Power distribution network frame multi-stage planning method considering multi-type source load evolution development

Abstract

The invention provides a power distribution network frame multi-stage planning method considering multi-type source load evolution development, and the method comprises the steps: analyzing the spatial-temporal evolution characteristics of a multi-type distributed power supply and a multi-type load based on the source load data of a to-be-planned region, building a multi-type source load evolution development prediction model, and carrying out the prediction of the multi-type source load evolution development; generating a source load typical scene oriented to a plurality of continuous planning level years; based on the typical source load scene of each planning level year, obtaining a multi-stage weak link evaluation result; a power distribution network frame multi-stage planning model is further constructed and solved, and a power distribution network frame multi-stage planning scheme is obtained; wherein the construction of the power distribution network frame multi-stage planning model comprises the following steps: establishing a multi-objective function, and taking the total planning cost, the voltage deviation and the line load rate as optimization objectives; constructing constraint conditions including an operation constraint, a network topology constraint and a multi-stage cooperative constraint; and topology continuity and evolution rationality of the power distribution network frame in multi-stage planning are realized.

Description

Technical Field

[0001] This invention belongs to the field of power system planning technology, specifically relating to a multi-stage planning method for distribution network structure that considers the evolution and development of multiple types of source loads. Background Technology

[0002] The power distribution network is gradually transforming from a simple power network that receives and distributes electricity to users into a new type of power network that integrates and interacts with power sources, grids, loads, and storage, and flexibly couples with the upper-level grid. Its role in promoting the local consumption of distributed power sources and supporting new types of loads is increasingly important. Meanwhile, driven by the rapid development of distributed energy, energy storage, and diverse loads, the power distribution network's source-load structure is changing more significantly, and its construction and operation modes are becoming more flexible and diverse. This poses a significant challenge to its coordinated planning, necessitating adjustments to power distribution network planning schemes to adapt to the evolution of multiple types of source loads.

[0003] In current power distribution network planning practices, traditional methods typically employ a single-stage static planning model, simplifying the planning problem into an optimization decision under a specific year or scenario, neglecting the inherent temporal evolution characteristics of the power system. While some existing technologies have introduced multi-stage planning concepts, they often employ independent planning or simple interconnections between stages, failing to establish an intrinsic evolutionary correlation mechanism between planning stages. Particularly when facing complex environments with high proportions of distributed generation and rapidly changing load patterns, traditional planning methods struggle to guarantee the consistency and adaptability of the planning scheme over time, easily leading to the failure of early investments or the need for premature modifications, resulting in resource waste.

[0004] In terms of source-load forecasting technology, existing distribution network planning often uses a single forecasting model to process source-load data, or only considers the static distribution characteristics of source loads. It lacks differentiated modeling of the output characteristics of different types of distributed energy (such as wind power and photovoltaics), and a precise grasp of the evolution patterns of diverse loads from residential, commercial, and industrial sectors. Furthermore, when dealing with the uncertainty of the spatiotemporal distribution of source loads, existing technologies often resort to simple methods such as increasing safety margins or setting extreme scenarios, failing to establish a highly adaptive planning decision-making mechanism at the level of source-load evolution mechanisms. This approach may not only lead to overly conservative planning schemes and increased unnecessary investment costs, but may also cause a significant decline in the technical and economic performance of the planning scheme when the actual evolution of source loads deviates from the forecast.

[0005] Furthermore, the evolutionary characteristics of distribution network topology have not been adequately considered in existing planning methods. Traditional methods often treat network structure adjustments between different planning stages as independent decision-making problems, neglecting the inheritance and evolutionary patterns of the power grid topology over time. In practical engineering applications, this disconnected planning approach easily leads to frequent and significant adjustments to the power grid structure, increasing construction and operation costs and potentially affecting power supply reliability. While complex network theory has been applied in power system analysis, effectively characterizing and utilizing the evolutionary characteristics of network topology within the framework of multi-stage distribution network planning, and establishing a planning model that balances technical economy and structural continuity, remains a pressing technical challenge.

[0006] Therefore, existing research on distribution network planning mainly focuses on the optimization planning of distribution networks under specific scenarios, often modifying and expanding specific lines based on the needs of the current development stage, without comprehensively considering the impact of the multi-stage evolution of sources and loads. Current technologies, when dealing with multi-stage planning problems, lack a systematic consideration of the evolutionary laws of the power grid topology and have failed to establish a planning decision-making mechanism adapted to the temporal characteristics of sources and loads, resulting in insufficient adaptability of planning schemes in the face of long-term evolutionary environments. Therefore, it is still necessary to comprehensively consider the resource endowments and future growth trends of various types of sources and loads, and to conduct research on multi-stage planning methods for distribution network architecture that consider the evolution of various types of sources and loads, providing technical support for the sustainable construction and development of distribution networks. Summary of the Invention

[0007] To address the shortcomings and deficiencies of existing technologies, this invention provides a multi-stage planning method for distribution network architecture that considers the evolution and development of multiple types of source loads, thereby solving the problems that existing distribution network architecture planning does not adequately consider the multi-stage evolution and development of source loads and is difficult to adapt to the dynamic changes in source load morphology under new power systems.

