CN111737924A - Method for selecting typical load characteristic transformer substation based on multi-source data - Google Patents
Method for selecting typical load characteristic transformer substation based on multi-source data Download PDFInfo
- Publication number
- CN111737924A CN111737924A CN202010822559.8A CN202010822559A CN111737924A CN 111737924 A CN111737924 A CN 111737924A CN 202010822559 A CN202010822559 A CN 202010822559A CN 111737924 A CN111737924 A CN 111737924A
- Authority
- CN
- China
- Prior art keywords
- substation
- load
- transformer
- industry
- clustering
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/04—Power grid distribution networks
Abstract
The invention relates to the technical field of power system load modeling, in particular to a method for selecting a typical load characteristic transformer substation based on multi-source data. The method comprises the steps of generally surveying load characteristics of a transformer substation of a target power grid under the same voltage level to obtain load characteristic data, wherein the load characteristic data comprises load type data and industry composition data; performing type clustering analysis on the transformer substations according to the load type data, and performing load classification on the transformer substations to obtain a plurality of transformer substation groups with similar load characteristics; performing industry clustering analysis on the transformer substation groups according to the industry composition data, and performing industry classification on the transformer substations; and selecting a typical transformer substation capable of representing the load characteristics according to the load classification and the industry classification. The invention applies the aggregation theory method to the selection of the typical sites, provides the basis for selecting the typical sites according to the membership relationship, and provides scientific basis for the modeler in the process of selecting the typical sites.
Description
Technical Field
The invention relates to the technical field of power system load modeling, in particular to a method for selecting a typical load characteristic transformer substation based on multi-source data.
Background
The load modeling work of the power system becomes a hotspot and key field of the power industry, the power system consists of a power plant, a power transmission network and a power load, the power load can be divided into industrial load, residential load, commercial load, agricultural load and other loads according to different power utilization main bodies, and in the research and application fields of the load modeling of the power system, the method for establishing the load model by the statistical synthesis method is widely used in the actual modeling work due to the advantages of clear physical model and high model precision. However, due to the characteristics of complexity, dispersion and randomness of the loads, the workload of completing detailed investigation and then integrating all the loads in the power system is too large, and the actual operability is not achieved. The comprehensive loads with load characteristics close to or similar to those of the same power grid at different load points and in different time periods are classified, and the classified load characteristics are described by using the same load model, so that the problem of workload redundancy of statistical comprehensive load modeling can be solved by selecting a typical transformer substation to replace the transformer substations with similar load characteristics to realize statistical comprehensive load modeling. The typical substation is selected according to personal experience, deviation is generated due to difference, and therefore the typical substation selection process needs to be achieved through a screening algorithm. The loads are divided into a plurality of types according to seasons, time, load levels or load composition, modeling is carried out on each type of load respectively, reasonable selection is achieved, the change of the difference of the loads is represented as continuity, and no distinct boundary exists between different objects, so that transformer substation classification has the characteristic of fuzziness, load characteristic classification and integration in the field of load modeling are necessary and feasible, and the load composition of the transformer substation can be selected as a characteristic vector to realize selection of a typical transformer substation. Industrial electricity occupies a large proportion in an electricity utilization structure of the whole society, so that load peak-valley difference is large, main electric equipment in different industries is different, so that load characteristics of different transformer substations are large, in order to further improve representativeness of a typical transformer substation, industry composition conditions in industrial loads are considered in a selection process, and the defects that a single general investigation result is random, and certain errors are caused when load classification and synthesis are carried out are overcome.
Therefore, it is necessary to provide a method for selecting a typical load characteristic substation more accurately.
Disclosure of Invention
The invention aims to at least solve one of the technical problems in the prior art and provides a method for selecting a typical load characteristic transformer substation based on multi-source data.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows: a method for selecting a typical load characteristic substation based on multi-source data comprises the following steps:
step 2, performing type clustering analysis on the transformer substations according to the load type data, and performing load classification on the transformer substations to obtain a plurality of transformer substation groups with similar load characteristics;
step 3, performing industry clustering analysis on the transformer substation groups according to industry composition data, and classifying the transformer substations in industry;
and 4, selecting a typical transformer substation capable of representing the load characteristics according to the load classification and the industry classification.
Further, in the step 1, the method specifically includes:
step 1.1, dividing survey ranges of a target power grid according to cities;
step 1.2, carrying out load general survey on all transformer substations in the same voltage class in the district of the city;
step 1.3, investigating load types borne by each transformer substation, active power consumed by each load type and occupied proportion to obtain load type data, wherein the load types comprise industrial loads, residential loads, commercial loads, agricultural loads and other loads;
and step 1.4, investigating the specific industry types and the occupied proportions in the industrial loads of all the substations to obtain industry composition data.
