A Method and Trading System for Energy Storage Right Trading Combining Discrete and Continuous Time

Abstract

The present invention discloses a method for trading energy storage rights that combines discrete and continuous time and its trading system. First, a hybrid discrete / continuous time two-layer optimization model is constructed. The upper layer makes discrete-time optimization decisions on the arbitrage plan and flexibility price of the energy storage power station, and the lower layer makes continuous-time optimization decisions on the output plans and flexibility orders of each renewable energy station. Secondly, a solution algorithm for the hybrid-time two-layer optimization is proposed. The continuous-time optimization of the lower layer is converted into discrete-time optimization by using an enhanced solution space transformation. Then, the KKT conditions are used to transform the two-layer optimization into a single-layer optimization form. Finally, the optimal flexibility trading plan is obtained by solving the model with software. The present invention can make full use of the flexibility of the energy storage power station to enhance the regulation ability of the renewable energy power station, reduce the risk of power supply shortage of the renewable energy station while increasing the revenue of the energy storage power station.

Description

Technical Field

[0001] The present invention belongs to the field of electrical engineering technology, and more specifically, relates to a hybrid discrete and continuous time energy storage rights trading method and trading system. Background Art

[0002] Renewable energy is widely considered the future's primary energy source and should be responsible for ensuring sufficient electricity supply. However, due to the uncertainty of their own output, they will be the primary demander of flexibility. There are two ways for renewable energy power plants to assume this responsibility: first, by discarding excess output and penalizing output shortfalls; second, by enhancing their regulatory capabilities. Compared to the first approach, the second is more proactive and can actively support power balance. The most direct approach is to equip renewable energy power plants with a certain amount of flexibility resources. Given the grid-wide uniformity of power balance, flexibility from other components can also be used to improve the operation of renewable energy power plants. Shared energy storage is a means of achieving flexibility trading and can be categorized into three types: community energy storage, virtual energy storage plants, and cloud energy storage. Cloud energy storage not only provides grid-level services but also allows for the sale of flexibility to other users, potentially improving the operation of renewable energy power plants. Therefore, exploring flexibility trading between energy storage plants and renewable energy plants is of great significance.

[0003] At present, energy storage power stations and renewable energy power stations are generally built into a unified model. On the one hand, it cannot fully utilize forecast information, resulting in more aggressive bidding and trading strategies, which in turn causes a higher risk of power shortages. On the other hand, the economic efficiency of power supply is not high. Summary of the Invention

[0004] In response to the above-mentioned defects or improvement needs of the existing technology, the present invention provides a hybrid discrete and continuous time energy storage rights trading method and trading system, which aims to make full use of forecast information and hedge the power supply shortage risk under the day-ahead bidding strategy of renewable energy power stations by purchasing flexibility, thereby improving the security and economy of power supply.

[0005] To achieve the above objectives, according to one aspect of the present invention, a hybrid discrete-time and continuous-time energy storage rights trading method is provided, comprising:

[0006] A discrete-time upper-layer energy storage operation model is constructed for the energy storage power station. This upper-layer energy storage operation model uses the arbitrage plan and the flexibility unit price as decision variables, aims to maximize the sum of arbitrage returns and flexibility trading returns, and sets constraints on an operation mode in which part of the energy storage power station's electricity is sold as flexibility to renewable energy power stations, while the remaining part is used for the arbitrage plan.

[0007] Construct a continuous-time lower-layer new energy power station operation model for a renewable energy power station. The lower-layer new energy power station operation model uses the bidding strategy and flexibility order as decision variables, aims to maximize the revenue of the renewable energy power station, and sets constraints based on the operation mode where the renewable energy power station can flexibly allocate the purchased flexibility to hedge the power supply shortage risk. Among them, the revenue of the renewable energy power station includes the predicted output energy revenue, flexibility expenditure, and expected power supply shortage penalty;

[0008] Use enhanced solution space transformation to convert the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model;

[0009] Use the KKT conditions to merge the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model to obtain a single-layer optimization model;

[0010] Solve for the decision variables of the single-layer optimization model.