[0008] This method first analyzes the spatiotemporal evolution characteristics of various types of distributed power sources (including distributed photovoltaic and wind power) and various types of loads (including residential, commercial, and industrial loads) in conjunction with the current status of the distribution network in the area to be planned. It then establishes a prediction model for the evolution and development of various types of source and load using a combination of data-driven and trend analysis. The data-driven part extracts the temporal features of source and load by improving the long short-term memory neural network and introduces appropriate feature parameters or temporal features to optimize the classification and prediction accuracy by combining the characteristics of different types of source and load. The trend analysis part characterizes the evolution trend of source and load by combining grey prediction with multi-scenario analysis. It introduces correction coefficients that are dynamically adjusted with the planning period for three scenarios: baseline, high growth, and conservative. Finally, it generates typical source and load scenarios for multiple continuous planning stages.

[0009] Subsequently, based on typical source and load scenarios at each planning stage, multi-scenario operation analysis of the distribution network was carried out. Data such as node voltage, line transmission power, and network loss were obtained through power flow calculation tools. Combined with a multi-dimensional evaluation index set (including node-level, line-level, and regional-level indicators) divided according to the distribution network level, a multi-stage quantitative evaluation of the weak links of the distribution network was conducted: first, the indicators were standardized according to preset principles, and then the comprehensive weight of the indicators was determined by combining the analytic hierarchy process and the entropy weight method. The comprehensive scores of nodes, lines, and regions were calculated according to the weights and divided into multiple weak levels to form a multi-stage weak link evaluation result.

[0010] Finally, based on the above evaluation results, complex network theory is used to characterize the topology evolution process of the distribution network between different planning stages. Three types of evolutionary characteristic indicators that reflect the role of node hubs, the local clustering characteristics of the network, and the degree of topological similarity between adjacent stages are extracted. Then, a multi-stage planning model of the distribution network structure covering multiple objectives and multiple constraints is established: the optimization objectives are total planning cost, voltage deviation, and line load rate. The constraints include operational constraints based on the physical characteristics of the distribution network, network topology constraints to ensure the rationality of the network structure, and multi-stage collaborative constraints based on topology evolution characteristic indicators (used to limit structural abrupt changes in the planning scheme between consecutive stages). After solving the model, a multi-stage planning scheme for the distribution network structure is obtained, which can realize the scientific, rational, and orderly planning of the distribution network structure and provide technical support for the sustainable construction and development of the distribution network under the widespread access of multiple types of sources and loads.

[0011] The specific technical solution adopted by this invention to solve its technical problem is as follows:

[0012] A multi-stage planning method for distribution network architecture that considers the evolution of multiple types of source loads includes:

[0013] Based on the source and load data of the area to be planned, the spatiotemporal evolution characteristics of multiple types of distributed power sources and multiple types of loads are analyzed. A prediction model for the evolution and development of multiple types of source and loads is established by combining data-driven and trend analysis methods, and typical source and load scenarios for multiple consecutive planning years are generated.

[0014] Based on the typical source and load scenarios of each planning level year, the distribution network is analyzed in multiple scenarios. Combined with a multi-dimensional evaluation index set, the weak links of the distribution network in each planning level year are quantitatively evaluated, and the evaluation results of weak links in multiple stages are obtained.

[0015] Based on the assessment results of the multi-stage weak links, a multi-stage planning model of the distribution network is constructed and solved to obtain a multi-stage planning scheme for the distribution network.

[0016] The construction of the multi-stage planning model for the power distribution network includes:

[0017] Establish a multi-objective function with total planning cost, voltage deviation, and line load rate as optimization objectives;

[0018] The constraints are constructed, including operational constraints based on the physical characteristics of the distribution network, network topology constraints to ensure the rationality of the network structure, and multi-stage collaborative constraints to limit structural changes in the planning scheme during consecutive planning years.

[0019] The multi-stage collaborative constraints are constructed based on complex network theory. By extracting the evolutionary characteristic indicators of the distribution network topology in adjacent planning years and transforming them into constraints, the topological continuity and evolutionary rationality of the distribution network structure in multi-stage planning are realized.

[0020] Furthermore, in the data-driven and trend analysis combined approach, the data-driven part is implemented using an improved Long Short-Term Memory (LSTM) neural network, specifically including:

[0021] The input, hidden, and output layer structures of the long short-term memory neural network are optimized and improved. A loss function is constructed by combining root mean square error with a regularization term to avoid overfitting of the prediction model. The loss function is associated with the true value, predicted value, and model weight parameters of the source load data samples.

