Differentiated dispatching method, device and equipment for renewable energy direct current system

Abstract

The application provides a renewable energy direct current system differentiation scheduling method, device and equipment, and relates to the technical field of renewable energy management. The method comprises the following steps: obtaining climate data of multiple regions in a set month, performing clustering analysis to divide the multiple regions into several typical regions; obtaining historical load data of the multiple regions in the set month, and generating a typical household load curve corresponding to each typical region according to the historical load data; based on the typical household load curve, a direct current micro-grid system model comprising a photovoltaic power generation unit, a wind power generation unit and an energy storage unit is constructed, and the model is subjected to multi-objective optimization based on NSGA-II to obtain a differentiated equipment configuration and operation scheduling strategy suitable for each typical region; wherein the multi-objective optimization takes the minimization of system economy and the maximization of user comfort as optimization objectives. The differentiated equipment configuration and operation scheduling strategy prepared by the application can balance the economy and user satisfaction simultaneously.

Description

Technical Field

[0001] This invention relates to the field of renewable energy management technology, and in particular to a method, apparatus and equipment for differentiated scheduling of renewable energy DC systems. Background Technology

[0002] Existing planning methods for rural hybrid renewable energy systems mostly focus on personalized designs for specific sites. When large-scale deployment is required, these methods become costly and inefficient due to a lack of universality and systematic approach. While some studies have attempted regional division, they typically rely on single resource indicators or static boundaries such as administrative boundaries or traditional climate zones. This crude zoning approach fails to deeply integrate the multi-dimensional climate and resource characteristics closely related to energy system performance, resulting in a severe disconnect between the zoning results and the actual modeling and optimization requirements of the energy system. It fails to accurately reflect the fundamental differences in energy supply and load demand across different regions, thus hindering effective regional differentiated planning.

[0003] Furthermore, with the improvement of rural living standards and the widespread use of flexible electrical equipment, user comfort has become a key indicator for measuring the success of a system. Focusing solely on economics often results in system configurations that fail to meet user comfort needs, leading to low user satisfaction and even the abandonment of renewable energy systems.

[0004] Therefore, how to balance the conflicting goals of economy and user satisfaction during the planning stage and formulate a system solution that is both economical and humane has become a pressing technical challenge for the development of rural energy systems. Summary of the Invention

[0005] This invention provides a method, apparatus, and equipment for differentiated scheduling of renewable energy DC systems to address the problem of simultaneously balancing the conflicting goals of economic efficiency and user satisfaction during the planning phase.

[0006] In a first aspect, embodiments of the present invention provide a differentiated dispatch method for a renewable energy DC system, comprising:

[0007] Climate data for multiple regions in a set month is acquired, cluster analysis is performed on the climate data, and the multiple regions are divided into several typical zones based on the cluster classification results; wherein, the climate data includes average temperature, wind speed and solar irradiance.

[0008] Historical load data for multiple regions in a set month are obtained, and typical household load curves corresponding to each typical zone are generated based on the historical load data.

[0009] Based on the typical household load curves, a DC microgrid system model including photovoltaic power generation units, wind power generation units, and energy storage units is constructed. The model is then subjected to multi-objective optimization using a Non-dominated Sorting Genetic Algorithm II (NSGA-II) to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. The multi-objective optimization aims to minimize system economics and maximize user comfort. System economics includes annualized equipment investment, operation and maintenance costs, and grid purchase and sale costs. User comfort is quantified based on flexible load satisfaction.

[0010] In one possible implementation, the cluster analysis of the climate data includes:

[0011] A multidimensional feature matrix is ​​constructed based on the climate data;

[0012] The K-medoids clustering algorithm is used to perform cluster analysis on the feature matrix.

[0013] In one possible implementation, the clustering analysis of the feature matrix using the K-medoids clustering algorithm includes:

[0014] Each characteristic parameter in the climate data is normalized.

[0015] K data points are randomly selected from the normalized data as the initial cluster centers;

[0016] Calculate the Euclidean distance between the remaining data points and the centers of each initial cluster, and assign each data point to the corresponding cluster according to the principle of closest distance;

[0017] Recalculate the cluster center for each cluster and update the initial cluster centers;

[0018] The iterative process assigns each data point to the corresponding cluster based on the nearest principle and updates the cluster center until the cluster center no longer changes or the preset number of iterations is reached, in order to obtain the final clustering classification result.

[0019] In one possible implementation, the designated month includes January and July.

[0020] In one possible implementation, generating typical household load curves for each typical zone based on the historical load data includes:

[0021] For households within a target typical zone, determine the activation probability distribution of at least one electrical device during the heating season and the non-heating season; wherein, the target typical zone is one of the typical zones.

[0022] Based on the activation probability distribution, the Monte Carlo method is used to simulate and generate the start-stop event sequence of each electrical device within the simulation period, and the start-stop event sequence is converted into a continuous instantaneous power curve according to the rated power of each electrical device.

[0023] At each time step, the instantaneous power curves of each electrical device in the target typical partition are superimposed to generate multiple simulated household total load curves.

[0024] From the multiple simulated total household load curves, the curve with the smallest root mean square error between it and the average load curve of all simulated curves is selected as the typical household load curve corresponding to the target typical zone.

[0025] In one possible implementation, the multi-objective optimization of the DC microgrid system model based on NSGA-II yields differentiated equipment configuration and operation scheduling strategies applicable to each typical zone, including:

[0026] Initialize the population and algorithm parameters; whereby each individual in the population represents a set of candidate solutions including optimization parameters, time window parameters, and target parameters; the optimization parameters include equipment configuration parameters and operation scheduling parameters; the equipment configuration parameters include the number of photovoltaic panels, the number of wind turbines, and the number of batteries; the operation scheduling parameters include flexible load mode, air conditioning temperature, and electric vehicle SOC level; the time window parameters include equipment start-up and shutdown time and charging / discharging period; the target parameters include investment cost, maintenance cost, electricity price, and user satisfaction;

[0027] Simulations are performed on each individual in the current population to calculate its annualized total cost under the system's economic objective and its comprehensive satisfaction value under the user comfort objective, and to determine whether the preset system operation constraints are met; wherein, the annualized total cost includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs;

[0028] Perform fast non-dominated sorting on the current population, divide all individuals into multiple non-dominated levels based on the annualized total cost value and the comprehensive satisfaction value, and calculate the crowding degree of individuals within the same non-dominated level;

[0029] Selection is performed based on the non-dominated hierarchy and crowding of individuals, and crossover and mutation operations are performed on the selected individuals to generate offspring populations;

[0030] The parent population is merged with the offspring population. Non-dominated sorting and crowding calculation are performed on the merged population again. Individuals with the same number as the parent population size are selected from the sorting and crowding based on the sorting and crowding to form a new generation population.

