Power distribution method, device, equipment, medium and product of energy storage system

Abstract

The application discloses a power distribution method, device, equipment, medium and product of an energy storage system, and the method comprises the following steps: collecting electrical parameters and environmental data of each energy storage unit of the energy storage system; performing electrochemical simulation on the electrical parameters and environmental data of the current energy storage unit to obtain electrochemical implicit characteristics, and fusing the electrical parameters, environmental data and electrochemical implicit characteristics of the current energy storage unit into a state matrix; obtaining health influence parameters of each energy storage unit and calculating health degrees, inputting each state matrix and each health degree into a power distribution model, taking minimization of total power loss and total health loss of the energy storage system as a target, and solving an initial power distribution scheme under a preset constraint condition, and inputting the initial power distribution scheme into a preset correction model for correction to obtain a target power distribution scheme of the energy storage system, so that the energy storage system is distributed with power according to the target power distribution scheme. The application can improve the accuracy of power distribution of the energy storage system under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of power control for energy storage systems, and more particularly to power distribution methods, devices, equipment, media, and products for energy storage systems. Background Technology

[0002] Energy storage systems, as key devices for mitigating power fluctuations from renewable energy grid connection, supporting stable microgrid operation, and enabling peak-valley arbitrage on the user side, are important technical means to ensure power supply reliability, enhance renewable energy absorption capacity, and reduce user electricity costs. In recent years, with the large-scale grid connection of intermittent power sources such as wind and solar power, and the rapid growth of new loads such as microgrids and electric vehicle charging facilities, the operating conditions of energy storage systems have become increasingly complex, placing higher demands on the accuracy and flexibility of power control. Energy storage systems typically consist of multiple energy storage units. If power allocation is unreasonable, it can not only lead to overload or underload of some energy storage units, accelerating their aging process, but also cause problems such as system response lag, overall efficiency decline, and even affect the power balance and frequency stability of the power grid.

[0003] Existing energy storage systems mostly employ proportional power allocation techniques based on fixed mathematical models. However, these techniques have several drawbacks: Firstly, the monitoring of energy storage units is limited to basic electrical parameters such as voltage and current, failing to integrate key influencing factors such as environment, aging, and charge / discharge cycle count. This results in the allocation model failing to fully reflect the actual operating status of each unit. Under complex conditions such as high-temperature environments or long-cycle charge / discharge cycles, the allocation model deviates from the actual operating conditions of the units, leading to low accuracy in power allocation. Secondly, this technique is poorly adaptable to changes in operating conditions. When the power grid experiences sudden fluctuations or rapid load changes, the adjustment mechanism of the fixed mathematical model lacks flexibility, and power adjustment exhibits significant lag, making it difficult to meet the rapid support requirements of modern power grids for energy storage systems. Under complex conditions, large power allocation deviations or even local overloads often occur, failing to meet the actual requirements of energy storage systems for accurate power control. Summary of the Invention

[0004] This invention provides a power distribution method, apparatus, equipment, medium, and product for energy storage systems, which can improve the accuracy of power distribution in energy storage systems under complex operating conditions.

[0005] In a first aspect, an embodiment of the present invention provides a power distribution method for an energy storage system, comprising: Collect electrical parameters and environmental data of each energy storage unit in the energy storage system; For each energy storage unit, the electrical parameters and environmental data corresponding to the current energy storage unit are input into a preset electrochemical simulation model for simulation to obtain the electrochemical latent characteristics of the current energy storage unit. The electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit are then fused to obtain a state matrix. Based on the obtained health impact parameters corresponding to each energy storage unit, the health status of each energy storage unit is calculated. The state matrix and the health status are input into a preset power allocation model. With the goal of minimizing the total power loss and total health loss of the energy storage system, the model is solved under preset constraints to obtain an initial power allocation scheme. The initial power allocation scheme is then input into a preset correction model for correction to obtain a target power allocation scheme for the energy storage system. Power is then allocated to the energy storage system according to the target power allocation scheme.

[0006] By collecting electrical parameters and environmental data from each energy storage unit in the energy storage system, fundamental data reflecting the electrical characteristics and operating environment of the energy storage units can be obtained without human intervention, ensuring the accuracy of power allocation from the data source. Inputting electrical parameters and environmental data into an electrochemical simulation model yields implicit electrochemical characteristics, overcoming the limitations of traditional sensors that can only monitor electrical parameters such as voltage and current. This allows for the acquisition of implicit information that cannot be directly measured, such as the internal chemical reaction state and thermal runaway risk of the energy storage unit, providing a more comprehensive state basis for power allocation and contributing to the accuracy of subsequent power control. Fusing electrical parameters, environmental data, and implicit electrochemical characteristics to obtain a state matrix integrates physical sensor data with simulation-generated implicit characteristics into a multi-dimensional state representation. This enables subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charge-discharge, reducing the accuracy deviation of power allocation. Calculating the health of each energy storage unit based on health impact parameters systematically integrates key factors related to the unit's health, including those that cannot be directly measured. The health status of the energy storage system is quantified into a calculable health index, providing precise individualized constraints for subsequent adjustments and improving the accuracy of power allocation. The state matrix and health index are input into the power allocation model, and the initial power allocation scheme is obtained under preset constraints with the goal of minimizing total power loss and total health loss. This allows for the integration of inter-unit power transmission losses and unit health losses into a unified optimization framework, ensuring that the initial scheme prioritizes units with good health while meeting total power requirements, preventing units with poor health from being allocated excessive power, thus reducing the risk of local overload and improving power allocation accuracy. The initial power allocation scheme is then input into a preset correction model for modification, resulting in a target power allocation scheme. Power allocation to the energy storage system is based on this target scheme. Pre-trained correction models for each energy storage unit can be used to individually verify and adjust the initial scheme, identifying and correcting allocations that do not match the actual capacity of the units. This ensures that the allocated power of each unit in the final scheme matches its current operating capacity, fundamentally solving allocation deviations caused by ignoring individual unit differences and ultimately achieving accurate power allocation. This application can improve the accuracy of power allocation in energy storage systems under complex operating conditions.

[0007] Furthermore, the step of inputting each of the state matrices and each of the health values ​​into a preset power allocation model, with the objective of minimizing the total power loss and total health loss of the energy storage system, is performed under preset constraints to obtain an initial power allocation scheme, specifically including: Obtain the grid dispatch power command of the energy storage system; The power grid dispatch power command is parsed to obtain power demand data; Based on the power demand data, a total power deviation threshold is determined; Based on the state matrices described above, several interaction coefficients between the energy storage units are determined. Based on the health status of each energy storage unit, the bias coefficient of each energy storage unit is determined. The total power deviation threshold, each of the interaction coefficients, each of the bias coefficients, and each of the health values ​​are input into the power allocation model to solve the problem using the quantum annealing algorithm, thereby obtaining the initial power allocation scheme that satisfies the minimum total power loss, the minimum total health loss, and the constraints of the energy storage system.

