Configuration optimization and power decision-making method of energy storage system in distribution network based on Archimedean optimization algorithm

Abstract

The present invention relates to the field of energy storage technology for distribution networks connected to distributed energy sources, and specifically discloses a distribution network energy storage system configuration optimization and power decision method based on the Archimedes optimization algorithm, which mainly solves the configuration scheme of the distribution network energy storage system connected to high-penetration distributed energy and the optimal charging and discharging power strategy during the operation cycle of the energy storage battery. The present invention first establishes a multi-objective energy storage optimization configuration and optimal charging and discharging power decision model that considers the minimum voltage deviation of the distribution network, the lowest line loss rate, and the optimal energy storage economy. Then, the optimization model is solved by establishing an Archimedes optimization algorithm. Finally, the effectiveness of the multi-objective optimization configuration and charging and discharging power decision instructions of the energy storage system on the IEEE33 node is verified through example analysis, and a variety of optimization algorithms are compared to prove that the energy storage optimization configuration and charging and discharging power decision based on the Archimedes optimization algorithm have good solution convergence performance and rationality of algorithm decision.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage for distribution networks connected to distributed energy sources, and in particular to a multi-objective optimization configuration and optimal charging and discharging power decision of energy storage systems for distribution networks based on an Archimedean optimization algorithm. Background Art

[0002] As a renewable energy source, wind and solar energy have received more and more attention from the energy industry in recent years. In the first half of 2020 alone, my country's wind power generation reached 237.9 billion kWh. With the proposal of the country's "dual carbon" goals and the grand blueprint it has drawn up, more and more wind and solar energy are connected to the distribution network in a distributed form. As the capacity of distributed energy connected to the distribution network increases, the distribution network is gradually developing towards a high penetration rate. However, distributed energy is extremely susceptible to the external environment, and its power has strong uncertainty and randomness, which will cause node voltage fluctuations; similarly, the increase in penetration will lead to high power output of DG, resulting in reverse power flow, which may cause serious voltage rise and other serious threats to the safe and stable operation of the distribution network system.

[0003] With the rapid advancement of material technology, the field of energy storage has been further developed. Extensive and flexible distributed energy storage is the main flexible and controllable resource of the distribution network, which can effectively smooth the fluctuation of distributed energy output power and effectively improve the absorption capacity of distributed energy. Therefore, how to reasonably and effectively configure the energy storage system in the distribution network and how to more optimally control the charging and discharging power of the energy storage system has become an urgent problem to be solved, which has very important practical and theoretical significance. Summary of the invention

[0004] The purpose of the present invention is to utilize the Archimedean optimization method to reasonably and effectively configure an energy storage system in a distribution network and to more optimally control the charging and discharging power of the energy storage system.

[0005] In order to achieve the above-mentioned invention object, the present invention provides the following technical solutions: A method for optimizing the configuration and power decision of a distribution network energy storage system based on an Archimedean optimization algorithm, the method comprising the following steps: Step 1: Build a distribution network and energy storage system model connected to distributed energy, including: Step (1-1) constructs the distribution network system operation model, which mainly includes the fluctuation deviation index of the distribution network node voltage and the network line loss rate during the operation of the distribution network; Step (1-2) constructs an energy storage system configuration model; it mainly includes the cost required for the full life cycle of energy storage and the output model of the energy storage system connected to a high-penetration distribution network; Steps (1-3) construct a high-penetration distribution network operation model, a configuration model for energy storage systems connected to the distribution network, and constraints, including: (1) node voltage constraints and line current constraints for high-penetration distribution networks connected to distributed energy; (2) power balance constraints; (3) flow constraints; (4) energy storage battery capacity and energy rate constraints; (5) energy storage battery SOC state constraints; (6) distributed energy output constraints; Step 2: Propose a specific multi-objective evaluation function, including: Step (2-1) defines a penalty objective function for node voltage over-limit behavior for the system safety of the distribution network nodes; Step (2-2) defines a penalty objective function for distribution network operation losses; Step (2-3) considers the location of the energy storage system in the distribution network and defines the capacity evaluation objective function; Step 3: Apply the Archimedean optimization algorithm to solve the multi-objective model optimization, including: Step (3-1) Initialize the parameters and optimization target of the Archimedean optimization algorithm; Step (3-2) initializes the positions of all wooden blocks, including all decision variables, and randomly assigns each wooden block an object volume and density, and an acceleration when running in water; Step (3-3) evaluates the initial population and selects the individual with the best fitness value to assign the best position, density, volume and acceleration, and updates the density and volume of each individual iteratively; Steps (3-4) define the transfer operator and density factor, gradually transforming the algorithm from the search phase to the utilization phase; Steps (3-5) use the mutual collisions between individual wooden blocks in the exploration phase to update the acceleration of individual wooden blocks; in the utilization phase, there is no mutual collision between wooden blocks, and the acceleration of individual wooden blocks is updated using the "utilization phase" formula; Steps (3-6) normalize the acceleration and calculate its percentage change; Steps (3-7) update the position of each individual and its associated attributes according to the transfer factor; Steps (3-8) use the objective function to update the fitness value of each individual, and record the individual with the best fitness value and its corresponding related attributes; Step 4: Output the optimal control decision, including: Step (4-1) outputs the energy storage system site selection and capacity determination decision; Step (4-2) outputs the optimal charging and discharging switching strategy for the energy storage system.

