A shared energy storage optimization control method based on IGDT and master-slave game

Abstract

The present invention relates to the technical field of shared energy storage scheduling, and discloses a shared energy storage optimization control method based on IGDT and master-slave game, comprising the following steps: establishing a corresponding mathematical model according to an integrated energy system; establishing a double-layer master-slave game model, solving it using a Kriging model, and using it as a deterministic model of the IGDT; establishing an IGDT model, and calculating uncertainty weights according to the output and energy consumption of a microgrid in the system; establishing a risk avoidance and risk opportunity model according to user needs. The present invention utilizes shared energy storage to handle the uncertainty of multiple uncertain variables such as photovoltaics, promotes the absorption of renewable energy, and reduces the cost of each microgrid in the integrated energy system. In addition, according to the strategic inclination of decision makers, a risk avoidance and risk opportunity model is established, and the influence of uncertain variables in the microgrid on the system and itself is considered, and the uncertainty weight is calculated, thereby further reducing the influence of uncertainty on the system.

Description

Technical Field

[0001] The present invention relates to the technical field of shared energy storage scheduling, and in particular to a shared energy storage optimization control method based on IGDT and master-slave game. Background Art

[0002] With the continuous growth of energy demand and the gradual transformation of energy structure, the application of distributed energy systems is becoming more and more extensive. Among them, shared energy storage and microgrid technology have become key technologies for achieving coordination and optimization of distributed energy systems.

[0003] Shared energy storage refers to connecting multiple different energy storage devices together in some way to form a centralized energy storage system to achieve energy sharing and optimization. Shared energy storage helps balance the volatility of renewable energy and improve the flexibility and stability of the system. Microgrids can buy and sell stored energy from shared energy storage when there is an imbalance in energy supply and demand to maintain the balance of the power system.

[0004] However, in the pursuit of sustainable development and the urgent need to reduce carbon emissions, the intermittent and uncertain nature of renewable energy will have a significant impact on the stable operation and reasonable scheduling of the system. Therefore, how to reduce the impact of uncertainty in renewable energy and optimize the regulation strategy of shared energy storage has become an issue worth considering. Summary of the invention

[0005] The purpose of the present invention is to provide a shared energy storage optimization and control method based on IGDT and master-slave game. Shared energy storage can handle the uncertainty of multiple uncertain variables such as photovoltaics, promote the consumption of renewable energy, and reduce the cost of each microgrid in the integrated energy system. According to the strategic tendency of decision makers, a risk avoidance and risk opportunity model is established, the impact of uncertain variables in the microgrid on the system and itself is considered, and the uncertainty weight is calculated, thereby further reducing the impact of uncertainty on the system.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A shared energy storage optimization control method based on IGDT and master-slave game includes the following steps:

[0008] Obtain relevant data, including the power and capacity configuration and charging and discharging efficiency of shared energy storage, forecast data such as photovoltaic and load output power within each microgrid, and grid electricity prices in different time periods, and establish a model corresponding to the integrated energy system;

[0009] A two-layer master-slave game model is established and solved using the Kriging model as the deterministic model of IGDT;

[0010] Establish an IGDT model and calculate the uncertainty weights based on the microgrid's output and energy consumption in the system;

[0011] Establish risk aversion and risk opportunity models based on user needs and solve them.

[0012] As a further solution of the present invention: the master-slave game model is divided into two layers according to different subjects:

[0013] The upper layer is dominated by shared energy storage operators, who are leaders in the master-slave game and aim to maximize their own gains;

[0014] The lower layer is based on each microgrid as the main body and is a follower in the master-slave game, with the goal of minimizing the operating costs of each microgrid itself.

