Power distribution network optimization method adaptive to large-capacity load transfer of power distribution network

Abstract

The application belongs to the technical field of power distribution network optimization, and discloses a power distribution network optimization method suitable for large-capacity load transfer of the power distribution network, and the specific steps of the optimization method of the power distribution network are as follows: step 1: data intelligent analysis and feature recognition; step 2: multi-dimensional feature matching and strategy generation; step 3: space intelligent transfer optimization; step 4: time intelligent transfer optimization; and step 5: cross-regional collaborative optimization. The application realizes the deep integration of advanced technologies such as big data, artificial intelligence and Internet of Things. The method can significantly improve the intelligent management level of the power distribution network, realize the accurate prediction and intelligent scheduling of the load, effectively alleviate the load pressure of the regional power distribution network, balance the load rate, significantly reduce the peak-valley difference, improve the overall operation efficiency and stability of the power grid, and has obvious advantages and benefits in improving the operation efficiency of the power grid, guaranteeing the power supply reliability, promoting the optimal allocation of resources and promoting energy transformation.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network optimization, and in particular relates to a distribution network optimization method adapted to large-capacity load transfer in the distribution network. Background Art

[0002] Traditional methods rely primarily on manual experience and limited historical data to forecast and dispatch load. This static, lagging management approach struggles to accurately capture real-time load changes and underlying trends, making it difficult to cope with sudden load peaks or troughs, which can easily lead to grid instability and even failures. Furthermore, existing approaches lack cross-regional coordination mechanisms, leaving regional power grids often operating independently, making it difficult to optimize the allocation and sharing of power resources. This siloed operating model not only limits competition and development in the power market but also leads to waste of power resources and increased transaction costs. In today's world of increasing energy scarcity and increasingly prominent environmental issues, this inefficient resource allocation approach clearly cannot meet the demands of sustainable development.

[0003] Furthermore, traditional approaches neglect demand-side response, a significant issue. As a crucial component of the power grid, the flexibility and responsiveness of the demand side are crucial for balancing grid loads and improving operational efficiency. However, existing approaches often focus solely on supply-side regulation while ignoring the potential of the demand side. This results in a single, ineffective load regulation method. This not only exacerbates the peak-to-valley problem but also limits the grid's ability to absorb renewable energy, hindering the progress of energy transition.

[0004] In summary, the drawbacks of existing large-capacity load management methods in distribution networks primarily include delayed forecasting and scheduling, insufficient cross-regional coordination, and neglected demand-side response. These issues not only impact the operational efficiency and stability of the power grid but also hinder the advancement of energy transformation and green development. Therefore, improvements and innovations are proposed. Summary of the Invention

[0005] The object of the present invention is to provide a distribution network optimization method that is adaptable to large-capacity load transfer in the distribution network, so as to solve the problems raised in the above background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a distribution network optimization method adapted to large-capacity load transfer in the distribution network, wherein the specific steps of the distribution network optimization method are as follows:

[0007] Step 1: Intelligent Data Analysis and Feature Identification: Utilizing big data and artificial intelligence technologies, we conduct in-depth mining of regional distribution network load fluctuation data. Combined with real-time load data from adjacent regional distribution networks, we intelligently identify regional distribution network load fluctuation patterns and the complementary load characteristics between adjacent regions, generating an accurate load characteristic map.

[0008] Step 2: Multi-dimensional feature matching and strategy generation: Based on the load characteristic map obtained in Step 1, a multi-dimensional feature matching algorithm is used to determine whether the distribution network load characteristics meet the preset spatial transfer optimization conditions or temporal transfer optimization conditions. Based on the matching results, the corresponding optimization strategy set is automatically generated.

[0009] Step 3: Spatial Intelligent Transfer Optimization: For areas that meet the spatial transfer optimization criteria, a reinforcement learning-based intelligent planning optimization method is used to automatically adjust the switch configuration in the substation and network structure to achieve intelligent load switching and balanced distribution. Furthermore, IoT technology is used to monitor network status in real time and dynamically adjust operating modes to ensure N-1 security and optimal load factor.