[0011] In one embodiment, the discrete-time upper-layer energy storage operation model includes:

[0012] Objective function:

[0013]

[0014] Among them, the subscripts i and t represent the renewable energy power station number and time period number respectively, T is the time period length, and R ESS is the total revenue of the energy storage power station, is the predicted electricity price at time period t, are the charging power and discharging power at time period t in the arbitrage plan respectively; is the flexibility unit price, is the flexibility order subscribed by the i-th renewable energy power station at time period t;

[0015] Constraint conditions:

[0016]

[0017] Among them, is the maximum energy available for the arbitrage plan at time period t, and are the maximum energy and minimum energy of the energy storage power station respectively; is the maximum power capacity available for the arbitrage plan at time period t, is the maximum power capacity of the energy storage power station.

[0018] In one embodiment, the discrete-time upper-layer energy storage operation model further includes constraint conditions for constraining the arbitrage process, which are:

[0019]

[0020]

[0021]

[0022]

[0023]

[0024] Among them, and are the energies in the t+1 period and the t period in the arbitrage plan respectively, and η c and η d are the charging efficiency and discharging efficiency of the energy storage power station respectively; and are the charging power and discharging power in the t+1 period in the arbitrage plan respectively, is the ramp limit for the energy storage power station to participate in energy arbitrage; I ea,t is the charging and discharging state during the arbitrage process, 1 represents charging, and 0 represents discharging; is the initial energy of the energy storage power station.

[0025] In one embodiment, the lower-layer new energy power station operation model for continuous time includes:

[0026] Objective function:

[0027]

[0028] Among them, τ is the time variable, H is the total scheduling period, represents the expected total revenue of the i-th renewable energy power station, is the day-ahead output plan of the i-th renewable energy power station at time τ, P r.s is the probability of the random scenario s, is the penalty coefficient for wind power shortage, is the power supply shortage of the i-th renewable energy power station at time τ;

[0029] Constraint conditions:

[0030]

[0031] Among them, is the real-time output of the i-th renewable energy power station at time τ, and are the charging power and discharging power of the purchased flexibility of the i-th renewable energy power station at time τ respectively, is the abandoned electricity of the i-th renewable energy power station at time τ.

[0032] In one embodiment, the lower-layer new energy power station operation model further includes constraint conditions for flexible scheduling constraints, which are:

[0033]

[0034]

[0035]

[0036] Among them, is the energy in the flexibility purchased by the i-th renewable energy power station at time τ under scenario s, and η c and η d are the charging efficiency and discharging efficiency of the energy storage power station respectively. is the ramp limit for the i-th renewable energy power station to order flexibility in the t time period.

[0037] In one embodiment, the discrete-time lower-layer new energy power station operation model obtained by enhanced solution space transformation is:

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044] Among them, is the vector composed of, represents the k-th interpolation coefficient of (·)(τ) in the t time period, W is a known constant matrix, and J is an enhanced matrix. are the powers P (·)t , corresponding to the 3rd and 4th control points in the t time period respectively, and (·)t .

[0045] In one embodiment, the process of obtaining a single-layer optimization model by using KKT condition conversion includes:

[0046] Simplify the discrete-time lower-layer new energy power station operation model to:

[0047]

[0048] s.t.Ai x i ≤B i :ω i

[0049] wherein, x i is a vector composed of decision variables in the lower-layer new energy power station operation model, A i , B i and F i are simplified coefficient matrices, and ω i is a dual variable vector;

[0050] Construct the Lagrange function:

[0051]

[0052] Obtain the KKT conditions of the lower-layer new energy power station operation model according to the Lagrange function:

[0053]

[0054] A i x i -B i ≤0⊥ω i ≥0

[0055] Reduce the bilevel optimization to a single-level optimization through the KKT conditions to obtain a single-level optimization model. The single-level optimization model is based on the discrete-time upper-layer energy storage operation model and adds the constraint conditions:

[0056] s.t. A i x i ≤B i :ω i

[0057]

[0058] A i x i -B i ≤0⊥ω i ≥0.

