An energy storage full-cycle collaborative decision-making method and system based on reverse recursion and dual-mode strategy

Abstract

This invention provides a collaborative decision-making method and system for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy. Starting from the end of the time axis, it traverses each continuous time period in reverse along the time axis. Based on the time period characteristic parameters of the current continuous time period and the SOC level corresponding to the start of the continuous time period, it evaluates the lowest cumulative cost of each SOC level at the start time, calculated from different SOC targets at the end time, under the feasible mode of the current continuous time period. Then, based on the cost information of all continuous time periods, it traverses each continuous time period in forward along the time axis, starting from the initial SOC level of the energy storage battery at the initial time. For each continuous time period, it selects the target strategy mode corresponding to the lowest cumulative cost. The selected target strategy modes are then connected in series. Through dynamic switching of dual modes, it takes into account both grid dispatch compliance and economic benefits, and finally achieves collaborative optimization of the entire lifecycle of energy storage over a long time scale.

Description

Technical Field

[0001] This invention relates to the field of charging and discharging technology, and in particular to a collaborative decision-making method and system for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy. Background Technology

[0002] With the large-scale integration of new energy sources into the power system and the widespread adoption of peak-valley pricing mechanisms, energy storage devices have become core equipment for smoothing out fluctuations in new energy output and improving the stability and economy of power grid operation. Their position in the power system is becoming increasingly prominent. However, the charging and discharging operation of energy storage devices is constrained by a combination of complex factors, including the uncertainty of revenue caused by peak-valley price fluctuations, dispatch mismatch caused by load forecasting deviations, physical limitations of transformer capacity, dynamic dispatching instructions such as peak shaving and valley filling on the grid side, and the constraints of the battery's own state of charge (SOC) on charging and discharging power and duration.

[0003] Several energy storage optimization control schemes have emerged in the existing technology. For example, the Chinese invention patent CN114944661B, "A Three-Stage Optimization Control Method for Microgrids Based on Rolling Optimization of Energy Storage System," discloses rolling optimization and three-stage optimization control technologies and has the basic logic of time-sharing decision-making. However, as a typical model predictive control scheme, its core rolling optimization logic usually only focuses on a limited short period of time in the future. This makes it difficult to predict mandatory dispatch instructions or peak electricity prices on a long time scale such as 48 hours in the present, and it is easy to cause global losses throughout the entire cycle due to short-sighted decision-making. At the same time, this technology does not take into account both instruction tracking and economics. The optimized dynamic coordination requirements are difficult to adapt to the complex scenarios driven by both grid dispatch instructions and electricity price fluctuations, and the overall economic optimization effect is limited over a long period of time. Although the Chinese invention patent with publication number CN112529727A, "Microgrid Energy Storage Dispatch Method, Device and Equipment Based on Deep Reinforcement Learning", proposes an energy storage dispatch scheme based on deep reinforcement learning, which can achieve a certain degree of optimized dispatch, the technical path it adopts is different from that of this invention. Moreover, the scheme does not address the problem of dynamic strategy switching under multiple constraints. It highly overlaps with the core technical problem solved by this invention but cannot provide an effective dynamic adaptation solution.

[0004] In addition, most existing energy storage control strategies still adopt single-objective optimization models or simple rule control based on fixed thresholds, failing to fully consider the synergistic effects of the aforementioned multiple constraints. In particular, it is difficult to maximize the overall economic benefits of energy storage sites over long timescales such as 48 hours. This deficiency directly leads to low utilization of energy storage assets, insufficient dispatch flexibility, and an inability to fully realize the value of energy storage devices. It is also difficult to adapt to the current high proportion of new energy power system operation requirements. Therefore, there is an urgent need for an energy storage control technology that can coordinate multiple constraints, take into account long-term economic efficiency, and has the ability to dynamically switch between command tracking and economic optimization to solve the shortcomings of existing technologies. Summary of the Invention

[0005] This invention provides a collaborative decision-making method and system for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy, in order to solve the problems mentioned in the background art.

[0006] A collaborative decision-making method for the entire lifecycle of energy storage based on inverse recursion and a bimodal strategy includes: S1: Obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on electricity price changes and power grid dispatch instructions, and define time period characteristic parameters for each continuous time period; S2: Obtain the continuous range of the state of charge (SOC) of the energy storage battery, and discretize the SOC into multiple SOC levels based on the continuous range. Each SOC level corresponds to a unique state of the energy storage battery. S3: Starting from the end of the time axis, traverse each continuous time period backward along the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC level corresponding to the start of the continuous time period, determine the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, evaluate the possible SOC level of the energy storage battery to the end of the current detection time period for each feasible mode, and output the action cost value of the corresponding feasible mode. After traversing all feasible modes, record the minimum action cost value at the SOC level corresponding to the start of the continuous time period. Repeat this operation until all SOC levels at the start of the continuous time period have been traversed. S4: Based on the cost information of all continuous time periods and the optimal strategy at each SOC level calculated in step S3, starting from the initial SOC level of the energy storage battery at the initial moment, traverse each continuous time period in the forward direction along the time axis, select the target strategy mode corresponding to the lowest cumulative cost, and connect the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle, and generate control commands based on the strategy sequence to drive the operation of the energy storage system.

[0007] Preferably, in step S1, based on electricity price changes and grid dispatch instructions along the time axis, the time axis is divided into multiple continuous time periods, including: Extract time points of state changes from electricity price changes and power grid dispatch instructions; The time points of state changes are sorted in chronological order, and the time axis is divided into multiple continuous and non-overlapping time periods using two adjacent nodes as boundaries.

[0008] Preferably, in step S2, the state of charge (SOC) is discretized into multiple SOC levels based on a continuous value range, including: The continuous value range is evenly divided according to the rule of equal spacing, resulting in multiple SOC levels.