Further, in the step 2, the method specifically includes:
step 2.1, selecting the load constitution of the transformer substation as a characteristic vector;
wherein the content of the first and second substances,representing the load composition of the ith transformer substation, wherein n is the number of the transformer substations, and m is the number of the characteristic indexes;
step 2.2, applying a polymerization theory method to the transformer substation load type classification, and improving a fuzzy clustering algorithm by adopting genetic simulated annealing;
and 2.3, calculating the similarity among the substations through fuzzy clustering, and converting the similarity among different substation load compositions into the magnitude of the membership degree for expression, wherein the membership degree is as follows:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the jth initial cluster center and,is the kth initial clustering center, K is the number of clustering centers,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.4, constructing a fuzzy similar matrix according to the distance matrix, randomly selecting c clustering centers through a genetic algorithm, initializing control parameters, initializing a population, calculating the membership and fitness of each individual according to the distance matrix, performing selection, crossing, mutation and other operations through the genetic algorithm to generate a new population, and replacing or receiving the old individual through a simulated annealing algorithm; wherein, the clustering center is:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.5, target functionThe distance square sum of each substation load and the corresponding clustering center is formed and then summed, and an objective function is outputA minimum cluster center and membership matrix, wherein the objective function is:
in the formula (I), the compound is shown in the specification,each substation load constitutes the sum of squared distances from its corresponding cluster center,is the degree of membership of the ith substation to the jth substation,is the feature vector of the ith substation,is the jth initial cluster center;
and 2.6, according to the maximum membership principle of the substation membership matrix and the clustering center, taking the membership degree between the substation and the clustering center as a basis for selecting a typical site, and carrying out load classification on the substation.
Further, in the step 3, the method specifically includes:
step 3.1, selecting industry components in the industrial load as feature vectors;
in the formula (I), the compound is shown in the specification,the method is characterized by comprising the following steps that (1) the industrial composition in the ith transformer substation industrial load is realized, p is the category of the transformer substation industrial load, and n is the number of the transformer substations;
step 3.2, performing data dimension reduction on the industry composition data in the industrial load through multidimensional scaling MDS, and mapping data point pairs with equal distances in a high-dimensional space in a low-dimensional space;
in the formula (I), the compound is shown in the specification,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the x-axis vector after the dimension reduction,the y-axis vector after dimensionality reduction, and n is the number of the transformer substations;
3.3, clustering the industry composition data subjected to dimensionality reduction by adopting a genetic simulated annealing improved fuzzy clustering algorithm;
step 3.4, outputting the objective functionThe objective function of the minimum clustering center and the membership matrix is as follows:
in the formula (I), the compound is shown in the specification,the distance square sum of each transformer substation industry composition and the corresponding clustering center is summed,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the jth clustering center, n is the number of the transformer substations, and T is the number of the clustering centers;
the membership matrix is:
in the formula (I), the compound is shown in the specification,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the j-th cluster center and,is the T-th clustering center, and T is the number of the clustering centers;
the clustering center is as follows:
in the formula (I), the compound is shown in the specification,is the j-th cluster center and,is the membership degree formed by the i-th substation and the j-th substation industry,the characteristic vector is formed by industries in the i-th transformer substation industrial load after dimensionality reduction, and n is the number of the transformer substations;
and 3.5, classifying the industry composition membership matrix and the corresponding clustering center matrix according to the maximum membership principle.
Further, in the step 4, for the transformer substation groups obtained by load classification, the transformer substation closest to the corresponding clustering center is selected from each group as a typical load characteristic transformer substation.
The invention has the beneficial effects that: compared with the prior art, the method has the advantages that the aggregation theory method is used for selecting the typical sites, the basis for selecting the typical sites is given according to the membership relationship, and scientific basis is provided for modelers in the process of selecting the typical sites; when load characteristic classification is carried out on the investigated target transformer substation, a classification principle is not set, so that the problem that the manually set classification principle is unreasonable is solved; the simulated annealing algorithm and the genetic algorithm are combined to improve the fuzzy clustering algorithm and are applied to selecting a typical transformer substation, so that the problem of inaccurate algorithm convergence caused by overlarge quantity space and improper initial value selection is solved; multilevel data such as load type composition of the transformer substation and industry composition in industrial load are organically combined in the process of selecting the typical station, so that the selected typical transformer substation can simultaneously well represent load characteristics and industry characteristics, and the problem of uniformity of typical basis in the process of screening the typical station is solved. A practical and efficient method is provided for selecting a typical transformer substation process in the field of electric power system load modeling based on a statistical synthesis method.
Drawings
Fig. 1 is a block flow diagram of a method for selecting a typical load characteristic substation based on multi-source data according to a preferred embodiment of the present invention;
FIG. 2 is a flow chart of the genetic simulated annealing algorithm to improve the fuzzy clustering algorithm in the preferred embodiment of the present invention.
Detailed Description
The technical solutions in the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments.
