Method and system for predicting structural form of power distribution network

Abstract

The invention discloses a power distribution network structure form prediction method and system. The method comprises the following steps: acquiring first power data; analyzing the first power data to obtain a first power distribution network feature and a second power distribution network feature; classifying the power distribution network region, and predicting a future first feature of the power distribution network by using different prediction models according to a classification result; and generating different region planning suggestions of the power distribution network based on the second feature of the power distribution network and the future first feature of the power distribution network. According to the method, the characteristics of the power distribution network are obtained through multi-dimensional analysis, the power distribution network regions are classified, and specific prediction models are used for different types of regions, so that the prediction accuracy and pertinence are improved; comprehensive, scientific and targeted planning schemes are provided for different regions of the power distribution network, the operation efficiency, reliability and adaptability of the power distribution network are effectively improved, and reasonable planning of the power distribution network is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution networks, and in particular, to a method and system for predicting the structural form of a distribution network. Background Art

[0002] With the in-depth adjustment of the global energy pattern and the booming development of local economies, the stability and efficiency of energy supply have become key issues. The development status of the distribution network is directly related to the stable operation of the local economic society and the smooth progress of the energy transformation. At present, the widespread access of distributed energy sources such as solar energy and wind energy in traditional distribution network areas, as well as the rapid growth of new loads such as electric vehicle charging demands, have made the operating environment of the distribution network increasingly complex. At the same time, the acceleration of new urbanization construction has put forward higher requirements for the power supply capacity, reliability and intelligent level of the distribution network. Therefore, in-depth research on the evolution mechanism of the distribution network structural form and accurate prediction of its development trend have become one of the core tasks of sustainable development.

[0003] However, there are many limitations in the existing research on distribution network planning and development. On the one hand, most studies only analyze the development of the distribution network from a single dimension, such as only focusing on the load growth trend in the time dimension, or only focusing on the grid layout optimization in the space dimension, but ignoring the mutual connection and synergy among multi-dimensional factors such as time, space, and resources. This makes it difficult to comprehensively grasp the evolution law of the distribution network structural form, and the planning scheme cannot adapt to the complex and changeable actual situation. On the other hand, the existing technologies lack sufficient consideration of the unique characteristics of the distribution network. There are problems such as uneven economic development levels, uneven resource distribution, and large geographical environment differences in some areas, but the traditional research methods have not effectively developed personalized planning strategies for these differences. In addition, in the face of the uncertainties brought by the access of new loads and distributed energy sources, the accuracy and reliability of the existing prediction models are insufficient, and they cannot provide strong support for the forward-looking planning of the distribution network. These problems seriously restrict the scientific planning and high-quality development of the distribution network, and urgently need innovative theories and methods to solve. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method and system for predicting the structural form of a distribution network to solve the problems that the existing research on distribution network planning and development lacks multi-dimensional comprehensive consideration, is difficult to grasp the evolution law of the distribution network structural form, and has insufficient prediction accuracy and cannot provide strong support for the scientific planning of the distribution network.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a method for predicting the structural form of a distribution network, including:

[0008] Obtain the first power data;

[0009] Analyze the first power data to obtain the first characteristics and the second characteristics of the distribution network;

[0010] Classify the distribution network area, and according to the classification results, use different prediction models to predict the first future characteristics of the distribution network;

[0011] Generate planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network.

[0012] As a preferred scheme of the distribution network structure form prediction method described in the present invention, wherein: analyzing the first power data to obtain the first characteristics of the distribution network includes:

[0013] Preprocess the first power data, and analyze the preprocessed first power data from three dimensions of time, space, and resources to obtain the first characteristics of the distribution network.

[0014] As a preferred scheme of the distribution network structure form prediction method described in the present invention, wherein: it further includes:

[0015] Based on the first characteristics of the distribution network, use a Markov chain with multi-dimensional feature weights to obtain the state transition process of the first characteristics of the distribution network;

[0016] According to the state transition process of the first characteristics of the distribution network, establish an evolution equation of the distribution network to obtain the second characteristics of the distribution network.

[0017] As a preferred scheme of the distribution network structure form prediction method described in the present invention, wherein: generating planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network includes:

[0018] Set optimization indicators, and generate an optimization objective function according to the optimization indicators;

[0019] Based on the second characteristics of the distribution network and the first future characteristics of the distribution network, use an intelligent algorithm to solve the optimization objective function to obtain planning suggestions for different regions of the distribution network.