[0010] Step 4: Time-based intelligent transfer optimization: For areas requiring only time-based transfer optimization, an innovative collaborative optimization mechanism for demand-side response and energy storage systems is introduced. Intelligent algorithms are used to predict future load trends, combined with user electricity behavior analysis to develop personalized energy storage charging and discharging plans. This plan also integrates new load types such as electric vehicle charging stations and thermal and cold storage systems to achieve smooth load transfer over time, reduce peak-to-valley fluctuations, and improve overall grid efficiency.

[0011] Step 5: Cross-regional collaborative optimization: Based on steps 3 and 4, a cross-regional collaborative optimization strategy is further introduced. By utilizing the complementary characteristics of the distribution networks in adjacent regions, cross-regional power trading and shared energy storage methods can be used to achieve load balancing and resource optimization configuration on a larger scale.

[0012] Preferably, the data intelligent analysis and feature recognition process also includes the prediction and evaluation of the impact of extreme weather and special events on holidays on load fluctuations, so as to improve the accuracy and adaptability of the optimization strategy.

[0013] Preferably, the load characteristics in step 1 specifically include:

[0014] The first step is data collection: load data from meters, substations, and power plants is collected from the power system. The time range of the data, including active power and reactive power, is determined based on the analysis requirements.

[0015] The second step is data preprocessing: remove outliers, missing values, and data points that do not meet the requirements; standardize and normalize the data for subsequent analysis; ensure that all data points are aligned on the time axis for time series analysis;

[0016] Third, feature extraction: Calculate the basic statistics of load data, such as mean, standard deviation, maximum, minimum, and peak-to-valley difference; extract the periodicity and trend characteristics of load data over time; and analyze the load correlation between different regions or different users.

[0017] The fourth step is pattern recognition: grouping load data based on similarity to identify different load patterns; using machine learning algorithms to classify load data and distinguish different types of loads; time series analysis: using time series models to predict and analyze load data trends;

[0018] Step 5: Spectrum generation: Based on the above data, a curve chart showing the change of load over time is drawn to intuitively display the load fluctuation; the depth of color is used to represent the size and distribution of the load, forming a heat map to show the spatial distribution characteristics of the load; the main components of the load data are extracted through the principal component analysis method, and drawn into a spectrum to show the main fluctuation characteristics of the load.

[0019] Preferably, the demand-side response and energy storage system collaborative optimization mechanism also includes personalized modeling and prediction of user electricity consumption behavior, guiding users to actively participate in load regulation through incentive mechanisms, thereby improving the flexibility and response speed of the user side.

[0020] Preferably, the cross-regional collaborative optimization strategy also includes real-time monitoring and forecasting of electricity market prices, guided by economic benefits, to achieve optimal decision-making for cross-regional electricity transactions and promote effective allocation and utilization of resources.

[0021] Preferably, the multi-dimensional feature matching algorithm formula in step 2 is specifically:

[0022]

[0023] Among them, x n Indicates the nth characteristic value of the current load characteristic, y n It represents the nth feature value of the preset optimization condition, and a is the total number of features.

[0024] Preferably, the intelligent algorithm in step 4 is specifically a time series prediction algorithm, which includes an autoregressive integrated moving average model and an exponential smoothing algorithm, specifically:

[0025] Autoregressive integrated moving average model: φ(A)(y n -c)=θ(A)∈ t

[0026] Among them, φ(A) and θ(A) are polynomials about the backshift operator A, y t is the value of the time series at time n, c is a constant term, ∈ n is the error term.

[0027] Exponential smoothing algorithm: S i =αx i +(1-α)S i-1

[0028] Among them, S i is the smoothed value at time i, x i is the actual observation value at time i, and α is the smoothing parameter (0<α<1).

