Electric energy dispatching method based on energy storage system

Abstract

The invention relates to the technical field of electric energy dispatching, in particular to an electric energy dispatching method based on an energy storage system. The method comprises the following steps: deploying a distributed sensor array in a battery module to collect multi-source heterogeneous data, and carrying out standardization processing to generate a standardized energy storage data set; performing tensor completion on the standardized energy storage data set to generate energy storage tensor completion data; time-space dimension analysis is carried out on the energy storage tensor complemented data, cross-modal feature fusion is carried out, and an energy storage cross-modal feature space is generated; therefore, through structural processing of multi-dimensional fusion data and dynamic feedback optimization of the digital twinborn model, the problems that traditional cold chain storage environment prediction lags behind and the model adaptability is poor are solved, and the precision of temperature prediction and the intelligent level of storage management are improved.

Description

Technical Field

[0001] This invention relates to the field of power dispatching technology, and in particular to a power dispatching method based on an energy storage system. Background Technology

[0002] Existing energy storage dispatch technologies primarily rely on static strategies or simplified models, failing to effectively consider multi-dimensional factors such as battery aging, frequency response, and dynamic load changes. This results in insufficient dispatch accuracy and adaptability, making it difficult to cope with high-frequency load fluctuations and abnormal grid disturbances. Data acquisition often focuses on a single dimension, such as voltage, current, or SOC, neglecting the comprehensive analysis of multi-modal data such as temperature and polarization losses. Furthermore, traditional filtering and interpolation methods are used to handle missing data and noise, lacking in-depth modeling of data continuity and spatial correlation, thus limiting the precision of dispatch. While digital twin technology has been proposed, it largely remains at the level of visualization and limited prediction, lacking dynamic optimization and adaptive update capabilities, making it difficult to cope with long-term losses and changes in charge and discharge paths. In terms of optimization strategies, existing methods typically focus solely on loss reduction or stability, lacking coordinated control of loss reduction and frequency response, and failing to construct effective game theory models, easily leading to an imbalance between system performance and loss reduction benefits. Summary of the Invention

[0003] Therefore, it is necessary to provide a power dispatching method based on energy storage systems to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a power dispatching method based on an energy storage system is provided, the method comprising the following steps: Step S1: Deploy a distributed sensor array within the battery module to collect multi-source heterogeneous data and perform protocol conversion to obtain an initial multimodal dataset; adjust the noise spectrum filtering threshold of the initial multimodal dataset and perform long-term drift compensation to generate a standardized energy storage dataset. Step S2: Perform tensor completion on the standardized energy storage dataset to generate energy storage tensor completion data; perform time-space dimension analysis on the energy storage tensor completion data and perform cross-modal feature fusion to generate energy storage cross-modal feature space; Step S3: Construct a digital twin of the battery health status based on the energy storage cross-modal feature space to obtain a digital twin model of the energy storage system; generate a game-theoretic scheduling strategy from the digital twin model of the energy storage system to obtain the energy dispatch strategy of the energy storage system. Step S4: Obtain real-time energy storage monitoring data; calculate the execution process deviation based on the real-time energy storage monitoring data and the energy storage system's power dispatch strategy. If the energy storage deviation data is greater than a preset threshold, generate SOC attenuation data; send the SOC attenuation data back to the energy storage system's digital twin model for network parameter optimization, and perform full-process scheduling analysis of the energy storage system to obtain full-process scheduling analysis data; generate a scheduling report based on the full-process scheduling analysis data to obtain a full-process report on energy storage system power dispatch.

[0005] The beneficial effects of this invention lie in the real-time acquisition of multi-dimensional heterogeneous data in a cold chain storage environment through the deployment of a multi-source sensor array. Furthermore, timestamp alignment technology ensures precise synchronization of different data sources in the time dimension, resulting in highly timely and complete timestamp-aligned cold chain storage data. Subsequently, through three-dimensional spatial mapping and the construction of a feature tensor matrix, spatial distribution information of temperature, humidity, and other environmental parameters is integrated to generate a fused feature matrix. This matrix reflects the multi-dimensional dynamic characteristics of the storage environment, providing a structured data foundation for subsequent analysis. Based on this, a pre-defined Long Short-Term Memory (LSTM) network model is used to predict temperature distribution on the fused feature matrix, capturing the time-series characteristics and spatial correlation of temperature changes, generating predicted temperature distribution data for the storage environment. This data is further combined with the fused feature matrix to construct a cold chain causal knowledge graph, revealing the causal relationships between environmental variables. A digital twin model is then constructed based on this graph, enabling multi-dimensional simulation and dynamic modeling of the storage system's state. The digital twin model then undergoes iterative error residual analysis. The system identifies the deviation between model predictions and actual monitoring, generates differential analysis data, and automatically adjusts compensation parameters through a feedback loop to optimize and fine-tune the model, forming a more accurate warehouse optimization digital twin model. Ultimately, this optimized model drives the construction of logistics and warehousing management reports, systematically presenting key indicators such as temperature prediction accuracy, environmental stability, and management efficiency, supporting scientific decision-making and precise scheduling. Overall, this method enhances the environmental perception capability and prediction accuracy of cold chain warehousing systems through a closed loop of data acquisition, synchronization, fusion, prediction, and feedback optimization, achieving dynamic adaptability and continuous optimization of the model. Therefore, this invention, through structured processing of multi-dimensional fused data and dynamic feedback optimization of the digital twin model, solves the problems of lagging environmental predictions and poor model adaptability in traditional cold chain warehousing, improving the accuracy of temperature prediction and the level of intelligent warehousing management.

[0006] Preferably, step S1 includes the following steps: Step S11: Deploy voltage and temperature sensors within the battery module to collect individual cell temperature and terminal voltage data at a frequency of 10Hz. The voltage sensor has an accuracy of 0.1%, and the temperature sensor has an accuracy of 0.5℃. Step S12: Deploy a high-frequency current sensor with a bandwidth of 20kHz on the DC side of the energy storage converter, and collect the charging and discharging current waveform at a frequency of 1kHz to generate energy storage charging and discharging waveform data. Step S13: Deploy a synchronous phase measurement unit at the grid connection point to collect grid frequency, voltage phase angle and harmonic distortion rate at a frequency of 50Hz, and generate raw grid phase data; Step S14: Convert the individual cell temperature data, terminal voltage data, energy storage charging and discharging waveform data, and grid phase raw data through a multi-protocol adapter to generate an initial multi-mode dataset; Step S15: Adjust the noise spectrum filtering threshold of the initial multimodal dataset and perform long-term drift compensation to generate a standardized energy storage dataset.

[0007] This invention achieves high-frequency (10Hz) dynamic acquisition of the operating status of individual cells by deploying high-precision voltage and temperature sensors within the battery module. This ensures that minute characteristics of temperature gradient changes and voltage fluctuations within each time slice are accurately captured, effectively supporting cell thermal management and aging assessment. Simultaneously, a high-frequency current sensor with a bandwidth of 20kHz is deployed on the DC side of the energy storage converter. Combined with a data sampling frequency of 1kHz, fine-grained waveform characteristics of the current during charging and discharging are obtained, accurately reflecting the dynamic response and power disturbance modes caused by load changes, forming highly timely energy storage charging and discharging waveform data. Furthermore, a synchronous phase measurement unit is introduced at the grid connection point to synchronously acquire key measurement parameters such as grid frequency, voltage phase angle, and harmonic distortion rate at a frequency of 50Hz. This not only provides data support for grid operation stability but also provides a data foundation for subsequent power quality analysis and frequency response optimization. These three types of heterogeneous data with different frequencies are uniformly converted through a multi-protocol adapter, solving the compatibility issues between different data formats and communication protocols, and are aggregated to generate an initial multimodal dataset. Subsequently, to address the issues of high-frequency noise and low-frequency signal drift during the data acquisition process, the system introduced a spectral domain filtering algorithm. By setting a threshold, the algorithm separates the high-pass and low-pass components of the signal, filtering out unstructured noise such as electromagnetic interference and thermal disturbances. Combined with time-domain analysis, long-term drift compensation is performed, ensuring that all data maintain physical consistency with the baseline values ​​in terms of amplitude. This ultimately results in a standardized energy storage dataset with unified standards and high data reliability. This dataset provides a crucial foundation for subsequent model training, state identification, and energy dispatch optimization, enabling multi-source fusion, high-dimensional accuracy, and spatiotemporal consistency.

[0008] Preferably, step S2, tensor completion of the standardized energy storage dataset, includes: The standardized energy storage dataset is mapped to a three-dimensional space and sorted by timestamp to obtain standardized three-dimensional mapping data for energy storage. Low-rank Tucker decomposition was used to repair temperature and current data in the energy storage three-dimensional mapping normalized data to obtain standard tensor repair data; The standard tensor repair data is used to construct a three-dimensional feature tensor of device × time × feature, and missing value detection is performed to obtain energy storage tensor completion data.

