A charging station participating in peak-valley regulation optimization method based on improved PSO and related device

Abstract

The application provides a charging station participating in distribution network peak-valley regulation optimization method based on improved PSO and related devices, and the method steps include: firstly, obtaining power load original data of a distribution network and power battery parameter data of centralized charging stations participating in regulation; secondly, obtaining predicted load power of each period and daily average load according to the power load original data; then, based on the load power of each period and the daily average load, presetting the direction of power exchange between the power battery and the distribution network in each period, and eliminating the correlation between the power batteries; finally, taking the charging and discharging power of each power battery corresponding to the optimized task after eliminating the correlation as a low-dimensional optimization problem, combining the parameter data, and solving by using an improved particle swarm algorithm to obtain the charging and discharging power strategy of each power battery in each period. The application can improve the solving efficiency of the distribution network peak-valley regulation by splitting the high-dimensional optimization task into a low-dimensional optimization problem and then solving by using the improved particle swarm algorithm.

Description

Technical Field

[0001] This invention belongs to the field of peak-valley regulation technology in power systems, specifically relating to an optimization method and related device for charging stations to participate in peak-valley regulation based on improved PSO. Background Technology

[0002] As the number of electric vehicles increases, their impact on the power system becomes increasingly significant. The charging demand for electric vehicles fluctuates dramatically in short periods, and this load fluctuation directly affects the stability of the power grid. Furthermore, the peak-to-valley load difference in the power grid is gradually widening, making traditional power regulation methods insufficient to meet the demands during peak load periods. Moreover, with the integration of renewable energy sources, grid dispatch faces even greater challenges.

[0003] Currently, research on centralized charging stations' participation in grid peak-valley regulation is still in its developmental stage. Existing technical solutions mainly alleviate grid load fluctuations by optimizing electric vehicle charging strategies, such as ordered charging strategies and V2G technology. However, these solutions are insufficient in rapidly responding to load fluctuations in a short period of time, and they also suffer from directional constraints in electric vehicle charging and discharging control, resulting in limited operational efficiency and economy. Furthermore, the particle swarm optimization algorithm used in existing solutions is prone to getting trapped in local optima in high-dimensional optimization problems, failing to efficiently solve complex grid regulation problems and not fully leveraging the potential of centralized charging stations in battery resource integration, management, and control. Summary of the Invention

[0004] In view of this, the present invention provides a method and related device for optimizing the peak-valley regulation of charging stations in the distribution network based on improved PSO, aiming to solve the search efficiency problem of existing schemes in high-dimensional optimization by improving the particle swarm optimization algorithm, and improve the system regulation capability and economy.

[0005] To achieve the above objectives, the technical solution provided by the present invention is as follows:

[0006] In a first aspect, the present invention provides a method for optimizing the peak-valley regulation of power distribution networks based on improved PSO (Power Distribution System Optimization), comprising the following steps:

[0007] Obtain raw data on the load power of the distribution network, as well as parameter data of the power batteries involved in regulation in the centralized charging station;

[0008] Based on the original load power data, the predicted load power and daily average load for each time period are calculated.

[0009] Based on the load power and daily average load for each time period, the direction of power exchange between the power battery and the distribution network is preset for each time period in order to eliminate the correlation between the power batteries;

[0010] After eliminating correlation, the charging and discharging power optimization task for each power battery is treated as a low-dimensional optimization problem. Combined with parameter data, an improved particle swarm optimization algorithm is used to solve the problem, thereby obtaining the charging and discharging power strategy for each power battery at each time period.

[0011] In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is solved by using the particle swarm optimization algorithm separately, and finally all the optimal solutions are integrated to obtain the global optimal solution.

[0012] Furthermore, the predicted load power for each time period is the average of the load power for the corresponding time period of the distribution network in the previous 7 days; the daily average load is the arithmetic mean of the predicted load power for 96 time periods on that day.

[0013] Furthermore, the improved particle swarm optimization algorithm adopts a sub-vector-based improved particle swarm optimization algorithm, and the solution process includes:

[0014] Map each power battery corresponding to a low-dimensional optimization problem to a particle and initialize all particle indices.

[0015] Initialize the position and velocity of each particle, and verify and correct the position and velocity of each particle according to the preset constraints; the position coordinates of the particles are the charging and discharging power of the power battery at each time period.

[0016] For each particle, a particle swarm optimization algorithm is executed based on a preset objective function and the particle's iterative update formula to iteratively update the particle's position and velocity;

[0017] Based on the particle index, all particles are traversed sequentially, and the iterative optimization process of the particle swarm optimization algorithm for each particle is repeated until the termination condition is met, at which point the individual optimal position of each particle is obtained.

[0018] By integrating the optimal individual positions of all particles, a global charging and discharging power strategy is obtained.

[0019] Furthermore, the improved particle swarm optimization algorithm aims to minimize grid load fluctuations, and the corresponding objective function is:

[0020]

[0021] In the formula, Let be the objective function. To predict the load power for period i, The average daily load Let j be the power exchanged between the j-th power battery and the distribution network during the predicted time period i. The number of power batteries involved in regulation.

[0022] Furthermore, the iterative update formula for particles is:

[0023]

[0024]

[0025] In the formula, and These are the velocity vectors of the i-th particle in the k-th generation and the (k+1)-th generation, respectively. and These are the position vectors of the i-th particle in the k-th and (k+1)-th generations, respectively. and Let be the individual optimal position vector and the group optimal position vector of the i-th particle in the k-th generation, respectively. and Given two random numbers between 0 and 1 in the k-th generation. For inertial weights, and All of these are learning factors.

[0026] Furthermore, the preset constraints include:

[0027] Charge / discharge direction constraints:

[0028]

[0029] In the formula, and sgn(·) represents the power exchanged between the j1-th and j2-th power batteries and the distribution network during the predicted time period i, respectively, and sgn(·) is the sign function.

[0030]

[0031] In the formula, P is the power exchanged between the power battery and the power distribution network, sgn(P) = 1 indicates that the power battery is discharging, sgn(P) = 0 indicates that the power battery is neither discharging nor charging, and sgn(P) = -1 indicates that the power battery is charging.

[0032] SOC constraints:

[0033]

[0034] In the formula, Let j be the state of charge of the j-th power battery during the predicted time period i. and These are the lower and upper limits of the state of charge, respectively;

[0035] Charge and discharge power limits:

[0036]

[0037] In the formula, Let j be the power exchanged between the j-th power battery and the distribution network during the predicted time period i. and These are the minimum and maximum power exchanges between the power battery and the power distribution network, respectively.

[0038] Number of iterations and convergence constraints:

[0039]

[0040]

[0041] In the formula, For the number of iterations, This represents the maximum number of iterations. and These are the global optimal solutions for the (t+1)th and tth iterations, respectively. To set a threshold

[0042] Furthermore, after iteratively updating the particle's position and velocity, if the particle position does not meet the preset constraints, the charging and discharging power of the power battery is adjusted directly to adjust the particle position. Specifically, this includes:

[0043] If the charging / discharging direction constraint is not met, the charging / discharging power of the power battery in the corresponding time period will be modified as follows:

[0044]

[0045] If the charging and discharging power limit constraint is not met, the charging and discharging power of the power battery for the corresponding time period will be modified as follows:

[0046]

[0047] If the SOC constraint is not met, the charging and discharging power of the power battery for the corresponding time period will be modified as follows:

[0048]

[0049] In the formula, The duration of a single time period. Let j be the actual usable capacity of the j-th power battery. Let j be the state of charge of the j-th power battery during the predicted time period i-1. Let be the power exchanged between the j-th power battery and the distribution network during the predicted time period i-1.