Referring to fig. 1-2, a preferred embodiment of the present invention, a method for selecting a typical load characteristic substation based on multi-source data, the method comprises the following steps:
step 2, performing type clustering analysis on the transformer substations according to the load type data, and performing load classification on the transformer substations to obtain a plurality of transformer substation groups with similar load characteristics;
step 3, performing industry clustering analysis on the transformer substation groups according to industry composition data, and classifying the transformer substations in industry;
and 4, selecting a typical transformer substation capable of representing the load characteristics according to the load classification and the industry classification.
In this embodiment, in the step 1, the method specifically includes:
step 1.1, dividing survey ranges of a target power grid according to cities;
step 1.2, carrying out load general survey on all transformer substations in the same voltage class in the district of the city;
step 1.3, investigating load types borne by each transformer substation, active power consumed by each load type and occupied proportion to obtain load type data, wherein the load types comprise industrial loads, residential loads, commercial loads, agricultural loads and other loads;
and step 1.4, investigating the specific industry types and the occupied proportions in the industrial loads of all the substations to obtain industry composition data.
In this embodiment, in the step 2, the method specifically includes:
step 2.1, selecting the load constitution of the transformer substation as a characteristic vector;
wherein the content of the first and second substances,representing the load composition of the ith transformer substation, wherein n is the number of the transformer substations, and m is the number of the characteristic indexes;
step 2.2, applying a polymerization theory method to the transformer substation load type classification, and improving a fuzzy clustering algorithm by adopting genetic simulated annealing;
and 2.3, calculating the similarity among the substations through fuzzy clustering, and converting the similarity among different substation load compositions into the magnitude of the membership degree for expression, wherein the membership degree is as follows:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the jth initial cluster center and,is the kth initial clustering center, K is the number of clustering centers,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.4, constructing a fuzzy similar matrix according to the distance matrix, randomly selecting c clustering centers through a genetic algorithm, initializing control parameters, initializing a population, calculating the membership and fitness of each individual according to the distance matrix, performing selection, crossing, mutation and other operations through the genetic algorithm to generate a new population, and replacing or receiving the old individual through a simulated annealing algorithm; wherein, the clustering center is:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.5, target functionThe distance square sum of each substation load and the corresponding clustering center is formed and then summed, and an objective function is outputA minimum cluster center and membership matrix, wherein the objective function is:
in the formula (I), the compound is shown in the specification,each substation load constitutes the sum of squared distances from its corresponding cluster center,is the degree of membership of the ith substation to the jth substation,is the feature vector of the ith substation,is the jth initial cluster center;
and 2.6, according to the maximum membership principle of the substation membership matrix and the clustering center, taking the membership degree between the substation and the clustering center as a basis for selecting a typical site, and carrying out load classification on the substation.
In this embodiment, in the step 3, the method specifically includes:
step 3.1, selecting industry components in the industrial load as feature vectors;
in the formula (I), the compound is shown in the specification,the method is characterized by comprising the following steps that (1) the industrial composition in the ith transformer substation industrial load is realized, p is the category of the transformer substation industrial load, and n is the number of the transformer substations;
step 3.2, performing data dimension reduction on the industry composition data in the industrial load through multidimensional scaling MDS, and mapping data point pairs with equal distances in a high-dimensional space in a low-dimensional space;
in the formula (I), the compound is shown in the specification,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the x-axis vector after the dimension reduction,the y-axis vector after dimensionality reduction, and n is the number of the transformer substations;
3.3, clustering the industry composition data subjected to dimensionality reduction by adopting a genetic simulated annealing improved fuzzy clustering algorithm;
step 3.4, outputting the objective functionThe objective function of the minimum clustering center and the membership matrix is as follows:
in the formula (I), the compound is shown in the specification,the distance square sum of each transformer substation industry composition and the corresponding clustering center is summed,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the jth cluster centerN is the number of the transformer substations, and T is the number of the clustering centers;
the membership matrix is:
in the formula (I), the compound is shown in the specification,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the j-th cluster center and,is the T-th clustering center, and T is the number of the clustering centers;
the clustering center is as follows:
in the formula (I), the compound is shown in the specification,is the j-th cluster center and,is the membership degree formed by the i-th substation and the j-th substation industry,the characteristic vector is formed by industries in the i-th transformer substation industrial load after dimensionality reduction, and n is the number of the transformer substations;
and 3.5, classifying the industry composition membership matrix and the corresponding clustering center matrix according to the maximum membership principle.
In this embodiment, in step 4, for the transformer substation groups obtained by load classification, the transformer substation closest to the corresponding clustering center is selected as the typical load characteristic transformer substation in each group.
