Energy storage system optimization method, device and equipment for low-voltage area and storage medium

Abstract

This invention relates to the field of power systems and discloses a method, apparatus, equipment, and storage medium for optimizing an energy storage system in a low-voltage distribution area. The method acquires current-time power load data, renewable energy generation data, meteorological data, and three-phase power grid data for the low-voltage distribution area. Using a pre-trained prediction model, it predicts the load and generation data for the next time point in the low-voltage distribution area based on the power load, generation, and meteorological data. Using a three-phase linear power grid model, it determines the balance parameters between the phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data. Finally, using an improved genetic algorithm, it optimizes the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and generation data. This invention can solve the problem of inflexible matching between existing energy storage system configuration and operation strategies, leading to low system operating efficiency and resource waste.

Description

Technical Field

[0001] This invention relates to the field of power systems, and in particular to a method, apparatus, equipment and storage medium for optimizing energy storage systems in low-voltage distribution areas. Background Technology

[0002] With the rapid development of renewable energy and the increasing complexity of power loads in low-voltage distribution areas, traditional power dispatching and energy storage management methods are facing severe challenges. Low-voltage distribution areas play a crucial role in power supply; however, due to issues such as large load fluctuations and unstable renewable energy output, the stability and economy of power supply often fail to meet actual needs.

[0003] Existing low-voltage distribution area energy storage systems generally adopt a static configuration approach and mostly employ simple rule-based scheduling strategies. Due to the lack of dynamic optimization and intelligent scheduling, the energy utilization efficiency of battery storage is low, and the configuration of energy storage systems cannot flexibly respond to load fluctuations and energy changes. This results in energy storage devices potentially being insufficient during load peaks or storing too much energy during off-peak periods, failing to effectively reduce the peak-to-valley difference in the power grid and affecting the stability of the power system.

[0004] Therefore, there is an urgent need for a low-voltage distribution area energy storage system that can dynamically optimize configuration based on real-time data, and improve the configuration accuracy and operating efficiency of the energy storage system through data-driven methods, thereby reducing unnecessary resource waste. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and storage medium for optimizing energy storage systems in low-voltage distribution areas, in order to solve the problem that existing low-voltage distribution area energy storage systems are inflexible in configuration, resulting in low system operating efficiency and resource waste.

[0006] The first aspect of this invention provides a method for optimizing an energy storage system in a low-voltage distribution area, comprising:

[0007] Acquire the working data of the low-voltage distribution area for the current time period, including power load data, renewable energy generation data, meteorological data, and three-phase power grid data;

[0008] Using a pre-trained prediction model, the load data and power generation data for the predicted time period in the low-voltage distribution area are predicted based on the power load data, the renewable energy generation data, and the meteorological data.

[0009] Using a three-phase linear power grid model, the balance parameters between phases in the three-phase power grid are determined based on the electrical parameters of each phase in the three-phase power grid data. Then, using an improved genetic algorithm, the three-phase energy storage configuration parameters in the energy storage system are optimized based on the balance parameters, predicted load data, and power generation data.

[0010] In one feasible implementation, the prediction model is a load prediction model based on Long Short-Term Memory (LSTM) networks and a power generation prediction model based on Support Vector Regression (SVR). The step of using the pre-trained prediction model to predict the load and power generation data of the low-voltage distribution area at the next time point based on the power load data, the renewable energy power generation data, and the meteorological data includes: inputting the power load data into the load prediction model to extract a first time-series feature of the load at each time point in the power load data; calculating the load data for the predicted time period in the low-voltage distribution area based on each of the first time-series features and the meteorological data; inputting the renewable energy power generation data into the power generation prediction model to extract a second time-series feature of the power generation at each time point in the renewable energy power generation data; and calculating the power generation data for the predicted time period in the low-voltage distribution area based on each of the second time-series features and the meteorological data.

[0011] In one feasible implementation, the three-phase linear power grid model is a model based on a three-phase power grid load flow algorithm. The step of using the three-phase linear power grid model to determine the balance parameters between phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data includes: using the three-phase power grid load flow algorithm to calculate the load flow of the three-phase power grid based on the voltage, current, and power factor of each phase in the three-phase power grid data; using the three-phase imbalance method to calculate the load percentage of each phase based on the load flow of the three-phase power grid, and using the load percentage of each phase as the balance parameter between phases in the three-phase power grid.

[0012] In one feasible implementation, the step of using an improved genetic algorithm to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and power generation data includes: obtaining all energy storage configuration schemes of the energy storage system for the current time period, and constructing an initial population based on all the energy storage configuration schemes, where each individual in the initial population is an energy storage configuration scheme, and each energy storage configuration scheme includes the type, capacity, charging and discharging time, and power of the energy storage device; calculating the fitness value of each individual based on the balance parameters, predicted load data, and power generation data; selecting individuals whose fitness values ​​meet a preset fitness threshold for crossover and mutation operations to generate a new population until convergence to the optimal solution, and optimizing the three-phase energy storage configuration parameters in the energy storage system based on the optimal solution, wherein the crossover operation involves selecting multiple individuals whose fitness values ​​meet the preset fitness threshold to exchange genes with each other, and the mutation operation involves performing crossover operations on multiple individuals from different spatial dimensions.

[0013] In one feasible implementation, the fitness value is calculated using the following formula:

[0014] F=

[0015] Where n is a certain prediction time period, For load power, For power generation capacity, For energy storage power, The fitness value of the previous time period.

[0016] In one feasible implementation, after optimizing the three-phase energy storage configuration parameters in the energy storage system using an improved genetic algorithm based on the balance parameters, predicted load data, and power generation data, the method further includes: generating a charging and discharging time sequence table for all devices in the energy storage system based on the optimized three-phase energy storage configuration parameters, predicted load data, and power generation data, wherein the charging and discharging time sequence table includes the charging and discharging time period, power capacity, and priority of the devices; obtaining real-time load data and power generation data for the predicted time period, and using an improved optimization strategy to select several devices from the charging and discharging time sequence table according to the balance parameters for dynamic scheduling planning to obtain a charging and discharging strategy, wherein the improved optimization strategy is a control strategy that adds linear programming to a fuzzy control algorithm; and scheduling each device in the energy storage system to perform charging and discharging operations based on the charging and discharging strategy.

[0017] In one feasible implementation, the step of selecting several devices from the charging / discharging time series table using an improved optimization strategy according to the balance parameters for dynamic scheduling planning to obtain a charging / discharging strategy includes: calculating the data deviation between real-time data and predicted data within the predicted time period, wherein the data deviation includes the difference between real-time load data and predicted load data, and the difference between real-time power generation data and predicted power generation data; constructing a linear programming objective function for each device based on the balance parameters and the data deviation, wherein the linear programming objective function is: ,in, The unit energy cost during the charge / discharge period. Let n be the absolute value of the charging and discharging power of the device, and t be the predicted time period. [0, n]; Find the optimal solution for the linear programming objective function, and select several devices from the charging and discharging timing table based on the optimal solution. Use a fuzzy control algorithm to fuzzify the parameters of the several devices to obtain the charging and discharging strategy.

[0018] A second aspect of the present invention provides a control device for an energy storage system in a low-voltage distribution area, comprising: an acquisition module for acquiring current operating data of the low-voltage distribution area, the operating data including power load data, renewable energy generation data, meteorological data, and three-phase power grid data; a prediction module for using a pre-trained prediction model to predict the load data and generation data of the low-voltage distribution area at the next time point based on the power load data, the renewable energy generation data, and the meteorological data; and an optimization module for using a three-phase linear power grid model to determine the balance parameters between the phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data, and using an improved genetic algorithm to optimize the three-phase energy storage configuration parameters in the energy storage system according to the balance parameters, the predicted load data, and the generation data.