[0015] As a further solution of the present invention: establishing a master-slave game upper model;

[0016] Based on adjusting the transaction price with the microgrid, selling and purchasing energy to maximize the economy, that is, through the formula:

[0017] max C SESS =C SESS,income -C SESS,OM

[0018]

[0019] Among them, C SESS,OM and C SESS,income is the maintenance cost of shared energy storage and the revenue from transactions with the microgrid, T and I are the operation cycle and the number of microgrid users; P t SESS,ch and P t SESS,dis The charging and discharging power of the shared energy storage; and is the power from shared energy storage to microgrid i and the power from microgrid i to shared energy storage at time t; SESS,OM is the operation and maintenance cost coefficient; and is the price at which microgrid i sells or buys electricity from shared energy storage at time t; Δt is the time interval.

[0020] As a further solution of the present invention: the process of acquiring the charging and discharging power of the shared energy storage is:

[0021] When operating, shared energy storage needs to constrain its charging and discharging power within the rated range, that is:

[0022]

[0023] Among them, P t SESS,ch and P t SESS,disis the charging and discharging power traded between the shared energy storage and all microgrids at time t; and is the rated charging and discharging power of the shared energy storage.

[0024] As a further solution of the present invention: under the charge state constraint of the shared energy storage, it is necessary to keep the shared energy storage from being overcharged or over-discharged, and to maintain the continuity of its energy storage, that is, the initial charge state of each cycle is the same, that is:

[0025] in: and is the charge state of the shared energy storage at time 0 and time t; SESS,ch and λ SESS,dis the efficiency of charging and discharging shared energy storage; Configure capacity for shared energy storage.

[0026] As a further solution of the present invention: establishing a master-slave game lower layer model;

[0027] Based on the fact that multiple microgrids can interact with shared energy storage and the power grid respectively, the energy transaction volume with shared energy storage and the power grid is adjusted to reduce the cost, that is, to construct the objective function:

[0028]

[0029] in, and They are the revenues from microgrid i, grid, and shared energy storage transactions respectively; is the power of energy transferred from A to B in microgrid i at time t; is the transaction electricity price at the corresponding time; MG,OM is the microgrid operation and maintenance coefficient.

[0030] As a further solution of the present invention: the power balance constraint of the microgrid in microgrid i is:

[0031]

[0032] in, and is the power of photovoltaic and load of microgrid i at time t;

[0033] The internal power constraint of the microgrid in microgrid i is:

[0034]

[0035] in, and is the maximum power of the microgrid i photovoltaic and load;

[0036] The charging and discharging power and signal constraints of the microgrid in microgrid i are:

[0037]

[0038] in, is the maximum power of energy transferred from A to B in microgrid i; is a binary variable, indicating the direction of energy transmission between A and B in microgrid i at time t. It indicates that the energy is transferred from the grid to microgrid i at the current moment.

[0039] As a further solution of the present invention: based on the IGDT model, the uncertain weight is calculated according to the output and energy consumption of the microgrid in the system;

[0040] Right now:

[0041] Among them, X and d are uncertainty variables and decision variables; U is the uncertainty set;

[0042] The uncertainty of the variables is:

[0043]

[0044] Establish uncertainty sets for PV and load:

[0045]

[0046] in, and is the predicted output power of load and photovoltaic; P t Load and P t PV is the actual output power of the load and photovoltaic; Load and PV is the uncertainty radius of load and PV;

[0047] ε=σ Load ξ Load +σ PV ξ PV

[0048] Among them, ε is the comprehensive uncertainty radius defined according to the uncertainty radius of load and photovoltaic, which is used to measure the uncertainty of microgrid i; σ Load and Load For and PV The corresponding weight of .

[0049] As a further solution of the present invention: establish a risk avoidance model, namely:

[0050] maxε

[0051]

[0052] Among them, ε is the comprehensive uncertainty radius; and are the operating costs of microgrid i after IGDT optimization and the operating costs of microgrid i in the deterministic model after only master-slave game optimization; θ RA is the risk aversion coefficient;

[0053] The system constraints include power constraints of equipment within the microgrid, charging and discharging action constraints, and shared energy storage charge state constraints.