[0050] Secondly, the present invention provides a charging station participation in distribution network peak-valley regulation optimization device based on improved PSO, comprising:

[0051] The data acquisition module is used to acquire raw load power data of the distribution network and parameter data of the power batteries involved in regulation in the centralized charging station.

[0052] The load parameter calculation module is used to calculate the predicted load power and daily average load for each time period based on the original load power data.

[0053] The power direction preset module is used to preset the direction of power exchange between the power battery and the distribution network in each time period based on the load power and daily average load in each time period, so as to eliminate the correlation between the power batteries.

[0054] The optimization solution module is used to treat the charging and discharging power optimization task corresponding to each power battery after eliminating correlation as a low-dimensional optimization problem, and combine it with parameter data to solve it using an improved particle swarm algorithm to obtain the charging and discharging power strategy of each power battery at each time period.

[0055] In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is solved by using the particle swarm optimization algorithm separately, and finally all the optimal solutions are integrated to obtain the global optimal solution.

[0056] Thirdly, the present invention provides a computer device, the device including a processor and a memory:

[0057] The memory is used to store computer programs and send the instructions of the computer programs to the processor;

[0058] The processor executes instructions from the computer program, such as the first aspect, a method for optimizing the peak-valley regulation of charging stations participating in the distribution network based on an improved PSO.

[0059] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements, as in the first aspect, a method for optimizing the peak-valley regulation of a charging station based on an improved PSO for participating in distribution network regulation.

[0060] In summary, this invention provides a method and related apparatus for optimizing charging station participation in distribution network peak-valley regulation based on an improved PSO (Particle Swarm Optimization) algorithm. The method includes the following steps: acquiring raw load power data of the distribution network and parameter data of the power batteries participating in regulation at the centralized charging station; calculating the predicted load power and daily average load for each time period based on the raw load power data; pre-setting the direction of power exchange between the power batteries and the distribution network for each time period based on the load power and daily average load to eliminate the correlation between the power batteries; treating the optimization task of charging and discharging power corresponding to each power battery after eliminating the correlation as a low-dimensional optimization problem, and solving it using an improved particle swarm optimization algorithm in conjunction with the parameter data to obtain the charging and discharging power strategy of each power battery in each time period; wherein, when solving the problem using the improved particle swarm optimization algorithm, each low-dimensional optimization problem is optimized separately using the particle swarm optimization algorithm, and finally, all optimal solutions are integrated to obtain the global optimal solution. This invention eliminates the correlation between power batteries by pre-setting the direction of power exchange between power batteries and the distribution network at different time periods. It breaks down the high-dimensional charging and discharging power optimization task into a low-dimensional optimization problem corresponding to each power battery. After using an improved particle swarm optimization algorithm to find the best solution in each low dimension, it integrates the global optimal solution. This solves the problems of insufficient response to load fluctuations, limited charging and discharging direction constraints, and the tendency of traditional particle swarm optimization algorithms to fall into local optima in high-dimensional optimization. It effectively improves the high-dimensional solution efficiency, system regulation capability, and operational economy of distribution network peak and valley regulation. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 A flowchart of a method for optimizing the peak-valley regulation of a charging station based on an improved PSO, provided in an embodiment of the present invention;

[0063] Figure 2 A flowchart of the improved particle swarm optimization algorithm provided in an embodiment of the present invention;

[0064] Figure 3 A block diagram of a charging station participating in distribution network peak-valley regulation optimization device based on improved PSO provided in an embodiment of the present invention;

[0065] Figure 4 This is a block diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0066] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0067] The background technology of this invention will be further introduced below.

[0068] The widespread adoption of electric vehicles (EVs) has brought new challenges to electricity demand. EV charging demand fluctuates dramatically in short periods, directly impacting grid stability. Furthermore, the peak-to-valley load difference in the grid is widening, making traditional power regulation methods insufficient to meet peak load demands. With the integration of renewable energy, grid dispatch faces even greater challenges. Currently, the global EV penetration rate continues to rise, and large-scale EV charging has become one of the main causes of grid load fluctuations. For example, during weekday evening peak hours, a large number of EVs charging can cause a significant short-term increase in local grid load, far exceeding the grid's normal carrying capacity, leading to voltage dips, frequency fluctuations, and in severe cases, even local grid paralysis. Simultaneously, the intermittent and random nature of renewable energy sources such as wind and solar power, coupled with their large-scale grid connection, further exacerbates grid load uncertainty—changes in solar intensity and wind speed fluctuations cause dramatic fluctuations in renewable energy output. The randomness of EV charging load combined with the volatility of renewable energy output exponentially increases the complexity of grid dispatch. Traditional power dispatch methods based on fixed-period regulation are no longer suitable for this dynamically changing load scenario, necessitating a flexible and efficient load regulation mechanism to balance grid supply and demand.

[0069] Against this backdrop, centralized charging stations (CCS) are gradually being put on the agenda as a new load regulation method. Centralized charging stations utilize the large-scale battery resources of electric vehicles to participate in peak-valley regulation of the power grid. They offer not only fast response and high efficiency but also reduced losses by decreasing power transmission distance, providing economic and environmental benefits. Compared to distributed charging piles, centralized charging stations have significant advantages such as centralized battery resources, high management efficiency, and strong regulation capabilities. They can typically accommodate dozens to hundreds of electric vehicles charging simultaneously and are equipped with a professional battery management system (BMS) and dispatch control system, enabling unified monitoring, management, and dispatch of all power batteries within the station. From the perspective of power grid operation, centralized charging stations can serve as virtual energy storage power stations for the power grid. Their large-scale power battery clusters can quickly absorb excess power from the grid (valley charging) and release power during peak grid load periods (peak discharge), achieving the spatial and temporal transfer of power and thus smoothing out the peak-valley difference in the power grid. A large-scale centralized charging station has ample energy storage capacity, capable of releasing sufficient power to the grid during peak hours, effectively alleviating local grid load pressure. Simultaneously, centralized charging stations are typically located near grid load centers or distribution network nodes, shortening power transmission distances and reducing transmission losses. This reduces energy waste and lowers grid operating costs, achieving both economic and environmental benefits. Furthermore, centralized charging stations can be integrated with renewable energy power plants, absorbing unstable outputs from wind and solar power, achieving efficient renewable energy utilization, and further increasing the proportion of clean energy in the grid.

[0070] Currently, research on centralized charging stations (CCS) participating in grid peak-valley regulation is still in its developmental stage. Existing technical solutions mainly focus on how to mitigate grid load fluctuations by optimizing electric vehicle charging strategies. Existing solutions are typically based on the following two methods:

[0071] 1. Orderly Charging Strategy: This approach optimizes the distribution of charging load by regulating the charging time of electric vehicles (EVs), minimizing excessive concentration of charging during peak hours. Specifically, it controls the start and stop times of EV charging at charging stations to distribute the grid load and balance grid load fluctuations. The advantage of this approach is its ability to effectively reduce peak grid load during charging. However, due to its reliance on time-based scheduling, its effectiveness is often limited by the sudden nature of EV charging demand, failing to respond quickly to load fluctuations and failing to fully consider the potential of EVs as a power source. Existing orderly charging strategies often employ a pre-set time window scheduling model, such as distributing charging demand from peak hours to off-peak hours. However, this scheduling method lacks flexibility and cannot cope with sudden grid load fluctuations—such as sudden increases in grid load due to extreme weather or load recovery after a sudden power outage. The response delay of orderly charging strategies is relatively long, making it difficult to meet the needs of real-time grid control. Furthermore, existing orderly charging strategies only focus on the "charging" stage, failing to fully explore the discharge potential of EV batteries, treating EVs merely as a load and ignoring their value as mobile energy storage units, leading to a waste of grid control resources. Meanwhile, orderly charging strategies often require users to adjust their charging time. If users have an urgent charging need, the scheduling strategy will not be able to be executed, further limiting its practical application effect.