The invention applies the aggregation theory method to the selection of the typical sites, gives the basis for selecting the typical sites according to the membership relationship, and provides scientific basis for the modeler in the process of selecting the typical sites; when load characteristic classification is carried out on the investigated target transformer substation, a classification principle is not set, so that the problem that the manually set classification principle is unreasonable is solved; the simulated annealing algorithm and the genetic algorithm are combined to improve the fuzzy clustering algorithm and are applied to selecting a typical transformer substation, so that the problem of inaccurate algorithm convergence caused by overlarge quantity space and improper initial value selection is solved; multilevel data such as load type composition of the transformer substation and industry composition in industrial load are organically combined in the process of selecting the typical station, so that the selected typical transformer substation can simultaneously well represent load characteristics and industry characteristics, and the problem of uniformity of typical basis in the process of screening the typical station is solved. A practical and efficient method is provided for selecting a typical transformer substation process in the field of electric power system load modeling based on a statistical synthesis method.
In order to facilitate an understanding of the present invention, the following provides a more detailed method of the present invention:
a method for selecting a typical load characteristic transformer substation based on multi-source data is characterized by comprising the following steps:
A. developing the general survey of the load characteristics of the transformer substation under the same voltage level of the target power grid, wherein the survey contents comprise: the load types and the occupied proportion of the load types in the power supply range of each transformer substation are as follows, the load types include 5 types: industrial, residential, commercial, agricultural and other loads; specific types and configurations of industrial loads, the industrial types including: mining, chemical, petroleum, paper, food processing, mechanical, transportation, electrical, electronics, textile, metal processing, rubber and plastic manufacturing, wood processing, tobacco industry, printing industry, leather industry, steel industry, electrified railroad, electrolytic aluminum, medicine, electroceramics, cement, photovoltaics, and others. (ii) a
A1. Dividing a survey range of a target power grid according to cities;
A2. carrying out load general investigation on all transformer substations in the same voltage level in the district of the local city;
A3. load type data survey content and form: investigating the active power consumed and the proportion thereof in the types of loads, wherein the types of loads comprise industrial loads, residential loads, commercial loads, agricultural loads and other loads;
A4. content and form of industrial load survey: investigating the specific type and the occupied proportion of the industrial load;
B. performing clustering analysis on the transformer substation according to the load type data to finish transformer substation load classification;
B1. selecting load constitution of a transformer substation as a characteristic vector;
B2. the aggregation theory method is used for transformer substation classification, a fuzzy clustering algorithm is improved by using genetic simulated annealing, and the detailed process is shown in an attached figure 2;
B3. calculating the similarity among the transformer substations through fuzzy clustering, and converting the similarity among different transformer substation load compositions into the magnitude of membership for expression;
B4. constructing a fuzzy similar matrix according to the distance matrix, randomly selecting c clustering centers through a genetic algorithm, initializing control parameters, initializing a population, calculating the membership degree and the fitness degree of each individual according to the distance matrix, performing selection, crossing, mutation and other operations through the genetic algorithm to generate a new population, and replacing or receiving old individuals through a simulated annealing algorithm; specifically, the specific process of the steps comprises:
recording n transformer substation objects to be classified as a setEach of the transformer substationsThere are m characteristic indexes (where power transformation is selected)Station 5 different types of load formation as eigenvectors), i.e.Can be usedDimensional property index vector to represent:,and the load structure of the ith transformer substation is shown, n is the number of the transformer substations, and m is the number of the characteristic indexes. All the characteristic indexes of m objects can form a matrix, and the matrix is recorded asBalance ofIs a matrix of characteristic indicators for X.
a. Data normalization
Because the dimension and the magnitude of the m characteristic indexes are not necessarily the same, the action of the characteristic indexes with a certain magnitude on classification can be highlighted in the operation process, and the action of the characteristic indexes with small magnitude is reduced or even eliminated. Therefore, the data needs to be standardized, scaled to fall into a small specific interval, unit limitation of the data is removed, and the data is converted into a dimensionless pure numerical value, so that indexes of different units or orders of magnitude can be compared and weighted conveniently.
Normalization treatment:
for forward sequenceTo carry outAs a result of the transformation as follows,then the new sequence after treatmentAll fall within the new interval [0,1 ]]Is internal and dimensionless, and has. The characteristic index matrix after normalization processing is as follows:
b. constructing fuzzy similar matrix, initializing membership degree matrix
Clustering is to identify the closeness between elements in the characteristic index matrix X according to a certain criterion, and classify objects close to each other into a class. For this purpose, the required voltage is in [0,1 ]]Number inShowing objects in XAndthe degree of closeness or similarity between them, i.e. the similarity factor. Matrix of characteristic indexesData in (1)Is standardized to obtain,Andthe degree of similarity between them is recorded asAnd in the interval [0,1 ]]In this way, a fuzzy similarity matrix between all objects can be obtained。
For the objectAnddegree of similarity therebetween (similarity coefficient)The determination is measured by the linear distance between the object point and the point in the multidimensional space, and the Euclidean distance is adopted for determination:
in the formula (I), the compound is shown in the specification,is thatAndthe euclidean distance between them,is the feature vector of the ith substation,is the feature vector of the jth substation,is the 1 st feature vector of the ith substation,is the 1 st eigenvector of the jth substation, and m is the eigenvector dimension;
c. fuzzy classification
Due to the fuzzy relation matrix between the objects constructed by the above methodsGenerally speaking, it is just a fuzzy similarity matrix, and not necessarily transitive. Therefore, a new fuzzy equivalent matrix is constructed from the R, and then dynamic clustering is performed based on the fuzzy equivalent matrix.