[0031] The simulation, sorting, genetic operations and population update steps are executed iteratively until the iteration termination condition is met. The final set of individuals in the first non-dominated level of the population is output as the Pareto optimal solution set.

[0032] For each typical partition, an equilibrium point is selected from the corresponding Pareto optimal solution set, and the equipment configuration parameters and operation scheduling parameters corresponding to the equilibrium point are determined as the differentiated equipment configuration and operation scheduling strategy for that typical partition.

[0033] In one possible implementation, before performing multi-objective optimization of the DC microgrid system model based on NSGA-II, constraints are set for the DC microgrid system model.

[0034] The constraints include: constraints on the number of photovoltaic panels installed, constraints on the number of wind turbines installed, constraints on the number of batteries installed, constraints on the charging and discharging power of batteries and the upper and lower limits of their state of charge, and constraints on the real-time power balance of the system at each scheduling moment.

[0035] In one possible implementation, the quantification of user comfort includes:

[0036] Determine a satisfaction calculation function for at least one type of flexible load; wherein, the flexible load includes air conditioning, electric vehicles, and movable loads;

[0037] Calculate the satisfaction values ​​of various flexible loads during the scheduling cycle;

[0038] Based on preset weighting coefficients, the satisfaction values ​​of various flexible loads are weighted and summed to obtain a quantitative index of user comfort.

[0039] Secondly, embodiments of the present invention provide a differentiated dispatching device for a renewable energy DC system, comprising:

[0040] The partitioning module is used to acquire climate data for multiple regions in a set month, perform cluster analysis on the climate data, and partition the multiple regions according to the clustering classification results to obtain several typical partitions; wherein, the climate data includes average temperature, wind speed and solar irradiance.

[0041] The load curve generation module is used to acquire historical load data of multiple regions in a set month, and generate typical household load curves corresponding to each typical zone based on the historical load data.

[0042] The scheduling module is used to construct a DC microgrid system model including photovoltaic power generation units, wind power generation units, and energy storage units based on the typical household load curve, and to perform multi-objective optimization on the DC microgrid system model based on NSGA-II to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. Among them, the multi-objective optimization aims to minimize system economy and maximize user comfort. The system economy includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs, and the user comfort is quantified based on flexible load satisfaction.

[0043] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0044] In this embodiment of the invention, by acquiring and clustering climate data such as average temperature, wind speed, and solar irradiance for multiple regions in a set month, the climate-energy characteristics of different regions can be accurately extracted. This breaks through the limitations of traditional personalized design, which lacks universality and the rough zoning that is out of touch with the needs of the energy system. Combined with historical load data, typical household load curves are generated to ensure that the system modeling fits the actual energy consumption scenario. Subsequently, a DC microgrid model containing photovoltaic, wind power, and energy storage units is constructed, and the NSGA-II algorithm is used to optimize the dual objectives of economy and user comfort. This achieves reasonable control of the annualized equipment investment, operation and maintenance, and grid purchase and sale costs of the system. At the same time, the user comfort experience is guaranteed by quantifying the flexible load satisfaction. The resulting differentiated equipment configuration and operation scheduling strategy realizes precise customization from macro resource assessment to micro system design, effectively improving the comprehensive performance and applicability of rural renewable energy DC systems. Attached Figure Description

[0045] Figure 1 This is an application scenario diagram of the differentiated scheduling method for renewable energy DC systems provided in an embodiment of the present invention;

[0046] Figure 2 This is a flowchart illustrating a differentiated dispatch method for a renewable energy DC system according to an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of a rural hybrid renewable energy DC microgrid system provided in one embodiment of this application;

[0048] Figure 4 This is a schematic diagram of a climate data clustering process provided in an embodiment of the present invention;

[0049] Figure 5This is a schematic diagram of the process for generating typical household load curves using the Monte Carlo stochastic simulation method according to an embodiment of the present invention;

[0050] Figure 6 This is a flowchart illustrating the multi-objective optimization of a DC microgrid system model based on NSGA-II, according to an embodiment of the present invention.

[0051] Figure 7 This is a schematic diagram of the structure of a differentiated dispatching device for a renewable energy DC system provided in an embodiment of the present invention. Detailed Implementation

[0052] This invention aims to propose a differentiated design method for rural hybrid renewable energy DC systems based on climate-resource zoning, so as to achieve precise and customized design from macro-resource assessment to micro-system configuration and operation scheduling.

[0053] First, a quantifiable, clusterable, and mappable climate-resource characteristic zoning method for energy system design is established at a national scale. Based on renewable energy endowment data such as climate conditions, solar radiation intensity, and wind speed across the country, a climate-energy characteristic matrix is ​​constructed. The K-medoids clustering algorithm is then used to perform typical regional divisions and feature extraction, providing a scientific basis for subsequent regionalized system design. Next, combining the energy characteristics of each typical region with rural residents' energy consumption behavior, Monte Carlo stochastic simulation is used to generate typical annual household load curves that conform to actual patterns, ensuring the representativeness and authenticity of the system model. Finally, considering the energy supply and demand characteristics of different regions, a multi-energy coupled DC microgrid model including photovoltaic power generation units, wind power generation units, and battery energy storage units is established. NSGA-II is introduced for multi-objective optimization design, with the dual optimization objectives of minimizing system economics (including annualized equipment investment, operation and maintenance, and grid purchase and sale costs) and maximizing user comfort (flexible load satisfaction).

[0054] Through collaborative optimization and iterative evolution, a solution set that achieves the optimal balance between economy and comfort is obtained. By combining the climate resource endowment and energy consumption differences in various regions, a highly adaptable and economically reasonable equipment capacity configuration and operation control strategy is formed, realizing the regional differentiated optimization design and comprehensive performance improvement of rural hybrid renewable energy DC systems.

[0055] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0056] Figure 1 This diagram illustrates an application scenario of the differentiated scheduling method for renewable energy DC systems provided in this embodiment of the invention. Figure 1As shown, monitoring centers for climate and load data are set up in different regions. Regional servers are responsible for collecting and storing climate and load data for their respective regions. A centralized server is set up to formulate differentiated dispatch schemes for the renewable energy DC system. This centralized server communicates with each regional server to obtain climate and load data from each region. It can perform cluster analysis based on the data from each region to determine differentiated dispatch schemes to adapt to regional differences. Furthermore, a database is set up corresponding to the centralized server to centrally store the data from each regional server, facilitating access by the centralized server, and also to store the differentiated dispatch schemes determined by the centralized server, making it convenient for each regional server to access them.