[0019] A third aspect of the present invention provides an electronic device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the electronic device to execute the energy storage system control method for the low-voltage substation described above.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described energy storage system control method for low-voltage substations.

[0021] The technical solution provided by this invention involves acquiring operational data for the current time period in a low-voltage distribution area. This operational data includes power load data, renewable energy generation data, meteorological data, and three-phase power grid data. Using a pre-trained prediction model, the load and generation data for the predicted time period in the low-voltage distribution area are predicted based on the power load data, renewable energy generation data, and meteorological data. A three-phase linear power grid model is used to determine the balance parameters between each phase in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data. An improved genetic algorithm is then used to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, the predicted load data, and the generation data. In this invention, by real-time acquisition and analysis of power load, generation data, meteorological information, and three-phase power grid data in a low-voltage distribution area, and by employing a three-phase power grid model and an improved genetic algorithm, the configuration and operation strategies of the energy storage system are dynamically adjusted to achieve configuration flexibility and improve the system's adaptability, economy, and operational efficiency. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of an embodiment of the energy storage system control method for low-voltage distribution areas in this invention;

[0023] Figure 2This is a schematic diagram of another embodiment of the energy storage system control method for low-voltage distribution areas in this invention;

[0024] Figure 3 This is a schematic diagram of one embodiment of the energy storage system control device for the low-voltage distribution area in this invention;

[0025] Figure 4 This is a schematic diagram of another embodiment of the energy storage system control device for the low-voltage substation in this invention;

[0026] Figure 5 This is a schematic diagram of one embodiment of the electronic device in this invention. Detailed Implementation

[0027] This invention provides a control method, device, equipment, and storage medium for an energy storage system in a low-voltage distribution area. Based on the analysis and prediction of real-time data, the configuration of the energy storage device is optimized using a three-phase linear power grid model and an improved genetic algorithm. This can maximize the satisfaction of the power demand of the low-voltage distribution area and avoid the problems of over-configuration or under-configuration that may exist in traditional static configuration methods.

[0028] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] It is understood that the executing entity of this invention can be an energy storage system optimization device for a low-voltage distribution area, or it can be a terminal, a local monitoring platform, or a server; the specific implementation is not limited here. This embodiment of the invention uses a server as the executing entity as an example for explanation. In the energy storage system of a low-voltage distribution area, there are multiple power generation devices and energy storage devices, which can be connected via a local area network (LAN). The devices report data to the server through the LAN, and the server performs unified scheduling and configuration optimization. This invention achieves low-voltage distribution area energy storage system configuration optimization by applying predictive models, three-phase power grid models, and improved genetic algorithms, thereby ensuring grid load balance and efficient utilization of the battery system.

[0030] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1One embodiment of the energy storage system control method for low-voltage distribution areas in this invention includes:

[0031] 101. Obtain the current time period's working data in the low-voltage distribution area. This working data includes power load data, renewable energy generation data, meteorological data, and three-phase power grid data.

[0032] Understandably, the current extreme operating data of low-voltage distribution areas obtained through smart meters, weather sensors, and power generation monitoring systems should be understood as the time length consisting of the point in time when the resource allocation optimization of the energy storage system is triggered and T time points prior to the triggering point, or the time for each energy storage and dispatch process. This operating data specifically includes, but is not limited to, the following:

[0033] Power load data: This includes the power consumption of each user. The data is collected once every minute. Of course, the accuracy can be optimized according to the actual configuration, and the time interval can be customized to be less than one minute or even shorter.

[0034] Renewable energy generation data: mainly acquiring real-time power generation from photovoltaic and wind power systems;

[0035] Meteorological data, such as temperature, wind speed, and solar intensity, affect load demand and power generation.

[0036] Three-phase power grid data: Obtain power grid parameters such as voltage, current, and power factor for each phase.

[0037] These data are uploaded to the server in real time and preprocessed, including data cleaning, missing value imputation, and normalization, to ensure the accuracy of subsequent predictive analysis.

[0038] 102. Using a pre-trained prediction model, predict the load and power generation data for the predicted time period in the low-voltage distribution area based on power load data, power generation data, and meteorological data.

[0039] It should be noted that the prediction time period can be the next time period after the current time period, or any time period after the current time period. The prediction model is actually a hybrid model containing two network structures, which are used to predict load data and power generation data respectively. Specifically, the prediction model is trained using a large amount of historical load data and historical power generation data respectively.

[0040] Specifically, when using pre-trained prediction models to forecast future load and power generation data, this includes:

[0041] Load data forecasting uses a model derived from a Long Short-Term Memory (LSTM) network to predict future load data over time. The input consists of current electricity load data and corresponding meteorological data. The LSTM model extracts load variation features from the input electricity load data to generate time-series features with long-term dependencies. Based on these time-series features and combined with meteorological data, the output is the load demand for the next time point, thus obtaining the corresponding load data.

[0042] Power Generation Forecast: This forecast uses a Support Vector Regression (SVR) model to predict the power generation of renewable energy sources (such as solar and wind power). The input consists of the current power generation data and the corresponding meteorological data (such as sunlight intensity and wind speed). SVR can regress future power generation based on this data, thus obtaining the power generation data.

[0043] By using the above two-step prediction, we can obtain the predicted load demand and power generation of the low-voltage distribution area in the future, providing key input for the optimization of the energy storage system.

[0044] 103. Using a three-phase linear power grid model, the balance parameters between phases in the three-phase power grid are determined based on the electrical parameters of each phase in the three-phase power grid data. An improved genetic algorithm is then used to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and power generation data.

[0045] This step can be understood as including two steps: load balancing analysis and configuration parameter optimization, wherein:

[0046] When using a three-phase power grid model for load balance analysis, the load flow of the three-phase power grid is first calculated: based on the voltage, current and power factor data of the three-phase power grid, the load flow algorithm is used to calculate the load distribution in the power grid, ensuring that the current and voltage of each phase are within a reasonable range.

[0047] Then, the balance parameters are calculated: the load proportion of each phase is calculated using the three-phase imbalance method to obtain the degree of load imbalance in the three-phase power grid. The calculated balance parameters are used as key reference data for optimizing energy storage configuration.

[0048] When using an improved genetic algorithm to optimize energy storage configuration, the first step is to create an initial population: Construct an initial population where each individual represents an energy storage configuration scheme, including the type, capacity, charging and discharging time period, and power of the energy storage device.

[0049] Then, the fitness is calculated: using the predicted load data, generation data, and balance parameters of the three-phase power grid, the fitness of each energy storage configuration is calculated.

[0050] Next, select individuals: select suitable individuals based on fitness values, perform crossover and mutation operations, and generate a new population.

[0051] Finally, convergence to the optimal solution: After several iterations, convergence to the optimal solution yields the configuration scheme of the energy storage device, including capacity, charging and discharging time, and power allocation.

[0052] In summary, by acquiring real-time data from low-voltage distribution areas and utilizing prediction modules, three-phase linear power grid models, and improved genetic algorithms, the parameters of the energy storage system are dynamically configured. This solves the problems of inflexible and unreasonable configuration of traditional energy storage systems, making the capacity, quantity, and operating mode of energy storage devices more precise. It also minimizes three-phase imbalance in the power grid, improves grid stability, and reduces the investment and operating costs of the energy storage system.

[0053] Please see Figure 2 Another embodiment of the energy storage system control method for low-voltage distribution areas in this invention includes:

[0054] 201. Obtain the working data of the energy storage system in the low-voltage distribution area.

[0055] It should be noted that the working data includes power load data, renewable energy generation data, meteorological data, and three-phase power grid data, among which:

[0056] Electricity load data can be collected in real time from low-voltage distribution areas using smart meters and sensors. This includes the real-time electricity demand of each electricity user unit (such as residential users, commercial users, etc.), and the data collection frequency can be set to 1 minute or less to ensure the timeliness of the data.