[0029] Preferably, the introduction of the cross-region collaborative optimization strategy in step 5 involves a mathematical model of the optimization problem, specifically:

[0030] Mathematical model of optimization problem:

[0031] in,

[0032] · is the electricity generation cost of region i in period t;

[0033] · is the inter-regional electricity transaction cost of region i in period t;

[0034] · is the power transmission loss cost of region i in period t;

[0035] · is the possible income of region i in period t;

[0036] T is the total number of time periods.

[0037] Preferably, the constraints in the mathematical model of the optimization problem are:

[0038] Electricity supply and demand balance:

[0039] in,

[0040] is the amount of power transferred from region j to region i in time period t;

[0041] is the power generation of region i in time period t;

[0042] and are the energy storage discharge and charge amounts of region i in time period t, respectively;

[0043] D i,t is the electricity demand of region i in period t.

[0044] Energy storage system constraints:

[0045] Among them, E i,t is the energy storage capacity of region i at the end of period t;

[0046] ηcha and ηdis are the charging and discharging efficiencies of the energy storage system, respectively;

[0047] Δt is the time interval;

[0048] E i,min and E i,max They are the minimum and maximum limits of energy storage capacity respectively.

[0049] The beneficial effects of the present invention are as follows:

[0050] By integrating cutting-edge technologies such as big data, artificial intelligence, and the Internet of Things, this invention can accurately capture and predict subtle changes in load, providing a solid foundation for intelligent scheduling of power grids.

[0051] Specifically, this method first uses intelligent analysis and feature recognition to deeply analyze the load characteristics of the distribution network, revealing its inherent laws and trends, providing a scientific basis for the subsequent development of optimization strategies. Subsequently, the introduction of multi-dimensional feature matching and strategy generation mechanisms makes the development of optimization strategies more flexible and diverse, allowing them to be tailored to the actual conditions of different regions, ensuring maximum optimization results.

[0052] In terms of spatial intelligent transfer optimization, this method demonstrates powerful spatial scheduling capabilities. Through intelligent planning optimization and real-time IoT monitoring, it achieves intelligent load switching and balanced distribution across regions, effectively alleviating load pressure in local areas and improving the load balance of the entire power grid. In terms of temporal intelligent transfer optimization, it cleverly utilizes demand-side response and energy storage system collaborative optimization mechanisms to guide users to actively participate in load regulation, achieving smooth load transfer across time, reducing peak-to-valley differences, and improving the overall operational efficiency of the power grid.

[0053] The proposed method also transcends regional constraints, enabling cross-regional collaborative optimization. Through innovative approaches such as real-time monitoring and forecasting of the power market, cross-regional power trading, and shared energy storage, it not only promotes the optimal allocation and utilization of power resources, but also reduces transaction costs and enhances the economic efficiency of the power grid. This cross-regional collaborative operation model undoubtedly opens up broader opportunities for the future development of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a simplified flow chart of the main steps of the distribution network optimization method of the present invention. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a distribution network optimization method that adapts to large-capacity load transfer in the distribution network. The specific steps of the distribution network optimization method are as follows:

[0057] Step 1: Intelligent Data Analysis and Feature Identification: Utilizing big data and artificial intelligence technologies, we conduct in-depth mining of regional distribution network load fluctuation data. Combined with real-time load data from adjacent regional distribution networks, we intelligently identify regional distribution network load fluctuation patterns and the complementary load characteristics between adjacent regions, generating an accurate load characteristic map.

[0058] The core of this step is to deeply mine the massive load fluctuation data of the regional distribution network and integrate it with the real-time load information of adjacent regions to form a comprehensive and accurate load characteristic map. We use advanced big data processing technologies and artificial intelligence algorithms, such as deep learning and cluster analysis, to analyze load data in multiple dimensions and at multiple levels. In particular, we introduce time series analysis technology to capture the cyclical and seasonal characteristics of load fluctuations. Combined with extreme weather prediction models and holiday impact assessment mechanisms, we conduct forward-looking analysis of load fluctuations under special events, thereby greatly improving the accuracy and foresight of optimization strategies.