[0057] See Figure 2 The flowchart illustrating the implementation of the differentiated scheduling method for renewable energy DC systems provided in this embodiment of the invention is described in detail below:

[0058] S201: Obtain climate data for multiple regions in a specified month, perform cluster analysis on the climate data, and divide the multiple regions into several typical zones based on the cluster classification results; among them, the climate data includes average temperature, wind speed and solar irradiance.

[0059] The execution subject of each embodiment of this application can be a server, processor, microprocessor, or other device with data processing capabilities. In actual implementation, the specific implementation method of the execution subject can be selected according to actual needs. This embodiment does not impose any special restrictions on this, as long as it is a device with data processing capabilities.

[0060] In practice, the regional scope corresponding to the climate data can be determined according to the needs. Optionally, the average temperature, wind speed and solar irradiance of a set month can be obtained nationwide to achieve the division of typical climate-energy regions nationwide.

[0061] In one possible implementation, the designated months include January and July. Specifically, by defining January and July as the designated months, corresponding to extreme winter and summer climate conditions respectively, the differences in climate characteristics across regions can be comprehensively and critically characterized. The clustering and partitioning results based on data from these months better reflect the energy supply and demand patterns of different regions under seasonal variations. This allows the subsequently generated typical load curves and system optimization schemes to better adapt to the climate characteristics of different seasons, further enhancing the adaptability and operational stability of the renewable energy DC system to different regional climates.

[0062] In other possible implementations, the months can be set to include January, February, July, and August, or the months can be set to include December, January, February, June, July, and August.

[0063] The data sets are expanded from January and July to include January, February, July, and August, or December, January, February, June, July, and August. This approach retains the core months corresponding to extreme winter and summer climates while incorporating key periods of the winter-summer transition. This allows for a more comprehensive and continuous capture of the evolution of climate characteristics in various regions, avoiding the potential omission of seasonal climate differences during transitional periods when relying solely on single-month data. The multi-dimensional feature matrix constructed based on this richer monthly climate data can more nuancedly reflect the differences in energy endowment across different regions during winter, summer, and the winter-summer transition period. This ensures that the clustering and zoning results not only reflect regional characteristics under extreme climates but also accurately cover the climate features of transitional seasons, making the zoning more closely aligned with the actual situation of annual climate fluctuations.

[0064] Furthermore, by performing cluster analysis on climate data, data redundancy can be effectively eliminated, and representative regional characteristic parameters can be extracted, enabling the system optimization design to be deployed in a targeted manner based on the resource endowment of different regions.

[0065] S202: Obtain historical load data for multiple regions in a set month, and generate typical household load curves for each typical zone based on the historical load data.

[0066] Historical load data for the specified months was obtained from the aforementioned regions. This historical load data includes operating power consumption records of various electrical devices and data related to user energy consumption behavior. For each typical zone, a corresponding typical household load curve was generated by combining its climate characteristics with historical load data.

[0067] During the generation process, the impact of regional climate on energy consumption behavior should be fully considered, and the randomness, seasonality and adjustability of load should be reproduced to ensure that the generated typical household load curve can truly reflect the actual energy consumption pattern of households in the region, and provide reliable data support for system modeling and optimization.

[0068] S203, based on typical household load curves, constructs a DC microgrid system model including photovoltaic power generation units, wind power generation units, and energy storage units. Based on NSGA-II, the DC microgrid system model is optimized in multiple objectives to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. Among them, the multi-objective optimization aims to minimize system economy and maximize user comfort. System economy includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs, while user comfort is quantified based on flexible load satisfaction.

[0069] Figure 3 This is a schematic diagram of a rural hybrid renewable energy DC microgrid system provided in an embodiment of this application, as shown below. Figure 3As shown, a model of a rural hybrid renewable energy DC microgrid system is constructed, incorporating photovoltaic and wind power generation units and batteries. This aims to promote the transformation of buildings from traditional single-energy consumption terminals to integrated energy nodes that combine energy production, consumption, and flexible regulation. The system uses a DC bus as its core hub, enabling unified access and coordinated scheduling of various energy inputs, different types of loads, and energy storage devices.

[0070] On the energy supply side, photovoltaic, wind power, and the power grid are connected to the busbars respectively, providing energy to various indoor loads and energy storage units. The three energy sources form a complementary operating mechanism, effectively improving the system's energy self-sufficiency and fluctuation suppression capabilities. Surplus electricity can be fed back to the public power grid, realizing the external output of green energy and further enhancing the system's sustainability, stability, and economy.

[0071] On the load side, the system configuration includes fixed loads and flexible loads. The former has stable power consumption characteristics, while the latter has time-adjustable or power-controllable characteristics, enabling intelligent start-up and dynamic adjustment based on renewable energy output, electricity price signals, or grid conditions, within the range permissible for user comfort. Through the proactive response of flexible loads, the system can effectively participate in demand-side management, improving overall operational flexibility and the absorption of renewable energy.

[0072] The energy storage system consists of stationary batteries and user-owned electric vehicles. The battery system stores energy when photovoltaic or wind power output is surplus, and releases it during peak load periods or when renewable energy output is insufficient, thus playing a role in peak shaving and valley filling and enhancing power supply reliability. Electric vehicles, as mobile energy storage units, can be integrated into the overall system energy management strategy, improving the dispatchability and comprehensive utilization efficiency of energy storage resources.

[0073] Among them, based on the typical household load curves of each typical zone, a DC microgrid system model integrating photovoltaic power generation units, wind power generation units and energy storage units is constructed. This model needs to achieve coordinated linkage between energy supply, load consumption and energy storage regulation to meet the diversified electricity needs of households.

[0074] In actual implementation, the multi-objective optimization process involves solving a bi-objective optimization function with complex constraints. The two objectives are to minimize the system's total life cycle cost (i.e., minimize system economy) and maximize user satisfaction. In the case of a given month, a key component of user satisfaction is the comfort of the user's living environment.

[0075] Optionally, the objective function can be represented as follows:

[0076]

[0077]

[0078] In the formula, Annual investment and construction costs for the equipment; Annual operating and maintenance costs of the equipment; The system's annual electricity purchase cost; This represents the system's annual electricity sales cost; Purchase cost per unit capacity of each piece of equipment; The rated power of each device is equal to the rated power of a single device multiplied by the number of devices. The equipment discount rate; Indicates the service life of the equipment.

[0079] In the objective function, the function representing user comfort is as follows:

[0080]

[0081] In the formula, Overall user satisfaction; , , The satisfaction levels were for electric vehicles, air conditioning, and portable loads, respectively. , , Satisfaction coefficients for electric vehicles, air conditioners, and movable loads are required. .

[0082] In practical implementation, a Pareto optimal solution set is obtained through iterative algorithmic evolution, which fully reflects the trade-off between economy and comfort. For each typical region, considering its climate-resource endowment and energy consumption habits, a suitable balance point is selected from the corresponding Pareto optimal solution set to determine the differentiated equipment configuration scheme and operation scheduling strategy for that region, thereby achieving optimal overall system performance.

