraint.
Pt＝Btθt……(31)
Figure BDA0003469387700000151
In the formula: ptInjecting a power vector for the node of the time period t; b istA node admittance matrix for time period t; thetatA node voltage phase angle vector for time period t; pL,tThe system active power load is the time period t.
2) And (5) carrying out power generation penetration power constraint on the new energy of the load node.
Figure BDA0003469387700000152
In the formula: pWTGn,maxThe maximum output power of the nth wind turbine generator set is obtained; n is a radical ofWTGThe method comprises the steps of (1) collecting a wind turbine generator set; pPVGo,maxThe maximum output power of the No. o photovoltaic unit; n is a radical ofPVGThe photovoltaic unit is a set; pLf,maxThe maximum penetration power of the load node f.
3) And (5) branch power flow constraint.
|Pu,t|≤Pu,max……(34)
In the formula: pu,tThe active power flow of the u-th branch in the time period t; pu,maxThe upper limit of the transmission power of the u-th branch.
4) And (5) restraining the upper and lower output limits of the conventional generator set.
PG,min≤PG,t≤PG,max……(35)
In the formula: pG,max、PG,minThe active output of the conventional generator set is the upper limit and the lower limit.
5) And (5) restricting the operation condition of the wind turbine generator.
0≤PWTG,t≤PWTG,max……(36)
In the formula PWTG,maxThe maximum output of the wind turbine generator is obtained.
6) And (5) restricting the operation condition of the photovoltaic unit.
0≤PPVG,t≤PPVG,max……(37)
In the formula PPVG,maxThe maximum output of the photovoltaic unit.
The embodiment realizes the minimization of 3 objective functions by combining the objective functions, the constraint conditions and the actual conditions of the wind-light output correlation on the basis of considering the investment cost, the economy and the environmental benefits of the power grid so as to achieve the optimal planning effect. The constructed multi-target power grid expansion planning model is as follows:
Figure BDA0003469387700000161
and adopting an optimization software package to solve in a Matlab R2016 environment, and outputting a power grid planning result.
Further, the probabilistic planning model can be summarized as:
Y_planning＝f_planning(X,T,PLAN)。
wherein, Y _ planning is the construction cost of the power grid, economic loss of wind and light abandoning and sewage discharge of different schemes. X ═ X1,x2,…,xn) Wind speed, illumination intensity, load data; t is a system state parameter, including parameters such as the commissioning state of a conventional generator set, a wind power plant grid-connected line and a photovoltaic power plant grid-connected line; the PLAN is the expansion condition of different power transmission lines.
In other embodiments, latin hypercube sampling may also be used. Latin Hypercube Sampling (LHS) is an extension of hierarchical sampling in multiple dimensions, originally proposed by McKay et al in 1979, whose accuracy is affected by the correlation between input random variables.
The LHS consists of two parts of sampling and sequencing, wherein the purpose of sampling is to enable sampling points to completely cover a variable distribution area, and the purpose of sequencing is to control the correlation between input random variables. Thereby obtaining a sufficiently accurate calculation result with fewer sample points.
When the method is applied to sampling of the multidimensional random variable in the embodiment, compared with Simple Random Sampling (SRS), the LHS method ensures that a sampling point can completely cover a distribution interval and cannot be repeated, so that the method has higher sampling efficiency and better robustness.
Let the cumulative distribution function of the random variable X be F (X), and I ═ F (X), then U is the random variable and is in [0,1]The distribution is uniform. To obtain N samples, [0,1]]The interval is divided into N parts to obtain N non-overlapping subintervals, and the length of each interval is 1/N. Randomly selecting a sample value from each subinterval by means of the inverse F of the cumulative distribution function F (X)-1(X) calculating the sample values. If the subinterval midpoint is selected, the sampling point may be calculated as follows:
Figure BDA0003469387700000162
in the formula, xiIs the sampled value, where i is 1,2, …, N.
According to the above descriptions about the hybrid rattan Copula model, the power transmission system planning model and the LHS method, a power transmission system extension planning method based on the hybrid rattan Copula can be summarized as follows:
according to the above descriptions about the hybrid rattan Copula model, the power transmission system planning model and the LHS method, a power transmission system extension planning method based on the hybrid rattan Copula can be summarized as follows:
step 1: inputting a historical data set D of random variables (mainly comprising continuous random variables such as wind speed, illumination and load) of a power system, clustering parameters K (set to be 2-10 types), sampling number N and the like, and outputting 9 clustering schemes by using a Mini Batch K-Means algorithm;