[0009] This invention maps standardized energy storage datasets to a three-dimensional space and uses timestamps as indices to perform temporal sorting of multidimensional data. It establishes a spatially structured data framework based on device number, acquisition time, and physical characteristics (such as voltage, temperature, and current), thereby constructing standardized three-dimensional mapping data for energy storage under a unified coordinate system. This effectively preserves the dynamic changes of various sensor data over time. Subsequently, low-rank Tucker tensor decomposition technology is used to repair the temperature and current subspaces in this three-dimensional structure. This process, based on the tensor rank constraint principle, represents the original data as the product of a core tensor and multiple factor matrices. By minimizing the tensor kernel norm or reconstruction error, redundancy between sensors and low-rank data are extracted, thus filling in abnormal missing measurement areas caused by sensor fluctuations, acquisition delays, or external interference, generating standard tensor repair data with both structural continuity and statistical consistency. Next, the system reconstructs the three-dimensional feature tensor based on this tensor data according to the logic of device × time × feature, further ensuring the data normalization in the device dimension, the equidistant spacing in the time dimension, and the stationarity in the feature dimension, laying a unified data foundation for subsequent processing. Based on this tensor, the system introduces a multi-scale missing value detection mechanism. Combining sliding window and contextual semantic analysis, it analyzes null values, pseudo-zero values, and outliers at different locations and performs interpolation-based completion processing according to the tensor space structure self-consistency rules. Finally, it generates energy storage tensor completion data that meets the requirements in terms of physical consistency, mathematical discriminability, and temporal logic, providing structurally complete and semantically clear data support for high-quality twin modeling, state diagnosis, and predictive control.

[0010] Preferably, step S2, which involves performing time-space dimension analysis on the energy storage tensor completion data and fusing cross-modal features, includes: Extract time-dimensional data from the energy storage tensor completion data to generate battery SOC data; Spatial dimension data is extracted from the energy storage tensor completion data to generate grid frequency fluctuation data; Align battery SOC data and grid frequency fluctuation data along the time axis to generate a synchronized energy storage dataset; Long-term dependency analysis is performed on battery SOC data to obtain battery dependency analysis data; charging and discharging behavior features are extracted from the battery dependency analysis data to obtain time step feature vectors. The grid frequency fluctuation data is used to calculate the unit topology impedance to obtain the grid topology impedance calculation data; the power loss relationship of the charging and discharging path is analyzed from the grid topology impedance calculation data, and the spatial distribution characteristics are extracted to obtain the spatial dimension feature matrix. The time-step feature vector and the spatial dimension feature matrix are concatenated by tensors to obtain the energy storage feature concatenation matrix; the energy storage feature concatenation matrix is ​​then synchronized using the synchronized energy storage dataset to obtain the energy storage scheduling feature fusion matrix. Based on the energy storage scheduling feature fusion matrix, the initial time-space weight allocation is performed, and the feature space is constructed to obtain the energy storage cross-modal feature space.

[0011] This invention extracts State of Charge (SOC) data, representing the remaining capacity of a battery, in the time dimension to depict the battery's charging and discharging evolution at each time step. Simultaneously, it extracts frequency fluctuation data, representing the dynamic characteristics of the power grid system, in the spatial dimension to capture the supply and demand disturbance characteristics of different spatial nodes. By aligning the SOC data and power grid frequency fluctuation data along the time axis, a synchronized energy storage dataset is constructed, ensuring the consistency of analysis of data from different sources under the same time reference. Next, long-term dependency modeling is performed on the SOC data. By analyzing the nonlinear correlation across time steps in the time series, the influence mechanism of the battery's historical state on future charging and discharging behavior is extracted, and a time-step feature vector is constructed to reflect its dynamic evolution trend along the time axis. Then, unit topological impedance calculation is performed on the power grid frequency fluctuation data. By constructing an electrical connection characteristic model between power grid nodes, the energy transmission loss of different nodes during the charging and discharging process is calculated, further extracting its spatial distribution characteristics and generating a spatial dimension feature matrix. Subsequently, the temporal step feature vectors and spatial dimension feature matrices are concatenated on a tensor structure to form an energy storage feature concatenation matrix. This matrix maintains the independence of the physical meaning of various features while ensuring overall semantic consistency. Then, the concatenated matrix is ​​reconstructed using a synchronized energy storage dataset under a unified time reference to generate an energy storage scheduling feature fusion matrix. Based on this, a time-space bidirectional initial weight allocation is performed, and a multi-scale feature weight field is constructed according to historical correlation strength and spatial node load response capabilities, completing the construction of the energy storage cross-modal feature space. This feature space fully integrates the evolutionary trends of the temporal dimension and the energy transfer mechanism of the spatial dimension, possessing strong robustness, cross-domain expressiveness, and scheduling interpretability.

[0012] Preferably, step S3 includes the following steps: Step S31: Perform multi-scale energy storage analysis based on the cross-modal feature space of energy storage, and construct a digital twin of the battery health status to obtain a digital twin model of the energy storage system; Step S32: Design a two-layer optimization framework based on the digital twin model of the energy storage system to obtain the two-layer optimization framework for energy storage, which includes an upper-layer loss reduction framework and a lower-layer energy consumption stability framework. Step S33: Generate a game-theoretic scheduling strategy based on the upper-level loss reduction framework and the lower-level energy consumption stability framework to obtain the energy storage system's power scheduling strategy.

[0013] This invention constructs a digital twin model of an energy storage system driven by multi-scale analysis, based on the data representation foundation of the cross-modal feature space of energy storage. This achieves a coupled modeling and hierarchical control mechanism throughout the entire process, from data fusion to scheduling optimization. At the data level, it first uses the cross-modal feature space as input, integrating multi-source information such as time step features, spatial node features, and energy consumption loss features. Through multi-scale energy storage analysis methods, it models the state evolution at different time granularities (e.g., hourly, daily) and spatial resolutions (e.g., node-level, regional-level), forming a joint expression of health state parameters such as battery state of charge, temperature gradient, and impedance change trends. This leads to the construction of a digital twin of the battery's health state with high fitting accuracy and traceability, mapping a predicted path of the battery's evolution from its current state to its future evolution. Subsequently, a two-layer optimization framework for the energy storage system was designed, centered on this digital twin. The upper-layer energy reduction framework focuses on executing a 24-hour rolling optimization charge-discharge plan based on electricity price, load forecasting, and battery energy efficiency data to minimize costs or maximize benefits. The lower-layer energy consumption stability framework, based on the dynamic feedback results of the digital twin, implements frequency domain and time domain response strategies for grid frequency disturbances, current abrupt changes, or charge-discharge fluctuations, ensuring the operational stability of the energy storage system and the grid system under dynamic coupling. Finally, based on the data input and output response parameters in this two-layer optimization framework, a power dispatch strategy for the energy storage system was jointly constructed using a game-theoretic modeling method involving energy reduction weight factors and stability weight factors. This ensures that dispatch commands meet the real-time frequency regulation requirements of the grid while also considering energy reduction objectives.

[0014] Preferably, step S31 includes the following steps: Step S311: Calculate ohmic polarization / concentration polarization loss based on the energy storage cross-modal characteristic space and generate electrochemical scale analysis data; Step S312: Perform thermodynamic analysis to predict capacity decay based on the energy storage cross-modal characteristic space, and generate thermodynamic scale analysis data; Step S313: Detect electrode stress cracks based on the energy storage cross-modal characteristic space, calculate the crack propagation rate, and generate mechanical scale analysis data; Step S314: Initialize the parameters of the electrochemical scale analysis data, thermodynamic scale analysis data, and mechanical scale analysis data, and construct a digital twin to generate a digital twin model of the energy storage system.

[0015] This invention improves the visualization capabilities and dynamic prediction accuracy of energy storage systems under complex operating conditions by deeply analyzing data across the cross-modal characteristic space of energy storage and constructing a unified analysis framework across multiple physical scales. At the data level, firstly, using fused three-dimensional tensor feature representations, the micro-scale electrochemical behavior of multi-modal signals such as time, voltage, current, temperature, and impedance is modeled. Using voltage-current response curves under unsteady-state conditions, the contributions of ohmic polarization and concentration polarization are separated, a polarization loss function is established, and electrochemical scale analysis data is generated. Secondly, in the thermodynamic dimension, combining the joint distribution characteristics of temperature and current under long-term operation, the thermal gradient tensor is used to map the degree of heat accumulation in different SOC intervals, predicting capacity decay trends and outputting thermodynamic scale analysis data. Further, in the mechanical dimension, stress response analysis is performed on the correlation tensor between current fluctuations and temperature changes under high-frequency sampling, detecting microcracks in electrode materials induced by cyclic loading. Through crack profile fitting and diffusion path modeling methods, the crack propagation rate is quantified, generating mechanical scale analysis data. The three types of scale analysis data mentioned above are used as the initial parameter set. After unified normalization and tensor concatenation, they are used to initialize the digital twin model of the energy storage system. This enables the model to not only reflect the current multi-physical states of the energy storage unit, but also to predict future operational risks. During the continuous updating of state data, the twin model can accurately characterize charging and discharging strategies, battery remaining life, and fault precursors based on the evolution trends of different physical mechanisms.

[0016] Preferably, step S33 includes the following steps: Based on the upper-level loss reduction framework and the lower-level energy consumption stability framework, a game-theoretic scheduling strategy is generated to obtain the energy storage system power scheduling strategy, which includes a loss reduction optimization strategy and a frequency response optimization strategy. Obtain electricity price data; extract load forecast data from the upper-level energy reduction framework to obtain energy storage load forecast data; and formulate a 24-hour linear charging and discharging plan based on the load forecast data and electricity price data to obtain energy reduction optimization strategies. Real-time grid frequency is extracted from the lower-level energy consumption stability framework to obtain real-time grid frequency data. When the real-time grid frequency data exceeds 0.2Hz, an energy storage correction command is triggered to make adjustments, resulting in a frequency response optimization strategy.