[0083] In this embodiment, by acquiring and clustering climate data such as average temperature, wind speed, and solar irradiance for multiple regions in a set month, the climate-energy characteristics of different regions can be accurately extracted. This breaks through the limitations of traditional personalized design, which lacks universality and the rough zoning that is out of touch with the energy system's needs. Combined with historical load data, typical household load curves are generated to ensure that the system modeling fits the actual energy consumption scenario. Subsequently, a DC microgrid model containing photovoltaic, wind power, and energy storage units is constructed, and the NSGA-II algorithm is used for dual-objective optimization of economy and user comfort. This achieves reasonable control of the system's annualized equipment investment, operation and maintenance, and grid purchase and sale costs, while ensuring user comfort through flexible load satisfaction quantification. The resulting differentiated equipment configuration and operation scheduling strategy realizes precise customization from macro resource assessment to micro system design, effectively improving the comprehensive performance and applicability of rural renewable energy DC systems.

[0084] In one possible implementation, cluster analysis is performed on the climate data, including:

[0085] Constructing a multidimensional feature matrix based on climate data;

[0086] The K-medoids clustering algorithm is used to perform cluster analysis on the feature matrix.

[0087] Among them, a multi-dimensional feature matrix is ​​constructed based on climate data. This matrix takes each region to be planned as an independent sample and uses average temperature, wind speed and solar irradiance as feature dimensions. It integrates scattered single-dimensional climate data into a structured analysis carrier, comprehensively characterizes the climate-resource integrated characteristics of each region, and provides a standardized data foundation for subsequent cluster analysis.

[0088] The K-medoids clustering algorithm was used to perform cluster analysis on the above multidimensional feature matrix. This algorithm uses actual data points as cluster centers, which can effectively reduce the impact of outliers on the clustering results and improve the accuracy of partitioning.

[0089] Among other possible implementation methods, clustering algorithms such as K-means clustering and fuzzy C-means clustering can be flexibly selected according to the actual scenario.

[0090] In this embodiment, by constructing a feature matrix containing multi-dimensional climate data and using the K-medoids clustering algorithm for analysis, the limitations of traditional single resource indicators or static boundary partitioning are overcome. This approach can deeply integrate climate and resource characteristics closely related to energy system performance, making the clustering partitioning results more in line with the actual needs of energy system modeling and optimized operation. This provides a scientific and targeted partitioning basis for the subsequent differentiated system design of each region, helping to achieve accurate matching of energy supply and demand characteristics in different regions.

[0091] In one possible implementation, the K-medoids clustering algorithm is used to perform cluster analysis on the feature matrix, including:

[0092] Each characteristic parameter in the climate data is normalized separately.

[0093] K data points are randomly selected from the normalized data as the initial cluster centers;

[0094] Calculate the Euclidean distance between the remaining data points and the centers of each initial cluster, and assign each data point to the corresponding cluster according to the principle of closest distance;

[0095] Recalculate the cluster center for each cluster and update the initial cluster centers;

[0096] The iterative process assigns each data point to the corresponding cluster based on the nearest principle and updates the cluster center until the cluster center no longer changes or the preset number of iterations is reached, in order to obtain the final clustering classification result.

[0097] During clustering, outdoor environmental features such as temperature, solar irradiance, and wind speed have different dimensions and numerical ranges. Directly using Euclidean distance for similarity calculations can lead to biases. To avoid this problem, it is necessary to normalize these features to achieve equivalent comparisons under the same dimension, thereby obtaining more reasonable cluster partitions and cluster centers. In specific implementation,

[0098] The initial data is normalized using the following formula:

[0099]

[0100] In the formula, This represents the normalized input data. This represents the original input data. This represents the minimum value of the original input data. This represents the maximum value of the original input data. After normalization, the amplitudes of the four types of data can be fixed within the range of [0,1].

[0101] The Euclidean distance is calculated as follows:

[0102]

[0103] In the formula, It is an n-dimensional data point. Let j be the center point of the j-th cluster. It is the feature dimension. Let be the coordinates of the center point of the j-th cluster in the i-th dimension.

[0104] To facilitate understanding of the specific process of cluster analysis of climate data Figure 4 A schematic diagram of a climate data clustering process is shown, including steps such as data input, data preprocessing, determining initial cluster centers, data point allocation, cluster center updating, iteration conditions and processes, and clustering result output. The example uses January and July as the set months.

[0105] In addition, such as Figure 4 As shown, during the process of updating the initial cluster center, a criterion function is calculated for each member point in the cluster, and the data point that minimizes the criterion function is selected as the new cluster center.

[0106] The mathematical form of the criterion function is:

[0107]

[0108] In the formula, The total squared error of the clustering results; For each cluster class The samples in.

[0109] In this embodiment, the characteristic parameters of the climate data are first normalized to effectively eliminate similarity calculation biases caused by different units and numerical ranges. Then, through a series of K-medoids clustering operations, such as randomly selecting initial cluster centers, allocating data points based on Euclidean distance, and iteratively updating cluster centers, the clustering process can be made more scientific and reasonable. The final clustering results can more accurately reflect the essential differences in climate-energy characteristics of different regions, and the extracted regional characteristic parameters are more representative, laying a solid foundation for the targeted deployment of subsequent system optimization design.

[0110] In one possible implementation, typical household load curves for each typical zone are generated based on historical load data, including:

[0111] For households within a target typical zone, determine the activation probability distribution of at least one electrical device during the heating season and the non-heating season; wherein, the target typical zone is one of the typical zones.

[0112] Based on the activation probability distribution, the Monte Carlo method is used to simulate and generate the start-stop event sequence of each electrical device in the simulation period, and the start-stop event sequence is converted into a continuous instantaneous power curve according to the rated power of each electrical device.

[0113] At each time step, the instantaneous power curves of each electrical device in the target typical partition are superimposed to generate multiple simulated household total load curves.

[0114] From multiple simulated total household load curves, the curve with the smallest root mean square error between it and the average load curve of all simulated curves is selected as the typical household load curve corresponding to the target typical area.

[0115] In this embodiment of the application, the Monte Carlo random simulation method is used to generate a highly reliable typical household load curve.

[0116] like Figure 5 A schematic flowchart illustrating the generation of typical household load curves using a Monte Carlo stochastic simulation method is shown in one embodiment. (Combined with...) Figure 5 The process of generating a typical household load curve is introduced, mainly including:

[0117] 1) Simulation Framework Initialization and Parameter Setting: First, establish the spatiotemporal scale and boundary conditions of the simulation, including the total simulation period, the duration of a single simulation, the time step, and the total number of simulations. The key setting is to define differentiated operating parameters for various types of electrical equipment during the "heating season" and "non-heating season" to accurately reflect the systematic impact of seasonal climate changes on energy consumption behavior.

