Quantitative battery life degradation prediction method based on mixed integer programming model

Abstract

The present invention relates to a quantitative battery life attenuation prediction method based on a mixed integer programming model, comprising the following steps: step S1: establishing a nonlinear life model based on preset constraints; step S2: using a multivariate Taylor expansion to linearize the nonlinear life model into a mixed integer linear programming model; step S3: obtaining a new battery power based on the mixed integer linear programming model; step S4: calculating the Euclidean distance, if the Euclidean distance ED is less than a preset threshold or the number of cycles n reaches a rated value, then the cycle is stopped, the battery power is output, and a prediction result is obtained; otherwise, a new battery power point is set, and the method returns to step S2. The present invention considers the factors affecting the battery life more comprehensively, can quickly and effectively predict the results, and has high reliability.

Description

Technical Field

[0001] The present invention relates to the field of battery life prediction, and in particular to a quantitative battery life attenuation prediction method based on a mixed integer programming model. Background Art

[0002] With the rapid development of electric vehicles and distributed energy, the power system has gradually become highly uncertain. Large-scale electrochemical energy storage systems on the grid side are widely used in the balance uncertainty of the power system due to their flexible configuration and easy maintenance.

[0003] With the large-scale installation and operation of electrochemical energy storage systems, the operating efficiency and life decay characteristics of electrochemical energy storage have an increasing impact on the reliability and economy of power system operation. Studies have now found that the charge and discharge rate, charge and discharge depth (DOD, Depth of discharge) and operating temperature of electrochemical energy storage have a nonlinear effect on its operating efficiency and life decay. However, the electrochemical energy storage operation model and control model commonly used in power systems are relatively simple, generally only considering the upper and lower limits of the storage device's charge and discharge power, the initial energy storage constraints of the energy storage device, and the upper and lower limits of the energy storage device's storage energy. This approach of abandoning model accuracy to gain computational advantages has defects such as unreasonable constraints and low model accuracy, which has seriously affected the full utilization of energy storage equipment and cannot quantify battery life decay. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a quantitative battery life attenuation prediction method based on a mixed integer programming model, which considers the factors affecting battery life more comprehensively, can predict the results quickly and effectively, and has high reliability.

[0005] To achieve the above object, the present invention adopts the following technical solution:

[0006] A quantitative battery life attenuation prediction method based on a mixed integer programming model comprises the following steps:

[0007] Step S1: Establishing a nonlinear life model based on preset constraints;

[0008] Step S2: Using the multivariate Taylor expansion method to linearize the nonlinear life model, the nonlinear life model is transformed into a mixed integer linear programming model.

[0009] Step S3: Obtaining a new battery power based on a mixed integer linear programming model;

[0010] Step S4: Calculate the Euclidean distance. If the Euclidean distance ED is less than a preset threshold or the number of cycles n reaches a rated value, stop the cycle, output the battery power, and obtain the prediction result; otherwise, set a new battery power point and return to step S2.

[0011] Furthermore, the preset constraint conditions are specifically:

[0012] The battery power cannot exceed its maximum charge and discharge power during operation:

[0013]

[0014]

[0015] α t +β t ≤1 (3)

[0016] Among them, P t bat Indicates the battery charging and discharging power at time t. If it is greater than zero, it is the battery charging power, and if it is less than zero, it is the battery discharging power; Indicates the upper and lower limits of charge and discharge power. greater than zero, Less than zero; α t , β t is a 0 or 1 variable that controls the charge and discharge state of the battery at time t; constraint (3) prohibits the battery from charging and discharging at the same time;

[0017] The SOC operating range during battery operation is expressed as

[0018]

[0019] S min ≤S t ≤S max (5)

[0020] Where S t is the SOC of the battery at the tth period, P t bat is the charge and discharge power of the battery, Q bat is the rated capacity of the battery, Δt is the duration of the period, S max , S min are the maximum and minimum values ​​of the SOC during battery operation; constraint (4) represents the SOC of the battery at time t, which is determined by the SOC of the battery at time t-1 and the power at time t;

[0021] Constraint (6) indicates that in order to prevent the battery from being damaged by a rapid change in battery power, the power change between the two previous dispatches cannot exceed ΔP, that is,

[0022]

[0023] Maximum loop constraint:

[0024]

[0025]

[0026] Among them, η t , γ t is a 0, 1 variable used to assist in calculating the number of battery conversions; η max is the maximum number of transitions from the charging state to the discharging state, γ max is the maximum number of transitions from the discharge state to the charge state, and the sum of the two is the maximum number of cycles; T is the total number of scheduling in the optimization scheduling process.