[0054] As a further solution of the present invention: establish a risk opportunity model, namely:

[0055] minε

[0056]

[0057] Among them, θ RS is the risk opportunity coefficient.

[0058] Beneficial effects of the invention: The invention starts from the economic perspective of each subject in the integrated energy system, takes into account the characteristics of shared energy storage and microgrid power trading that can absorb renewable energy, and establishes a two-layer master-slave game optimization scheduling model with shared energy storage as the leader and multiple microgrids as followers. The model can make full use of shared energy storage for energy transmission between microgrids, reduce the operating cost of microgrids, and also alleviate the volatility of renewable energy to a certain extent;

[0059] The present invention also proposes a risk aversion and risk opportunity model based on IGDT in response to the demand preferences of decision makers, so as to meet the different preferences of decision makers for system economy and stability. Furthermore, by calculating and determining the weights, the influence of uncertain variables on the system and microgrid can be effectively reflected, so that the optimization of IGDT scheduling is more reasonable and can better meet the needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The present invention will be further described below in conjunction with the accompanying drawings.

[0061] Figure 1 It is a flow chart of a shared energy storage optimization control method based on IGDT and master-slave game according to an embodiment of the present invention;

[0062] Figure 2 It is a comprehensive energy system structure diagram and a master-slave game architecture in a shared energy storage optimization control method based on IGDT and master-slave game in an embodiment of the present invention;

[0063] Figure 3It is a flow chart of solving a double-layer master-slave game model using a 3Kriging metamodel in a shared energy storage optimization control method based on IGDT and master-slave game in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] See also Figure 1-Figure 3 As shown, the present invention is a shared energy storage optimization control method based on IGDT and master-slave game, comprising the following steps:

[0066] S1: Establish a corresponding mathematical model based on the integrated energy system.

[0067] S2: Establish a two-layer master-slave game model and use the Kriging model to solve it as the deterministic model of IGDT.

[0068] S3: Establish an IGDT model and calculate the uncertain weights based on the output and energy consumption of the microgrid in the system.

[0069] S4: Establish risk aversion and risk opportunity models based on user needs and solve them.

[0070] Next, the method of the present invention is described in detail.

[0071] S1: Obtain relevant data, including the power and capacity configuration and charging and discharging efficiency of shared energy storage, forecast data such as photovoltaic and load output power within each microgrid, and grid electricity prices in each period, and establish a model corresponding to the integrated energy system.

[0072] S2: Establish a two-layer master-slave game model and use the Kriging model to solve it as the deterministic model of IGDT.

[0073] The master-slave game model in S2 is divided into two layers according to different subjects: the upper layer is based on the shared energy storage operator, which is the leader of the master-slave game and aims to maximize its own gains; the lower layer is based on each microgrid, which is the follower of the master-slave game and aims to minimize the operating costs of each microgrid.

[0074] S2.1 Establishing the upper-level model of master-slave game

[0075] Objective Function

[0076] Shared energy storage can trade energy with multiple microgrids, and the main expenditure is its operation and maintenance cost. Therefore, the objective function can be as shown in formula (1), which maximizes its own economy by adjusting the transaction electricity price with the microgrid and selling and purchasing energy.

[0077] maxC SESS =C SESS,income -C SESS,OM (1)

[0078]

[0079] Where: C SESS,OM and C SESS,income is the maintenance cost of shared energy storage and the revenue from transactions with the microgrid. T and I are the operation cycle and the number of microgrid users; P t SESS,ch and P t SESS,dis The charging and discharging power of the shared energy storage; and is the power from shared energy storage to microgrid i and the power from microgrid i to shared energy storage at time t; SESS,OM is the operation and maintenance cost coefficient; and is the price at which microgrid i sells or buys electricity from shared energy storage at time t; Δt is the time interval.

[0080] Constraints

[0081] When building a scheduling model for shared energy storage, its charging and discharging constraints and state of charge constraints also need to be considered.

[0082] Charge and discharge power constraints

[0083] Shared energy storage needs to constrain its charging and discharging power within the rated range during operation.