[0017] This invention, within an upper-level energy conservation framework, uses historical and real-time electricity price data to model the correlation between the system and the historical operating trajectory of energy storage load. By performing high-dimensional time series prediction of load characteristics, it extracts predicted energy storage load data for the next 24 hours. This load prediction process combines short-cycle fluctuations and long-cycle trend factors to establish a time-step distribution feature vector, which is then cross-linearly modeled with electricity price data. This ultimately generates a continuous 24-hour linear charging and discharging behavior sequence, constituting an energy conservation optimization strategy. This allows the system to prioritize charging during low-price periods and discharging during high-price periods during intraday scheduling, thereby minimizing energy costs. Secondly, within a lower-level energy consumption stability framework, instantaneous changes in grid frequency are extracted using synchronous phase measurement data with a real-time sampling frequency of 50Hz. This constructs a grid frequency time series. When an abnormal event with a frequency deviation exceeding 0.2Hz occurs in this series, a frequency response mechanism is immediately triggered, invoking pre-set energy storage correction commands to rapidly adjust battery charging and discharging power. This correction strategy relies on regression analysis of energy storage response data under historical frequency disturbances to form a response boundary model, enabling rapid absorption and mitigation of frequency disturbances. Finally, the two strategies are fused using tensor-level weights to generate an energy dispatch strategy for the energy storage system that combines optimal energy reduction with frequency regulation capabilities. This strategy retains predictability across multiple time scales while possessing dynamic adaptability with a second-level response.

[0018] Preferably, step S4 includes the following steps: Step S41: Obtain real-time energy storage monitoring data; Step S42: Calculate the execution process deviation based on real-time energy storage monitoring data and energy storage system power dispatch strategy. If the energy storage deviation data is greater than the preset threshold, generate SOC attenuation data. Step S43: The SOC attenuation data is transmitted back to the digital twin model of the energy storage system for network parameter optimization, generating an optimized digital twin model of the energy storage system; Step S44: Based on the optimized digital twin model of the energy storage system, perform full-process scheduling analysis of the energy storage system and generate a report to obtain a full-process report on the power scheduling of the energy storage system.

[0019] This invention achieves high-precision modeling of the energy storage system's operating status and enhanced traceability of the entire scheduling process by introducing an execution process deviation feedback mechanism and a dynamic optimization process using a digital twin model. This improves the system's operational transparency and predictive control capabilities from a data perspective. First, the system acquires real-time energy storage monitoring data through a sensor network deployed at key nodes of the energy storage subsystem. Data types include high-frequency operating indicators such as battery module SOC, terminal voltage, instantaneous power, and charge / discharge command response delay. After entering the data processing link, this data is dynamically compared with the currently executing energy storage system power scheduling strategy to construct a time-aligned mapping between the expected response sequence and the actual feedback sequence, thereby calculating the execution process deviation. By performing statistical sliding window analysis on the deviation data, the mean fluctuation and relative deviation rate are extracted. When the deviation exceeds a preset threshold (e.g., 5% or SOC fluctuation exceeding 2%), the system considers this deviation as a sign of an internal degradation trend in the energy storage system and derives SOC decay data accordingly. This attenuation data is not treated as an isolated indicator, but rather fed back into the digital twin model of the energy storage system to participate in the updating and optimization of the twin network parameters. The backpropagation (BP) iterative algorithm is used to correct the activation function weights and state transition functions related to SOC prediction in the twin model, thereby generating an optimized digital twin model that more closely reflects the actual state. This model not only retains the physical modeling capability for prior knowledge such as battery thermal behavior and electrochemical response, but also integrates the dynamic feedback characteristics of operational data, realizing the self-evolution of the twin. Finally, based on the optimized digital twin model, the system generates a full-process report on the energy storage system's power dispatch. This report can cover the entire structural topology of the energy storage system in the spatial dimension, trace the evolution path of execution deviations in the temporal dimension, and make quantitative predictions of operational trends over a certain period, providing high-precision reference data for operation and maintenance personnel.

[0020] Preferably, step S42 includes the following steps: Step S421: Calculate the deviation value based on real-time energy storage monitoring data and energy storage system power dispatch strategy, and generate energy storage deviation value data; Step S422: Calculate the SOC deviation based on real-time energy storage monitoring data and energy storage system power dispatch strategy to obtain energy storage SOC deviation data; Step S423: If the energy storage deviation data is greater than 5% and the energy storage SOC deviation data is greater than 2%, then the compensation coefficient for error propagation path backtracking is generated, and the SOC attenuation data is obtained.

[0021] This invention effectively improves the accuracy and response efficiency of identifying abnormal operating states and tracking SOC (State of Charge) degradation trends in energy storage systems at the data level by introducing a multi-deviation quantification mechanism and error propagation path backtracking analysis. First, the system uses the comparison between real-time energy storage monitoring data and the current scheduling strategy to quantify and calculate both the overall scheduling deviation and the battery SOC deviation. Specifically, the energy storage deviation data is constructed by sampling and normalizing the difference between real-time charging / discharging power and the target power to build a time-series offset function, reflecting the actual response deviation of the scheduling system at the execution layer. Simultaneously, the SOC deviation data is extracted based on the point-to-point error between the predicted and actual observed SOC curves, combined with the volatility and difference gradient changes within the current time window, to extract a dynamic SOC residual feature vector. This dual-deviation index system enables parallel perception of scheduling deviations and battery aging trends. Furthermore, when the energy storage deviation exceeds 5% and the SOC deviation exceeds 2%—a combined threshold condition—the system activates an error propagation path backtracking mechanism. This mechanism traces the causal relationships between multi-stage control commands causing the deviation, historical charging and discharging rate adjustments, battery pack thermal behavior trends, and instantaneous grid frequency disturbances. Through principal component analysis and error sensitivity assessment, it identifies key nodes and major influencing variables that caused the deviation. Based on this, the system constructs an error compensation coefficient matrix based on path backtracking and combines historical decay patterns to correct and fit the current deviation, ultimately generating SOC decay data with both physical rationality and data-driven accuracy. This SOC decay data not only boasts high accuracy and fast response but can also serve as feedback input to optimize digital twin parameters, supporting subsequent state assessment and predictive analysis.

[0022] This specification provides a power dispatching system based on an energy storage system for executing the aforementioned power dispatching method based on an energy storage system. The power dispatching system based on an energy storage system includes: The multi-source heterogeneous sensing standardization module is used to deploy a distributed sensor array within the battery module to collect multi-source heterogeneous data and perform protocol conversion to obtain an initial multimodal dataset. The initial multimodal dataset is then subjected to noise spectrum adjustment filtering thresholds and long-term drift compensation to generate a standardized energy storage dataset. The cross-modal feature fusion module is used to perform tensor completion on the standardized energy storage dataset to generate energy storage tensor completion data; the energy storage tensor completion data is then analyzed in terms of time and space dimensions, and cross-modal feature fusion is performed to generate an energy storage cross-modal feature space; The digital twin modeling and game-theoretic scheduling module is used to construct a digital twin of the battery health status based on the cross-modal feature space of energy storage, thereby obtaining a digital twin model of the energy storage system; and to generate a game-theoretic scheduling strategy from the digital twin model of the energy storage system, thereby obtaining the energy scheduling strategy of the energy storage system. The real-time deviation analysis and closed-loop optimization module is used to acquire real-time energy storage monitoring data; based on the real-time energy storage monitoring data and the energy storage system's power dispatch strategy, it calculates the execution process deviation. If the energy storage deviation data is greater than a preset threshold, it generates SOC attenuation data; the SOC attenuation data is sent back to the energy storage system's digital twin model for network parameter optimization, and a full-process scheduling analysis of the energy storage system is performed to obtain full-process scheduling analysis data; based on the full-process scheduling analysis data, a scheduling report is generated to obtain a full-process report on the energy storage system's power dispatch.