[0027] Furthermore, the nonlinear life model is established as follows: According to Miner's linear fatigue cumulative damage principle, the capacity loss caused by a single battery small cycle is shown in formula (10), and the total battery capacity decay is obtained by linearly accumulating the battery capacity decay caused by each battery small cycle, that is,

[0028]

[0029] in, It represents the capacity decay caused by the t-th battery dispatch in the dispatch cycle, T is the total number of dispatches in the dispatch cycle, is the total capacity decay during the dispatch period:

[0030]

[0031] in, is the influence coefficient of battery charge and discharge rate on battery capacity decay in the t-th scheduling, Q t is the charge and discharge amount of the battery during the tth scheduling process, and q is the coefficient to be determined;

[0032] The expression (11) shows that

[0033]

[0034] Among them C t is the ratio of the battery charge and discharge current to the rated current during the t-th scheduling, where a, b, c, and d are fitting coefficients;

[0035] The current ratio is the ratio of the battery charging and discharging power to the battery rated power.

[0036]

[0037] Among them I t , I R , P t , P R Respectively represent the charging or discharging current of the battery in the tth scheduling cycle, the rated current of the battery, the charging and discharging power of the battery in the tth scheduling, and the rated power of the battery;

[0038] Substituting formula (12) into formula (11) yields the coefficient About power P t The expression

[0039] Q i is the charge and discharge capacity of the battery in the i-th small cycle, and the calculation formula is as follows

[0040]

[0041] Where V is the grid-connected voltage of the energy storage device, and the grid-connected voltage remains constant; ΔT is the duration of the i-th scheduling;

[0042] Combining formula (13) with (12), (11), and (10), we can obtain the expression (14) for the capacity decay caused by the battery during a scheduling cycle:

[0043]

[0044] Formula (14) fully considers the nonlinear effect of battery discharge depth and discharge rate on battery life; a, b, c, d are the coefficients obtained after simplification;

[0045] Establish a quantitative model that comprehensively considers the effects of discharge depth DOD, charge and discharge rate, and cycle number on the life attenuation of electrochemical energy storage devices, that is, a nonlinear life model

[0046]

[0047] Furthermore, the step S2 is specifically as follows:

[0048] Using Taylor's formula for linearization, at power P t Taylor expansion is performed at this point, and the first two terms of the expansion are intercepted to obtain the linear expression of the polynomial function at this point:

[0049]

[0050] in represents the linearized result of battery capacity decay caused by the t-th scheduling, a t , b t Formula (15) in P t The coefficients obtained after Taylor expansion of the power point.

[0051] From formula (9), we can see that the total battery capacity decay is equal to the linear accumulation of the capacity decay caused by each small cycle. In the process of optimal scheduling, the battery power does not change during the scheduling period. Therefore, each scheduling period can be regarded as a small cycle.

[0052]

[0053] T is the total scheduling time, t is the tth scheduling period, is the capacity decay caused by the battery power in period t;

[0054] Taking microgrid dispatch as an example, considering the full life cycle operation of the energy storage battery, the following constraints are added to the original model. Constraint (18) indicates that the SOC of the battery before and after the dispatch cycle is equal, that is, the starting SOC needs to be equal to the ending SOC

[0055]

[0056] Constraints (19) and (20) represent the upper and lower limits of the primary network switching power, where is the upper limit of the power exchanged between the integrated energy system and the main grid, P t buy , P t sell They represent the power purchased from the main grid during the t-th dispatch period and the power sold to the main grid by the microgrid during the t-th dispatch period; The variable is 0, 1, indicating the state of electricity purchase and sale of the integrated energy system at each moment.

[0057]

[0058]

[0059] Microgrids cannot both purchase and sell electricity, so

[0060]

[0061] The power balance of the microgrid can be expressed as

[0062] P t buy +P t w +P t v =P t load +P t bat +P t sell (twenty two)

[0063] Among them, P tw , P t v , P t load , respectively represent the wind turbine power generation, photovoltaic power generation, and microgrid load in the tth scheduling period;

[0064] The objective function is shown in (23),

[0065]

[0066] Specifically, after Taylor expansion and linearization, (23) is expressed as

[0067]

[0068] in They represent the time-of-use electricity price of the main grid respectively; w represents the battery capacity attenuation cost coefficient, which is a scalar in this paper; a t , b t Represents the coefficient of Taylor expansion linearization of battery power at time t.

