Energy storage scheduling method and system, computer equipment and computer program product

Abstract

The invention is suitable for the technical field of energy scheduling, and provides an energy storage scheduling method and system, computer equipment and a computer program product, and the method comprises the steps: carrying out the state synchronization of a digital twin model and a physical energy storage system; performing simulation prediction based on the digital twinborn model to obtain predicted time sequence data; and simulation iteration is carried out based on the operation data of the digital twin model and the prediction time sequence data, and a target scheduling decision sequence is output, and the target scheduling decision sequence is used for indicating energy storage scheduling of the physical energy storage system. According to the energy storage scheduling method provided by the invention, accurate scheduling of the physical energy storage system can be realized, so that the benefit of the physical energy storage system in the life cycle is maximized.

Description

Technical Field

[0001] This application belongs to the field of energy dispatching technology, and in particular relates to an energy storage dispatching method, system, computer equipment and computer program product. Background Technology

[0002] In recent years, with the acceleration of the global energy transition and the continuous expansion of renewable energy installations, electrochemical energy storage, as a key technology for absorbing fluctuating renewable energy, has entered the stage of large-scale application. According to global energy data, the cumulative installed capacity of global electrochemical energy storage exceeded 200GW in 2024, with the Chinese market experiencing an annual growth rate exceeding 100%. Against this backdrop, improving the economic benefits and operational efficiency of energy storage systems has become a focus of industry attention, and the importance of energy storage operation and maintenance scheduling technology is increasingly prominent.

[0003] Compared with traditional generator sets, electrochemical energy storage systems have significantly different operating characteristics, which makes it difficult to directly apply traditional power system dispatching theories and methods to energy storage systems.

[0004] Therefore, there is an urgent need for an energy storage scheduling method to achieve precise scheduling of energy storage systems and maximize the benefits of energy storage systems throughout their entire life cycle. Summary of the Invention

[0005] In view of this, embodiments of this application provide an energy storage scheduling method, system, computer equipment, and computer program product to achieve precise scheduling of physical energy storage systems, thereby maximizing the benefits of the physical energy storage systems throughout their life cycle.

[0006] The first aspect of this application provides an energy storage scheduling method, including:

[0007] Synchronize the state of the digital twin model with the physical energy storage system;

[0008] Simulation prediction is performed based on the digital twin model to obtain predicted time series data;

[0009] Simulation iterations are performed based on the operational data of the digital twin model and the predicted time series data, and a target scheduling decision sequence is output. The target scheduling decision sequence is used to instruct the energy storage scheduling of the physical energy storage system.

[0010] In one implementation of the first aspect, the state synchronization of the digital twin model with the physical energy storage system includes:

[0011] Obtain operational status data of the physical energy storage system, including operational status data of the battery pack system and the power conversion system within the physical energy storage system;

[0012] Obtain the operating environment data of the physical energy storage system;

[0013] The operating status data and operating environment data of the physical energy storage system are mapped to the operating data of the digital twin model.

[0014] In one implementation of the first aspect, the step of performing simulation prediction based on the digital twin model to obtain predicted time series data includes:

[0015] Collect external data from multiple sources and map it to the basic input vector of the digital twin model;

[0016] Multiple scenario input vectors are generated based on different time-sharing price prediction sequences, different load prediction curves, and the basic input vector;

[0017] Each of the multiple scene input vectors is input into the digital twin model to obtain the predicted time series data corresponding to each scene input vector.

[0018] In one implementation of the first aspect, each scenario input vector includes a charge / discharge strategy, and different charge / discharge strategies correspond to different depths of discharge (DOD), charge / discharge rates (CR), ambient temperatures (T), and average states of charge (SOC). avg There is at least one difference;

[0019] The step of inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector includes:

[0020] For each scenario input vector, the depth of discharge (DOD), charge / discharge rate (CR), ambient temperature (T), and average state of charge (SOC) corresponding to the charge / discharge strategy are respectively assigned. avg Input the battery health state loss assessment model, based on ΔS=f(DOD,CR,T,SOC) avg The battery health state loss ΔS corresponding to each of the scenario input vectors is obtained;

[0021] The battery health state loss ΔS corresponding to each scenario input vector is input into the economic cost conversion model, based on... The economic loss cost C(ΔS) corresponding to the loss of health status ΔS is obtained.