step 2: obtaining a DBI index value corresponding to each K value by applying formulas (1) to (8) according to the clustering scheme obtained in the step 1), and outputting an optimal clustering number KbestAnd its corresponding scene partitioning scheme C ═ C (C)1,C2,…,CK)；
Step 3: input partition set C ═ C1,C2,…,CK) Respectively fitting the probability density function of the multidimensional variable of each scene by a nonparametric kernel density estimation method, and acquiring an edge distribution function F (x) by using equations (11) and (12)i)。
Step 4: root of herbaceous plantEdge distribution function F (x) obtained according to step 1)i). Selecting the optimal parameters of each alternative Copula function by using the Copula function carried by the MATLAB, and calculating an alternative Copula function C by using formulas (13) and (14)pAnd empirical Copula function CnThe Euclidean distance between the cumulative distribution functions is calculated, and each scene outputs three optimal Copula functions and parameters thereof;
step 5: setting the total number O of samples, and generating an independent variable set L of each scene by using an LHS method by using a formula (9)iWherein i ═ 1,2, …, K;
step 6: the independent variable set L obtained according to step 5)iInputting the multi-dimensional variable sampling points into a rattan Copula model constructed in each scene, respectively sampling (please refer to the step of obtaining the multi-dimensional variable sampling points in section 2 of 4.2 above), and outputting a sampling sample Z;
step 7: according to the sampling sample Z obtained in the step 6), calculating the AD distance under each scene by using formulas (23) and (24), selecting a rattan Copula model with small AD distance as a model of the scene, and outputting a sample data set U;
step 8: obtaining an edge distribution function by using formulas (11) and (12) according to the sample data set U obtained in the step 7) through a nonparametric kernel density method, and converting the sample data into a required 'wind-solar-load' sample set X according to an inverse function of estimated edge distribution;
step 8: and inputting data such as the 'wind, light and load' sample set X and the power system parameter T into a power transmission system planning model, and outputting a power grid planning result.
In this regard, the present application also presents an example validation, specifically:
firstly, an improved IEEE-6 node system is adopted to verify the rationality and the validity of a model, and on the basis of the IEEE-6 node system in the document [1], a wind power plant and a photovoltaic power station are added to perform wind-solar output correlation analysis, as shown in figure 4.
The 3 buses 1, 3 and 6 of the system are connected with a power supply, wherein the power supply G1 is 50MW, the power supply G2 is 165MW, the power supply G3 is 545MW, the scale of a wind farm is 120MW, and the scale of a photovoltaic power station is 120 MW; the cost of a single circuit line per unit length is 80 ten thousand yuan/km.
Firstly, wind power station access node 2 and photovoltaic power station access node 5 are considered in wind and light grid-connected planning, and power grid extension planning is not involved. In the scheme without considering the wind-solar output correlation, system standard data are adopted, and wind-solar output is independent; in the scheme of considering the wind-solar output correlation, a wind-solar output correlation model (referred to as a correlation-considered comparison method) is constructed by adopting the wind-solar output correlation model analysis (referred to as a correlation-considered method herein) established in the text and a large-scale wind-solar complementary power grid expansion planning considering the correlation [ J ]. Table 2 lists the results of the objective function calculations for 3 cases.
TABLE 2 comparison of results
Figure BDA0003469387700000181
As can be seen from the table 2, the fluctuation and uncertainty of independent output of the wind power plant and the photovoltaic power station can be relieved by considering the correlation of wind and light output in the system, the new energy consumption capability is improved to some extent, economic loss of abandoned wind and abandoned light is obviously reduced, the system discharge capacity is also obviously reduced, the construction cost of a power grid is reduced, and certain economic benefit is brought. The method has higher modeling precision, is better than the objective function obtained by a correlation comparison method, is more favorable for accurately simulating the correlation of wind-light joint distribution, and further increases the accuracy of the scheme of power transmission network extension planning.