[0006] Compared with the prior art, the present invention has the following beneficial effects: the present method can effectively cope with the fluctuating output conditions of distributed energy resources, and effectively and reasonably optimize the configuration of the energy storage system and make decisions on the energy storage charging and discharging power for the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 is a flow chart of the method of the present invention; Figures 2 to 6 This is the result of the example analysis based on the IEEE-33 node system. Wind power is connected to nodes 9, 20, and 25 respectively, and photovoltaic power is connected to node 28. Lithium batteries are selected as the energy storage system components. Figure 2 The four optimization algorithms are Whale Optimization Algorithm (WOA), Particle Swarm Optimization Algorithm (PSO), Genetic Algorithm (GA) and Archimedean Optimization Algorithm (AOA). They solve the energy storage optimization configuration and energy storage power decision model. The current function curve changes with the number of iterations. Figure 2 As shown, the horizontal axis is the number of iteration steps and the vertical axis is the objective function value; Figure 3 The average amplitude of node voltage in all periods corresponding to the optimization configurations and control strategies of the four optimization algorithms, from top to bottom, are Archimedean algorithm (AOA), particle swarm optimization algorithm (PSO), whale optimization algorithm (WOA) and genetic algorithm (GA); Figure 4 and Figure 5 They are schematic diagrams of the decision output power and SOC state of the energy storage system in the full period cycle under the Archimedean algorithm (AOA); Figure 6 The statistical mean of the average load of all nodes in the whole period corresponding to the four algorithms is obtained. DETAILED DESCRIPTION

[0008] The present invention will be further described below in conjunction with the accompanying drawings.

[0009] The present invention provides a distribution network energy storage system configuration optimization and power decision-making method based on the Archimedean optimization algorithm, such as Figure 1 As shown, the following steps are included: Step 1: Build a distribution network and energy storage system model connected to distributed energy, including: Step (1-1) constructs the distribution network system operation model, which mainly includes the fluctuation deviation index of the distribution network node voltage and the network line loss rate during the distribution network operation: (1) Fluctuation deviation index of distribution network node voltage (2) Distribution network line loss rate Where N is the number of nodes, M is the number of branches; T is the statistical calculation period; are the voltage of the i-th node and the reference voltage at time t, respectively. is the maximum allowable deviation; Represents the network loss consumption of branch ij at time t.

[0010] Step (1-2) constructs an energy storage system configuration model, which mainly includes the cost required for the full life cycle of energy storage and the output model of the energy storage system connected to a high-penetration distribution network: (1) Life cycle cost of energy storage system Where: They are the equivalent annual investment cost and annual operation and maintenance cost of the energy storage system respectively; They are the annual low-storage, high-generation arbitrage and government subsidies for electricity sales; is the fixed investment cost of a single energy storage system; is the number of energy storage units installed; are the configured capacity and configured power of the nth energy storage system respectively; a, b are the unit power cost and unit capacity cost of the energy storage system; is the annual capital recovery rate; r is the discount rate; y is the useful life; is the operation and maintenance coefficient.