[0118] 2) Generation of Device Event Sequences Based on Probability Distributions: For each type of electrical equipment (such as air conditioners, water heaters, electric vehicles, etc.), based on its specific daily activation probability distribution model, Monte Carlo random sampling is used to generate the specific activation time and frequency of each device within a day. This process iterates over all devices and all simulated days, ultimately outputting a discrete device start-up and shutdown event sequence spanning a long time.

[0119] 3) Synthesis and Aggregation of Load Power Curves: The discrete device start-stop events generated in the previous step are converted into continuous instantaneous power curves based on the rated power and operating characteristics of each device. Subsequently, at each time step, the power values ​​of all devices are algebraically superimposed to synthesize the total load curve for a single household. This process generates a complete annual time-series load data for each independent simulation.

[0120] 4) After completing numerous simulations and obtaining a sample set of load curves, statistical methods are used to select the most representative annual load curve. This invention defines the average load curve of all simulation results as the expected baseline, and calculates the root mean square error (RMSE) between each simulation curve and this baseline. Finally, the simulation curve with the smallest RMS error is selected as the typical annual load curve output. This selection criterion ensures that the selected curve faithfully reflects the system's average expected load while preserving the random fluctuation characteristics of the actual load to the greatest extent possible.

[0121] In this embodiment, the activation probability distribution of electrical equipment is set for typical zones during the heating and non-heating seasons. The Monte Carlo method is used to simulate and generate equipment start-up and shutdown event sequences, which can quantify the fuzzy user energy consumption behavior into an analyzable probability distribution, reproduce the randomness and seasonality of the load. Then, by superimposing instantaneous power curves and selecting the curve with the smallest root mean square error with the average load curve as the typical household load curve, the generated load curve is ensured to have extremely high reliability and representativeness, accurately characterizing the temporal characteristics and adjustable potential of flexible loads, and providing reliable and realistic core data support for subsequent multi-objective optimization.

[0122] In one possible implementation, multi-objective optimization is performed on the DC microgrid system model based on NSGA-II to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical region, including:

[0123] Initialize the population and algorithm parameters; each individual in the population represents a set of candidate solutions including optimization parameters, time window parameters, and target parameters; optimization parameters include equipment configuration parameters and operation scheduling parameters; equipment configuration parameters include the number of photovoltaic panels, wind turbines, and batteries; operation scheduling parameters include flexible load mode, air conditioning temperature, and electric vehicle SOC level; time window parameters include equipment start-up and shutdown time and charging / discharging period; target parameters include investment cost, maintenance cost, electricity price, and user satisfaction.

[0124] Simulations are performed on each individual in the current population to calculate its annualized total cost under the system's economic objective and its comprehensive satisfaction value under the user comfort objective, and to determine whether the preset system operation constraints are met; whereby the annualized total cost includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs;

[0125] Perform fast non-dominated sorting on the current population, divide all individuals into multiple non-dominated levels based on the annualized total cost and overall satisfaction value, and calculate the crowding degree of individuals within the same non-dominated level;

[0126] Selection is performed based on the non-dominated hierarchy and crowding of individuals, and crossover and mutation operations are performed on the selected individuals to generate offspring populations;

[0127] The parent and offspring populations are merged, and non-dominated sorting and crowding calculation are performed on the merged population again. Based on the sorting and crowding, individuals with the same number as the parent population are selected to form a new generation population.

[0128] The simulation, sorting, genetic operations and population update steps are executed iteratively until the iteration termination condition is met. The final set of individuals in the first non-dominated level of the population is output as the Pareto optimal solution set.

[0129] For each typical partition, an equilibrium point is selected from its corresponding Pareto optimal solution set. The equipment configuration parameters and operation scheduling parameters corresponding to the equilibrium point are determined as the differentiated equipment configuration and operation scheduling strategy for that typical partition.

[0130] like Figure 6 A schematic diagram of the process for multi-objective optimization of a DC microgrid system model based on NSGA-II is shown according to one embodiment.

[0131] Building upon the aforementioned embodiments, this embodiment employs the NSGA-II algorithm to solve a bi-objective optimization problem with complex constraints. Through iterative evolution, this algorithm ultimately outputs a Pareto optimal solution set under given constraints, which comprehensively represents the trade-off between economy and comfort. For each typical climate-resource zone, from its corresponding Pareto optimal solution set, a most adaptive equilibrium point is selected as the solution based on local climate, resource endowment, and energy consumption habits. This solution explicitly provides the optimal equipment capacity configuration and system operation strategy. Furthermore, based on this operation strategy, detailed electricity consumption plans for flexible loads are optimized (such as preferred charging times for electric vehicles and start-up and shutdown times for water heaters), thereby achieving lean coordination and matching between the source and load sides, forming the final regionally differentiated design.

[0132] like Figure 6 As shown, this is the initialization phase for the population and algorithm parameters. Algorithm parameters include population size and number of iterations. Population individuals are determined using typical partitions, and each population individual includes at least one set of candidate solutions containing optimization parameters, time window parameters, and target parameters, with the goal of optimizing device capacity configuration and system operation strategy.

[0133] In one specific embodiment, the criteria for evaluating the population objective function are as follows:

[0134] Objective 1: Total Cost = Annual Equipment Investment Cost + Annual Equipment Operation and Maintenance Cost + Annual Electricity Purchase Cost from Grid - Electricity Sales Revenue

[0135] Objective 2: User satisfaction = 0.192 × Air conditioning satisfaction + 0.634 × Flexible load satisfaction + 0.174 × Electric vehicle satisfaction

[0136] In addition, the following four aspects should be considered when calculating the degree of constraint violation:

[0137] (1) Satisfaction constraint: Overall user satisfaction ≥ 0.2

[0138] (2) Cost constraint: Annualized total cost ≤ 5000 yuan

[0139] (3) Equipment constraints: ≤30 photovoltaic units, ≤5 wind turbines, ≤5 storage batteries

[0140] (4) Time constraints: Each device shall operate within the allowed time window.

[0141] In this embodiment, by initializing a population containing various parameters such as equipment configuration and operation scheduling, simulation calculations are performed on individuals, and genetic operations are carried out in combination with non-dominated sorting and crowding analysis. After multiple rounds of iteration, a Pareto optimal solution set is obtained. Then, suitable equilibrium points are selected for each typical partition, which can efficiently search for the optimal trade-off between economy and comfort. This solution not only fully matches the climate resource endowment and energy consumption habits of the corresponding region, but also clarifies the specific equipment configuration parameters and operation scheduling strategies, realizing precise coordination of source, load and storage. This makes the system planning scheme both scientifically consistent and decision-making flexible, significantly improving the practicality of the project.