[0057] Renewable energy data can be collected using weather sensors and online monitoring platforms to gather real-time and forecasted power generation data from renewable energy sources such as solar photovoltaic and wind power. For example, the power output of solar panels is affected by factors such as sunlight intensity and temperature, while the output of wind turbines is affected by factors such as wind speed and wind direction.

[0058] Meteorological data (such as temperature, wind speed, humidity, and light intensity) can be collected through weather stations or third-party meteorological data platforms. Meteorological data has a significant impact on load demand and renewable energy generation, and is especially important in predicting load fluctuations in low-voltage distribution areas.

[0059] Three-phase power grid data can also be acquired using smart meters and sensors. This three-phase power grid data includes three-phase current and three-phase voltage. The three-phase current and three-phase voltage of the low-voltage distribution area are collected in real time and used to analyze the load distribution and three-phase imbalance of the power grid in the distribution area. The electrical characteristics of each phase of the power grid can be accurately obtained.

[0060] In this embodiment, after collecting the aforementioned data, the process further includes transmitting and storing the collected data. Specifically, the data is transmitted to a server via a wireless or wired network and stored in the server's database. It should be noted that high-speed encryption is used during data transmission to ensure low latency and high reliability, and to prevent the loss of critical data.

[0061] Furthermore, before storage, the data undergoes cleaning and standardization to ensure its quality and consistency. Cleaning primarily involves removing missing and outlier values ​​and filling in missing load data. Standardization can be understood as data normalization, which involves normalizing the collected power load data (e.g., Min-Max normalization) to unify the data range to a standard interval, ensuring data stability and accuracy.

[0062] 202. Forecasting of load data and power generation data.

[0063] Specifically, predictions are made using prediction models, which are load prediction models based on Long Short-Term Memory (LSTM) networks and power generation prediction models based on Support Vector Regression (SVR).

[0064] For load data prediction, the power load data is input into the load prediction model to extract the first time-series feature of the load at each time point in the power load data; based on each of the first time-series features and the meteorological data, the load data for the predicted time period in the low-voltage distribution area is calculated. This LSTM model specifically comprises the following parts: a preprocessing layer, an input layer, a candidate layer, an update layer, and an output layer, wherein:

[0065] The preprocessing layer consists of forget gates, and its calculation formula is as follows: , where f t The output of the forget gate is σ, which is the sigmoid activation function, and W is the output of the forget gate. f It is a weight matrix, which uses empirical values ​​set through testing, b f It is a bias term, h t−1 It is the hidden state of the previous time step, x t This is the current power load data.

[0066] After the current power load data is input into the preprocessing layer, it is processed by the Sigmoid activation function to remove environmental factors from the power load data, and then output to the input layer.

[0067] In the input layer, the Sigmoid activation function is used to perform a secondary activation process on the preprocessed power load data. This secondary activation process is the same as that of the preprocessing layer, except that a weight matrix is ​​added to further optimize the removal of environmental factors.

[0068] After the secondary activation, the power load data passes through the candidate layer and then the update layer to record and update the features before being output to the output layer, thus outputting the first time-series feature.

[0069] The candidate layer here can be understood as candidate memory units, and its formula is:

[0070] ,in, These are candidate memory features.

[0071] The update layer can be understood as updating memory units, and its formula is: , where C t It is the memory feature of the current time step, C t−1 It is a memory feature from the previous time step.

[0072] Based on the above layers, the prediction formula for the load forecasting model is obtained:

[0073] ,in This is the load forecast value for the next time point. This is load data from past times. This is the temperature data at the current time. Here is the humidity data at the current time point, and f is a function of the load forecasting model.

[0074] Extract the time features, lag features, and trend features from the power load data, form a feature matrix (i.e., the first time series feature) from the time features, lag features, and trend features, and input it into the load forecasting model to obtain the load forecast result for a future time period (e.g., 1 minute later).

[0075] For power generation data prediction, the power generation data of the renewable energy source is input into the power generation prediction model to extract the second time-series feature of power generation at each time point in the renewable energy power generation data; based on each of the second time-series features and the meteorological data, the power generation data for the predicted time period in the low-pressure area is calculated; the calculation formula of the power generation prediction model is as follows:

[0076] Where G(x) is the predicted power generation, For kernel functions (such as Gaussian kernels). is a Lagrange multiplier, and b is a bias term.

[0077] It's important to note that kernel functions map input data to a high-dimensional feature space, enabling nonlinear problems to be linearized within this higher-dimensional space. Through this mapping, SVR can solve complex nonlinear regression problems. The role of kernel functions is to transform data from the original space to a high-dimensional space, where classification or regression can be performed linearly without explicit transformation calculations.

[0078] Regarding the selection of the kernel function, the linear relationship of power generation at different time points in the power generation data is first analyzed. In the power generation forecasting of renewable energy (such as solar and wind power), power generation often exhibits nonlinear characteristics, and the relationship between different features (such as meteorological data, historical power generation data, etc.) is complex. Therefore, a kernel function that can capture nonlinear relationships is selected based on this relationship. In this invention, the Gaussian radial basis function (RBF kernel) is selected. Due to its parameterized form and good characteristics, such a kernel function can effectively map the input data to a high-dimensional space, enabling the model to linearize when dealing with nonlinear problems in high-dimensional space.

[0079] After determining the kernel function as the RBF kernel function, it is also necessary to determine the core parameters and other hyperparameters. The core parameters are determined using cross-validation, and the specific steps are as follows:

[0080] Select a parameter range (e.g., from 0.1 to 10 for testing).

[0081] Within this scope, cross-validation is used to train and validate the model, and the prediction error (such as mean squared error, MSE) corresponding to the value of each core parameter is calculated.

[0082] Choose the value that minimizes the verification error as the final core parameter.

[0083] Other hyperparameters include the penalty parameter and the fault tolerance, both of which are determined using cross-validation. Specifically:

[0084] The determination of the penalty parameters includes:

[0085] The range of candidate values ​​selected: Usually, the selected values ​​are within a certain range, such as

[10] −3 10 3 ].

[0086] Divide the training data into training and validation sets: Divide the training data into multiple small subsets, use one subset as the validation set, and use the remaining subsets as the training set.

[0087] Grid search: Train and validate the model within a range of candidate values, and select the value that minimizes the validation error.

[0088] Select the optimal value: Obtain the optimal value through cross-validation.

[0089] Determining the fault tolerance rate includes:

[0090] Selection based on empirical values: Generally, the value range can be between [0.01, 0.1]. It is common to start with a smaller value (such as 0.01) and then adjust it based on the results of cross-validation.

[0091] Cross-validation: Similar to the penalty parameter, cross-validation is used to determine the optimal value. Grid search is also a commonly used selection method.

[0092] The lag characteristics, meteorological characteristics, and periodic characteristics of the renewable energy power generation data are extracted, and the above-mentioned lag characteristics, meteorological characteristics, and periodic characteristics are formed into a feature matrix (i.e., the second time series characteristics), which is then input into the power generation prediction model to obtain the power generation prediction results for a future period of time (e.g., 1 minute later).

[0093] 203. Calculate the balance parameters of a three-phase power grid using a three-phase power grid model.

[0094] By using a three-phase power grid model and real-time power grid data, load flow is calculated, and the current, voltage, and power factor of each phase are obtained. The load proportion of each phase is calculated using the three-phase imbalance method, yielding the balance parameters of the three-phase power grid load, providing a basis for energy storage system optimization.