[0059] In one embodiment,

[0060] Replace the complementary slackness condition A i x i -B i ≤0⊥ω i ≥0 in the KKT conditions with the strong duality condition, which is:

[0061]

[0062] The flexibility unit price is expanded in binary to obtain the discrete-time flexibility unit price, which is:

[0063]

[0064] where is the minimum unit of the flexibility unit price, β p,k is a 0-1 variable, and N BE is the number of bits of the binary expansion;

[0065] Using the big M method, the variables obtained by introducing the binary expansion are given extremely large M as coefficients, and the non-linear problem is converted into a mixed-integer linear problem for solution.

[0066] In one embodiment, it is solved by software GUROBI based on the MATLAB platform.

[0067] According to another aspect of the present invention, a hybrid discrete and continuous-time energy storage right trading system is provided, including:

[0068] A first model construction unit for constructing a discrete-time upper-layer energy storage operation model for an energy storage power station. The upper-layer energy storage operation model uses the arbitrage plan and the flexibility unit price as decision variables, aims to maximize the sum of the arbitrage revenue and the flexibility trading revenue, and sets constraints on the operation mode in which a part of the power of the energy storage power station is sold as flexibility to a renewable energy power station and the other part is used for the arbitrage plan;

[0069] A second model construction unit for constructing a continuous-time lower-layer new energy power station operation model for a renewable energy power station. The lower-layer new energy power station operation model uses the bidding strategy and the flexibility order as decision variables, aims to maximize the revenue of the renewable energy power station, and sets constraints on the operation mode in which the renewable energy power station can flexibly control the purchased flexibility to hedge the power supply shortage risk. Among them, the revenue of the renewable energy power station includes the predicted output energy revenue, the flexibility expenditure, and the expected power supply shortage penalty;

[0070] A discrete-time conversion unit for converting the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model by using enhanced solution space transformation;

[0071] A single-layer optimization model conversion unit for merging the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model by using the KKT conditions to obtain a single-layer optimization model;

[0072] A solution unit for solving the decision variables for the single-layer optimization model.

[0073] Generally speaking, compared with the prior art by the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0074] 1. This application constructs two different operation models for energy storage power stations and renewable energy power stations respectively. On the one hand, the electricity of the energy storage power station is used for arbitrage to obtain benefits, and on the other hand, it is sold to the renewable energy power station for its flexible scheduling, which is called selling flexibility. The renewable energy power station selects a bidding strategy according to the day-ahead prediction information, that is, declares the future output plan. However, the real-time output of renewable energy usually differs from the output plan, and the excess will be abandoned, and the shortage will be punished. Considering the power supply reliability, the penalty coefficient is much greater than the wind abandonment coefficient, and a conservative bidding strategy is usually selected, that is, the declared planned output is relatively low to avoid the risk of power supply shortage as much as possible. Selecting a conservative bidding strategy can reduce the penalty, but it will lead to an increase in the abandonment volume and cause energy waste. In this application, by establishing a flexibility transaction between the energy storage power station and the renewable energy power station, the renewable energy power station can order some flexibility from the energy storage power station to hedge the risk of power supply shortage. Therefore, the bidding strategy can be optimized, which can not only increase the income of the energy storage power station, but also reduce the power supply shortage risk of the renewable energy power station and reduce the wind abandonment energy, and overall improve the security and economy of power supply.

[0075] 2. For the renewable energy power station, a continuous-time lower-layer new energy power station operation model is constructed. Compared with the discrete-time model, the continuous-time model can better utilize the prediction information to make decisions, identify more power supply shortage risks, select a more sensible day-ahead bidding and trading strategy for the renewable energy power station, and reduce the power supply shortage risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is a flowchart of the steps of a hybrid discrete and continuous-time energy storage right trading method according to an embodiment;

[0077] Figure 2 It is a block diagram of a hybrid discrete and continuous-time energy storage right trading system according to an embodiment;

[0078] FIG. 3(a) is a schematic diagram of a random scenario of the first renewable energy power station according to an embodiment;