[0020] As a preferred scheme of the distribution network structure form prediction method described in the present invention, wherein: classifying the distribution network area includes:

[0021] Adopt an improved fuzzy C-means clustering algorithm to divide the distribution network area into urban suburban type, industrial park type, agricultural rural type, and tourist scenic area type.

[0022] As a preferred solution of the distribution network structure form prediction method described in the present invention, wherein: according to the classification result, different prediction models are used to predict the first future feature of the distribution network, including:

[0023] Using a hybrid model dominated by LSTM to predict the first future feature of the distribution network in the suburban area of the city;

[0024] Using a hybrid model dominated by GCN to predict the first future feature of the distribution network in the industrial park type;

[0025] Using a hybrid model mainly based on seasonal prediction to predict the first future feature of the distribution network in the rural and agricultural type;

[0026] Using a hybrid model mainly based on fluctuation prediction to predict the first future feature of the distribution network in the tourist scenic area type.

[0027] As a preferred solution of the distribution network structure form prediction method described in the present invention, wherein: the optimization objective function is expressed as:

[0028] minF(x) = [f1(x), f2(x),..., f m (x)]

[0029] Wherein, F(x) is the objective function, and f1(x), f2(x), f m (x) are optimization indicators;

[0030] Based on the result of the objective function, generate distribution network planning suggestions for different regions including three levels: grid structure optimization, equipment configuration, and operation control.

[0031] In a second aspect, the present invention provides a distribution network structure form prediction system, including:

[0032] A data acquisition module for acquiring the first power data;

[0033] A feature acquisition module for analyzing the first power data to obtain the first distribution network feature and the second distribution network feature;

[0034] A division prediction module for classifying the distribution network area, and according to the classification result, using different prediction models to predict the first future feature of the distribution network;

[0035] An optimization suggestion module for generating distribution network planning suggestions for different regions based on the second distribution network feature and the first future feature of the distribution network.

[0036] In a third aspect, the present invention provides a computing device, including:

[0037] A memory and a processor;

[0038] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the distribution network structure form prediction method are implemented.

[0039] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the distribution network structure form prediction method.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention analyzes and obtains the characteristics of the distribution network from the dimensions of time, space, and resources, comprehensively grasps the current situation of the distribution network, obtains the second feature based on the first feature, and more deeply understands the evolution law of the distribution network; uses an improved fuzzy C-means clustering algorithm to classify the distribution network regions, and uses specific prediction models for different types of regions, improving the accuracy and pertinence of the prediction; sets optimization indicators to generate an objective function, which can provide a comprehensive, scientific, and targeted planning scheme for different regions of the distribution network, effectively improving the operation efficiency, reliability, and adaptability of the distribution network, and promoting the reasonable planning of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0042] Figure 1 It is a schematic diagram of the overall process logic of the distribution network structure form prediction method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] Embodiment 1

[0045] Referring to Figure 1 , which is an embodiment of the present invention, a distribution network structure form prediction method is provided, including:

[0046] S100: Obtain the first power data;

[0047] S102: Analyze the first power data to obtain the first characteristics and the second characteristics of the distribution network;

[0048] S104: Classify the distribution network areas. According to the classification results, use different prediction models to predict the first future characteristics of the distribution network;

[0049] S106: Generate planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network.

[0050] It should be noted that by obtaining the first power data and analyzing the characteristics of the distribution network from the time, space, and resource dimensions, the current situation of the distribution network can be comprehensively grasped. Based on this, the second characteristics are obtained by using a Markov chain with multi-dimensional feature weights and establishing an evolution equation, so as to more deeply understand the evolution law of the distribution network. The improved fuzzy C-means clustering algorithm is used to classify the distribution network areas, and specific prediction models are used for different types of areas, significantly improving the accuracy and pertinence of the prediction. The objective function is generated by setting optimization indicators, and the planning suggestions obtained by using intelligent algorithms cover three levels: grid structure optimization, equipment configuration, and operation control, which can provide comprehensive, scientific, and targeted planning schemes for different regions of the distribution network, effectively improving the operation efficiency, reliability, and adaptability of the distribution network, and promoting the reasonable planning of the distribution network.