[0059] Step 2: Multi-dimensional feature matching and strategy generation: Based on the load characteristic map obtained in step 1, a multi-dimensional feature matching algorithm is used to determine whether the distribution network load characteristics meet the preset spatial transfer optimization conditions (such as high load fluctuation and strong complementarity) or time transfer optimization conditions (such as single high load fluctuation). Based on the matching results, the corresponding optimization strategy set is automatically generated;

[0060] Based on the load characteristic map constructed in Step 1, we designed an efficient multi-dimensional feature matching algorithm. This algorithm comprehensively considers the spatial distribution, temporal characteristics, fluctuation intensity, and complementary characteristics of adjacent regions to achieve accurate feature matching. Once a match is successful, the system automatically triggers the strategy generation engine, which rapidly generates a set of optimization strategies based on preset optimization objectives and constraints, such as load factor balance and N-1 safety. These strategies not only cover various optimization methods such as spatial and temporal shifting, but also utilize algorithmic optimization to ensure that each strategy achieves optimal load distribution and maximizes economic benefits while ensuring safe and stable operation of the power grid.

[0061] Step 3: Spatial Intelligent Transfer Optimization: For areas that meet the spatial transfer optimization criteria, a reinforcement learning-based intelligent planning optimization method is used to automatically adjust the switch configuration in the substation and network structure to achieve intelligent load switching and balanced distribution. Furthermore, IoT technology is used to monitor network status in real time and dynamically adjust operating modes to ensure N-1 security and optimal load factor.

[0062] For areas that meet the requirements for optimal spatial load shifting, we employ an intelligent planning and optimization method based on reinforcement learning. This method simulates and learns from various complex scenarios in power grid operation. Through trial and error and optimization, it finds the optimal load shedding and balanced distribution plan. In practice, we combine reinforcement learning algorithms with the physical model of the power grid to achieve intelligent load shifting by adjusting substation switch configurations and optimizing network structure. Furthermore, we leverage IoT technology to build a real-time monitoring system for comprehensive, around-the-clock monitoring of the power grid's operating status. This ensures that potential safety hazards can be promptly identified and addressed during the load shift process, safeguarding the grid's N-1 security.

[0063] Step 4: Time-based intelligent transfer optimization: For areas requiring only time-based transfer optimization, an innovative collaborative optimization mechanism for demand-side response and energy storage systems is introduced. Intelligent algorithms are used to predict future load trends, combined with user electricity behavior analysis to develop personalized energy storage charging and discharging plans. This plan also integrates new load types such as electric vehicle charging stations and thermal and cold storage systems to achieve smooth load transfer over time, reduce peak-to-valley fluctuations, and improve overall grid efficiency.

[0064] In terms of intelligent time-based transfer optimization, we have innovatively introduced a collaborative optimization mechanism combining demand-side response and energy storage systems. By using intelligent algorithms to accurately predict future load trends and analyzing user electricity usage behavior, we can tailor personalized energy storage charging and discharging plans for each user. At the same time, we actively integrate new load types such as electric vehicle charging stations and thermal and cold storage systems, achieving smooth load transfer over time through intelligent scheduling. This collaborative optimization mechanism not only effectively reduces peak-to-valley differences and improves the overall efficiency of the power grid, but also encourages users to actively participate in load regulation through incentive mechanisms, enhancing user-side flexibility and responsiveness.

[0065] Step 5: Cross-regional collaborative optimization: Based on steps 3 and 4, a cross-regional collaborative optimization strategy is further introduced. By utilizing the complementary characteristics of the distribution networks in adjacent regions, cross-regional power trading and shared energy storage methods can be used to achieve load balancing and resource optimization configuration on a larger scale.