[0011] (2) Energy storage system output model Where: is the output power of the energy storage unit at time t; are the discharge and charge efficiencies of the energy storage battery respectively.

[0012] Steps (1-3) construct a high-penetration distribution network operation model, a configuration model for energy storage systems connected to the distribution network, and constraints. Specifically, they include: (1) node voltage constraints and line current constraints for high-penetration distribution networks connected to distributed energy; (2) power balance constraints; (3) flow constraints; (4) energy storage battery capacity and energy rate constraints; (5) energy storage battery SOC state constraints; and (6) distributed energy output constraints.

[0013] Step 2: Propose a specific multi-objective evaluation function, including: Step (2-1) defines the node voltage over-limit penalty objective function for the system safety of the distribution network nodes: Step (2-2) formulates a penalty objective function for distribution network operation losses: Step (2-3) considers the location of the energy storage system in the distribution network and defines the capacity evaluation objective function: Step 3: Apply the Archimedean optimization algorithm to solve the multi-objective model optimization, including: Step (3-1) Initialize the parameters and optimization target of the Archimedean optimization algorithm: in are the search upper bound and search lower bound of each decision variable, respectively. Represents the i-th group in the object.

[0014] Step (3-2) initializes the positions of all wooden blocks, including all decision variables, and initializes the object volume and density of each wooden block, as well as the acceleration running in water: Step (3-3) evaluates the initial population and selects the individual with the best fitness value to assign the best position , optimal density , optimal volume , optimal acceleration , update the density and volume of each individual iteratively: in These are, respectively, the best density and volume of objects discovered so far.

[0015] Steps (3-4) define the transfer operator and density factor, gradually transforming the algorithm from the search phase to the utilization phase: The transfer TF increases gradually with time until it reaches 1. Where t and t max are the current iteration number and the maximum iteration number respectively. Similarly, the density reduction factor d also helps the Archimedean algorithm to perform global to local search. When the density factor is used, it decreases over time.

[0016] Steps (3-5) use the mutual collisions between individual wooden blocks in the exploration phase to update the acceleration of individual wooden blocks; in the utilization phase, there is no mutual collision between wooden blocks, and the acceleration of individual wooden blocks is updated using the "utilization phase" formula: (1) Exploration stage: If , then there is a collision between objects, and a random material is selected to update the object acceleration of iteration T+1: (2) Utilization stage: If , there is no collision between the objects, and the object acceleration at iteration t+1 is updated using the following formula: in The acceleration, density, and volume corresponding to a random material.

[0017] Steps (3-6) normalize the acceleration and calculate its percentage change: where u and l are normalized ranges and are set to 0.9 and 0.1, respectively. The norm determines the percentage of the step size that each block will change. If block i is far from the global optimum, the acceleration value will be high - this means that the object will be in the exploration phase; otherwise, it is in the exploitation phase. This illustrates how the search transitions from the exploration phase to the exploitation phase. In normal cases, the acceleration factor starts with a large value and decreases over time. This helps the search agent move towards the global optimal solution while moving away from the local solution. However, it is worth mentioning that there may still be some search agents that need to stay in the exploration phase more time than normal. Therefore, the Archimedean algorithm achieves a balance between exploitation and exploitation.

[0018] Steps (3-7) update the position of each individual and its associated attributes according to the transfer factor: Among them C 1 , C 2 The constant values ​​are 2 and 6 respectively.

[0019] Steps (3-8) use the objective function to update the fitness value of each individual, record the individual with the best fitness value and its corresponding related attributes, and repeat the iterative cycle until the convergence criterion is reached.

[0020] Step 4: Output the optimal control decision, including: Step (4-1): Output energy storage system site selection and capacity determination decision; Step (4-2): Output the optimal charging and discharging power strategy of the energy storage system.

[0021] Figure 2The four optimization algorithms are Whale Optimization Algorithm (WOA), Particle Swarm Optimization Algorithm (PSO), Genetic Algorithm (GA) and Archimedean Optimization Algorithm (AOA). They solve the energy storage optimization configuration and energy storage power decision model. The current function curve changes with the number of iterations. Figure 2 As shown, it can be seen that compared with other algorithms, the Archimedes optimization algorithm can quickly and effectively converge to a better target value.