As described above, the transitive closure of the fuzzy similarity matrix RIs a fuzzy equivalence matrix. To be provided withThe clustering method based on classification is called fuzzy transmission closed-packet method.
The method comprises the following specific steps: (1) transfer closure for solving fuzzy similarity matrix R by using square self-synthesis method(ii) a (2) Selecting confidence level values appropriatelyIn the range of [0,1]To find outIs cut matrixWhich is an equivalent Boole matrix on X. Then press againstClassifying, the obtained classification isAnd equivalence classification on the horizontal.
is a confidence level value forIf, ifThen is atObject at confidence levelAnd objectFall into the same category.
Degree of membership:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the jth initial cluster center and,is the k-th initial cluster center and,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
clustering center:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
d. initializing control parameters, including a weighting index, a maximum iteration number and an evaluation objective function in a fuzzy C-means clustering algorithm, a population size, a maximum evolution number, a cross probability and a variation probability in a genetic algorithm, and simulating an annealing initial temperature, a temperature cooling coefficient and a termination temperature in an annealing algorithm;
B5. an objective functionThe distance square sum of each substation load and the corresponding clustering center is formed and then summed, and an objective function is outputThe minimum clustering center and membership matrix;
an objective function:
in the formula (I), the compound is shown in the specification,each substation load constitutes the sum of squared distances from its corresponding cluster center,is the degree of membership of the ith substation to the jth substation,is the feature vector of the ith substation,is the jth initial cluster center;
B6. according to the maximum membership principle of the substation membership matrix and the clustering center, the classification of the substations is completed, the membership degree between the substation and the clustering center is used as a basis for selecting typical sites, and a specific substation load classification principle is not set manually;
C. for each group of substations, carrying out substation industry classification on the substations according to the industry composition data;
step 3.1, selecting industry components in the industrial load as feature vectors;
in the formula (I), the compound is shown in the specification,is the industry composition in the i-th substation industrial load, p is the class of the substation industrial load, n is the variationThe number of power stations;
step 3.2, performing data dimension reduction on the industry composition data in the industrial load through multidimensional scaling MDS, and mapping data point pairs with equal distances in a high-dimensional space in a low-dimensional space;
in the formula (I), the compound is shown in the specification,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the x-axis vector after the dimension reduction,the y-axis vector after dimensionality reduction, and n is the number of the transformer substations;
3.3, clustering the industry composition data subjected to dimensionality reduction by adopting a genetic simulated annealing improved fuzzy clustering algorithm;
step 3.4, outputting the objective functionThe objective function of the minimum clustering center and the membership matrix is as follows:
in the formula (I), the compound is shown in the specification,the distance square sum of each transformer substation industry composition and the corresponding clustering center is summed,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the jth clustering center, n is the number of the transformer substations, and T is the number of the clustering centers;
the membership matrix is:
in the formula (I), the compound is shown in the specification,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the j-th cluster center and,is the T-th clustering center, and T is the number of the clustering centers;
the clustering center is as follows:
in the formula (I), the compound is shown in the specification,is the j-th cluster center and,is the membership degree formed by the i-th substation and the j-th substation industry,the characteristic vector is formed by industries in the i-th transformer substation industrial load after dimensionality reduction, and n is the number of the transformer substations;
and 3.5, classifying the industry composition membership matrix and the corresponding clustering center matrix according to the maximum membership principle.
D. Selecting a typical transformer substation;
D1. for the load classification, selecting the transformer substation closest to the corresponding clustering center as a typical load characteristic in each class;
D2. the selected typical transformer substation can simultaneously well represent load characteristics and industrial characteristics, and the uniformity of typicality in the typical site selection process is guaranteed.
The method comprises the steps of combining load composition data with industrial composition data in industrial loads to finish selecting a typical transformer substation, and finishing the aggregation grouping of the transformer substations based on the load composition by selecting the load composition as a characteristic vector to ensure that the transformer substations in each group have similar load compositions; secondly, selecting industry composition data as a characteristic vector for each group of substations based on the grouping condition in the first step, and finishing the aggregation grouping of the substations based on the industry composition in the industrial load so as to classify the substations with similar industry compositions into one class; and thirdly, selecting representative transformer substations in each group of transformer substations as typical transformer substations according to the relation between the clustering centers and the membership degrees, so that the screened typical transformer substations can show the composition proportion of 5 different load types of the transformer substations and can also show the industry composition characteristics in industrial loads. The problem that the stability of results is affected due to different initialization is optimized through a simulated annealing algorithm, the accuracy of a fuzzy clustering algorithm is improved through iteration of a genetic algorithm, the simulated annealing algorithm is combined with the genetic algorithm to improve a fuzzy C mean value clustering method, the fuzzy C mean value clustering method is a clustering analysis method based on an objective function, the closeness degree, namely membership degree, of all elements is calculated according to the objective function and serves as a method for dividing similar transformer substations, and a clustering center can represent the load characteristics of the transformer substations.