[0066] Building on steps 3 and 4, we further proposed a cross-regional collaborative optimization strategy. This strategy leverages the complementary nature of adjacent regional distribution networks, achieving load balancing and resource optimization across a wider range through cross-regional power trading and shared energy storage. We established an efficient power market trading mechanism, leveraging big data analytics to monitor and forecast power market prices in real time, ensuring that trading decisions leverage the latest market price information to maximize economic benefits. Furthermore, we strengthened information sharing and collaboration mechanisms across cross-regional power grids, enhancing the security and reliability of the entire power grid through joint response to emergencies and coordinated planning for power grid development.

[0067] The process of intelligent data analysis and feature recognition also includes the prediction and evaluation of the impact of extreme weather and special events on holidays on load fluctuations, so as to improve the accuracy and adaptability of the optimization strategy.

[0068] In terms of special event prediction and assessment, we have introduced advanced weather forecast models and holiday consumption behavior analysis technology, combining historical data and real-time information to accurately predict and assess load fluctuations during extreme weather and holidays;

[0069] The load characteristics in step 1 specifically include:

[0070] The first step is data collection: load data from meters, substations, and power plants is collected from the power system. The time range of the data, including active power and reactive power, is determined based on the analysis requirements.

[0071] The second step is data preprocessing: remove outliers, missing values, and data points that do not meet the requirements; standardize and normalize the data for subsequent analysis; ensure that all data points are aligned on the time axis for time series analysis;

[0072] Third, feature extraction: Calculate the basic statistics of load data, such as mean, standard deviation, maximum, minimum, and peak-to-valley difference; extract the periodicity and trend characteristics of load data over time; and analyze the load correlation between different regions or different users.

[0073] The fourth step is pattern recognition: grouping load data based on similarity to identify different load patterns; using machine learning algorithms to classify load data and distinguish different types of loads; time series analysis: using time series models to predict and analyze load data trends;

[0074] Step 5: Spectrum generation: Based on the above data, a curve chart showing the change of load over time is drawn to intuitively display the load fluctuation; the depth of color is used to represent the size and distribution of the load, forming a heat map to show the spatial distribution characteristics of the load; the main components of the load data are extracted through the principal component analysis method, and drawn into a spectrum to show the main fluctuation characteristics of the load.

[0075] The coordinated optimization mechanism of demand-side response and energy storage systems also includes personalized modeling and prediction of user electricity consumption behavior, guiding users to actively participate in load regulation through incentive mechanisms, and improving the flexibility and response speed of the user side.

[0076] During the personalized modeling and prediction process, we collect and analyze user electricity usage data in a continuously expanding breadth and depth, including but not limited to electricity usage habits, device preferences, electricity price sensitivity, and other information. Through advanced machine learning algorithms, we can achieve more accurate electricity usage behavior predictions and personalized energy storage charging and discharging plan formulation.

[0077] Among them, the cross-regional collaborative optimization strategy also includes real-time monitoring and forecasting of electricity market prices, guided by economic benefits, to achieve optimal decision-making for cross-regional electricity transactions and promote the effective allocation and utilization of resources.

[0078] For real-time monitoring and forecasting of electricity market prices, we leverage big data analytics to build a power market monitoring network covering multiple adjacent regions. This enables high-frequency, high-precision market price forecasts, providing strong support for cross-regional power trading. We are also strengthening communication and collaboration mechanisms across interregional power grids to jointly address the challenges and opportunities of grid development.

[0079] The multi-dimensional feature matching algorithm formula in step 2 is specifically as follows:

[0080]

[0081] Among them, x n Indicates the nth characteristic value of the current load characteristic, y n It represents the nth feature value of the preset optimization condition, and a is the total number of features.

[0082] The intelligent algorithm in step 4 is specifically a time series prediction algorithm, which includes an autoregressive integrated moving average model and an exponential smoothing algorithm, specifically:

[0083] Autoregressive integrated moving average model: φ(A)(y n -c)=θ(A)∈ t

[0084] Among them, φ(A) and θ(A) are polynomials about the backshift operator A, y tis the value of the time series at time n, c is a constant term, ∈ n is the error term.