[0023] The beneficial effects of this invention lie in the real-time acquisition of multi-dimensional heterogeneous data in a cold chain storage environment through the deployment of a multi-source sensor array. Furthermore, timestamp alignment technology ensures precise synchronization of different data sources in the time dimension, resulting in highly timely and complete timestamp-aligned cold chain storage data. Subsequently, through three-dimensional spatial mapping and the construction of a feature tensor matrix, spatial distribution information of temperature, humidity, and other environmental parameters is integrated to generate a fused feature matrix. This matrix reflects the multi-dimensional dynamic characteristics of the storage environment, providing a structured data foundation for subsequent analysis. Based on this, a pre-defined Long Short-Term Memory (LSTM) network model is used to predict temperature distribution on the fused feature matrix, capturing the time-series characteristics and spatial correlation of temperature changes, generating predicted temperature distribution data for the storage environment. This data is further combined with the fused feature matrix to construct a cold chain causal knowledge graph, revealing the causal relationships between environmental variables. A digital twin model is then constructed based on this graph, enabling multi-dimensional simulation and dynamic modeling of the storage system's state. The digital twin model then undergoes iterative error residual analysis. The system identifies the discrepancies between model predictions and actual monitoring, generates differential analysis data, and automatically adjusts compensation parameters through a feedback loop to optimize and fine-tune the model, resulting in a more accurate optimized digital twin model for warehousing. Ultimately, this optimized model drives the construction of logistics and warehousing management reports, systematically presenting key indicators such as temperature prediction accuracy, environmental stability, and management efficiency, supporting scientific decision-making and precise scheduling. Overall, this method, through a closed loop of data collection, synchronization, fusion, prediction, and feedback optimization, enhances the environmental awareness and prediction accuracy of the cold chain warehousing system, achieving dynamic adaptability and continuous optimization of the model. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the steps of a power dispatching method based on an energy storage system. Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S4. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] The technical method of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0026] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0027] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0028] To achieve the above objectives, please refer to Figures 1 to 2 A power dispatching method based on an energy storage system, the method comprising the following steps: Step S1: Deploy a distributed sensor array within the battery module to collect multi-source heterogeneous data and perform protocol conversion to obtain an initial multimodal dataset; adjust the noise spectrum filtering threshold of the initial multimodal dataset and perform long-term drift compensation to generate a standardized energy storage dataset. Step S2: Perform tensor completion on the standardized energy storage dataset to generate energy storage tensor completion data; perform time-space dimension analysis on the energy storage tensor completion data and perform cross-modal feature fusion to generate energy storage cross-modal feature space; Step S3: Construct a digital twin of the battery health status based on the energy storage cross-modal feature space to obtain a digital twin model of the energy storage system; generate a game-theoretic scheduling strategy from the digital twin model of the energy storage system to obtain the energy dispatch strategy of the energy storage system. Step S4: Obtain real-time energy storage monitoring data; calculate the execution process deviation based on the real-time energy storage monitoring data and the energy storage system's power dispatch strategy. If the energy storage deviation data is greater than a preset threshold, generate SOC attenuation data; send the SOC attenuation data back to the energy storage system's digital twin model for network parameter optimization, and perform full-process scheduling analysis of the energy storage system to obtain full-process scheduling analysis data; generate a scheduling report based on the full-process scheduling analysis data to obtain a full-process report on energy storage system power dispatch.

[0029] In this embodiment of the invention, reference Figure 1 The diagram shown is a flowchart illustrating the steps of a power dispatching method based on an energy storage system according to the present invention. In this example, the power dispatching method based on an energy storage system includes the following steps: Step S1: Deploy a distributed sensor array within the battery module to collect multi-source heterogeneous data and perform protocol conversion to obtain an initial multimodal dataset; adjust the noise spectrum filtering threshold of the initial multimodal dataset and perform long-term drift compensation to generate a standardized energy storage dataset. In this embodiment of the invention, key physical quantity monitoring points are selected within the battery module, and multiple types of sensing units, such as temperature, voltage, current, stress, and humidity sensors, are deployed to achieve synchronous acquisition of multi-source physical signals during battery operation. These sensor nodes are connected to the edge acquisition terminal via wired or wireless communication, forming a distributed data acquisition network, and data synchronization and fusion are achieved through timestamps and sampling frequencies. Since the acquired data exhibits significant heterogeneity in terms of physical source, unit of measurement, time scale, and spatial distribution, data cleaning is first required to remove missing values, outliers, and invalid sampling points. Then, numerical scale standardization and feature pattern unification are performed based on methods such as normalization, standard deviation scaling, and wavelet transform to ensure comparability and processability of different types of data. On this basis, the standardized data from all sensor nodes within the same time window are structured and organized to construct a multi-dimensional tensor data representation structure, with time series as the main axis and spatial nodes and sensor types as auxiliary axes, forming a standardized energy storage dataset with completeness, temporal sequence, and structured characteristics. This dataset not only preserves the correlation between multiple physical information sources, but also has scalability, providing a unified and standardized data foundation for subsequent tensor completion, modal fusion, and twin construction.

[0030] Step S2: Perform tensor completion on the standardized energy storage dataset to generate energy storage tensor completion data; perform time-space dimension analysis on the energy storage tensor completion data and perform cross-modal feature fusion to generate energy storage cross-modal feature space; In this embodiment of the invention, the standardized energy storage dataset is typically represented as a sparse tensor with a three-dimensional time-space-modal structure. Some observations in this tensor are missing due to sensor momentary failures, communication delays, or environmental interference. To improve data integrity, the tensor structure first needs to be completed. The completion process can employ tensor decomposition methods based on low-rank constraints, such as CANDECOMP / PARAFAC decomposition (CP decomposition) or Tucker decomposition, to separate the kernel tensor and factor matrix of the original sparse tensor, and estimate the missing data by minimizing the reconstruction error. Alternatively, a neural completion method based on deep learning can be introduced, such as constructing a temporal convolutional network (TCN) or a spatiotemporal graph convolutional network (ST-GCN), utilizing the continuity of the data in the time and space dimensions for nonlinear interpolation and reconstruction. The completed tensor data, as a complete multidimensional data structure, provides the basis for conducting multidimensional joint analysis. Building upon this foundation, time series trends are further extracted using sliding time window techniques. Simultaneously, spatial dependency analysis is performed by combining the spatial topology of sensor nodes. For example, dynamic time warping (DTW) is used to align the time series of different nodes, or graph neural networks are used to model spatial correlations. After completing the joint time-space modeling, the features of different physical quantities such as voltage, current, temperature, and displacement along the modal axes are combined. Through a multimodal fusion network controlled by principal component analysis (PCA), canonical correlation analysis (CCA), or attention mechanisms, the feature vectors of each modality are mapped to a unified embedding space, and a high-dimensional interaction matrix is ​​constructed. This achieves collaborative representation and fusion of cross-modal features, ultimately forming a cross-modal feature space for energy storage with rich semantic structure and strong expressive power. This space supports subsequent digital twin construction and intelligent scheduling strategy generation.

[0031] Step S3: Construct a digital twin of the battery health status based on the energy storage cross-modal feature space to obtain a digital twin model of the energy storage system; generate a game-theoretic scheduling strategy from the digital twin model of the energy storage system to obtain the energy dispatch strategy of the energy storage system. In this embodiment of the invention, a structured modeling and state prediction mechanism based on multidimensional heterogeneous features is used to achieve holographic mapping and game-theoretic optimization control of the operating state of the energy storage system. From a data perspective, the energy storage cross-modal feature space contains high-dimensional feature tensors generated by tensor completion and spatiotemporal fusion, covering multi-source modal indicators such as voltage, current, temperature, strain, and insulation state, while maintaining continuity and consistency in both time and space dimensions. For this high-dimensional feature space, a feature-driven digital twin oriented towards health state representation must first be constructed. A deep neural network architecture is typically used to model this feature space, including but not limited to multilayer perceptrons (MLP), convolutional neural networks (CNN), or hybrid structures combining graph neural networks (GNN) and recurrent neural networks (RNN). Through training, the model acquires the ability to identify potential failure modes, degradation processes, and lifespan trends within the battery module. During model training, historical operating state labels or expert-annotated data serve as supervision signals to minimize the error function between the predicted and actual states, gradually achieving convergence of the digital twin model parameters. Building upon this foundation, a digital twin of the health status serves as an interactive bridge between the virtual and physical systems. Reinforcement learning or game theory optimization strategies (such as policy gradient methods based on Markov decision processes, Q-learning algorithms, or Stackelberg game models) are introduced. Based on the health status evolution trends predicted by the twin model, strategies for load allocation, power regulation, and balance control of each energy storage unit are generated, outputting a set of executable power dispatch actions. This strategy generation process uses maximizing resource utilization, minimizing system degradation, and optimizing response time as objective functions. Multiple rounds of strategy game solving are conducted within the feasible solution space, and the results are fed back to the physical system to guide the allocation and execution of energy storage tasks, forming a closed-loop scheduling mechanism that combines virtual and physical systems.

[0032] Step S4: Obtain real-time energy storage monitoring data; calculate the execution process deviation based on the real-time energy storage monitoring data and the energy storage system's power dispatch strategy. If the energy storage deviation data is greater than a preset threshold, generate SOC attenuation data; send the SOC attenuation data back to the energy storage system's digital twin model for network parameter optimization, and perform full-process scheduling analysis of the energy storage system to obtain full-process scheduling analysis data; generate a scheduling report based on the full-process scheduling analysis data to obtain a full-process report on energy storage system power dispatch.

[0033] In this embodiment of the invention, multi-source indicator data streams during system operation are collected in real time through edge computing nodes or embedded monitoring units. These data include, but are not limited to, individual battery cell voltage, current, temperature, state of charge (SOC), insulation resistance, and energy flow distribution. This data is transmitted to a local or cloud-based time-series database via high-frequency sampling and undergoes time-series alignment, outlier removal, and unit normalization to form cleaned real-time energy storage monitoring data. Subsequently, this real-time data is input into the energy storage system power dispatch strategy module generated by game theory. By comparing the deviation between the actual operating results and the expected output of the strategy, the execution process deviation is calculated using evaluation functions such as mean square error (MSE), dynamic time warping (DTW), or Bayesian error boundary-based evaluation functions. If the deviation exceeds a preset dynamic tolerance threshold range, the deviation is further analyzed to identify the root cause of performance degradation, and SOC (State of Charge) decay data is quantified using indicators such as battery capacity decay, increased internal resistance, or thermal runaway trend. This type of SOC decay data not only exhibits temporal correlation but also manifests as a high-dimensional sparse sequence related to historical states, environmental factors, and usage patterns. Therefore, it requires modeling using attention mechanisms or variational autoencoder (VAE) structures to enhance its ability to describe system state degradation paths. Subsequently, the SOC decay data is fed back into the energy storage system's digital twin model. By adjusting the neural network parameters, state transition probability matrix, or dynamic discriminator weights, the twin achieves adaptive updates to environmental and degradation changes. Finally, monitoring data, deviation assessment results, strategy adjustment records, and optimized state parameters are integrated throughout the process. A unified information integration module generates a full-process report on energy dispatch for the energy storage system. This report not only possesses timeliness and traceability but also serves as data support for subsequent strategy verification and health assessment.