[0142] In one possible implementation, before performing multi-objective optimization of the DC microgrid system model based on NSGA-II, it also includes setting constraints for the DC microgrid system model.

[0143] The constraints include: constraints on the number of photovoltaic panels installed, constraints on the number of wind turbines installed, constraints on the number of batteries installed, constraints on the charging and discharging power of batteries and the upper and lower limits of their state of charge, and constraints on the real-time power balance of the system at each scheduling moment.

[0144] In actual implementation, the constraints are expressed as follows:

[0145] Constraint ①: Limitation on the number of photovoltaic panels installed.

[0146]

[0147] In the formula, This refers to the number of photovoltaic panels installed. This represents the maximum number of photovoltaic panels.

[0148] Constraint ②: Number of wind turbines installed

[0149]

[0150] In the formula, This refers to the number of fans to be installed. This represents the maximum number of fans that can be installed.

[0151] Constraint ③ Battery Constraint:

[0152]

[0153] In the formula, This refers to the number of batteries installed. This represents the maximum number of batteries that can be installed.

[0154]

[0155] In the formula, Power for charging the battery; This refers to the battery discharge power. This represents the maximum charging power of the battery. This refers to the maximum discharge power of the battery. Let t be the state of charge of the battery at time t; This represents the battery's maximum state of charge. This represents the minimum state of charge of the battery.

[0156] Constraint ④ Real-time power balance constraint

[0157]

[0158] In the formula, For electrical load; Power for charging electric vehicles; Power for charging the battery; The power output sold to the power grid; Photovoltaic power generation; The power generation capacity of the wind turbine; The discharge power of electric vehicle batteries; This refers to the battery discharge power. Power purchased from the power grid.

[0159] In this embodiment, constraints are set for the number of photovoltaic panels, wind turbines, and batteries installed in the DC microgrid system model, as well as constraints for the charging and discharging power and upper and lower limits of the state of charge of the batteries and the real-time power balance of the system. This effectively limits the safe boundary of system operation, avoids abnormal operation such as equipment overload and over-temperature, and ensures that the system is always within a safe and stable operating range during the optimization process. At the same time, the optimization results meet the feasibility requirements of actual engineering deployment. On the basis of ensuring a balance between economy and comfort, the reliability and durability of system operation are further improved.

[0160] In one possible implementation, the quantification of user comfort includes:

[0161] Determine the satisfaction calculation function for at least one type of flexible load; where flexible loads include air conditioning, electric vehicles, and movable loads;

[0162] Calculate the satisfaction values ​​of various flexible loads during the scheduling cycle;

[0163] Based on preset weighting coefficients, the satisfaction values ​​of various flexible loads are weighted and summed to obtain a quantitative index of user comfort.

[0164] Optionally, the loads that can be moved include rice cookers, electric kettles, induction cookers, washing machines, etc.

[0165] Objective function ①, system cost:

[0166]

[0167]

[0168] In the formula, Annual investment and construction costs for the equipment; Annual operating and maintenance costs of the equipment; The system's annual electricity purchase cost; This represents the system's annual electricity sales cost; Purchase cost per unit capacity of each piece of equipment; The rated power of each device is equal to the rated power of a single device multiplied by the number of devices. The equipment discount rate; Indicates the service life of the equipment.

[0169] Objective function ②, user satisfaction:

[0170]

[0171] In the formula, For overall user satisfaction, , , These are, respectively, satisfaction levels for electric vehicles, air conditioning, and movable loads. , , Satisfaction coefficients for electric vehicles, air conditioning, and movable loads. Requirements. .

[0172] Electric vehicle satisfaction:

[0173]

[0174] In the formula, The state of charge that users expect to achieve when the electric vehicle dispatch ends; This represents the state of charge of the electric vehicle at the end of the scheduling cycle.

[0175] Air conditioner satisfaction:

[0176]

[0177] In the formula, Set the temperature for the air conditioner; The optimal temperature; The human body's sensitivity to cold is taken as 0.66; The human body's sensitivity to cold is taken as 0.99.

[0178] Satisfaction with transferable loads:

[0179]

[0180] In the formula, The earliest tolerance time; This is the latest tolerance period; Preferred start time; Preferred end time; This refers to the actual start time of the load. The penalty coefficient for advance payment is set to 0.7; The delay penalty coefficient is set to 0.4.

[0181] In this embodiment, by determining the satisfaction calculation functions for various flexible loads such as air conditioners, electric vehicles, and movable loads, and calculating their satisfaction values ​​separately, the user comfort requirements that were originally vague are transformed into quantifiable and calculable specific indicators. This makes the comfort objectives in multi-objective optimization clearer and more in line with the actual user experience, ensuring that the optimization process can accurately balance economic costs and user needs. This makes the final system solution not only economical and reasonable, but also effectively improves the user's experience and acceptance of renewable energy systems.

[0182] Based on the above embodiments, this invention, by constructing a technical system of "data-driven zoning - load probability modeling - multi-objective collaborative optimization," realizes a paradigm shift in rural DC microgrid planning from generalized design to refined customization. Its technical advantages are mainly reflected in innovations at three levels: systematic design, planning framework, and optimization model.

[0183] This invention proposes a differentiated planning paradigm for large-scale heterogeneous regions, constructing a systematic methodological framework of "scene identification - standardized modeling - differentiated output," achieving a methodological leap from traditional "case-by-case customization" to "systematic and precise design." Through dynamic climate-resource zoning, heterogeneous regions across the country are categorized into a limited number of typical scenarios, and Pareto optimal solution sets are generated for each scenario within a unified optimization framework. This method ensures that the system configurations (such as the ratio of photovoltaic to wind turbine capacity, energy storage scale, etc.) and operating strategies of different regions reflect their resource endowments and load differences. By quantitatively balancing cost and satisfaction within the solution set, it provides planning schemes for each region that combine scientific consistency and decision-making flexibility, thus systematically resolving the long-standing contradiction between accuracy and generalizability in rural energy planning at the methodological level.

[0184] At the planning level, this invention proposes a dynamic zoning technique strongly coupled with the energy system model, effectively addressing the problem of insufficient representativeness of input data in regional energy planning. Unlike traditional static zoning schemes, this method constructs a multi-dimensional feature matrix based on key climate and resource parameters such as average monthly temperature, solar irradiance, and wind speed, and employs the K-medoids clustering algorithm for combined feature analysis. The significant advantages of this method are: the zoning feature parameters have a clear physical correlation with photovoltaic power output, wind turbine power generation efficiency, and building heating and cooling load requirements, ensuring that the zoning results accurately reflect system performance differences; simultaneously, the algorithm uses actual sample points as cluster centers, ensuring that the selected typical areas have a real engineering background and statistical representativeness. This dynamic zoning method improves the scientific rigor and diversity of model inputs from the outset, providing reliable boundary conditions and an engineering adaptability foundation for subsequent system optimization.