[0079] FIG. 3(b) is a schematic diagram of a random scenario of the second renewable energy power station according to an embodiment;

[0080] FIG. 3(c) is a schematic diagram of a random scenario of the third renewable energy power station according to an embodiment;

[0081] Figure 4 It is a distribution diagram of predicted electricity prices according to an embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0082] To make the objectives, technical solutions and advantages of the present invention more clear and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0083] As Figure 1 shown is a step flow chart of a method for trading energy storage rights that combines discrete and continuous time in an embodiment, which mainly includes the following steps:

[0084] Step S100: Construct a discrete-time upper-layer energy storage operation model for the energy storage power station and a continuous-time lower-layer new energy power station operation model for the renewable energy power station.

[0085] The flexibility trading studied in the present invention occurs between the energy storage power station and the renewable energy power station, where the energy storage power station sells flexibility to obtain higher profits, and the renewable energy power station purchases flexibility to hedge the risk of power supply shortage. In flexibility trading, the energy storage power station determines the flexibility price and transmits it to the renewable energy power station for its flexibility order decision, and the flexibility order in turn affects the flexibility allocation of the energy storage power station.

[0086] For the energy storage power station, flexibility trading synergizes with its energy arbitrage. According to the predicted electricity price, the energy storage power station arranges its charge and discharge plan for energy arbitrage. Except for the flexibility occupied by the arbitrage plan, other flexibility can be allocated to different renewable energy power stations to obtain additional profits.

[0087] For the renewable energy power station, flexibility trading synergizes with its energy bidding. According to the predicted information, the renewable energy power station declares its output curve and benefits from the winning bid volume. Considering that the real-time output of renewable energy is usually different from the output plan, the excess will be abandoned and the shortage will be penalized. Considering the power supply reliability, the penalty coefficient is much larger than the wind abandonment coefficient, and a conservative bidding strategy is generally introduced. To improve the energy utilization rate and hedge the risk of power supply shortage, the renewable energy power station can order some flexibility from the energy storage power station and choose a less conservative bidding strategy.

[0088] (1.1) Upper-layer energy storage operation model

[0089] The upper-layer optimization takes the energy storage power station as the core, determines the arbitrage plan and flexibility price to maximize the profit of the energy storage power station, and constrains the operation of the energy storage power station. The upper layer adopts discrete-time optimization, and its model is as follows:

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097] Among them, the subscripts i and t represent the renewable energy power station and the time period number respectively. One energy storage power station can sell flexibility to multiple renewable energy power stations at the same time. The flexibility purchased by each renewable energy power station can be different. i represents the number of the renewable energy power station, and i = 1, 2, 3,.... Each scheduling period H can be divided into multiple time periods, and the length of each time period is T. t represents the number of the time period, and t = 1, 2, 3,.... R ESS is the total revenue of the energy storage power station, is the predicted electricity price in the t-th time period, are the charging power and discharging power in the t-th time period in the arbitrage plan respectively; is the unit price of flexibility, is the flexibility order ordered by the i-th renewable energy power station in the t-th time period. is the maximum energy available for the t-th time period in the arbitrage plan, and are the maximum energy and minimum energy of the energy storage power station respectively; is the maximum power capacity available for the t-th time period in the arbitrage plan, is the maximum power capacity of the energy storage power station. and are the energies in the (t + 1)-th time period and the t-th time period in the arbitrage plan respectively, η c and η d are the charging efficiency and discharging efficiency of the energy storage power station respectively; and are the charging power and discharging power in the (t + 1)-th time period in the arbitrage plan respectively, is the ramp-up limit for the energy storage power station to participate in energy arbitrage; I ea,t is the charge and discharge state during the arbitrage process. 1 represents charging, and 0 represents discharging; is the initial energy of the energy storage power station.