[0085] Exponential smoothing algorithm: S i =αx i +(1-α)S i-1

[0086] Among them, S i is the smoothed value at time i, x i is the actual observation value at time i, and α is the smoothing parameter (0<α<1).

[0087] Among them, the introduction of cross-regional collaborative optimization strategy in step 5 involves the mathematical model of the optimization problem, specifically:

[0088] Mathematical model of optimization problem:

[0089] in,

[0090] · is the electricity generation cost of region i in period t;

[0091] · is the inter-regional electricity transaction cost of region i in period t;

[0092] · is the power transmission loss cost of region i in period t;

[0093] · is the possible income of region i in period t;

[0094] T is the total number of time periods.

[0095] Among them, the constraints in the mathematical model of the optimization problem are:

[0096] Electricity supply and demand balance:

[0097] in,

[0098] is the amount of power transferred from region j to region i in time period t;

[0099] is the power generation of region i in time period t;

[0100] and are the energy storage discharge and charge amounts of region i in time period t, respectively;

[0101] D i,t is the electricity demand of region i in period t.

[0102] Energy storage system constraints:

[0103] where E i,t is the energy storage power at the end of time period t in zone i;

[0104] ηchaand ηdisare the charging and discharging efficiencies of the energy storage system, respectively;

[0105] Δt is the time interval;

[0106] E i,min and E i,max are the minimum and maximum limits of the energy storage power, respectively.

[0107] It has to be noted that, in the present document, the terms "first", "second", etc. merely serve to identify different entities or actions and do not necessarily require or imply any actual relationship or order between these entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0108] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous further modifications and changes can be apparent to one skilled in the art without departing from the spirit and scope of the application, which is defined by the appended claims and their equivalents.

Claims

1. A distribution network optimization method adapted to large-capacity load transfer in a distribution network, characterized by: The specific steps of the distribution network optimization method are as follows: Step 1: Intelligent Data Analysis and Feature Identification: Utilizing big data and artificial intelligence technologies, we conduct in-depth mining of regional distribution network load fluctuation data. Combined with real-time load data from adjacent regional distribution networks, we intelligently identify regional distribution network load fluctuation patterns and the complementary load characteristics between adjacent regions, generating an accurate load characteristic map. Step 2: Multi-dimensional feature matching and strategy generation: Based on the load characteristic map obtained in step 1, a multi-dimensional feature matching algorithm is used to determine whether the load characteristics of the distribution network meet the preset spatial transfer optimization conditions or temporal transfer optimization conditions. According to the matching results, the corresponding optimization strategy set is automatically generated; Step 3: Spatial Intelligent Transfer Optimization: For areas that meet the spatial transfer optimization criteria, a reinforcement learning-based intelligent planning optimization method is used to automatically adjust the switch configuration in the substation and network structure to achieve intelligent load switching and balanced distribution. Furthermore, IoT technology is used to monitor network status in real time and dynamically adjust operating modes to ensure N-1 security and optimal load factor. Step 4: Time-based intelligent transfer optimization: For areas requiring only time-based transfer optimization, an innovative collaborative optimization mechanism for demand-side response and energy storage systems is introduced. Intelligent algorithms are used to predict future load trends, combined with user electricity behavior analysis to develop personalized energy storage charging and discharging plans. This plan also integrates new loads such as electric vehicle charging stations and thermal and cooling storage systems to achieve smooth load transfer over time, reduce peak-to-valley variations, and improve overall grid efficiency. Step 5: Cross-regional collaborative optimization: Building on steps 3 and 4, a cross-regional collaborative optimization strategy is introduced. This leverages the complementary nature of adjacent regional distribution networks, enabling cross-regional power trading and shared energy storage to achieve load balancing and resource optimization across a wider range. The load characteristics in step 1 specifically include: The first step is data collection: load data from meters, substations, and power plants are collected from the power system. The time range of the data, including active power and reactive power, is determined based on the analysis requirements. The second step is data preprocessing: remove outliers, missing values, and data points that do not meet the requirements; standardize and normalize the data for subsequent analysis; ensure that all data points are aligned on the time axis for time series analysis; The third step is feature extraction: calculate the basic statistics of load data, such as mean, standard deviation, maximum, minimum, and peak-to-valley difference; extract the periodicity and trend characteristics of load data over time; Analyze the load correlation between different areas or different users; The fourth step is pattern recognition: grouping load data based on similarity to identify different load patterns; using machine learning algorithms to classify load data and distinguish different types of loads; time series analysis: using time series models to predict and analyze load data trends; Step 5: Graph generation: Based on the above data, a load variation curve is drawn over time to visually display the load fluctuation. The load size and distribution are represented by color depth, forming a heat map to show the spatial distribution characteristics of the load. The main components of the load data are extracted through principal component analysis and plotted into a spectrum to show the main fluctuation characteristics of the load. The introduction of the cross-region collaborative optimization strategy in step 5 involves a mathematical model of the optimization problem, specifically: Mathematical model of optimization problem: in, · is the electricity generation cost of region i in period t; · is the inter-regional electricity transaction cost of region i in period t; · is the power transmission loss cost of region i in period t; · is the possible income of region i in period t; T is the total number of time periods.

