A photovoltaic and energy storage optimal configuration method and device based on a snowflake network

Abstract

The present disclosure relates to the technical field of power grid optimization, and provides a photovoltaic and energy storage optimization configuration method and device based on a snowflake network. The method comprises determining typical scenarios of photovoltaic and load correlation and occurrence probability of each typical scenario based on historical load demand data and historical photovoltaic output data; determining the access position and capacity of photovoltaic and energy storage by using a constructed double-layer optimization configuration model of photovoltaic and energy storage comprising a planning layer and a running layer; calculating the constraint condition and objective function of the planning layer based on the optimization result of the running layer, and optimizing the access position and capacity of photovoltaic and energy storage; and based on the load curve corresponding to the typical scenario with the occurrence probability greater than a set value and the photovoltaic output curve and the optimized result of the planning layer, optimizing the output condition of photovoltaic and energy storage in each period according to the objective function and constraint condition of the running layer, and feeding back the optimization result to the planning layer. The adaptability and flexibility of the snowflake network can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of power grid optimization technology, and in particular to a method and apparatus for optimizing the configuration of photovoltaic and energy storage based on snowflake grids. Background Technology

[0002] To adapt to the global energy transition trend, the power system is actively promoting new energy power generation technologies, especially the widespread application of renewable energy represented by photovoltaics.

[0003] Photovoltaics (PV) offers advantages such as cleanliness, environmental friendliness, and low maintenance costs, making PV grid integration a significant trend in power distribution network development. However, as the proportion of PV power in the grid gradually increases, its output becomes highly volatile and uncertain due to weather conditions and other factors, posing new challenges to the operation and stability of the power system. To mitigate the adverse effects of PV grid integration, the integration of energy storage systems becomes crucial. Energy storage systems can store excess PV energy during periods of low demand and release it during peak demand, thereby achieving energy balance regulation and ensuring the stable operation of the power grid.

[0004] Snowflake network is a new type of active distribution network. It takes a 10kV ring network box as the core node and consists of 8 10kV lines from 4 substations or 6 10kV lines from 3 substations to form an independent feeder cluster. The different feeders inside are interconnected, which can give full play to its strong load transfer capability.

[0005] With the large-scale integration of photovoltaic (PV) power and the gradual opening of the electricity market, the Snowflake Grid faces the problem of strong fluctuations in power output and demand. These fluctuations not only affect the stable operation of the grid but also increase the difficulty of grid planning and management. Therefore, a photovoltaic and energy storage optimization scheme based on the Snowflake Grid is needed to improve its adaptability and flexibility. Summary of the Invention

[0006] To address the aforementioned issues, this disclosure provides a method and apparatus for optimizing the configuration of photovoltaic and energy storage based on snowflake grids, which can improve the adaptability and flexibility of snowflake grids.

[0007] Firstly, a method for optimizing the configuration of photovoltaic and energy storage based on snowflake grids is provided, the method comprising: Obtain historical load demand data and photovoltaic power output data from Xuehua.com; Based on historical load demand data and historical photovoltaic output data, typical scenarios of the correlation between photovoltaic power and load and the probability of occurrence of each typical scenario are determined. The connection location and capacity of photovoltaic and energy storage are determined by using the constructed two-layer optimization configuration model of photovoltaic and energy storage. The model includes a planning layer and an operation layer. The planning layer uses system economic indicators as the objective function, and the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and investment costs for photovoltaics and energy storage. The planning layer is used to calculate the constraints and objective function of the planning layer based on the optimization results of the operation layer, and to optimize the access location and capacity of photovoltaic and energy storage. The operation layer is used to optimize the output of photovoltaic and energy storage in each time period based on the load curve corresponding to typical scenarios with a probability greater than a set value, the photovoltaic output curve, the optimized access location and capacity of photovoltaic and energy storage, and the objective function and constraints of the operation layer. The optimization results are then fed back to the planning layer.

[0008] Furthermore, using the constructed two-layer optimization configuration model for photovoltaics and energy storage, the access locations and capacities of photovoltaics and energy storage are determined, including: The capacity and access location of photovoltaics, and the capacity and access location of energy storage are used as particles in the improved particle swarm optimization algorithm. The particle positions and velocities are initialized and input into the execution layer. The positions and velocities of the particle swarm are updated using the optimized access locations and capacities of photovoltaics and energy storage from the planning layer. The updated positions and velocities of the particle swarm are then input into the execution layer until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or until the set number of iterations is reached.

[0009] Furthermore, the system economic indicator F2 is:

[0010]

[0011] in, For the investment costs of photovoltaics and energy storage, For operating costs, Indicates the discount rate; This refers to the lifespan of distributed photovoltaic and energy storage equipment; This represents the investment cost per unit capacity of distributed photovoltaic power generation. This represents the investment cost per unit capacity of energy storage. This represents the total installed capacity of distributed photovoltaic power. This indicates the total installed capacity of energy storage.

[0012] Furthermore, the operating cost F1 and the voltage offset index VDI are:

[0013]

[0014]

[0015]

[0016]

[0017]

[0018] in, Indicates operating costs; Indicates network loss cost; This indicates the cost of purchasing electricity from the main grid; T represents the cost of curtailment; N represents the total number of time slots; and T represents the number of photovoltaic power plants. Indicates the price of photovoltaic power generation; Let represent the power generation of the i-th photovoltaic power station at time t; Indicates the price of network loss; This represents the network loss at time t; This indicates the main grid electricity purchase price; This represents the main network purchase volume at time t; Indicates the penalty price for solar power curtailment; U represents the amount of abandoned electricity from the i-th photovoltaic power station at time t; i,t U represents the actual voltage value of node i at time t obtained from probabilistic power flow; i,N Let be the rated voltage of node i at time t.