[0022] In one implementation of the first aspect, the scenario input vector includes power load rate P, ambient temperature T, and equipment aging degree A;

[0023] The step of inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector includes:

[0024] Perform the following steps for each of the plurality of scene input vectors:

[0025] Obtain the power load rate P, ambient temperature T, and equipment aging level A from the scenario input series;

[0026] The efficiency η of the power conversion system corresponding to the scenario input vector is calculated based on η(P, T, A) = base_η(P)·f_T(T)·f_A(A), where base_η(P) is the basic efficiency curve, which describes the efficiency curve of the power conversion system with load variation under standard temperature and initial equipment conditions, f_T(T) is the temperature influence factor, and f_A(A) is the aging influence factor.

[0027] The efficiency η of the power conversion system corresponding to each of the multiple scene input vectors is obtained.

[0028] In one implementation of the first aspect, the simulation iteration based on the operational data of the digital twin model and the predicted time series data, and the output of the target scheduling decision sequence, includes:

[0029] The running data of the digital twin model and the predicted time series data are concatenated into an observation vector;

[0030] The observation vector is input into the agent's current policy network to obtain candidate scheduling instructions;

[0031] The candidate scheduling instructions are input into the digital twin model for simulation, and a reward sequence is generated based on the simulation results;

[0032] The estimated value of the agent's value network is calculated based on the reward sequence;

[0033] Gradient descent is performed to update the value network based on the error between the target value and the estimated value.

[0034] The current policy network is updated based on the advantage function of the value network;

[0035] The process of returning to the step of inputting the observation vector into the agent's current policy network to obtain candidate scheduling instructions continues until a preset iteration condition is triggered.

[0036] The candidate scheduling instructions are output as the target scheduling decision sequence.

[0037] In one implementation of the first aspect, after performing simulation iterations based on the operational data of the digital twin model and the predicted time series data, and outputting the target scheduling decision sequence, the method further includes:

[0038] Energy storage scheduling instructions are generated based on the target scheduling decision sequence, and the energy storage scheduling instructions are transmitted to the physical energy storage system.

[0039] A second aspect of this application provides an energy storage dispatch system, including:

[0040] The state synchronization module is used to synchronize the state of the digital twin model with the physical energy storage system;

[0041] The simulation prediction module is used to perform simulation prediction based on the digital twin model to obtain predicted time series data;

[0042] The scheduling decision module is used to perform simulation iterations based on the running data of the digital twin model and the predicted time series data, and output the target scheduling decision sequence.

[0043] A third aspect of this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.

[0044] A fourth aspect of this application provides a computer program product including a computer program that, when run, causes the method described in the first aspect to be executed.

[0045] The beneficial effect of the first aspect of the embodiments of this application is that: by synchronizing the state of the digital twin model with the physical energy storage system, the digital twin model has the latest operating data of the physical energy storage system. Then, based on the state-synchronized digital twin model, simulation prediction is performed to obtain predicted time series data. Simulation iteration is performed based on the operating data of the digital twin model and the predicted time series data, and a target scheduling decision sequence is output. The target scheduling decision sequence is used to instruct the energy storage scheduling of the physical energy storage system, so as to achieve precise scheduling of the physical energy storage system and maximize the benefits of the physical energy storage system throughout its life cycle.

[0046] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram illustrating the implementation process of the energy storage scheduling method provided in the embodiments of this application;

[0049] Figure 2 This is a schematic diagram illustrating the implementation process of the energy storage scheduling method provided in the embodiments of this application;

[0050] Figure 3 This is a schematic diagram illustrating the implementation process of the energy storage scheduling method provided in the embodiments of this application;

[0051] Figure 4 This is an example diagram of the energy storage dispatching system provided in the embodiments of this application;

[0052] Figure 5 This is a schematic diagram of the computer device provided in an embodiment of this application;

[0053] Figure 6 This is a schematic diagram of a computer program product provided in an embodiment of this application. Detailed Implementation

[0054] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0055] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0056] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0057] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0058] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0059] References to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0060] This application provides an energy storage scheduling method for achieving precise scheduling of physical energy storage systems, thereby maximizing the benefits of the physical energy storage systems throughout their lifecycle. The energy storage scheduling method provided in this application synchronizes the state of a digital twin model with the physical energy storage system, ensuring the digital twin model has the latest operational data of the physical energy storage system. Then, simulation prediction is performed based on the synchronized digital twin model to obtain predicted time-series data. Simulation iterations are then performed based on the operational data of the digital twin model and the predicted time-series data, and a target scheduling decision sequence is output. This target scheduling decision sequence instructs the energy storage scheduling of the physical energy storage system, thereby achieving precise scheduling of the physical energy storage system and maximizing its benefits throughout its lifecycle.