[0008] By inputting the state matrix and health status into the power allocation model, and solving for the initial power allocation scheme under preset constraints with the goal of minimizing total power loss and total health loss, the power transmission loss between units and the health loss of the units themselves can be incorporated into a unified optimization framework. This allows the initial scheme to prioritize the use of units with good health status while meeting the total power demand, and to avoid allocating excessive power to units with poor health status. This reduces the risk of local overload from the source and improves the accuracy of power allocation.

[0009] Furthermore, the initial power allocation scheme obtained by using the quantum annealing algorithm to solve for the minimum total power loss, minimum total health loss, and the constraints of the energy storage system specifically includes: The health status is converted into the spin state of each qubit corresponding to each energy storage unit, and the spin state of each qubit is integrated to obtain the initial qubit spin state vector of the energy storage system. The initial qubit spin state vector is randomly perturbed using the quantum tunneling effect to generate several random qubit spin state vectors. The spin state vectors of each random qubit are screened using the aforementioned constraints to obtain a number of candidate qubit spin state vectors, wherein the constraints include the total power deviation threshold. For each candidate qubit spin state vector, the current candidate qubit spin state vector, each interaction coefficient, and each bias coefficient are substituted into a preset objective function for calculation to obtain the Hamiltonian value corresponding to the current candidate qubit spin state vector. The objective function is constructed using the total power loss and total health loss of the energy storage system. Based on the Hamiltonian values, the spin state vector of the target qubit is determined until the preset iteration termination condition is met, thus obtaining the initial power allocation scheme.

[0010] By converting the health status of each energy storage unit into qubit spin state variables and integrating them into an initial qubit spin state vector, the health status of the energy storage units can be mapped to the quantum computing space, enabling subsequent solutions to make power allocation decisions based on health status. Randomly perturbing the initial qubit spin state vector using the quantum tunneling effect overcomes the limitation of traditional annealing algorithms easily getting trapped in local optima, efficiently searching for more candidate solutions in the solution space. Finally, using constraints to screen the random qubit spin state vector ensures that candidate solutions meet practical operational constraints such as the total power deviation threshold. Infeasible solutions are eliminated; the spin state vectors, interaction coefficients, and bias coefficients of candidate qubits are substituted into the objective function constructed with total power loss and total health loss to calculate the Hamiltonian value. This quantizes the power allocation problem into a computable physical energy value, achieving a unified representation of multi-objective optimization; the spin state vector of the target qubit is determined iteratively based on the Hamiltonian value, which can gradually approach the global optimal solution by minimizing the Hamiltonian. Thus, under the premise of satisfying multiple constraints, the total power loss and total health loss are minimized simultaneously, ensuring the accuracy of the initial power allocation scheme from the algorithm level.

[0011] Furthermore, the step of inputting the initial power allocation scheme into a preset correction model for correction to obtain the target power allocation scheme of the energy storage system specifically includes: The initial power allocation scheme and the health status of each energy storage unit are input into a preset correction model. The initial power allocation value of each energy storage unit in the initial power allocation scheme is calculated using the preset maximum allowable output power value corresponding to each energy storage unit, so as to obtain the power deviation value corresponding to each energy storage unit. The power deviation value and health status of each energy storage unit are respectively input into a preset long short-term memory network for processing to obtain the correction coefficient corresponding to each energy storage unit. The initial power allocation scheme is modified using the aforementioned correction coefficients to obtain the target power allocation scheme.

[0012] This process involves inputting the initial power allocation scheme into a preset correction model for modification, resulting in the target power allocation scheme. The correction model pre-trained locally by each energy storage unit can be used to individually verify and adjust the initial scheme, identify and correct allocations in the initial scheme that do not match the actual capacity of the unit, and ensure that the allocated power of each unit in the final scheme matches its current operating capacity. This fundamentally solves the allocation deviation caused by ignoring individual differences of units, and ultimately achieves accurate power allocation.

[0013] Furthermore, the process of fusing the electrical parameters, environmental data, and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix specifically includes: The electrical parameters, environmental data, and electrochemical latent features of the current energy storage unit are input into a preset deep neural network to integrate the data through the input layer of the deep neural network to obtain an input feature vector. The deep neural network includes the input layer, several hidden layers, and an output layer. The input feature vector is input into each of the hidden layers for processing to obtain intermediate feature vectors; The intermediate feature vector is input into the output layer and linearly mapped to obtain the state matrix.

[0014] This method of fusing electrical parameters, environmental data, and electrochemical latent features to obtain a state matrix integrates physical sensing data with simulation-generated latent features into a multi-dimensional state representation. This allows subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charging and discharging, thereby reducing the accuracy deviation of power allocation.

[0015] Furthermore, the calculation of the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit specifically includes: Obtain the health impact parameters corresponding to each of the energy storage units, wherein the health impact parameters include the depth of charge and discharge, aging characteristic data, and the number of charge and discharge cycles; The charge / discharge depth, aging characteristic data, and charge / discharge cycle count of each energy storage unit are input into a preset health assessment function to calculate the health of each energy storage unit.

[0016] Calculating the health status of each energy storage unit based on health impact parameters in this way can systematically integrate the key influencing factors related to the health status of the unit, quantify the health status that cannot be directly measured into a calculable health status index, provide a precise individualized constraint basis for subsequent corrections, and improve the accuracy of power allocation.

[0017] Secondly, an embodiment of the present invention provides a power distribution device for an energy storage system, comprising a first module, a second module and a third module; The first module is used to collect electrical parameters and environmental data of each energy storage unit in the energy storage system; The second module is used to input the electrical parameters and environmental data corresponding to the current energy storage unit into a preset electrochemical simulation model for simulation of each of the energy storage units, to obtain the electrochemical latent characteristics of the current energy storage unit, and to fuse the electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix; The third module is used to calculate the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit, input each state matrix and each health status into a preset power allocation model, and solve the problem under preset constraints with the goal of minimizing the total power loss and total health loss of the energy storage system to obtain an initial power allocation scheme. The initial power allocation scheme is then input into a preset correction model for correction to obtain a target power allocation scheme for the energy storage system, so as to allocate power to the energy storage system according to the target power allocation scheme.