[0095] In this embodiment, the three-phase power grid model is a model trained using a three-phase power grid load flow algorithm. Based on this, when calculating the balance parameters, the three-phase power grid load flow algorithm is specifically used to calculate the load flow of the three-phase power grid based on the voltage, current, and power factor of each phase in the three-phase power grid data. Using the three-phase imbalance method, the load ratio of each phase is calculated based on the load flow of the three-phase power grid, and the load ratio of each phase is used as the balance parameter between the phases in the three-phase power grid.

[0096] In practical applications, current and voltage data are extracted from three-phase power grid data and converted into corresponding vectors. Ohm's theorem is then used to construct a balance calculation formula: V = Z * I, where V is the voltage vector, Z is the impedance matrix, and I is the current vector. Based on the results calculated using this formula, a three-phase power grid model is used to analyze the load imbalance of each phase of the power grid. According to the analyzed load imbalance, balance parameters such as current, voltage, and power factor in the three-phase power grid are determined, and feedback signals are provided for subsequent energy storage system optimization.

[0097] In another feasible implementation, when analyzing the load imbalance degree of each phase of the power grid, the current deviation and voltage deviation of each phase of the power grid are calculated, and the three-phase imbalance factor is used as an evaluation index for power grid balance. Balance parameters are then determined based on this evaluation index. The formula for calculating the imbalance factor is as follows:

[0098] ,in, For current deviation, voltage deviation, and power difference, This is the three-phase current imbalance factor. Let be the current in the i-th phase.

[0099] The three-phase power grid load flow algorithm is actually used to process the voltage, phase angle, and power distribution of each phase in the power grid, specifically including:

[0100] Node voltage calculation: The relationship between the voltage magnitude and phase angle of each phase can be calculated using the load current algorithm. Based on the phase difference relationship between voltage and current, the balance of current, power, and voltage in the three-phase power grid is solved.

[0101] Power Calculation: Based on the voltage, current, and power factor of each phase, calculate the active power (P) and reactive power (Q) of each phase (A, B, C) using the following formulas:

[0102]

[0103]

[0104] in, These represent the voltage, current, and power factor of phase A, respectively.

[0105] Based on the current, voltage, and power factors of the three-phase power grid, multiple calculations are performed until the calculation results converge, and the voltage, current, and power distribution of each node and each phase are obtained, which is the load flow of the three-phase power grid.

[0106] Then, based on the load flow calculation results, the differences in current, voltage, and power for each phase are calculated, thereby deriving the three-phase imbalance of the power grid, specifically including:

[0107] Current imbalance (ΔI):

[0108]

[0109] This formula is used to assess the difference between three-phase currents. A large current imbalance may lead to grid failure or equipment overload.

[0110] Voltage imbalance (ΔV):

[0111]

[0112] Voltage imbalance can affect the normal operation of electrical equipment, leading to reduced energy efficiency and increased equipment losses.

[0113] Power imbalance: The power imbalance is calculated by comparing the active and reactive power of each phase. A large power imbalance may indicate that a phase is receiving too much or too little power.

[0114] Finally, based on the load flow calculation results and the analysis of the three-phase unbalance, the load percentage of each phase is calculated. The calculation method for the load percentage is as follows:

[0115]

[0116] Among them, P A P B P C These are the active power of phases A, B, and C in a three-phase power grid.

[0117] These load percentages reflect the proportion of load for each phase in the entire power grid, providing a basis for the configuration of energy storage devices. For phases with a large load percentage, energy storage devices will be allocated more to these phases to achieve load balancing in the power grid.

[0118] Determining load balancing parameters: The calculated load percentage for each phase is used as the balancing parameter in subsequent energy storage configuration optimization. Based on the load percentage, the charging and discharging strategies of the energy storage equipment can be dynamically adjusted through optimization algorithms to achieve load balance between phases.

[0119] 204. Improve the genetic algorithm to optimize energy storage configuration.

[0120] Specifically, an improved genetic algorithm (IGA) is used to optimize the configuration of energy storage devices. The population is initialized, fitness values ​​are calculated, and crossover and mutation operations are performed to select the optimal energy storage configuration. This process considers the capacity, charging / discharging time, and power allocation of the energy storage devices.

[0121] It should be noted that the improved genetic algorithm used in this invention, compared to the traditional genetic algorithm, mainly improves the following aspects:

[0122] Individual selection strategy: By setting a preset balance value, excellent individuals are selected for subsequent crossover and mutation operations, thereby improving the convergence speed and optimization effect of the algorithm.

[0123] Crossover and mutation operations: Through carefully designed crossover and mutation operations, the diversity of the population is increased, which helps the algorithm find the optimal solution in a wider search space and escape local optima.

[0124] These improvements enable the improved genetic algorithm to achieve higher efficiency and accuracy in optimizing the configuration parameters of three-phase energy storage systems.

[0125] In practical applications, the optimization of the three-phase energy storage configuration parameters in the energy storage system can be generally divided into the following steps:

[0126] Step 1, construct the initial population;

[0127] First, generate energy storage configuration schemes: Based on parameters such as the type, capacity, charging and discharging time, and power of the energy storage equipment, generate a series of possible energy storage configuration schemes.

[0128] Then, construct the initial population: use the generated energy storage configuration schemes as individuals in the initial population, with each individual representing an energy storage configuration scheme. The size of the initial population should be determined based on the complexity of the problem and computational resources.

[0129] Step 2, the iterative process of the genetic algorithm;

[0130] First, calculate the fitness value: For each individual in the initial population, calculate its fitness value according to the fitness function, that is, calculate the adaptability of each individual in balancing grid load and energy storage.

[0131] Then, the selection operation: based on the fitness value, select a group of excellent individuals as parents for subsequent crossover and mutation operations; roulette wheel selection, tournament selection, and other strategies can be used for the selection operation.

[0132] It should be noted that, firstly, the fitness of each individual is evaluated, and the fitness function for each individual is set as follows:

[0133]

[0134] Where F(x) i ) is an individual x i Total fitness, f j (xi) is the individual x i The fitness value ω on the objective function j j is the weight of objective function j, and M is the number of objective functions.

[0135] Then, in the selection process, the fitness value of each individual is mapped to a selection probability. Assuming there are N individuals, the individual selection probability is:

[0136]

[0137] Wherein, P(x i ) is an individual x i The probability of being selected.

[0138] Finally, a weighted roulette wheel selection algorithm is used to select individuals from the current population to enter the next generation based on probability. That is, based on the fitness value of each individual, a corresponding interval is set in the roulette wheel, and individuals are randomly selected from the population.

[0139] Furthermore, crossover operations: crossover operations are performed on selected parent individuals to generate new individuals; crossover operations can employ strategies such as single-point crossover, double-point crossover, and uniform crossover.

[0140] During the crossover process, it should be ensured that the genes of the new individual (i.e., some parameters of the energy storage configuration scheme) are within a reasonable range and meet the constraints of the power grid.

[0141] Assume there are two parent individuals and The crossover operation is as follows:

[0142] Generate crossover offspring: For each gene and A random exchange strategy is adopted:

[0143]

[0144] Where, r j It is a uniformly random number, and r j ∈[0,1]. If r j If <0.5, then the parent generation will be... The gene is passed on to the offspring y1, otherwise it will be passed on to the parent generation. The genes of the parent are passed on to offspring y1. Similarly, the genes of the other offspring y2 come from the exchange between the parents.

[0145] Crossover probability: The crossover operation only occurs when certain probability conditions are met; the crossover probability is P. c (Usually Pc∈[0.6,0.9), that is:

[0146] , where C cross C is the number of times the crossover operation is performed. total It is the total number of individuals in the population.

[0147] Furthermore, mutation operations: Mutation operations are performed on newly generated individuals to further increase the diversity of the population; mutation operations can employ strategies such as random mutation and Gaussian mutation.