[0061] The energy storage scheduling method provided in this application can be applied to computer devices such as mobile phones, tablets, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptops, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), servers, and cloud servers. This application does not impose any restrictions on the specific type of computer device.

[0062] like Figure 1 As shown in the figure, this application provides an energy storage scheduling method, including:

[0063] Step S1: Synchronize the state of the digital twin model with the physical energy storage system.

[0064] In applications, the physical energy storage system, or physical energy storage system layer, mainly includes a battery pack system, a PCS (Power Conversion System), an environmental monitoring system, and a field control system. The battery pack system comprises multiple battery modules, each equipped with a temperature sensor, current sensor, and voltage sensor; the PCS system includes power conversion units and monitoring units; the environmental monitoring system monitors parameters such as ambient temperature and humidity; and the field control system executes scheduling commands and provides feedback on the execution status.

[0065] In applications, a digital twin mapping layer enables data interaction and state synchronization between the physical energy storage system layer and the digital twin model. The digital twin mapping layer includes a data acquisition interface, a data preprocessing module, a state mapping module, and a command transmission module. This layer acquires real-time operational data from physical devices in the physical energy storage system using industrial communication protocols (such as Modbus and IEC61850), and after data preprocessing such as data cleaning and anomaly detection, maps it to the digital twin model.

[0066] Step S2: Perform simulation prediction based on the digital twin model to obtain predicted time series data.

[0067] In applications, simulations based on digital twin models are used to obtain predicted time-series data for future predetermined times, such as the next 6 hours, 12 hours, or 24 hours. It can be understood that the next 6 hours refers to the 6 hours starting from the point in time when the simulation prediction is performed based on the digital twin model.

[0068] In applications, the predicted time-series data includes power trajectory, efficiency curve, battery health status degradation, and economic benefit prediction.

[0069] Step S3: Based on the running data of the digital twin model and the predicted time series data, perform simulation iteration and output the target scheduling decision sequence, which is used to indicate the energy storage scheduling of the physical energy storage system.

[0070] In application, simulation iterations are performed based on the operational data and predicted time series data of the digital twin model until the target scheduling decision sequence is obtained and output. Based on this target scheduling decision sequence, the physical energy storage system can be scheduled to achieve precise scheduling of the physical energy storage system and maximize the benefits of the physical energy storage system throughout its life cycle.

[0071] In one embodiment, step S1, synchronizing the state of the digital twin model with the physical energy storage system, includes:

[0072] Step S11: Obtain the operating status data of the physical energy storage system, including the operating status data of the battery pack system and the power conversion system in the physical energy storage system.

[0073] In applications, the operational status data of a battery pack system includes the voltage, current, temperature, state of charge, state of health, and cycle count of the battery pack or individual cells within it. The operational status data of a power conversion system (PCS) includes power load rate, ambient temperature, equipment aging level, efficiency, voltage level, operating mode, and fault information. The operational status data of a physical energy storage system can be directly collected through a sensor cluster within the physical energy storage system.

[0074] Step S12: Obtain the operating environment data of the physical energy storage system.

[0075] In applications, the operating environment data of physical energy storage systems includes real-time ambient temperature, humidity, irradiance, wind speed, grid status, electricity price information, and load information of the environment in which the physical energy storage system is located. Specifically, in scenarios with integrated photovoltaic or wind power, irradiance and wind speed are acquired through photovoltaic irradiance sensors and anemometers; grid status, including grid voltage and grid frequency, is acquired in real time through SCADA (Supervisory Control and Data Acquisition) systems or DCS (Distributed Control System) systems to determine grid-connected / off-grid switching conditions.

[0076] Step S13: Map the operating status data of the physical energy storage system and the operating environment data to the operating data of the digital twin model.

[0077] In application, the operational status and environmental data of the physical energy storage system are preprocessed, including data cleaning, missing value imputation, and outlier removal, and then mapped to the operational data of the digital twin model. This synchronizes the digital twin model with the physical energy storage system's state. After synchronization, the digital twin system's operational data includes the latest operational status and environmental data of the physical energy storage system.

[0078] like Figure 2 As shown, in one embodiment, step S2, which involves performing simulation prediction based on the digital twin model to obtain predicted time series data, includes:

[0079] Step S21: Collect multi-source external data and map it to the basic input vector of the digital twin model.

[0080] In applications, multi-source external data includes external environmental data of the physical energy storage system and power load rate P, ambient temperature T, and equipment aging level A of the power conversion system. Mapping the processed multi-source external data to the input vector x(t) of the digital twin model yields:

[0081] x(t)=[P(t),T(t),A(t),Hum(t),Irr(t),V grid (t),f grid (t),p price (t),p load (t)];

[0082] Where P, T, and A represent the power load rate of the power conversion system, ambient temperature, and equipment aging degree, respectively, and Irr represents the illuminance and V. grid f is the grid voltage. grid p is the power grid frequency. price For grid electricity price, p load For the grid load, the above data is time-series data with respect to t.