[0018] The first module collects electrical parameters and environmental data from each energy storage unit in the energy storage system. This allows for the acquisition of fundamental data reflecting the electrical characteristics and operating environment of the energy storage units without human intervention, ensuring the accuracy of power allocation from the data source. The second module inputs the electrical parameters and environmental data into an electrochemical simulation model to obtain electrochemical implicit characteristics. This overcomes the limitations of traditional sensors that can only monitor electrical parameters such as voltage and current, acquiring implicit information such as the internal chemical reaction state and thermal runaway risk of the energy storage unit, which cannot be directly measured. This provides a more comprehensive state basis for power allocation, contributing to the accuracy of subsequent power control. The fusion of electrical parameters, environmental data, and electrochemical implicit characteristics yields a state matrix, integrating physical sensor data with simulation-generated implicit characteristics into a multi-dimensional state representation. This allows subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charge and discharge, reducing the accuracy deviation of power allocation. The third module calculates the health of each energy storage unit based on health impact parameters, systematically integrating key factors related to the health of the unit. The intangible health status is quantified into a calculable health index, providing precise individualized constraints for subsequent adjustments and improving the accuracy of power allocation. The state matrix and health index are input into the power allocation model, and the initial power allocation scheme is obtained under preset constraints with the goal of minimizing total power loss and total health loss. This allows for the inclusion of power transmission losses between units and the health loss of each unit within a unified optimization framework. This ensures that the initial scheme prioritizes units with good health while meeting total power requirements, preventing units with poor health from being allocated excessive power, thus reducing the risk of local overload and improving the accuracy of power allocation. The initial power allocation scheme is then input into a preset correction model for refinement, resulting in a target power allocation scheme. Power allocation to the energy storage system is based on this target scheme. Pre-trained correction models for each energy storage unit can be used to individually verify and adjust the initial scheme, identifying and correcting allocations that do not match the actual capacity of the units. This ensures that the allocated power of each unit in the final scheme matches its current operating capacity, fundamentally solving allocation deviations caused by ignoring individual unit differences and ultimately achieving accurate power allocation.

[0019] Thirdly, another embodiment of the present invention provides a terminal device, including: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction that causes the processor to perform the operation of the power distribution method of the energy storage system.

[0020] Fourthly, another embodiment of the present invention provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus where the computer-readable storage medium is located to perform a power distribution method for an energy storage system.

[0021] Fifthly, another embodiment of the present invention provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements a power distribution method for an energy storage system. Attached Figure Description

[0022] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating one embodiment of a power allocation method for an energy storage system provided in this application; Figure 2 This is a flowchart illustrating steps S201 to S206 provided in this application; Figure 3 This is a schematic diagram of the power distribution device of an energy storage system provided in this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0026] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0029] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0030] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application according to the specific circumstances.

[0031] In the field of power control of energy storage systems, energy storage systems are key equipment for smoothing up new energy fluctuations, supporting microgrids and peak-valley arbitrage. However, existing proportional power allocation technology monitoring is limited to basic parameters such as voltage and current, without incorporating key factors such as temperature and aging. This results in low allocation accuracy under complex operating conditions, and the fixed mathematical model adjustment is lagging, making it difficult to adapt to sudden fluctuations in the power grid. It is prone to power deviation and local overload, and cannot meet the actual needs for precise power allocation.

[0032] See Figure 1 In order to improve the accuracy of power distribution of energy storage systems under complex operating conditions, an embodiment of the present invention provides a power distribution method for energy storage systems, including steps S101 to S103. Step S101: Collect electrical parameters and environmental data of each energy storage unit in the energy storage system; In some embodiments, electrical parameters and environmental data of each energy storage unit in the energy storage system are collected, specifically including: continuously collecting the voltage, current, ambient temperature and ambient humidity of each energy storage unit through voltage sensors, current sensors, temperature sensors and humidity sensors, respectively, to obtain electrical parameters and environmental data, which are then digitized and transmitted by the data acquisition unit.

[0033] It should be noted that electrical parameters include, but are not limited to, voltage and current, and environmental data include, but are not limited to, ambient temperature and humidity.

[0034] Step S102: For each energy storage unit, input the electrical parameters and environmental data corresponding to the current energy storage unit into a preset electrochemical simulation model for simulation to obtain the electrochemical latent characteristics of the current energy storage unit. Then, fuse the electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix. In some embodiments, for each of the energy storage units, the electrical parameters and environmental data corresponding to the current energy storage unit are input into a preset electrochemical simulation model for simulation to obtain the electrochemical latent characteristics of the current energy storage unit. Specifically, this includes: establishing an electrochemical simulation model using COMSOL Multiphysics (a numerical simulation software), which includes parameterized settings for electrode material properties and electrolyte characteristics. The electrode material properties include lithium-ion diffusion coefficient and conductivity, and the electrolyte characteristics include ionic conductivity and viscosity. The voltage, current, ambient temperature, and humidity of each energy storage unit are input into the model as boundary conditions. By solving electrochemical control equations, such as the charge conservation equation, the mass conservation equation, and the Butler-Volmer kinetic equation (equations describing electrode reaction kinetics), the internal chemical reaction state and thermal runaway risk information of each energy storage unit are simulated and generated as electrochemical latent characteristics.

[0035] It should be noted that the generation frequency of the latent features is consistent with the sampling frequency of the physical sensor, which can be set to 1kHz to ensure that the simulation data and the measured data can be aligned and fused in time.

[0036] In some embodiments, fusing the electrical parameters, environmental data, and electrochemical latent features corresponding to the current energy storage unit to obtain a state matrix specifically includes: inputting the electrical parameters, environmental data, and electrochemical latent features of the current energy storage unit into a preset deep neural network, and integrating the data through the input layer of the deep neural network to obtain an input feature vector, wherein the deep neural network includes the input layer, several hidden layers, and an output layer; processing the input feature vector by inputting it into each of the hidden layers to obtain an intermediate feature vector; and linearly mapping the intermediate feature vector into the output layer to obtain the state matrix. Specifically, the electrical parameters, environmental data, and electrochemical latent features of the current energy storage unit are input in parallel to the input layer of the deep neural network. The input layer integrates these data into an input feature vector, the total dimension of which is equal to the sum of the dimensions of the electrical parameters, environmental data, and electrochemical latent features. The first hidden layer uses 64 neurons to perform nonlinear transformation and feature extraction on the input feature vector, outputting a first intermediate feature vector. The second hidden layer uses 32 neurons to further extract features and reduce the dimensionality of the first intermediate feature vector, outputting a second intermediate feature vector. The third hidden layer uses 16 neurons to abstract and compress the second intermediate feature vector, outputting a third intermediate feature vector. The output layer uses a linear mapping function to map the third intermediate feature vector into a state matrix of the target dimension.

[0037] In some embodiments, the construction process of a deep neural network specifically includes: the deep neural network adopts a 3-layer hidden layer structure, including an input layer and an output layer; the number of neurons in the input layer is equal to the sum of the electrical parameter dimension, the environmental data dimension, and the electrochemical feature dimension; the number of neurons in the three hidden layers are set to 64, 32, and 16, respectively; the output layer dimension is equal to the state matrix dimension; all hidden layers have an activation function, and the activation function is ReLU.