[0185] At the optimization level, this invention constructs a multi-objective collaborative optimization mechanism that integrates flexible load satisfaction evaluation. This optimization model quantifies the operational characteristics of key flexible loads such as air conditioners, electric vehicles, and water heaters into user satisfaction indicators, which, together with the system's annualized total cost, constitute a dual-objective optimization system. By employing NSGA-II for multi-dimensional search and evolution, a Pareto-optimal solution set representing the optimal trade-off between economy and comfort is obtained. This result not only provides a single optimal solution but also forms a visualized decision space, supporting planners in making quantitative trade-offs and flexible choices between cost and experience based on regional development policies, investment capacity, and energy demand, thereby significantly improving the engineering practicality and decision support value of planning results.

[0186] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0187] To verify the effectiveness of the differentiated scheduling of renewable energy DC systems provided in this application, four typical partitions (A, B, C, and D) were selected from the K-medoids clustering results. The costs and satisfaction levels corresponding to the equipment configuration and operation scheduling strategies determined based on this application's scheme and traditional schemes were statistically analyzed. The corresponding statistical results are shown in the table below:

[0188] Table 1. Cost and Satisfaction Statistics

[0189]

[0190] As shown in the table above, the annualized costs of scenarios A, B, and C all decreased significantly, with reductions of 12.35%, 32.60%, and 23.85%, respectively; while the cost of scenario D increased by 1.96% compared to before optimization. Meanwhile, user satisfaction improved in all scenarios, with increases of 12.04%, 13.75%, 11.74%, and 7.93%, respectively. Scenario B showed the highest cost reduction rate and satisfaction increase, mainly due to its relatively abundant renewable energy resources, high frequency of flexible equipment use, and significant adjustability potential, allowing optimized scheduling to fully realize the dual benefits of energy saving and improved comfort. Conversely, scenario D suffers from scarce renewable resources and weak load adjustability, limiting its optimization space and resulting in increased costs alongside improved satisfaction, exhibiting a clear trade-off between benefits and costs.

[0191] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0192] Figure 7 A schematic diagram of the structure of the differentiated dispatch device for a renewable energy DC system provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below:

[0193] like Figure 7 As shown, the differentiated dispatching device 4 for renewable energy DC systems includes:

[0194] The partitioning module 701 is used to acquire climate data for multiple regions in a set month, perform cluster analysis on the climate data, partition the multiple regions according to the cluster classification results, and obtain several typical partitions; among them, the climate data includes average temperature, wind speed and solar irradiance.

[0195] The load curve generation module 702 is used to acquire historical load data of multiple regions in a set month, and generate typical household load curves corresponding to each typical zone based on the historical load data.

[0196] The scheduling module 703 is used to construct a DC microgrid system model containing photovoltaic power generation units, wind power generation units, and energy storage units based on typical household load curves, and to perform multi-objective optimization on the DC microgrid system model based on NSGA-II to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. Among them, the multi-objective optimization aims to minimize system economy and maximize user comfort. System economy includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs, while user comfort is quantified based on flexible load satisfaction.

[0197] In one possible implementation, partitioning module 701 is specifically used to construct a multidimensional feature matrix based on climate data; and to perform cluster analysis on the feature matrix using the K-medoids clustering algorithm.

[0198] In one possible implementation, partitioning module 701 is specifically used for:

[0199] Each characteristic parameter in the climate data is normalized separately.

[0200] K data points are randomly selected from the normalized data as the initial cluster centers;

[0201] Calculate the Euclidean distance between the remaining data points and the centers of each initial cluster, and assign each data point to the corresponding cluster according to the principle of closest distance;

[0202] Recalculate the cluster center for each cluster and update the initial cluster centers;

[0203] The iterative process assigns each data point to the corresponding cluster based on the nearest principle and updates the cluster center until the cluster center no longer changes or the preset number of iterations is reached, in order to obtain the final clustering classification result.

[0204] In one possible implementation, the load curve generation module 702 is specifically used for:

[0205] For households within a target typical zone, determine the activation probability distribution of at least one electrical device during the heating season and the non-heating season; wherein, the target typical zone is one of the typical zones.

[0206] Based on the activation probability distribution, the Monte Carlo method is used to simulate and generate the start-stop event sequence of each electrical device in the simulation period, and the start-stop event sequence is converted into a continuous instantaneous power curve according to the rated power of each electrical device.

[0207] At each time step, the instantaneous power curves of each electrical device in the target typical partition are superimposed to generate multiple simulated household total load curves.

[0208] From multiple simulated total household load curves, the curve with the smallest root mean square error between it and the average load curve of all simulated curves is selected as the typical household load curve corresponding to the target typical area.

[0209] In one possible implementation, the scheduling module 703 is specifically used for:

[0210] Initialize the population and algorithm parameters; each individual in the population represents a set of candidate solutions including optimization parameters, time window parameters, and target parameters; optimization parameters include equipment configuration parameters and operation scheduling parameters; equipment configuration parameters include the number of photovoltaic panels, wind turbines, and batteries; operation scheduling parameters include flexible load mode, air conditioning temperature, and electric vehicle SOC level; time window parameters include equipment start-up and shutdown time and charging / discharging period; target parameters include investment cost, maintenance cost, electricity price, and user satisfaction.

[0211] Simulations are performed on each individual in the current population to calculate its annualized total cost under the system's economic objective and its comprehensive satisfaction value under the user comfort objective, and to determine whether the preset system operation constraints are met; whereby the annualized total cost includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs;

[0212] Perform fast non-dominated sorting on the current population, divide all individuals into multiple non-dominated levels based on the annualized total cost and overall satisfaction value, and calculate the crowding degree of individuals within the same non-dominated level;

[0213] Selection is performed based on the non-dominated hierarchy and crowding of individuals, and crossover and mutation operations are performed on the selected individuals to generate offspring populations;

[0214] The parent and offspring populations are merged, and non-dominated sorting and crowding calculation are performed on the merged population again. Based on the sorting and crowding, individuals with the same number as the parent population are selected to form a new generation population.

[0215] The simulation, sorting, genetic operations and population update steps are executed iteratively until the iteration termination condition is met. The final set of individuals in the first non-dominated level of the population is output as the Pareto optimal solution set.

[0216] For each typical partition, an equilibrium point is selected from its corresponding Pareto optimal solution set. The equipment configuration parameters and operation scheduling parameters corresponding to the equilibrium point are determined as the differentiated equipment configuration and operation scheduling strategy for that typical partition.