[0009] Preferably, in step S3, the specific method for determining the action cost value of transferring the energy storage battery to the end of the current detection period for each feasible mode is as follows: The comprehensive cost incurred during a continuous period is determined by operating according to the feasible mode, the current discrete SOC level, the current specific energy storage action, and the net load forecast curve adapted to the current energy storage environment. The sum of the cumulative cost of subsequent continuous periods and the comprehensive cost of this continuous period is taken as the cumulative action cost value of this continuous period. Among them, the current specific energy storage action is the specific operation instruction selected from the set of energy storage actionable actions that can directly drive the operation of the equipment; The predicted net load forecast curve is a random variable that follows the probability distribution of a family of net load curves predicted based on historical data. The net load forecast curve has time as the horizontal axis and the difference between the total electricity load of the photovoltaic storage system users and the actual output of new energy as the vertical axis, which fully represents the future power supply and demand balance of the system. The total cost includes basic electricity costs, transition costs, and control incentive and penalty costs.

[0010] Preferably, in step S4, traversing each consecutive time period in the forward direction along the time axis and selecting the target strategy pattern corresponding to the lowest cumulative cost for each consecutive time period includes: Obtain the combined range of two non-overlapping strategy classes that fully cover all available energy storage actions over a continuous period, all available energy storage actions, and all net load forecast curves; Select the target strategy pattern that corresponds to the lowest cumulative cost for each consecutive time period.

[0011] Preferably, in step S3, starting from the end of the time axis, each consecutive time period is traversed backwards along the time axis. Based on the time period characteristic parameters of the current continuous time period and the SOC level corresponding to the start of the continuous time period, the feasibility of the command tracking strategy mode and the economic optimization strategy mode under the constraints of the current time period characteristic parameters is determined. Then, for each feasible mode, the possible SOC level of the energy storage battery to be transferred to the end of the current detection time period is evaluated, including: When the time period feature parameters meet the instruction execution requirements, it is determined that the current time period feature parameters meet the instruction tracking strategy mode; when the time period feature parameters meet the security restriction requirements, it is determined that the current time period feature parameters meet the economic optimization strategy mode. A first policy subset is pre-established based on the economic optimization policy model, and a second policy subset is established based on the instruction tracking policy model. Traverse each consecutive time period backward along the time axis. At each time step, based on the feasible mode, dynamically determine the action vector in the first or second strategy subset. Based on the preset reward mechanism, perform weighted evaluation on the output of different strategy subsets to determine the possible SOC level of the energy storage battery to be transferred to the end of the current detection period.

[0012] The preferred method for determining the overall cost is as follows: Calculate the cumulative cost function at each time step under the state of charge node for each consecutive time period, and determine the comprehensive cost based on the output of the cumulative cost function.

[0013] Preferably, the first strategy subset is a set of strategies for energy storage actions and their corresponding energy storage effects, which takes meeting the rigid requirements of power grid dispatch as the primary criterion, ensuring the overall operation of energy storage sites and conforming to the preset control standards issued by the power grid. The second strategy subset is a set of strategies for energy storage actions and their corresponding energy storage effects, which takes maximizing the economic benefits of energy storage throughout its entire life cycle as the primary criterion, and optimizes charging and discharging timing, power and duration, and makes full use of the arbitrage space brought about by peak-valley electricity price differences and fluctuations in new energy output.

[0014] A collaborative decision-making system for the entire lifecycle of energy storage based on inverse recursion and a dual-modal strategy includes: The time period segmentation module is used to obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on the electricity price changes and grid dispatch instructions of the time axis, and define time period characteristic parameters for each continuous time period; The SOC division module is used to obtain the continuous range of values ​​of the state of charge (SOC) of the energy storage battery, and to discretize the SOC into multiple SOC levels based on the continuous range of values. Each SOC level corresponds to a unique state of the energy storage battery. The gear cost assessment module is used to traverse each continuous time period backward from the end of the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC gear corresponding to the start of the continuous time period, it determines the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, for each feasible mode, it evaluates the possible SOC gear of the energy storage battery to the end of the current detection time period and outputs the action cost value of the corresponding feasible mode. After traversing all feasible modes, it records the minimum action cost value at the SOC gear corresponding to the start of the continuous time period. This operation is repeated until all SOC gears at the start of the continuous time period have been traversed. The optimal sequence determination module is used to calculate the cost information of all continuous time periods and the optimal strategy at each SOC level. Starting from the initial SOC level of the energy storage battery at the initial moment, it traverses each continuous time period in the forward direction along the time axis, selects the target strategy mode corresponding to the lowest cumulative cost, and concatenates the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Based on the strategy sequence, it generates control commands to drive the operation of the energy storage system.