[0038] In some embodiments, the training process of the deep neural network specifically includes: collecting full life-cycle operation data of the energy storage unit as a training sample set, totaling no less than 1000 sets; each training sample set contains input data and label data, the input data being electrical parameters, environmental data, and latent features generated by electrochemical simulation collected at the same time, and the label data being the labeled state matrix corresponding to that time; before training, normalizing all input data to map the data range of each dimension to the interval between zero and one, so as to eliminate the influence between different dimensions; during training, the normalized training samples are sequentially input into the deep neural network, and the mean square error between the network output and the label data is calculated as the loss function value; the backpropagation algorithm is used to calculate the gradient of the loss function with respect to the weights of each layer of the network, and the Adam optimization algorithm is used to update the network weights; each iteration uses a mini-batch training method, and after each iteration of the full training data is completed, the network performance is evaluated on the validation set. When the loss function value of the validation set no longer decreases for 10 consecutive iterations, training is stopped and the current network parameters are saved as the finally trained deep neural network.

[0039] It should be noted that after the deep neural network training is completed, the time synchronization between physical sensor data and electrochemical simulation data is achieved through a network time protocol, with the time deviation controlled within 1 millisecond. When the deviation between physical sensor data and simulation data exceeds 10%, a weighted fusion strategy is activated, with the weight of physical data set to 0.7 and the weight of simulation data set to 0.3, and a sensor calibration reminder is triggered. When the electrochemical simulation process malfunctions, it can automatically switch to a physical sensor data priority mode, generating a simplified state matrix based solely on voltage, current, and environmental data.

[0040] This method of fusing electrical parameters, environmental data, and electrochemical latent features to obtain a state matrix integrates physical sensing data with simulation-generated latent features into a multi-dimensional state representation. This allows subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charging and discharging, thereby reducing the accuracy deviation of power allocation.

[0041] Step S103: Based on the obtained health impact parameters corresponding to each energy storage unit, calculate the health status of each energy storage unit, input each state matrix and each health status into a preset power allocation model, and solve the problem under preset constraints with the goal of minimizing the total power loss and total health loss of the energy storage system to obtain an initial power allocation scheme. Then, input the initial power allocation scheme into a preset correction model for correction to obtain the target power allocation scheme of the energy storage system, and allocate power to the energy storage system according to the target power allocation scheme.

[0042] In some embodiments, calculating the health level of each energy storage unit based on the acquired health impact parameters of each energy storage unit specifically includes: acquiring the health impact parameters of each energy storage unit, wherein the health impact parameters include charge / discharge depth, aging characteristic data, and charge / discharge cycle count; inputting the charge / discharge depth, aging characteristic data, and charge / discharge cycle count of each energy storage unit into a preset health evaluation function to calculate the health level of each energy storage unit. Specifically, the health evaluation function models the health level of the energy storage unit based on the biological immune system mechanism. This function takes the charge / discharge cycle count, aging characteristic data, and charge / discharge depth as inputs, wherein the aging characteristic data includes aging-related information such as capacity decay rate and internal resistance change; the health evaluation function comprehensively analyzes the unit's usage history reflected by the charge / discharge cycle count, the capacity decay and internal resistance change trends reflected by the aging characteristic data, and the impact of charge / discharge depth on electrode material aging, and outputs a health index that quantitatively assesses the health status of the unit; this health index is used to characterize the current health status of the energy storage unit, with a higher value indicating a healthier unit and a lower value indicating a more severe degree of aging.

[0043] It should be noted that the health assessment function is based on the biological immune system modeling the health of energy storage units, and is trained using machine learning regression algorithms, such as support vector regression. The model is trained using a large amount of historical data to learn the mapping relationship between charge / discharge cycle count, capacity decay rate and internal resistance changes in aging characteristic data, and charge / discharge depth and health. When the health index is lower than the preset health threshold, a gene repair program is triggered to adjust the charge / discharge depth. This can be done by adjusting the charge / discharge depth to protect battery electrode materials and slow down aging, using linear or non-linear adjustment strategies. For example, the charge / discharge depth is adjusted proportionally according to the degree to which the health deviates from the threshold. After the gene repair program is activated, the adjustment range of the charge / discharge depth does not exceed 20% of the original depth per cycle, and the number of adjustments per day is ≤3. The health threshold is dynamically adjusted according to the type of energy storage unit, such as 80% for lithium batteries and 70% for lead-acid batteries.

[0044] In some embodiments, the formula for calculating the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit specifically includes: Health assessment formula: ; Gene repair formula: ; In the formula, Indicators representing the health status of energy storage units; Indicates the number of charge / discharge cycles, reflecting the cell's usage history; The vector formed from aging characteristic data contains aging-related information such as capacity decay rate and internal resistance change; Indicates the depth of charge / discharge; h is the health assessment function; This indicates the set health threshold. Below this value, the gene repair process is triggered; This indicates the adjusted new depth of charge and discharge; represents the original charge / discharge depth; m represents the gene repair function.

[0045] For example, before calculating health status, the process includes: drawing on the mechanisms of the biological immune system to establish a health gene library for the energy storage unit, that is, encoding features such as charge-discharge cycle data into DNA sequences to simulate the impact of power strategies on lifespan; the DNA sequence length of the health gene library is 256 bits, of which the first 64 bits encode the number of charge-discharge cycles, the middle 128 bits encode aging characteristic data, and the last 64 bits encode the charge-discharge depth; when it is necessary to simulate the impact of power strategies on lifespan in the future, the Monte Carlo method can be used to perform more than 1000 iterative simulations.

[0046] Calculating the health status of each energy storage unit based on health impact parameters in this way can systematically integrate the key influencing factors related to the health status of the unit, quantify the health status that cannot be directly measured into a calculable health status index, provide a precise individualized constraint basis for subsequent corrections, and improve the accuracy of power allocation.

[0047] See Figure 2 In some embodiments, the step of inputting each of the state matrices and each of the health values ​​into a preset power allocation model, with the goal of minimizing the total power loss and total health loss of the energy storage system, is to solve the problem under preset constraints to obtain an initial power allocation scheme, specifically including steps S201 to S206. Step S201: Obtain the grid dispatch power command of the energy storage system; In some embodiments, obtaining the grid dispatch power command of the energy storage system specifically includes: receiving the grid dispatch power command from the grid dispatch center through a communication interface.

[0048] Step S202: Parse the power grid dispatching command to obtain power demand data; In some embodiments, parsing the power grid dispatch power command to obtain power demand data specifically includes: extracting the total power target value that the energy storage system needs to provide in the next dispatch cycle from the power grid dispatch power command.