[0022] To address the need for classification and forecasting of various types of source loads, and in combination with the output characteristics of wind power and photovoltaic power, as well as the electricity consumption characteristics of residential, commercial, and industrial loads, characteristic parameters or time-series characteristics adapted to each type of source load are introduced to improve the relevance and accuracy of classification and forecasting.

[0023] Furthermore, in the data-driven and trend analysis combined approach, the trend analysis portion is implemented using a combination of grey prediction and scenario analysis, specifically including:

[0024] The historical data of the source load are accumulated to generate an intermediate sequence. A linear differential equation containing the development coefficient and the gray action parameter is established based on the intermediate sequence, and the parameters are solved by the least squares method.

[0025] The predicted values ​​of the intermediate sequence are obtained based on the solved differential equations, and then the predicted values ​​of the original source load sequence are obtained by cumulative subtraction and restoration.

[0026] The grey prediction results are corrected by introducing scenario coefficients for the baseline scenario, high-growth scenario, and conservative scenario. Each scenario coefficient is dynamically adjusted according to the difference between the planning level year and the baseline year. The prediction value of each scenario is the product of the grey prediction result and the corresponding scenario coefficient.

[0027] Furthermore, the typical source-load scenario for multiple consecutive planning horizontal years is generated, specifically including:

[0028] Based on the multi-type source load evolution and development prediction model, hourly source load power curves are generated for the entire year of each planning level.

[0029] For each planning year, the hourly source load power curves are divided by season to form the initial source load scenario set for each season. Each initial scenario records the hourly power data of multiple types of source loads on a single day.

[0030] A quantitative method is used to characterize the difference between any two initial scenarios. Then, the initial source load scenario set for each season is reduced by a scenario reduction method to obtain the typical source load scenarios for each season in each planning year.

[0031] Furthermore, the multi-dimensional evaluation index set is divided into the following levels according to the distribution network: node-level indicators, line-level indicators, and regional-level indicators:

[0032] The node-level indicators are used to evaluate the structural rationality and load distribution balance of distribution network nodes, and are calculated based on the total number of nodes, the number of branch lines connected to each node, and node load data.

[0033] The line-level indicators are used to evaluate the load status and energy consumption level of distribution network lines, and are calculated based on line transmission power, rated transmission capacity and network loss data.

[0034] The regional indicators are used to evaluate the source-load matching degree and distributed power absorption capacity of a designated area of ​​the distribution network, and are calculated based on the power output, load power and abandoned power data of the area.

[0035] Furthermore, a multi-stage quantitative assessment of weak links in the distribution network is conducted, specifically including:

[0036] The evaluation indicators are standardized according to the principle of "maximizing positive indicators and minimizing negative indicators", and the original indicator values ​​are transformed into standardized values ​​with a unified range of values.

[0037] The comprehensive weight of each evaluation indicator is determined by combining subjective weight and objective weight. The subjective weight is determined by the analytic hierarchy process (AHP), and the objective weight is calculated by the entropy weight method.

[0038] For distribution network nodes, lines, and designated areas, evaluation indicators of the corresponding levels are selected and weighted according to comprehensive weights to obtain a comprehensive score;

[0039] The weakness of distribution network nodes, lines and areas is divided into multiple levels based on the comprehensive score, and the level classification results constitute the multi-stage weakness assessment results.

[0040] Furthermore, the evolutionary characteristic indicators are extracted from the distribution network topology based on complex network theory and are divided into three categories according to function:

[0041] The first type of index characterizes the pivotal role of a node in the distribution network topology and is calculated based on the distribution characteristics of the shortest path between nodes.

[0042] The second type of index characterizes the local clustering characteristics of the distribution network topology and is calculated based on the connection relationship between node neighbors;

[0043] The third type of index characterizes the similarity of distribution network topologies in adjacent planning years and is used to determine the stationarity of topology evolution.

[0044] Furthermore, the topological evolution constraints in the multi-stage cooperative constraints include:

[0045] Node stability constraint: Limit the change in the hub function index of the same node in adjacent planning years to not exceed a preset threshold;

[0046] Clustering coefficient evolution constraint: restrict the local clustering characteristics of the distribution network in the current planning level year to be no less than the reasonable attenuation range of the previous planning level year;

[0047] Topology similarity constraint: The topology similarity between adjacent planning year distribution networks is limited to a preset lower limit to ensure topology continuity.

[0048] Furthermore, when conducting multi-scenario operation analysis of the distribution network, the node voltage, line transmission power, and network loss data of the distribution network under different source and load scenarios in each planning level year are obtained through power flow calculation tools, providing data support for subsequent quantitative assessment of weak links.

[0049] And a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method described above.

