Energy storage device operation strategy decision method containing charge and discharge complementary constraints
Patent Information
- Application Number
- CN202611074475.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-20
- Publication Date
- 2026-10-27
AI Technical Summary
[0004]本发明的目的在于克服现有储能能量管理方法中混合整数规划（mixed-integerprogramming，MIP）模型求解效率较低、连续松弛模型可能产生物理上不可执行的同时充放电（simultaneous charging and discharging，SCD），以及缺少松弛精确性评估方法等问题，提供了一种含充放电互补约束的储能装置运行策略决策方法，能够减少0-1整型变量数量和模型求解规模，降低计算资源占用并缩短求解时间；同时保持与完整混合整数规划模型相同的最优运行成本，保证调度结果满足储能运行的物理约束，从而实现一种兼具求解高效性、运行经济性和物理可执行性的广义储能能量管理方法
[0088]本发明能够根据SCD边际成本和最大允许SCD功率定量计算松弛间隙，并判断连续松弛模型是否与广义储能MIP模型具有相同的最优目标函数值；仅在负边际成本时段设置充放电模式变量为0-1整型变量，减少整型变量数量和模型求解规模；对于零边际成本时段存在的SCD，通过调度恢复在保持SoC轨迹和最优目标函数值不变的情况下予以消除。由此，在提高模型求解效率的同时，获得与完整MIP模型具有相同最优运行成本且满足充放电物理约束的广义储能调度方案。
Smart Images
Figure CN122992916A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of energy storage system optimization scheduling and operation control, and in particular to a method for decision-making on the operation strategy of an energy storage device with complementary charging and discharging constraints. Background Technology
[0002] To reduce fossil fuel consumption and carbon emissions, the proportion of renewable energy sources such as wind power and solar power in the power system is continuously increasing. Broadly defined energy storage, with its rapid power regulation and cross-time energy transfer capabilities, can improve the absorption of renewable energy, achieve peak shaving and valley filling, and reduce system operating costs, making it an important flexibility resource for new power systems. However, as the scale of energy storage integration and dispatch cycles continue to increase, balancing the physical feasibility, operational economy, and solution efficiency of energy storage dispatch has become a significant challenge for energy storage management.
[0003] Existing energy storage management typically treats charging power, discharging power, and state of charge (SoC) for each scheduling period as continuous decision variables. Due to limitations imposed by the energy conversion process and the operating modes of power conversion equipment, energy storage can only be in a charging, discharging, or shutdown state at any given time, and cannot simultaneously charge and discharge. Therefore, existing methods usually use 0-1 integer variables to represent the charging and discharging operation modes of energy storage, and formulate the energy storage scheduling problem as a mixed-integer programming (MIP) problem. Each 0-1 integer variable corresponds to one of the two possible modes: charging or discharging; for variables containing... A single energy storage device can hold at most [number] energy storage units during a single scheduling period. The number of charging and discharging mode combinations increases rapidly with the increase in the amount of energy storage and the scheduling period, leading to a significant increase in solution time and computational resource consumption. To reduce computational load, existing techniques typically relax 0-1 integer variables into continuous variables, transforming the MIP problem into a linear optimization problem. However, the relaxed model may produce physically unenforceable simultaneous charging and discharging decisions, resulting in a lower objective function value than the original MIP model. Existing research mainly judges the accuracy of relaxation through specific sufficient conditions; when these conditions are not met, it cannot be determined whether the relaxed model has the same optimal value as the original MIP model. In addition, existing methods lack a unified evaluation index for relaxation gaps, making it impossible to determine whether simultaneous charging and discharging changes the optimal value or whether it can be directly eliminated. Therefore, how to improve solution efficiency while ensuring the optimality and physical feasibility of energy storage scheduling remains a key issue in energy storage management. Summary of the Invention
[0004] The purpose of this invention is to overcome the problems of low solution efficiency of mixed-integer programming (MIP) models, the potential for physically unexecutable simultaneous charging and discharging (SCD) in continuous relaxation models, and the lack of relaxation accuracy evaluation methods in existing energy storage management methods. This invention provides an energy storage device operation strategy decision-making method with complementary charging and discharging constraints. This method reduces the number of 0-1 integer variables and the scale of the model solution, lowers computational resource consumption, and shortens solution time. Simultaneously, it maintains the same optimal operating cost as the complete mixed-integer programming model, ensuring that the scheduling results meet the physical constraints of energy storage operation. Thus, it achieves a generalized energy storage management method that combines high solution efficiency, economical operation, and physical executability.
[0005] To achieve the above objectives, the technical solution provided by this invention is: a method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints, comprising the following steps:
[0006] 1) Obtain the charging and discharging cost coefficient, charging and discharging efficiency, charging and discharging power upper limit, SoC upper and lower limits, initial SoC and scheduling cycle of generalized energy storage, and establish a generalized energy storage MIP model that includes charging and discharging mode variables;
[0007] 2) Relax the charging and discharging mode variables into continuous variables, construct and solve the continuous relaxation model, and obtain the solution results of the continuous relaxation model, including the SoC, charging power and discharging power of each scheduling period;
[0008] 3) Based on the solution results of the continuous relaxation model obtained in step 2), calculate the SoC change, reference charging power and reference discharging power of adjacent scheduling periods, and decompose the charging power and discharging power in the solution results of the continuous relaxation model into reference charging and discharging power and SCD power; calculate the SCD marginal cost and relaxation gap of each scheduling period, and determine whether the continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value based on the relaxation gap;
[0009] 4) Determine if partial integerization is needed: If the solution result of the continuous relaxation model contains a SCD that can reduce the objective function value, then partial integerization is needed. That is, the charging and discharging mode variables for all negative marginal cost periods are set to 0-1 integer variables, while the charging and discharging mode variables for other scheduling periods remain continuous variables. Construct and solve a partial MIP model, and use the solution result of the partial MIP model as the final optimization solution result, including the SoC, charging power, and discharging power for each scheduling period. If the solution result of the continuous relaxation model does not contain a SCD that can reduce the objective function value, then partial integerization is not needed, and the solution result of the continuous relaxation model obtained in step 2) is directly used as the final optimization solution result.
[0010] 5) Check if there is a Scheduled Discharge (SCD) in each zero-marginal-cost period in the final optimization solution. If an SCD exists, recalculate the baseline charging power and baseline discharging power based on the SoC change of the corresponding period, and replace the charging power and discharging power of the corresponding period in the final optimization solution with the baseline charging power and baseline discharging power. The solution results of other scheduling periods remain unchanged. If there is no SCD in any zero-marginal-cost period, no scheduling recovery is performed, and the final optimization solution results are directly retained. Thus, the final SoC, charging power, and discharging power of each scheduling period are obtained.