[0051] In the embodiment of the present application, the above step S102 includes the following sub-steps A1-A2;

[0052] In A1: Preprocess the first power data, and analyze the preprocessed first power data from the three dimensions of time, space, and resources to obtain the first characteristics of the distribution network.

[0053] In an alternative embodiment, the first power data may be historical load data, economic growth data, population change data, geographic information data, distribution network topology data, load distribution data, power source distribution data, renewable energy resource data, land resource data, etc.

[0054] In an alternative embodiment, the preprocessing operations include data cleaning and data standardization;

[0055] Exemplarily, for the detection and processing of outliers in the first power data, this embodiment uses the Z-score method to detect outliers, and the calculation formula is as follows:

[0056]

[0057] Where X is the value of the data point; μ is the average value of the data set; σ is the standard deviation of the data set.

[0058] Exemplarily, for the detection and processing of missing values in the first power data, this embodiment uses the time series interpolation method for linear interpolation. The linear interpolation steps are as follows:

[0059] Suppose there are two known time points and corresponding data values: (t1, y1) and (t2, y2), and it is necessary to estimate the value y at time t, where t1 ≤ t ≤ t2. Then there is:

[0060]

[0061] Exemplarily, perform min-max standardization processing on various types of data to unify the numerical range to the interval [0, 1]. The formula for min-max standardization is as follows:

[0062]

[0063] Among them, X is the value of the original data point, X min is the minimum value in the dataset. X max is the maximum value in the dataset, and X norm is the value of the standardized data point.

[0064] In the embodiment of the present application, the first distribution network feature includes the distribution network features and correlation analysis results obtained by analyzing through three dimensions of time, space, and resources.

[0065] Analyzing from the three dimensions of time, space, and resources, the analysis methods for obtaining the first distribution network feature can be time series decomposition method, spatial autocorrelation analysis method, geographically weighted regression method, comprehensive resource index method, data envelopment analysis method, and can also include wavelet analysis method, grey relational analysis method, principal component analysis method, clustering analysis method, analytic hierarchy process method, etc.

[0066] Exemplarily, the process of obtaining the first distribution network feature by analyzing from the three dimensions of time, space, and resources includes:

[0067] ① Analyze from the time dimension

[0068] Construct a time series decomposition model to decompose the distribution network load and structure features into a trend term, a seasonal term, and a random term:

[0069] Y t = T t + S t + R t

[0070] Among them, Y t is the observed value at time t, T t is the trend term, S t is the seasonal term, and R t is the random term.

[0071] The STL (Seasonal and Trend decomposition using Loess) method is used for decomposition. The local weighted regression scatterplot smoothing method is used to extract the trend term; the detrended data is used for seasonal decomposition of the seasonal term; the random term is calculated by subtracting the trend term and the seasonal term from the original data.

[0072] ②Conduct factor analysis from the spatial dimension

[0073] The global Moran's I and local Moran's I are used to evaluate the spatial correlation of the characteristics of the distribution network from the spatial dimension.

[0074] The global Moran's I is an important statistic for measuring spatial autocorrelation. It describes the average degree of association between all spatial units in the entire region and the surrounding areas. Its calculation formula is as follows:

[0075]

[0076] Among them, n is the total number of spatial units; w ij is the spatial weight, which reflects the mutual influence between spatial units; y i , y j represents the attribute values of the i-th and j-th spatial units; is the mean of the attribute values of all spatial units; S o is the sum of the spatial weight matrix.

[0077] The value range of Moran's I is between [-1, 1]. A value close to 1 indicates positive spatial autocorrelation, that is, similar values tend to cluster together; a value close to -1 indicates negative spatial autocorrelation, that is, similar values tend to be far away from each other; a value close to 0 indicates that the attribute values of spatial units are randomly distributed in space without obvious spatial autocorrelation.

[0078] Spatial heterogeneity analysis: The Geographically Weighted Regression (GWR) model is used to analyze the influence of different geographical locations on the structure and form of the distribution network. The Geographically Weighted Regression model is a local spatial regression analysis method that allows model parameters to vary with geographical location, thereby being able to capture the non-stationarity of spatial data. The basic form of the GWR model can be expressed as:

[0079]

[0080] Among them, y i is the dependent variable; x ik is the value of the k-th independent variable at the i-th observation point; β k (u i , vi ) is the regression coefficient of the k-th independent variable at the geographical location (u i , v i ).