[0015] Compared with the prior art, the present invention has achieved the following beneficial effects: By acquiring a timeline with a preset future time scale, and based on electricity price changes and grid dispatch instructions, the timeline is divided into multiple continuous periods. Characteristic parameters are defined for each continuous period, transforming previously difficult-to-quantify constraints such as electricity price fluctuations and dispatch instructions into fixed characteristic parameters for each period (e.g., electricity price, whether it is under control, target power). This provides a clear environmental input basis for subsequent strategy decisions. The characteristic parameters for each period are independently defined, ensuring that subsequent strategy evaluations can accurately match the core needs of different periods (e.g., focusing on instruction tracking during controlled periods and economic charging during low-price periods). The continuous range of the state of charge (SOC) of the energy storage battery is acquired, and based on this continuous range, the SOC is discretized into multiple SOC levels. Each SOC level corresponds to a unique state of the energy storage battery, transforming the infinite-dimensional continuous state space into a finite-dimensional discrete state space. This makes traversing all possible states during the reverse recursion computationally feasible, solving the problem of excessive computational power consumption in decision-making algorithms under continuous states. Furthermore, each SOC level corresponds to a unique position state, ensuring that all possible energy states of the energy storage battery can be accurately represented, avoiding decision-making biases caused by state omissions, and guaranteeing the comprehensiveness of cost calculations during subsequent reverse recursion. This provides a data foundation for globally optimal decision-making. By starting from the end of the time axis and traversing backwards along the time axis for each continuous period, based on the period feature parameters of the currently detected continuous period and the SOC corresponding to the starting point of the continuous period... The system assesses the feasibility of both the command tracking strategy and the economic optimization strategy under the current time period's characteristic parameter constraints. For each feasible mode, it evaluates the possible SOC levels at which the energy storage battery could be transferred to the end of the current detection time period, simultaneously outputting the action cost value for each feasible mode. After traversing all feasible modes, the minimum action cost value is recorded at the SOC level corresponding to the start of the consecutive time period. This process is repeated until all SOC levels at the start of the consecutive time period have been traversed, enabling advance prediction of constraints and benefits in subsequent time periods. It supports both command tracking and economic optimization modes, allowing for flexible selection based on time period characteristics. This satisfies the rigid requirements of grid dispatch commands while maximizing the economic benefits of energy storage assets. The system evaluates the action of the dual-modal strategy. Cost provides an objective and comparable quantitative basis for subsequent optimal strategy selection, avoiding cost assessment bias caused by fuzzy decision-making. Based on the cost information obtained in step S3 for all continuous time periods, starting from the initial SOC level of the energy storage battery at the initial moment, each continuous time period is traversed forward along the time axis. The target strategy mode corresponding to the lowest sum of cost and energy storage state value for the next continuous time period is selected. These selected target strategy modes are then concatenated to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Ultimately, this maximizes the utilization rate of energy storage assets, minimizes electricity costs, and maximizes the dispatch command compliance rate over a long time scale, solving the problem of low asset utilization efficiency caused by the fragmentation of traditional strategies.Ultimately, by dynamically switching between two modes, both grid dispatch compliance and economic benefits were balanced, achieving synergistic optimization of energy storage throughout its entire lifecycle over a long timescale.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0017] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a collaborative decision-making method for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy, as described in an embodiment of the present invention. Figure 2 This is a structural diagram of a full-cycle collaborative decision-making system for energy storage based on reverse recursion and a dual-modal strategy, as described in an embodiment of the present invention. Detailed Implementation

[0019] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] Example 1: This embodiment of the invention provides a collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy, such as... Figure 1 As shown, it includes: S1: Obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on electricity price changes and power grid dispatch instructions, and define time period characteristic parameters for each continuous time period; S2: Obtain the continuous range of the state of charge (SOC) of the energy storage battery, and discretize the SOC into multiple SOC levels based on the continuous range. Each SOC level corresponds to a unique state of the energy storage battery. S3: Starting from the end of the time axis, traverse each continuous time period backward along the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC level corresponding to the start of the continuous time period, determine the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, evaluate the possible SOC level of the energy storage battery to the end of the current detection time period for each feasible mode, and output the action cost value of the corresponding feasible mode. After traversing all feasible modes, record the minimum action cost value at the SOC level corresponding to the start of the continuous time period. Repeat this operation until all SOC levels at the start of the continuous time period have been traversed. S4: Based on the cost information of all continuous time periods and the optimal strategy at each SOC level calculated in step S3, starting from the initial SOC level of the energy storage battery at the initial moment, traverse each continuous time period in the forward direction along the time axis, select the target strategy mode corresponding to the lowest cumulative cost, and connect the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle, and generate control commands based on the strategy sequence to drive the operation of the energy storage system.

[0021] In this embodiment, the optimal strategy at each SOC level is the feasible mode corresponding to the minimum action cost value.

[0022] In this embodiment, the determination of multiple SOC levels is based on the SOC state discretization result of S2. First, the safe continuous range of the energy storage battery SOC is determined, and then the range is divided into finite levels that do not overlap according to equidistant or adaptive rules. All these discretized levels together constitute all possible SOC states at the beginning of each continuous time period, that is, all SOC levels at the beginning of that time period.

[0023] In this embodiment, the future preset time scale is, for example, the next 48 hours.

[0024] In this embodiment, the cost assessment includes the assessment of transition costs, action costs, etc.

[0025] In this embodiment, the continuous value range is 0%-100%.

[0026] In this embodiment, the cost information is the action cost value under different strategy modes for each consecutive time period.

[0027] In this embodiment, the feasible mode is at least one of the instruction tracking strategy mode and the economic optimization strategy mode.

[0028] In this embodiment, the SOC level is one of the non-overlapping intervals divided from the continuous SOC value range of the energy storage battery, serving as a quantitative benchmark node for the energy storage capacity status at the current time period.

[0029] In this embodiment, the time period characteristic parameters include, for example, the electricity price corresponding to the time period, whether the time period is a controlled time period, the target power value corresponding to the controlled time period, and the preset core control objectives (instruction tracking or economic optimization) and switching trigger conditions for each consecutive time period.

[0030] In this embodiment, the SOC level corresponds to a unique position state of the energy storage battery. This position state, i.e., the discrete SOC level, serves as the core application logic in the reverse cost calculation step, acting as the benchmark anchor point for state transition and cost evaluation, and determining the initial state benchmark for the recursive evaluation.

[0031] In this embodiment, if all SOC gears at the starting point of a continuous time period are traversed, it is considered that the strategy for this time period has been calculated. The process is then recursively repeated in reverse along the time axis until all continuous time periods have been traversed.