[0049] Step S203: Based on the power demand data, determine the total power deviation threshold; In some embodiments, determining a total power deviation threshold based on the power demand data specifically includes: setting a total power deviation threshold according to the total power target value in the power demand data, such as taking 5% of the target value as the threshold.

[0050] It should be noted that the total power deviation threshold represents the maximum allowable deviation between the total output power of the energy storage cluster and the power demand of the grid.

[0051] Step S204: Determine several interaction coefficients between each energy storage unit based on each state matrix. In some embodiments, based on the state matrices, several interaction coefficients between the energy storage units are determined, specifically including: extracting the current operating state information of each energy storage unit from each state matrix, and combining the obtained experimental data on power transmission loss between energy storage units to determine the interaction coefficient between every two energy storage units.

[0052] It should be noted that the interaction coefficient ranges from -0.2 to +0.2, and this coefficient reflects the correlation constraints in the power distribution between units.

[0053] Step S205: Based on the health status of each energy storage unit, determine the bias coefficient of each energy storage unit; In some embodiments, the bias coefficient of each energy storage unit is determined based on the health level of each unit. Specifically, this includes: setting an initial bias coefficient and determining the final bias coefficient according to the rule that the health level is positively correlated with the bias coefficient. For every 10% increase in health level, the bias coefficient increases by 0.1.

[0054] It should be noted that the bias coefficient reflects the impact of the energy storage unit's own characteristics on power allocation priority.

[0055] Step S206: Input the total power deviation threshold, each of the interaction coefficients, each of the bias coefficients, and each of the health values ​​into the power allocation model, and solve it using the quantum annealing algorithm to obtain the initial power allocation scheme that satisfies the minimum total power loss, the minimum total health loss, and the constraints of the energy storage system.

[0056] In some embodiments, inputting the total power deviation threshold, each of the interaction coefficients, each of the bias coefficients, and each of the health status into the power allocation model specifically includes: inputting the total power deviation threshold, each of the interaction coefficients, each of the bias coefficients, and each of the health status into the power allocation model for further solution.

[0057] It should be noted that the power allocation model involves multiple constraints, including the maximum charging and discharging power limit of a single energy storage unit and the deviation between the total power of the energy storage cluster and the grid demand not exceeding a total power deviation threshold.

[0058] In some embodiments, the step of using the quantum annealing algorithm to obtain the initial power allocation scheme that minimizes the total power loss and total health loss of the energy storage system and satisfies the constraints specifically includes: converting each health level into a qubit spin state quantity corresponding to each energy storage unit, and integrating the qubit spin state quantities to obtain the initial qubit spin state vector of the energy storage system; randomly perturbing the initial qubit spin state vector using the quantum tunneling effect to generate several random qubit spin state vectors; and applying the constraints to each random qubit spin state vector... The process involves screening to obtain several candidate qubit spin state vectors, where the constraint condition includes the total power deviation threshold. For each candidate qubit spin state vector, the current candidate qubit spin state vector, each interaction coefficient, and each bias coefficient are substituted into a preset objective function for calculation to obtain the Hamiltonian value corresponding to the current candidate qubit spin state vector. The objective function is constructed using the total power loss and total health loss of the energy storage system. Based on each Hamiltonian value, the target qubit spin state vector is determined until a preset iteration termination condition is met, resulting in the initial power allocation scheme.Specifically, the mapping rule between qubit spin states and power allocation strategy is as follows: spin state +1 corresponds to the unit outputting 100% of the base power, and spin state -1 corresponds to outputting 0% of the base power, i.e., shutdown. Intermediate states dynamically adjust the output ratio through quantum tunneling probability. At the start of the solution process, the health status of each energy storage unit is converted into corresponding qubit spin state quantities according to the rules. Units with higher health status are mapped to a +1 state, and units with lower health status are mapped to a -1 state. The qubit spin state quantities of all units are integrated into an initial qubit spin state vector. The initial qubit spin state vector is randomly perturbed using the quantum tunneling effect, enabling the system to cross energy barriers in the solution space, thereby escaping the local optimum region and generating several random qubit spin state vectors. Constraints are then applied to the generated random qubit spin states... Vectors are filtered, eliminating those that do not satisfy the multiple constraints problem, and retaining those that do satisfy the constraints as candidate qubit spin state vectors. For each candidate qubit spin state vector, it is substituted with each interaction coefficient and each bias coefficient into the objective function constructed based on the total power loss and total health loss of the energy storage system to calculate the Hamiltonian value corresponding to the candidate vector. The Hamiltonian values ​​corresponding to each candidate vector are compared, and the candidate vector with the smallest Hamiltonian value (to satisfy the objectives of minimizing total power loss and total health loss) is selected as the target qubit spin state vector for the current round. It is determined whether the preset iteration termination condition is met. If not, the quantum annealing algorithm steps are continued based on the current target vector. If the condition is met, the power allocation strategy corresponding to the current target qubit spin state vector is output as the initial power allocation scheme.

[0059] It should be noted that the initial temperature of quantum annealing can be set to 1000 Kelvin, the annealing rate is 0.95 per step, and the number of iterations is not less than 1000. The preset iteration termination condition can be set to the Hamiltonian value changing less than a preset threshold in multiple consecutive iterations, or the number of iterations reaching the preset maximum number of iterations.

[0060] In some embodiments, the formula for solving the initial power allocation scheme using the quantum annealing algorithm to minimize the total power loss and total health loss of the energy storage system while satisfying the constraints specifically includes: Objective function: ; In the formula, Represents the spin state vector of a quantum bit; This represents the spin state of the qubit corresponding to the i-th energy storage unit. The value can be +1 or -1; H(x) is the objective Hamiltonian, representing the objective function of the power allocation problem. The optimal power allocation scheme is obtained by minimizing H(x) by adjusting x; N represents the total number of energy storage units. This represents the interaction coefficient between the i-th and j-th energy storage units; This represents the bias coefficient relative to the i-th energy storage unit.

[0061] It should be noted that the first term in the objective function Using the interaction coefficient To quantify the power transmission losses between energy storage units, The value is determined based on experimental data of power transmission loss and the state matrix; the greater the loss, the higher the value. The larger the value, the better it guides the algorithm to choose an allocation scheme with lower power transmission loss when minimizing the objective function; the second term in the objective function Through bias coefficient To quantify the health and loss of each energy storage unit. The value is positively correlated with health; the higher the health, the better. The larger the value, the more the algorithm is guided to prioritize power allocation to units in good health when minimizing the objective function; by minimizing the objective function, the total power loss and total health loss of the energy storage system can be minimized simultaneously.