[0081] The core of the GWR model is the spatial weight matrix, which determines how the regression coefficient of each observation point is affected by neighboring observation points.

[0082] ③Conduct factor analysis from the resource dimension

[0083] Factor analysis of the resource dimension is an important method for evaluating the resource status and utilization efficiency of a region. The resource endowment evaluation method is to construct a comprehensive resource index (CRI) to evaluate the resource endowment of each distribution network fishing area; CRI is calculated by combining the weighted sum of different resource indices, expressed as:

[0084] CRI = ω1 * ERI + ω2 * RRI + ω3 * LRI

[0085] Among them, ERI is the electric power resource index, which measures the availability and reliability of electric power resources; RRI is the renewable energy resource index, which reflects the richness and development potential of renewable energy resources; LRI is the land resource index, which evaluates the quality and availability of land resources; ω1, ω2, ω3 are the corresponding weights, which are allocated according to the importance of each resource index.

[0086] The analysis of resource utilization efficiency uses the data envelopment analysis (DEA) method to evaluate the resource utilization efficiency of the regional distribution network. DEA is a non-parametric method used to evaluate the relative efficiency in the production process. The efficiency evaluation index of the DEA model is defined as:

[0087]

[0088] Among them, θ is the efficiency value, which represents the efficiency of resource utilization; e is a non-Archimedean infinitesimal, which is used to ensure the accuracy of the efficiency value; is the input slack variable, which represents the redundancy of the i-th input; is the output slack variable, which represents the shortage of the r-th output.

[0089] It should be noted that obtaining the first characteristic of the distribution network provides basic data for analyzing the dynamic evolution mechanism of the second characteristic of the distribution network.

[0090] In A2: Based on the first characteristic of the distribution network, use the Markov chain with multi-dimensional feature weights to obtain the state transition process of the first characteristic of the distribution network; according to the state transition process of the first characteristic of the distribution network, establish an evolution equation of the distribution network to obtain the second characteristic of the distribution network.

[0091] In this embodiment, the second characteristics of the distribution network include the distribution network structure form characteristics, the evolution law of the distribution network structure form, and the critical transformation characteristics.

[0092] Exemplarily, obtaining the first characteristic state transition process of the distribution network by using a Markov chain with multi-dimensional characteristic weights includes:

[0093] Define the state transition matrix P:

[0094] P = [p ij

[0095] where p ij represents the probability of transitioning from state i to state j, satisfying ∑ j P ij = 1.

[0096] Through the state transition matrix P, obtain the state probability vector π(t) expressed as:

[0097] π(t + 1) = π(t) * P

[0098] This equation indicates that the state probability at the next time step is the product of the current state probability and the state transition matrix.

[0099] Exemplarily, according to the first characteristic state transition process of the distribution network, establish an evolution equation of the distribution network to obtain the second characteristics of the distribution network, including:

[0100] In this embodiment, construct an evolution dynamics equation based on the Lotka-Volterra equation to describe the competition and cooperation relationship between different types of distribution network structures. The evolution equation of the proportion xi of the i-th structure type is:

[0101]

[0102] where r i is the natural growth rate of the i-th structure type; K i is the environmental capacity, indicating the maximum proportion that the system can support; a ij is the interaction coefficient between types i and j.

[0103] Through the state transition process and the evolution equation, obtain the distribution network structure form characteristics;

[0104] In this embodiment, use the early warning signal method to obtain the critical transformation point of the distribution network structure form. Among them, the main indicators include the autocorrelation coefficient (AR1) and the coefficient of variation (CV), expressed as:

[0105]

[0106] where z t is the time series data;​ is the mean of the time series; σ is the standard deviation of the time series, and μ is the mean of the time series.

[0107] It should be noted that the dynamic evolution mechanism studies the change of the distribution network structure form over time. In this step, multi-dimensional feature weights are introduced on the basis of the traditional Markov chain, and an improved state transition probability calculation method is proposed. This helps to understand how the distribution network structure responds to various internal and external factors and predict its future development trend; at the same time, the evolution law and critical transformation characteristics of the distribution network structure form are obtained, providing a theoretical basis for differential planning.