[0032] The beneficial effects of the above design scheme are as follows: By obtaining a time axis with a preset future time scale, and based on the electricity price changes and grid dispatch instructions on the time axis, the time axis is divided into multiple continuous time periods. For each continuous time period, characteristic parameters are defined, transforming previously difficult-to-quantify constraints such as electricity price fluctuations and dispatch instructions into fixed characteristic parameters for each time period (e.g., electricity price, whether it is controlled, target power). This provides a clear environmental input basis for subsequent strategy decisions. The characteristic parameters for each time period are defined independently, ensuring that subsequent strategy evaluations can accurately match the core needs of different time periods (e.g., focusing on instruction tracking during controlled periods and focusing on economic charging during low-price periods). Furthermore, by obtaining the continuous range of the state of charge (SOC) of the energy storage battery, and discretizing the SOC into multiple SOCs based on this continuous range, the design scheme achieves the following benefits. The OC (Open Center) level corresponds to a unique state of the energy storage battery for each SOC level, transforming the infinite-dimensional continuous state space into a finite-dimensional discrete state space. This makes it possible to traverse all possible states during the reverse recursion process, solving the problem of excessive computational power consumption in decision-making algorithms under continuous states. Furthermore, each SOC level corresponds to a unique position state, ensuring that all possible charge states of the energy storage battery can be accurately represented, avoiding decision-making biases caused by state omissions, and guaranteeing the comprehensiveness of cost calculations during subsequent reverse recursion. This provides a data foundation for globally optimal decision-making. Starting from the end of the time axis, each continuous time period is traversed backward along the time axis, based on the time period characteristic parameters of the currently detected continuous time period and the continuous time... The system determines the feasibility of both the command tracking strategy and the economic optimization strategy under the current time period's characteristic parameter constraints, starting from the beginning of each segment. For each feasible mode, it evaluates the possible SOC levels at which the energy storage battery could be transferred to the end of the current detection time period, while simultaneously outputting the action cost value for each feasible mode. After traversing all feasible modes, the minimum action cost value is recorded at the SOC level corresponding to the beginning of the continuous time period. This process is repeated until all SOC levels at the beginning of the continuous time period have been traversed, enabling advance prediction of constraints and benefits in subsequent time periods. It supports both command tracking and economic optimization modes, allowing for flexible selection based on time period characteristics, satisfying the rigid requirements of grid dispatch commands while maximizing the value of energy storage assets. The economic benefits are assessed by evaluating the action costs of the dual-modal strategy, providing an objective and comparable quantitative basis for subsequent optimal strategy selection. This avoids cost assessment bias caused by fuzzy decision-making. Based on the cost information of all continuous time periods calculated in step S3, starting from the initial SOC level of the energy storage battery at the initial moment, each continuous time period is traversed forward along the time axis. The target strategy mode corresponding to the lowest sum of cost and energy storage state value for the next continuous time period is selected. The selected target strategy modes are then concatenated to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Ultimately, this maximizes the utilization rate of energy storage assets, minimizes electricity costs, and maximizes the dispatch command compliance rate over a long time scale.This approach addresses the low asset utilization efficiency caused by the fragmentation of traditional strategies. Ultimately, by employing dual-modal dynamic switching, it balances grid dispatch compliance with economic benefits, achieving synergistic optimization of energy storage throughout its entire lifecycle over long timescales.

[0033] Example 2: Based on Example 1, this embodiment of the invention provides a collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy. In step S1, based on electricity price changes and grid dispatch instructions along the time axis, the time axis is divided into multiple continuous time periods, including: Extract time points of state changes from electricity price changes and power grid dispatch instructions; The time points of state changes are sorted in chronological order, and the time axis is divided into multiple continuous and non-overlapping time periods using two adjacent nodes as boundaries.

[0034] In this embodiment, the time point of state change is, for example, the node corresponding to the time of electricity price switching or the time when the scheduling instruction takes effect / fails.

[0035] The beneficial effects of the above design scheme are: by dividing the time axis into multiple continuous time periods based on electricity price changes and grid dispatch instructions, the segmented time periods have the characteristics of being continuous and non-overlapping and having unique features, which can accurately adapt to the dynamic determination requirements of the dual-modal strategy.

[0036] Example 3: Based on Example 1, this embodiment of the invention provides a collaborative decision-making method for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy. In step S2, the state of charge is discretized into multiple SOC levels based on a continuous value range, including: The continuous value range is evenly divided according to the rule of equal spacing, resulting in multiple SOC levels.

[0037] In this embodiment, the interval width is the interval range / the preset number of gears.

[0038] The beneficial effects of the above design scheme are: by discretizing the state of charge into multiple SOC levels based on the continuous value range, each SOC level corresponds to a unique state of the energy storage battery, the infinite-dimensional continuous state space is transformed into a finite-dimensional discrete state space, making it possible to traverse all possible states in the reverse recursion process, thus solving the problem of excessive computational power consumption of decision-making algorithms in continuous states.

[0039] Example 4: Based on Example 1, this embodiment of the invention provides a collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy. In step S3, the specific determination method for evaluating the action cost value of transferring the energy storage battery to the end of the current detection period for each feasible mode is as follows: The comprehensive cost incurred during a continuous period is determined by operating according to the feasible mode, the current discrete SOC level, the current specific energy storage action, and the net load forecast curve adapted to the current energy storage environment. The sum of the cumulative cost of subsequent continuous periods and the comprehensive cost of this continuous period is taken as the cumulative action cost value of this continuous period. Among them, the current specific energy storage action is the specific operation instruction selected from the set of energy storage actionable actions that can directly drive the operation of the equipment; The predicted net load forecast curve is a random variable that follows the probability distribution of a family of net load curves predicted based on historical data. The net load forecast curve has time as the horizontal axis and the difference between the total electricity load of the photovoltaic storage system users and the actual output of new energy as the vertical axis, which fully represents the future power supply and demand balance of the system. The total cost includes basic electricity costs, transition costs, and control incentive and penalty costs.

[0040] In this embodiment, the transition cost is the smoothing adjustment cost generated during the dual-mode strategy switching (such as the loss cost caused by sudden power changes).

[0041] In this embodiment, the control of reward and punishment costs includes, for example, the grid subsidy or reward obtained by executing the instruction tracking mode; and the penalty cost incurred if the target is not met.