[0022] Figure 3 The average amplitude of node voltage in all periods corresponding to the optimization configuration and power control strategy of the four optimization algorithms are Archimedes algorithm (AOA), particle swarm optimization algorithm (PSO), whale optimization algorithm (WOA) and genetic algorithm (GA) from top to bottom. As shown in the figure, the voltage distribution curve determined by the Archimedes optimization algorithm is more stable and closer to the standard value than that of other algorithms.

[0023] Figure 4 and Figure 5 Schematic diagrams of the decision output power and SOC state of the energy storage system in the full period under the Archimedean algorithm (AOA).

[0024] Figure 6 The statistical mean of the average load of all nodes in all time periods under the four algorithms are obtained. It can be seen that the load fluctuation level under the Archimedes optimization algorithm is smaller, which effectively improves the load fluctuation of the power grid.

[0025] The following table shows the corresponding energy storage system configuration solutions under the four algorithms: It can be concluded that the energy storage system optimization configuration and power decision-making based on the Archimedes optimization algorithm have better results.

[0026] In summary, the present invention constructs a multi-objective evaluation system with the minimum voltage fluctuation deviation, the minimum network loss consumption and the optimal economic efficiency of energy storage planning in the distribution network, and uses the established Archimedes optimization algorithm to solve the model, and finally obtains the optimal energy storage configuration scheme that can enable the stable operation of the distribution network under the condition of frequent distributed output fluctuations and the optimal charging and discharging power strategy during the operation cycle of the energy storage battery. By comparing multiple algorithms, it is proved that the optimization scheduling model based on Archimedes optimization has good optimization and convergence performance.

[0027] The above-mentioned example analysis is an example analysis of the patent protection content, and does not impose any form of limitation on the present invention. Any modification, equivalent substitution, improvement, etc. made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A distribution network energy storage system configuration optimization and power decision-making method based on Archimedes optimization algorithm, characterized in that: The method comprises the following steps: Step 1: Build a distribution network and energy storage system model connected to distributed energy, including: Step (1-1) constructs a distribution network system operation model, which mainly includes the fluctuation deviation index of the distribution network node voltage and the network line loss rate during the operation of the distribution network; Step (1-2) constructs an energy storage system configuration model, which mainly includes: the cost required for the full life cycle of energy storage, and the output model of the energy storage system connected to a high-penetration distribution network; Steps (1-3) construct a high-penetration distribution network operation model, a configuration model for energy storage systems connected to the distribution network, and constraints, including: (1) node voltage constraints and line current constraints for high-penetration distribution networks connected to distributed energy; (2) power balance constraints; (3) flow constraints; (4) energy storage battery capacity and energy rate constraints; (5) energy storage battery SOC state constraints; (6) distributed energy output constraints; Step 2: Propose a specific multi-objective evaluation function: Step (2-1) defines a penalty objective function for node voltage over-limit behavior for the system safety of the distribution network nodes; Step (2-2) setting a penalty objective function for distribution network operation losses; Step (2-3) considers the location of the energy storage system in the distribution network and sets the evaluation objective function; Step 3: Apply the Archimedean optimization algorithm to solve the multi-objective model optimization, including: Step (3-1) Initialize the parameters and optimization target of the Archimedean optimization algorithm; Step (3-2) initializes the positions of all wooden blocks, including all decision variables, and initializes the object volume and density randomly assigned to each wooden block, as well as the acceleration running in water; Step (3-3) evaluates the initial population and selects the individual with the best fitness value to assign the best position, density, volume and acceleration, and updates the density and volume of each individual iteratively; Steps (3-4) define the transfer operator and density factor, gradually transforming the algorithm from the search phase to the utilization phase; Steps (3-5) update the acceleration of individual wooden blocks according to the mutual collisions between individual wooden blocks in the exploration phase. In the utilization phase, there are no mutual collisions between wooden blocks, and the acceleration of individual wooden blocks is updated using the "utilization phase" formula. Steps (3-6) normalize the acceleration and calculate its percentage change; Steps (3-7) update the position of each individual and its associated attributes according to the transfer factor; Steps (3-8) use the objective function to update the fitness value of each individual, and record the individual with the best fitness value and its corresponding related attributes; Step 4: Output optimal control decision; Step (4-1) outputs the energy storage system site selection and capacity determination decision; Step (4-2) outputs the charging and discharging power decision of the energy storage system.