[0072] 2. V2G (Vehicle to Grid) Technology: This technology utilizes the electric vehicle's battery to charge during periods of low grid load and discharge in the reverse during periods of high grid load, transferring the stored energy from the battery back to the grid to balance the grid load. This solution provides a certain level of power support through the owner's electric vehicle battery, especially during peak grid periods. However, V2G technology still faces a series of technical and commercialization obstacles, including limitations in battery life, challenges in two-way communication and control between the electric vehicle and the grid, and directional constraints on charging and discharging. Furthermore, the high cost of battery management and charging infrastructure hinders its widespread adoption. From a technical perspective, the core bottleneck of V2G technology lies in battery life degradation. The number of charge-discharge cycles for electric vehicles is limited, and frequent bidirectional charging and discharging accelerates battery aging, leading to lower participation from vehicle owners and charging station operators. In terms of communication and control, V2G technology requires real-time bidirectional communication between electric vehicles, charging stations, and the power grid dispatch center, demanding extremely low communication latency to ensure timely execution of charging and discharging commands. However, current communication technologies struggle to reliably meet this requirement, exhibiting issues such as command transmission delays and signal interruptions, potentially leading to a disconnect between charging / discharging strategies and grid demands. Furthermore, V2G technology faces strict constraints on the directionality of charging and discharging, requiring precise control of the charging and discharging status of individual electric vehicles to prevent situations where some vehicles are charging while others are discharging simultaneously. Existing control algorithms struggle to achieve precise coordinated control of a large number of vehicles. From a commercial perspective, V2G technology necessitates the modification of existing charging infrastructure by adding bidirectional charging and discharging modules, leading to substantial investments on a large scale. Simultaneously, the current lack of a sound electricity pricing mechanism and revenue distribution model prevents charging station operators and vehicle owners from obtaining reasonable economic returns, further hindering the widespread adoption of V2G technology.

[0073] Given the load curve, a key issue to address is how operators of centralized charging stations can determine the charging and discharging time and power of each battery within the station to improve the load curve. While the two aforementioned solutions address the peak-valley load difference to some extent, neither can respond quickly to load fluctuations and both suffer from directional constraints in electric vehicle charging and discharging control, limiting operational efficiency and economy. Furthermore, although existing solutions achieve some degree of grid load regulation, they fail to provide efficient solutions for high-dimensional optimization problems. In particular, particle swarm optimization (PSO) algorithms tend to get trapped in local optima as the dimension increases, leading to a significant decrease in performance in high-dimensional problems involving centralized charging stations participating in distribution network peak-valley regulation. Moreover, existing solutions often fail to fully consider the advantages of centralized charging stations as auxiliary grid regulation resources, especially their potential in battery resource integration, management, and control. In practical applications, the optimization problem of centralized charging stations participating in peak-valley regulation of power distribution networks is a typical high-dimensional nonlinear optimization problem. Its optimization dimension equals the product of the number of power batteries and the number of time periods. As the number of power batteries increases, the optimization dimension rises significantly, making it difficult for traditional optimization algorithms to solve efficiently. In high-dimensional spaces, the traditional particle swarm optimization (PSO) algorithm faces the problem of a significantly expanded search range for particles, leading to chaotic search directions and slower convergence speeds. It often gets stuck in local optima before finding the global optimum, resulting in the optimized charging and discharging strategy failing to achieve the best peak-valley regulation effect. For example, in the regulation scenario of large-scale centralized charging stations, the convergence speed drops significantly when using the traditional PSO algorithm, and the final optimized solution deviates considerably from the global optimum, failing to meet the accuracy requirements of power grid regulation. In addition, most existing solutions adopt a "decentralized control" model, which means that each electric vehicle is charged and discharged separately. This does not make full use of the resource integration advantages of centralized charging stations, resulting in low control efficiency. For example, the charging and discharging strategies of multiple electric vehicles conflict with each other, resulting in some vehicles charging and some vehicles discharging, causing secondary energy conversion losses and reducing the economic efficiency of grid control.

[0074] Particle Swarm Optimization (PSO) algorithm continuously updates the velocity and position of particles to ultimately obtain the optimal position of the particles, which is the optimal solution to the optimization problem. The information of particle I can be represented by a D-dimensional vector, with its position as... The speed is The position and velocity vectors of other particles are similar, and the update equations for velocity and position are as follows:

[0075] (1)

[0076] (2)

[0077] Where ω is the inertia weight; It is the D-dimensional velocity vector generated by particle I through k iterations; It is the D-dimensional position vector generated by particle I through k iterations; It is the local optimum value of particle I after k iterations; It is the global optimal value of the entire population after k iterations; c1 and c2 are learning factors; rand1 and rand2 are random numbers between [0, 1].

[0078] Although the PSO algorithm is less affected by the dimensionality of the optimization problem, its search performance still deteriorates as the dimensionality increases, making it difficult to escape local optima during the search process. In the problem of centralized charging stations participating in peak-valley regulation of power distribution networks, the particle dimensionality increases 96 times with the number of power batteries in the centralized charging station. This dramatic increase in dimensionality affects the particle optimization effect, therefore, it is necessary to seek more efficient algorithms for solving high-dimensional problems. Traditional particle swarm optimization algorithms have the following four drawbacks:

[0079] 1. Slow response speed: Existing charging regulation strategies (such as those using traditional particle swarm optimization algorithms), especially ordered charging strategies, mainly rely on preset time windows for charging scheduling. As a result, the system struggles to respond promptly to sudden fluctuations in grid load, limiting the potential of electric vehicles as a flexible regulatory resource.

[0080] 2. Local Optimality Problem: Traditional particle swarm optimization (PSO) algorithms are prone to getting trapped in local optima in high-dimensional optimization problems, especially when the dimensionality of the problem is high, resulting in a significant decrease in search performance. In the scenario of centralized charging stations participating in grid regulation, the dimensionality of PSO increases by 96 times with the increase in the number of power batteries, leading to a decrease in optimization performance and difficulty in obtaining the global optimum.

[0081] 3. Lack of consideration for charging and discharging directionality constraints: Existing solutions rarely consider the directional constraints of electric vehicle charging and discharging. These constraints mean that electric vehicles participating in regulation should maintain a consistent charging or discharging state within the same time period, avoiding simultaneous charging and discharging to prevent ineffective secondary energy conversion. Current solutions neglect this constraint, potentially leading to low grid regulation efficiency.

[0082] 4. Low efficiency in solving high-dimensional optimization problems: As the number of batteries in a centralized charging station increases, the dimensionality of the optimization problem rises significantly. Traditional particle swarm optimization (PSO) algorithms struggle to maintain good search performance in such high-dimensional environments. With increasing dimensionality, the search space of the algorithm also expands, making traditional PSO algorithms prone to getting trapped in local optima and unable to efficiently solve complex power grid control problems.