[0011] 6) Output the final SoC, charging power and discharging power of each scheduling period obtained in step 5) to form a generalized energy storage scheduling scheme that satisfies the charging and discharging complementary constraints.
[0012] Furthermore, in step 1), the set of scheduling periods is defined as follows: To minimize the energy storage operating cost over the entire scheduling cycle, the following generalized energy storage MIP model is established:
[0013] (1);
[0014] The SoC transfer constraint for energy storage is:
[0015] (2);
[0016] The charging power constraint for energy storage is:
[0017] (3);
[0018] The discharge power constraint for energy storage is:
[0019] (4);
[0020] The charging and discharging modes of energy storage are constrained as follows:
[0021] (5);
[0022] The SoC boundary constraints for energy storage are:
[0023] (6);
[0024] The terminal SoC constraints for energy storage are:
[0025] (7);
[0026] In the formula, This represents the optimal objective function value for the generalized energy storage MIP model. and The first The charging power and discharging power during the time period are both continuous variables; , , The first , , The SoC over a given time period is a continuous variable; For the first The charging / discharging mode variable for each time period is a 0-1 integer variable, where... This indicates that charging is permitted. This indicates that discharge is permitted; and The first Charging cost coefficient and discharging cost coefficient for different time periods; These are charging efficiency and discharging efficiency, respectively. and These are the upper limits for charging power and discharging power, respectively. The scheduling time interval; Given an initial SoC; and The first The lower and upper limits of SoC for the time period; for ease of subsequent model representation, the ratio of the upper limit of charging power to the upper limit of discharging power is defined as: .
[0027] Furthermore, in step 2), the 0-1 integer variables in equation (5) are... relaxation is satisfied Continuous variables; in the aforementioned ratio The continuous relaxation forms of equations (3) and (4) are equivalently expressed as follows:
[0028] (8);
[0029] Equation (8) has the same charge and discharge power feasible domain as the continuous relaxation forms of Equations (3) and (4); Equations (1), (2), (6), (7), and (8) together constitute a continuous relaxation model without 0-1 integer variables. This continuous relaxation model is a linear programming model, which is solved using a linear programming solver to obtain the optimal solution of the continuous relaxation model. The optimal solution includes the SoC, charging power, and discharging power of each scheduling period.
[0030] Furthermore, in step 3), the optimal objective function value of the continuous relaxation model is denoted as... Since the feasible region of charge and discharge power of this continuous relaxation model includes the feasible region of the generalized energy storage MIP model, it satisfies... ,when When this is the case, the optimal objective function value of the continuous relaxation model is said to be equivalent to that of the generalized energy storage MIP model;
[0031] Define the overall charge-discharge efficiency as and define the first The change in SoC over the period is For a given SoC change Construct a reference charging power that can realize the variation of the SoC and satisfy the complementary charging and discharging constraints. and reference discharge power :
[0032] (9);
[0033] In the formula, This means that the change is taken when the SoC change is positive, otherwise it is taken as zero. This indicates that when the SoC change is negative, its absolute value is taken; otherwise, it is zero. At that time, the reference discharge power is zero, and the SoC is increased only through charging; when At this time, the reference charging power is zero, and the SoC reduction is achieved only through discharging; therefore, the reference charging power and the reference discharging power will not be positive at the same time, and can produce the same SoC change as the original charging and discharging power.
[0034] Based on reference charging power and reference discharge power The charge-discharge power obtained from the continuous relaxation model is decomposed into the reference charge-discharge power and the SCD power, as follows:
[0035] For any set of relaxed charging power and discharging power that satisfies equations (2) and (8), it can be expressed as:
[0036] (10);
[0037] In the formula, Indicates the first The SCD power of the SoC, i.e., the charging power, remains unchanged during the time period; by It can be seen that the increased charging power and discharge power The effects on the SoC cancel each other out, i.e., satisfy... The range of values for the SCD power is:
[0038] (11);
[0039] Given SoC variation Under these circumstances, the maximum allowable SCD power is:
[0040] (12);
[0041] In the formula, This indicates that, under the condition of meeting the charging and discharging power limits, the first... The maximum SCD power that the remaining power capacity during the time period can support; for a given relaxation charge / discharge power, Able to be The power is uniquely determined; through the above power decomposition, the charge and discharge power in the continuous relaxation model is decomposed into a reference charge and discharge power that realizes SoC variation and satisfies complementary constraints, and an SCD power that does not change SoC.
[0042] Furthermore, in step 3), for a given SoC change... The reference charging power and reference discharging power obtained from equation (9) satisfy the charging-discharging complementarity constraint, and their corresponding operating costs are defined as the reference operating costs:
[0043] (13);
[0044] Substituting the charge / discharge power decomposition relationship expressed in equation (10) into the objective function of equation (1), we obtain the first... Operating costs for a given period:
[0045] (14);
[0046] According to equation (14), the first... The marginal cost of SCD during the time period is:
[0047] (15);
[0048] In the formula, This represents the change in the objective function value caused by increasing the SCD power by one unit; when At this time, increasing the SCD power will increase operating costs; when At this time, the SCD power does not change the operating cost; when Increasing the SCD power will reduce the operating cost of the continuous relaxation model;
[0049] For a given SoC variation, the generalized energy storage MIP model enforces the satisfaction of the charge-discharge complementary constraint. The continuous relaxation model allows The value takes values within the range given in equation (11); therefore, the continuous relaxation model takes values in the first... The optimal operating cost for the time period is:
[0050] (16);
[0051] In the formula, The time-period running cost of the continuous relaxation model given the SoC variation;
[0052] Definition of the first relaxation intervals during the period The difference between the baseline operating cost of the generalized energy storage MIP model and the operating cost of the continuous relaxation model:
[0053] (17);
[0054] In the formula, The relaxation gap is determined by the marginal cost of SCD. and maximum allowable SCD power Together, they determine the quantitative representation of the objective function value that the continuous relaxation model can reduce using SCD under a given SoC variation;
[0055] From equation (17), we can obtain that the first... The condition for the relaxation interval of a time period to be equal to zero is:
[0056] (18).