[0084]

[0085]

[0086] Where: P t SESS,ch and P t SESS,dis is the charging and discharging power traded between the shared energy storage and all microgrids at time t; and is the rated charging and discharging power of the shared energy storage.

[0087] State of Charge Constraints

[0088] This patent mainly considers the state of charge constraints of shared energy storage to ensure that the shared energy storage will not be overcharged or over-discharged, and maintain the continuity of its energy storage, that is, the initial state of charge of each cycle is the same.

[0089]

[0090] Where: and is the charge state of the shared energy storage at time 0 and time t; SESS,ch and λ SESS,dis the efficiency of charging and discharging shared energy storage; Configure capacity for shared energy storage.

[0091] S2.2 Establishing the lower-level model of master-slave game

[0092] Objective Function

[0093] In this scenario, there are multiple microgrids that can interact with shared energy storage and the grid respectively, and reduce costs by adjusting the energy transaction volume with shared energy storage and the grid. The objective function is as follows:

[0094]

[0095] Where: and They are the revenues from microgrid i, grid, and shared energy storage transactions respectively; is the power of energy transferred from A to B in microgrid i at time t; is the transaction electricity price at the corresponding time; MG,OM is the microgrid operation and maintenance coefficient.

[0096] Constraints

[0097] The main constraints considered in microgrid i are the power balance constraint and internal power constraint of the microgrid. Since it is necessary to ensure that the energy transmission between microgrid i and the power grid and shared energy storage cannot be charged and discharged at the same time, binary variables are used to constrain it.

[0098] 1) Power balance constraints

[0099]

[0100] Where: and is the power of photovoltaic and load of microgrid i at time t.

[0101] 2) Internal power constraints

[0102]

[0103] Where: and is the maximum power of the microgrid i photovoltaic and load.

[0104] 3) Charge and discharge power and signal constraints

[0105]

[0106] Where: is the maximum power of energy transferred from A to B in microgrid i; is a binary variable, indicating the direction of energy transmission between A and B in microgrid i at time t, such as It means that at the current moment, energy is sent from the power plant to the microgrid i.

[0107] S2.3 Solving the two-layer master-slave game model using the Kriging metamodel

[0108] The Kriging metamodel is essentially a proxy model. A proxy model is a modeling and prediction method that can predict results close to the original model based on test points. Its core is to collect sample points through a special sampling method, construct a mathematical model that is similar to the original model, and fit the original complex model instead. In this patent, the solution can be obtained through the following steps:

[0109] Input relevant system data, such as photovoltaic forecast and load forecast, power purchase and sale prices from the power grid to the microgrid, etc.;

[0110] Initialize the sampling point R using Latin hypercube sampling e ;

[0111] Substitute the sampled electricity price into the lower model to obtain the electricity trading volume P e , get the initial data set;

[0112] Establish the Kriging metamodel of the lower model based on the initial data;

[0113] Bring the sample data set into the upper objective function, find the objective function value corresponding to each group of electricity price sample points, and divide it into i key areas;

[0114] In the key area, the optimal electricity price R of the upper objective function is searched by particle swarm algorithm. e,best ;

[0115] For the current optimal electricity price, the power transaction P solved by the lower optimization model is e,best Add it to the sample library as a new sample point, substitute it into the upper objective function to find the upper benefit, and compare it with the current optimal value. If the benefit is higher, replace the current optimal value, otherwise it remains unchanged;

[0116] The maximum value of the local optimal values ​​in all key areas is taken as the global optimal solution;

[0117] Determine whether the target meets the convergence conditions. If it does, output the optimal solution. Otherwise, return to step 4.

[0118] S3: Establish an IGDT model and calculate the uncertain weights based on the output and energy consumption of the microgrid in the system.

[0119] This patent uses IGDT to deal with the uncertainty of photovoltaics and loads. It has a low need for uncertain information, and decision makers can obtain operation scheduling strategies under different objectives based on their preferences.