[0083] Based on this, the present invention provides a method and related apparatus for optimizing charging stations to participate in distribution network peak-valley regulation based on an improved PSO algorithm. The aim is to solve the search efficiency problem of existing schemes in high-dimensional optimization by improving the particle swarm optimization algorithm, thereby enhancing the system's regulation capability and economic efficiency. Some embodiments of the present invention are described in detail below.

[0084] Please see Figure 1 This embodiment provides a method for optimizing the peak-valley regulation of charging stations in the distribution network based on improved PSO, including the following steps:

[0085] S1: Obtain raw data of the load power of the distribution network, as well as parameter data of the power batteries involved in the regulation in the centralized charging station.

[0086] It should be noted that the raw load power data refers to the measured power data collected at different time dimensions during the historical operation of the distribution network, which reflects the size of the power load on the grid side.

[0087] A centralized charging station refers to a charging station operation platform that has large-scale power battery charging facilities, can provide charging services for multiple power batteries in a centralized manner, and can participate in the dispatch and control of the power distribution network.

[0088] In this embodiment, the power battery refers to the energy storage type power battery that participates in the peak-valley regulation of the distribution network in the centralized charging station, which can realize the storage and release of electrical energy.

[0089] Parameter data refers to the inherent technical parameters and operational constraints of the power battery itself.

[0090] In this step, the original measured load power data on the distribution network side and the full parameter data of the power batteries involved in the regulation on the charging station side can be collected through the distribution network monitoring system and the charging station power battery management system, respectively.

[0091] Specifically, raw load power data for the distribution network can be collected through devices such as smart meters and load monitoring terminals within the network. The collection frequency should be consistent with the grid dispatch frequency. The collected data includes active power, reactive power, voltage, and frequency at each node, covering the main load nodes of the distribution network to ensure comprehensiveness and accuracy. This raw data needs to be preprocessed to remove outliers and fill in missing data using appropriate methods to ensure data integrity. The parameter data of the power batteries in centralized charging stations mainly includes two categories: first, inherent technical parameters, such as the rated capacity, rated voltage, charge / discharge rate, internal resistance, and cycle life of the power battery; second, operational constraint parameters, such as the upper and lower limits of state of charge (SOC), maximum charge / discharge power, and charge / discharge efficiency. This parameter data can be collected in real time through the battery management system (BMS) of the charging station and updated periodically to ensure that the parameter data reflects the real-time operating status of the power battery. In addition, historical charge / discharge data of the power battery also needs to be collected, including charge / discharge power and SOC changes at different times.

[0092] S2: Based on the original load power data, the predicted load power and daily average load for each time period are calculated.

[0093] It should be noted that the load power for each time period refers to the load power value of the distribution network within each consecutive time interval after dividing a single day into several time intervals. It should also be noted that grid dispatching is performed every 15 minutes, and to facilitate unified grid dispatching, the charging and discharging strategies of centralized charging stations are also changed every 15 minutes. In this invention, the time length is one day; therefore, for each power battery, the charging and discharging power for 96 time periods needs to be known, meaning that a single day can be divided into 96 time periods.

[0094] Daily average load refers to the arithmetic average of the load power of the distribution network during all time periods on a single day.

[0095] In this step, based on the original load power data, the predicted load power of the distribution network for each time period of the day is calculated; then, the predicted load power for all time periods is summed and averaged using the arithmetic mean method to obtain the daily average load of the distribution network.

[0096] Load power forecasting employs a historical data-based average forecasting method. This method is characterized by its simplicity, stability, and real-time performance, meeting the needs of real-time power grid control. The specific forecasting process is as follows: First, the raw load power data for the corresponding recent time period is extracted. After removing outliers, the average value is calculated and used as the forecasted load power for the corresponding time period on a future day. This forecasting method effectively avoids the impact of short-term load fluctuations because, in the short term, the power load of the distribution network usually exhibits strong regularity. Using the recent average value effectively smooths out load fluctuations caused by individual extreme situations, ensuring that the forecast accuracy meets the requirements of practical applications. The calculation of the daily average load involves adding the forecasted load power for all time periods of the day and dividing by the total number of time periods to obtain the daily average load value. This value serves as the core basis for determining the peak and valley periods of the distribution network—when the forecasted load power for a certain time period is higher than the daily average load, it is determined to be a peak period; when it is lower than the daily average load, it is determined to be a valley period; and when it is close to the daily average load, it is determined to be a flat period.

[0097] S3: Based on the load power and daily average load for each time period, the direction of power exchange between the power battery and the distribution network is preset for each time period in order to eliminate the correlation between the power batteries.

[0098] It should be noted that the direction of power exchange between the power battery and the distribution network refers to the direction of power transmission between the power battery and the distribution network, which includes two states: charging (the power battery absorbs power from the distribution network, and power flows from the distribution network to the battery) and discharging (the power battery releases power to the distribution network, and power flows from the battery to the distribution network).

[0099] The correlation between power batteries refers to the coupling of their charging and discharging behaviors when multiple power batteries participate in power grid regulation.

[0100] Eliminating correlation means decoupling the charging and discharging behaviors of multiple power batteries, so that the charging and discharging power optimization of a single power battery is not affected by the behavior of other power batteries, thus achieving independent optimization.

[0101] In this step, the power exchange direction is preset based on the load power and daily average load for each time period. For example, if the predicted load power of the distribution network is higher than the daily average load (peak period), the power exchange direction between all power batteries and the distribution network is preset to discharge during that time period; if the predicted load power of the distribution network is lower than the daily average load (valley period), the power exchange direction between all power batteries and the distribution network is preset to charge during that time period. By using this unified direction preset that is tied to the peak and valley characteristics of the distribution network, the disordered coupling of the charging and discharging behaviors of multiple power batteries is avoided, mutual interference between power batteries is eliminated, and the optimization tasks of each power battery are decoupled.

[0102] Specifically, the core logic of power direction preset is "peak-segment discharge, valley-segment charging," ensuring that the charging and discharging behavior of the power battery is consistent with the grid load demand, maximizing the peak-valley regulation effect. In practice, the predicted load power for all time periods of the day is first compared with the daily average load to determine the peak-valley attributes for each time period: For peak periods, the power exchange direction of all power batteries is preset to discharge, at which time the power batteries release electrical energy to the grid, relieving grid load pressure; for valley periods, the power exchange direction of all power batteries is preset to charge, at which time the power batteries absorb excess electrical energy from the grid, realizing energy storage; for flat periods, the direction can be preset according to the actual grid demand, usually preset to charging, to reserve electrical energy to cope with sudden loads. This unified direction preset can effectively eliminate the correlation between power batteries—because all power batteries have the same charging and discharging direction in the same time period, the charging and discharging power optimization of a single power battery does not need to consider the behavior of other power batteries, but only needs to meet its own parameter constraints and distribution network regulation objectives, thus decomposing the high-dimensional optimization problem of multiple batteries into a low-dimensional optimization problem of a single battery.

[0103] S4: After eliminating correlation, the charging and discharging power optimization task for each power battery is treated as a low-dimensional optimization problem. Combined with parameter data, an improved particle swarm optimization algorithm is used to solve the problem, thereby obtaining the charging and discharging power strategy for each power battery at each time period. In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is optimized separately using the particle swarm optimization algorithm, and finally all optimal solutions are integrated to obtain the global optimal solution.