[0062] By converting the health status of each energy storage unit into qubit spin state variables and integrating them into an initial qubit spin state vector, the health status of the energy storage units can be mapped to the quantum computing space, enabling subsequent solutions to make power allocation decisions based on health status. Randomly perturbing the initial qubit spin state vector using the quantum tunneling effect overcomes the limitation of traditional annealing algorithms easily getting trapped in local optima, efficiently searching for more candidate solutions in the solution space. Finally, using constraints to screen the random qubit spin state vector ensures that candidate solutions meet practical operational constraints such as the total power deviation threshold. Infeasible solutions are eliminated; the spin state vectors, interaction coefficients, and bias coefficients of candidate qubits are substituted into the objective function constructed with total power loss and total health loss to calculate the Hamiltonian value. This quantizes the power allocation problem into a computable physical energy value, achieving a unified representation of multi-objective optimization; the spin state vector of the target qubit is determined iteratively based on the Hamiltonian value, which can gradually approach the global optimal solution by minimizing the Hamiltonian. Thus, under the premise of satisfying multiple constraints, the total power loss and total health loss are minimized simultaneously, ensuring the accuracy of the initial power allocation scheme from the algorithm level.

[0063] By inputting the state matrix and health status into the power allocation model, and solving for the initial power allocation scheme under preset constraints with the goal of minimizing total power loss and total health loss, the power transmission loss between units and the health loss of the units themselves can be incorporated into a unified optimization framework. This allows the initial scheme to prioritize the use of units with good health status while meeting the total power demand, thus avoiding the allocation of excessive power to units with poor health status. This reduces the risk of local overload from the source and improves the accuracy of power allocation.

[0064] In some embodiments, the step of inputting the initial power allocation scheme into a preset correction model for correction to obtain the target power allocation scheme of the energy storage system specifically includes: inputting the initial power allocation scheme and the health status of each energy storage unit into the preset correction model, so as to calculate the deviation of the initial power allocation value of each energy storage unit in the initial power allocation scheme using the preset maximum allowable output power value corresponding to each energy storage unit, and obtaining the power deviation value corresponding to each energy storage unit; inputting the power deviation value and the health status corresponding to each energy storage unit into a preset long short-term memory network for processing to obtain the correction coefficient corresponding to each energy storage unit; and using the correction coefficients to correct the initial power allocation scheme to obtain the target power allocation scheme. Specifically, the initial power allocation scheme is input into a preset correction model to determine the initial power allocation value for each energy storage unit, and the maximum allowable output power value is determined based on the health status of each energy storage unit. The initial power allocation value is compared with the maximum allowable output power value, and the difference between the two is calculated as the power deviation value. The power deviation value and the health status are input into a preset long short-term memory network, which performs time-dependent processing on the data and outputs the correction coefficients corresponding to each energy storage unit. The initial power allocation values ​​of the corresponding energy storage units in the initial power allocation scheme are corrected using the correction coefficients to obtain the target power allocation scheme.

[0065] For example, the construction and training process of a Long Short-Term Memory (LSTM) network specifically includes: using a two-layer LSM network, setting the input as power deviation value and health level, and the output as correction coefficient; distributing the network to each energy storage unit, with each energy storage unit collecting local operating data as training samples to construct a local training dataset, with a training data volume of no less than 100 sets per unit; each energy storage unit training the network locally based on the local training dataset, with 10 training iterations per round; after completing local training, each energy storage unit uploading its local model parameters to the server, with a parameter upload frequency of once every 5 minutes; the server using a federated averaging algorithm to aggregate the parameters uploaded by each unit to form a global model; during global model aggregation, the FedAvg algorithm is used, with a learning rate set to 0.01, a batch size of 32, and the parameter weight of a single energy storage unit not exceeding 15% of the total weight, and the aggregation iteration of the global model being no less than 5 times per hour.

[0066] For example, the relevant formulas for the construction and training process of Long Short-Term Memory (LSTM) networks specifically include: Aggregate formula: ; In the formula, This represents the global model parameters after the (t+1)th training round; K represents the number of energy storage units. This represents the size of the local training dataset for the k-th energy storage unit; This represents the local model parameters obtained by the k-th energy storage unit after the t-th round of training.

[0067] It should be noted that each energy storage unit and the server use blockchain technology to achieve encrypted transmission of parameters, forming a decentralized federated learning network. When a failure occurs in the power allocation process, each energy storage unit activates a local emergency power allocation scheme, which allocates power based on historical global model parameters until the network is restored.

[0068] This process involves inputting the initial power allocation scheme into a preset correction model for modification, resulting in the target power allocation scheme. The correction model pre-trained locally by each energy storage unit can be used to individually verify and adjust the initial scheme, identify and correct allocations in the initial scheme that do not match the actual capacity of the unit, and ensure that the allocated power of each unit in the final scheme matches its current operating capacity. This fundamentally solves the allocation deviation caused by ignoring individual differences of units, and ultimately achieves accurate power allocation.

[0069] In some embodiments, power allocation of the energy storage system according to the target power allocation scheme specifically includes: sending the allocated power value of each energy storage unit in the target power allocation scheme to the controller of the corresponding energy storage unit, and the controller adjusting the output power or charging power of the energy storage unit according to the received allocated power value, so that each energy storage unit operates according to the target power allocation scheme.

[0070] For example, during the power allocation process, power-related data is acquired in real time. When a power deviation is detected (such as a deviation between the actual allocated power and the scheme), the circuit connection can be autonomously reconfigured through a superconducting switch and an adaptive topology circuit: the superconducting zero-resistance characteristic is used to quickly compensate for power fluctuations, the response time of the superconducting switch is no more than 10 microseconds, and the reconfiguration time of the adaptive topology circuit is no more than 1 millisecond; the superconducting switch uses YBCO high-temperature superconducting material, and the trigger threshold is that it conducts when the detected power deviation reaches 5% of the rated power; the reconfiguration logic of the adaptive topology circuit is as follows: when the power deviation is positive, i.e., the output is insufficient, it switches from a series topology to a parallel topology to increase the total power; when the power deviation is negative, i.e., the output is excessive, it switches to a star topology to shunt losses; during the adjustment process, the adjustment action, power deviation correction data, and circuit topology change information are recorded in real time, and synchronous feedback is used for health assessment and state mapping update.

[0071] For example, when the microgrid experiences a sudden load increase of 20%, the workflow is as follows: A sudden load change is detected using voltage and current sensors. Electrochemical simulation is used to generate a latent feature indicating a 3% increase in internal resistance, forming a multi-dimensional state matrix. Based on this state matrix and a health threshold of 80%, a preliminary scheme is generated to increase the output of each of the three energy storage units by 7%. Verification using a modified model reveals that unit 1's health is only 75%, insufficient to withstand a 7% increase. Therefore, a correction instruction is fed back, revising the scheme to increase unit 1 by 5%, and units 2 and 3 by 8% each. A superconducting switch is used to switch to a parallel topology, executing the revised instruction and recording data such as the 2% deviation correction for unit 1. Based on the feedback data, the gene bank is updated, encoding the charge-discharge cycle characteristics of unit 1 into a DNA sequence corresponding to a higher cycle count, prioritizing a reduction in its load during the next allocation round.