[0148] During the mutation process, it is also necessary to ensure that the genes of the new individuals are within a reasonable range and meet the constraints of the power grid.

[0149] For individual x=(x 1 ,x 2 ,...,xk The mutation operation is as follows:

[0150] Randomly select genes for mutation: Assume that the mutation probability of each gene is P during the mutation operation. m Given the mutation probability, the individual's gene x i changed to The specific formula is:

[0151]

[0152] Where, r i It is a random number, if r i <P m Then gene x i It is replaced with a new random value (the range of values ​​depends on the specific problem), otherwise it remains unchanged.

[0153] Mutation probability: The probability P of mutation m Typically lower (e.g., P) m =0.01 to P m =0.1), to avoid excessively disturbing the already found excellent solutions.

[0154] Furthermore, the population is updated: new individuals generated through crossover and mutation operations are added to the population, replacing some individuals with lower fitness; after updating the population, the next round of iteration is carried out.

[0155] Step 3: Convergence judgment and output of the optimal solution;

[0156] First, convergence judgment: During the iteration process, the average fitness value of the population or the fitness value of the best individual is calculated periodically; if the average fitness value or the fitness value of the best individual does not change significantly within a certain number of generations, or reaches the preset threshold, the algorithm is considered to have converged to the optimal solution.

[0157] Then, the optimal solution is output: when the algorithm converges to the optimal solution, the optimal solution is output as the three-phase energy storage configuration parameters in the energy storage system; at the same time, detailed information such as the fitness value of the optimal solution, the type, capacity, charging and discharging time and power of the energy storage device can be output.

[0158] Step 4, follow-up processing and verification;

[0159] First, verify the optimal solution: apply the obtained optimal solution to the actual energy storage system and conduct simulation or experimental verification; by comparing the simulation or experimental results with the expected goals (such as the effect of load peak shaving and valley filling, the economics of the energy storage system, etc.), evaluate the feasibility and effectiveness of the optimal solution.

[0160] Then, optimization and adjustment: Based on the verification results, the optimal solution is adjusted and optimized as necessary to improve its performance and effectiveness in practical applications.

[0161] In this embodiment, all energy storage configuration schemes of the energy storage system at the current moment are obtained, and an initial population is constructed based on all the energy storage configuration schemes. Each individual in the initial population is an energy storage configuration scheme, and each energy storage configuration scheme includes the type, capacity, charging and discharging time, and power of the energy storage device. Based on the balance parameters, predicted load data, and power generation data, the fitness value of each individual is calculated. Based on the fitness value, individuals whose balance meets the preset balance value are selected for crossover and mutation operations to generate a new population until convergence to the optimal solution. The three-phase energy storage configuration parameters in the energy storage system are optimized based on the optimal solution. The crossover operation involves selecting multiple individuals whose balance reaches the preset balance value to exchange genes with each other, and the mutation operation involves crossover operations on multiple individuals from different spatial dimensions.

[0162] It should be noted that after calculating the fitness value of an individual for a certain time period based on the fitness function described above, the subsequent calculation of the fitness value for the next time period can be directly based on the fitness value of the previous time period. That is, the fitness function described above can be simplified to the following formula:

[0163] F= ,

[0164] Where n is a certain prediction time period, For load power, For power generation capacity, For energy storage power, This represents the fitness value from the previous time period.

[0165] Furthermore, the maintenance costs of equipment with different configurations also vary. Therefore, in the configuration optimization process, in addition to optimizing parameters, it also includes maximizing the utilization efficiency of the individual devices after parameter optimization, specifically including:

[0166] Maximizing Charge / Discharge Efficiency Costs: In the optimization of energy storage systems, the differences in charge / discharge efficiency among different energy storage devices directly impact operating costs. Energy storage devices with lower charge / discharge efficiency require more energy to complete the same charge / discharge task, thus increasing the operating costs of the power system. Therefore, when calculating fitness, the charge / discharge efficiency of each energy storage configuration must be considered, and the charge / discharge loss cost must be calculated using the following formula:

[0167]

[0168] in, Costs incurred during the charging and discharging process. and These represent the charging power and discharging power for each time period, respectively. and , respectively, are the charging and discharging efficiencies of the energy storage device, k3 is the coefficient of charging and discharging loss cost, and t is a certain time point in the prediction period, t∈[0,N], and the prediction period contains N time points.

[0169] The fitness function can be expressed by the following comprehensive formula:

[0170] ,in, and The fitness value represents the weight of each cost item. The higher the fitness value, the better the energy storage configuration scheme performs in balancing grid load and energy storage.

[0171] 205. Dynamic scheduling and charge / discharge optimization.

[0172] Based on the optimized energy storage configuration parameters, a charging and discharging timing table is generated. Then, dynamic scheduling optimization is performed using fuzzy control algorithms and linear programming (LP) based on real-time grid load and generation capacity changes. Finally, the charging and discharging strategy of the energy storage equipment is adjusted to ensure grid load balance and reduce energy storage costs. This can be achieved through the following steps:

[0173] Step 1: Define fuzzy variables and membership functions;

[0174] Determine the input and output variables: Based on the requirements of the control system, determine the variables that need to be controlled, such as data deviation and the charging and discharging power of the equipment as input variables, and the specific parameters of the charging and discharging strategy as output variables.

[0175] Selecting a membership function: Choose an appropriate membership function for each fuzzy variable, such as a triangular, trapezoidal, or Gaussian function; the membership function is used to describe the fuzziness of the variable's values ​​and convert clear values ​​into fuzzy values.

[0176] Define fuzzy sets: Based on the range of values ​​for the input and output variables, define a series of fuzzy sets, such as "small", "medium", "large", etc. Each fuzzy set corresponds to a membership function.

[0177] Step 2: Establish a fuzzy rule base;

[0178] Formulate fuzzy rules: Based on expert experience and system requirements, formulate a series of fuzzy control rules.

[0179] The rules are similar in form to conditional statements like "if...then...", such as "if the data deviation is small and the device power is low, then the charging and discharging strategy is stable".

[0180] Storing rules: The defined fuzzy rules are stored in the rule base for use in subsequent reasoning.

[0181] Step 3, fuzzify the input variables;

[0182] Calculate membership degree: Use the membership degree function to convert the sharp values ​​of the input variables into fuzzy values, that is, calculate the membership degree of the input variables in each fuzzy set.

[0183] Constructing a fuzzy set: Based on the calculated membership degrees, construct a fuzzy set of the input variables.

[0184] Step 4, fuzzy reasoning;

[0185] Matching rules: Based on the fuzzy set of input variables, match the corresponding rules in the rule base.

[0186] Calculate the output fuzzy set: Use inference methods (such as max-min inference) to calculate the fuzzy set of the output variables. The inference process may involve the combination and operation of multiple rules.

[0187] Step 5: Deblur the output variables;

[0188] Choose a deblurring method: Select an appropriate deblurring method based on actual needs, such as the centroid method, the maximum membership method, or the weighted average method.

[0189] Calculate clear output values: Use the selected defuzzification method to convert the fuzzy set of output variables into clear control output values.

[0190] Step 6, Application and Optimization;

[0191] Implement control: Apply the calculated clear output to the actual system to realize the dynamic scheduling and planning of the charging and discharging strategy.

[0192] Monitoring and Adjustment: In practical applications, continuously monitor changes in system performance and output variables.

[0193] Adjust the membership function, fuzzy set, and rule base based on the monitoring results to optimize control performance.