[0098] In the above model, the objective function (1) consists of arbitrage revenue and trading revenue, and is calculated as shown in its first and second terms. Equation (2) represents the flexibility allocation between energy arbitrage and flexibility trading. Equations (3)-(7) constrain the arbitrage process, where (3) tracks the SoC, which is the state-of-charge transfer constraint of the energy storage for adjacent time periods, and (4)-(7) represent the remaining flexibility for energy arbitrage, where (4) is the charge and discharge power constraint for the energy storage power station to participate in arbitrage, (5) restricts simultaneous charging and discharging, (6) restricts the energy participating in arbitrage, and (7) is the charge and discharge quantity balance constraint at the beginning and end of the trading cycle.

[0099] (1.2) Lower-level new energy station operation model

[0100] The lower-level optimization takes the renewable energy power station as the core, and decides the bidding strategy and flexibility order of the renewable energy power station, aiming to maximize the revenue of each renewable energy power station and constraining the operation of the renewable energy power station. The lower-level adopts continuous-time optimization, and the specific model is as follows:

[0101]

[0102]

[0103]

[0104]

[0105]

[0106] where τ is the time variable, H is the total scheduling period, represents the expected total revenue of the i-th renewable energy power station, is the day-ahead output plan of the i-th renewable energy power station at time τ, P r.s is the probability of the random scenario s, is the penalty coefficient for wind power shortage, is the power supply shortage of the i-th renewable energy power station at time τ, is the real-time output of the i-th renewable energy power station at time τ, and are the charging power and discharging power of the flexibility purchased by the i-th renewable energy power station at time τ, respectively, is the curtailed electricity of the i-th renewable energy power station at time τ. is the energy in the flexibility purchased by the i-th renewable energy power station at time τ under scenario s, η c and η d are the charging efficiency and discharging efficiency of the energy storage power station, respectively, is the ramping limit for the flexibility ordered by the i-th renewable energy power station in period t.

[0107] In the above model, the objective function (8) consists of three parts: energy revenue, flexibility expenditure, and expected power supply shortage penalty. Equation (9) represents the operation of each renewable energy power station. Equations (10)-(12) constrain the flexibility invocation, which is restricted by the day-ahead flexibility order.

[0108] In continuous-time form, the lower-layer optimization can capture more renewable energy output prediction information, which helps to select wise bidding and trading strategies to reduce the power supply shortage risk of renewable energy power stations. However, new algorithms are needed to solve the proposed hybrid discrete / continuous-time bi-level optimization model.

[0109] Step S200: Use the enhanced solution space transformation to convert the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model.

[0110] First, use the enhanced solution space transformation method to convert the continuous-time optimization of the lower layer into discrete-time optimization, which can eliminate the time variable τ. Divide the entire scheduling period H into several segments, each with a time length of T. The continuous-time optimization problem (8)-(12) of the lower layer can be converted into a discrete-time optimization problem (13)-(18).

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117] where is the vector composed of represents the k-th interpolation coefficient of (·)(τ) at time period t, W is a known constant matrix, J is an enhanced matrix, are the powers P corresponding to the 3rd and 4th control points in period t respectively (·)t , are the powers P corresponding to the 1st and 2nd control points in period t respectively (·)tEquation (18) constrains the first-order continuity between adjacent time periods, ensuring a smooth connection of the power curves for adjacent time periods. B0, B1, B2, and B3 are the serial numbers of the control points, and there are 4 control points numbered 0 - 3 in a time period t.

[0118] It can be seen that the transformed discrete-time optimization (13)-(18) is a general linear programming problem.

[0119] Step S300: Incorporate the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model using the KKT conditions to obtain a single-layer optimization model.

[0120] Reduce the two-layer optimization problem to a single-layer optimization. For renewable energy power station i, the transformed lower-layer optimization is abbreviated as follows:

[0121]

[0122] s.t. A i x i ≤ B i : ω i (20)

[0123] In the formula, x i is a vector composed of variables, and A i , B i and F i are the corresponding coefficient matrices. ω i is the dual variable vector.

[0124] Use the KKT (Karush-Kuhn-Tucker) conditions to transform the lower-layer optimization, and the derivation is as follows:

[0125]

[0126]

[0127] A i x i - B i ≤ 0 ⊥ ω i ≥ 0 (23)

[0128] In the formula, L i is the constructed Lagrange function.