[0034] Preferably, step S1 includes the following steps: Step S11: Deploy voltage and temperature sensors within the battery module to collect individual cell temperature and terminal voltage data at a frequency of 10Hz. The voltage sensor has an accuracy of 0.1%, and the temperature sensor has an accuracy of 0.5℃. Step S12: Deploy a high-frequency current sensor with a bandwidth of 20kHz on the DC side of the energy storage converter, and collect the charging and discharging current waveform at a frequency of 1kHz to generate energy storage charging and discharging waveform data. Step S13: Deploy a synchronous phase measurement unit at the grid connection point to collect grid frequency, voltage phase angle and harmonic distortion rate at a frequency of 50Hz, and generate raw grid phase data; Step S14: Convert the individual cell temperature data, terminal voltage data, energy storage charging and discharging waveform data, and grid phase raw data through a multi-protocol adapter to generate an initial multi-mode dataset; Step S15: Adjust the noise spectrum filtering threshold of the initial multimodal dataset and perform long-term drift compensation to generate a standardized energy storage dataset.

[0035] In this embodiment of the invention, a voltage sensor with an accuracy of 0.1% and a temperature sensor with an accuracy of 0.5°C are deployed within the battery module. Individual cell terminal voltage and temperature data are collected at a frequency of 10Hz. The resulting signals form a low-frequency sampling state trajectory on the time axis, primarily used to evaluate the balance and thermal stability between cells. Secondly, a high-frequency current sensor with a bandwidth of 20kHz is deployed on the DC side of the energy storage converter. Charging and discharging current waveforms are collected at a frequency of 1kHz, capturing dynamic behaviors such as high-frequency disturbances and electromagnetic transient responses. This data is presented in time-series form and combined with methods such as Fourier transform or wavelet packet decomposition for time-frequency domain feature analysis. Simultaneously, a synchronous phasor measurement unit (PMU) is installed at the grid connection point to synchronously collect the grid frequency, voltage phase angle, and harmonic distortion rate at a frequency of 50Hz, forming a grid time-series vector containing phasor information, used to describe the dynamic interaction between the system and the grid. The three types of data sources mentioned above have different sampling frequencies, data dimensions, and encoding formats. A multi-protocol adapter is needed to uniformly encapsulate and convert the data carried by communication protocols such as Modbus, CAN, and IEC 61850. Negotiation and handshake, and data stream packaging are completed at the physical and network layers to ultimately generate an initial multimodal dataset with unified timestamps and a consistent structure. After completing protocol compatibility processing, to ensure the stability and comparability of the data in subsequent modeling processes, noise spectrum analysis is performed on the initial multimodal dataset. Based on the power spectral density (PSD) calculation results, a filtering threshold is set, and adaptive filters or bandpass filters are used to reduce high-frequency interference signals. Simultaneously, to address slow variable shifts caused by factors such as temperature drift and electrochemical aging during long-term use, drift correction methods such as least squares fitting of drift trends, incremental moving average compensation, or exponentially weighted moving average (EWMA) are used to correct signal drift. Finally, the above processing generates a standardized energy storage dataset with uniform scale, low noise, and high stability, providing a consistent basic input for subsequent data modeling and state estimation.

[0036] Preferably, step S2, tensor completion of the standardized energy storage dataset, includes: The standardized energy storage dataset is mapped to a three-dimensional space and sorted by timestamp to obtain standardized three-dimensional mapping data for energy storage. Low-rank Tucker decomposition was used to repair temperature and current data in the energy storage three-dimensional mapping normalized data to obtain standard tensor repair data; The standard tensor repair data is used to construct a three-dimensional feature tensor of device × time × feature, and missing value detection is performed to obtain energy storage tensor completion data.

[0037] In this embodiment of the invention, the standardized multi-source energy storage dataset is reorganized based on a unified timestamp, and a three-dimensional mapping operation is performed according to the three dimensions of "device number - sampling time - measurement characteristics" to construct a three-dimensional sparse tensor with energy storage units as spatial indices, physical time series as evolution indices, and state parameters such as temperature, voltage, and current as feature indices. In this three-dimensional structure, there are problems of inconsistent sampling frequencies and missing data in different dimensions, especially under high-frequency measurements, where temperature and current signals are easily affected by sensor failures or communication delays, resulting in null values. To mathematically complete these data gaps, the Tucker low-rank tensor decomposition method is selected. First, the three-dimensional tensor is decomposed into a core tensor and multiple mode matrices through high-order singular value decomposition (HOSVD). Then, rank constraint optimization is performed in the tensor quantum space, and the missing regions are numerically estimated by minimizing the tensor reconstruction error. This process relies on the low-rank structure assumption of the observed data, thereby achieving global collaborative repair of missing values ​​in the temperature and current dimensions. After tensor decomposition and numerical imputation, the reconstructed tensors are reorganized into a device × time × feature format, forming a structurally complete three-dimensional feature tensor for subsequent modeling and analysis. Subsequently, a missing value detection mechanism is introduced into the tensor data. Combined with a mask matrix to mark the initial missing region and the repair region, the integrity of the tensor, the repair error, and the consistency of the statistical distribution are verified. Outliers with large errors are removed, or the repair algorithm parameters are further refined to obtain the final energy storage tensor complete data, thus providing a data foundation for feature fusion and state modeling.

[0038] Preferably, step S2, which involves performing time-space dimension analysis on the energy storage tensor completion data and fusing cross-modal features, includes: Extract time-dimensional data from the energy storage tensor completion data to generate battery SOC data; Spatial dimension data is extracted from the energy storage tensor completion data to generate grid frequency fluctuation data; Align battery SOC data and grid frequency fluctuation data along the time axis to generate a synchronized energy storage dataset; Long-term dependency analysis is performed on battery SOC data to obtain battery dependency analysis data; charging and discharging behavior features are extracted from the battery dependency analysis data to obtain time step feature vectors. The grid frequency fluctuation data is used to calculate the unit topology impedance to obtain the grid topology impedance calculation data; the power loss relationship of the charging and discharging path is analyzed from the grid topology impedance calculation data, and the spatial distribution characteristics are extracted to obtain the spatial dimension feature matrix. The time-step feature vector and the spatial dimension feature matrix are concatenated by tensors to obtain the energy storage feature concatenation matrix; the energy storage feature concatenation matrix is ​​then synchronized using the synchronized energy storage dataset to obtain the energy storage scheduling feature fusion matrix. Based on the energy storage scheduling feature fusion matrix, the initial time-space weight allocation is performed, and the feature space is constructed to obtain the energy storage cross-modal feature space.

[0039] In this embodiment of the invention, key state variables are extracted from the completed three-dimensional energy storage tensor along the time axis to construct a battery SOC (State of Charge) data sequence in a time series manner, capturing the dynamic change trend of the battery's state of charge. Then, frequency fluctuation data at corresponding locations on the grid side and equipment side are extracted along the spatial axis. Based on continuous time slices, the volatility is calculated, and a spatially distributed frequency response matrix is ​​constructed to form grid frequency volatility data. After the initial extraction of the two modal data, they are mapped to a unified time axis using methods such as interpolation alignment and window synchronization to generate a synchronized energy storage dataset to support subsequent joint modeling. Then, time dependency modeling is performed on the battery SOC data. A sliding window mechanism is used to extract local time dependency features, and an autocorrelation structure or long-term memory structure is further constructed. Statistical analysis and frequency domain transformation (such as Fourier or wavelet transform) are used to capture the long-range correlation of the battery state over time, obtaining battery dependency analysis data. Based on this, behavioral features such as rate, amplitude changes, and peak-valley transitions during the charging and discharging process are extracted to construct a time-recursive time step feature vector. Simultaneously, based on the electrical topology and node location relationships, a grid unit impedance matrix is ​​constructed using grid frequency fluctuation data. Energy losses along the charging and discharging paths between nodes are analyzed using impedance solving algorithms to generate grid topology impedance calculation data. Then, spatial distribution features are extracted using high-dimensional feature encoding or spatial grouping strategies to form a spatial dimension feature matrix. Subsequently, the time-step feature vectors and the spatial dimension feature matrix are tensor-concatenated to construct a unified-dimensional energy storage feature concatenation matrix. The feature concatenation matrix is ​​then standardized, normalized, and timestamped using a synchronized energy storage dataset, generating an energy storage scheduling feature fusion matrix with alignment consistency and modal complementarity. Finally, initial weight allocation for the time and spatial dimensions is performed on the fusion matrix. Attention mechanisms, location embedding, or weight regularization strategies are employed to construct a joint feature representation space, forming a cross-modal energy storage feature space that can be used for subsequent modeling.