[0083] Step S22: Generate multiple scenario input vectors based on different time-sharing price prediction sequences, different load prediction curves, and the basic input vector.

[0084] Step S23: Input each of the multiple scene input vectors into the digital twin model to obtain the predicted time series data corresponding to each scene input vector.

[0085] In applications, the predicted time-series data includes the predicted battery health status loss and corresponding economic loss costs over a future period, as well as the efficiency of the power conversion system.

[0086] In one embodiment, each scenario input vector includes a charge / discharge strategy, and different charge / discharge strategies correspond to different depths of discharge (DOD), charge / discharge rates (CR), ambient temperatures (T), and average states of charge (SOC). avg There is at least one difference.

[0087] In applications, the depth of discharge (DOD) for each charge / discharge strategy refers to the ratio of the amount of electricity discharged in one discharge cycle to the battery's rated capacity. A larger DOD results in higher chemical and mechanical stress on the battery, leading to accelerated state of health (SOH) degradation. The charge / discharge rate (C-rate, CR) represents the rate at which the battery is charged and discharged according to a multiple of its rated capacity. A higher C-rate results in faster ion movement within the battery, which also accelerates SOH degradation. The ambient temperature (T) refers to the temperature of the environment in which the battery is located, or it can be the battery temperature. High or low temperatures accelerate electrode / electrolyte side reactions, affecting cycle life and calendar life. For every 10°C change in temperature, the SOH degradation rate can double or multiply. The average state of charge (SOC_avg) refers to the average state of charge (SOC) maintained by the battery during one cycle. If the battery remains in a high SOC range (e.g., >80%) or a low SOC range (e.g., <20%) for an extended period, it will lead to accelerated SOH degradation.

[0088] Step S23 involves inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector, including:

[0089] Step S231: For each scene input vector, the depth of discharge (DOD), charge / discharge rate (CR), ambient temperature (T), and average state of charge (SOC) corresponding to the charge / discharge strategy are respectively... avg Input the battery health state loss assessment model, based on ΔS=f(DOD,CR,T,SOC) avg The battery health state loss ΔS corresponding to each scenario input vector is obtained.

[0090] In application, ΔS = f(DOD, CR, T, SOC) avg )

[0091] =k_base·f_DOD(DOD)·f_CR(CR)·f_T(T)·f_SOC(SOC_avg);

[0092] Among them, f_DOD(DOD)=a_1·DOD^2+b_1·DOD+c_1;

[0093] f_CR(CR)=a_2·CR^2+b_2·CR+c_2;

[0094] f_T(T)=a_3·exp(b_3·(T-T_ref));

[0095] f_SOC(SOC_avg)=a_4·(SOC_avg-0.5)^2+b_4.

[0096] In the above formula, k_base is the baseline decay coefficient, representing the baseline SOH loss in a single cycle under standard conditions (e.g., DOD = 50%, CR = 1C, T = 25℃, SOC_avg = 50%). It is used as the normalization coefficient of the model to adjust for differences in different battery types or experimental data. If a battery loses 0.1% in a single cycle under standard conditions, then k_base = 0.001. f_DOD (DOD) is the nonlinear effect function of depth of discharge (DOD) on SOH loss. The higher the DOD, the more the battery active materials participate in the reaction, leading to increased mechanical stress and chemical side reactions. The quadratic term (a_1·DOD^2) reflects the nonlinear acceleration effect of high DOD. By fitting constants a_1, b_1, and c_1 to experimental data, for example, a_1 = 0.01, b_1 = 0.05, c_1 = 0.1, then when DOD = 100%, f_DOD(DOD) = 0.01·1 + 0.05·1 + 0.1 = 0.16; f_CR(CR) is the effect function of charge / discharge rate (CR) on SOH loss, and the quadratic term (a_2·CR^2) reflects the nonlinear acceleration effect of high rate. By fitting constants a_2, b_2, and c_2 to experimental data, for example, a_2 = 0.02, b_2 = 0.1, c_2 = 0.05, then when CR = 1C, f_CR(CR) = 0.02·1 + 0.1·1 + 0.05 = 0.17; f_T(T) is the exponential effect function of ambient temperature (T) on SOH loss, where T_ref is the reference temperature (e.g., 25°C). ℃), b_3 reflects temperature sensitivity. The constants a_3 and b_3 are fitted through experimental data. For example, a_3 = 1, b_3 = 0.1, T_ref = 25℃, then when T = 35℃, f_T(T) = exp(0.1·10) ≈ 2.718; f_SOC(SOC_avg) is the effect function of the average state of charge (SOC_avg) on ​​SOH loss. The quadratic term a_4·(SOC_avg-0.5)^2 reflects the nonlinear acceleration effect of SOC deviating from the median value (50%). The constants a_4 and b_4 are fitted through experimental data. For example, a_4 = 0.05, b_4 = 0.1, then when SOC_avg = 70%, f_SOC(SOC_avg) = 0.05·(0.7_0.5)2 + 0.1 = 0.05·0.04 + 0.1 = 0.102.