[0019] Furthermore, the constraints at the operational layer include: system power flow constraints, system operation constraints, energy storage system charging and discharging power constraints, photovoltaic power output constraints, upper and lower limits of battery charging state constraints for energy storage, and energy storage location constraints. Among these, the energy storage location constraint refers to the location variable of the energy storage belonging to the energy storage candidate location. The energy storage candidate location is obtained by selecting the energy storage location based on the tie-line distribution information of the snowflake network.

[0020] Furthermore, the constraints at the planning level include: system power flow constraints, system operation constraints, upper and lower limits on the number of photovoltaic power sources connected, and upper and lower limits on the number of energy storage access groups.

[0021] Furthermore, based on historical load demand data and historical photovoltaic output data, typical scenarios of photovoltaic-load correlation and the probability of occurrence of each typical scenario are determined, including: Based on historical load demand data and historical photovoltaic output data, the probability density functions of load demand and photovoltaic output are determined. By combining Latin hypercube sampling with Nataf transform, correlation sampling is performed on the probability density function of load demand and the probability density function of photovoltaic output to obtain several time-series scenarios of load and photovoltaic. Clustering algorithms were used to reduce the number of time-series load and photovoltaic scenarios to obtain a set of typical scenarios. By using the number of time-series scenarios participating in the typical scenario clustering and the total number of time-series scenarios, the probability of occurrence of each typical scenario is calculated.

[0022] Secondly, a photovoltaic and energy storage optimized configuration device based on a snowflake grid is provided, the device comprising: The acquisition unit is used to acquire historical load demand data and photovoltaic output data of the Snowflakes grid. The probability determination unit is used to determine typical scenarios of the correlation between photovoltaics and load, and the probability of occurrence of each typical scenario, based on historical load demand data and historical photovoltaic output data. An optimization configuration unit is used to determine the access locations and capacities of photovoltaic (PV) and energy storage using a constructed two-layer optimization configuration model. The model includes a planning layer and an operation layer. The planning layer uses system economic indicators as objective functions, while the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and PV and energy storage investment costs. The planning layer calculates the constraints and objective functions based on the optimization results of the operation layer and optimizes the access locations and capacities of PV and energy storage. The operation layer optimizes the output of PV and energy storage at various time periods based on the load curves corresponding to typical scenarios with a probability greater than a set value, the PV output curves, and the optimized access locations and capacities of PV and energy storage, using the objective functions and constraints of the operation layer, and feeds the optimization results back to the planning layer.

[0023] Furthermore, the optimization configuration unit is specifically used to use the photovoltaic capacity and access location, and the energy storage capacity and access location as particles in the improved particle swarm algorithm, initialize the particle position and velocity, and input them into the running layer; use the photovoltaic and energy storage access locations and capacities optimized by the planning layer to update the position and velocity of the particle swarm, and input the updated position and velocity of the particle swarm into the running layer until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or reaches the set number of iterations.

[0024] Furthermore, the system economic indicator F2 is:

[0025]

[0026] in, For the investment costs of photovoltaics and energy storage, For operating costs, Indicates the discount rate; This refers to the lifespan of distributed photovoltaic and energy storage equipment; This represents the investment cost per unit capacity of distributed photovoltaic power generation. This represents the investment cost per unit capacity of energy storage. This represents the total installed capacity of distributed photovoltaic power. This indicates the total installed capacity of energy storage.

[0027] Furthermore, the operating cost F1 and the voltage offset index VDI are:

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] in, Indicates operating costs; Indicates network loss cost; This indicates the cost of purchasing electricity from the main grid; T represents the cost of curtailment; N represents the total number of time slots; and T represents the number of photovoltaic power plants. Indicates the price of photovoltaic power generation; Let represent the power generation of the i-th photovoltaic power station at time t; Indicates the price of network loss; This represents the network loss at time t; This indicates the main grid electricity purchase price; This represents the main network purchase volume at time t; Indicates the penalty price for solar power curtailment; U represents the amount of abandoned electricity from the i-th photovoltaic power station at time t; i,t U represents the actual voltage value of node i at time t obtained from probabilistic power flow; i,N Let be the rated voltage of node i at time t.

[0034] Furthermore, the constraints at the operational layer include: system power flow constraints, system operation constraints, energy storage system charging and discharging power constraints, photovoltaic power output constraints, upper and lower limits of battery charging state constraints for energy storage, and energy storage location constraints. Among these, the energy storage location constraint refers to the location variable of the energy storage belonging to the energy storage candidate location. The energy storage candidate location is obtained by selecting the energy storage location based on the tie-line distribution information of the snowflake network.

[0035] Furthermore, the constraints at the planning level include: system power flow constraints, system operation constraints, upper and lower limits on the number of photovoltaic power sources connected, and upper and lower limits on the number of energy storage access groups.

[0036] Furthermore, the probability determination unit is specifically used to determine the probability density functions of load demand and photovoltaic output based on historical load demand data and historical photovoltaic output data. It combines Latin hypercube sampling with Nataf transform to perform correlation sampling on the probability density functions of load demand and photovoltaic output, obtaining several time-series scenarios of load and photovoltaic power. A clustering algorithm is then used to reduce these time-series scenarios to a set of typical scenarios. Finally, the probability of occurrence of each typical scenario is calculated using the number of time-series scenarios participating in the typical scenario clustering and the total number of time-series scenarios.