Based on the foregoing power transmission system planning method considering wind-light correlation, the present application provides a power transmission system planning device considering wind-light correlation, please refer to fig. 5, where fig. 5 is a schematic structural diagram of an embodiment of a power transmission system extension planning device in the present application, and in this embodiment, the power transmission system extension planning device may include a data obtaining module 210, a clustering module 220, a mixed rattan Copula model module 230, and a planning result module 240.
An obtaining data module 210, configured to obtain historical data sets of power system parameters and random variables;
the clustering module 220 is used for performing clustering classification and scene division on the historical data set according to a Mini Batch K-Means clustering method;
a mixed rattan Copula model module 230, configured to determine parameters and types of Copula functions according to each scene data after scene division, generate a sample corresponding to each scene data by using a rattan structure, and determine an optimal rattan structure in each scene by using an AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model;
and the planning result module 240 is configured to input the wind/solar load sample set and the power system parameters into the power transmission system planning model, and finally output a power grid planning result.
Based on the foregoing power transmission system planning method considering the wind-solar correlation, the present application also provides an electronic device, as shown in fig. 6, where fig. 6 is a schematic structural diagram of an embodiment of the electronic device of the present application. The electronic device 300 may comprise a memory 31 and a processor 32, the memory 31 being connected to the processor 32, the memory 31 having stored therein a computer program, the computer program implementing the method of any of the above embodiments when executed by the processor 32. The steps and principles thereof have been described in detail in the above method and will not be described in detail herein.
In the present embodiment, the processor 32 may also be referred to as a Central Processing Unit (CPU). The processor 32 may be an integrated circuit chip having signal processing capabilities. The processor 32 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Based on the above power transmission system planning method considering the wind-solar correlation, the application also provides a computer-readable storage medium. Referring to fig. 7, fig. 7 is a schematic structural diagram of an embodiment of a computer-readable storage medium according to the present application. The computer-readable storage medium 400 has stored thereon a computer program 41, the computer program 41 implementing the method of any of the above embodiments when executed by a processor. The steps and principles thereof have been described in detail in the above method and will not be described in detail herein.
Further, the computer-readable storage medium 400 may be various media that can store program codes, such as a usb disk, a removable hard disk, a read-only memory (ROM), a Random Access Memory (RAM), a magnetic tape, or an optical disk.
In summary, the present application proposes a method, an apparatus, a device and a storage medium for planning a power transmission system considering wind-light correlation, wherein the method includes: acquiring historical data sets of power system parameters and random variables; performing cluster classification and scene division on the historical data set according to a Mini Batch K-Means clustering method; determining Copula function parameters and function types, and establishing a mixed rattan Copula model by taking the AD distance as a judgment standard of the rattan structure; inputting data after scene division into a mixed rattan Copula model, and judging an optimal rattan structure to obtain a wind-solar-lotus sample set; and inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result. By the mode, scene division is carried out on the multi-dimensional data by utilizing a Mini Batch K-Means algorithm, so that the operation time is greatly reduced; the AD distance is used as a judgment standard of the rattan structure, the modeling precision of the hybrid rattan Copula is improved, and the reliability of the power transmission system extension planning method is further improved.
It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not limiting of the application. In addition, for convenience of description, only a part of structures related to the present application, not all of the structures, are shown in the drawings. The step numbers used herein are also for convenience of description only and are not intended as limitations on the order in which the steps are performed. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The terms "first", "second", etc. in this application are used to distinguish between different objects and not to describe a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The above description is only for the purpose of illustrating embodiments of the present application and is not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application or are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.