2. A distribution network optimization method adapted to large-capacity load transfer in a distribution network according to claim 1, characterized in that: The data intelligent analysis and feature recognition process also includes the prediction and evaluation of the impact of extreme weather and special events on holidays on load fluctuations, so as to improve the accuracy and adaptability of the optimization strategy.

3. The method for optimizing a distribution network adapted to large-capacity load transfer according to claim 1, characterized in that: The demand-side response and energy storage system collaborative optimization mechanism also includes personalized modeling and prediction of user electricity consumption behavior, guiding users to actively participate in load regulation through incentive mechanisms, thereby improving the flexibility and response speed of the user side.

4. The method for optimizing a distribution network adapted to large-capacity load transfer according to claim 1, characterized in that: The cross-regional collaborative optimization strategy also includes real-time monitoring and forecasting of electricity market prices, guided by economic benefits, to achieve optimal decision-making for cross-regional electricity transactions and promote the effective allocation and utilization of resources.

5. The method for optimizing a distribution network adapted to large-capacity load transfer according to claim 1, characterized in that: The multi-dimensional feature matching algorithm formula in step 2 is specifically: Among them, x n Indicates the nth characteristic value of the current load characteristic, y n It represents the nth feature value of the preset optimization condition, and a is the total number of features.

6. The method for optimizing a distribution network adapted to large-capacity load transfer according to claim 1, characterized in that: The intelligent algorithm in step 4 is specifically a time series prediction algorithm, which includes an autoregressive integrated moving average model and an exponential smoothing algorithm, specifically: Autoregressive integrated moving average model: φ(A)(y n -c)=θ(A)∈ t Among them, φ(A) and θ(A) are polynomials about the backshift operator A, y t is the value of the time series at time n, c is a constant term, ∈ t is the error term, Exponential smoothing algorithm: S i =αx i +(1-α)S i-1 Among them, S i is the smoothed value at time i, x i is the actual observation value at time i, and α is the smoothing parameter (0<α<1).

7. The method for optimizing a distribution network adapted to large-capacity load transfer according to claim 1, characterized in that: The constraints in the mathematical model of the optimization problem are: Electricity supply and demand balance: in, is the amount of power transferred from region j to region i in time period t; is the power generation of region i in time period t; and are the energy storage discharge and charge amounts of region i in time period t, respectively; D i,t is the electricity demand of region i in period t, Energy storage system constraints: Among them, E i,t is the energy storage capacity of region i at the end of period t; ηcha and ηdis are the charging and discharging efficiencies of the energy storage system, respectively; Δt is the time interval; E i,min and E i,max They are the minimum and maximum limits of energy storage capacity respectively.

Images