[0057] Furthermore, in step 3), all SoC trajectories that satisfy the SoC boundary, charge / discharge power limit, and terminal SoC constraint are combined into a feasible SoC trajectory set. The SoC change in each time period satisfies:
[0058] (19);
[0059] Combining Equations (6), (7), and (19), a set of feasible SoC trajectories is obtained. Any SoC trajectory in the model can be realized using the reference charge / discharge power corresponding to equation (9); therefore, the generalized energy storage MIP model and the continuous relaxation model have the same set of feasible SoC trajectories; in the set of feasible SoC trajectories Above, the optimal objective function values of the generalized energy storage MIP model and the continuous relaxation model are expressed as follows:
[0060] (20);
[0061] The continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value, i.e. If and only if the continuous relaxation model has an optimal SoC trajectory ,in Representing the initial SoC, the relaxation interval for each scheduling period is zero, i.e., satisfying:
[0062] (twenty one);
[0063] in, , Indicates the first , The optimal SOC for the time period, as expressed in equation (21), is equivalent to:
[0064] (twenty two);
[0065] Equations (21) and (22) constitute the necessary and sufficient condition for the equivalence of the optimal objective function value of the continuous relaxation model and the generalized energy storage MIP model, and are used to determine whether the continuous relaxation result needs to be solved by partial integerization.
[0066] According to equations (17) and (18), the accuracy of the continuous relaxation model in each scheduling period is determined by the marginal cost of SCD. and maximum allowable SCD power Jointly decided; for the sake of simplicity, the following will be... Recorded as And the actual SCD power in the optimal solution of the continuous relaxation model is denoted as Specifically, this includes the following five situations:
[0067] a. When At that time, increasing the SCD power will improve the objective function value; therefore, regardless of Whether it is greater than zero, the optimal solution of the continuous relaxation model satisfies... SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0068] b. When and At this point, given the SoC change, all charging or discharging power has been used, and there is no remaining power capacity available for SCD; therefore, the optimal solution of the continuous relaxation model satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0069] c. When and At this time, SCD will not change the objective function value, and the optimal solution of the continuous relaxation model can be found. to Arbitrary SCD power can be chosen between these values; therefore, SCD may occur in the continuous relaxation result, but this SCD can be directly eliminated while keeping the SoC variation and the objective function value unchanged. The results of continuous relaxation over time remain equivalent to the generalized energy storage MIP model;
[0070] d. When and At this time, although increasing the SCD power can reduce the objective function value, there is no remaining power capacity available for the SCD given the SoC change; therefore, the optimal solution of the continuous relaxation model still satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0071] e. when and At this point, increasing the SCD power can reduce the objective function value, and there is remaining power capacity available for the SCD; therefore, the continuous relaxation model will choose the maximum SCD power, i.e., satisfy... At this point, the objective function value of the continuous relaxation model is lower than that of the generalized energy storage MIP model. The results of continuous relaxation over time are not equivalent to the generalized energy storage MIP model;
[0072] In conclusion, only when and Only when both conditions are met will SCD produce a non-zero relaxation gap; otherwise, the continuous relaxation result does not have SCD, or the existing SCD will not change the objective function value and can be directly eliminated.
[0073] Furthermore, in step 4), the SCD marginal cost for each scheduling period is calculated according to equation (15), and the scheduling period is divided into a set of negative marginal cost periods. and zero marginal cost time set :
[0074] (twenty three);
[0075] After solving the continuous relaxation model, the actual SCD power in each scheduling period is calculated based on the reference charge and discharge power obtained from equation (9):
[0076] (twenty four);
[0077] when At that time, the first The time period does not have an SCD; when At that time, the first The time period has a SCD; based on the set of time periods with negative marginal cost. and actual SCD power Construct the set of time periods that require partial integerization for solution:
[0078] (25);
[0079] Determine Set Is it empty? If If the value is empty, it means that the continuous relaxation optimal solution obtained in step 2) does not contain an SCD that can reduce the objective function value, and we jump to step 5).
[0080] When set When not empty, in the set For the corresponding time periods, set 0-1 integer variables and solve part of the MIP model; for other scheduling time periods, keep the charging and discharging mode variables as continuous variables.
[0081] (26);
[0082] Replace equation (5) with equation (26), keep the objective function equation (1) and other constraints unchanged, construct and solve a partial MIP model, and obtain the solution results of the partial MIP model, including the SoC, charging power and discharging power of each scheduling period; after completing the solution, recalculate the set of zero marginal cost periods according to equation (24). SCD power at different time periods;
[0083] When set When the set is empty, partial MIP model solving is not performed, and the continuous relaxation model solution obtained in step 2) is determined as the final optimization solution; when the set is empty... When the solution is not empty, the partial MIP model solution obtained in step 4) is determined as the final optimized solution; the final optimized solution is uniformly denoted as... ,in , and These represent the first and second halves of the final optimization solution. The SoC, charging power, and discharging power for each time period; based on the final optimization solution, the SCD power for each zero marginal cost time period is calculated according to equation (24), and denoted as... .
[0084] Furthermore, in step 5), the zero marginal cost period in the final optimization solution is scheduled for recovery: for any period that satisfies... and For each scheduling period, the SoC change during that period is calculated based on the final optimization solution. And calculate the reference charging power according to equation (9). and reference discharge power ; to the reference charging power and reference discharge power The optimal charging power for this period was determined. and optimal discharge power , that is to say and Complete the SCD elimination for this zero marginal cost period; for other scheduling periods that do not require scheduling recovery, retain the charging power and discharging power from the final optimization solution, i.e. and Scheduling recovery does not change the SoC trajectory in the final optimization solution. Therefore, for all scheduling periods, the final SoC satisfies... This yields the final SoC, charging power, and discharging power for each scheduling period. Since the elimination of SCD during the zero marginal cost period does not change the SoC trajectory and objective function value, the result after scheduling recovery maintains the optimal objective function value of the final optimization solution and satisfies the charge-discharge complementary constraint throughout all scheduling periods.
[0085] Furthermore, in step 6), the output generalized energy storage scheduling scheme includes the final SoC, charging power, and discharging power for each scheduling period, i.e. The generalized energy storage scheduling scheme satisfies the charging and discharging complementarity constraint in all scheduling periods. Generalized energy storage refers to energy storage devices or systems that can absorb energy through the charging process and release energy through the discharging process, and whose operating status can be described by charging power, discharging power, SoC, charging and discharging efficiency, power limit and energy boundary. These include electrochemical energy storage, pumped hydro storage, compressed air energy storage, flywheel energy storage, thermal energy storage systems, electric-hydrogen-electric hydrogen energy storage systems, and electric vehicle hybrid systems with bidirectional energy regulation capabilities.
[0086] Furthermore, this method solves the continuous relaxation model and the partial MIP model at most once; when the time period set that triggers the partial integerization solution is empty, the continuous relaxation model is solved only once; when the time period set is a non-empty set, the charging and discharging mode variables are set as 0-1 integer variables only in all negative marginal cost time periods, thereby reducing the number of integer variables and the scale of model solution.