[0194] In this embodiment, after optimizing the three-phase energy storage configuration parameters of the energy storage system using an improved genetic algorithm based on the balance parameters, predicted load data, and power generation data, the method further includes:

[0195] Based on the optimized three-phase energy storage configuration parameters, predicted load data and power generation data, a charging and discharging time sequence table for all devices in the energy storage system is generated, wherein the charging and discharging time sequence table includes the charging and discharging time period, power capacity and priority of the devices;

[0196] Real-time load data and power generation data for the predicted time period are obtained, and several devices are selected from the charging and discharging time series table according to the balance parameters using an improved optimization strategy to perform dynamic scheduling planning, thereby obtaining a charging and discharging strategy. The improved optimization strategy is a control strategy that adds linear programming to the fuzzy control algorithm.

[0197] Based on the charging and discharging strategy, the devices in the energy storage system are scheduled to perform charging and discharging operations.

[0198] In practical applications, based on predicted load data, power generation data, and grid load flow analysis results, a charging and discharging time sequence table for energy storage devices is generated to optimize the selection of charging and discharging time periods, specifically including the selection of battery charging periods and the priority setting of discharging periods.

[0199] Assuming a low-voltage distribution area, after optimizing the three-phase energy storage configuration parameters and analyzing predicted load and power generation data, the following charging and discharging sequence table for the energy storage system equipment is obtained. Taking a group of energy storage devices as an example, device A's charging period is 0:00-6:00, with a charging capacity of 500kWh and a high priority; device B's discharging period is 18:00-22:00, with a discharging capacity of 300kWh and a medium priority. This information forms the basis for the charging and discharging sequence arrangement of all devices in the energy storage system, providing initial planning for subsequent dynamic scheduling.

[0200] During the forecast period (e.g., 18:00-22:00), load and power generation data for the low-voltage distribution area are collected in real time using smart meters and a power generation monitoring system. Assume that the real-time load data shows that at 19:00, the actual load is 50kW higher than the forecast load; and the real-time power generation data shows that the photovoltaic power generation is 30kW lower than the forecast.

[0201] Furthermore, after obtaining the charge / discharge timing table, a fuzzy control algorithm (FLC) is used to dynamically adjust the charge / discharge strategy of the energy storage device based on real-time grid load, generation capacity, and grid imbalance data. The rules of the fuzzy control are as follows: Where M is the number of rules, and Output(t) is the energy storage device's charging and discharging decision. For fuzzy rules, Input variables include real-time load and power generation; based on dynamically adjusted charging and discharging strategies, the charging and discharging power allocation of energy storage devices is optimized to minimize grid load fluctuations and maximize the operating efficiency of the energy storage system.

[0202] The step of using an improved optimization strategy to dynamically schedule and plan the charge / discharge timing table according to the balance parameters to obtain a charge / discharge strategy includes:

[0203] The data deviation between real-time data and predicted data within the prediction period is calculated. This data deviation includes the difference between real-time load data and predicted load data, as well as the difference between real-time power generation data and predicted power generation data. Based on the balance parameters and the data deviation, a linear programming objective function for each device is constructed. The linear programming objective function is: ,in, The unit energy cost during the charge / discharge period. Let n be the absolute value of the charging and discharging power of the device, and n be the prediction time period. Find the optimal solution for the linear programming objective function, and select several devices from the charging and discharging time series table based on the optimal solution. Use a fuzzy control algorithm to fuzzify the parameters of the several devices to obtain the charging and discharging strategy.

[0204] The objective function of this linear programming method employs absolute value constraints on charging and discharging power to eliminate negative power fluctuations during charging and discharging, ensuring smooth charging and discharging processes and reducing energy loss during battery charging and discharging. This method can effectively improve the efficiency of energy storage systems, especially when charging and discharging frequently change, avoiding damage to the battery due to excessive power fluctuations.

[0205] Data deviations were calculated: at 19:00, the real-time load data was 800kW, while the predicted load data was 750kW, a difference of 50kW; the real-time power generation data was 100kW, while the predicted power generation data was 130kW, a difference of -30kW. These data deviations reflect the differences between the actual power situation and the predicted situation, and are an important basis for subsequent dispatching strategies.

[0206] Constructing a linear programming objective function: Given that the unit energy cost during the charging / discharging period (e.g., 19:00-20:00) is 0.5 yuan / kWh, construct a linear programming objective function for each device based on balance parameters (e.g., phase A needs additional power to balance the grid) and data deviations. For devices A and B, substitute their possible charging / discharging power into this function. Considering the charging / discharging capacity limitations of the devices (e.g., the maximum discharge power of device A is 100kW, and the maximum discharge power of device B is 80kW), determine the optimal combination of charging / discharging power under the conditions of satisfying grid balance and device constraints through calculation.

[0207] Solving for the optimal solution and selecting equipment: A linear programming algorithm is used to solve the objective function to obtain the optimal solution. Assume the optimal solution indicates that equipment A should discharge at a power of 50kW, and equipment B should discharge at a power of 50kW. Based on this result, equipment A and equipment B are selected from the charging / discharging timing table to participate in this scheduling.

[0208] Fuzzy processing yields the charging and discharging strategy: Fuzzy control algorithms are used to fuzzify the charging and discharging parameters of devices A and B. For example, based on real-time grid load fluctuations and the current remaining power of the devices, parameters such as charging and discharging power and time are fuzzified into fuzzy sets such as "high," "medium," and "low." For device A, its discharge power of 50kW is fuzzified as "medium," and the discharge time is adjusted according to the remaining power and grid demand. A similar fuzzification process is performed on device B. Finally, by combining these fuzzified parameters, a specific charging and discharging strategy is obtained. For example, device A continuously discharges at a medium discharge power from 19:00 to 20:00, and device B also discharges at a medium discharge power during the same time period, dynamically adjusting according to real-time conditions to ensure grid load balance and efficient operation of the energy storage system.

[0209] Furthermore, after obtaining the charging and discharging strategy, the process also includes: solving for the optimal energy storage device scheduling strategy based on constraints such as energy storage capacity and charging and discharging time windows to maximize the economic benefits of the energy storage system; providing real-time feedback on the scheduling results and adjusting according to changes in grid load to ensure that the charging and discharging energy of the energy storage devices is minimized, while reducing the operating costs of the grid.

[0210] Understandably, this constraint specifically includes:

[0211] The charging power of energy storage devices must not exceed their maximum capacity. ;

[0212] The discharge power of energy storage devices must not exceed their maximum discharge capacity. ;

[0213] The charging and discharging process of energy storage devices must meet the requirements of grid load balance, that is:

[0214] ,in, Let i be the charging power during the i-th time period. The discharge power in the j-th time period is used to ensure that the energy storage system effectively balances the grid load;

[0215] By setting the above constraints, we can ensure that the charging and discharging behavior of the energy storage device does not exceed its capacity, and maximize the operating efficiency of the energy storage device while satisfying the grid load balance.

[0216] At this point, the fuzzy control rules used in the dynamic scheduling of the charging and discharging strategies of energy storage devices using the fuzzy control algorithm (FLC) specifically include, but are not limited to, any of the following:

[0217] Input conditions include: grid load fluctuations, changes in power generation, and the current status of energy storage devices (such as remaining capacity, charging and discharging efficiency, etc.).

[0218] Output conditions include: charging power, discharging power, and charging / discharging priority of the energy storage device;

[0219] By utilizing fuzzy control algorithms, complex and uncertain power grid load fluctuations can be flexibly handled, and the charging and discharging strategies of energy storage devices can be adjusted based on real-time data to ensure stable system operation.

[0220] In summary, fuzzy control algorithms can be used to quickly set fuzzy rules for load, power generation and grid imbalance, and output charging and discharging decisions, making this method applicable to energy storage systems with various hardware configurations. Furthermore, linear programming (LP) is used to minimize the total energy loss and cost of charging and discharging of energy storage devices. Constraints include energy storage capacity and charging and discharging limits, which can mitigate the problem of system lag caused by parameter fluctuations.