[0037] Thirdly, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When a processor executes a program stored in memory, it implements the steps of the above method.

[0038] Fourthly, a computer storage medium is provided, wherein a computer program is stored therein, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0039] Compared with the prior art, this disclosure has the following advantages: The solution of this disclosure embodiment, based on typical scenario data obtained by probability statistics, can better integrate various distributed power sources by rationally planning and optimizing the configuration of photovoltaic systems and energy storage systems, thereby realizing the large-scale consumption of photovoltaic energy, improving the adaptability and flexibility of the Snowflake Network, optimizing the utilization of Snowflake Network resources, and enhancing the power stability and power quality for users.

[0040] Other features and advantages of this disclosure will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the disclosure. The objects and other advantages of this disclosure may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1A schematic flowchart of a photovoltaic and energy storage optimization configuration method based on a snowflake grid according to an embodiment of the present disclosure is shown. Figure 2 A schematic flowchart of another photovoltaic and energy storage optimization configuration method based on snowflake grid according to an embodiment of the present disclosure is shown; Figure 3 A network topology diagram according to an embodiment of the present disclosure is shown; Figure 4 Data on photovoltaic output in various scenarios and time periods according to embodiments of this disclosure are shown; Figure 5 Data on load demand in various scenarios and time periods according to embodiments of this disclosure are shown; Figure 6 The photovoltaic output data for a typical operating scenario according to embodiments of this disclosure are shown at various time periods. Figure 7 The load demand data for a typical operating scenario according to embodiments of the present disclosure are shown for each time period; Figure 8 The probability of occurrence of 20 typical operating scenarios according to embodiments of this disclosure is shown; Figure 9 A diagram showing the intraday distribution network optimization results obtained according to an embodiment of this disclosure is provided. Figure 10 A diagram illustrating the energy storage operation status according to an embodiment of this disclosure is shown; Figure 11 A schematic diagram of an optimized configuration of photovoltaic power source and energy storage according to an embodiment of the present disclosure is shown. Figure 12 A schematic diagram of fitness curves according to an embodiment of the present disclosure is shown. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0044] Figure 1 and Figure 2 A schematic flowchart of a photovoltaic and energy storage optimization configuration method based on a snowflake grid according to an embodiment of the present disclosure is shown, as follows: Figure 1 As shown in the figure, the photovoltaic and energy storage optimization configuration method based on snowflake grid in this disclosure includes the following steps: S1. Input the topology data of the snowflake network, historical load demand data, and historical photovoltaic output data.

[0045] The topology data of the Snowflake network includes: the line connection status of the Snowflake network, and the line connection information includes the distribution information of tie lines.

[0046] S2. Determine the probability density function of load demand based on historical load demand data, and determine the probability density function of photovoltaic output based on historical photovoltaic output data.

[0047] The probability density function of load demand can be described by the probability density of active power and reactive power: (1) (2) In equations (1) and (2), and These represent the average values ​​of active and reactive power, respectively. and These represent the variances of active and reactive power, respectively.

[0048] The probability density function of photovoltaic power output can be described by the probability density function of the active power output by the photovoltaic system: (3) (4) In formulas (3) and (4), Γ(·) represents the Gamma function; the shape parameters α and β of the Beta distribution are obtained by maximum likelihood fitting based on historical data of light intensity; in formula (3), P is the photovoltaic active power output. η is the maximum output of photovoltaic active power; A is the total area of ​​photovoltaic cells; η is the total conversion efficiency of the photovoltaic power station; r max The maximum light intensity during the study period.

[0049] S3: By combining Latin hypercube sampling with Nataf transform, correlation sampling is performed on the probability density function of load demand and the probability density function of photovoltaic output to obtain several time-series scenarios of load and photovoltaic.

[0050] The basic operation process of Latin hypercube sampling consists of two steps: sampling and sorting.

[0051] sampling: 1) Given S random variables , , ..., , for , , ..., For any random variable in the set, its probability distribution function is: (5) Assuming the sampling size is N, each random variable The probability space is divided into N equally spaced, non-overlapping intervals, which yields... , ,…, ; 2) Randomly select a value from each sub-interval as... Sampled values; 3) Calculate X using the inverse function shown in formula (6). k The nth sample value, where rand represents a random number between 0 and 1; (6) The above steps yield an initial sampling matrix B of order S×N that can fully reflect the overall distribution of the random variables.

[0052] arrangement: The sample values ​​of random variables in matrix B obtained through sampling exhibit a monotonically increasing trend. Therefore, there is a strong correlation between the random variables in matrix B, which may not match the actual correlation between random variables. To adjust this correlation, the order of the sampled random variable values ​​needs to be changed to minimize the correlation between the sampled values ​​of independent random variables. In the arrangement process, a K×N order matrix I is first created, where each row represents the position of the element in the corresponding row of the initial sampling matrix B. Following the instructions of order matrix I, the elements of matrix B are rearranged to form the final sampling matrix.

[0053] For situations where photovoltaic power output and load demand do not follow a normal distribution, the inverse Nataf transform is used to transform the sample data that does not follow a normal distribution into a space that conforms to a normal distribution, making it convenient to use the Latin hypercube sampling (LHS) method to generate multi-scenario data that take into account correlations.