[0087] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0088] This invention can quantitatively calculate the relaxation gap based on the marginal cost of SCD and the maximum allowable SCD power, and determine whether the continuous relaxation model has the same optimal objective function value as the generalized energy storage MIP model. It sets the charging and discharging mode variables as 0-1 integer variables only during periods of negative marginal cost, reducing the number of integer variables and the model solution scale. For SCDs existing during periods of zero marginal cost, they are eliminated through scheduling recovery while maintaining the SoC trajectory and optimal objective function value. Thus, while improving model solution efficiency, a generalized energy storage scheduling scheme with the same optimal operating cost as the complete MIP model and satisfying charging and discharging physical constraints is obtained. Attached Figure Description
[0089] Figure 1 This is an overall flowchart of the method of the present invention. Detailed Implementation
[0090] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.
[0091] like Figure 1 As shown in the figure, this embodiment discloses a method for making operational strategy decisions for energy storage devices with complementary charge-discharge constraints, the specific details of which are as follows:
[0092] 1) Obtain the charging and discharging cost coefficient, charging and discharging efficiency, charging and discharging power upper limit, SoC upper and lower limits, initial SoC and scheduling cycle of generalized energy storage, and establish a generalized energy storage MIP model that includes charging and discharging mode variables.
[0093] In a specific example, the set of scheduling periods is defined as follows: To minimize the energy storage operating cost over the entire scheduling cycle, the following generalized energy storage MIP model is established:
[0094] (1);
[0095] The SoC transfer constraint for energy storage is:
[0096] (2);
[0097] The charging power constraint for energy storage is:
[0098] (3);
[0099] The discharge power constraint for energy storage is:
[0100] (4);
[0101] The charging and discharging modes of energy storage are constrained as follows:
[0102] (5);
[0103] The SoC boundary constraints for energy storage are:
[0104] (6);
[0105] The terminal SoC constraints for energy storage are:
[0106] (7);
[0107] In the formula, This represents the optimal objective function value for the generalized energy storage MIP model. and The first The charging power and discharging power during the time period are both continuous variables; , , The first , , The SoC over a given time period is a continuous variable; For the first The charging / discharging mode variable for each time period is a 0-1 integer variable, where... This indicates that charging is permitted. This indicates that discharge is permitted; and The first Charging cost coefficient and discharging cost coefficient for different time periods; These are charging efficiency and discharging efficiency, respectively. and These are the upper limits for charging power and discharging power, respectively. The scheduling time interval; Given an initial SoC; and The first The lower and upper limits of SoC for the time period; for ease of subsequent model representation, the ratio of the upper limit of charging power to the upper limit of discharging power is defined as: .
[0108] 2) Relax the charging and discharging mode variables into continuous variables, construct and solve the continuous relaxation model, and obtain the solution results of the continuous relaxation model, including the SoC, charging power and discharging power of each scheduling period.
[0109] In a specific example, the 0-1 integer variable in equation (5) relaxation is satisfied Continuous variables; in the aforementioned ratio The continuous relaxation forms of equations (3) and (4) are equivalently expressed as follows:
[0110] (8);
[0111] Equation (8) has the same charge and discharge power feasible domain as the continuous relaxation forms of Equations (3) and (4); Equations (1), (2), (6), (7), and (8) together constitute a continuous relaxation model without 0-1 integer variables. This continuous relaxation model is a linear programming model, which is solved using a linear programming solver to obtain the optimal solution of the continuous relaxation model. The optimal solution includes the SoC, charging power, and discharging power of each scheduling period.
[0112] 3) Based on the solution results of the continuous relaxation model obtained in step 2), calculate the SoC change, reference charging power and reference discharging power of adjacent scheduling periods, and decompose the charging power and discharging power in the solution results of the continuous relaxation model into reference charging and discharging power and SCD power; calculate the SCD marginal cost and relaxation gap of each scheduling period, and determine whether the continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value based on the relaxation gap.
[0113] In a specific example, let the optimal objective function value of the continuous relaxation model be... Since the feasible region of charge and discharge power of this continuous relaxation model includes the feasible region of the generalized energy storage MIP model, it satisfies... ,when When this is the case, the optimal objective function value of the continuous relaxation model is said to be equivalent to that of the generalized energy storage MIP model;
[0114] Define the overall charge-discharge efficiency as and define the first The change in SoC over the period is For a given SoC change Construct a reference charging power that can realize the variation of the SoC and satisfy the complementary charging and discharging constraints. and reference discharge power :
[0115] (9);
[0116] In the formula, This means that the change is taken when the SoC change is positive, otherwise it is taken as zero. This indicates that when the SoC change is negative, its absolute value is taken; otherwise, it is zero. At that time, the reference discharge power is zero, and the SoC is increased only through charging; when At this time, the reference charging power is zero, and the SoC reduction is achieved only through discharging; therefore, the reference charging power and the reference discharging power will not be positive at the same time, and can produce the same SoC change as the original charging and discharging power.
[0117] Based on reference charging power and reference discharge power The charge-discharge power obtained from the continuous relaxation model is decomposed into the reference charge-discharge power and the SCD power, as follows:
[0118] For any set of relaxed charging power and discharging power that satisfies equations (2) and (8), it can be expressed as:
[0119] (10);
[0120] In the formula, Indicates the first The SCD power of the SoC, i.e., the charging power, remains unchanged during the time period; by It can be seen that the increased charging power and discharge power The effects on the SoC cancel each other out, i.e., satisfy... The range of values for the SCD power is:
[0121] (11);
[0122] Given SoC variation Under these circumstances, the maximum allowable SCD power is:
[0123] (12);
[0124] In the formula, This indicates that, under the condition of meeting the charging and discharging power limits, the first... The maximum SCD power that the remaining power capacity during the time period can support; for a given relaxation charge / discharge power, Able to be The power is uniquely determined; through the above power decomposition, the charge and discharge power in the continuous relaxation model is decomposed into a reference charge and discharge power that realizes SoC variation and satisfies complementary constraints, and an SCD power that does not change SoC.