[0097] Step S232: Input the battery health state loss ΔS corresponding to each scenario input vector into the economic cost conversion model, based on... The economic loss cost C(ΔS) corresponding to the loss of health status ΔS is obtained.

[0098] In applications, C bat C represents the unit capacity cost of the battery (yuan / kWh), which is the initial investment cost per kWh of capacity. For example, if the unit price of a lithium battery is 1500 yuan / kWh, then C... bat =1500; CAP bat The total battery capacity (kWh) is the rated energy storage capacity of the battery system. If the energy storage system capacity is 1000kWh, then CAP... bat =1000; ΔS is the SOH loss per cycle (dimensionless, range 0-1), calculated by the battery health status loss assessment model. If ΔS = 0.001 (i.e., 0.1%), then the SOH loss per cycle is 0.1%. EOL This is the battery retirement threshold (usually 60%), meaning the battery needs to be retired when the State of Hypoxia (SOH) drops to this value. EOL If the value is 0.6, then the total acceptable SOH decay is 40%.

[0099] In applications, the above steps can be used to accurately assess the full lifecycle cost of an energy storage system under different charging strategies, providing SOH loss cost support for energy storage scheduling.

[0100] In one embodiment, the scenario input vector includes power load rate P, ambient temperature T, and equipment aging degree A;

[0101] Step S23 involves inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector, including:

[0102] Step S24: Perform the following steps for each of the plurality of scene input vectors:

[0103] Step S241: Obtain the power load rate P, ambient temperature T, and equipment aging degree A from the scenario input series.

[0104] Step S242: Calculate the efficiency η of the power conversion system corresponding to the scene input vector based on η(P, T, A) = base_η(P)·f_T(T)·f_A(A), where base_η(P) is the basic efficiency curve, which describes the efficiency curve of the power conversion system with load changes under standard temperature and initial equipment conditions, f_T(T) is the temperature influence factor, and f_A(A) is the aging influence factor.

[0105] In applications, the initial device state can be understood as the state of the device under new operating conditions.

[0106] In application, the basic efficiency curve adopts a piecewise polynomial model:

[0107]

[0108] The temperature influence factor adopts an exponential decay model: f_T(T)=1-k_T·(T-T_ref) 2 ;

[0109] The aging influencing factor adopts a linear decay model: f_A(A)=1-k_A·A;

[0110] Where k_T and k_A are attenuation coefficients, and T_ref is the reference temperature.

[0111] Step S243: Obtain the efficiency η of the power conversion system corresponding to each of the multiple scene input vectors.

[0112] like Figure 3 As shown, in one embodiment, step S3, which involves performing simulation iterations based on the operational data of the digital twin model and the predicted time-series data, and outputting a target scheduling decision sequence, includes:

[0113] Step S31: The running data of the digital twin model and the predicted time series data are concatenated into an observation vector.

[0114] In applications, operational data such as battery SOC / SOH, PCS output power, current, voltage, and ambient temperature, along with power trajectory, efficiency curve, SOH degradation, and economic benefit predictions for the next 24 hours, are concatenated into a complete observation vector *st*. The policy network receives this observation vector *st* as input and outputs an action distribution or deterministic action (such as charge / discharge power setting).

[0115] Step S32: Input the observation vector into the agent's current policy network to obtain candidate scheduling instructions.

[0116] Step S33: Input the candidate scheduling instruction into the digital twin model for simulation, and generate a reward sequence based on the simulation results.

[0117] In applications, candidate scheduling instructions are input into a digital twin simulation model for simulation to evaluate their impact on immediate benefits, immediate energy consumption, and future degradation costs.

[0118] Step S34: Calculate the estimated value of the agent's value network based on the reward sequence.

[0119] Step S35: Based on the error between the target value of the value network and the estimated value, perform gradient descent to update the value network.