[0072] For example, when the photovoltaic output power drops by 30% due to cloud cover, the workflow is as follows: Voltage fluctuations and a sudden increase in internal temperature of 2 degrees Celsius are detected within 10 milliseconds, generating a state matrix containing information such as the risk level of thermal runaway; based on this state matrix, three energy storage units are initially allocated 10% power each; through model verification, it is found that unit 3 is located in a high-temperature region and has a health level of 75%, so the scheme is revised to supplement unit 1 and unit 2 with 12% each, and unit 3 with 6%; the backup branch is quickly activated via a superconducting switch, completing power compensation within 3 milliseconds, and the data showing a 2% increase in power deviation correction is recorded; the high temperature and high load characteristics of unit 3 are encoded into a new DNA sequence, and its load is reduced by 10% in the next allocation round.

[0073] For example, in a scenario where charging occurs during off-peak hours when electricity prices are low, the workflow is as follows: Data on grid voltage stability and ambient temperature (15 degrees Celsius) are integrated to generate the implicit characteristic of the optimal charging efficiency range; combined with the constraint that shallow charging and discharging are beneficial to lifespan, a preliminary scheme is generated to charge each energy storage unit to 80% capacity; after aggregating the local models of each energy storage unit, the scheme is revised so that units with a health level above 90% are charged to 85%, and units with a health level between 70% and 90% are charged to 75%; the charging current is adjusted according to the revised instructions, and data showing an average charging efficiency improvement of 5% compared to the full charge / discharge mode is recorded.

[0074] By collecting electrical parameters and environmental data from each energy storage unit in the energy storage system, fundamental data reflecting the electrical characteristics and operating environment of the energy storage units can be obtained without human intervention, ensuring the accuracy of power allocation from the data source. Inputting electrical parameters and environmental data into an electrochemical simulation model yields implicit electrochemical characteristics, overcoming the limitations of traditional sensors that can only monitor electrical parameters such as voltage and current. This allows for the acquisition of implicit information that cannot be directly measured, such as the internal chemical reaction state and thermal runaway risk of the energy storage unit, providing a more comprehensive state basis for power allocation and contributing to the accuracy of subsequent power control. Fusing electrical parameters, environmental data, and implicit electrochemical characteristics to obtain a state matrix integrates physical sensor data with simulation-generated implicit characteristics into a multi-dimensional state representation. This enables subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charge-discharge, reducing the accuracy deviation of power allocation. Calculating the health of each energy storage unit based on health impact parameters systematically integrates key factors related to the unit's health, including those that cannot be directly measured. The health status of the energy storage system is quantified into a calculable health index, providing precise individualized constraints for subsequent adjustments and improving the accuracy of power allocation. The state matrix and health index are input into the power allocation model, and the initial power allocation scheme is obtained under preset constraints with the goal of minimizing total power loss and total health loss. This allows for the integration of inter-unit power transmission losses and unit health losses into a unified optimization framework, ensuring that the initial scheme prioritizes units with good health while meeting total power requirements, preventing units with poor health from being allocated excessive power, thus reducing the risk of local overload and improving power allocation accuracy. The initial power allocation scheme is then input into a preset correction model for modification, resulting in a target power allocation scheme. Power allocation to the energy storage system is based on this target scheme. Pre-trained correction models for each energy storage unit can be used to individually verify and adjust the initial scheme, identifying and correcting allocations that do not match the actual capacity of the units. This ensures that the allocated power of each unit in the final scheme matches its current operating capacity, fundamentally solving allocation deviations caused by ignoring individual unit differences and ultimately achieving accurate power allocation. This application can improve the accuracy of power allocation in energy storage systems under complex operating conditions.

[0075] See Figure 3 Based on the above method embodiments, corresponding device embodiments are provided; One embodiment of the present invention provides a power distribution device for an energy storage system, including a first module 100, a second module 200 and a third module 300; The first module 100 is used to collect electrical parameters and environmental data of each energy storage unit in the energy storage system; The second module 200 is used to input the electrical parameters and environmental data corresponding to the current energy storage unit into a preset electrochemical simulation model for simulation of each of the energy storage units, to obtain the electrochemical latent characteristics of the current energy storage unit, and to fuse the electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix; The third module 300 is used to calculate the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit, input each state matrix and each health status into a preset power allocation model, and solve the problem under preset constraints with the goal of minimizing the total power loss and total health loss of the energy storage system to obtain an initial power allocation scheme. The initial power allocation scheme is then input into a preset correction model for correction to obtain a target power allocation scheme for the energy storage system, so as to allocate power to the energy storage system according to the target power allocation scheme.

[0076] The first module collects electrical parameters and environmental data from each energy storage unit in the energy storage system. This allows for the acquisition of fundamental data reflecting the electrical characteristics and operating environment of the energy storage units without human intervention, ensuring the accuracy of power allocation from the data source. The second module inputs the electrical parameters and environmental data into an electrochemical simulation model to obtain electrochemical implicit characteristics. This overcomes the limitations of traditional sensors that can only monitor electrical parameters such as voltage and current, acquiring implicit information such as the internal chemical reaction state and thermal runaway risk of the energy storage unit, which cannot be directly measured. This provides a more comprehensive state basis for power allocation, contributing to the accuracy of subsequent power control. The fusion of electrical parameters, environmental data, and electrochemical implicit characteristics yields a state matrix, integrating physical sensor data with simulation-generated implicit characteristics into a multi-dimensional state representation. This allows subsequent power allocation to comprehensively consider the external electrical characteristics and internal chemical state of the energy storage unit, accurately reflecting the actual operating state of each unit under complex conditions such as high temperature, high humidity, and long-cycle charge and discharge, reducing the accuracy deviation of power allocation. The third module calculates the health of each energy storage unit based on health impact parameters, systematically integrating key factors related to the health of the unit. The intangible health status is quantified into a calculable health index, providing precise individualized constraints for subsequent adjustments and improving the accuracy of power allocation. The state matrix and health index are input into the power allocation model, and the initial power allocation scheme is obtained under preset constraints with the goal of minimizing total power loss and total health loss. This allows for the inclusion of power transmission losses between units and the health loss of each unit within a unified optimization framework. This ensures that the initial scheme prioritizes units with good health while meeting total power requirements, preventing units with poor health from being allocated excessive power, thus reducing the risk of local overload and improving the accuracy of power allocation. The initial power allocation scheme is then input into a preset correction model for refinement, resulting in a target power allocation scheme. Power allocation to the energy storage system is based on this target scheme. Pre-trained correction models for each energy storage unit can be used to individually verify and adjust the initial scheme, identifying and correcting allocations that do not match the actual capacity of the units. This ensures that the allocated power of each unit in the final scheme matches its current operating capacity, fundamentally solving allocation deviations caused by ignoring individual unit differences and ultimately achieving accurate power allocation.