[0217] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the above method embodiments. Exemplarily, the electronic device can be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, etc., and is not limited thereto.

[0218] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0219] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A differentiated dispatch method for a renewable energy DC system, characterized in that, include: Climate data for multiple regions in a set month is acquired, cluster analysis is performed on the climate data, and the multiple regions are divided into several typical zones based on the cluster classification results; wherein, the climate data includes average temperature, wind speed and solar irradiance. Historical load data for multiple regions in a set month are obtained, and typical household load curves corresponding to each typical zone are generated based on the historical load data. Based on the typical household load curves, a DC microgrid system model including photovoltaic power generation units, wind power generation units, and energy storage units is constructed. The model is then subjected to multi-objective optimization using the non-dominated sorting genetic algorithm NSGA-II to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. The multi-objective optimization aims to minimize system economics and maximize user comfort. System economics includes annualized equipment investment, operation and maintenance costs, and grid purchase and sale costs. User comfort is quantified based on flexible load satisfaction. The step of generating typical household load curves for each typical zone based on the historical load data includes: For households within a target typical zone, determine the activation probability distribution of at least one electrical device during the heating season and the non-heating season; wherein, the target typical zone is one of the typical zones. Based on the activation probability distribution, the Monte Carlo method is used to simulate and generate the start-stop event sequence of each electrical device within the simulation period, and the start-stop event sequence is converted into a continuous instantaneous power curve according to the rated power of each electrical device. At each time step, the instantaneous power curves of each electrical device in the target typical partition are superimposed to generate multiple simulated household total load curves. From the multiple simulated total household load curves, the curve with the smallest root mean square error between it and the average load curve of all simulated curves is selected as the typical household load curve corresponding to the target typical zone. The quantification of user comfort includes: Determine a satisfaction calculation function for at least one type of flexible load; wherein, the flexible load includes air conditioning, electric vehicles, and movable loads; Calculate the satisfaction values ​​of various flexible loads during the scheduling cycle; Based on preset weighting coefficients, the satisfaction values ​​of various flexible loads are weighted and summed to obtain a quantitative index of user comfort.

2. The differentiated dispatch method for renewable energy DC systems according to claim 1, characterized in that, The cluster analysis of the climate data includes: A multidimensional feature matrix is ​​constructed based on the climate data; The K-medoids clustering algorithm is used to perform cluster analysis on the feature matrix.

3. The differentiated dispatch method for renewable energy DC systems according to claim 2, characterized in that, The K-medoids clustering algorithm is used to perform clustering analysis on the feature matrix, including: Each characteristic parameter in the climate data is normalized. K data points are randomly selected from the normalized data as the initial cluster centers; Calculate the Euclidean distance between the remaining data points and the centers of each initial cluster, and assign each data point to the corresponding cluster according to the principle of closest distance; Recalculate the cluster center for each cluster and update the initial cluster centers; The iterative process assigns each data point to the corresponding cluster based on the nearest principle and updates the cluster center until the cluster center no longer changes or the preset number of iterations is reached, in order to obtain the final clustering classification result.

4. The differentiated dispatch method for renewable energy DC systems according to claim 1, characterized in that, The designated months include January and July.

5. The differentiated dispatch method for renewable energy DC systems according to claim 1, characterized in that, The multi-objective optimization of the DC microgrid system model based on NSGA-II yields differentiated equipment configuration and operation scheduling strategies applicable to each typical region, including: Initialize the population and algorithm parameters; whereby each individual in the population represents a set of candidate solutions including optimization parameters, time window parameters, and target parameters; the optimization parameters include equipment configuration parameters and operation scheduling parameters; the equipment configuration parameters include the number of photovoltaic panels, the number of wind turbines, and the number of batteries; the operation scheduling parameters include flexible load mode, air conditioning temperature, and electric vehicle SOC level; the time window parameters include equipment start-up and shutdown time and charging and discharging period; the target parameters include investment cost, maintenance cost, electricity price, and user satisfaction; Simulations are performed on each individual in the current population to calculate its annualized total cost under the system's economic objective and its comprehensive satisfaction value under the user comfort objective, and to determine whether the preset system operation constraints are met; wherein, the annualized total cost includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs; Perform fast non-dominated sorting on the current population, divide all individuals into multiple non-dominated levels based on the annualized total cost value and the comprehensive satisfaction value, and calculate the crowding degree of individuals within the same non-dominated level; Selection is performed based on the non-dominated hierarchy and crowding of individuals, and crossover and mutation operations are performed on the selected individuals to generate offspring populations; The parent population is merged with the offspring population. Non-dominated sorting and crowding calculation are performed on the merged population again. Individuals with the same number as the parent population size are selected from the sorting and crowding based on the sorting and crowding to form a new generation population. The simulation, sorting, genetic operations and population update steps are executed iteratively until the iteration termination condition is met. The final set of individuals in the first non-dominated level of the population is output as the Pareto optimal solution set. For each typical partition, an equilibrium point is selected from the corresponding Pareto optimal solution set, and the equipment configuration parameters and operation scheduling parameters corresponding to the equilibrium point are determined as the differentiated equipment configuration and operation scheduling strategy for that typical partition.

6. The differentiated dispatch method for renewable energy DC systems according to claim 1, characterized in that, Before performing multi-objective optimization of the DC microgrid system model based on NSGA-II, the method also includes setting constraints for the DC microgrid system model. The constraints include: constraints on the number of photovoltaic panels installed, constraints on the number of wind turbines installed, constraints on the number of batteries installed, constraints on the charging and discharging power of batteries and the upper and lower limits of their state of charge, and constraints on the real-time power balance of the system at each scheduling moment.

7. A renewable energy DC system differentiated dispatching apparatus for executing the differentiated dispatching method for renewable energy DC systems according to any one of claims 1 to 6, characterized in that, include: The partitioning module is used to acquire climate data for multiple regions in a set month, perform cluster analysis on the climate data, and partition the multiple regions according to the clustering classification results to obtain several typical partitions; wherein, the climate data includes average temperature, wind speed and solar irradiance. The load curve generation module is used to acquire historical load data of multiple regions in a set month, and generate typical household load curves corresponding to each typical zone based on the historical load data. The scheduling module is used to construct a DC microgrid system model including photovoltaic power generation units, wind power generation units, and energy storage units based on the typical household load curve, and to perform multi-objective optimization on the DC microgrid system model based on NSGA-II to obtain differentiated equipment configuration and operation scheduling strategies applicable to each typical zone. Among them, the multi-objective optimization aims to minimize system economy and maximize user comfort. The system economy includes annualized equipment investment, operation and maintenance, and grid purchase and sale costs, and the user comfort is quantified based on flexible load satisfaction.

8. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 6.

Images