[0054] Let the input random variable be ,remember The probability density function of (i=1,…,n) and cumulative distribution function Its correlation coefficient matrix D X Represented as: (7) In equation (7) Represents random variables and The correlation coefficient between them.

[0055] Transformation by equal probability: (8) Applying the Nataf transform formula: (9) Obtain the correlation coefficient matrix D of random variable Y Y The off-diagonal elements, in equation (9) , Representing random variables respectively , The mean; Indicates the correlation coefficient Standard normal random , The joint probability density function.

[0056] Through the above steps, the random variable is transformed from the input variable space to the standard normal space, and then the transformation to the independent standard normal space is completed.

[0057] D Y Perform Choleskey decomposition to obtain D Y =D Y =EE T , where E is a lower triangular matrix. Then we can obtain a random vector: (10) In equation (10), Z is a random vector in the independent standard normal space.

[0058] S4. Clustering algorithm is used to reduce the number of time-series load and photovoltaic scenarios to obtain a set of typical scenarios, and the probability of occurrence of each typical scenario is calculated.

[0059] The probability of occurrence of a typical scenario is obtained by comparing the number of time-series scenarios participating in the typical scenario cluster with the total number of time-series scenarios.

[0060] Specifically, the photovoltaic and load samples of each time period can be regarded as a scenario, that is, the time series scenario of load and photovoltaic. The improved k-means algorithm - k-means++ is used to reduce the scenario, thereby deterministically describing the randomness and volatility of source load.

[0061] The k-means++ clustering method is an iterative process aimed at minimizing the sum of squared errors across the entire dataset. The algorithm includes the following steps 1)-4): 1) Specify the number of clusters K, and select K data objects in the dataset as initial cluster centers. Each selected data object will serve as the center of K clusters. 2) For the remaining data objects in the dataset, the data objects are assigned to the clusters represented by the cluster centers with the greatest similarity, based on the criterion that the closer they are, the greater their similarity. 3) Calculate the mean of all data objects in each cluster and use the calculated mean as the new cluster center of the cluster. Then calculate the sum of squared distances from all data objects to the cluster center of their respective clusters, which is the sum of squared errors (SEE) of the entire dataset. 4) Determine whether the sum of squared errors and cluster centers of the entire dataset have changed. If they have changed, return to step 2) and start the next iteration. If they have not changed, the clustering ends and the clustering results are output.

[0062] The formula for calculating the Euclidean distance between data points is as follows: (11) Where x represents a data object; b i The i-th cluster center is represented by m; the dimension of the data is represented by x. j Let b represent the j-th attribute value of data point x. ij b i The j-th attribute value.

[0063] The formula for calculating the sum of squared errors (SEE) for the entire dataset is as follows: (12) S4. Using the constructed photovoltaic and energy storage dual-layer optimization configuration model, determine the access location and capacity of photovoltaic and energy storage.

[0064] Among them, the two-layer optimization configuration model is as follows: Figure 2 As shown, it includes a planning layer and an operation layer. The planning layer uses system economic indicators as the objective function, and the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and photovoltaic and energy storage investment costs. The planning layer is used to calculate the constraints and objective function of the planning layer based on the optimization results of the operation layer, and to optimize the access location and capacity of photovoltaic and energy storage. The operation layer is used to optimize the output of photovoltaic and energy storage in each time period based on the load curve corresponding to typical scenarios with a probability greater than a set value, the photovoltaic output curve, the optimized access location and capacity of photovoltaic and energy storage, and the objective function and constraints of the operation layer. The optimization results are then fed back to the planning layer.

[0065] This step S4 specifically includes: S41. Set the photovoltaic capacity and access location, and the energy storage capacity and access location to the particles of the improved particle swarm algorithm, initialize the particle position and velocity, and input them into the running layer. The particle positions and velocities in the improved particle swarm optimization algorithm are initialized, and the access locations and capacities of photovoltaic and energy storage in the planning layer are set and used as inputs for the operation layer. S42. Based on the load curve, photovoltaic output curve, and position and velocity of the particle population corresponding to typical scenarios with a probability greater than the set value, the operation layer optimizes the output of photovoltaic and energy storage in each time period according to the objective function and constraints of the operation layer, and feeds the optimization results back to the planning layer. S43. The planning layer calculates the constraints and objective function of the planning layer based on the optimization results of the operation layer, and optimizes the access location and capacity of photovoltaic and energy storage. S44. Update the position and velocity of the particle population using the optimized access locations and capacities of photovoltaic and energy storage in the planning layer. Input the updated position and velocity of the particle population into the operation layer and repeat steps S42 to S43 until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or until the set number of iterations is reached.

[0066] The objective function of the runtime model includes the runtime cost F1 and the voltage offset index VDI as follows: (13) (14) (15) (16) (17) in, Indicates operating costs; Indicates network loss cost; This indicates the cost of purchasing electricity from the main grid; T represents the cost of curtailment; N represents the total number of time slots; and T represents the number of photovoltaic power plants. Indicates the price of photovoltaic power generation; Let represent the power generation of the i-th photovoltaic power station at time t; Indicates the price of network loss; This represents the network loss at time t; This indicates the main grid electricity purchase price; This represents the main network purchase volume at time t; Indicates the penalty price for solar power curtailment; This represents the amount of abandoned electricity generated by the i-th photovoltaic power station at time t.