[0104] It should be noted that the charge and discharge power optimization task refers to determining the specific values ​​of the charge and discharge power of a single power battery in each time period under a preset power exchange direction, so that the value can meet both the parameter constraints of the power battery itself and the peak and valley regulation target of the distribution network.

[0105] Low-dimensional optimization problems refer to the optimization problems of charging and discharging power of a single power battery after eliminating correlations. The optimization dimension is only related to the number of time periods, which is much lower than the high-dimensional problems of joint optimization of multiple power batteries, and the difficulty of solving them is greatly reduced.

[0106] In this embodiment, the improved particle swarm optimization algorithm refers to improving the traditional particle swarm optimization algorithm to perform independent optimization for each dimension. That is, it abandons the traditional PSO method of performing overall optimization for high-dimensional problems and instead performs PSO optimization independently for multiple low-dimensional problems.

[0107] Particle swarm optimization refers to a stochastic optimization algorithm based on swarm intelligence. It maps the solution of an optimization problem to particles and uses the velocity and position of the particles in the solution space to iteratively update the solution and find the optimal solution that satisfies the optimization objective and constraints.

[0108] The global optimal solution refers to the set of charging and discharging power strategies for multiple power batteries at different times after integrating the low-dimensional optimal solutions of all individual power batteries. It is the optimal solution that can achieve the overall goal of charging stations participating in the peak-valley regulation of the distribution network.

[0109] In this step, the charging and discharging power optimization task of each power battery after eliminating correlation can be defined as an independent low-dimensional optimization problem. Then, the collected power battery parameter data is used as a constraint condition, and the optimization objective function is set in combination with the goal of distribution network peak and valley regulation (such as smoothing peak and valley loads and reducing regulation costs). The particle swarm optimization algorithm is used to independently optimize each low-dimensional optimization problem to obtain the optimal charging and discharging power strategy of a single power battery in each time period that satisfies the constraints and objectives. Finally, the low-dimensional optimal solutions of all power batteries are integrated to form a globally optimal charging and discharging power strategy for multiple power batteries to participate in distribution network peak and valley regulation.

[0110] Furthermore, for a complex problem that is difficult to solve directly, the best approach is to break it down into multiple simpler, easier-to-solve problems. This embodiment draws on this idea and employs an improved PSO algorithm (such as Improved PSO based on Sub-vector). The basic idea is to divide the high-dimensional vector to be optimized into multiple low-dimensional sub-vectors, and then use the PSO algorithm to optimize each sub-vector, ultimately obtaining the optimal solution for the high-dimensional vector.

[0111] One problem with this improved algorithm is whether the high-dimensional optimization problem can be decomposed into multiple low-dimensional problems for optimization and solution—that is, a problem of separability. Whether a high-dimensional problem can be decomposed into multiple low-dimensional problems for optimization and solution mainly depends on the correlation between the sub-vectors. If the sub-vectors are completely uncorrelated, then the high-dimensional optimization problem can be decomposed into multiple low-dimensional optimization problems for optimization and solution; if the sub-vectors are correlated, then it cannot be decomposed into multiple low-dimensional problems for optimization and solution.

[0112] In this embodiment, for a single power battery, its dimensions (e.g., 96 dimensions) are correlated because the power exchanged with the distribution network during time period i is related to the previous i-1 time periods. For the entire centralized charging station, the power batteries (m boxes) participating in the peak-valley regulation of the distribution network are also correlated because the power batteries participating in the peak-valley regulation of the distribution network must satisfy the charging and discharging direction constraints. Therefore, all dimensions (96×m dimensions) of the centralized charging station's participation in the peak-valley regulation of the distribution network are correlated, and it cannot be divided into multiple low-dimensional problems for optimization unless some method is used to make the sub-vectors uncorrelated.

[0113] While the correlations between individual power batteries cannot be eliminated, the correlations between power batteries themselves can be eliminated. The correlation between power batteries arises because the direction of power exchange between the power batteries and the distribution network must be consistent at any given time. If the direction of power exchange between the power batteries and the distribution network is predetermined for each time period, the correlation between the power batteries disappears. This allows the high-dimensional optimization problem of centralized charging stations participating in distribution network peak-valley regulation to be solved by dividing it into multiple low-dimensional optimization problems. In this embodiment, the direction setting can be:

[0114] (3)

[0115] That is, when the load power P during a certain period of time Li Greater than the average daily load power P aver When the power battery is in a state of discharge, it discharges; otherwise, it charges. While this method cannot obtain the theoretically optimal solution, the optimal solution obtained using this method is much better than that obtained using the PSO algorithm. Furthermore, the optimal solution obtained using this method significantly reduces load fluctuations in the distribution network, fully meeting the requirements of practical applications. It should be further noted that the core purpose of this directional setting is to eliminate the correlation between power batteries, laying the foundation for decomposing high-dimensional problems. Although this setting cannot obtain the theoretically absolute optimal solution, it can significantly reduce the difficulty of the solution and improve the solution efficiency while ensuring the control effect, making it more in line with the needs of practical engineering applications. Extensive experimental verification shows that after adopting this directional setting, the peak-valley difference suppression effect of the distribution network is significantly improved, far superior to the traditional PSO algorithm, and the solution time is greatly shortened, meeting the needs of real-time grid control. In addition, this directional setting can effectively avoid the problem of chaotic charging and discharging directions of power batteries at the same time, reduce secondary energy conversion losses, improve the economic efficiency of control, extend the cycle life of power batteries, and further reduce the operating costs of charging stations.

[0116] This embodiment provides an optimization method for charging stations participating in distribution network peak-valley regulation based on an improved PSO (Power Swarm Optimization) algorithm. This method addresses the challenges of solving high-dimensional joint optimization problems caused by the correlation of charging and discharging behaviors between batteries when multiple power batteries participate in distribution network peak-valley regulation at traditional charging stations. It also addresses the issues of premature convergence and low optimization efficiency inherent in traditional PSO algorithms. The method proposes a hierarchical optimization approach of first decoupling and then performing multidimensional optimization. Firstly, by combining the quantitative comparison results of predicted load power and daily average load in different time periods of the distribution network, the power exchange direction between the power batteries and the distribution network is preset. This eliminates the correlation of charging and discharging behaviors between multiple power batteries from a regulatory logic perspective, thus decomposing the originally complex multi-battery joint high-dimensional charging and discharging power optimization problem into a more manageable solution. The solution is derived from multiple independent, lower-dimensional optimization problems involving single batteries. An improved particle swarm optimization (PSO) algorithm is then developed, abandoning the traditional approach of optimizing the entire system for high-dimensional problems. Instead, the PSO algorithm is executed independently for each low-dimensional optimization problem. Finally, the global optimum is obtained by integrating all the optimal low-dimensional solutions. This improvement significantly reduces the algorithm's complexity, improves optimization efficiency and accuracy, and effectively avoids the problem of PSO getting trapped in local optima in high-dimensional optimization scenarios. This allows the optimized charging and discharging power strategy to efficiently smooth out peak-valley load fluctuations in the distribution network, significantly improving the practical engineering application effect of centralized charging stations in distribution network peak-valley regulation. Furthermore, through centralized control and optimized scheduling strategies, rapid response and regulation of the grid load are achieved. Especially when grid load fluctuations are severe, the battery discharge state of the charging station can be adjusted in a timely manner, thereby ensuring the stable operation of the grid.