[0125] For a given SoC variation The reference charging power and reference discharging power obtained from equation (9) satisfy the charging-discharging complementarity constraint, and their corresponding operating costs are defined as the reference operating costs:
[0126] (13);
[0127] Substituting the charge / discharge power decomposition relationship expressed in equation (10) into the objective function of equation (1), we obtain the first... Operating costs for a given period:
[0128] (14);
[0129] According to equation (14), the first... The marginal cost of SCD during the time period is:
[0130] (15);
[0131] In the formula, This represents the change in the objective function value caused by increasing the SCD power by one unit; when At this time, increasing the SCD power will increase operating costs; when At this time, the SCD power does not change the operating cost; when Increasing the SCD power will reduce the operating cost of the continuous relaxation model;
[0132] For a given SoC variation, the generalized energy storage MIP model enforces the satisfaction of the charge-discharge complementary constraint. The continuous relaxation model allows The value takes values within the range given in equation (11); therefore, the continuous relaxation model takes values in the first... The optimal operating cost for the time period is:
[0133] (16);
[0134] In the formula, The time-period running cost of the continuous relaxation model given the SoC variation;
[0135] Definition of the first relaxation intervals during the period The difference between the baseline operating cost of the generalized energy storage MIP model and the operating cost of the continuous relaxation model:
[0136] (17);
[0137] In the formula, The relaxation gap is determined by the marginal cost of SCD. and maximum allowable SCD power Together, they determine the quantitative representation of the objective function value that the continuous relaxation model can reduce using SCD under a given SoC variation;
[0138] From equation (17), we can obtain that the first... The condition for the relaxation interval of a time period to be equal to zero is:
[0139] (18).
[0140] The set of feasible SoC trajectories comprises all SoC trajectories that satisfy SoC boundaries, charge / discharge power limits, and terminal SoC constraints. The SoC change in each time period satisfies:
[0141] (19);
[0142] Combining Equations (6), (7), and (19), a set of feasible SoC trajectories is obtained. Any SoC trajectory in the model can be realized using the reference charge / discharge power corresponding to equation (9); therefore, the generalized energy storage MIP model and the continuous relaxation model have the same set of feasible SoC trajectories; in the set of feasible SoC trajectories Above, the optimal objective function values of the generalized energy storage MIP model and the continuous relaxation model are expressed as follows:
[0143] (20);
[0144] The continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value, i.e. If and only if the continuous relaxation model has an optimal SoC trajectory ,in Representing the initial SoC, the relaxation interval for each scheduling period is zero, i.e., satisfying:
[0145] (twenty one);
[0146] in, , Indicates the first , The optimal SOC for the time period, as expressed in equation (21), is equivalent to:
[0147] (twenty two);
[0148] Equations (21) and (22) constitute the necessary and sufficient condition for the equivalence of the optimal objective function value of the continuous relaxation model and the generalized energy storage MIP model, and are used to determine whether the continuous relaxation result needs to be solved by partial integerization.
[0149] According to equations (17) and (18), the accuracy of the continuous relaxation model in each scheduling period is determined by the marginal cost of SCD. and maximum allowable SCD power Jointly decided; for the sake of simplicity, the following will be... Recorded as And the actual SCD power in the optimal solution of the continuous relaxation model is denoted as Specifically, this includes the following five situations:
[0150] a. When At that time, increasing the SCD power will improve the objective function value; therefore, regardless of Whether it is greater than zero, the optimal solution of the continuous relaxation model satisfies... SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0151] b. When and At this point, given the SoC change, all charging or discharging power has been used, and there is no remaining power capacity available for SCD; therefore, the optimal solution of the continuous relaxation model satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0152] c. When and At this time, SCD will not change the objective function value, and the optimal solution of the continuous relaxation model can be found. to Arbitrary SCD power can be chosen between these values; therefore, SCD may occur in the continuous relaxation result, but this SCD can be directly eliminated while keeping the SoC variation and the objective function value unchanged. The results of continuous relaxation over time remain equivalent to the generalized energy storage MIP model;
[0153] d. When and At this time, although increasing the SCD power can reduce the objective function value, there is no remaining power capacity available for the SCD given the SoC change; therefore, the optimal solution of the continuous relaxation model still satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model;
[0154] e. when and At this point, increasing the SCD power can reduce the objective function value, and there is remaining power capacity available for the SCD; therefore, the continuous relaxation model will choose the maximum SCD power, i.e., satisfy... At this point, the objective function value of the continuous relaxation model is lower than that of the generalized energy storage MIP model. The results of continuous relaxation over time are not equivalent to the generalized energy storage MIP model;
[0155] In conclusion, only when and Only when both conditions are met will SCD produce a non-zero relaxation gap; otherwise, the continuous relaxation result does not have SCD, or the existing SCD will not change the objective function value and can be directly eliminated.
[0156] 4) Determine if partial integerization is needed: If the solution result of the continuous relaxation model contains a SCD that can reduce the objective function value, then partial integerization is needed. That is, the charging and discharging mode variables for all negative marginal cost periods are set to 0-1 integer variables, while the charging and discharging mode variables for other scheduling periods remain continuous variables. Construct and solve a partial MIP model, and use the solution result of the partial MIP model as the final optimization solution result, including the SoC, charging power, and discharging power for each scheduling period. If the solution result of the continuous relaxation model does not contain a SCD that can reduce the objective function value, then partial integerization is not needed, and the solution result of the continuous relaxation model obtained in step 2) is directly used as the final optimization solution result.
[0157] In a specific example, the SCD marginal cost of each scheduling period is calculated according to equation (15), and the scheduling period is divided into a set of negative marginal cost periods. and zero marginal cost time set :
[0158] (twenty three);
[0159] After solving the continuous relaxation model, the actual SCD power in each scheduling period is calculated based on the reference charge and discharge power obtained from equation (9):
[0160] (twenty four);
[0161] when At that time, the first The time period does not have an SCD; when At that time, the first The time period has a SCD; based on the set of time periods with negative marginal cost. and actual SCD power Construct the set of time periods that require partial integerization for solution:
[0162] (25);
[0163] Determine Set Is it empty? If If the value is empty, it means that the continuous relaxation optimal solution obtained in step 2) does not contain an SCD that can reduce the objective function value, and we jump to step 5).
[0164] When set When not empty, in the set For the corresponding time periods, set 0-1 integer variables and solve part of the MIP model; for other scheduling time periods, keep the charging and discharging mode variables as continuous variables.
[0165] (26);
[0166] Replace equation (5) with equation (26), keep the objective function equation (1) and other constraints unchanged, construct and solve a partial MIP model, and obtain the solution results of the partial MIP model, including the SoC, charging power and discharging power of each scheduling period; after completing the solution, recalculate the set of zero marginal cost periods according to equation (24). SCD power at different time periods;
[0167] When set When the set is empty, partial MIP model solving is not performed, and the continuous relaxation model solution obtained in step 2) is determined as the final optimization solution; when the set is empty... When the solution is not empty, the partial MIP model solution obtained in step 4) is determined as the final optimized solution; the final optimized solution is uniformly denoted as... ,in , and These represent the first and second halves of the final optimization solution. The SoC, charging power, and discharging power for each time period; based on the final optimization solution, the SCD power for each zero marginal cost time period is calculated according to equation (24), and denoted as... .