[0067] The voltage deviation index measures the voltage stability of a distribution network at a specific moment. It represents the degree to which the node voltage deviates from the rated voltage. The higher the index, the worse the voltage stability. Its calculation formula is as follows: (18) In equation (18), VDI represents the voltage offset index, U i,t U represents the actual voltage value of node i at time t obtained from probabilistic power flow; i,N Let be the rated voltage of node i at time t.

[0068] The constraints at the operational layer include static safety constraints and other limitations. Static safety constraints specifically cover system power flow constraints and system operation constraints. Other constraints include energy storage system charging and discharging power constraints, photovoltaic power output constraints, upper and lower limits of battery charging state constraints, and energy storage location constraints.

[0069] The expression for the system power flow constraint is: (19) (20) (twenty one) (twenty two) (twenty three) (twenty four) In formulas (19)~(24), Let P represent the set of all branches in the snowflake network, where k represents node k. ik Q represents the active power on branch ik. ik This represents the reactive power on branch ik. and Let be the resistance and reactance of branch ij; Let be the current in branch ij; Let be the active power flowing from node i to node j on the branch. Let be the reactive power flowing from node i to node j on the branch; Let be the reactive power flowing from node j to node i on the branch; Let be the voltage at node i. Let J be the voltage at node j. Let be the sum of the active power injected into node i. , , These represent the active power injected into the energy storage, the photovoltaic power, and the power consumed by the load at node i, respectively. Let i be the sum of reactive power injected into node i. , , These represent the reactive power injected into the energy storage, photovoltaic, and consumed by the load at node i, respectively.

[0070] The expression for the system's operational constraints is: (25) (26) In formulas (25)~(26), and These are the minimum and maximum permissible node voltage values ​​of the system, respectively. This represents the maximum permissible branch current value of the system.

[0071] The charging and discharging power constraints of the energy storage system are: (27) (28) Formula (27)~28) is in the middle , and These represent the upper and lower limits of the charging / discharging power and energy storage transmission power of the i-th energy storage unit during time period t, respectively.

[0072] The output constraint of photovoltaic power sources is: (29) In equation (29) Upper limit of photovoltaic power output The upper and lower limits of the battery state of charge for energy storage are constrained as follows: (30) In equation (30), SOC represents the battery charging state; The battery charging state of the i-th energy storage unit during the t-th time period; and These represent the upper and lower limits of the battery's state of charge for energy storage.

[0073] The location constraints for energy storage are: (31) In equation (31), For the location variable of energy storage, This is the set of potential energy storage locations determined based on the distribution of interconnecting lines in the snowflake network.

[0074] Existing optimization methods are mostly applicable only to microgrids or single feeders, with relatively simple modeling and no consideration of the interconnections between different feeders, making them unsuitable for snowflake grids to maximize their capabilities. The alternative energy storage locations determined in this invention based on the interconnection distribution of the snowflake grid are used as constraints in the optimization process, making them suitable for snowflake grids to fully utilize their capabilities.

[0075] The planning function of the planning layer model is the system economic index F2, which includes the investment cost C2 for photovoltaic and energy storage and the operating cost F1: (32) (33) In equations (32) and (33), Indicates the discount rate; This refers to the lifespan of distributed photovoltaic and energy storage equipment; This represents the investment cost per unit capacity of distributed photovoltaic power generation. This represents the investment cost per unit capacity of energy storage. This represents the total installed capacity of distributed photovoltaic power. This indicates the total installed capacity of energy storage.

[0076] It should be noted that F2 here is the fitness function of the improved particle swarm optimization algorithm.

[0077] The constraints at the planning level include static safety constraints and other limitations. Static safety constraints are consistent with those at the operation level. Other limitations include upper and lower limits on the number of photovoltaic power sources and energy storage units that can be connected.

[0078] The upper and lower limits for the number of distributed photovoltaic (PV) installations are as follows: (34) In equation (34), NPV represents the number of photovoltaic access groups; and These represent the upper and lower limits of the number of photovoltaic (PV) modules that can be connected to the grid.

[0079] The upper and lower limits of the number of energy storage access groups are constrained as follows: (35) In equation (35), NESS represents the number of photovoltaic access groups; and These represent the upper and lower limits of the number of photovoltaic (PV) modules that can be connected to the grid.

[0080] The embodiments of the present invention will be illustrated below with a specific example.

[0081] This paper uses a snowflake net for example verification. The network topology diagram is as follows: Figure 3 As shown.

[0082] The candidate nodes for photovoltaic (PV) power supply are {2, 3, 4, 8, 12, 20, 27, 32}; the candidate nodes for energy storage (ESS) power supply are {3, 6, 9, 18, 20, 28, 30}. The rated capacity of a single PV power supply is 100kW, with a maximum of 20 PV units per node and a minimum of 5 PV units per node. The rated capacity of each energy storage unit is 50kW, with a maximum of 20 energy storage units per node and a minimum of 6 energy storage units per node. PV power output is from 06:00 to 17:00 daily. The economic lifespan of both PV and energy storage is set at 10 years. The energy storage charge / discharge efficiency is 0.9, the maximum energy storage SOC is 0.9, the minimum energy storage SOC is 0.1, and the initial energy storage SOC is 0.5. Other parameters are shown in Table 1 below.

[0083] The State of Charge (SOC) is the percentage of a battery's rated capacity remaining. It reflects the battery's remaining capacity and indicates its ability to continue operating. The SOC is 0 when the battery is fully discharged and 1 when it is fully charged; it is typically expressed as 0–100%.