[0120] The IGDT model consists of a system model and an uncertainty model, which can be expressed by equations (15) and (16):

[0121]

[0122] Where: X and d are uncertainty variables and decision variables; U is the uncertainty set

[0123] The uncertainty of a variable can usually be expressed as

[0124]

[0125] S3.1 Establishing the uncertainty set of photovoltaic and load

[0126]

[0127] Where: and is the predicted output power of load and photovoltaic; P t Load and P t PV is the actual output power of the load and photovoltaic; Load and PV is the uncertainty radius of load and PV.

[0128] ε=σ Load ξ Load +σ PV ξ PV (19)

[0129] Where: ε is the comprehensive uncertainty radius defined according to the uncertainty radius of load and photovoltaic, which is used to measure the uncertainty of microgrid i; σ Load and σ PV For Load and PV The corresponding weight of .

[0130] S3.2 Calculate the uncertainty weight σ based on the output and energy consumption of the microgrid in the system Load and σ PV

[0131] To calculate the weight, we must not only consider the impact of the photovoltaic and load uncertainty of microgrid i on the microgrid itself, but also its impact on the system. Therefore, we first calculate the output and energy consumption of the microgrid in the system, as shown in equations (20)-(23)

[0132]

[0133] Secondly, considering the impact of PV and load uncertainties on the microgrid itself, the calculations are as shown in equations (24)-(25):

[0134]

[0135] σ Load +σ PV =1 (25)

[0136] S4: Establish risk aversion and risk opportunity models based on user needs and solve them.

[0137] In S4, decision makers can adopt different decision-making methods according to their needs, establish risk aversion models or risk opportunity models, and solve strategies that meet their needs. The above model is a nonlinear mixed integer model and can be solved by commercial solvers such as GUROBI.

[0138] Building a risk aversion model

[0139] The risk aversion model can ensure that the optimization target is within an acceptable range and seek the maximum uncertainty radius of uncertain variables. The larger the uncertainty radius, the less sensitive the strategy is to uncertain fluctuations, and the more stable the system is with better robustness. That is, within an acceptable range, the economy is sacrificed to obtain system stability.

[0140] maxε (26)

[0141]

[0142] Where: ε is the comprehensive uncertainty radius; and are the operating costs of microgrid i after IGDT optimization and the operating costs of microgrid i in the deterministic model after only master-slave game optimization; θ RA is the risk aversion coefficient; the system constraints include power constraints of equipment in the microgrid, charging and discharging action constraints, and shared energy storage charge state constraints.

[0143] Build risk opportunity model

[0144] The purpose of the risk-opportunity model is to seek a better optimization target as much as possible. By minimizing the uncertainty radius as much as possible, the possibility of obtaining the best optimization target is increased. That is, better economic performance is achieved by sacrificing the stability of the system.

[0145] minε (28)

[0146]

[0147] Where: θ RS is the risk opportunity coefficient.

[0148] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A shared energy storage optimization control method based on IGDT and master-slave game, characterized in that: The following steps are involved: Obtain relevant data, including the power and capacity configuration and charging and discharging efficiency of shared energy storage, forecast data such as photovoltaic and load output power within each microgrid, and grid electricity prices in different time periods, and establish a model corresponding to the integrated energy system; A two-layer master-slave game model is established and solved using the Kriging model as the deterministic model of IGDT; Establish an IGDT model and calculate the uncertainty weights based on the microgrid's output and energy consumption in the system; Build risk avoidance and risk opportunity models based on user needs.

2. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 1, characterized in that: The master-slave game model is divided into two layers according to different subjects: The upper layer is dominated by shared energy storage operators, who are leaders in the master-slave game and aim to maximize their own gains; The lower layer is based on each microgrid as the main body and is a follower in the master-slave game, with the goal of minimizing the operating costs of each microgrid itself.

3. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 2, characterized in that: Establish the upper-level model of master-slave game; Based on adjusting the transaction price with the microgrid, selling and purchasing energy to maximize the economy, that is, through the formula: max C SESS =C SESS,income -C SESS，OM Among them, C SESS,OM and C SESS,income is the maintenance cost of shared energy storage and the revenue from transactions with the microgrid, T and I are the operation cycle and the number of microgrid users; P t SESS,ch and P t SESS,dis The charging and discharging power of the shared energy storage; and is the power from shared energy storage to microgrid i and the power from microgrid i to shared energy storage at time t; SESS,OM is the operation and maintenance cost coefficient; and is the price at which microgrid i sells or buys electricity from shared energy storage at time t; Δt is the time interval.

4. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 3, characterized in that: The process of obtaining the charging and discharging power of shared energy storage is as follows: When operating, shared energy storage needs to constrain its charging and discharging power within the rated range, that is: Among them, P t SESS,ch and P t SESS,dis is the charging and discharging power traded between the shared energy storage and all microgrids at time t; and is the rated charging and discharging power of the shared energy storage.

5. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 4, characterized in that: Under the state of charge constraint of shared energy storage, it is necessary to prevent the shared energy storage from being overcharged or over-discharged, and to maintain the continuity of its energy storage, that is, the initial state of charge of each cycle is the same, that is: in: and is the charge state of the shared energy storage at time 0 and time t; SESS,ch and λ SESS,dis the efficiency of charging and discharging shared energy storage; Configure capacity for shared energy storage.

6. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 2, characterized in that: Establish the lower-level model of master-slave game; Based on the fact that multiple microgrids can interact with shared energy storage and the power grid respectively, the energy transaction volume with shared energy storage and the power grid is adjusted to reduce the cost, that is, to construct the objective function: in, and They are the revenues from microgrid i, grid, and shared energy storage transactions respectively; is the power of energy transferred from A to B in microgrid i at time t; is the transaction electricity price at the corresponding time; MG,OM is the microgrid operation and maintenance coefficient.

7. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 6, characterized in that: The power balance constraint of the microgrid in microgrid i is: in, and is the power of photovoltaic and load of microgrid i at time t; The internal power constraint of the microgrid in microgrid i is: in, and is the maximum power of the microgrid i photovoltaic and load; The charging and discharging power and signal constraints of the microgrid in microgrid i are: in, is the maximum power of energy transferred from A to B in microgrid i; is a binary variable, indicating the direction of energy transmission between A and B in microgrid i at time t. It indicates that the energy is transferred from the grid to microgrid i at the current moment.

8. The shared energy storage optimization control method based on IGDT and master-slave game according to claim 1 is characterized in that: Based on the IGDT model, the uncertain weights are calculated according to the output and energy consumption of the microgrid in the system; Right now: Among them, X and d are uncertainty variables and decision variables; U is the uncertainty set; The uncertainty of the variables is: Establish uncertainty sets for PV and load: in, and is the predicted output power of load and photovoltaic; P t Load and P t PV is the actual output power of the load and photovoltaic; Load and PV is the uncertainty radius of load and PV; e=s Load x Load +s PV x PV in, ε is the comprehensive uncertainty radius defined according to the uncertainty radius of load and photovoltaic, which is used to measure the uncertainty of microgrid i; σ Load and Load For and PV The corresponding weight of .

9. The shared energy storage optimization control method based on IGDT and master-slave game according to claim 1 is characterized in that: Build a risk aversion model, namely: maxε Among them, ε is the comprehensive uncertainty radius; and are the operating costs of microgrid i after IGDT optimization and the operating costs of microgrid i in the deterministic model after only master-slave game optimization; θ RA is the risk aversion coefficient; The system constraints include power constraints of equipment within the microgrid, charging and discharging action constraints, and shared energy storage charge state constraints.

10. A shared energy storage optimization control method based on IGDT and master-slave game according to claim 9, characterized in that: Build a risk opportunity model, namely: minε Among them, θ RS is the risk opportunity coefficient.

Images