[0040] Preferably, step S3 includes the following steps: Step S31: Perform multi-scale energy storage analysis based on the cross-modal feature space of energy storage, and construct a digital twin of the battery health status to obtain a digital twin model of the energy storage system; Step S32: Design a two-layer optimization framework based on the digital twin model of the energy storage system to obtain the two-layer optimization framework for energy storage, which includes an upper-layer loss reduction framework and a lower-layer energy consumption stability framework. Step S33: Generate a game-theoretic scheduling strategy based on the upper-level loss reduction framework and the lower-level energy consumption stability framework to obtain the energy storage system's power scheduling strategy.

[0041] In this embodiment of the invention, based on the cross-modal feature space of energy storage, a multi-scale spatiotemporal decomposition method is used to perform hierarchical analysis of the energy storage system operation data, including the identification of battery charge change patterns at short time scales and aging trends at long time scales. Simultaneously, the state evolution path is constructed by combining the joint distribution characteristics of key variables such as temperature, current, and voltage. By mapping these multi-scale time series to a dynamic Bayesian network or a graph convolution-based temporal coding structure, a dynamic characterization of the battery health state can be achieved. This allows for the training and generation of a digital twin model reflecting the battery performance degradation law and abnormal behavior response, ultimately forming a digital twin model of the energy storage system. Subsequently, a two-layer optimization framework is constructed based on this digital twin model. The upper-layer loss reduction framework uses profit maximization or cost minimization as the objective function, constructing a mixed-integer linear programming or Lagrange relaxation optimization model based on predicted market prices, load demand, and energy storage state distribution. The lower-layer energy consumption stability framework focuses on energy balance and system dynamic constraints during real-time operation, constructing a second-order state-space model or robust optimization structure to ensure that stability conditions such as voltage, current, and frequency are met during actual scheduling. The upper and lower layers establish a coupled feedback channel through an iterative update mechanism. The upper layer provides a candidate solution space for the scheduling strategy, while the lower layer returns stability evaluation indicators and provides feedback to correct boundary conditions, forming a closed-loop optimization iterative system. Finally, non-cooperative game theory or Stackelberg game modeling methods are introduced into this two-layer optimization framework. Considering the resource competition relationships among multiple stakeholders, including battery modules, grid nodes, and external loads, a Nash equilibrium solution model or evolutionary game strategy learning algorithm is constructed to achieve dynamic balance of the scheduling strategy among different stakeholders, thereby generating an energy storage system power scheduling strategy oriented towards global efficiency and local stability. This process highly relies on the temporal consistency, spatial integrity, and modal synergy of energy storage data, ensuring the rationality and effectiveness of the game-based scheduling mechanism at the data level.

[0042] Preferably, step S31 includes the following steps: Step S311: Calculate ohmic polarization / concentration polarization loss based on the energy storage cross-modal characteristic space and generate electrochemical scale analysis data; Step S312: Perform thermodynamic analysis to predict capacity decay based on the energy storage cross-modal characteristic space, and generate thermodynamic scale analysis data; Step S313: Detect electrode stress cracks based on the energy storage cross-modal characteristic space, calculate the crack propagation rate, and generate mechanical scale analysis data; Step S314: Initialize the parameters of the electrochemical scale analysis data, thermodynamic scale analysis data, and mechanical scale analysis data, and construct a digital twin to generate a digital twin model of the energy storage system.

[0043] In this embodiment of the invention, at the electrochemical scale, key variables such as instantaneous voltage, current density, estimated internal resistance, and temperature field changes are extracted from the cross-modal feature space. Based on the extended Butler-Volmer equation framework and the RC submodule response function in the battery equivalent circuit model, the energy loss caused by ohmic polarization and concentration polarization is quantified. By introducing impedance spectral mapping and multi-node voltage response differential analysis, hidden variables such as electrolyte conductivity and electrode interface reaction kinetic parameters can be further separated to form electrochemical scale analysis data. Secondly, in the thermodynamic scale analysis, an unsteady-state heat conduction model is constructed using continuously acquired temperature gradient distribution, current waveform, and energy input / output time series in the cross-modal feature space. Combined with the heat capacity and thermal conductivity parameters of the battery material, finite difference time step progression is performed to deduce the core temperature rise trend and local overheating probability density function. The key thermal driving mechanism causing capacity decay is predicted through an energy conservation model, generating thermodynamic scale analysis data. Next, in terms of mechanical scale analysis, high-dimensional feature patterns such as anomalous internal resistance growth, voltage step changes, and charge / discharge cycle compression in the cross-modal feature space are utilized. A crack detection network is trained using XRD parameters or finite element simulation results to calculate the maximum principal stress and shear stress distribution of the electrode sheet under cyclic loading. Crack propagation rate is estimated based on Paris's law or empirical crack growth models, thus constructing a mechanical stress-driven crack evolution dataset. Finally, the electrochemical, thermodynamic, and mechanical scale analysis data are initialized and fused using high-dimensional tensor parameter alignment and physical constraint embedding methods to generate cross-scale consistent input feature tensors. A digital twin model of the energy storage system with a structure-behavior-performance three-layer mapping mechanism is then constructed, providing multi-scale coupling support for subsequent scheduling simulations and performance predictions.

[0044] Of particular importance, step S32 includes the following: Step S32: Design a two-layer optimization framework based on the digital twin model of the energy storage system to obtain the two-layer optimization framework for energy storage, which includes an upper-layer loss reduction framework and a lower-layer energy consumption stability framework. The formula for the upper-level loss-reduction frame is:

[0045] in, For time indexing; This represents the total number of scheduling periods; The unit price of electricity purchased / sold by the power grid during time period t; The power exchanged with the grid during time period t (kW) represents the power purchased (positive values) and sold (negative values). This is the battery aging cost coefficient; Battery capacity degradation during time period t; This refers to the battery's rated capacity. The formula for lower-level energy consumption stability is:

[0046] in, The actual frequency of the power grid during time period t; The state of charge of the battery during time period t; This is the SOC reference value; This is the SOC deviation penalty coefficient.

[0047] In this embodiment of the invention, coordinated control is achieved through hierarchical processing of time-series energy reduction targets and real-time stability constraints. Specifically, the technical means involves using battery capacity degradation prediction data output from a digital twin model. (temperature collected in real time) Current (And historical charge-discharge cycle data generated by regression using an LSTM network) and power time series data exchanged with the power grid. (Source: SCADA system 15-minute granular electricity pricing) (Based on the current load forecast data), the upper-level energy reduction framework quantifies battery loss as an energy reduction cost item. (in The aging cost conversion factor is dynamically updated through battery lifecycle cost amortization and replacement, and is compared with the cost of electricity purchase and sale. The objective function is constructed using a linear weighted average, and a branch-and-bound algorithm is used to solve a 96-time-period mixed-integer programming problem. The lower-level energy consumption stability framework is then connected in real time to the millisecond-level grid frequency data stream collected by the PMU. And the second-level SOC data reported by BMS, the frequency deviation L2 norm is calculated. and SOC tracking error (in The safety window is dynamically set to 40%-70% based on the optimization results from the upper layer, and a regularization coefficient is introduced. The weights of both are balanced (automatically reduced when frequency fluctuations exceed a threshold). (by relaxing the SOC constraints), a quadratic programming problem is ultimately formed and solved online using the interior-point method, outputting a power correction command. The actual operating data is fed back to the digital twin at the energy storage converter execution layer to achieve closed-loop parameter calibration.

[0048] Preferably, step S33 includes the following steps: Based on the upper-level loss reduction framework and the lower-level energy consumption stability framework, a game-theoretic scheduling strategy is generated to obtain the energy storage system power scheduling strategy, which includes a loss reduction optimization strategy and a frequency response optimization strategy. Obtain electricity price data; extract load forecast data from the upper-level energy reduction framework to obtain energy storage load forecast data; and formulate a 24-hour linear charging and discharging plan based on the load forecast data and electricity price data to obtain energy reduction optimization strategies. Real-time grid frequency is extracted from the lower-level energy consumption stability framework to obtain real-time grid frequency data. When the real-time grid frequency data exceeds 0.2Hz, an energy storage correction command is triggered to make adjustments, resulting in a frequency response optimization strategy.

[0049] In this embodiment of the invention, within the upper-layer energy reduction framework, the system receives historical electricity load data and meteorological forecasting factors (such as temperature, humidity, and holiday indicators). It then constructs an energy storage load forecasting model using an integrated time-series regression method (such as a sliding window-based LSTM model or a SARIMA model), outputting time-segmented electricity demand curves for the next 24 hours, thus forming energy storage load forecasting data. Simultaneously, the system incorporates electricity price time-series data provided by the electricity market and performs trend analysis and price fluctuation range identification. Based on these two types of data, the objective function is set to "minimize energy reduction costs." Under the premise of satisfying energy storage device capacity and power constraints, minimum SOC limits, and lifetime degradation constraints, a 24-hour linear programming model is constructed, setting charging and discharging commands for each hour. The energy reduction optimization strategy is obtained by solving this linear programming model. Secondly, within the lower-layer energy consumption stability framework, the system accesses the frequency data stream generated by the grid synchronization phasor measurement unit (PMU) in real time, sampling at a frequency of 10 points per second. Median filtering and trend removal are performed first, followed by differential analysis and threshold judgment of the rate of change of frequency (ROCOF). When a transient disturbance exceeding ±0.2Hz with a frequency deviation of 50Hz from the center value is detected, the frequency response mechanism is triggered. Based on the magnitude and direction of the frequency deviation, the current energy loss reduction optimization strategy is offset and corrected. This correction dynamically adjusts the charging and discharging power based on a predefined proportional-integral adjustment factor, and the correction command is issued within the strategy control frame (e.g., within 1 minute), thereby generating the frequency response optimization strategy. Finally, the energy loss reduction optimization strategy and the frequency response optimization strategy are jointly scheduled through a game-theoretic coordination mechanism to ensure that the system achieves a dynamic balance between energy loss reduction and grid stability, generating a complete energy storage system power dispatch strategy and sending it to the execution control unit.