[0084] Table 1 Planning-related parameters

[0085] The correlation coefficient between illuminance and load can be calculated from historical data of illuminance and load at the same access location, with a correlation coefficient of 0.3 set between load and illuminance. Using Latin hypercube sampling and Nataf inverse transform, the sampled values ​​of photovoltaic output and load at corresponding times are used to construct a scenario, generating a total of 500 time-series scenarios. The data for photovoltaic output and load demand in each scenario and time period are as follows: Figure 4 and Figure 5 As shown.

[0086] A 24×2×500-order photovoltaic output-load correlation sample matrix was constructed through sampling, where the data at each time point corresponds to a specific scenario. To reduce the computational burden, this paper uses a clustering algorithm to perform cluster analysis on 500 scenario samples based on similarity. The number of clusters was set to 20, and the K-means clustering method was applied to process the 24×2×500-dimensional photovoltaic-load data. The photovoltaic power output (single 100kW) and load demand data for the 20 typical operating scenarios obtained by clustering at each time point are as follows: Figure 6 and Figure 7 As shown.

[0087] Depend on Figure 6 and Figure 7As can be seen, the data in each of the clustered typical operating scenarios exhibit significant differences, effectively representing the original data. By analyzing the number of original vectors contained in the cluster centers, the probability of each typical scenario can be calculated, thereby enabling probability simulation and feature representation of diverse operating scenarios. Figure 8 The figure shows the probability of occurrence for 20 typical operating scenarios.

[0088] Based on the proposed two-level probabilistic programming model, the optimized configuration is performed, and the intraday distribution network optimization results and energy storage operation status are as follows: Figure 9 and Figure 10 As shown, the optimized configuration results are as follows: Figure 11 As shown.

[0089] Figure 10 The left vertical axis of the bar chart corresponds to the charging and discharging power of energy storage at different times of the day; the right vertical axis of the line chart corresponds to the state of charge of energy storage at different times of the day.

[0090] The final optimized capacity planning result is a photovoltaic capacity of 1.1MW (11 units) and an energy storage capacity of 0.4MW (8 units); the optimal grid connection location scheme for the photovoltaic power supply and energy storage is as follows: Figure 11 As shown.

[0091] As the number of iterations increases, the Particle Swarm Optimization (PSO) algorithm typically converges to or near the optimal solution. This can be observed by watching the fitness value gradually improve with increasing iterations, eventually reaching a stable level. Typically, a fitness curve is plotted during algorithm execution, showing how the current optimal fitness value of the swarm changes with the number of iterations. Analyzing this curve can determine whether the algorithm has converged or whether parameters (such as the number of iterations, inertia weight, acceleration coefficient, etc.) need to be adjusted to obtain better results. If the fitness curve shows that the algorithm converges before reaching the maximum number of iterations, the number of iterations can be reduced to save computational resources. Conversely, if the fitness improvement is slow, it may be necessary to increase the number of iterations or adjust other parameters to promote a better search. The fitness curve for this example is shown below. Figure 12 As shown.

[0092] Depend on Figure 12 It can be seen that when the number of iterations is 15, the algorithm converges close to the optimal solution. In the example in this paper, the number of iterations is set to 50, and the optimization results are good. The final solution satisfies the objective well.

[0093] In summary, this invention addresses the stability issues arising from large-scale photovoltaic (PV) power generation connected to the Snowflakes grid by proposing a Snowflakes-based method for optimizing the configuration of PV and energy storage. First, based on historical load demand and PV output data, scenarios considering the correlation between PV and load are generated using Latin hypercube sampling and Nataf inverse transform. These scenarios more realistically reflect the actual system's operating state. Subsequently, K-means++ clustering is used to reduce the generated scenarios, effectively decreasing computational complexity and time costs. To optimize the configuration of the PV and energy storage system, this invention constructs a two-layer optimization model. The planning layer focuses on system economic indicators, emphasizing the system's economic efficiency and return on investment; the operation layer focuses on system operating costs and voltage deviation index, aiming to ensure efficient system operation and stability, and reduce potential voltage deviation problems during operation. The entire optimization process is solved using the Particle Swarm Optimization (PSO) algorithm. Through iterative processes at both the operation and planning layers, the connection locations and capacity configurations of PV and energy storage are ultimately determined, thereby improving the overall system efficiency. The research results show that the proposed method can not only effectively address the fluctuations in photovoltaic (PV) output but also significantly optimize the configuration of PV and energy storage, thereby improving the operational efficiency and reliability of the distribution network. The optimized configuration scheme disclosed herein is of great significance for achieving a high proportion of renewable energy access, such as PV. Through the application of this method, the Xuehua distribution network can operate more stably and efficiently in the face of the fluctuations in PV power generation. This not only provides valuable experience and technical reference for the future development of power systems but also offers a practical solution for real-world engineering applications.

[0094] Based on the above method, this disclosure also provides a photovoltaic and energy storage optimization configuration device based on a snowflake grid, corresponding to the above method. The device includes an acquisition unit, an occurrence probability determination unit, and an optimization configuration unit, wherein: The acquisition unit is used to acquire historical load demand data and photovoltaic output data of the Snowflakes grid. The probability determination unit is used to determine typical scenarios of the correlation between photovoltaics and load, and the probability of occurrence of each typical scenario, based on historical load demand data and historical photovoltaic output data. An optimization configuration unit is used to determine the access locations and capacities of photovoltaic (PV) and energy storage using a constructed two-layer optimization configuration model. The model includes a planning layer and an operation layer. The planning layer uses system economic indicators as objective functions, while the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and PV and energy storage investment costs. The planning layer calculates the constraints and objective functions based on the optimization results of the operation layer and optimizes the access locations and capacities of PV and energy storage. The operation layer optimizes the output of PV and energy storage at various time periods based on the load curves corresponding to typical scenarios with a probability greater than a set value, the PV output curves, and the optimized access locations and capacities of PV and energy storage, using the objective functions and constraints of the operation layer, and feeds the optimization results back to the planning layer.