[0069] Furthermore, the Euclidean distance is specifically:

[0070]

[0071] in is the new battery power obtained in step S3, P t is the initial power.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] 1. The present invention does not require additional acquisition equipment, which is not only simple to implement but also effectively improves the reliability of the system

[0074] 2. The present invention comprehensively considers the number of cycles, discharge depth, and discharge rate. It can be simplified into a polynomial about power P and number of cycles without losing accuracy, thereby quantifying the battery capacity decay;

[0075] 3. The present invention transforms the nonlinear battery life model into a mixed integer programming problem based on the method of successive linearization and prediction-correction of multivariate Taylor expansion, thereby reducing the difficulty of model solution and improving the prediction speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a flow chart of the method of the present invention;

[0077] Figure 2 The convergence of battery power in one embodiment of the present invention

[0078] Figure 31 is a battery SOC timing diagram in one embodiment of the present invention. DETAILED DESCRIPTION

[0079] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0080] Please refer to Figure 1 The present invention provides a quantitative battery life attenuation prediction method based on a mixed integer programming model, comprising the following steps:

[0081] The first step is to establish a nonlinear mixed integer programming model for the battery.

[0082] In the second step, let the rated power of the battery be the initial power of the battery at each moment, P t Indicates that the number of loops is represented by n and is equal to zero.

[0083] The third step is to calculate the objective function (5-14) at the battery power point P t Perform Taylor expansion and take the first two terms of the expansion to obtain (5-16), and obtain the mixed integer linear programming model.

[0084] Step 4: Solve the mixed integer linear programming model to obtain the new battery power

[0085] Step 5: Calculate the Euclidean distance

[0086] Step 6: Update the number of loops n=n+1.

[0087] Step 7: If the Euclidean distance ED is less than the preset threshold or the number of cycles n reaches the rated value, the cycle is stopped and the battery power of step 3 is output. The model is solved. Otherwise, the new battery power point And go back to step 3.

[0088] In this embodiment, the factors considered in the established quantitative battery life attenuation model include the number of cycles of the energy storage device, the upper and lower limits of the charge and discharge rate, the battery power ramp rate, and the upper and lower limit constraints of the SOC of the battery energy storage device.

[0089] The battery power cannot exceed its maximum charge and discharge power during operation:

[0090]

[0091]

[0092] α t +β t ≤1 (3)

[0093] Among them, P tbat Indicates the battery charging and discharging power at time t. If it is greater than zero, it is the battery charging power, and if it is less than zero, it is the battery discharging power; Indicates the upper and lower limits of charge and discharge power. greater than zero, Less than zero; α t , β t is a 0 or 1 variable that controls the charge and discharge state of the battery at time t; constraint (3) prohibits the battery from charging and discharging at the same time;

[0094] The SOC operating range during battery operation is expressed as

[0095]

[0096] S min ≤S t ≤S max (5)

[0097] Where S t is the SOC of the battery at the tth period, P t bat is the charge and discharge power of the battery, Q bat is the rated capacity of the battery, Δt is the duration of the period, S max , S min are the maximum and minimum values ​​of the SOC during battery operation; constraint (4) represents the SOC of the battery at time t, which is determined by the SOC of the battery at time t-1 and the power at time t;

[0098] Constraint (6) indicates that in order to prevent the battery from being damaged by a rapid change in battery power, the power change between the two previous dispatches cannot exceed ΔP, that is,

[0099]

[0100] Maximum loop constraint:

[0101]

[0102]

[0103] Among them, η t , γ t is a 0, 1 variable used to assist in calculating the number of battery conversions; η max is the maximum number of transitions from the charging state to the discharging state, γ max is the maximum number of transitions from the discharge state to the charge state, and the sum of the two is the maximum number of cycles; T is the total number of scheduling in the optimization scheduling process.