[0042] In this embodiment, the specific instructions include, but are not limited to, charging at a certain power, discharging at a certain power, and maintaining the current power level.

[0043] The beneficial effects of the above design scheme are as follows: By operating according to the feasible mode, the current discrete SOC level, the current specific energy storage action, and the net load prediction curve adapted to the current energy storage environment, the comprehensive cost generated in a continuous period is used as the sum of the cumulative cost of subsequent continuous periods and the comprehensive cost of this continuous period. This incorporates the influencing factors of dual-modal strategy, real-time energy storage status, actual actions, and environmental supply and demand status into a unified calculation framework, ensuring that each decision scenario corresponds to a unique and accurate cost value, providing reliable data support for globally optimal decision-making. By defining energy storage actions as operational commands that can directly drive equipment operation, cost calculations are matched one-to-one with actual equipment control actions, ensuring that optimized strategies can be directly implemented. By selecting a net load forecast curve that is compatible with the current energy storage environment, the anti-interference capability of cost assessment is improved, effectively covering multiple possible states of future power supply and demand, reducing the impact of forecast errors on cost assessment, and adapting to forecast uncertainty. By integrating basic electricity costs, transition costs, and control reward and penalty costs to obtain a comprehensive cost, the true cost and benefits of energy storage operation are fully restored, achieving a dynamic balance between economic efficiency and grid dispatch compliance.

[0044] Example 5: Based on Example 1, this embodiment of the invention provides a collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy. In step S4, each consecutive time period is traversed forward along the time axis, and the target strategy mode corresponding to the lowest cumulative cost in each consecutive time period is selected, including: Obtain the combined range of two non-overlapping strategy classes that fully cover all available energy storage actions over a continuous period, all available energy storage actions, and all net load forecast curves; Select the target strategy pattern that corresponds to the lowest cumulative cost for each consecutive time period.

[0045] In this embodiment, since the cumulative cost at the SOC level at the beginning of each consecutive time period has been recorded in the reverse recursion stage, the target strategy mode corresponding to the lowest cumulative cost can be directly selected in the current forward recursion stage using the recorded information.

[0046] In this embodiment, the constraints include electricity price, grid dispatch instructions, and net load forecasting scenarios.

[0047] The beneficial effects of the above design scheme are as follows: By acquiring the combination range of two non-overlapping strategy classes that completely cover all energy storage actions, all energy storage actions, and all net load forecast curves for continuous time periods, a comprehensive scenario coverage for strategy selection is achieved, constructing a full-dimensional feasible combination space. This ensures that every potentially optimal strategy can be included in the evaluation, avoiding the risk of missing the optimal solution. By selecting the target strategy mode corresponding to the lowest cumulative cost for each continuous time period, the advantages of the dual-modal strategy are transformed into directly comparable quantitative data, ensuring the scientific nature and consistency of strategy selection, avoiding decision fluctuations caused by human intervention, and achieving automatic selection of the instruction tracking mode with the lower total cost; conversely, the economic mode is prioritized, achieving a dynamic optimal balance between economic efficiency and grid compliance. This solves the problem of insufficient adaptability of traditional fixed priority strategies. The minimum expected total cost obtained through reverse recursion calculation allows the future value calculated through reverse recursion to be directly connected to the evaluation indicators of forward selection, forming a complete logical closed loop from reverse recursion prediction of global cost to forward selection of the optimal strategy.

[0048] Example 6: Based on Example 1, this embodiment of the invention provides a collaborative decision-making method for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy. In step S3, starting from the end of the time axis, each continuous time period is traversed backward along the time axis. Based on the time period characteristic parameters of the current continuous time period and the SOC level corresponding to the start of the continuous time period, the feasibility of the command tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters is determined. Then, for each feasible mode, the possible SOC level of the energy storage battery to be transferred to the end of the current detection time period is evaluated, including: When the time period feature parameters meet the instruction execution requirements, it is determined that the current time period feature parameters meet the instruction tracking strategy mode; when the time period feature parameters meet the security restriction requirements, it is determined that the current time period feature parameters meet the economic optimization strategy mode. A first policy subset is pre-established based on the economic optimization policy model, and a second policy subset is established based on the instruction tracking policy model. Traverse each consecutive time period backward along the time axis. At each time step, based on the feasible mode, dynamically determine the action vector in the first or second strategy subset. Based on the preset reward mechanism, perform weighted evaluation on the output of different strategy subsets to determine the possible SOC level of the energy storage battery to be transferred to the end of the current detection period.

[0049] In this embodiment, the time period characteristic parameters meet the command execution requirements, such as the discharge command requiring SOC ≥ safety lower limit and the charging command requiring total load not to exceed transformer capacity. The time period characteristic parameters meet the safety restriction requirements, such as the charging and discharging actions conforming to the SOC safety range and equipment power limits.

[0050] In this embodiment, a time step refers to the smallest processing unit of a continuous time period during the reverse traversal process. It corresponds one-to-one with the continuous time periods divided in step S1. Each time step corresponds to a complete continuous time period. The reverse traversal needs to complete the strategy adaptation and cost calculation for each time step one by one. For example, if the 48-hour time axis is divided into 96 continuous time periods at 30-minute intervals, then each 30-minute time period is a time step (numbered from time step 1 to time step 96, where time step 96 is the end time period of the optimization cycle. The reverse traversal starts from time step 96 and processes time step 95, time step 94, and so on until time step 1. Within each time step, the complete process of strategy subset activation, action evaluation, end point SOC level determination, and cost calculation is completed independently.