[0129] Therefore, the two-layer optimization is reduced to a single-layer optimization:

[0130]

[0131] In one embodiment, it further includes transforming the above model into a mixed-integer linear programming by the strong duality theory and the binary expansion linearization method. There are some non-linear terms in Equation (24), including those in Equation (1) in Equation (5) and the complementary slackness condition (23), which will lead to computational difficulties. Therefore, some linearization techniques are adopted for the above non-linear terms.

[0132] 1) Strong duality condition: Considering that the transformed lower-level optimization (19)-(20) is convex, the complementary slackness condition (23) can be replaced by the strong duality condition (25).

[0133]

[0134] It can be seen that after replacing the complementary slackness condition, except for the bilinear term the other parts of Equation (25) are linear.

[0135] 2) Binary expansion: In order to linearize the above bilinear term, the flexibility unit price is pre-expanded in binary. For example, the power capacity price is discretized into Equation (26).

[0136]

[0137] where is the minimum unit of the power capacity price, β pk is a 0, 1 variable. N BE is the number of bits of the binary expansion. Considering that in practice, the price usually has a minimum unit, such as $0.01, Equation (26) is reasonable.

[0138] 3) Big M method: Through binary expansion, new bilinear terms are introduced. For example, can be replaced by Equation (27).

[0139]

[0140] The new bilinear term can be linearized by the big M method. By introducing artificial variables and using a very large M as the coefficient, the problem is transformed into a mixed-integer linear problem for solution, such as Equation (28).

[0141]

[0142] The bilinear term in Equation (5) can be directly linearized by the big M method.

[0143] Using the above linearization techniques, Equation (24) can be converted into a mixed-integer linear programming that is easy to solve with existing commercial solvers.

[0144] It should be noted that steps S200 and S300 are to optimize and merge the initial model. Among them, the enhanced solution space transformation and the KKT condition are both disclosed in the prior art. Therefore, the specific process introduced in the above embodiments can be adopted, or the traditional existing method can be used for processing. On the premise of clearly using the enhanced solution space transformation and the KKT condition for processing, the specific processing process is not limited. Those skilled in the art process the above model based on the enhanced solution space transformation and the KKT condition in the prior art.

[0145] Step S400: Solve the decision variables for the single-layer optimization model.

[0146] Equation (24) is the flexibility trading model proposed in this paper between the energy storage power station and the renewable energy power station. This model can be solved through the commercial software GUROBI based on the MATLAB platform to obtain the flexibility trading results between the energy storage power station and the renewable energy power station, so as to increase the revenue of the energy storage power station and reduce the power supply shortage risk of the renewable energy power station.

[0147] Correspondingly, the present application also protects a hybrid discrete and continuous time energy storage right trading system, as Figure 2 shown, which mainly includes:

[0148] The first model construction unit is used to construct a discrete-time upper-layer energy storage operation model for the energy storage power station. The upper-layer energy storage operation model takes the arbitrage plan and the flexibility unit price as decision variables, aims to maximize the sum of the arbitrage revenue and the flexibility trading revenue, and sets constraints on the operation mode of taking a part of the power of the energy storage power station as flexibility to sell to the renewable energy power station and the other part for the operation of the arbitrage plan;

[0149] The second model construction unit is used to construct a continuous-time lower-layer new energy power station operation model for the renewable energy power station. The lower-layer new energy power station operation model takes the bidding strategy and the flexibility order as decision variables, aims to maximize the revenue of the renewable energy power station, and sets constraints on the operation mode that the renewable energy power station can flexibly control the purchased flexibility to hedge the power supply shortage risk. Among them, the revenue of the renewable energy power station includes the predicted output energy revenue, the flexibility expenditure, and the expected power supply shortage penalty;

[0150] The discrete-time conversion unit is used to convert the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model by using the enhanced solution space transformation;

[0151] The single-layer optimization model conversion unit is used to merge the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model by using the KKT condition to obtain a single-layer optimization model;

[0152] A solution unit for solving decision variables for a single-layer optimization model.