[0104] In this embodiment, a nonlinear life model is established as follows: According to Miner's linear fatigue cumulative damage principle, the capacity loss caused by a single battery small cycle is shown in formula (10), and the total battery capacity decay is obtained by linearly accumulating the battery capacity decay caused by each battery small cycle, that is,

[0105]

[0106] in, It represents the capacity decay caused by the t-th battery dispatch in the dispatch cycle, T is the total number of dispatches in the dispatch cycle, is the total capacity decay during the dispatch period:

[0107]

[0108] in, is the influence coefficient of battery charge and discharge rate on battery capacity decay in the t-th scheduling, Q t is the charge and discharge amount of the battery during the tth scheduling process, and q is the coefficient to be determined;

[0109] The expression (11) shows that

[0110]

[0111] Among them C t is the ratio of the battery charge and discharge current to the rated current during the t-th scheduling, where a, b, c, and d are fitting coefficients;

[0112] The current ratio is the ratio of the battery charging and discharging power to the battery rated power.

[0113]

[0114] Among them I t , I R , P t , P R Respectively represent the charging or discharging current of the battery in the tth scheduling cycle, the rated current of the battery, the charging and discharging power of the battery in the tth scheduling, and the rated power of the battery;

[0115] Substituting formula (12) into formula (11) yields the coefficient About power P t The expression

[0116] Q i is the charge and discharge capacity of the battery in the i-th small cycle, and the calculation formula is as follows

[0117]

[0118] Where V is the grid-connected voltage of the energy storage device, and the grid-connected voltage remains constant; Δt is the duration of the i-th scheduling;

[0119] Combining formula (13) with (12), (11), and (10), we can obtain the expression (14) for the capacity decay caused by the battery during a scheduling cycle:

[0120]

[0121] Formula (14) fully considers the nonlinear effect of battery discharge depth and discharge rate on battery life; a, b, c, d are the coefficients obtained after simplification;

[0122] Establish a quantitative model that comprehensively considers the effects of discharge depth DOD, charge and discharge rate, and cycle number on the life attenuation of electrochemical energy storage devices, that is, a nonlinear life model

[0123]

[0124] Constraints

[0125]

[0126]

[0127] α t +β t ≤1 (3)

[0128]

[0129] S min ≤S t ≤S max (5)

[0130]

[0131]

[0132]

[0133] The objective function formula (15) in the battery life model has strong nonlinearity and cannot be solved directly, so it needs further processing.

[0134] Preferably, the Taylor formula is used for linearization, and the power P t Taylor expansion is performed at this point, and the first two terms of the expansion are intercepted to obtain the linear expression of the polynomial function at this point:

[0135]

[0136] in represents the linearized result of battery capacity decay caused by the t-th scheduling, a t , b t Formula (15) in P t The coefficients obtained after Taylor expansion of the power point.

[0137] From formula (9), we can see that the total battery capacity decay is equal to the linear accumulation of the capacity decay caused by each small cycle. In the process of optimal scheduling, the battery power does not change during the scheduling period. Therefore, each scheduling period can be regarded as a small cycle.

[0138]

[0139] T is the total scheduling time, t is the tth scheduling period, is the capacity decay caused by the battery power in period t;

[0140] Through Taylor expansion, the original nonlinear mixed integer programming model, which is complex to calculate and difficult to solve, is converted into a mixed integer linear programming model.

[0141] Then, by solving the linear model multiple times, the battery power is predicted and corrected to approximate the result of the original nonlinear model. Finally, the battery power converges to a stable value and the model solution is completed.

[0142] Taking microgrid dispatch as an example, considering the full life cycle operation of energy storage batteries, the following constraints are added to the original model to construct patent application scenarios. Considering the full life cycle operation of energy storage batteries, the following constraints are added to the original model. Constraint (18) indicates that the SOC of the battery is equal before and after the dispatch cycle, that is, the starting SOC needs to be equal to the ending SOC

[0143]

[0144] Constraints (19) and (20) represent the upper and lower limits of the primary network switching power, where is the upper limit of the power exchanged between the integrated energy system and the main grid, P t buy , P t sell They represent the power purchased from the main grid during the t-th dispatch period and the power sold to the main grid by the microgrid during the t-th dispatch period; The variable is 0, 1, indicating the state of electricity purchase and sale of the integrated energy system at each moment.