[0051] In this embodiment, determining the possible SOC level of the energy storage battery at the end of the current detection period refers to the starting point of the current detection time step. After selecting a specific action based on the activated strategy subset, the SOC value of the energy storage battery at the end of the period is calculated by combining the energy storage charging and discharging power, action execution time, battery charging and discharging efficiency, predicted net load curve, and the current starting SOC level. For example, assuming that the continuous range of energy storage battery SOC (0%-100%) is discretely divided into 11 levels (level 1: 0%, level 2: 10%, up to level 11: 100%), and the current detection time step is time step k (uncontrolled period, activating the first...). A subset economic optimization model is adopted, with the starting SOC level at level 3 (20%). The action of charging at 60kW power for 30 minutes is selected (given that the battery charge-discharge efficiency is 90% and the charging amount corresponds to an SOC increase of 18%). The calculated lower limit of the starting SOC is 20% + 18% × 90% = 36.2%, which falls within the range of level 4 to level 5 (30%-40%). Therefore, the endpoint SOC corresponding to the current action is determined to be 36.2%, and it is found to be between level 4 and level 5. Then, the cumulative cost at the endpoint SOC = 36.2% is calculated by linear interpolation of the cumulative cost of level 4 and level 5.

[0052] In this embodiment, the command tracking strategy mode prioritizes grid dispatch commands and ensures grid safety through three core actions: First, it transforms safety commands such as peak shaving, frequency regulation, and power limiting into rigid constraints on charging and discharging power and duration; second, it combines multiple constraint verifications such as transformer capacity and battery SOC safety range to avoid secondary risks such as overload and voltage drop caused by energy storage operations; and third, it monitors the grid status in real time, quickly corrects deviations, and responds to sudden fluctuations to ensure that the operation of the power station meets the grid safety requirements.

[0053] The beneficial effects of the above design scheme are as follows: By pre-establishing a first strategy subset based on an economic optimization strategy model and a second strategy subset based on an instruction tracking strategy model, actions are pre-classified into two non-overlapping and fully covered subsets. Reverse traversal only needs to evaluate actions within the currently activated subset, eliminating redundant screening and significantly improving computational efficiency. By traversing each consecutive time period backward along the time axis, at each time step, the action vector in the first or second strategy subset is dynamically determined based on the feasible mode. The outputs of different strategy subsets are weighted and evaluated based on a preset reward mechanism to determine the possible SOC level for the energy storage battery to be transferred to the end of the current detection period. The existence of grid dispatch instructions is used as the activation basis. The second strategy subset is activated during controlled periods to ensure grid compliance, while the first strategy subset is activated during uncontrolled periods to maximize benefits. This solves the problem of insufficient adaptability of a single strategy and achieves accurate matching of core objectives in different time periods. Through the preset reward mechanism, the outputs of different subsets are weighted and evaluated, transforming compliance value and economic value into directly comparable quantitative indicators. This transforms strategy selection from subjective judgment to data-driven selection, ensuring the objectivity and optimality of the selection results.

[0054] Example 7: Based on Example 6, this embodiment of the invention provides a collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy. The specific method for determining the comprehensive cost is as follows: Calculate the cumulative cost function at each time step under the state of charge node for each consecutive time period, and determine the comprehensive cost based on the output of the cumulative cost function.

[0055] In this embodiment, the cumulative cost function refers to the function model in the reverse recursive calculation module used to quantify the optimal cost of a certain charged state node over the entire cycle at a specific time step. Its core logic is to integrate the comprehensive cost of the current time step with the energy storage state value of the next time step, and output the minimum expected cumulative cost corresponding to the charged state node, which serves as the core quantitative basis for selecting the optimal strategy mode in the current period.

[0056] The beneficial effects of the above design scheme are: calculating the cumulative cost function at each time step under the state of charge node in each continuous period, determining the comprehensive cost based on the output of the cumulative cost function, achieving global optimization throughout the entire cycle, and improving the quantitative accuracy of action cost value and the completeness of scenario coverage.

[0057] Example 8: Based on Example 6, this embodiment of the invention provides a collaborative decision-making method for the entire life cycle of energy storage based on reverse recursion and a dual-modal strategy. The first strategy subset is a set of energy storage actions and their corresponding energy storage effects formulated with the primary criterion of meeting the rigid requirements of grid dispatch and ensuring that the overall operation of the energy storage station conforms to the preset control standards issued by the grid. The second strategy subset is a set of energy storage actions and their corresponding energy storage effects formulated with the primary criterion of maximizing the economic benefits of energy storage throughout the entire life cycle, by optimizing the timing of charging and discharging, power and duration, and making full use of the arbitrage space brought about by the peak-valley electricity price difference and the fluctuation of new energy output.

[0058] The beneficial effects of the above design scheme are: by clearly defining the boundaries of strategy subset classification, conflicts between economic benefits and grid compliance objectives are avoided, the stability of strategy execution is improved, a structured foundation is provided for dynamic switching between dual modes, corresponding subsets can be quickly activated as needed, adapting to different operating scenarios and improving strategy flexibility.

[0059] Example 9: This embodiment of the invention provides a collaborative decision-making system for the entire lifecycle of energy storage based on reverse recursion and a dual-modal strategy, such as... Figure 2 As shown, it includes: The time period segmentation module is used to obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on the electricity price changes and grid dispatch instructions of the time axis, and define time period characteristic parameters for each continuous time period; The SOC division module is used to obtain the continuous range of values ​​of the state of charge (SOC) of the energy storage battery, and to discretize the SOC into multiple SOC levels based on the continuous range of values. Each SOC level corresponds to a unique state of the energy storage battery. The gear cost assessment module is used to traverse each continuous time period backward from the end of the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC gear corresponding to the start of the continuous time period, it determines the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, for each feasible mode, it evaluates the possible SOC gear of the energy storage battery to the end of the current detection time period and outputs the action cost value of the corresponding feasible mode. After traversing all feasible modes, it records the minimum action cost value at the SOC gear corresponding to the start of the continuous time period. This operation is repeated until all SOC gears at the start of the continuous time period have been traversed. The optimal sequence determination module is used to calculate the cost information of all continuous time periods and the optimal strategy at each SOC level. Starting from the initial SOC level of the energy storage battery at the initial moment, it traverses each continuous time period in the forward direction along the time axis, selects the target strategy mode corresponding to the lowest cumulative cost, and concatenates the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Based on the strategy sequence, it generates control commands to drive the operation of the energy storage system.