One specific example is provided below:
the typical load site selection method provided by the invention comprises the following steps:
1. developing general load characteristic survey aiming at all 220kV transformer substations carried by the power grid in Jiangxi, and collecting 160 groups of effective data in the general load characteristic survey;
2. classifying the transformer substation according to the load composition;
fuzzy C-means clustering is improved through a genetic simulated annealing algorithm, loads are selected to form characteristic vectors, 160 transformer substations are divided into 9 classes, and the clustering center of each class is shown in table 1;
TABLE 1 clustering center List
According to the query clustering center and the substation membership matrix, the substations are grouped according to the maximum membership principle, and the substation grouping result is shown in the following table 2;
TABLE 2 grouping of substations
3. Correspondingly sorting out the industry components in the industry components corresponding to each group of substations according to the grouping condition of the substations, and mapping 25-dimensional industry component data in a two-dimensional space in an equidistance manner by multi-dimensional scaling, wherein the data are shown in a table below;
table 3.1 mapping table is composed of the first group of substation industries
Table 3.2 second group of substation industries form mapping table
Table 3.3 mapping table is formed by the third group of substation industries
Table 3.4 mapping table for industry composition of fourth group of substations
Table 3.5 fifth group substation industry composition mapping table
Table 3.6 sixth group of substations industry composition mapping table
Table 3.7 seventh group substation industry composition mapping table
Table 3.8 eighth group substation industry composition mapping table
Table 3.9 ninth group substation industry composition mapping table
4. The C mean value clustering algorithm is improved through a genetic simulated annealing algorithm to cluster the industry composition data of each group, and the clustering center of each group is shown in the following table 4;
TABLE 4 Cluster centers of each group
5. Selecting a transformer substation of a clustering center as a typical transformer substation which represents the most representative industry composition characteristics in the load characteristic transformer substations, and referring to the following table 5;
TABLE 5 typical substations
The above additional technical features can be freely combined and used in superposition by those skilled in the art without conflict.
It is to be understood that the present invention has been described with reference to certain embodiments, and that various changes in the features and embodiments, or equivalent substitutions may be made therein by those skilled in the art without departing from the spirit and scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.
Claims (5)
1. A method for selecting a typical load characteristic transformer substation based on multi-source data is characterized by comprising the following steps:
step 1, carrying out load characteristic general survey on a transformer substation of a target power grid under the same voltage level to obtain load characteristic data, wherein the load characteristic data comprises load type data and industry composition data;
step 2, performing type clustering analysis on the transformer substations according to the load type data, and performing load classification on the transformer substations to obtain a plurality of transformer substation groups with similar load characteristics;
step 3, performing industry clustering analysis on the transformer substation groups according to industry composition data, and classifying the transformer substations in industry;
and 4, selecting a typical transformer substation capable of representing the load characteristics according to the load classification and the industry classification.
2. The method for selecting the substation with the typical load characteristics based on the multi-source data according to claim 1, wherein the method comprises the following steps: in the step 1, the method specifically includes:
step 1.1, dividing survey ranges of a target power grid according to cities;
step 1.2, carrying out load general survey on all transformer substations in the same voltage class in the district of the city;
step 1.3, investigating load types borne by each transformer substation, active power consumed by each load type and occupied proportion to obtain load type data, wherein the load types comprise industrial loads, residential loads, commercial loads, agricultural loads and other loads;
and step 1.4, investigating the specific industry types and the occupied proportions in the industrial loads of all the substations to obtain industry composition data.