[0095] Furthermore, the optimization configuration unit is specifically used to use the photovoltaic capacity and access location, and the energy storage capacity and access location as particles in the improved particle swarm algorithm, initialize the particle position and velocity, and input them into the running layer; use the photovoltaic and energy storage access locations and capacities optimized by the planning layer to update the position and velocity of the particle swarm, and input the updated position and velocity of the particle swarm into the running layer until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or reaches the set number of iterations.

[0096] Furthermore, the probability determination unit is specifically used to determine the probability density functions of load demand and photovoltaic output based on historical load demand data and historical photovoltaic output data. It combines Latin hypercube sampling with Nataf transform to perform correlation sampling on the probability density functions of load demand and photovoltaic output, obtaining several time-series scenarios of load and photovoltaic power. A clustering algorithm is then used to reduce these time-series scenarios to a set of typical scenarios. Finally, the probability of occurrence of each typical scenario is calculated using the number of time-series scenarios participating in the typical scenario clustering and the total number of time-series scenarios.

[0097] Based on the same inventive concept as the above disclosure, this disclosure also provides an electronic device. The electronic device of this disclosure includes at least one processor and at least one memory electrically connected to the processor. The memory is electrically connected to the processor, wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described above.

[0098] It should be noted that the electrical connection between the above-mentioned units does not necessarily mean the connection between lines. The indirect connection method can be applied to the embodiments of this disclosure as long as it achieves the purpose of this disclosure.

[0099] Based on the same inventive concept, this disclosure also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the steps of the above method.

[0100] Although the present disclosure 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; and 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 disclosure.

Claims

1. A method for optimizing the configuration of photovoltaic and energy storage based on snowflake grids, characterized in that, The method includes, Obtain historical load demand data and photovoltaic power output data from Xuehua.com; Based on historical load demand data and historical photovoltaic output data, typical scenarios of the correlation between photovoltaic power and load and the probability of occurrence of each typical scenario are determined. The connection location and capacity of photovoltaic and energy storage are determined by using the constructed two-layer optimization configuration model of photovoltaic and energy storage. The model includes a planning layer and an operation layer. The planning layer uses system economic indicators as the objective function, and the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and investment costs for photovoltaics and energy storage. The planning layer is used to calculate the constraints and objective function of the planning layer based on the optimization results of the operation layer, and to optimize the access location and capacity of photovoltaic and energy storage. The operation layer is used to optimize the output of photovoltaic and energy storage in each time period based on the load curve and photovoltaic output curve corresponding to typical scenarios with a probability greater than a set value, as well as the optimized access location and capacity of photovoltaic and energy storage, and based on the objective function and constraints of the operation layer, and feeds the optimization results back to the planning layer. The constraints of the operating layer include energy storage location constraints, which are the location variables of the energy storage belonging to the energy storage candidate locations. The energy storage candidate locations are obtained by selecting energy storage locations based on the tie-line distribution information of the snowflake network.

2. The method according to claim 1, characterized in that, Using the constructed two-layer optimization configuration model for photovoltaics and energy storage, the access locations and capacities of photovoltaics and energy storage are determined, including: The capacity and access location of photovoltaics, and the capacity and access location of energy storage are used as particles in the improved particle swarm optimization algorithm. The particle positions and velocities are initialized and input into the execution layer. The positions and velocities of the particle swarm are updated using the optimized access locations and capacities of photovoltaics and energy storage from the planning layer. The updated positions and velocities of the particle swarm are then input into the execution layer until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or until the set number of iterations is reached.

3. The method according to claim 2, characterized in that, The system's economic indicator F2 is: in, For the investment costs of photovoltaics and energy storage, For operating costs, Indicates the discount rate; This refers to the lifespan of distributed photovoltaic and energy storage equipment; This represents the investment cost per unit capacity of distributed photovoltaic power generation. This represents the investment cost per unit capacity of energy storage. This represents the total installed capacity of distributed photovoltaic power. This indicates the total installed capacity of energy storage.

4. The method according to any one of claims 1-3, characterized in that, The operating cost F1 and voltage offset index VDI are: in, Indicates operating costs; Indicates network loss cost; This indicates the cost of purchasing electricity from the main grid; T represents the cost of curtailment; N represents the total number of time slots; and T represents the number of photovoltaic power plants. Indicates the price of photovoltaic power generation; Let represent the power generation of the i-th photovoltaic power station at time t; Indicates the price of network loss; This represents the network loss at time t; This indicates the main grid electricity purchase price; This represents the main network purchase volume at time t; Indicates the penalty price for solar power curtailment; U represents the amount of electricity wasted by the i-th photovoltaic power station at time t; i,t U represents the actual voltage value of node i at time t obtained from probabilistic power flow; i,N Let be the rated voltage of node i at time t.

5. The method according to any one of claims 1-3, characterized in that, The constraints at the operational layer also include: system power flow constraints, system operation constraints, energy storage system charging and discharging power constraints, photovoltaic power output constraints, and upper and lower limits of battery charging state constraints for energy storage.