[0117] In one embodiment of the present invention, the predicted load power for each time period is the average value of the load power of the corresponding time period of the distribution network in the previous 7 days; the daily average load is the arithmetic average of the predicted load power of 96 time periods on that day.

[0118] Specifically, the predicted load power P for period i Li Equal to the average load power of the corresponding period in the previous 7 days, daily average load P aver The calculation formula is as follows:

[0119] (4)

[0120] It should be noted that the predicted load power P for time period i Li Using the average load power of the corresponding period in the previous 7 days is not very accurate (in fact, no processing method can be absolutely accurate; as long as it is a prediction, there will definitely be errors). However, it can basically meet the requirements of practical applications. After all, in the short term, the electricity load generally will not change much, and taking the average load power of the previous 7 days can greatly reduce the impact of individual extreme cases.

[0121] like Figure 2As shown, in one embodiment of the present invention, the improved particle swarm optimization algorithm employs the Improved PSO based on Sub-vector algorithm, and the solution process includes:

[0122] Step 1: Map the power battery corresponding to each low-dimensional optimization problem to a particle and initialize all particle indices.

[0123] Each independent low-dimensional power battery charging and discharging power optimization problem is mapped to a single particle in the particle swarm optimization algorithm. Then, the distribution network load prediction data, power battery rated capacity / charge and discharge rate / state of charge (SOC) upper and lower limits and other constraint parameters are read, as well as the inertia weight, cognitive factor, social factor and other control parameters of the particle swarm optimization algorithm, and the particle index variable is initialized (e.g., I=1).

[0124] Step 2: Initialize the position and velocity of each particle, and verify and correct the position and velocity of each particle according to the preset constraints; the position coordinates of the particles are the charging and discharging power of the power battery at each time period.

[0125] For a particle with index I, initialize its position vector (dimension equal to the number of time periods, each component corresponding to the charging and discharging power of the power battery in that time period) and velocity vector (dimension consistent with the position vector, initial value usually set to a small random number); then, in combination with preset conditions such as the upper and lower limits of the charging and discharging power of the power battery and the SOC change rate constraint, perform compliance correction on the initialized position and velocity to ensure that the initial state of the particle is within the feasible solution space and avoid subsequent iterations from falling into the invalid solution region.

[0126] For example, when using the IPSOS algorithm to solve the problem of centralized charging stations participating in the peak-valley regulation of the distribution network, the power battery is the particle in the algorithm. The charging and discharging power of each power battery in the centralized charging station at each time period is used as the position coordinate of the particle, as shown in the following formula:

[0127] (5)

[0128] Step 3: For each particle, execute the particle swarm optimization algorithm based on the preset objective function and the particle's iterative update formula to iteratively update the particle's position and velocity.

[0129] First, the fitness value of the current particle is calculated based on a preset fitness function (such as a weighted combination function aimed at minimizing the peak-to-valley difference in the distribution network and minimizing control costs), thereby evaluating the merits of its corresponding charging and discharging strategy; then, the individual optimal position (p) of the particle is updated. best (Update if current fitness is better than the historical best) and the population optimal position (g) best(If the current particle fitness is better than the global historical best, then update it); then update the particle's velocity and position using the velocity-position iteration formula of the particle swarm algorithm; if the updated position exceeds the charging and discharging power constraint limit, then correct the position and velocity again, and recalculate the fitness value to complete one iteration closed loop.

[0130] Step 4: Based on the particle index, traverse all particles sequentially, repeat the iterative optimization process of the particle swarm optimization algorithm for each particle, until the termination condition is met and the individual optimal position of each particle is obtained.

[0131] Based on the particle index, the iteration process of step 3 is repeated for each particle in turn: if the current particle I meets the termination condition (such as reaching the maximum number of iterations or the fitness change rate being less than the convergence threshold), its individual optimal position (i.e., the optimal charging and discharging power vector of the corresponding power battery) is output and saved; if the termination condition is not met, the iteration continues; after the current particle completes the iteration, the particle index I is incremented (I=I+1), and the process is switched to the next particle to repeat the optimization process until all particles (index I equals the total number of particles) have completed the iteration and output their respective individual optimal positions.

[0132] Step 5: Integrate the optimal individual positions of all particles to obtain the global charging and discharging power strategy.

[0133] The optimal individual positions (p) of all particles best The dimensions of the low-dimensional vectors (corresponding to the charging and discharging power of a single power battery during a given time period) are concatenated to form a global high-dimensional solution matrix, which is the set of charging and discharging power strategies of all power batteries involved in the regulation at each time period.

[0134] The improved particle swarm optimization (PSO) algorithm proposed in this embodiment transforms a high-dimensional problem into multiple low-dimensional subproblems through subvector decomposition. This effectively avoids the tendency of PSO to get trapped in local optima in high-dimensional problems, improving global search capabilities and enabling the discovery of better solutions. This improved PSO algorithm based on subvector decomposition allows the invention to efficiently solve high-dimensional optimization problems, avoiding the performance degradation of traditional PSO as dimensionality increases. It provides a more efficient solution for the optimization participation of centralized charging stations in peak-valley regulation of power distribution networks.

[0135] In a further embodiment of the present invention, the iterative update formula for particles is:

[0136] (6)

[0137] (7)

[0138] In the formula, I is the particle number, that is, the serial number of the power battery, I = 1, 2, ..., m; and These are the velocity vectors of the i-th particle in the k-th generation and the (k+1)-th generation, respectively. and These are the position vectors of the i-th particle in the k-th and (k+1)-th generations, respectively. and Let be the individual optimal position vector and the group optimal position vector of the i-th particle in the k-th generation, respectively. and Given two random numbers between 0 and 1 in the k-th generation. For inertial weights, and These are the cognitive coefficient and the social coefficient, respectively.

[0139] In a further embodiment of the present invention, the objective of centralized charging stations participating in distribution network peak-valley regulation is to reduce the load peak-valley difference, minimize load fluctuations, and achieve the function of load smoothing. The objective function for minimizing load fluctuations (i.e., the objective function of the improved particle swarm optimization algorithm) is as follows:

[0140] (8)

[0141] In the formula, Let be the power grid load fluctuation function, which is the objective function of this optimization problem; i is the time period number; The predicted load power for time period i; 1 is the daily average load; j is the serial number of the power battery in the centralized charging station; m is the number of power batteries in the centralized charging station that participate in the peak-valley regulation of the distribution network. The size of m is determined by the operator of the centralized charging station according to the specific situation. The principle is to ensure that as many power batteries as possible participate in the peak-valley regulation of the distribution network while ensuring the battery swapping needs of electric vehicles. Let be the power exchanged between the j-th power battery and the distribution network during the predicted time period i.

[0142] In a further embodiment of the present invention, the constraints on centralized charging stations participating in peak-valley regulation of the power distribution network mainly include three aspects: constraints on the charging and discharging direction of the power battery; constraints on the charging and discharging power of the power battery; and constraints on the state of charge (SOC) of the power battery. In addition, the execution of the improved PSO algorithm also needs to meet the constraints on the number of iterations and convergence conditions. Therefore, the preset constraints may include:

[0143] (1) Charging and discharging direction constraints:

[0144] (9)

[0145] In the formula, and Let be the power exchanged between the j1-th and j2-th power batteries and the distribution network during the predicted time period i, respectively, and sgn(·) be the sign function; i = 1, 2, ..., 96; j1 = 1, 2, ..., m; j2 = 1, 2, ..., m;

[0146] (10)

[0147] In the formula, P represents the power exchanged between the power battery and the distribution network. sgn(P) = 1 indicates that the power battery is discharging, sgn(P) = 0 indicates that the power battery is neither discharging nor charging, and sgn(P) = -1 indicates that the power battery is charging. Equation (10) ensures that within the same time period, there will not be a situation where some power batteries are charging while others are discharging, thus avoiding the problem of secondary energy conversion and ensuring the economic efficiency of the power battery's participation in the peak-valley regulation of the distribution network. By considering the directional constraints of electric vehicle charging and discharging, the simultaneous charging and discharging of electric vehicles within the same time period is avoided, reducing energy conversion losses and improving the efficiency and economy of grid regulation.