3. The method for selecting the substation with the typical load characteristics based on the multi-source data according to claim 1, wherein the method comprises the following steps: in the step 2, the method specifically includes:
step 2.1, selecting the load constitution of the transformer substation as a characteristic vector;
wherein the content of the first and second substances,the load composition of the ith transformer substation is shown, n is the number of the transformer substations, and m is the number of the characteristic indexes;
step 2.2, applying a polymerization theory method to the transformer substation load type classification, and improving a fuzzy clustering algorithm by adopting genetic simulated annealing;
and 2.3, calculating the similarity among the substations through fuzzy clustering, and converting the similarity among different substation load compositions into the magnitude of the membership degree for expression, wherein the membership degree is as follows:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the jth initial cluster center and,is the kth initial clustering center, K is the number of clustering centers,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.4, constructing a fuzzy similar matrix according to the distance matrix, randomly selecting c clustering centers through a genetic algorithm, initializing control parameters, initializing a population, calculating the membership and fitness of each individual according to the distance matrix, performing selection, crossing, mutation and other operations through the genetic algorithm to generate a new population, and replacing or receiving the old individual through a simulated annealing algorithm; wherein, the clustering center is:
in the formula (I), the compound is shown in the specification,is the degree of membership of the ith substation to the jth substation,is the eigenvector of the ith substation, and m is the characteristic dimension of the substation;
step 2.5, target functionThe distance square sum of each substation load and the corresponding clustering center is formed and then summed, and an objective function is outputA minimum cluster center and membership matrix, wherein the objective function is:
in the formula (I), the compound is shown in the specification,each substation load constitutes the sum of squared distances from its corresponding cluster center,is the degree of membership of the ith substation to the jth substation,is the feature vector of the ith substation,is the jth initial clustering center, and K is the number of clustering centers;
and 2.6, according to the maximum membership principle of the substation membership matrix and the clustering center, taking the membership degree between the substation and the clustering center as a basis for selecting a typical site, and carrying out load classification on the substation.
4. The method for selecting the substation with the typical load characteristics based on the multi-source data according to claim 1, wherein the method comprises the following steps: in the step 3, the method specifically includes:
step 3.1, selecting industry components in the industrial load as feature vectors;
in the formula (I), the compound is shown in the specification,the method is characterized by comprising the following steps that (1) the industrial composition in the ith transformer substation industrial load is realized, p is the category of the transformer substation industrial load, and n is the number of the transformer substations;
step 3.2, performing data dimension reduction on the industry composition data in the industrial load through multidimensional scaling MDS, and mapping data point pairs with equal distances in a high-dimensional space in a low-dimensional space;
in the formula (I), the compound is shown in the specification,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the x-axis vector after the dimension reduction,the y-axis vector after dimensionality reduction, and n is the number of the transformer substations;
3.3, clustering the industry composition data subjected to dimensionality reduction by adopting a genetic simulated annealing improved fuzzy clustering algorithm;
step 3.4, outputting the objective functionIs the smallest cluster center and membership matrix, whose objective function is:
in the formula (I), the compound is shown in the specification,the distance square sum of each transformer substation industry composition and the corresponding clustering center is summed,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the jth clustering center, n is the number of the transformer substations, and T is the number of the clustering centers;
the membership matrix is:
in the formula (I), the compound is shown in the specification,is the membership degree formed by the i-th substation and the j-th substation industry,is a characteristic vector formed by industries in the i-th substation industrial load after dimensionality reduction,is the jth clusterThe center of the device is provided with a central hole,is the T-th clustering center, and T is the number of the clustering centers;
the clustering center is as follows:
in the formula (I), the compound is shown in the specification,is the j-th cluster center and,is the membership degree formed by the i-th substation and the j-th substation industry,the characteristic vector is formed by industries in the i-th transformer substation industrial load after dimensionality reduction, and n is the number of the transformer substations;
and 3.5, classifying the industry composition membership matrix and the corresponding clustering center matrix according to the maximum membership principle.
5. The method for selecting the substation with the typical load characteristics based on the multi-source data according to claim 1, wherein the method comprises the following steps: in the step 4, for the transformer substation groups obtained by load classification, the transformer substation closest to the corresponding clustering center is selected from each group as a typical load characteristic transformer substation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010822559.8A CN111737924B (en) | 2020-08-17 | 2020-08-17 | Method for selecting typical load characteristic transformer substation based on multi-source data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010822559.8A CN111737924B (en) | 2020-08-17 | 2020-08-17 | Method for selecting typical load characteristic transformer substation based on multi-source data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111737924A true CN111737924A (en) | 2020-10-02 |