[0221] 206. Scheduling execution and real-time feedback.

[0222] Based on dynamic scheduling results, the charging and discharging operations of energy storage devices are executed in real time. The system monitors grid load changes and the operating status of energy storage devices in real time, adjusting charging and discharging strategies to cope with load fluctuations. A feedback mechanism ensures the operating efficiency of energy storage devices and grid stability.

[0223] Specifically, this step includes: continuously optimizing the configuration of the energy storage system based on real-time grid load, power generation data, and energy storage device status monitoring data to ensure that the energy storage system can flexibly respond to different load scenarios; adjusting the capacity allocation and charging / discharging strategies of the energy storage devices when grid load fluctuates significantly to reduce the impact of grid imbalance on power system stability; and conducting long-term optimized scheduling based on the health status and lifespan of the energy storage devices to ensure that the devices maintain high operating efficiency and stability throughout their lifespan.

[0224] In this embodiment of the invention, by real-time collection and analysis of power load, power generation data, meteorological information and three-phase power grid data of low-voltage distribution areas, a three-phase power grid model and an improved genetic algorithm are used to dynamically adjust the configuration and operation strategy of the energy storage system, so as to achieve configuration flexibility and improve the system's adaptability, economy and operating efficiency.

[0225] A three-phase power grid model is used to realize load flow analysis and energy storage optimization of the three-phase power grid, which minimizes the three-phase imbalance of the power grid, improves the stability of the power grid, and reduces the investment and operating costs of the energy storage system.

[0226] By utilizing fuzzy control algorithms and linear programming, the charging and discharging strategies of energy storage devices can be flexibly adjusted based on factors such as changes in grid load and fluctuations in renewable energy generation, thereby improving the responsiveness and adaptability of the energy storage system.

[0227] The control method for the energy storage system in the low-voltage distribution area in this embodiment of the invention has been described above. The control device for the energy storage system in the low-voltage distribution area in this embodiment of the invention is described below. Please refer to [link / reference]. Figure 3 and 4 An embodiment of the energy storage system control device for low-voltage distribution areas in this invention includes:

[0228] The acquisition module 310 is used to acquire the working data of the current time period in the low-voltage distribution area. The working data includes power load data, renewable energy power generation data, meteorological data and three-phase power grid data.

[0229] The prediction module 320 is used to predict the load data and power generation data of the low-voltage distribution area during the prediction period based on the power load data, the renewable energy power generation data and the meteorological data using a pre-trained prediction model.

[0230] The optimization module 330 is used to determine the balance parameters between the phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data using a three-phase linear power grid model, and to optimize the three-phase energy storage configuration parameters in the energy storage system using an improved genetic algorithm based on the balance parameters, predicted load data and power generation data.

[0231] Optionally, the prediction model is a load prediction model based on Long Short-Term Memory Network (LSTM) and a power generation prediction model based on Support Vector Regression (SVR).

[0232] The prediction module 320 includes:

[0233] The first prediction unit 321 is used to input the power load data into the load prediction model, extract the first time-series feature of the load at each time point in the power load data, and calculate the load data of the low-voltage area at the next time point based on each of the first time-series features and the meteorological data.

[0234] The second prediction unit 322 is used to input the power generation data of the renewable energy into the power generation prediction model, extract the second time series feature of power generation at each time point in the power generation data of the renewable energy, and calculate the power generation data of the low-pressure area at the next time point based on each second time series feature and the meteorological data.

[0235] Optionally, the three-phase linear power grid model is a model based on the three-phase power grid load flow algorithm;

[0236] The optimization module 330 includes: a calculation unit 331, used for:

[0237] Using the aforementioned three-phase power grid load flow algorithm, the load flow of the three-phase power grid is calculated based on the voltage, current, and power factor of each phase in the three-phase power grid data.

[0238] Using the three-phase imbalance method, the load proportion of each phase is calculated based on the load flow of the three-phase power grid, and the load proportion of each phase is used as the balance parameter between the phases in the three-phase power grid.

[0239] The optimization module 330 includes: an optimization unit 332, used for...

[0240] Obtain all energy storage configuration schemes of the energy storage system at the current moment, and construct an initial population based on all the energy storage configuration schemes. Each individual in the initial population is an energy storage configuration scheme, and each energy storage configuration scheme includes the type, capacity, charging and discharging time and power of the energy storage device.

[0241] Based on the balance parameters, predicted load data, and power generation data, the fitness value of each individual is calculated;

[0242] Individuals whose fitness values ​​meet the preset fitness threshold are selected for crossover and mutation operations to generate a new population until convergence to the optimal solution. Based on the optimal solution, the three-phase energy storage configuration parameters in the energy storage system are optimized. The crossover operation involves selecting multiple individuals whose fitness values ​​meet the preset fitness threshold to exchange genes with each other, and the mutation operation involves crossover operations on multiple individuals from different spatial dimensions.

[0243] Optionally, the fitness value is calculated using the following formula:

[0244] F= Where n is a certain prediction time period, For load power, For power generation capacity, For energy storage power, This represents the fitness value from the previous time period.

[0245] Optionally, the device further includes a scheduling module 340, used for:

[0246] Based on the optimized three-phase energy storage configuration parameters, predicted load data and power generation data, a charging and discharging time sequence table for all devices in the energy storage system is generated, wherein the charging and discharging time sequence table includes the charging and discharging time period, power capacity and priority of the devices;

[0247] Real-time load data and power generation data for the predicted time period are obtained, and several devices are selected from the charging and discharging time series table according to the balance parameters using an improved optimization strategy to perform dynamic scheduling planning, thereby obtaining a charging and discharging strategy. The improved optimization strategy is a control strategy that adds linear programming to the fuzzy control algorithm.

[0248] Based on the charging and discharging strategy, the devices in the energy storage system are scheduled to perform charging and discharging operations.

[0249] Optionally, the scheduling module 340 is specifically used for:

[0250] The data deviation between real-time data and predicted data within the prediction period is calculated. The data deviation includes the difference between real-time load data and predicted load data, as well as the difference between real-time power generation data and predicted power generation data.

[0251] Based on the balance parameters and the data deviation, a linear programming objective function is constructed for each device, wherein the linear programming objective function is: ,in, The unit energy cost during the charge / discharge period. Here, n represents the absolute value of the charging and discharging power of the device, and n is the predicted time period.

[0252] Find the optimal solution for the linear programming objective function, and select several devices from the charging and discharging timing table based on the optimal solution. Use a fuzzy control algorithm to fuzzify the parameters of the several devices to obtain the charging and discharging strategy.

[0253] In this embodiment of the invention, the current power load data, renewable energy generation data, meteorological data, and three-phase power grid data of a low-voltage distribution area are acquired. A pre-trained prediction model is used to predict the load and generation data of the low-voltage distribution area at the next time point based on the power load data, generation data, and meteorological data. A three-phase linear power grid model is used to determine the balance parameters between the phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data. An improved genetic algorithm is then used to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, the predicted load data, and the generation data. This addresses the problem of inflexible matching between existing energy storage system configuration and operation strategies, leading to low system efficiency and resource waste.

[0254] above Figure 3 and Figure 4 The energy storage system control device of the low-voltage substation in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The electronic equipment in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0255] See Figure 5 As shown, the electronic device includes a processor 500 and a memory 501. The memory 501 stores machine-executable instructions that can be executed by the processor 500. The processor 500 executes the machine-executable instructions to implement the energy storage system control method of the low-voltage substation described above.

[0256] Furthermore, Figure 5 The electronic device shown also includes a bus 502 and a communication interface 503. The processor 500, the communication interface 503 and the memory 501 are connected via the bus 502.