Claims (10)

1. A method for planning a power transmission system in consideration of wind-solar correlation, comprising:
acquiring historical data sets of power system parameters and random variables;
performing cluster classification and scene division on the historical data set according to a Mini Batch K-Means clustering method;
respectively determining parameters and types of Copula functions according to each scene data after scene division, generating a sample corresponding to each scene data by using a rattan structure, and judging an optimal rattan structure under each scene by adopting an AD distance;
establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model;
and inputting the wind-solar-load sample set and the power system parameters into a power transmission system planning model, and finally outputting a power grid planning result.
2. The method according to claim 1, wherein the stochastic variables include wind speed, light, and load, and wherein the cluster classification and scene division of the historical data set according to the Mini Batch K-Means clustering method comprises:
inputting a historical data set D of continuous random variables of the power system about wind speed, illumination and load, clustering parameters K and sampling data N, and outputting 9 clustering schemes by using a Mini Batch K-Means clustering method;
calculating to obtain DBI index value corresponding to each K value, and outputting the optimal clustering number KbestAnd its corresponding scene division set C ═ (C)1,C2,…,CK)。
3. The method for planning a power transmission system considering wind-solar correlation according to claim 2, further comprising:
inputting the scene division set C ═ (C ═ C)1,C2,…,CK) Respectively fitting the probability density function of the multidimensional variable of each scene by a nonparametric kernel density estimation method to obtain an edge distribution function F (x)i)；
According to the edge distribution function F (x)i) Selecting the optimal parameters of each alternative Copula function by using the Copula function carried by the MATLAB, and calculating an alternative Copula function CpAnd empirical Copula function CnThe Euclidean distance between the cumulative distribution functions is calculated, and each scene outputs three optimal Copula functions and parameters thereof;
setting the total number O of samples, and generating an independent variable set L of each scene by using an LHS methodiWherein i ═ 1,2, …, K;
set L of the independent variablesiInputting a rattan Copula model constructed in each scene, respectively sampling, and outputting a sampling sample Z;
and calculating the AD distance in each scene according to the sampling sample Z, and selecting a rattan Copula model with small AD distance as a mixed rattan Copula model of the scene.
4. The method for planning a power transmission system considering wind-solar correlation according to claim 3, further comprising:
and obtaining a sample data set U according to the mixed rattan Copula model, obtaining an edge distribution function through a nonparametric kernel density method, and converting the sample data set U into a required wind-solar-load sample set X according to an inverse function of estimated edge distribution.
5. A method for planning a power transmission system taking account of wind-solar correlation according to claim 3, wherein said independent variable sets LiInputting the rattan Copula model built by each scene and sampling respectively, and outputting a sampling sample Z, wherein the method comprises the following steps:
set L of the independent variablesiRandomly generating an n-dimensional sample set which is subjected to independent uniform distribution;
and according to the structural column writing equations of the C rattan and the D rattan, the first-dimension sample points are uniformly distributed sampling points, sampling points corresponding to the second dimension are obtained by utilizing the property of conditional distribution, the sample points of the first dimension are sequentially obtained one by one according to the previous one-dimension sample points until all the sample points of the n dimensions are obtained, and finally the sampling sample Z is generated.
6. The method of claim 1, wherein the inputting the set of wind-solar-load samples and the power system parameters into a power transmission system planning model further comprises:
establishing a power transmission system planning model according to the objective function and the constraint condition;
the target function comprises a power grid construction cost value, a wind abandoning and light abandoning economic loss value and the discharge capacity of a conventional power grid unit, and the constraint conditions comprise power balance constraint, load node new energy power generation penetration power constraint, branch flow constraint, conventional power generator unit output upper and lower limit constraint, wind power unit operation condition constraint, photovoltaic unit operation condition constraint and equipment operation state 0-1 constraint.
7. The method according to claim 6, wherein the inputting the wind-solar-load sample set and the power system parameters into a power transmission system planning model and finally outputting a power grid planning result comprises:
under the constraint condition, when the minimization is realized by the 3 objective functions, the optimal power grid planning result is output.
8. A power transmission system planning apparatus considering wind-solar correlation, comprising:
the data acquisition module is used for acquiring historical data sets of the parameters and the random variables of the power system;
the clustering module is used for carrying out clustering classification and scene division on the historical data set according to a Mini Batch K-Means clustering method;
the mixed rattan Copula model module is used for respectively determining parameters and types of Copula functions according to each scene data after scene division, generating samples corresponding to each scene data by using rattan structures, and judging the optimal rattan structure in each scene by adopting AD distance; establishing a mixed rattan Copula model according to the optimal rattan structure under each scene, and generating a wind-light-load sample set according to the proportion of each scene data in the mixed rattan Copula model;
and the planning result module is used for inputting the wind-solar load sample set and the power system parameters into a power transmission system planning model and finally outputting a power grid planning result.
9. An electronic device, comprising a memory and a processor, the memory being coupled to the processor, the memory storing a computer program which, when executed by the processor, implements a method of power transmission system planning that takes into account wind-solar correlations as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that a computer program is stored which, when executed, implements the method of power transmission system planning taking into account wind-solar correlations as claimed in any one of claims 1 to 7.
CN202210039670.9A 2022-01-13 2022-01-13 Power transmission system planning method and device considering wind-light correlation and storage medium Pending CN114386700A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210039670.9A CN114386700A (en) 2022-01-13 2022-01-13 Power transmission system planning method and device considering wind-light correlation and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210039670.9A CN114386700A (en) 2022-01-13 2022-01-13 Power transmission system planning method and device considering wind-light correlation and storage medium