[0148] (2) SOC constraint:

[0149] The lifespan of a power battery is not only related to its charging and discharging power, but also to its state of charge (SOC). Overcharging or over-discharging a power battery will significantly shorten its lifespan. Therefore, when power batteries participate in peak-valley regulation of the power distribution network, they should meet the following constraints:

[0150] (11)

[0151] In the formula, Let j be the state of charge of the j-th power battery during the predicted time period i. and These are the lower and upper limits of the state of charge, respectively;

[0152] (3) Charge and discharge power limits:

[0153] If the charging and discharging power of the power battery exceeds the allowable value, its lifespan will be significantly shortened, and it may even cause an accident. Therefore, the power exchanged between the power battery and the power distribution network must meet the following constraints:

[0154] (12)

[0155] In the formula, Let j be the power exchanged between the j-th power battery and the distribution network during the predicted time period i. and These represent the minimum power (a negative value indicates the power battery is charging) and the maximum power (a positive value indicates the power battery is discharging) exchanged between the power battery and the power distribution network, respectively.

[0156] (4) Iteration number and convergence constraints:

[0157] (13)

[0158] (14)

[0159] In the formula, For the number of iterations, This represents the maximum number of iterations. and These are the global optimal solutions for the (t+1)th and tth iterations, respectively. A threshold is set. When the change in the global optimal solution g(t) over several consecutive generations is less than the set threshold... Stop iterating when the time is right.

[0160] The aforementioned constraints primarily involve physical and operational conditions such as the directionality of battery charging and discharging, SOC range, battery power limitations, number of iterations, and convergence conditions. The setting of the objective function and constraints provides a comprehensive optimization framework for improving the particle swarm optimization algorithm. Furthermore, the optimization strategies employed in the aforementioned embodiments can enhance the feasibility of solutions and global optimal performance.

[0161] In a further embodiment of the present invention, after iteratively updating the position and velocity of the particles, if the particle position does not meet the preset constraints, processing is required. There are generally two methods for processing constraints: using a penalty function or direct modification. This embodiment uses the direct modification method. This embodiment processes three constraints:

[0162] (1) Constraints on the charging and discharging direction of the power battery.

[0163] If the generated new position does not satisfy the charging / discharging direction constraint, i.e., fsgn2(P) Li P ij If <0, then modify the charging and discharging power of the power battery for the corresponding time period (i.e., modify the position of the particles) according to the following formula:

[0164] (15)

[0165] (2) Power battery charging and discharging power constraints.

[0166] When generating a new position in each iteration, it is determined whether the constraint condition is met. If not, the charging and discharging power of the power battery in the corresponding time period is modified according to the following formula (that is, the position of the particle is modified):

[0167] (16)

[0168] (3) Power battery SOC constraint.

[0169] (17)

[0170] In the formula, The duration of a single time period is 15 minutes in this embodiment; Let be the actual usable capacity of the j-th power battery, representing the total energy released when a fully charged power battery discharges to the minimum allowable discharge voltage under certain conditions. Let j be the state of charge of the j-th power battery during the predicted time period i-1. Let be the power exchanged between the j-th power battery and the distribution network during the predicted time period i-1.

[0171] Based on the design of the foregoing embodiments, it is evident that, compared to the prior art, the advantages of the present invention are as follows:

[0172] 1. Rapid Response to Grid Load Fluctuations. This invention achieves rapid response to grid load fluctuations by utilizing centralized charging stations in grid load regulation. During sudden increases in grid load, centralized charging stations can quickly adjust the charging and discharging states of batteries, rapidly supplying or absorbing energy from the grid, effectively alleviating grid load pressure and ensuring grid stability. Compared to traditional orderly charging strategies, this invention can adjust grid load in a shorter time, greatly improving grid dispatch efficiency and flexibility.

[0173] 2. Efficient optimization algorithm, avoiding local optima. This invention employs an improved particle swarm optimization (PSO) algorithm based on subvector decomposition, which can efficiently solve high-dimensional optimization problems and significantly reduce the probability of the algorithm getting trapped in local optima in high-dimensional space. This optimization method effectively improves the global search capability of the optimization process by decomposing the high-dimensional problem into multiple low-dimensional subproblems, ensuring that a global optimum can be found and avoiding the performance degradation problem commonly seen in traditional algorithms in high-dimensional cases.

[0174] 3. Improve grid regulation efficiency and reduce energy loss. This invention introduces a charge / discharge directionality constraint to ensure that all electric vehicles participating in regulation maintain the same charge / discharge state within the same time period, avoiding conflicts between charging and discharging. This innovative measure effectively prevents secondary energy conversion, reduces energy loss, and improves the overall efficiency of grid load regulation. Furthermore, this constraint optimizes battery charging and discharging strategies, improving the system's economy and operational stability.

[0175] 4. High efficiency in large-scale systems. As the number of batteries in centralized charging stations increases, the dimensionality of the optimization problem grows exponentially, making it difficult for traditional optimization methods to efficiently solve problems in large-scale systems. The improved particle swarm optimization algorithm of this invention maintains high efficiency even with a rapidly increasing dimensionality, enabling its widespread application in large-scale electric vehicle charging stations and power grid systems. This advantage makes this invention suitable not only for small-scale systems but also for providing reliable solutions in large-scale power grid control.

[0176] Based on the above advantages, this invention can significantly improve the flexibility of power grid regulation, optimize the solution efficiency of the algorithm, and reduce the system operating cost. It provides a new and efficient solution for centralized charging stations to participate in power grid peak-valley regulation and has good market application prospects.

[0177] Based on the same inventive concept, this application also provides an improved PSO-based charging station participation in distribution network peak-valley regulation optimization device for implementing the above-mentioned improved PSO-based charging station participation in distribution network peak-valley regulation optimization method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in the embodiments of the improved PSO-based charging station participation in distribution network peak-valley regulation optimization device provided below can be found in the limitations of the improved PSO-based charging station participation in distribution network peak-valley regulation optimization method described above, and will not be repeated here.

[0178] Please see Figure 3 This invention also provides a charging station participation in distribution network peak-valley regulation optimization device based on improved PSO, comprising:

[0179] The data acquisition module is used to acquire raw load power data of the distribution network and parameter data of the power batteries involved in regulation in the centralized charging station.

[0180] The load parameter calculation module is used to calculate the predicted load power and daily average load for each time period based on the original load power data.

[0181] The power direction preset module is used to preset the direction of power exchange between the power battery and the distribution network in each time period based on the load power and daily average load in each time period, so as to eliminate the correlation between the power batteries.

[0182] The optimization solution module is used to treat the charging and discharging power optimization task corresponding to each power battery after eliminating correlation as a low-dimensional optimization problem, and combine it with parameter data to solve it using an improved particle swarm algorithm to obtain the charging and discharging power strategy of each power battery at each time period.