2. The method according to claim 1, characterized in that In the steps (3-4), the density factor and the transfer factor are defined to effectively control the optimization and exploration processes of the algorithm, and to reasonably balance the two stages to achieve more effective optimization.

3. The method according to claim 1, characterized in that In the step (3-1), the objectives are normalized during the multi-objective comprehensive optimization process to avoid falling into a local optimal solution during the algorithm optimization process.

4. The method according to claim 1, characterized in that: In the step (1-1), (1) Fluctuation deviation index of distribution network node voltage (2) Distribution network line loss rate Where N is the number of nodes, M is the number of branches; T is the statistical calculation period; are the voltage of the i-th node and the reference voltage at time t, respectively. is the maximum allowable deviation; Represents the network loss consumption of branch ij at time t.

5. The method according to claim 1, characterized in that In the steps (1-2), (1) Life cycle cost of energy storage system Where: They are the equivalent annual investment cost and annual operation and maintenance cost of the energy storage system respectively; They are the annual low-storage, high-generation arbitrage and government subsidies for electricity sales; is the fixed investment cost of a single energy storage system; is the number of energy storage units installed; are the configured capacity and configured power of the nth energy storage system respectively; a, b are the unit power cost and unit capacity cost of the energy storage system; is the annual capital recovery rate; r is the discount rate; y is the useful life; is the operation and maintenance coefficient; (2) Energy storage system output model Where: is the output power of the energy storage unit at time t; are the discharge and charge efficiencies of the energy storage battery respectively.

6. The method according to claim 1, characterized in that In the step 2, Step (2-1) defines the node voltage over-limit penalty objective function for the system safety of the distribution network nodes: Step (2-2) formulates a penalty objective function for distribution network operation losses: Step (2-3) considers the location of the energy storage system in the distribution network and defines the capacity evaluation objective function:

7. The method according to claim 1, characterized in that In step 3, Step (3-1) Initialize the parameters and optimization target of the Archimedean optimization algorithm: in are the search upper bound and search lower bound of each decision variable, respectively. represents the i-th group in the object; Step (3-2) initializes the positions of all wooden blocks, including all decision variables, and initializes the object volume and density of each wooden block, as well as the acceleration running in water: Step (3-3) evaluates the initial population and selects the individual with the best fitness value to assign the best position , optimal density , optimal volume , optimal acceleration , update the density and volume of each individual iteratively: in are respectively the best density and volume of objects discovered so far; Steps (3-4) define the transfer operator and density factor, gradually transforming the algorithm from the search phase to the utilization phase: The transfer TF increases gradually with time until it reaches 1; t and t max They are the current number of iterations and the maximum number of iterations respectively; Steps (3-5) use the mutual collisions between individual wooden blocks in the exploration phase to update the acceleration of individual wooden blocks; in the utilization phase, there is no mutual collision between wooden blocks, and the "utilization phase" formula is used to update the acceleration of individual wooden blocks: (1) Exploration stage: If , then there is a collision between objects, and a random material is selected to update the object acceleration of iteration T+1: (2) Utilization stage: If , there is no collision between the objects, and the object acceleration at iteration t+1 is updated using the following formula: in The acceleration, density, and volume corresponding to a random material; Steps (3-6) normalize the acceleration and calculate its percentage change: where u and l are normalized ranges and are set to 0.9 and 0.1, respectively; The norm determines the percentage of the step size that each block will change; Steps (3-7) update the position of each individual and its associated attributes according to the transfer factor: Where C1 and C2 are constant values ​​2 and 6 respectively; Steps (3-8) use the objective function to update the fitness value of each individual, record the individual with the best fitness value and its corresponding related attributes, and repeat the iterative cycle until the convergence criterion is reached.

Images