[0050] As an example of the present invention, reference is made to Figure 2 As shown, step S4 in this example includes: Step S41: Obtain real-time energy storage monitoring data; Step S42: Calculate the execution process deviation based on real-time energy storage monitoring data and energy storage system power dispatch strategy. If the energy storage deviation data is greater than the preset threshold, generate SOC attenuation data. Step S43: The SOC attenuation data is transmitted back to the digital twin model of the energy storage system for network parameter optimization, generating an optimized digital twin model of the energy storage system; Step S44: Based on the optimized digital twin model of the energy storage system, perform full-process scheduling analysis of the energy storage system and generate a report to obtain a full-process report on the power scheduling of the energy storage system.

[0051] In this embodiment of the invention, multi-type sensor arrays deployed on battery modules, power conversion devices, and the grid side acquire real-time energy storage monitoring data at high frequency. This data includes parameters such as individual cell SOC, voltage, current, temperature, charge / discharge status, energy storage device operating status, and grid frequency fluctuations, all synchronized with a unified timestamp. Based on this, the system invokes the currently active energy storage system's power dispatch strategy, comparing the expected control signal output by the strategy with the actual collected energy storage behavior data step-by-step. It then constructs an execution process deviation measurement function based on indicators such as mean square error and mean absolute error to quantify the difference between strategy execution and device response. If the deviation exceeds a preset dynamic threshold (automatically generated from the fluctuation confidence interval within the operating cycle), the state decay analysis module is activated. This module extracts the SOC decay rate curve from the deviation accumulation trend and generates SOC decay data through surface fitting and parametric regression methods. Subsequently, the SOC decay data is used as feedback input to the state estimation subnetwork in the digital twin model of the energy storage system. A weighted update mechanism (such as Bayesian optimization or stochastic gradient update) is used to iteratively reconstruct the network weights and input mapping parameters, thus forming an optimized digital twin model of the energy storage system. After the digital twin model optimization is completed, the system calls the full-process scheduling simulation module to perform charging and discharging path replanning and stability verification against the current system load forecast, electricity price data, and frequency disturbance levels. A multi-objective scheduling solution algorithm generates multi-period operating trajectories and control commands. Finally, the strategy execution results, deviation correction records, SOC evolution trends, scheduling load response, and topology changes are presented in a graphical and textual format to construct a full-process report on energy storage system power dispatch, providing fundamental data support for subsequent operational evaluation and strategy fine-tuning.

[0052] Preferably, step S42 includes the following steps: Step S421: Calculate the deviation value based on real-time energy storage monitoring data and energy storage system power dispatch strategy, and generate energy storage deviation value data; Step S422: Calculate the SOC deviation based on real-time energy storage monitoring data and energy storage system power dispatch strategy to obtain energy storage SOC deviation data; Step S423: If the energy storage deviation data is greater than 5% and the energy storage SOC deviation data is greater than 2%, then the compensation coefficient for error propagation path backtracking is generated, and the SOC attenuation data is obtained.

[0053] In this embodiment of the invention, by comparing real-time energy storage monitoring data with the current scheduling strategy of the energy storage system step by step, the differences between the actual response values ​​and expected scheduling values ​​of key control variables such as voltage, current, and power are extracted, a state variable difference sequence is constructed, and the comprehensive energy storage deviation value data is calculated based on the time-weighted residual function. Subsequently, the system further performs independent error modeling on the SOC state, analyzes the cumulative error in the SOC estimation path through a bidirectional integral model or Kalman filter, compares the calculated theoretical SOC trajectory with the collected actual SOC data, and generates energy storage SOC deviation data. After the deviation data is acquired, the system sets a threshold judgment mechanism. If the comprehensive energy storage deviation value exceeds 5% and the SOC deviation exceeds 2%, the error propagation path analysis module is activated to construct an error path map based on causal inference and residual propagation mechanism. This map uses a multi-scale dynamic Bayesian network or graph convolution mechanism to analyze the influence strength of the deviation source node on the SOC output result, generates error propagation channels across variable dimensions, and extracts the influence path with strong stability by combining the statistical behavior of historical operating intervals. Based on this, a compensation coefficient generation model is constructed. This model performs hierarchical weighting based on the current path propagation weight and deviation magnitude, outputting SOC correction factors in multiple dimensions. Finally, it is superimposed with the original SOC estimate to form the corrected SOC attenuation data, which is used for subsequent digital twin parameter updates and scheduling strategy adjustments. This method forms a complete deviation diagnosis and correction chain from data differencing, path identification, causal modeling to compensation generation, improving the state observability and response stability of energy storage systems in dynamic environments.

[0054] Especially important is step S43: Step S431: The SOC attenuation data is fed back to the digital twin model of the energy storage system for loss parameter optimization to obtain the model loss function; Step S432: Retrain the PINN grid parameters of the digital twin model of the energy storage system based on the model loss function to obtain the model grid parameter data; Step S433: Reconstruct the feature space of the model grid parameter data to obtain the optimized digital twin model of the energy storage system.

[0055] In this embodiment of the invention, by introducing a backpropagation mechanism for SOC decay data and combining the structural advantages of Physical-Informed Neural Networks (PINN), dynamic updates and structural reconstruction of the digital twin model of the energy storage system at the data level are achieved. Specifically, step S431 first compares the SOC decay data generated by the deviation analysis and error backtracking process as the target output with the current output value of the model to construct a composite loss function that includes time series residuals, nonlinear energy loss functions, and battery aging influencing factors. This loss function not only retains the fitting accuracy requirements of the data-driven layer but also incorporates the physical constraints existing in the battery's thermal-electrical-mechanical coupling mechanism, thereby providing a structural optimization target for subsequent network updates. Step S432, based on the constructed loss function, retrains the grid parameters in the PINN model. This process includes not only the backpropagation update of weights and biases in traditional neural networks but also the re-encoding of physical constraint boundary conditions in the spatial-temporal discrete region of the PINN model and the re-optimization of its gradient response function. During training, the system uses the SOC decay trend curve as the main supervisory signal to guide the neural network in updating the feature mapping relationships related to battery health in specific grid regions. It particularly focuses on the error distribution characteristics of current density gradient, voltage stability distribution, and temperature evolution path within multi-scale grid cells, using adaptive gradient descent and L-BFGS optimizers to achieve high-precision weight adjustment and structure fitting. Step S433 further reconstructs the updated model grid parameter data in a high-dimensional feature space, rebuilding the interdependencies and nonlinear correlations between features, and outputting an optimized digital twin model of the energy storage system with the ability to recognize the current system state. During this reconstruction, methods such as Tensor Principal Component Analysis (TPCA) and Laplace eigenmaps are used to extract significant factors in the model's higher-order structure to ensure that the updated twin model maintains physical consistency while possessing high predictive generalization ability, thereby achieving iterative evolution of the model under data-physical dual-driven conditions.

[0056] This specification provides a power dispatching system based on an energy storage system for executing the aforementioned power dispatching method based on an energy storage system. The power dispatching system based on an energy storage system includes: The multi-source heterogeneous sensing standardization module is used to deploy a distributed sensor array within the battery module to collect multi-source heterogeneous data and perform protocol conversion to obtain an initial multimodal dataset. The initial multimodal dataset is then subjected to noise spectrum adjustment filtering thresholds and long-term drift compensation to generate a standardized energy storage dataset. The cross-modal feature fusion module is used to perform tensor completion on the standardized energy storage dataset to generate energy storage tensor completion data; the energy storage tensor completion data is then analyzed in terms of time and space dimensions, and cross-modal feature fusion is performed to generate an energy storage cross-modal feature space; The digital twin modeling and game-theoretic scheduling module is used to construct a digital twin of the battery health status based on the cross-modal feature space of energy storage, thereby obtaining a digital twin model of the energy storage system; and to generate a game-theoretic scheduling strategy from the digital twin model of the energy storage system, thereby obtaining the energy scheduling strategy of the energy storage system. The real-time deviation analysis and closed-loop optimization module is used to acquire real-time energy storage monitoring data; based on the real-time energy storage monitoring data and the energy storage system's power dispatch strategy, it calculates the execution process deviation. If the energy storage deviation data is greater than a preset threshold, it generates SOC attenuation data; the SOC attenuation data is sent back to the energy storage system's digital twin model for network parameter optimization, and a full-process scheduling analysis of the energy storage system is performed to obtain full-process scheduling analysis data; based on the full-process scheduling analysis data, a scheduling report is generated to obtain a full-process report on the energy storage system's power dispatch.