[0108] In the embodiment of the present application, the above step S104 includes the following steps B1-B2:

[0109] In B1: The improved fuzzy C-means clustering algorithm is used to divide the distribution network area into suburban type, industrial park type, rural agricultural type and tourist scenic area type.

[0110] In the embodiment of the present application, based on multi-dimensional factors such as economic development level, population density, geographical location, etc., the fuzzy C-means clustering algorithm is used to classify the distribution network area. The fuzzy C-means clustering algorithm is as follows:

[0111]

[0112] Among them, is the membership degree of sample i to cluster j, m is the fuzzy index, usually taking a value greater than 1 to ensure that the distribution of membership degrees is more dispersed, x i is the sample point, c j is the cluster center.

[0113] In B2: The LSTM-dominated hybrid model is used to predict the first future feature of the distribution network in the suburban type; the GCN-dominated hybrid model is used to predict the first future feature of the distribution network in the industrial park type; the hybrid model mainly based on seasonal prediction is used to predict the first future feature of the distribution network in the rural agricultural type; the hybrid model mainly based on fluctuation prediction is used to predict the first future feature of the distribution network in the tourist scenic area type.

[0114] For different regions, personalized prediction of the distribution network structure form is carried out, and the prediction function is:

[0115] Y k = f k (X k , θ k )

[0116] Among them, Y k is the prediction result of the kth type of distribution network area, f k is the corresponding prediction function, X kis the input feature, θ k are the model parameters.

[0117] Exemplarily, an LSTM model is adopted to process the time - series data, electric - vehicle charging load data, and distributed photovoltaic power generation data in the first power data. A time attention layer is added to optimize the structure of the gating unit to obtain the time - series prediction value. A graph convolutional network (GCN) model is used to process the distribution network topology structure information, electric - vehicle charging load data, and distributed photovoltaic power generation data in the first power data. A spatial attention mechanism is added to optimize the design of the graph convolutional layer to obtain the prediction value of the distribution network node features. The outputs of the GCN and LSTM models are combined. For different regions, a dynamic weight fusion method based on the attention mechanism is introduced to obtain a hybrid model to obtain more accurate prediction results.

[0118] Exemplarily, the process of obtaining the electric - vehicle charging load data includes:

[0119] The Monte Carlo simulation method is used to generate the electric - vehicle charging load curve and the single - vehicle charging power curve:

[0120] P(t)=P max (1 - e -t / τ )

[0121] where P max is the maximum charging power, and τ is the charging time constant.

[0122] Considering the influence of different charging strategies (such as valley - price charging, random charging, etc.), a charging load model is constructed, which involves simulating the charging behaviors of a large number of electric vehicles and aggregating their charging power curves to predict the overall charging load.

[0123] Exemplarily, the process of distributed photovoltaic power generation data includes:

[0124] Simulating the daily sunlight intensity change based on the beta distribution:

[0125]

[0126] where x is the normalized time (0≤x≤1), α and β are shape parameters, and B(α,β) is the beta function.

[0127] Photovoltaic power calculation:

[0128] P = η*A*R*f(x;α,β)

[0129] where η is the photovoltaic conversion efficiency, A is the panel area, and R is the irradiance intensity.

[0130] The Gaussian mixture method is used to aggregate different types of loads:

[0131]

[0132] Among them, K is the number of mixed components, and w i is the weight of the i-th component, and N(x|μ i ,∑i) is a multi-dimensional Gaussian distribution, where μ i is the mean vector and ∑i is the covariance matrix.

[0133] In the embodiment of the present application, the above step S106 includes the following steps C1 - C2:

[0134] In C1: Set the optimization index, and generate an optimization objective function according to the optimization index; based on the second feature of the distribution network and the first future feature of the distribution network, use an intelligent algorithm to solve the optimization objective function to obtain the planning suggestions for different regions of the distribution network.

[0135] In an alternative embodiment, the intelligent algorithm can be a genetic algorithm, a multi-objective optimization method, a particle swarm optimization algorithm, a simulated annealing algorithm, etc.