[0050] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0051] Compared with the prior art, the present invention and its preferred embodiments have at least the following beneficial effects:

[0052] It can more accurately adapt to the dynamic evolution characteristics of various types of loads, solving the problem that existing planning does not adequately consider the multi-stage development of loads. Through a prediction model that combines data-driven and trend analysis, it not only uses an improved long short-term memory neural network to capture the temporal patterns of loads, but also covers different evolution trends through multi-scenario grey prediction. Furthermore, it can introduce adaptive features to optimize prediction accuracy for wind power, photovoltaics, and various types of loads. The resulting multi-stage typical load scenarios are more in line with actual planning needs, providing a precise source load input basis for subsequent planning.

[0053] This approach enables multi-dimensional and multi-stage quantitative assessment of weak links in the distribution network, avoiding the limitations of existing assessment methods that are singular or static. Through a hierarchical set of assessment indicators at the node, line, and regional levels, combined with a weighted calculation method that integrates subjective and objective factors, it comprehensively covers dimensions such as the rationality of the distribution network structure, power supply quality, operational balance, and economy. Furthermore, it dynamically assesses and classifies weaknesses according to planning stages, accurately pinpointing network shortcomings at different stages and providing targeted optimization directions for the planning model.

[0054] To ensure the continuity and rational evolution of the distribution network structure in multi-stage planning, and to overcome the shortcomings of insufficient connection between existing planning stages and the susceptibility to structural abrupt changes, this study utilizes complex network theory to extract evolutionary characteristic indicators reflecting the hub role of nodes, network aggregation characteristics, and topological similarity. These indicators are then transformed into multi-stage collaborative constraints, effectively limiting drastic changes in the planning scheme between consecutive stages. This ensures a smooth transition of the network structure with the evolution of source loads, reducing the cost waste and power outage risks caused by frequent modifications.

[0055] The constructed multi-objective programming model balances economy and security, and is more in line with the actual construction needs of distribution networks than existing single-objective programming models. The model uses total planning cost, voltage deviation, and line load rate as optimization objectives, while incorporating distribution network physical operation constraints and network topology constraints. It can control construction and operation and maintenance costs while ensuring stable power supply voltage and balanced line load, thus achieving scientific planning of the distribution network.

[0056] In summary, this invention forms a complete technical chain of source load prediction, weakness assessment, topology constraints, and multi-stage modeling, which can provide a sustainable planning solution for new distribution networks with wide access to multiple types of source loads and provide technical support for the orderly construction of distribution networks under the new power system. Attached Figure Description

[0057] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0058] Figure 1 This is a flowchart illustrating the prediction process for the evolution and development of multiple types of source loads according to an embodiment of the present invention.

[0059] Figure 2 This is a diagram of a multi-dimensional evaluation index set for power distribution networks according to an embodiment of the present invention. Detailed Implementation

[0060] To make the features and advantages of the present invention more apparent and understandable, specific embodiments are described below in detail:

[0061] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0062] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0063] To address the problem that existing distribution network planning does not adequately consider the multi-stage evolution of source loads, this invention proposes a multi-stage planning method for distribution network that considers the evolution of multiple types of source loads.

[0064] To achieve the multi-stage planning scheme for the distribution network that considers the evolution and development of multiple types of source loads in this embodiment, its core multi-type source load evolution and development prediction process is as follows: Figure 1 As shown, it includes:

[0065] First, based on the current status of the power distribution network in the area to be planned, we analyze the spatiotemporal distribution characteristics of various types of distributed power sources (including distributed photovoltaic and wind power) and various types of loads (including residential, commercial and industrial loads), and extract the key characteristics (including time characteristics, meteorological characteristics and socio-economic characteristics) that affect the growth and evolution of power sources and loads.

[0066] Next, a prediction model combining data-driven and trend analysis is constructed. The data-driven part extracts the temporal features of source loads by improving the long short-term memory neural network, and introduces adaptive parameters or features to optimize the prediction accuracy by combining the characteristics of different types of source loads. The trend analysis part generates the basic trend of source load evolution through gray prediction, and introduces dynamic correction coefficients for three scenarios: baseline, high growth, and conservative.

[0067] The two types of prediction results are then merged according to preset weights to obtain a comprehensive prediction result for multiple types of source loads.

[0068] Finally, based on the comprehensive forecast results, typical source-load scenarios for each stage of the distribution network in terms of seasons are generated for the short-term, medium-term and long-term planning stages, providing basic input for subsequent multi-scenario power flow analysis and weak link assessment of the distribution network.

[0069] Based on this, the implementation process of the multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads in this embodiment of the invention is as follows:

[0070] 1) Combine the analysis of the current situation of the area to be planned to determine the spatiotemporal distribution characteristics of various types of distributed power sources and loads. Based on data-driven and trend analysis methods, establish prediction models for the evolution and development of various types of power sources and loads, and generate typical power source and load scenarios for the short-term, medium-term and long-term planning stages of the distribution network.

[0071] 2) Based on the typical source-load multi-stage scenarios generated in step 1), Matpower is used to perform multi-scenario power flow analysis of the distribution network, and a multi-dimensional evaluation index set is constructed to conduct multi-stage quantitative evaluation of the weak links of the distribution network.