[0257] The memory 501 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 503 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc. The bus 502 may be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0258] The processor 500 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 500 or by instructions in software form. The processor 500 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 501. The processor 500 reads the information in memory 501 and, in conjunction with its hardware, completes the method steps of the aforementioned embodiment.

[0259] The present invention also provides an electronic device, the computer device including a memory and a processor, the memory storing computer-readable instructions, which, when executed by the processor, cause the processor to perform the steps of the energy storage system control method for the low-voltage substation in the above embodiments.

[0260] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the energy storage system control method for the low-voltage substation.

[0261] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0262] 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, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0263] The above-described 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 an energy storage system in a low-voltage distribution area, characterized in that, Includes the following steps: Data acquisition steps: Acquire the working data of the low-voltage distribution area for the current time period, including power load data, renewable energy generation data, meteorological data, and three-phase power grid data; Data prediction steps: Using a pre-trained prediction model, the load data and power generation data for the predicted time period in the low-voltage distribution area are predicted based on the power load data, the renewable energy power generation data, and the meteorological data; wherein, the prediction model includes a load prediction model based on long short-term memory networks and a power generation prediction model based on support vector regression. Configuration optimization steps: Using a three-phase linear power grid model, the balance parameters between each phase in the three-phase power grid are determined based on the electrical parameters of each phase in the three-phase power grid data. Then, using an improved genetic algorithm, the three-phase energy storage configuration parameters in the energy storage system are optimized based on the balance parameters, predicted load data, and power generation data. The operation and scheduling of the energy storage system are controlled based on the optimized parameters. The three-phase linear power grid model is a model based on the three-phase power grid load flow algorithm, and the balance parameters are the load proportion of each phase. The step of using an improved genetic algorithm to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and power generation data, and controlling the operation scheduling of the energy storage system based on the optimized parameters includes: constructing multiple energy storage configuration schemes using an improved genetic algorithm; calculating the fitness of the energy storage configuration schemes based on the predicted load data and power generation data; selecting one of the multiple energy storage configuration schemes based on the fitness and performing crossover and mutation operations until convergence to the optimal one; and controlling the operation scheduling of the energy storage system based on the load ratio of each phase and the optimal energy storage configuration scheme.

2. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 1, characterized in that, The data prediction step includes: The power load data is input into the load prediction model, and the first time-series feature of the load at each time point in the power load data is extracted; The load data for the predicted time period in the low-pressure area is calculated based on each of the first time-series features and the meteorological data. The power generation data of the renewable energy source is input into the power generation prediction model, and the second time-series feature of power generation at each time point in the power generation data of the renewable energy source is extracted; Based on the second time-series characteristics and the meteorological data, the power generation data for the predicted time period in the low-pressure area is calculated.

3. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 1, characterized in that, The step of determining the balance parameters between each phase in the three-phase power grid in the configuration optimization step includes: Using the aforementioned three-phase power grid load flow algorithm, the load flow of the three-phase power grid is calculated based on the voltage, current, and power factor of each phase in the three-phase power grid data. Using the three-phase imbalance method, the load proportion of each phase is calculated based on the load flow of the three-phase power grid, and the load proportion of each phase is used as the balance parameter between the phases in the three-phase power grid.

4. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 1, characterized in that, The configuration optimization step optimizes the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and power generation data, including: Obtain all energy storage configuration schemes of the energy storage system for the current time period, and construct an initial population based on all the energy storage configuration schemes. Each individual in the initial population is an energy storage configuration scheme, and each energy storage configuration scheme includes the type, capacity, charging and discharging time and power of the energy storage device. Based on the predicted load and power generation data, the fitness value of each individual is calculated; Individuals whose fitness values ​​meet the preset fitness threshold are selected for crossover and mutation operations to generate a new population until convergence to the optimal solution. Based on the optimal solution and the balance parameters, the three-phase energy storage configuration parameters in the energy storage system are optimized. The crossover operation involves selecting multiple individuals whose fitness values ​​meet the preset fitness threshold to exchange genes with each other, and the mutation operation involves crossover operations on multiple individuals from different spatial dimensions.

5. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 4, characterized in that, The formula for calculating the fitness value is: F= Where n is a certain prediction time period, For load power, For power generation, For energy storage power, This represents the fitness value from the previous time period.

6. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 1, characterized in that, Following the configuration optimization step, the following is also included: Based on the optimized three-phase energy storage configuration parameters, predicted load data and power generation data, a charging and discharging time sequence table for all devices in the energy storage system is generated, wherein the charging and discharging time sequence table includes the charging and discharging time period, power capacity and priority of the devices; Real-time load data and power generation data for the predicted time period are obtained, and several devices are selected from the charging and discharging time series table according to the balance parameters using an improved optimization strategy to perform dynamic scheduling planning, thereby obtaining a charging and discharging strategy. The improved optimization strategy is a control strategy that adds linear programming to the fuzzy control algorithm. Based on the charging and discharging strategy, the devices in the energy storage system are scheduled to perform charging and discharging operations.

7. The method for optimizing the energy storage system in a low-voltage distribution area according to claim 6, characterized in that, The improved optimization strategy selects several devices from the charging / discharging timing table according to the balance parameters for dynamic scheduling planning, resulting in a charging / discharging strategy, including: The data deviation between real-time data and predicted data within the prediction period is calculated. The data deviation includes the difference between real-time load data and predicted load data, as well as the difference between real-time power generation data and predicted power generation data. Construct the linear programming objective function for each device, wherein the linear programming objective function is: ,in, The unit energy cost during the charge / discharge period. Here, n represents the absolute value of the charging and discharging power of the device, and n is the predicted time period. Find the optimal solution for the linear programming objective function, and select several devices from the charging and discharging timing table based on the optimal solution. Use a fuzzy control algorithm to fuzzify the parameters of the several devices to obtain the charging and discharging strategy.

8. An energy storage system optimization device for a low-voltage distribution area, characterized in that, The device includes: The data acquisition module is used to acquire the working data of the low-voltage distribution area for the current time period. The working data includes power load data, renewable energy power generation data, meteorological data and three-phase power grid data. The data prediction module is used to predict the load data and power generation data of the low-voltage distribution area for the predicted time period based on the power load data, the power generation data of the renewable energy and the meteorological data using a pre-trained prediction model; wherein the prediction model is a load prediction model based on long short-term memory network and a power generation prediction model based on support vector regression. The configuration optimization module is used to determine the balance parameters between phases in the three-phase power grid based on the electrical parameters of each phase in the three-phase power grid data using a three-phase linear power grid model, and to optimize the three-phase energy storage configuration parameters in the energy storage system using an improved genetic algorithm based on the balance parameters, predicted load data, and power generation data, and to control the operation of the energy storage system based on the optimized parameters. The three-phase linear power grid model is a model based on the three-phase power grid load flow algorithm, and the balance parameters are the load proportion of each phase. The step of using an improved genetic algorithm to optimize the three-phase energy storage configuration parameters in the energy storage system based on the balance parameters, predicted load data, and power generation data, and controlling the operation scheduling of the energy storage system based on the optimized parameters includes: constructing multiple energy storage configuration schemes using an improved genetic algorithm; calculating the fitness of the energy storage configuration schemes based on the predicted load data and power generation data; selecting one of the multiple energy storage configuration schemes based on the fitness and performing crossover and mutation operations until convergence to the optimal one; and controlling the operation scheduling of the energy storage system based on the load ratio of each phase and the optimal energy storage configuration scheme.

9. An electronic device, characterized in that, The electronic device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the electronic device to execute the energy storage system optimization method for low-voltage substations as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements the energy storage system optimization method for low-voltage distribution areas as described in any one of claims 1-7.

Images