[0060] In this embodiment, the future preset time scale is, for example, the next 48 hours.

[0061] In this embodiment, the cost assessment includes the assessment of transition costs, action costs, etc.

[0062] In this embodiment, the continuous value range is 0%-100%.

[0063] In this embodiment, the cost information is the action cost value under different strategy modes for each consecutive time period.

[0064] In this embodiment, the SOC level is one of the non-overlapping intervals divided from the continuous SOC value range of the energy storage battery, serving as a quantitative benchmark node for the energy storage capacity status at the current time period.

[0065] In this embodiment, the time period characteristic parameters include, for example, the electricity price corresponding to the time period, whether the time period is a controlled time period, the target power value corresponding to the controlled time period, and the preset core control objectives (instruction tracking or economic optimization) and switching trigger conditions for each consecutive time period.

[0066] In this embodiment, the SOC level corresponds to a unique position state of the energy storage battery. This position state, i.e., the discrete SOC level, serves as the core application logic in the reverse cost calculation step, acting as a benchmark anchor point for state transition and cost evaluation, and determining the initial state benchmark for the recursive evaluation.

[0067] The beneficial effects of the above design scheme are as follows: By obtaining a time axis with a preset future time scale, and based on the electricity price changes and grid dispatch instructions on the time axis, the time axis is divided into multiple continuous time periods. For each continuous time period, time-specific parameters are defined, transforming previously difficult-to-quantify constraints such as electricity price fluctuations and dispatch instructions into fixed characteristic parameters for each time period (e.g., electricity price, whether it is controlled, target power). This provides a clear environmental input basis for subsequent strategy decisions. The characteristic parameters for each time period are defined independently, ensuring that subsequent strategy evaluations can accurately match the core needs of different time periods (e.g., focusing on instruction tracking during controlled periods and focusing on economic charging during low-price periods). By obtaining the continuous range of the state of charge (SOC) of the energy storage battery, and discretizing the SOC based on this continuous range... Multiple SOC levels, each corresponding to a unique state of the energy storage battery, transform the infinite-dimensional continuous state space into a finite-dimensional discrete state space. This makes traversing all possible states during the reverse recursion computationally feasible, solving the problem of excessive computational power consumption in decision-making algorithms under continuous states. Furthermore, each SOC level corresponds to a unique position state, ensuring that all possible charge states of the energy storage battery can be accurately represented, avoiding decision-making biases caused by state omissions, and guaranteeing the comprehensiveness of cost calculations during subsequent reverse recursion. This provides a data foundation for globally optimal decision-making. Starting from the end of the time axis, each continuous time period is traversed backward along the time axis, based on the time period characteristic parameters of the currently detected continuous time period and... The system identifies a State of Charge (SOC) level at the start of a continuous time period. It determines the feasibility of both the command tracking strategy and the economic optimization strategy under the constraints of the current time period's characteristic parameters. For each feasible mode, it evaluates the possible SOC levels at which the energy storage battery could be transferred to the end of the current detection time period, simultaneously outputting the action cost value for each feasible mode. After traversing all feasible modes, it records the minimum action cost value at the SOC level corresponding to the start of the continuous time period. This process is repeated until all SOC levels at the start of the continuous time period have been traversed, enabling advance prediction of constraints and benefits in subsequent time periods. It supports both command tracking and economic optimization modes, allowing for flexible selection based on time period characteristics. This satisfies both the rigid requirements of grid dispatch commands and the optimal SOC level. The economic benefits of large-scale energy storage assets are assessed by evaluating the action costs of a dual-modal strategy, providing an objective and comparable quantitative basis for subsequent optimal strategy selection. This avoids cost assessment biases caused by fuzzy decision-making. Based on cost information across all continuous time periods, starting from the initial SOC level of the energy storage battery at its initial moment, each continuous time period is traversed forward along the time axis. The target strategy mode corresponding to the lowest sum of cost and energy storage state value for the next continuous time period is selected. The selected target strategy modes are then concatenated to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Ultimately, this maximizes the utilization rate of energy storage assets, minimizes electricity costs, and maximizes the dispatch command compliance rate over a long time scale.This approach addresses the low asset utilization efficiency caused by the fragmentation of traditional strategies. Ultimately, by employing dual-modal dynamic switching, it balances grid dispatch compliance with economic benefits, achieving synergistic optimization of energy storage throughout its entire lifecycle over long timescales.

[0068] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this application and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A collaborative decision-making method for the entire lifecycle of energy storage based on inverse recursion and a dual-modal strategy, characterized in that, include: S1: Obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on electricity price changes and power grid dispatch instructions, and define time period characteristic parameters for each continuous time period; S2: Obtain the continuous range of the state of charge (SOC) of the energy storage battery, and discretize the SOC into multiple SOC levels based on the continuous range. Each SOC level corresponds to a unique state of the energy storage battery. S3: Starting from the end of the time axis, traverse each continuous time period backward along the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC level corresponding to the start of the continuous time period, determine the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, evaluate the possible SOC level of the energy storage battery to the end of the current detection time period for each feasible mode, and output the action cost value of the corresponding feasible mode. After traversing all feasible modes, record the minimum action cost value at the SOC level corresponding to the start of the continuous time period. Repeat this operation until all SOC levels at the start of the continuous time period have been traversed. S4: Based on the cost information of all continuous time periods and the optimal strategy at each SOC level calculated in step S3, starting from the initial SOC level of the energy storage battery at the initial moment, traverse each continuous time period in the forward direction along the time axis, select the target strategy mode corresponding to the lowest cumulative cost, and connect the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle, and generate control commands based on the strategy sequence to drive the operation of the energy storage system.

2. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 1, characterized in that, In step S1, based on electricity price changes and grid dispatch instructions along the time axis, the time axis is divided into multiple continuous time periods, including: Extract time points of state changes from electricity price changes and power grid dispatch instructions; The time points of state changes are sorted in chronological order, and the time axis is divided into multiple continuous and non-overlapping time periods using two adjacent nodes as boundaries.

3. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 1, characterized in that, In S2, the state of charge (SOC) is discretized into multiple SOC levels based on a continuous value range, including: The continuous value range is evenly divided according to the rule of equal spacing, resulting in multiple SOC levels.

4. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 1, characterized in that, In S3, the specific method for determining the action cost value of transferring the energy storage battery to the end of the current detection period for each feasible mode is as follows: The comprehensive cost incurred during a continuous period is determined by operating according to the feasible mode, the current discrete SOC level, the current specific energy storage action, and the net load forecast curve adapted to the current energy storage environment. The sum of the cumulative cost of subsequent continuous periods and the comprehensive cost of this continuous period is taken as the cumulative action cost value of this continuous period. Among them, the current specific energy storage action is the specific operation instruction selected from the set of energy storage actionable actions that can directly drive the operation of the equipment; The predicted net load forecast curve is a random variable that follows the probability distribution of a family of net load curves predicted based on historical data. The net load forecast curve has time as the horizontal axis and the difference between the total electricity load of the photovoltaic storage system users and the actual output of new energy as the vertical axis, which fully represents the future power supply and demand balance of the system. The total cost includes basic electricity costs, transition costs, and control incentive and penalty costs.

5. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 1, characterized in that, In step S4, each consecutive time period is traversed forward along the time axis, and the target strategy pattern corresponding to the lowest cumulative cost for each consecutive time period is selected, including: Obtain the combined range of two non-overlapping strategy classes that fully cover all available energy storage actions over a continuous period, all available energy storage actions, and all net load forecast curves; Select the target strategy pattern that corresponds to the lowest cumulative cost for each consecutive time period.

6. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 1, characterized in that, In step S3, starting from the end of the time axis, each continuous time period is traversed backwards along the time axis. Based on the time period characteristic parameters of the current continuous time period and the SOC level corresponding to the start of the continuous time period, the feasibility of the command tracking strategy mode and the economic optimization strategy mode under the constraints of the current time period characteristic parameters is determined. Then, for each feasible mode, the possible SOC level of the energy storage battery to be transferred to the end of the current detection time period is evaluated, including: When the time period feature parameters meet the instruction execution requirements, it is determined that the current time period feature parameters meet the instruction tracking strategy mode; when the time period feature parameters meet the security restriction requirements, it is determined that the current time period feature parameters meet the economic optimization strategy mode. A first policy subset is pre-established based on the economic optimization policy model, and a second policy subset is established based on the instruction tracking policy model. Traverse each consecutive time period backward along the time axis. At each time step, based on the feasible mode, dynamically determine the action vector in the first or second strategy subset. Based on the preset reward mechanism, perform weighted evaluation on the output of different strategy subsets to determine the possible SOC level of the energy storage battery to be transferred to the end of the current detection period.

7. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 4, characterized in that, The specific method for determining the comprehensive cost is as follows: Calculate the cumulative cost function at each time step under the state of charge node for each consecutive time period, and determine the comprehensive cost based on the output of the cumulative cost function.

8. The energy storage full-cycle collaborative decision-making method based on reverse recursion and a dual-modal strategy according to claim 6, characterized in that, The first strategy subset is a set of strategies for energy storage actions and their corresponding energy storage effects, which take meeting the rigid requirements of power grid dispatch as the primary criterion and ensure the overall operation of energy storage sites in accordance with the preset control standards issued by the power grid. The second strategy subset is a set of strategies for energy storage actions and their corresponding energy storage effects, which take maximizing the economic benefits of energy storage throughout its entire life cycle as the primary criterion and optimize charging and discharging timing, power and duration, and make full use of the arbitrage space brought about by peak-valley electricity price differences and fluctuations in new energy output.

9. A collaborative decision-making system for the entire lifecycle of energy storage based on reverse recursion and a bimodal strategy, specifically used to implement the collaborative decision-making method for the entire lifecycle of energy storage based on reverse recursion and a bimodal strategy as described in claim 1, characterized in that, include: The time period segmentation module is used to obtain a time axis with a future preset time scale, divide the time axis into multiple continuous time periods based on the electricity price changes and grid dispatch instructions of the time axis, and define time period characteristic parameters for each continuous time period; The SOC division module is used to obtain the continuous range of values ​​of the state of charge (SOC) of the energy storage battery, and to discretize the SOC into multiple SOC levels based on the continuous range of values. Each SOC level corresponds to a unique state of the energy storage battery. The gear cost assessment module is used to traverse each continuous time period backward from the end of the time axis. Based on the time period characteristic parameters of the current continuous time period and a SOC gear corresponding to the start of the continuous time period, it determines the feasibility of the instruction tracking strategy mode and the economic optimization strategy mode under the constraint of the current time period characteristic parameters. Then, for each feasible mode, it evaluates the possible SOC gear of the energy storage battery to the end of the current detection time period and outputs the action cost value of the corresponding feasible mode. After traversing all feasible modes, it records the minimum action cost value at the SOC gear corresponding to the start of the continuous time period. This operation is repeated until all SOC gears at the start of the continuous time period have been traversed. The optimal sequence determination module is used to calculate the cost information of all continuous time periods and the optimal strategy at each SOC level. Starting from the initial SOC level of the energy storage battery at the initial moment, it traverses each continuous time period in the forward direction along the time axis, selects the target strategy mode corresponding to the lowest cumulative cost, and concatenates the selected target strategy modes in sequence to obtain the optimal charging and discharging strategy sequence covering the entire preset time scale optimization cycle. Based on the strategy sequence, it generates control commands to drive the operation of the energy storage system.

Images