[0136] In the embodiment of the present application, using the multi-objective optimization method to solve the optimization objective function is expressed as:

[0137] minF(x)=[f1(x),f2(x),...,f m (x)]

[0138] s.t.g i (x)≤0,i=1,2,...,p

[0139] h j (x)=0,j=1,2,...,q

[0140] x l ≤x≤x u

[0141] Among them, F(x) is the objective function, f1(x), f2(x), f m (x) are the optimization indices, g i (x) and h j (x) are the inequality and equality constraints respectively, and x l ,x u are the lower and upper bounds of the decision variable x respectively.

[0142] In an alternative embodiment, the optimization indices include indices such as economy, reliability, and environmental friendliness.

[0143] In an alternative embodiment, based on the results of the objective function, generate planning suggestions for different regions of the distribution network including three levels: grid structure optimization, equipment configuration, and operation control;

[0144] Exemplarily, in terms of the optimization of the grid structure, a renovation plan for the weak links after the "N-1" verification is proposed for the reinforcement of the backbone network, including specific measures such as building 2 new 110 kV substations, expanding 3 35 kV substations, and building 25 km of new 110 kV lines. In terms of the planning of the connection channels, a "hand-in-hand" ring network structure is designed based on the reliability requirements to enable the 10 kV lines in adjacent areas to be used as backups for each other. For the optimization of the power source layout, considering the characteristics of new energy access, 50 MW of distributed photovoltaic power and 30 MW of wind farms are added in the load center area.

[0145] In terms of the equipment configuration suggestions, the transformer capacity configuration adopts a hierarchical configuration strategy. The main transformer selects the S9-M-315 kVA model, and the distribution transformer selects the S11-M-100 kVA and other model specifications. For the layout of the switchgear, intelligent ring network cabinets are arranged at key nodes, vacuum circuit breakers with measurement and control functions are set, and protection devices such as lightning arresters and fuses are equipped. The reactive power compensation device configuration adopts a combination of local compensation and centralized compensation, and capacitor banks with ratings ranging from 10 kVar to 50 kVar are installed in each distribution substation.

[0146] In terms of the operation control suggestions, the voltage control strategy adopts a sub-region combined voltage regulation scheme, and the voltage qualification rate of over 95% is achieved by adjusting the taps of on-load tap-changing transformers and the switching of reactive power compensation devices; the load transfer plan designs an intelligent decision-making mechanism of "pre-plan library + expert system", which can quickly realize load transfer in case of equipment maintenance and faults; the emergency response plan formulates a standardized disposal process for typical scenarios such as natural disasters and equipment failures.

[0147] In C2: Evaluate and optimize the four hybrid models.

[0148] In the embodiment of the present application, the mean absolute percentage error, root mean square error, and coefficient of determination are used to evaluate and optimize the model.

[0149] The above is a schematic solution of a method for predicting the structure form of a distribution network in this embodiment. It should be noted that the technical solution of the system for predicting the structure form of the distribution network belongs to the same concept as the technical solution of the above method for predicting the structure form of the distribution network. For the details not described in the technical solution of the system for predicting the structure form of the distribution network in this embodiment, reference can be made to the description of the technical solution of the method for predicting the structure form of the distribution network above.

[0150] The system for predicting the structure form of the distribution network in this embodiment includes:

[0151] A data acquisition module for acquiring the first power data;

[0152] A feature acquisition module for analyzing the first power data to obtain the first feature and the second feature of the distribution network;

[0153] A division prediction module, which is used to classify the distribution network area. According to the classification results, different prediction models are used to predict the first future feature of the distribution network.

[0154] An optimization suggestion module, which is used to generate planning suggestions for different regions of the distribution network based on the second feature of the distribution network and the first future feature of the distribution network.

[0155] This embodiment also provides a computing device, which is applicable to the situation of predicting the structure form of the distribution network, including:

[0156] A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for predicting the structure form of the distribution network as proposed in the above embodiment.

[0157] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by the processor, it implements the method for predicting the structure form of the distribution network as proposed in the above embodiment.

[0158] The storage medium proposed in this embodiment and the method for predicting the structure form of the distribution network proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0159] Through the above description of the implementation manners, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course, it can also be implemented by hardware. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes contributions to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FLASH), a hard disk, or an optical disc of a computer, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0160] It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solution of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solution of the present invention, and it should be covered by the scope of the claims of the present invention.

[0161] Embodiment 2

[0162] Referring to Tables 1-2, based on the previous embodiment, this embodiment provides an application case of a method for constructing a probability prediction model for large power users' load to illustrate the feasibility and beneficial effects of our solution.