[0072] 3) Based on the multi-stage quantitative evaluation results obtained in step 2), the evolution process of the distribution network topology between different stages is characterized by complex network theory. Then, a multi-stage planning model of the distribution network structure covering multiple objectives and multiple constraints is established. The multi-stage planning scheme of the distribution network structure is obtained by solving the distribution network structure planning model.

[0073] It should be noted that the design of the solution part is not an innovative feature of this invention. Based on the modeling provided in this embodiment, those skilled in the art can select well-known methods in the field, such as genetic algorithms and particle swarm optimization algorithms, to complete the solution according to the actual situation and technical conditions.

[0074] As a preferred embodiment, step 1) establishes a multi-type source-load evolution prediction model based on data-driven and trend analysis methods, specifically including:

[0075] (2.1) Based on the multi-dimensional data of source and load, analyze the composition of various types of distributed power sources (distributed photovoltaic and wind power) and loads (residential, commercial, and industrial), and extract and summarize the key characteristics affecting the growth and evolution of source and load, including: time characteristics, meteorological characteristics, and socio-economic characteristics, to form a feature matrix. ,in Let N be the set of real numbers N×M, where N and M represent the number of samples and the number of features of the source payload data, respectively.

[0076] (2.2) Construct a multi-type source-load evolution prediction model that combines data-driven prediction and trend analysis prediction, specifically including the following steps:

[0077] (2.2.1) Establish a data-driven prediction model, which includes the following steps:

[0078] (2.2.1a) An improved long short-term memory neural network (LSTM) is used to extract the temporal features of the source load data and optimize the design of the prediction model structure, including the input layer, hidden layer and output layer;

[0079] (2.2.1b) The loss function is constructed by combining the root mean square error with a regularization term to avoid overfitting of the prediction model;

[0080]

[0081] in, , and are the true value and predicted value of the p-th sample, respectively; θ is the regularization coefficient; These are the model weight parameters.

[0082] (2.2.1c) Optimize the classification model for multi-type source load forecasting: For wind power and photovoltaic forecasting, Weibull distribution and Beta distribution parameters are introduced to improve the accuracy of wind power and photovoltaic forecasting respectively; For residential, commercial and industrial loads, holiday and seasonal features are introduced to improve the accuracy of classified load forecasting.

[0083] (2.2.2) The grey prediction method combined with scenario analysis is used to characterize the trend characteristics of source load evolution, and a prediction model based on trend analysis is established, which specifically includes the following steps:

[0084] (2.2.1a) Construction of the grey prediction model: historical data of source load By accumulating the data, we obtain the 1-AGO sequence. Establish a first-order linear differential equation Where a and b are the development coefficient and the grey action parameter, respectively, and these two parameters can be solved using the least squares method; solving the differential equation yields the prediction formula for the 1-AGO sequence. Where e is a mathematical constant; and the predicted value of the original sequence is obtained by cumulative subtraction and restoration. .

[0085] (2.2.1b) Scenario Trend Correction: Three scenario coefficients are introduced to correct the grey prediction results, categorized into a baseline scenario, a high-growth scenario, and a conservative scenario. Among these, the baseline scenario correction coefficient... Predicted value High growth scenario correction factor Predicted value , where y, These represent the planning year and the base year, respectively; conservative scenario adjustment factor. Predicted value .

[0086] (2.2.3) Combining the data-driven prediction results from (2.2.1) with the trend analysis prediction results from (2.2.2) yields a comprehensive prediction result:

[0087]

[0088] in, , and The results are the predictions of the Long Short-Term Memory Neural Network Prediction Model, the Grey Prediction Model, and the combined prediction results, respectively. , These are the weights for the Long Short-Term Memory Neural Network Prediction Model and the Grey Prediction Model, respectively.

[0089] As a preferred embodiment, step 1) generates typical source-load scenarios for the short-term, medium-term, and long-term planning stages of the distribution network, specifically including:

[0090] (3.1) Based on the source-load evolution and development prediction model, the hourly source-load power curves for each stage of the year are predicted and generated for the short-term, medium-term and long-term planning stages.

[0091] (3.2) For different planning stages, the annual source load power data is divided according to different seasons and on a daily basis to obtain the initial source load scenario sets for the transition season, summer, and winter. , , Mt, Ms, and Mw represent the number of initial scenarios for each season, and each initial scenario contains daily power data of multiple types of source loads.

[0092] (3.3) Using Euclidean distance The differences between any two scenarios are represented, and the initial scenarios for each season are reduced using the inverse scenario reduction method to obtain the typical scenarios for each season within the planning year of stage t. , , .