[0120] Step S36: Update the current policy network based on the advantage function of the value network.

[0121] In applications, the current policy network can be updated using policy gradient ascent or by employing pruning strategies such as PPO.

[0122] Step S37: Return to the step of inputting the observation vector into the agent's current policy network to obtain candidate scheduling instructions, until the preset iteration condition is triggered.

[0123] In the application, the preset iteration condition is that the value network and / or the current policy network converge.

[0124] Step S38: Output the candidate scheduling instructions as the target scheduling decision sequence.

[0125] In applications, the target scheduling decision sequence can achieve a balance between the local and global benefits of physical energy storage systems.

[0126] In applications, multiple agents with different optimization objectives can be trained simultaneously based on the running data of the digital twin model and the predicted time series data. The coordination agent can then coordinate the agents with different optimization objectives to improve the efficiency and performance of the training agents.

[0127] In applications, multi-agent systems can be constructed based on agents and a coordinating agent with the following three different optimization objectives, and each agent in the system can be trained separately:

[0128] The efficiency optimization agent is responsible for optimizing the efficiency of the power conversion system and minimizing energy loss; the health optimization agent is responsible for optimizing the battery health state and minimizing SOH loss; the profit optimization agent is responsible for optimizing economic profits and maximizing market revenue; and the coordination agent is responsible for coordinating the decisions of the above three types of agents and balancing multi-objective optimization.

[0129] The system state space of a multi-agent system includes: battery SOC state, PCS operating parameters, ambient temperature, grid state, market electricity price information, etc., and the action space includes charging and discharging power settings, operating mode selection (such as peak-valley arbitrage, ancillary services, etc.), and charging and discharging strategy parameters (such as rate limit, SOC range, etc.).

[0130] The reward function for the coordinating agent is designed as follows: R = w1·R_revenue - w2·R_loss - w3·R_degradation;

[0131] Where R_revenue represents the market revenue reward for the revenue-optimizing agent, R_loss represents the energy loss penalty for the efficiency-optimizing agent, R_degradation represents the battery degradation penalty for the health-optimizing agent, and w1, w2, and w3 are weighting coefficients that can be dynamically adjusted according to user preferences.

[0132] In one embodiment, after performing simulation iterations based on the operational data of the digital twin model and the predicted time series data, and outputting the target scheduling decision sequence, step S3 further includes:

[0133] Step S4: Generate energy storage scheduling instructions based on the target scheduling decision sequence, and transmit the energy storage scheduling instructions to the physical energy storage system.

[0134] In application, the target scheduling decision sequence can be understood as the optimal scheduling decision sequence. It is transformed into actual energy storage scheduling instructions, including the charging and discharging power setpoints, charging and discharging mode switching, and related strategy parameters at each moment. These instructions are then sent to the physical energy storage system through the instruction transmission module to ensure that the physical energy storage system executes energy storage scheduling instructions consistent with the optimal strategy in the simulation environment of the digital twin model, thereby maximizing the benefits of the physical energy storage system throughout its entire life cycle.

[0135] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0136] This application also provides an energy storage scheduling system for executing the steps described in the above-described energy storage scheduling method embodiments. The energy storage scheduling system can be a virtual appliance within a computer device, run by the computer device's processor, or it can be the computer device itself.

[0137] like Figure 4 As shown, the energy storage dispatching system 400 provided in this application embodiment includes:

[0138] The state synchronization module 401 is used to synchronize the state of the digital twin model with the physical energy storage system;

[0139] The simulation prediction module 402 is used to perform simulation prediction based on the digital twin model to obtain predicted time series data;

[0140] The scheduling decision module 403 is used to perform simulation iteration based on the operating data of the digital twin model and the predicted time series data, and output a target scheduling decision sequence, which is used to instruct the energy storage scheduling of the physical energy storage system.

[0141] In one embodiment, the state synchronization module 401 is used for:

[0142] Obtain operational status data of the physical energy storage system, including operational status data of the battery pack system and the power conversion system within the physical energy storage system;

[0143] Obtain the operating environment data of the physical energy storage system;

[0144] The operating status data and operating environment data of the physical energy storage system are mapped to the operating data of the digital twin model.

[0145] In one embodiment, the simulation prediction module 402 is used for:

[0146] Collect external data from multiple sources and map it to the basic input vector of the digital twin model;

[0147] Multiple scenario input vectors are generated based on different time-sharing price prediction sequences, different load prediction curves, and the basic input vector;

[0148] Each of the multiple scene input vectors is input into the digital twin model to obtain the predicted time series data corresponding to each scene input vector.