[0145]

[0146]

[0147] Microgrids cannot both purchase and sell electricity, so

[0148]

[0149] The power balance of the microgrid can be expressed as

[0150] P t buy +P t w +P t v =P t load +P t bat +P t sell (twenty two)

[0151] Among them, P t w , P t v , P t load , respectively represent the wind turbine power generation, photovoltaic power generation, and microgrid load in the tth scheduling period;

[0152] The objective function is shown in (23),

[0153]

[0154] Specifically, after Taylor expansion and linearization, (23) is expressed as

[0155]

[0156] in They represent the time-of-use electricity price of the main grid respectively; w represents the battery capacity attenuation cost coefficient, which is a scalar in this paper; a t , b t It represents the coefficient of Taylor expansion linearization of battery power at time t. The optimization goal is to minimize the comprehensive energy operation cost.

[0157] Embodiment 1:

[0158] In this embodiment, the constraint conditions are set by the method of the present invention, and the influence of the number of cycles, discharge depth, and discharge rate on the battery life are comprehensively considered to establish a nonlinear life model and quantify the battery life decay. The nonlinear model that was originally difficult to solve is transformed into a mixed integer linear programming model by using the multivariate Taylor expansion to linearize it successively, thereby improving the solution speed. Compared with the previous method, this method considers the factors affecting the battery life more comprehensively, has a high model accuracy, and converts the nonlinear model into a mixed integer linear programming with a fast solution speed and stable results.

[0159] Figure 2It shows that as the number of iterations increases, the Euclidean distance between the previous and next solutions approaches zero, and the model converges. This machine is configured with CPUi7-6700HQ, 2.60GHZ, 8GB running memory, and software version Matlab2020a+Gurobi. The actual solution time for each iteration is 0.02s, and the total running time is 0.16s. For the same model, the solution time of the intelligent algorithm is 50.34s, which shows that the method of converting the nonlinear model into a mixed integer linear programming model in this patent has a shorter solution time than the traditional method.

[0160] The larger the weight coefficient, the more attention is paid to battery capacity decay. When the weight coefficient is 0, it degenerates into the traditional method. As can be seen from Table 1, compared with the traditional method, the cost of the method of this patent is only increased by 4.42%, but the capacity decay is reduced by 32.42%, and the number of cycles is reduced by 37.5%. From the perspective of the full life cycle cost, the proposed method can delay battery decay and extend the comprehensive energy service life without basically increasing the cost. If the weight coefficient is changed from 100 to 1000, the cost increases by 5.66%, the capacity decay is reduced by 34.21%, and the number of cycles is reduced by 40%. This shows that the weight coefficient can be flexibly changed according to actual conditions, which shows that the method proposed in this patent is more flexible than the traditional method and can give full play to the battery performance according to actual conditions. And the linearization error is less than 1%, which means that the linearized function does not deviate from the original function and the linearization method is feasible. By Figure 3 It can be seen that compared with the traditional method, the battery SOC state of this patent changes more slowly. And the degree of relaxation of the battery SOC state changes is different with different weights, which shows that this patent can give full play to the performance of the battery according to the actual situation.

[0161] Table 1 Comparison of battery capacity attenuation

[0162]