6. The method according to any one of claims 1-3, characterized in that, The constraints at the planning level include: system power flow constraints, system operation constraints, upper and lower limits on the number of photovoltaic power sources connected, and upper and lower limits on the number of energy storage access groups.

7. The method according to any one of claims 1-3, characterized in that, Based on historical load demand data and historical photovoltaic (PV) output data, typical scenarios of PV-load correlation and the probability of occurrence for each typical scenario are determined, including: Based on historical load demand data and historical photovoltaic output data, the probability density functions of load demand and photovoltaic output are determined. By combining Latin hypercube sampling with Nataf transform, correlation sampling is performed on the probability density function of load demand and the probability density function of photovoltaic output to obtain several time-series scenarios of load and photovoltaic. Clustering algorithms were used to reduce the number of time-series load and photovoltaic scenarios to obtain a set of typical scenarios. By using the number of time-series scenarios participating in the typical scenario clustering and the total number of time-series scenarios, the probability of occurrence of each typical scenario is calculated.

8. A photovoltaic and energy storage optimized configuration device based on snowflake grid, characterized in that, The device includes an acquisition unit, an occurrence probability determination unit, and an optimization configuration unit, wherein: The acquisition unit is used to acquire historical load demand data and photovoltaic output data of the Snowflakes grid. The probability determination unit is used to determine typical scenarios of the correlation between photovoltaics and load, and the probability of occurrence of each typical scenario, based on historical load demand data and historical photovoltaic output data. An optimization configuration unit is used to determine the access locations and capacities of photovoltaic (PV) and energy storage using a constructed two-layer optimization configuration model. The model includes a planning layer and an operation layer. The planning layer uses system economic indicators as objective functions, while the operation layer uses operating costs and voltage deviation index as objective functions. The system economic indicators include operating costs and PV and energy storage investment costs. The planning layer calculates the constraints and objective functions based on the optimization results of the operation layer and optimizes the access locations and capacities of PV and energy storage. The operation layer optimizes the output of PV and energy storage at various time periods based on the load curves corresponding to typical scenarios with a probability greater than a set value, the PV output curves, and the optimized access locations and capacities of PV and energy storage, using the objective functions and constraints of the operation layer, and feeds the optimization results back to the planning layer. The constraints of the operating layer include energy storage location constraints, which are the location variables of the energy storage belonging to the energy storage candidate locations. The energy storage candidate locations are obtained by selecting energy storage locations based on the tie-line distribution information of the snowflake network.

9. The apparatus according to claim 8, characterized in that, The optimization configuration unit is specifically used to use the photovoltaic capacity and access location, and the energy storage capacity and access location as particles in the improved particle swarm algorithm, initialize the particle position and velocity, and input them into the execution layer; use the photovoltaic and energy storage access locations and capacities optimized by the planning layer to update the position and velocity of the particle swarm, and input the updated position and velocity of the particle swarm into the execution layer until the objective function of the planning layer reaches its minimum value under the constraints of the planning layer, or reaches the set number of iterations.

10. The apparatus according to claim 9, characterized in that, The system's economic indicator F2 is: in, For the investment costs of photovoltaics and energy storage, For operating costs, Indicates the discount rate; This refers to the lifespan of distributed photovoltaic and energy storage equipment; This represents the investment cost per unit capacity of distributed photovoltaic power generation. This represents the investment cost per unit capacity of energy storage. This represents the total installed capacity of distributed photovoltaic power. This indicates the total installed capacity of energy storage.

11. The apparatus according to any one of claims 8-10, characterized in that, The operating cost F1 and voltage offset index VDI are: in, Indicates operating costs; Indicates network loss cost; This indicates the cost of purchasing electricity from the main grid; T represents the cost of curtailment; N represents the total number of time slots; and T represents the number of photovoltaic power plants. Indicates the price of photovoltaic power generation; Let represent the power generation of the i-th photovoltaic power station at time t; Indicates the price of network loss; This represents the network loss at time t; This indicates the main grid electricity purchase price; This represents the main network purchase volume at time t; Indicates the penalty price for solar power curtailment; U represents the amount of electricity wasted by the i-th photovoltaic power station at time t; i,t U represents the actual voltage value of node i at time t obtained from probabilistic power flow; i,N Let be the rated voltage of node i at time t.

12. The apparatus according to any one of claims 8-10, characterized in that, The constraints at the operational layer also include: system power flow constraints, system operation constraints, energy storage system charging and discharging power constraints, photovoltaic power output constraints, and upper and lower limits of battery charging state constraints for energy storage.

13. The apparatus according to any one of claims 8-10, characterized in that, The constraints at the planning level include: system power flow constraints, system operation constraints, upper and lower limits on the number of photovoltaic power sources connected, and upper and lower limits on the number of energy storage access groups.

14. The apparatus according to any one of claims 8-10, characterized in that, The probability determination unit is specifically used to determine the probability density functions of load demand and photovoltaic output based on historical load demand data and historical photovoltaic output data. It combines Latin hypercube sampling with Nataf transform to perform correlation sampling on the probability density functions of load demand and photovoltaic output, obtaining several time-series load and photovoltaic scenarios. A clustering algorithm is then used to reduce these time-series scenarios to a set of typical scenarios. Finally, the probability of occurrence for each typical scenario is calculated using the number of time-series scenarios participating in the typical scenario clustering and the total number of time-series scenarios.

15. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method according to any one of claims 1-7.

16. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed by a processor, implements the steps of any one of the methods described in claims 1-7.

Images