[0168] 5) Check if there is a Scheduled Discharge (SCD) in each zero-marginal-cost period in the final optimization solution. If an SCD exists, recalculate the baseline charging power and baseline discharging power based on the SoC change of the corresponding period, and replace the charging power and discharging power of the corresponding period in the final optimization solution with the baseline charging power and baseline discharging power. The solution results of other scheduling periods remain unchanged. If there is no SCD in any zero-marginal-cost period, no scheduling recovery is performed, and the final optimization solution results are directly retained. Thus, the final SoC, charging power, and discharging power of each scheduling period are obtained.
[0169] In a specific example, scheduling recovery is performed on the zero marginal cost period in the final optimization solution: for any period that satisfies and For each scheduling period, the SoC change during that period is calculated based on the final optimization solution. And calculate the reference charging power according to equation (9). and reference discharge power ; to the reference charging power and reference discharge power The optimal charging power for this period was determined. and optimal discharge power , that is to say and Complete the SCD elimination for this zero marginal cost period; for other scheduling periods that do not require scheduling recovery, retain the charging power and discharging power from the final optimization solution, i.e. and Scheduling recovery does not change the SoC trajectory in the final optimization solution. Therefore, for all scheduling periods, the final SoC satisfies... This yields the final SoC, charging power, and discharging power for each scheduling period. Since the elimination of SCD during the zero marginal cost period does not change the SoC trajectory and objective function value, the result after scheduling recovery maintains the optimal objective function value of the final optimization solution and satisfies the charge-discharge complementary constraint throughout all scheduling periods.
[0170] 6) Output a generalized energy storage scheduling scheme, including the final SoC, charging power, and discharging power for each scheduling period, i.e. The generalized energy storage scheduling scheme satisfies the charging and discharging complementarity constraint in all scheduling periods. Generalized energy storage refers to energy storage devices or systems that can absorb energy through the charging process and release energy through the discharging process, and whose operating status can be described by charging power, discharging power, SoC, charging and discharging efficiency, power limit and energy boundary. These include electrochemical energy storage, pumped hydro storage, compressed air energy storage, flywheel energy storage, thermal energy storage systems, electric-hydrogen-electric hydrogen energy storage systems, and electric vehicle hybrid systems with bidirectional energy regulation capabilities.
[0171] Below, we will conduct experimental verification of the energy storage device operation strategy decision-making method described above in this embodiment, based on Tables 1 to 4.
[0172] Table 1
[0173]
[0174] Table 1 shows the specific parameters of the equipment used in the experiment. As can be seen from the experimental results (Tables 2-4), this embodiment involved two optimization solutions. The first solution was a continuous relaxation model, and the results are shown in Table 2. Simultaneous charging and discharging occurred in the third and fifth time periods. The simultaneous charging and discharging in the third time period reduced the objective function value of the relaxation model, resulting in an optimal operating cost of −35.5919 for the continuous relaxation model; the simultaneous charging and discharging in the fifth time period did not change the objective function value.
[0175] Table 2
[0176]
[0177] Based on the solution results of the continuous relaxation model, only the 0-1 charge / discharge mode variable was set for the third time period corresponding to the negative marginal cost. A second optimization solution was then constructed and performed, i.e., a partial MIP model was solved. The results are shown in Table 3. The optimal operating cost of the partial MIP model is −30.4900, which is the same as the complete MIP model with 0-1 charge / discharge mode variables set for all time periods. At this point, simultaneous charge / discharge in the third time period has been eliminated, while simultaneous charge / discharge in the fifth time period still exists without affecting the optimal operating cost.
[0178] Table 3
[0179]
[0180] After completing the above two optimization solutions, the baseline charge / discharge power is directly calculated based on the SoC change in the 5th time period, and the original charge / discharge power is replaced with the baseline charge / discharge power. This scheduling recovery process does not require solving the optimization model again. The recovery results are shown in Table 4. Simultaneous charging and discharging in the 5th time period is eliminated, the SoC trajectory and optimal operating cost remain unchanged, the optimal operating cost of the final scheduling scheme is still −30.4900, and the charge / discharge complementarity constraint is satisfied in all scheduling periods.
[0181] Table 4
[0182]
[0183] Experimental results show that this invention achieves the same optimal operating cost as the full MIP model by solving only one continuous relaxation model and one partial MIP model. Compared to the full MIP model, which requires setting 0-1 variables in all five scheduling periods, this invention only sets 0-1 variables in one necessary period, reducing the number of integer variables and the scale of the model solution, while ensuring the optimality and physical executability of the energy storage scheduling results.
[0184] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.
Claims
1. A method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints, characterized in that, Includes the following steps: 1) Obtain the charging and discharging cost coefficient, charging and discharging efficiency, charging and discharging power upper limit, SoC upper and lower limits, initial SoC and scheduling cycle of generalized energy storage, and establish a generalized energy storage MIP model that includes charging and discharging mode variables; 2) Relax the charging and discharging mode variables into continuous variables, construct and solve the continuous relaxation model, and obtain the solution results of the continuous relaxation model, including the SoC, charging power and discharging power of each scheduling period; 3) Based on the solution results of the continuous relaxation model obtained in step 2), calculate the SoC change, reference charging power and reference discharging power of adjacent scheduling periods, and decompose the charging power and discharging power in the solution results of the continuous relaxation model into reference charging and discharging power and SCD power; calculate the SCD marginal cost and relaxation gap of each scheduling period, and determine whether the continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value based on the relaxation gap; 4) Determine whether partial integerization is needed: When there is an SCD in the solution of the continuous relaxation model that can reduce the objective function value, partial integerization is needed. That is, the charging and discharging mode variables of all negative marginal cost periods are set as 0-1 integer variables, while the charging and discharging mode variables of other scheduling periods are kept as continuous variables. Construct and solve a partial MIP model, and obtain the solution results of the partial MIP model as the final optimization solution results, including the SoC, charging power and discharging power of each scheduling period; When there is no SCD in the solution of the continuous relaxation model that can reduce the objective function value, there is no need for partial integerization, and the solution of the continuous relaxation model obtained in step 2) is directly used as the final optimization solution. 5) Check whether SCD exists in each zero marginal cost period in the final optimization solution; If SCD exists, the reference charging power and reference discharging power are recalculated based on the SoC change in the corresponding time period, and the reference charging power and reference discharging power in the corresponding time period in the final optimization solution are replaced with the reference charging power and reference discharging power. The solution results for other scheduling time periods remain unchanged. If there is no SCD in any zero marginal cost period, then no scheduling recovery is performed, and the final optimization solution result is directly retained; thus, the final SoC, charging power and discharging power of each scheduling period are obtained; 6) Output the final SoC, charging power and discharging power of each scheduling period obtained in step 5) to form a generalized energy storage scheduling scheme that satisfies the charging and discharging complementary constraints.
2. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 1, characterized in that, In step 1), the set of scheduling periods is defined as follows: To minimize the energy storage operating cost over the entire scheduling cycle, the following generalized energy storage MIP model is established: （1）； The SoC transfer constraint for energy storage is: （2）； The charging power constraint for energy storage is: （3）； The discharge power constraint for energy storage is: （4）； The charging and discharging modes of energy storage are constrained as follows: （5）； The SoC boundary constraints for energy storage are: （6）； The terminal SoC constraints for energy storage are: （7）； In the formula, This represents the optimal objective function value for the generalized energy storage MIP model. and The first The charging power and discharging power during the time period are both continuous variables; , , The first , , The SoC over a given time period is a continuous variable; For the first The charging / discharging mode variable for each time period is a 0-1 integer variable, where... This indicates that charging is permitted. This indicates that discharge is permitted; and The first Charging cost coefficient and discharging cost coefficient for different time periods; These are charging efficiency and discharging efficiency, respectively. and These are the upper limits for charging power and discharging power, respectively. The scheduling time interval; Given an initial SoC; and The first The lower and upper limits of SoC for the time period; for ease of subsequent model representation, the ratio of the upper limit of charging power to the upper limit of discharging power is defined as: .
3. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 2, characterized in that, In step 2), the 0-1 integer variables in equation (5) are... relaxation is satisfied Continuous variables; in the aforementioned ratio The continuous relaxation forms of equations (3) and (4) are equivalently expressed as follows: （8）； Equation (8) has the same charge and discharge power feasible domain as the continuous relaxation forms of Equations (3) and (4); Equations (1), (2), (6), (7), and (8) together constitute a continuous relaxation model without 0-1 integer variables. This continuous relaxation model is a linear programming model, which is solved using a linear programming solver to obtain the optimal solution of the continuous relaxation model. The optimal solution includes the SoC, charging power, and discharging power of each scheduling period.
4. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 3, characterized in that, In step 3), let the optimal objective function value of the continuous relaxation model be denoted as . Since the feasible region of charge and discharge power of this continuous relaxation model includes the feasible region of the generalized energy storage MIP model, it satisfies... ,when When this is the case, the optimal objective function value of the continuous relaxation model is said to be equivalent to that of the generalized energy storage MIP model; Define the overall charge-discharge efficiency as and define the first The change in SoC over the period is ; For a given SoC variation Construct a reference charging power that can realize the variation of the SoC and satisfy the complementary charging and discharging constraints. and reference discharge power : （9）； In the formula, This means that the change is taken when the SoC change is positive, otherwise it is taken as zero. This indicates that when the SoC change is negative, its absolute value is taken; otherwise, it is zero. At that time, the reference discharge power is zero, and the SoC is increased only through charging; when At this time, the reference charging power is zero, and the SoC reduction is achieved only through discharging; therefore, the reference charging power and the reference discharging power will not be positive at the same time, and can produce the same SoC change as the original charging and discharging power. Based on reference charging power and reference discharge power The charge-discharge power obtained from the continuous relaxation model is decomposed into the reference charge-discharge power and the SCD power, as follows: For any set of relaxed charging power and discharging power that satisfies equations (2) and (8), it can be expressed as: （10）； In the formula, Indicates the first The SCD power of the SoC, i.e., the charging power, remains unchanged during the time period; by It can be seen that the increased charging power and discharge power The effects on the SoC cancel each other out, i.e., satisfy... The range of values for the SCD power is: （11）； Given SoC variation Under these circumstances, the maximum allowable SCD power is: （12）； In the formula, This indicates that, under the condition of meeting the charging and discharging power limits, the first... The maximum SCD power that the remaining power capacity during the time period can support; for a given relaxation charge / discharge power, Able to be The power is uniquely determined; through the above power decomposition, the charge and discharge power in the continuous relaxation model is decomposed into a reference charge and discharge power that realizes SoC variation and satisfies complementary constraints, and an SCD power that does not change SoC.
5. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 4, characterized in that, In step 3), for a given SoC change... The reference charging power and reference discharging power obtained from equation (9) satisfy the charging-discharging complementarity constraint, and their corresponding operating costs are defined as the reference operating costs: （13）； Substituting the charge / discharge power decomposition relationship expressed in equation (10) into the objective function of equation (1), we obtain the first... Operating costs for a given period: （14）； According to equation (14), the first... The marginal cost of SCD during the time period is: （15）； In the formula, This represents the change in the objective function value caused by increasing the SCD power by one unit; when At this time, increasing the SCD power will increase operating costs; when At this time, the SCD power does not change the operating cost; when Increasing the SCD power will reduce the operating cost of the continuous relaxation model; For a given SoC variation, the generalized energy storage MIP model enforces the satisfaction of the charge-discharge complementary constraint. The continuous relaxation model allows The value takes values within the range given in equation (11); therefore, the continuous relaxation model takes values in the first... The optimal operating cost for the time period is: （16）； In the formula, The time-period running cost of the continuous relaxation model given the SoC variation; Definition of the first relaxation intervals during the period The difference between the baseline operating cost of the generalized energy storage MIP model and the operating cost of the continuous relaxation model: （17）； In the formula, The relaxation gap is determined by the marginal cost of SCD. and maximum allowable SCD power Together, they determine the quantitative representation of the objective function value that the continuous relaxation model can reduce using SCD under a given SoC variation; From equation (17), we can obtain that the first... The condition for the relaxation interval of a time period to be equal to zero is: （18）。 6. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 5, characterized in that, In step 3), all SoC trajectories that satisfy the SoC boundary, charge / discharge power limit, and terminal SoC constraint are combined into a feasible SoC trajectory set. The SoC change in each time period satisfies: （19）； Combining Equations (6), (7), and (19), a set of feasible SoC trajectories is obtained. Any SoC trajectory in the model can be realized using the reference charge / discharge power corresponding to equation (9); therefore, the generalized energy storage MIP model and the continuous relaxation model have the same set of feasible SoC trajectories; in the set of feasible SoC trajectories Above, the optimal objective function values of the generalized energy storage MIP model and the continuous relaxation model are expressed as follows: （20）； The continuous relaxation model and the generalized energy storage MIP model have the same optimal objective function value, i.e. If and only if the continuous relaxation model has an optimal SoC trajectory ,in Representing the initial SoC, the relaxation interval for each scheduling period is zero, i.e., satisfying: （21）； in, , Indicates the first , The optimal SOC for the time period, as expressed in equation (21), is equivalent to: （22）； Equations (21) and (22) constitute the necessary and sufficient condition for the equivalence of the optimal objective function value of the continuous relaxation model and the generalized energy storage MIP model, and are used to determine whether the continuous relaxation result needs to be solved by partial integerization. According to equations (17) and (18), the accuracy of the continuous relaxation model in each scheduling period is determined by the marginal cost of SCD. and maximum allowable SCD power Jointly decided; for the sake of simplicity, the following will be... Recorded as And the actual SCD power in the optimal solution of the continuous relaxation model is denoted as Specifically, this includes the following five situations: a. When At that time, increasing the SCD power will improve the objective function value; therefore, regardless of Whether it is greater than zero, the optimal solution of the continuous relaxation model satisfies... SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model; b. When and At this point, given the SoC change, all charging or discharging power has been used, and there is no remaining power capacity available for SCD; therefore, the optimal solution of the continuous relaxation model satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model; c. When and At this time, SCD will not change the objective function value, and the optimal solution of the continuous relaxation model can be found. to Arbitrary SCD power can be chosen between these values; therefore, SCD may occur in the continuous relaxation result, but this SCD can be directly eliminated while keeping the SoC variation and the objective function value unchanged. The results of continuous relaxation over time remain equivalent to the generalized energy storage MIP model; d. When and At this time, although increasing the SCD power can reduce the objective function value, there is no remaining power capacity available for the SCD given the SoC change; therefore, the optimal solution of the continuous relaxation model still satisfies SCD will not occur, the first The results of continuous relaxation over time are equivalent to the generalized energy storage MIP model; e. when and At this point, increasing the SCD power can reduce the objective function value, and there is remaining power capacity available for the SCD; therefore, the continuous relaxation model will choose the maximum SCD power, i.e., satisfy... At this point, the objective function value of the continuous relaxation model is lower than that of the generalized energy storage MIP model. The results of continuous relaxation over time are not equivalent to the generalized energy storage MIP model; In conclusion, only when and Only when both conditions are met will SCD produce a non-zero relaxation gap; otherwise, the continuous relaxation result does not have SCD, or the existing SCD will not change the objective function value and can be directly eliminated.
7. The method for decision-making on operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 6, characterized in that, In step 4), the SCD marginal cost of each scheduling period is calculated according to equation (15), and the scheduling period is divided into a set of negative marginal cost periods. and zero marginal cost time set : （23）； After solving the continuous relaxation model, the actual SCD power in each scheduling period is calculated based on the reference charge and discharge power obtained from equation (9): （24）； when At that time, the first The time period does not have an SCD; when At that time, the first SCD exists during this time period; Based on the set of periods with negative marginal cost and actual SCD power Construct the set of time periods that require partial integerization for solution: （25）； Determine Set Is it empty? If If the value is empty, it means that the continuous relaxation optimal solution obtained in step 2) does not contain an SCD that can reduce the objective function value, and we jump to step 5). When set When not empty, in the set For the corresponding time periods, set 0-1 integer variables and solve part of the MIP model; for other scheduling time periods, keep the charging and discharging mode variables as continuous variables. （26）； Replace equation (5) with equation (26), keep the objective function equation (1) and other constraints unchanged, construct and solve a partial MIP model, and obtain the solution results of the partial MIP model, including the SoC, charging power and discharging power of each scheduling period; After completing the solution, recalculate the set of zero marginal cost periods according to equation (24). SCD power at different time periods; When set When the set is empty, partial MIP model solving is not performed, and the continuous relaxation model solution obtained in step 2) is determined as the final optimization solution; when the set is empty... When the solution is not empty, the partial MIP model solution obtained in step 4) is determined as the final optimized solution; the final optimized solution is uniformly denoted as... ,in , and These represent the first and second halves of the final optimization solution. The SoC, charging power, and discharging power for each time period; based on the final optimization solution, the SCD power for each zero marginal cost time period is calculated according to equation (24), and denoted as... .
8. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 7, characterized in that, In step 5), the zero marginal cost period in the final optimization solution is scheduled for recovery: for any period that satisfies... and For each scheduling period, the SoC change during that period is calculated based on the final optimization solution. And calculate the reference charging power according to equation (9). and reference discharge power ; to the reference charging power and reference discharge power The optimal charging power for this period was determined. and optimal discharge power , that is to say and Complete the SCD elimination for this zero marginal cost period; for other scheduling periods that do not require scheduling recovery, retain the charging power and discharging power from the final optimization solution, i.e. and Scheduling recovery does not change the SoC trajectory in the final optimization solution. Therefore, for all scheduling periods, the final SoC satisfies... This yields the final SoC, charging power, and discharging power for each scheduling period. Since the elimination of SCD during the zero marginal cost period does not change the SoC trajectory and objective function value, the result after scheduling recovery maintains the optimal objective function value of the final optimization solution and satisfies the charge-discharge complementary constraint throughout all scheduling periods.
9. The method for decision-making on the operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 8, characterized in that, In step 6), the output generalized energy storage scheduling scheme includes the final SoC, charging power, and discharging power for each scheduling period, i.e. The generalized energy storage scheduling scheme satisfies the charging and discharging complementarity constraint in all scheduling periods. Generalized energy storage refers to energy storage devices or systems that can absorb energy through the charging process and release energy through the discharging process, and whose operating status can be described by charging power, discharging power, SoC, charging and discharging efficiency, power limit and energy boundary. These include electrochemical energy storage, pumped hydro storage, compressed air energy storage, flywheel energy storage, thermal energy storage systems, electric-hydrogen-electric hydrogen energy storage systems, and electric vehicle hybrid systems with bidirectional energy regulation capabilities.
10. The method for decision-making on operation strategy of an energy storage device with complementary charge-discharge constraints according to claim 9, characterized in that, This method solves the continuous relaxation model and the partial MIP model at most once; when the time period set that triggers the partial integerization solution is empty, the continuous relaxation model is solved only once; when the time period set is not empty, the charging and discharging mode variables are set as 0-1 integer variables only in all negative marginal cost time periods, thereby reducing the number of integer variables and the scale of model solution.