[0093] As a preferred embodiment, step 2) involves using a single-dimensional multi-dimensional evaluation index set to conduct a multi-stage quantitative evaluation of the weak links in the distribution network, specifically including:

[0094] (4.1) Construction of a multi-dimensional evaluation index set, the obtained evaluation index set is as follows: Figure 2 As shown, it includes:

[0095]

[0096] (4.2) Based on the above multi-dimensional evaluation indicators, a multi-stage quantitative evaluation of the weak links in the distribution network is carried out, specifically including the following steps:

[0097] (4.2.1) Standardization of indicators: The evaluation indicators are standardized according to the principle of "maximizing positive indicators and minimizing negative indicators", and the original indicator values ​​are transformed into standardized values ​​in the range of [0,1].

[0098] (4.2.2) Calculation of subjective and objective weights: The subjective weights of each indicator are determined by the analytic hierarchy process, the objective weights are calculated by the entropy weight method, and the subjective and objective weights are combined by the linear weighting method.

[0099] (4.2.3) Calculation of comprehensive score for node / line / region: For distribution network node i, four node-level indicators, namely node connectivity, load balance, voltage deviation and voltage fluctuation rate, are selected and weighted and summed to calculate the comprehensive score; for distribution network line ij, two line-level indicators, namely line load rate and line loss rate, are selected and weighted and summed to calculate the comprehensive score; for distribution network region Ω, two region-level indicators, namely source-load matching degree and curtailment rate, are selected and weighted and summed to calculate the comprehensive score.

[0100] (4.2.4) Classification of weak links: The weak links of power grid nodes / lines / areas are classified according to the calculated comprehensive score, including four levels: good, minor, moderate and serious.

[0101] (4.2.5) For the short-term, medium-term and long-term planning stages of the distribution network, quantitative assessments are carried out according to the process of (4.2.1)-(4.2.4) respectively, and assessment results of weak links of the distribution network under multiple planning stages are formed.

[0102] As a preferred embodiment, step 3) uses complex network theory to characterize the distribution network topology evolution process between different stages, specifically including:

[0103] (5.1) Represent the distribution network at each planning stage as an undirected weighted network diagram. ,in Let n be the set of nodes in stage t (including power nodes, load nodes, connection nodes, etc.), and n be the total number of nodes. Let be the set of edges in stage t ( This indicates that there is a line connection between nodes i and j. (Indicates no connection); Let be the edge weight matrix for stage t. ,in , Let be the rated transmission capacity and length of line ij in stage t, respectively.

[0104] (5.2) To further characterize the evolution of distribution network topology at different stages, the following topology evolution characteristic indicators are established:

[0105] (5.2.1) Betweenness Centrality Index

[0106]

[0107] in, This represents the total number of shortest paths between nodes s and z. The number of shortest paths passing through node i, and the betweenness centrality index. It represents the pivotal role of a node in the network.

[0108] (5.2.2) Average clustering coefficient index

[0109]

[0110] in, The actual number of connections between node i's neighboring nodes; For the connection branch of node i in stage t Average clustering coefficient index Characterizes the local clustering properties of the network.

[0111] (5.2.3) Topological similarity index

[0112]

[0113] Among them, topological similarity index Characterize the topological similarity of the network at stage t and stage t-1.

[0114] As a preferred embodiment, step 3) establishes a multi-stage planning model for the distribution network structure, encompassing multiple objectives and constraints. Specifically, this includes: based on the representation of the distribution network topology evolution process, establishing a multi-stage planning model for the distribution network structure with multiple objectives (lowest total planning cost, lowest voltage deviation, and lowest line load rate) and constraints (network topology constraints, multi-stage evolution constraints, and distribution network operation constraints). Specifically, this includes:

[0115] (6.1) Objective function

[0116]

[0117] Among them, F is the comprehensive goal of the multi-stage planning of the distribution network structure, which includes three parts: F1, F2, and F3; F1 is the total planning cost indicator, F2 is the voltage deviation indicator, and F3 is the line load rate indicator.

[0118] (6.1.1) The overall planning cost index F1 is:

[0119]

[0120] Where t=1, 2, and 3 represent the short-term, medium-term, and long-term stages of distribution network planning, respectively; The benchmark rate of return; , , These are the construction investment cost, operation and maintenance cost, and network loss cost of the distribution network structure planning.

[0121] Construction investment cost of power distribution network planning for:

[0122]

[0123] in, , These represent the newly added line length and newly added transformer capacity between nodes i and j in stage t, respectively. , These represent the construction investment costs per unit length of line and per unit capacity of transformer, respectively.

[0124] Operation and maintenance costs of power distribution network planning for:

[0125]

[0126] in, , These are the annual operation and maintenance costs per unit length of line and per unit capacity of transformer, respectively.

[0127] Network loss cost in distribution network planning for:

[0128]

[0129] in, For grid loss electricity price; This refers to the annual power loss of the distribution network.

[0130] (6.1.2) The voltage deviation index F2 is:

[0131]

[0132] in, Let i be the actual voltage of node i in stage t and time period tt. This is the node reference voltage.