Publications (1)

Publication Number Publication Date
CN114386700A true CN114386700A (en) 2022-04-22

Family

ID=81202754

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210039670.9A Pending CN114386700A (en) 2022-01-13 2022-01-13 Power transmission system planning method and device considering wind-light correlation and storage medium

Country Status (1)

Country Link
CN (1) CN114386700A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116776613A (en) * 2023-06-27 2023-09-19 国家电网有限公司华东分部 Wind-light output scene reconstruction system
CN116993032A (en) * 2023-09-28 2023-11-03 国网山西省电力公司运城供电公司 Distribution network planning method, distribution network planning device, storage medium and computer equipment
CN117410991A (en) * 2023-12-15 2024-01-16 国网湖北省电力有限公司经济技术研究院 Power balancing method, system, equipment and storage medium for distributed resources

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116776613A (en) * 2023-06-27 2023-09-19 国家电网有限公司华东分部 Wind-light output scene reconstruction system
CN116776613B (en) * 2023-06-27 2024-02-06 国家电网有限公司华东分部 Wind-light output scene reconstruction system
CN116993032A (en) * 2023-09-28 2023-11-03 国网山西省电力公司运城供电公司 Distribution network planning method, distribution network planning device, storage medium and computer equipment
CN116993032B (en) * 2023-09-28 2024-01-19 国网山西省电力公司运城供电公司 Distribution network planning method, distribution network planning device, storage medium and computer equipment
CN117410991A (en) * 2023-12-15 2024-01-16 国网湖北省电力有限公司经济技术研究院 Power balancing method, system, equipment and storage medium for distributed resources
CN117410991B (en) * 2023-12-15 2024-03-01 国网湖北省电力有限公司经济技术研究院 Power balancing method, system, equipment and storage medium for distributed resources

Similar Documents

Publication Publication Date Title
Hu et al. A new clustering approach for scenario reduction in multi-stochastic variable programming
CN114386700A (en) Power transmission system planning method and device considering wind-light correlation and storage medium
CN102693451B (en) Method for predicting ammonia process flue gas desulphurization efficiency based on multiple parameters
CN109308571B (en) Distribution line variable relation detection method
CN109886464B (en) Low-information-loss short-term wind speed prediction method based on optimized singular value decomposition generated feature set
CN111401755B (en) Multi-new-energy output scene generation method, device and system based on Markov chain
CN107612016B (en) Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy
Yang et al. Hybrid prediction method for wind speed combining ensemble empirical mode decomposition and Bayesian ridge regression
CN110380444B (en) Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula
Cheng et al. Enhanced state estimation and bad data identification in active power distribution networks using photovoltaic power forecasting
CN110889603A (en) Power system economic dispatching method considering wind power correlation based on PCA-Copula theory
CN111027775A (en) Step hydropower station generating capacity prediction method based on long-term and short-term memory network
CN111162537B (en) Latin hypercube sampling method probability load flow calculation method based on combined Copula function
CN115000968A (en) Offshore wind power access system planning method based on game and wind power output correlation
CN115456312A (en) Short-term power load prediction method and system based on octyl geometric modal decomposition
CN108694475B (en) Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model
CN114266416A (en) Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium
CN113659631A (en) Wind-solar power station group output description method considering time-varying characteristics
CN110991741B (en) Section constraint probability early warning method and system based on deep learning
CN109447512B (en) Large power grid reliability assessment method based on uniform design
CN115051363B (en) Distribution network area user change relation identification method and device and computer storage medium
CN113627655B (en) Method and device for simulating and predicting pre-disaster fault scene of power distribution network
CN114971090A (en) Electric heating load prediction method, system, equipment and medium
Puiu et al. Principled data completion of network constraints for day ahead auctions in power markets
Putra et al. Modeling of high uncertainty photovoltaic generation in quasi dynamic power flow on distribution systems: A case study in Java Island, Indonesia

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