[0163] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made according to the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A quantitative battery life attenuation prediction method based on a mixed integer programming model, characterized in that: The following steps are involved: Step S1: Establishing a nonlinear life model based on preset constraints; Step S2: Using the multivariate Taylor expansion method to linearize the nonlinear life model, the nonlinear life model is transformed into a mixed integer linear programming model. Step S3: Obtaining a new battery power based on a mixed integer linear programming model; Step S4: Calculate the Euclidean distance. If the Euclidean distance ED is less than a preset threshold or the number of cycles n reaches a rated value, stop the cycle, output the battery power, and obtain a prediction result; otherwise, set a new battery power point and return to step S2; The preset constraints are specifically: The battery power cannot exceed its maximum charge and discharge power during operation: α t +β t ≤1 (3) Among them, P t bat Indicates the battery charging and discharging power at time t. If it is greater than zero, it is the battery charging power, and if it is less than zero, it is the battery discharging power; Indicates the upper and lower limits of charge and discharge power. greater than zero, Less than zero; α t , β t is a 0 or 1 variable that controls the charge and discharge state of the battery at time t; constraint (3) prohibits the battery from charging and discharging at the same time; The SOC operating range during battery operation is expressed as S min ≤S t ≤S max (5) Where S t is the SOC of the battery at the tth period, P t bat is the charge and discharge power of the battery, Q bat is the rated capacity of the battery, Δt is the duration of the period, S max , S min are the maximum and minimum values ​​of the SOC during battery operation; constraint (4) represents the SOC of the battery at time t, which is determined by the SOC of the battery at time t-1 and the power at time t; Constraint (6) indicates that in order to prevent the battery from being damaged by a rapid change in battery power, the power change between the two previous dispatches cannot exceed ΔP, that is, Maximum loop constraint: Among them, η t , γ t is a 0, 1 variable used to assist in calculating the number of battery conversions; η max is the maximum number of transitions from the charging state to the discharging state, γ max is the maximum number of transitions from the discharge state to the charge state, and the sum of the two is the maximum number of cycles; T is the total number of scheduling in the optimization scheduling process; The nonlinear life model is established as follows: According to Miner's linear fatigue cumulative damage principle, the capacity loss caused by a single battery small cycle is shown in formula (10), and the total battery capacity decay is obtained by linearly accumulating the battery capacity decay caused by each battery small cycle, that is, in, It represents the capacity decay caused by the t-th battery dispatch in the dispatch cycle, T is the total number of dispatches in the dispatch cycle, is the total capacity decay during the dispatch period: in, is the influence coefficient of battery charge and discharge rate on battery capacity decay in the t-th scheduling, Q t is the charge and discharge amount of the battery during the tth scheduling process, and q is the coefficient to be determined; The expression is as shown in (11), Among them C t is the ratio of the battery charge and discharge current to the rated current of the t-th scheduling, where a, b, c, d are fitting coefficients; the current ratio is the ratio of the battery charge and discharge power to the battery rated power, that is, Among them I t , I R , P t , P R Respectively represent the charging or discharging current of the battery in the tth scheduling cycle, the rated current of the battery, the charging and discharging power of the battery in the tth scheduling, and the rated power of the battery; Substituting formula (12) into formula (11) yields the coefficient k t cyc About power P t The expression Q i is the charge and discharge capacity of the battery in the i-th small cycle, and the calculation formula is as follows Where V is the grid-connected voltage of the energy storage device, and the grid-connected voltage remains constant; Δt is the duration of the i-th scheduling; Combining formula (13) with (12), (11), and (10), we can obtain the expression (14) for the capacity decay caused by the battery during a scheduling cycle: Formula (14) fully considers the nonlinear effect of battery discharge depth and discharge rate on battery life; a', b', c', d' are the coefficients obtained after simplification; Establish a quantitative model that comprehensively considers the effects of discharge depth DOD, charge and discharge rate, and cycle number on the life attenuation of electrochemical energy storage devices, that is, a nonlinear life model The step S2 is specifically as follows: Using Taylor's formula for linearization, at power P t Taylor expansion is performed at this point, and the first two terms of the expansion are intercepted to obtain the linear expression of the polynomial function at this point: in represents the linearized result of battery capacity decay caused by the t-th scheduling, a t , b t Formula (15) in P t The coefficients obtained after Taylor expansion of the power point; From formula (9), we can see that the total battery capacity decay is equal to the linear accumulation of the capacity decay caused by each small cycle. In the process of optimal scheduling, the battery power does not change during the scheduling period. Therefore, each scheduling period is regarded as a small cycle. T is the total scheduling time, t is the tth scheduling period, is the capacity decay caused by the battery power in period t; Considering the full life cycle operation of the energy storage battery, the following constraints are added to the original model. Constraint (18) indicates that the SOC of the battery is equal before and after the scheduling cycle, that is, the starting SOC needs to be equal to the ending SOC Constraints (19) and (20) represent the upper and lower limits of the primary network switching power, where is the upper limit of the power exchanged between the integrated energy system and the main grid, P t buy , P t sell They represent the power purchased from the main grid during the t-th dispatch period and the power sold to the main grid by the microgrid during the t-th dispatch period; The variable is 0, 1, indicating the state of electricity purchase and sale of the integrated energy system at each moment. Microgrids cannot both purchase and sell electricity, so The power balance of the microgrid is expressed as P t buy +P t w +P t v =P t load +P t bat +P t sell (22) Among them, P t w , P t v , P t load , respectively represent the wind turbine power generation, photovoltaic power generation, and microgrid load in the tth scheduling period; The objective function is shown in (23), Specifically, after Taylor expansion and linearization, (23) is expressed as in They represent the time-of-use electricity price of the main grid; w represents the battery capacity attenuation cost coefficient; a t , b t Represents the coefficient of Taylor expansion linearization of battery power at time t; The Euclidean distance is specifically: in is the new battery power obtained in step S3, P t is the initial power.

Images