[0057] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0058] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for electrical energy dispatching based on energy storage system, characterized in that, The method comprises the following steps: Step S1: deploying a distributed sensing array inside the battery module to collect multi-source heterogeneous data, and performing protocol conversion to obtain an initial multi-modal data set; performing noise spectrum adjustment filtering threshold on the initial multi-modal data set, and performing long-term drift compensation to generate a standardized energy storage data set; Step S2: performing tensor completion on the standardized energy storage data set to generate energy storage tensor completion data; performing time-space dimension analysis on the energy storage tensor completion data, and performing cross-modal feature fusion to generate an energy storage cross-modal feature space; Step S3: constructing a battery health state digital twin based on the energy storage cross-modal feature space to obtain an energy storage system digital twin model; performing game scheduling strategy generation on the energy storage system digital twin model to obtain an energy storage system power scheduling strategy; Step S4: obtaining real-time energy storage monitoring data; Performing execution process deviation calculation based on the real-time energy storage monitoring data and the energy storage system power scheduling strategy, and if the energy storage deviation data is greater than a preset threshold, generating SOC attenuation data; Returning the SOC attenuation data to the energy storage system digital twin model for network parameter optimization, and performing energy storage system full-process scheduling analysis to obtain full-process scheduling analysis data; Generating a scheduling report based on the full-process scheduling analysis data to obtain an energy storage system power scheduling whole-process report.

2. The method of claim 1, wherein, Step S1 comprises the following steps: Step S11: deploying a voltage sensor and a temperature sensor inside the battery module to collect single cell temperature data and terminal voltage data at a frequency of 10 Hz, wherein the accuracy of the voltage sensor is 0.1%, and the accuracy of the temperature sensor is 0.5℃; Step S12: deploying a high-frequency current sensor with a bandwidth of 20 kHz on the DC side of the energy storage converter to collect charge and discharge current waveforms at a frequency of 1 kHz to generate energy storage charge and discharge waveform data; Step S13: deploying a synchronous phase measurement unit at a grid access point to collect grid frequency, voltage phase angle and harmonic distortion rate at a frequency of 50 Hz to generate grid phase raw data; Step S14: performing protocol conversion on the single cell temperature data, the terminal voltage data, the energy storage charge and discharge waveform data and the grid phase raw data through a multi-protocol adapter to generate an initial multi-modal data set; Step S15: performing noise spectrum adjustment filtering threshold on the initial multi-modal data set, and performing long-term drift compensation to generate a standardized energy storage data set.

3. The method of claim 1, wherein, The tensor completion of the standardized energy storage data set in step S2 comprises: Mapping the standardized energy storage data set to a three-dimensional space for timestamp sorting to obtain energy storage three-dimensional mapping standardized data; Repairing the temperature and current data of the energy storage three-dimensional mapping standardized data by low-rank Tucker decomposition to obtain standard tensor repair data; Constructing a three-dimensional feature tensor of device-time-feature from the standard tensor repair data, and performing missing value detection to obtain energy storage tensor completion data.

4. The method of claim 1, wherein, The time-space dimension analysis and cross-modal feature fusion of the energy storage tensor completion data in step S2 comprise: Extracting time dimension data from the energy storage tensor completion data to generate battery SOC data; Extracting spatial dimension data from the energy storage tensor completion data to generate grid frequency fluctuation rate data; The battery SOC data and the grid frequency fluctuation rate data are time axis aligned to generate a synchronized energy storage data set; The battery SOC data is subjected to long-term dependence relationship analysis to obtain battery dependence relationship analysis data; the battery dependence relationship analysis data is subjected to charge and discharge behavior feature extraction to obtain a time step feature vector; The grid frequency fluctuation rate data is subjected to unit topology impedance calculation to obtain grid topology impedance calculation data; the grid topology impedance calculation data is subjected to charge and discharge path electric energy loss relationship analysis and spatial distribution feature extraction to obtain a spatial dimension feature matrix; The time step feature vector and the spatial dimension feature matrix are subjected to tensor splicing to obtain an energy storage feature splicing matrix; the energy storage feature splicing matrix is subjected to synchronization processing by using the synchronized energy storage data set to obtain an energy storage scheduling feature fusion matrix; The energy storage scheduling feature fusion matrix is subjected to time-space initial weight distribution and feature space construction to obtain an energy storage cross-modal feature space.

5. The electrical energy dispatch method based on energy storage system according to claim 1, wherein, Step S3 comprises the following steps: Step S31: based on the energy storage cross-modal feature space, multi-scale energy storage analysis is performed, and a battery health state digital twin is constructed to obtain an energy storage system digital twin model; Step S32: based on the energy storage system digital twin model, a double-layer optimization framework is designed to obtain an energy storage double-layer optimization framework, wherein the energy storage double-layer optimization framework comprises an upper-layer consumption reduction framework and a lower-layer energy consumption stability framework; Step S33: based on the upper-layer consumption reduction framework and the lower-layer energy consumption stability framework, a game scheduling strategy is generated to obtain an energy storage system electric energy scheduling strategy.

6. The electrical energy dispatch method based on energy storage system according to claim 5, wherein, Step S31 comprises the following steps: Step S311: based on the energy storage cross-modal feature space, ohmic polarization / diffusion polarization loss calculation is performed to generate electrochemical scale analysis data; Step S312: based on the energy storage cross-modal feature space, thermodynamic analysis of capacity attenuation prediction is performed to generate thermodynamic scale analysis data; Step S313: based on the energy storage cross-modal feature space, electrode stress crack detection is performed, and crack diffusion rate measurement is calculated to generate mechanical scale analysis data; Step S314: the electrochemical scale analysis data, the thermodynamic scale analysis data and the mechanical scale analysis data are subjected to parameter initialization, and a digital twin is constructed to generate an energy storage system digital twin model.

7. The electric energy dispatching method based on the energy storage system according to claim 5, characterized in that, Step S33 comprises the following: Based on the upper-layer consumption reduction framework and the lower-layer energy consumption stability framework, a game scheduling strategy is generated to obtain an energy storage system electric energy scheduling strategy, wherein the energy storage system electric energy scheduling strategy comprises a consumption reduction optimization strategy and a frequency response optimization strategy; Obtain electricity price data; perform load prediction extraction on the upper-layer consumption reduction framework to obtain energy storage load prediction data; based on the load prediction data and the electricity price data, perform 24-hour linear charge and discharge planning to obtain the consumption reduction optimization strategy; Real-time grid frequency data is obtained by performing real-time grid frequency extraction on the lower-layer energy consumption stability framework; when the real-time grid frequency data exceeds 0.2 Hz, an energy storage correction instruction is triggered for adjustment to obtain the frequency response optimization strategy.

8. The electrical energy dispatch method based on energy storage system according to claim 1, wherein, Step S4 comprises the following steps: Step S41: obtain real-time energy storage monitoring data; Step S42: based on real-time energy storage monitoring data and energy storage system electric energy scheduling strategy, execute process deviation calculation, if the energy storage deviation data is greater than the preset threshold, generate SOC attenuation data; Step S43: return the SOC attenuation data to the energy storage system digital twin model for network parameter optimization, and generate an optimized energy storage system digital twin model; Step S44: based on the optimized energy storage system digital twin model, perform energy storage system full-process scheduling analysis, and generate a report, obtaining an energy storage system electric energy scheduling full-process report.

9. The electrical energy dispatch method based on energy storage system according to claim 1, wherein, Step S42 includes the following steps: Step S421: based on real-time energy storage monitoring data and energy storage system electric energy scheduling strategy, calculate the deviation value, and generate energy storage deviation value data; Step S422: based on real-time energy storage monitoring data and energy storage system electric energy scheduling strategy, calculate the SOC deviation, and obtain energy storage SOC deviation data; Step S423: if the energy storage deviation value data is greater than 5% and the energy storage SOC deviation data is greater than 2%, perform error propagation path backtracking compensation coefficient generation, and obtain SOC attenuation data.

10. An electrical energy dispatch system based on an energy storage system, characterized by, A method for performing the energy storage system-based electric energy scheduling method as claimed in claim 1, the energy storage system-based electric energy scheduling system comprising: A multi-source heterogeneous perception standardization module for deploying a distributed sensing array inside a battery module to collect multi-source heterogeneous data, and performing protocol conversion to obtain an initial multi-modal data set; performing noise spectrum adjustment filtering threshold on the initial multi-modal data set, and performing long-term drift compensation to generate a standardized energy storage data set; A cross-modal feature fusion module for performing tensor completion on the standardized energy storage data set to generate energy storage tensor completion data; performing time-space dimension analysis on the energy storage tensor completion data, and performing cross-modal feature fusion to generate an energy storage cross-modal feature space; A digital twin modeling and game scheduling module for constructing a battery health state digital twin based on the energy storage cross-modal feature space to obtain an energy storage system digital twin model; performing game scheduling strategy generation on the energy storage system digital twin model to obtain an energy storage system electric energy scheduling strategy; A real-time deviation analysis and closed-loop optimization module for obtaining real-time energy storage monitoring data; based on real-time energy storage monitoring data and energy storage system electric energy scheduling strategy, performing execution process deviation calculation, if the energy storage deviation data is greater than the preset threshold, generating SOC attenuation data; returning the SOC attenuation data to the energy storage system digital twin model for network parameter optimization, and performing energy storage system full-process scheduling analysis to obtain full-process scheduling analysis data; based on the full-process scheduling analysis data, performing scheduling report generation to obtain an energy storage system electric energy scheduling full-process report.

Images