[0163] Taking a certain county in Guizhou Province as an example for experimental verification, the comparative experimental results between the method of the present invention and traditional methods are as follows:

[0164] Table 1 Comparison of prediction accuracies in different types of distribution network regions

[0165]

[0166]

[0167] Table 2 Comparison of planning effects of different methods

[0168]

[0169] The experimental results show that, compared with traditional methods, the method of the present invention has significant advantages in both prediction accuracy and planning effect. The prediction accuracy of the method of the present invention is increased by an average of 43.6%, the power supply reliability is increased by 0.19 percentage points, the voltage qualification rate is increased by 2.41 percentage points, the line loss rate is reduced by 17.52%, the equipment utilization rate is increased by 11.12 percentage points, the economic benefit is increased by 23.43%, and the adaptability score of the planning scheme is increased by 14.2 points.

Claims

1. A prediction method for the structural form of a distribution network, characterized in that, Including: Obtain the first power data; Analyze the first power data to obtain the first characteristics and the second characteristics of the distribution network; Classify the distribution network regions. According to the classification results, use different prediction models to predict the first future characteristics of the distribution network; Generate planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network.

2. The method for predicting the structural form of a distribution network according to claim 1, wherein Analyze the first power data to obtain the first characteristics of the distribution network, including: Preprocess the first power data, and analyze the preprocessed first power data from three dimensions of time, space, and resources to obtain the first characteristics of the distribution network.

3. The method for predicting the structural form of a distribution network according to claim 2, wherein It also includes: Based on the first characteristics of the distribution network, use a Markov chain with multi-dimensional feature weights to obtain the state transition process of the first characteristics of the distribution network; According to the state transition process of the first characteristics of the distribution network, establish an evolution equation of the distribution network to obtain the second characteristics of the distribution network.

4. The method for predicting the structural form of a distribution network according to claim 3, wherein, Generate planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network, including: Set optimization indicators, and generate an optimization objective function according to the optimization indicators; Based on the second characteristics of the distribution network and the first future characteristics of the distribution network, use an intelligent algorithm to solve the optimization objective function to obtain planning suggestions for different regions of the distribution network.

5. The method for predicting the structural form of a distribution network according to claim 4, wherein Classify the distribution network regions, including: Adopt an improved fuzzy C-means clustering algorithm to divide the distribution network regions into urban suburban type, industrial park type, agricultural rural type, and tourist scenic area type.

6. The method for predicting the structural form of a distribution network according to claim 5, characterized in that, According to the classification results, use different prediction models to predict the first future characteristics of the distribution network, including: Adopt a hybrid model dominated by LSTM to predict the first future characteristics of the distribution network in the urban suburban type; Adopt a hybrid model dominated by GCN to predict the first future characteristics of the distribution network in the industrial park type; Adopt a hybrid model mainly based on seasonal prediction to predict the first future characteristics of the distribution network in the agricultural rural type; Adopt a hybrid model mainly based on volatility prediction to predict the first future characteristics of the distribution network in the tourist scenic area type.

7. The method for predicting the structural form of a distribution network according to claim 4, characterized in that, The optimization objective function is expressed as: minF(x) = [f1(x), f2(x),..., f m (x)] Among them, F(x) is the objective function, and f1(x), f2(x), f m (x) are the optimization metrics; Based on the results of the objective function, generate planning suggestions for different regions of the distribution network including three levels: grid structure optimization, equipment configuration, and operation control.

8. A system applying the distribution network structure form prediction method as described in any one of claims 1 - 7, characterized in that, Including: A data acquisition module for obtaining the first power data; A feature acquisition module for analyzing the first power data to obtain the first characteristics and the second characteristics of the distribution network; A division and prediction module for classifying the distribution network regions. According to the classification results, use different prediction models to predict the first future characteristics of the distribution network; An optimization suggestion module for generating planning suggestions for different regions of the distribution network based on the second characteristics of the distribution network and the first future characteristics of the distribution network.

9. An electronic device, including: A memory and a processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the distribution network structure form prediction method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium stores computer-executable instructions that, when executed by a processor, implement the steps of the distribution network structure form prediction method according to any one of claims 1 to 7.

Images