[0149] In one embodiment, each scenario input vector includes a charge / discharge strategy, and different charge / discharge strategies correspond to different depths of discharge (DOD), charge / discharge rates (CR), ambient temperatures (T), and average states of charge (SOC). avg There is at least one difference;

[0150] The simulation prediction module 402 is used for:

[0151] For each scenario input vector, the depth of discharge (DOD), charge / discharge rate (CR), ambient temperature (T), and average state of charge (SOC) corresponding to the charge / discharge strategy are respectively assigned. avg Input the battery health state loss assessment model, based on ΔS=f(DOD,CR,T,SOC) avg The battery health state loss ΔS corresponding to each of the scenario input vectors is obtained;

[0152] The battery health state loss ΔS corresponding to each scenario input vector is input into the economic cost conversion model, based on... The economic loss cost C(ΔS) corresponding to the loss of health status ΔS is obtained.

[0153] In one embodiment, the scenario input vector includes power load rate P, ambient temperature T, and equipment aging degree A;

[0154] The simulation prediction module 402 is used for:

[0155] Perform the following steps for each of the plurality of scene input vectors:

[0156] Obtain the power load rate P, ambient temperature T, and equipment aging level A from the scenario input series;

[0157] The efficiency η of the power conversion system corresponding to the scenario input vector is calculated based on η(P, T, A) = base_η(P)·f_T(T)·f_A(A), where base_η(P) is the basic efficiency curve, which describes the efficiency curve of the power conversion system with load variation under standard temperature and initial equipment conditions, f_T(T) is the temperature influence factor, and f_A(A) is the aging influence factor.

[0158] The efficiency η of the power conversion system corresponding to each of the multiple scene input vectors is obtained.

[0159] In one embodiment, the scheduling decision module 403 is configured to:

[0160] The running data of the digital twin model and the predicted time series data are concatenated into an observation vector;

[0161] The observation vector is input into the agent's current policy network to obtain candidate scheduling instructions;

[0162] The candidate scheduling instructions are input into the digital twin model for simulation, and a reward sequence is generated based on the simulation results;

[0163] The estimated value of the agent's value network is calculated based on the reward sequence;

[0164] Gradient descent is performed to update the value network based on the error between the target value and the estimated value.

[0165] The current policy network is updated based on the advantage function of the value network;

[0166] The process of returning to the step of inputting the observation vector into the agent's current policy network to obtain candidate scheduling instructions continues until a preset iteration condition is triggered.

[0167] The candidate scheduling instructions are output as the target scheduling decision sequence.

[0168] In one embodiment, the energy storage scheduling system 400 further includes an instruction transmission module, configured to: after performing simulation iteration based on the operating data of the digital twin model and the predicted time series data, and outputting a target scheduling decision sequence, generate an energy storage scheduling instruction according to the target scheduling decision sequence, and transmit the energy storage scheduling instruction to the physical energy storage system.

[0169] In applications, the modules in an energy storage scheduling system can be software program modules, or they can be implemented through different logic circuits integrated in a processor, or they can be implemented through multiple distributed processors.

[0170] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 5 As shown, the computer device 5 of this embodiment includes: at least one processor 50 ( Figure 5 (Only one is shown) a processor, a memory 51, and a computer program 52 stored in the memory 51 and executable on the at least one processor 50, wherein the processor 50 executes the computer program 52 to implement the steps in any of the above-described energy storage scheduling method embodiments.

[0171] The computer device may include, but is not limited to, a processor 50 and a memory 51. Those skilled in the art will understand that... Figure 5 The computer device 5 is merely an example and does not constitute a limitation on the computer device 5. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0172] The processor 50 can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0173] In some embodiments, the memory 51 may be an internal storage unit of the computer device 5, such as a hard disk or memory of the computer device 5. In other embodiments, the memory 51 may be an external storage device of the computer device 5, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 5. Furthermore, the memory 51 may include both internal and external storage units of the computer device 5. The memory 51 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 51 can also be used to temporarily store data that has been output or will be output.

[0174] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0175] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0176] like Figure 6 As shown, this application embodiment provides a computer program product 6, including a computer program 52. When the computer program 52 is run, the steps in the above-described energy storage scheduling method embodiments are executed.

[0177] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a device / terminal equipment, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0178] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0179] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0180] In the embodiments provided in this application, it should be understood that the disclosed computer devices and methods can be implemented in other ways. For example, the computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0181] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0182] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An energy storage dispatching method, characterized in that, include: Synchronize the state of the digital twin model with the physical energy storage system; Simulation prediction is performed based on the digital twin model to obtain predicted time series data; Simulation iterations are performed based on the operational data of the digital twin model and the predicted time series data, and a target scheduling decision sequence is output. The target scheduling decision sequence is used to instruct the energy storage scheduling of the physical energy storage system.