[0133] (6.1.3) The line load factor index F3 is:

[0134]

[0135] in, Let be the actual transmission power of line ij in the t-th period of the t-th stage.

[0136] (6.2) Constraints

[0137] (6.2.1) The network topology constraints are:

[0138]

[0139] in, , , These represent the constraints on the number of distribution network lines, the degree of nodes, and the connectivity constraints, respectively.

[0140] (6.2.2) Multi-stage evolution constraints include topology evolution constraints and multi-stage planning constraints:

[0141] The topological evolution constraints are:

[0142]

[0143] in, , , These represent the node stability constraints, clustering coefficient evolution constraints, and topological similarity constraints in the multi-stage planning of the distribution network, respectively. This represents the upper limit of the change in betweenness; The attenuation coefficient; This represents the lower bound of topological similarity.

[0144] Multi-stage planning constraints are:

[0145]

[0146] in, , , These represent the investment continuity constraint, capacity expansion constraint, and equipment reuse constraint in the multi-stage planning of the distribution network, respectively. This represents the upper limit of the investment growth rate. Reserve a capacity allowance; To represent the binary variable for the newly added line / transformer in stage t, .

[0147] (6.2.3) The topology constraints for the distribution network operation are:

[0148]

[0149] in, , For power balance constraints in distribution network operation, , These represent the active and reactive power outputs of the power source connected to node i during the t-th time period in stage t. , These represent the active and reactive loads of node i at stage t and time period tt, respectively. , These are the node voltage constraints and line transmission power constraints for distribution network operation, respectively. , These are the upper and lower limits of the node voltage, respectively.

[0150] Based on the technical solutions provided by the above embodiments of the present invention, the problem of insufficient consideration of the impact of the multi-stage evolution of source loads in the current distribution network planning can be effectively solved:

[0151] First, a multi-type source-load evolution prediction model is established based on data-driven and trend analysis methods, generating typical source-load scenarios for the short-term, medium-term, and long-term planning stages of the distribution network. Second, Matpower is used to conduct multi-scenario power flow analysis of the distribution network, and a multi-dimensional evaluation index set is constructed to quantitatively evaluate the weak links of the distribution network in multiple stages. Then, complex network theory is used to characterize the distribution network topology evolution process between different stages, thereby establishing a multi-stage planning model for the distribution network structure that covers multiple objectives and constraints, and solving for the multi-stage planning scheme of the distribution network structure. This invention can overcome the shortcomings of traditional distribution network structure planning, which focuses on specific scenarios and development stages and fails to fully consider the multi-stage evolution of source loads. It provides a scheme reference for the planning and construction of distribution networks with widespread access to multiple types of source loads, realizing the scientific, rational, and orderly planning of the distribution network structure (e.g., through multi-stage collaborative constraints, the topological similarity between adjacent stages is significantly improved, avoiding planning scheme jumps), and providing technical support for the sustainable construction and development of new distribution systems.

[0152] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0153] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0154] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0155] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0156] This invention is not limited to the preferred embodiment described above. Anyone inspired by this invention can derive other forms of multi-stage planning methods for distribution network architecture that consider the evolution of multiple types of source loads. All equivalent changes and modifications made within the scope of the claims of this invention shall fall within the scope of this invention.

Claims

1. A multi-stage planning method for distribution network architecture considering the evolution of multiple types of source loads, characterized in that, include: Based on the source and load data of the area to be planned, the spatiotemporal evolution characteristics of multiple types of distributed power sources and multiple types of loads are analyzed. A prediction model for the evolution and development of multiple types of source and loads is established by combining data-driven and trend analysis methods, and typical source and load scenarios for multiple consecutive planning years are generated. Based on the typical source and load scenarios of each planning level year, the distribution network is analyzed in multiple scenarios. Combined with a multi-dimensional evaluation index set, the weak links of the distribution network in each planning level year are quantitatively evaluated, and the evaluation results of weak links in multiple stages are obtained. Based on the assessment results of the multi-stage weak links, a multi-stage planning model of the distribution network is constructed and solved to obtain a multi-stage planning scheme for the distribution network. The construction of the multi-stage planning model for the power distribution network includes: Establish a multi-objective function with total planning cost, voltage deviation, and line load rate as optimization objectives; The constraints are constructed, including operational constraints based on the physical characteristics of the distribution network, network topology constraints to ensure the rationality of the network structure, and multi-stage collaborative constraints to limit structural changes in the planning scheme during consecutive planning years. The multi-stage collaborative constraints are constructed based on complex network theory. By extracting the evolutionary characteristic indicators of the distribution network topology in adjacent planning years and transforming them into constraints, the topological continuity and evolutionary rationality of the distribution network structure in multi-stage planning are realized.

2. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: In the data-driven and trend analysis combined approach, the data-driven part is implemented using an improved Long Short-Term Memory (LSTM) neural network, specifically including: The input, hidden, and output layer structures of the long short-term memory neural network are optimized and improved. A loss function is constructed by combining root mean square error with a regularization term to avoid overfitting of the prediction model. The loss function is associated with the true value, predicted value, and model weight parameters of the source load data samples. To address the need for classification and forecasting of various types of source loads, and in combination with the output characteristics of wind power and photovoltaic power, as well as the electricity consumption characteristics of residential, commercial, and industrial loads, characteristic parameters or time-series characteristics adapted to each type of source load are introduced to improve the relevance and accuracy of classification and forecasting.

3. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: In the data-driven and trend analysis combined approach, the trend analysis portion employs a combination of grey prediction and scenario analysis, specifically including: The historical data of the source load are accumulated to generate an intermediate sequence. A linear differential equation containing the development coefficient and the gray action parameter is established based on the intermediate sequence, and the parameters are solved by the least squares method. The predicted values ​​of the intermediate sequence are obtained based on the solved differential equations, and then the predicted values ​​of the original source load sequence are obtained by cumulative subtraction and restoration. The grey prediction results are corrected by introducing scenario coefficients for the baseline scenario, high-growth scenario, and conservative scenario. Each scenario coefficient is dynamically adjusted according to the difference between the planning level year and the baseline year. The prediction value of each scenario is the product of the grey prediction result and the corresponding scenario coefficient.

4. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: Generating the typical source load scenario for multiple consecutive planning horizontal years specifically includes: Based on the multi-type source load evolution and development prediction model, hourly source load power curves are generated for the entire year of each planning level. For each planning year, the hourly source load power curves are divided by season to form the initial source load scenario set for each season. Each initial scenario records the hourly power data of multiple types of source loads on a single day. A quantitative method is used to characterize the difference between any two initial scenarios. Then, the initial source load scenario set for each season is reduced by a scenario reduction method to obtain the typical source load scenarios for each season in each planning year.

5. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: The multi-dimensional evaluation index set is divided into three levels according to the distribution network: node-level indicators, line-level indicators, and regional-level indicators. The node-level indicators are used to evaluate the structural rationality and load distribution balance of distribution network nodes, and are calculated based on the total number of nodes, the number of branch lines connected to each node, and node load data. The line-level indicators are used to evaluate the load status and energy consumption level of distribution network lines, and are calculated based on line transmission power, rated transmission capacity and network loss data. The regional indicators are used to evaluate the source-load matching degree and distributed power absorption capacity of a designated area of ​​the distribution network, and are calculated based on the power output, load power and abandoned power data of the area.

6. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: A multi-stage quantitative assessment of weak links in the distribution network is conducted, specifically including: The evaluation indicators are standardized according to the principle of "maximizing positive indicators and minimizing negative indicators", and the original indicator values ​​are transformed into standardized values ​​with a unified range of values. The comprehensive weight of each evaluation indicator is determined by combining subjective weight and objective weight. The subjective weight is determined by the analytic hierarchy process (AHP), and the objective weight is calculated by the entropy weight method. For distribution network nodes, lines, and designated areas, evaluation indicators of the corresponding levels are selected and weighted according to comprehensive weights to obtain a comprehensive score; The weakness of distribution network nodes, lines and areas is divided into multiple levels based on the comprehensive score, and the level classification results constitute the multi-stage weakness assessment results.

7. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: The evolutionary characteristic indicators are extracted from the distribution network topology based on complex network theory and are divided into three categories according to function: The first type of index characterizes the pivotal role of a node in the distribution network topology and is calculated based on the distribution characteristics of the shortest path between nodes. The second type of index characterizes the local clustering characteristics of the distribution network topology and is calculated based on the connection relationship between node neighbors; The third type of index characterizes the similarity of distribution network topologies in adjacent planning years and is used to determine the stationarity of topology evolution.

8. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: The topological evolution constraints in the multi-stage cooperative constraints include: Node stability constraint: Limit the change in the hub function index of the same node in adjacent planning years to not exceed a preset threshold; Clustering coefficient evolution constraint: Limit the local clustering characteristics of the distribution network in the current planning year to be no less than the reasonable attenuation range of the previous planning year; Topology similarity constraint: The topology similarity between adjacent planning year distribution networks is limited to a preset lower limit to ensure topology continuity.

9. The multi-stage planning method for distribution network structure considering the evolution of multiple types of source loads as described in claim 1, characterized in that: When conducting multi-scenario operation analysis of the distribution network, the node voltage, line transmission power and network loss data of the distribution network under different source and load scenarios in each planning year are obtained through power flow calculation tools, which provides data support for subsequent quantitative assessment of weak links.

10. A computer device, characterized in that, The invention includes a processor and a non-transitory computer-readable storage medium storing a computer program, wherein the processor, when executing the computer program, implements the method of claims 1-9.

Images