[0183] In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is solved by using the particle swarm optimization algorithm separately, and finally all the optimal solutions are integrated to obtain the global optimal solution.

[0184] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0185] Reference Figure 4 The present invention also provides a computer device, including: a memory and a processor, and a computer program stored in the memory. When the computer program is executed on the processor, it implements the method for optimizing the peak-valley regulation of the distribution network based on the improved PSO as described in any of the above methods.

[0186] The computer device may be a desktop computer, laptop, handheld computer, or cloud server, etc. This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 4 The examples of computer devices are merely examples and do not constitute a limitation on computer devices. They may include more or fewer components than shown in the illustration, or combinations of certain components, or different components. For example, they may also include input / output devices, network access devices, etc.

[0187] The processor referred to can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0188] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard drive or RAM. In other embodiments, the memory may be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory may include both internal and external storage units of the computer device. The memory is used to store the operating system, applications, boot loader, data, and other programs, such as the program code of the computer program. The memory can also be used to temporarily store data that has been output or will be output.

[0189] This invention also provides a computer-readable storage medium storing a computer program thereon. When the computer program is run by a processor, it implements the method for optimizing the peak-valley regulation of a charging station based on an improved PSO, as described in any of the above methods.

[0190] In this embodiment, if the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0191] This invention provides a computer program product, including a computer program that, when executed by a processor, implements the method described above for optimizing the peak-valley regulation of charging stations participating in distribution networks based on improved PSO.

[0192] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0193] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0194] In the embodiments disclosed in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0195] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the participation of charging stations in distribution network peak-valley regulation based on improved PSO, characterized in that, Includes the following steps: Obtain raw data on the load power of the distribution network, as well as parameter data of the power batteries involved in regulation in the centralized charging station; Based on the original load power data, the predicted load power and daily average load for each time period are calculated. Based on the load power and daily average load for each time period, the direction of power exchange between the power battery and the distribution network for each time period is preset to eliminate the correlation between the power batteries. After eliminating correlation, the charging and discharging power optimization task for each power battery is treated as a low-dimensional optimization problem. Combined with the parameter data, an improved particle swarm optimization algorithm is used to solve the problem, thereby obtaining the charging and discharging power strategy for each power battery at each time period. In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is solved by using the particle swarm optimization algorithm separately, and finally all the optimal solutions are integrated to obtain the global optimal solution.

2. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 1, characterized in that, The predicted load power for each time period is the average load power of the distribution network for the corresponding time period over the previous 7 days; the daily average load is the arithmetic mean of the predicted load power for the 96 time periods on that day.

3. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 1, characterized in that, The improved particle swarm optimization algorithm adopts a sub-vector-based improved particle swarm optimization algorithm, and the solution process includes: Map each power battery corresponding to the low-dimensional optimization problem to a particle and initialize all particle indices. Initialize the position and velocity of each particle, and verify and correct the position and velocity of each particle according to the preset constraints; the position coordinates of the particles are the charging and discharging power of the power battery at each time period. For each particle, a particle swarm optimization algorithm is executed based on a preset objective function and the particle's iterative update formula to iteratively update the particle's position and velocity; Based on the particle number, all particles are traversed sequentially, and the iterative optimization process of the particle swarm optimization algorithm for each particle is repeated until the termination condition is met, at which point the individual optimal position of each particle is obtained. By integrating the optimal individual positions of all particles, a global charging and discharging power strategy is obtained.

4. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 3, characterized in that, The objective of the improved particle swarm optimization algorithm is to minimize grid load fluctuations, and the corresponding objective function is: In the formula, Let be the objective function. To predict the load power for period i, The daily average load, Let j be the power exchanged between the j-th power battery and the distribution network during the predicted time period i. The number of power batteries involved in regulation.

5. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 3, characterized in that, The iterative update formula for the particles is: In the formula, and These are the velocity vectors of the i-th particle in the k-th generation and the (k+1)-th generation, respectively. and These are the position vectors of the i-th particle in the k-th and (k+1)-th generations, respectively. and Let be the individual optimal position vector and the group optimal position vector of the i-th particle in the k-th generation, respectively. and Given two random numbers between 0 and 1 in the k-th generation. For inertial weights, and All of these are learning factors.

6. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 3, characterized in that, The preset constraints include: Charge / discharge direction constraints: In the formula, and sgn(·) represents the power exchanged between the j1-th and j2-th power batteries and the distribution network during the predicted time period i, respectively, and sgn(·) is the sign function. In the formula, P is the power exchanged between the power battery and the power distribution network, sgn(P) = 1 indicates that the power battery is discharging, sgn(P) = 0 indicates that the power battery is neither discharging nor charging, and sgn(P) = -1 indicates that the power battery is charging. SOC constraints: In the formula, Let j be the state of charge of the j-th power battery during the predicted time period i. and These are the lower and upper limits of the state of charge, respectively; Charge and discharge power limits: In the formula, Let j be the power exchanged between the j-th power battery and the distribution network during the predicted time period i. and These are the minimum and maximum power exchanges between the power battery and the power distribution network, respectively. Number of iterations and convergence constraints: In the formula, For the number of iterations, This represents the maximum number of iterations. and These are the global optimal solutions for the (t+1)th and tth iterations, respectively. To set a threshold.

7. The method for optimizing charging station participation in distribution network peak-valley regulation based on improved PSO according to claim 6, characterized in that, After iteratively updating the particle's position and velocity, if the particle position does not meet the preset constraints, a direct modification method is used to adjust the charging and discharging power of the power battery, thereby adjusting the particle position. Specifically, this includes: If the charging / discharging direction constraint is not met, the charging / discharging power of the power battery in the corresponding time period will be modified as follows: If the charging and discharging power limit constraint is not met, the charging and discharging power of the power battery for the corresponding time period will be modified as follows: If the SOC constraint is not met, the charging and discharging power of the power battery for the corresponding time period will be modified as follows: In the formula, The duration of a single time period. Let j be the actual usable capacity of the j-th power battery. Let j be the state of charge of the j-th power battery during the predicted time period i-1. Let be the power exchanged between the j-th power battery and the distribution network during the predicted time period i-1.

8. A charging station participating in distribution network peak-valley regulation optimization device based on improved PSO, characterized in that, include: The data acquisition module is used to acquire raw load power data of the distribution network and parameter data of the power batteries involved in regulation in the centralized charging station. The load parameter calculation module is used to calculate the predicted load power and daily average load for each time period based on the original load power data. The power direction preset module is used to preset the direction of power exchange between the power battery and the distribution network in each time period based on the load power and daily average load in each time period, so as to eliminate the correlation between the power batteries. The optimization solution module is used to treat the charging and discharging power optimization task corresponding to each power battery after eliminating correlation as a low-dimensional optimization problem, and combine the parameter data to solve it using an improved particle swarm algorithm to obtain the charging and discharging power strategy of each power battery in each time period. In the improved particle swarm optimization algorithm, each low-dimensional optimization problem is solved by using the particle swarm optimization algorithm separately, and finally all the optimal solutions are integrated to obtain the global optimal solution.

9. A computer device, characterized in that, The device includes a processor and a memory: The memory is used to store computer programs and send the instructions of the computer programs to the processor; The processor executes, according to the instructions of the computer program, a method for optimizing the peak-valley regulation of a charging station based on an improved PSO as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for optimizing the peak-valley regulation of a charging station based on an improved PSO as described in any one of claims 1-7.

Images