CN111737924B CN111737924B (en) | 2021-03-02 |
Family
ID=72658496
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010822559.8A Active CN111737924B (en) | 2020-08-17 | 2020-08-17 | Method for selecting typical load characteristic transformer substation based on multi-source data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111737924B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113344073A (en) * | 2021-06-02 | 2021-09-03 | 云南电网有限责任公司电力科学研究院 | Daily load curve clustering method and system based on fusion evolution algorithm |
CN113822511A (en) * | 2020-11-16 | 2021-12-21 | 全球能源互联网研究院有限公司 | Multistage power quality assessment method and device based on fuzzy clustering analysis |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102999876A (en) * | 2012-11-16 | 2013-03-27 | 中国电力科学研究院 | Selection method for typical load characteristic substations |
CN103279803A (en) * | 2013-04-27 | 2013-09-04 | 深圳供电局有限公司 | Load modeling method and system based on comprehensive information theory and modern interior point theory |
CN103646354A (en) * | 2013-11-28 | 2014-03-19 | 国家电网公司 | Effective index FCM and RBF neural network-based substation load characteristic categorization method |
CN104268402A (en) * | 2014-09-25 | 2015-01-07 | 国家电网公司 | Power system load clustering method based on fuzzy c-means algorithm |
CN104299115A (en) * | 2014-11-11 | 2015-01-21 | 国网重庆市电力公司电力科学研究院 | Intelligent substation secondary system state analysis method based on fuzzy C-mean clustering algorithm |
US20160092611A1 (en) * | 2014-09-26 | 2016-03-31 | State Grid Corporation Of China | Method for constructing real-time solar irradiation metering network of gigawatts level photovoltaic power generation base |
-
2020
- 2020-08-17 CN CN202010822559.8A patent/CN111737924B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102999876A (en) * | 2012-11-16 | 2013-03-27 | 中国电力科学研究院 | Selection method for typical load characteristic substations |
CN103279803A (en) * | 2013-04-27 | 2013-09-04 | 深圳供电局有限公司 | Load modeling method and system based on comprehensive information theory and modern interior point theory |
CN103646354A (en) * | 2013-11-28 | 2014-03-19 | 国家电网公司 | Effective index FCM and RBF neural network-based substation load characteristic categorization method |
CN104268402A (en) * | 2014-09-25 | 2015-01-07 | 国家电网公司 | Power system load clustering method based on fuzzy c-means algorithm |
US20160092611A1 (en) * | 2014-09-26 | 2016-03-31 | State Grid Corporation Of China | Method for constructing real-time solar irradiation metering network of gigawatts level photovoltaic power generation base |
CN104299115A (en) * | 2014-11-11 | 2015-01-21 | 国网重庆市电力公司电力科学研究院 | Intelligent substation secondary system state analysis method based on fuzzy C-mean clustering algorithm |
Non-Patent Citations (2)
Title |
---|
周开乐等: "基于改进模糊C 均值算法的电力负荷特性分类", 《电力系统保护与控制》 * |
王帅: "基于数据挖掘和GSA-BP多模型神经网络的微网短期负荷预测", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113822511A (en) * | 2020-11-16 | 2021-12-21 | 全球能源互联网研究院有限公司 | Multistage power quality assessment method and device based on fuzzy clustering analysis |
CN113822511B (en) * | 2020-11-16 | 2023-10-20 | 全球能源互联网研究院有限公司 | Multistage electric energy quality assessment method and device based on fuzzy cluster analysis |
CN113344073A (en) * | 2021-06-02 | 2021-09-03 | 云南电网有限责任公司电力科学研究院 | Daily load curve clustering method and system based on fusion evolution algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN111737924B (en) | 2021-03-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111199016B (en) | Daily load curve clustering method for improving K-means based on DTW | |
CN106845717B (en) | Energy efficiency evaluation method based on multi-model fusion strategy | |
CN110991786B (en) | 10kV static load model parameter identification method based on similar daily load curve | |
CN106203867A (en) | Grid division methods based on power distribution network assessment indicator system and cluster analysis | |
CN107301604A (en) | Multi-model fusion estimation system | |
CN109902953A (en) | A kind of classification of power customers method based on adaptive population cluster | |
CN111737924B (en) | Method for selecting typical load characteristic transformer substation based on multi-source data | |
CN106503919A (en) | A kind of power distribution network evaluation methodology based on power supply zone characteristic | |
CN111461921B (en) | Load modeling typical user database updating method based on machine learning | |
CN108694470A (en) | A kind of data predication method and device based on artificial intelligence | |
CN115829105A (en) | Photovoltaic power prediction method based on historical data feature search | |
CN110738232A (en) | grid voltage out-of-limit cause diagnosis method based on data mining technology | |
CN103869102B (en) | A kind of large regional power grid load statistics and sorting technique | |
CN111428766B (en) | Power consumption mode classification method for high-dimensional mass measurement data | |
CN111553568A (en) | Line loss management method based on data mining technology | |
Haghdadi et al. | Clustering-based optimal sizing and siting of photovoltaic power plant in distribution network | |
CN110991638B (en) | Generalized load modeling method based on clustering and neural network | |
Wang et al. | Application of clustering technique to electricity customer classification for load forecasting | |
CN110348604A (en) | A kind of linear regression power predicating method and system based on electricity consumption Specialty aggregation | |
CN114372835B (en) | Comprehensive energy service potential customer identification method, system and computer equipment | |
CN113392877B (en) | Daily load curve clustering method based on ant colony algorithm and C-K algorithm | |
CN112464168B (en) | Comprehensive energy potential user targeting evaluation and extraction method | |
CN110852628A (en) | Rural medium and long term load prediction method considering development mode influence | |
Bai et al. | A novel improved approach for fast and accurate load clustering in power system | |
He et al. | Method for determining comprehensive weight vector based on multiple linear fitting |
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 |