[0077] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the power distribution method of an energy storage system provided by any of the above-described method embodiments of the present invention.

[0078] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0079] Based on the above-described embodiment of the power allocation method for an energy storage system, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power allocation method for an energy storage system according to any embodiment of the present invention.

[0080] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0081] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0082] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0083] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a power distribution method for an energy storage system as described in any of the above-described method embodiments of the present invention.

[0084] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0085] Based on the above-described method embodiments, another embodiment of the present invention provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements a power distribution method for an energy storage system.

[0086] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A power distribution method for an energy storage system, characterized in that, include: Collect electrical parameters and environmental data of each energy storage unit in the energy storage system; For each energy storage unit, the electrical parameters and environmental data corresponding to the current energy storage unit are input into a preset electrochemical simulation model for simulation to obtain the electrochemical latent characteristics of the current energy storage unit. The electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit are then fused to obtain a state matrix. Based on the obtained health impact parameters corresponding to each energy storage unit, the health status of each energy storage unit is calculated. The state matrix and the health status are input into a preset power allocation model. With the goal of minimizing the total power loss and total health loss of the energy storage system, the model is solved under preset constraints to obtain an initial power allocation scheme. The initial power allocation scheme is then input into a preset correction model for correction to obtain a target power allocation scheme for the energy storage system. Power is then allocated to the energy storage system according to the target power allocation scheme.

2. The power distribution method for an energy storage system as described in claim 1, characterized in that, The process involves inputting each of the state matrices and health parameters into a preset power allocation model, aiming to minimize the total power loss and total health loss of the energy storage system, and solving the problem under preset constraints to obtain an initial power allocation scheme. Specifically, this includes: Obtain the grid dispatch power command of the energy storage system; The power grid dispatch power command is parsed to obtain power demand data; Based on the power demand data, a total power deviation threshold is determined; Based on the state matrices described above, several interaction coefficients between the energy storage units are determined. Based on the health status of each energy storage unit, the bias coefficient of each energy storage unit is determined. The total power deviation threshold, each of the interaction coefficients, each of the bias coefficients, and each of the health values ​​are input into the power allocation model to solve the problem using the quantum annealing algorithm, thereby obtaining the initial power allocation scheme that satisfies the minimum total power loss, the minimum total health loss, and the constraints of the energy storage system.

3. The power distribution method for an energy storage system as described in claim 2, characterized in that, The initial power allocation scheme obtained by solving the problem using the quantum annealing algorithm, which satisfies the minimum total power loss, the minimum total health loss, and the constraints of the energy storage system, specifically includes: The health status is converted into the spin state of each qubit corresponding to each energy storage unit, and the spin state of each qubit is integrated to obtain the initial qubit spin state vector of the energy storage system. The initial qubit spin state vector is randomly perturbed using the quantum tunneling effect to generate several random qubit spin state vectors. The spin state vectors of each random qubit are screened using the aforementioned constraints to obtain a number of candidate qubit spin state vectors, wherein the constraints include the total power deviation threshold. For each candidate qubit spin state vector, the current candidate qubit spin state vector, each interaction coefficient, and each bias coefficient are substituted into a preset objective function for calculation to obtain the Hamiltonian value corresponding to the current candidate qubit spin state vector. The objective function is constructed using the total power loss and total health loss of the energy storage system. Based on the Hamiltonian values, the spin state vector of the target qubit is determined until the preset iteration termination condition is met, thus obtaining the initial power allocation scheme.

4. The power distribution method for an energy storage system as described in claim 1, characterized in that, The step of inputting the initial power allocation scheme into a preset correction model for correction to obtain the target power allocation scheme of the energy storage system specifically includes: The initial power allocation scheme and the health status of each energy storage unit are input into a preset correction model. The initial power allocation value of each energy storage unit in the initial power allocation scheme is calculated using the preset maximum allowable output power value corresponding to each energy storage unit, so as to obtain the power deviation value corresponding to each energy storage unit. The power deviation value and health status of each energy storage unit are respectively input into a preset long short-term memory network for processing to obtain the correction coefficient corresponding to each energy storage unit. The initial power allocation scheme is modified using the aforementioned correction coefficients to obtain the target power allocation scheme.

5. The power distribution method for an energy storage system as described in claim 1, characterized in that, The process of fusing the electrical parameters, environmental data, and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix specifically includes: The electrical parameters, environmental data, and electrochemical latent features of the current energy storage unit are input into a preset deep neural network to integrate the data through the input layer of the deep neural network to obtain an input feature vector. The deep neural network includes the input layer, several hidden layers, and an output layer. The input feature vector is input into each of the hidden layers for processing to obtain intermediate feature vectors; The intermediate feature vector is input into the output layer and linearly mapped to obtain the state matrix.

6. The power distribution method for an energy storage system as described in claim 1, characterized in that, The calculation of the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit specifically includes: Obtain the health impact parameters corresponding to each of the energy storage units, wherein the health impact parameters include the depth of charge and discharge, aging characteristic data, and the number of charge and discharge cycles; The charge / discharge depth, aging characteristic data, and charge / discharge cycle count of each energy storage unit are input into a preset health assessment function to calculate the health of each energy storage unit.

7. A power distribution device for an energy storage system, characterized in that, It includes Module 1, Module 2, and Module 3; The first module is used to collect electrical parameters and environmental data of each energy storage unit in the energy storage system; The second module is used to input the electrical parameters and environmental data corresponding to the current energy storage unit into a preset electrochemical simulation model for simulation of each of the energy storage units, to obtain the electrochemical latent characteristics of the current energy storage unit, and to fuse the electrical parameters, environmental data and electrochemical latent characteristics corresponding to the current energy storage unit to obtain a state matrix; The third module is used to calculate the health status of each energy storage unit based on the acquired health impact parameters corresponding to each energy storage unit, input each state matrix and each health status into a preset power allocation model, and solve the problem under preset constraints with the goal of minimizing the total power loss and total health loss of the energy storage system to obtain an initial power allocation scheme. The initial power allocation scheme is then input into a preset correction model for correction to obtain a target power allocation scheme for the energy storage system, so as to allocate power to the energy storage system according to the target power allocation scheme.

8. A terminal device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation of the power distribution method of the energy storage system as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus containing the computer-readable storage medium to perform the power distribution method of the energy storage system as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the communication device, the power distribution method of the energy storage system as described in any one of claims 1 to 6 is implemented.

Images