2. The energy storage dispatching method as described in claim 1, characterized in that, The process of synchronizing the state of the digital twin model with the physical energy storage system includes: Obtain operational status data of the physical energy storage system, including operational status data of the battery pack system and the power conversion system within the physical energy storage system; Obtain the operating environment data of the physical energy storage system; The operating status data and operating environment data of the physical energy storage system are mapped to the operating data of the digital twin model.

3. The energy storage dispatching method as described in claim 1, characterized in that, The simulation prediction based on the digital twin model to obtain predicted time series data includes: Collect external data from multiple sources and map it to the basic input vector of the digital twin model; Multiple scenario input vectors are generated based on different time-sharing price prediction sequences, different load prediction curves, and the basic input vector; Each of the multiple scene input vectors is input into the digital twin model to obtain the predicted time series data corresponding to each scene input vector.

4. The energy storage dispatching method as described in claim 3, characterized in that, Each scenario input vector includes a charge / discharge strategy, and the depth of discharge (DOD), charge / discharge rate (CR), ambient temperature (T), and average state of charge (SOC) correspond to different charge / discharge strategies. avg There is at least one difference; The step of inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector includes: For each scenario input vector, the depth of discharge (DOD), charge / discharge rate (CR), ambient temperature (T), and average state of charge (SOC) corresponding to the charge / discharge strategy are respectively assigned. avg Input the battery health state loss assessment model, based on ΔS=f(DOD,CR,T,SOC) avg The battery health state loss ΔS corresponding to each of the scenario input vectors is obtained; The battery health state loss ΔS corresponding to each scenario input vector is input into the economic cost conversion model, based on... The economic loss cost C(ΔS) corresponding to the loss of health status ΔS is obtained.

5. The energy storage dispatching method as described in claim 3, characterized in that, The scenario input vector includes power load rate P, ambient temperature T, and equipment aging level A; The step of inputting each of the multiple scene input vectors into the digital twin model to obtain the predicted time-series data corresponding to each scene input vector includes: Perform the following steps for each of the plurality of scene input vectors: Obtain the power load rate P, ambient temperature T, and equipment aging level A from the scenario input series; The efficiency η of the power conversion system corresponding to the scenario input vector is calculated based on η(P, T, A) = base_η(P)·f_T(T)·f_A(A), where base_η(P) is the basic efficiency curve, which describes the efficiency curve of the power conversion system with load variation under standard temperature and initial equipment conditions, f_T(T) is the temperature influence factor, and f_A(A) is the aging influence factor. The efficiency η of the power conversion system corresponding to each of the multiple scene input vectors is obtained.

6. The energy storage dispatching method as described in claim 2, characterized in that, The simulation iteration based on the operational data of the digital twin model and the predicted time series data, and the output of the target scheduling decision sequence, includes: The running data of the digital twin model and the predicted time series data are concatenated into an observation vector; The observation vector is input into the agent's current policy network to obtain candidate scheduling instructions; The candidate scheduling instructions are input into the digital twin model for simulation, and a reward sequence is generated based on the simulation results; The estimated value of the agent's value network is calculated based on the reward sequence; Gradient descent is performed to update the value network based on the error between the target value and the estimated value. The current policy network is updated based on the advantage function of the value network; The process of returning to the step of inputting the observation vector into the agent's current policy network to obtain candidate scheduling instructions continues until a preset iteration condition is triggered. The candidate scheduling instructions are output as the target scheduling decision sequence.

7. The energy storage dispatching method according to any one of claims 1 to 6, characterized in that, After performing simulation iterations based on the operational data of the digital twin model and the predicted time series data, and outputting the target scheduling decision sequence, the method further includes: Energy storage scheduling instructions are generated based on the target scheduling decision sequence, and the energy storage scheduling instructions are transmitted to the physical energy storage system.

8. An energy storage dispatch system, characterized in that, include: The state synchronization module is used to synchronize the state of the digital twin model with the physical energy storage system; The simulation prediction module is used to perform simulation prediction based on the digital twin model to obtain predicted time series data; The scheduling decision module is used to perform simulation iterations based on the operating data of the digital twin model and the predicted time series data, and output a target scheduling decision sequence, which is used to instruct the energy storage scheduling of the physical energy storage system.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, Includes a computer program, which, when run, causes the method described in any one of claims 1 to 7 to be performed.

Images