[0153] The above trading system is used to execute the trading method introduced above. Each unit therein is used to implement each step in the corresponding method. The specific function implementation of each unit can be referred to the above introduction and will not be elaborated here.

[0154] The following uses a specific embodiment to verify the advantages of this application.

[0155] In flexible trading, consider an energy storage power station with a capacity of 40 MW / 80 MWh and three renewable energy power stations with a capacity of 80 MW each. For the energy storage power station, the ramping capacity is assumed to be infinite, and the charge-discharge efficiency is set at 95%. The minimum SoC is 20%, and the initial SoC is 60%. For each renewable energy power station, according to the power output prediction and historical error, 5 random scenarios with a 5-minute accuracy are generated. The prediction and historical error data are both from the US PJM market, and the generated scenarios are as attached Figures 3(a) to 3(c) ... The predicted electricity prices used to calculate the energy storage arbitrage and renewable energy generation revenue are given in the attachment Figure 4 ... After the day-ahead bidding and trading strategy decisions of the renewable energy power stations are completed, a 5-minute economic dispatch is performed to verify the real-time power supply shortage risk. To ensure power supply reliability, referring to the load shedding penalty, the penalty coefficient for power supply shortage is set at 1000 $ / MWh.

[0156] Compare the revenue situations of the traditional energy storage power station and renewable energy power stations (RPP) without flexible trading and those with flexible trading in this invention. The results are shown in Table 1 below.

[0157] Table 1 Simulation Results

[0158]

[0159] According to the simulation results, through flexible trading, the energy storage power station sells part of its flexibility to the renewable energy power stations, the total revenue is significantly increased, and the power shortage risk is significantly reduced. Through hybrid discrete / continuous time optimization, more power supply shortage risks can be identified from the prediction information, the flexibility value is evaluated higher, and more sensible bidding and trading strategies are selected. Based on the above analysis, the effectiveness of the proposed method in promoting flexible trading to increase the revenue of energy storage power stations and reduce the power supply shortage risk of renewable energy power stations can be verified.

[0160] Those skilled in the art can easily understand that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for trading energy storage rights that mixes discrete and continuous time, characterized in that, Including: Construct a discrete-time upper-layer energy storage operation model for the energy storage power station. The upper-layer energy storage operation model uses the arbitrage plan and the flexibility unit price as decision variables, aims to maximize the sum of the arbitrage income and the flexibility trading income, and sets the constraint conditions for the operation mode in which a part of the power of the energy storage power station is sold as flexibility to the renewable energy power station and the other part is used for the arbitrage plan; Construct a continuous-time lower-layer new energy power station operation model for the renewable energy power station. The lower-layer new energy power station operation model uses the bidding strategy and the flexibility order as decision variables, aims to maximize the income of the renewable energy power station, and sets the constraint conditions for the operation mode in which the renewable energy power station can flexibly control the purchased flexibility to hedge the power supply shortage risk. Among them, the income of the renewable energy power station includes the predicted output energy income, the flexibility expenditure, and the expected power supply shortage penalty; Convert the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model by using the enhanced solution space transformation; Merge the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model by using the KKT conditions to obtain a single-layer optimization model; Solve the decision variables for the single-layer optimization model; The process of obtaining the single-layer optimization model by using the KKT conditions conversion includes: Simplify the discrete-time lower-layer new energy power station operation model to: s.t.A i x i ≤B i :ω i where x i is a vector composed of decision variables in the operation model of the lower-layer new energy power station, A i , B i and F i are simplified coefficient matrices, ω i is a vector of dual variables, and i represents the number of the new energy power station; Construct the Lagrange function L i : Obtain the KKT conditions of the lower-layer new energy power station operation model according to the Lagrange function; A i x i -B i ≤0⊥ω i ≥0 Reduce the double-layer optimization to a single-layer optimization through the KKT conditions to obtain a single-layer optimization model. The single-layer optimization model is based on the discrete-time upper-layer energy storage operation model and adds the constraint conditions: s.t.A i x i ≤B i :ω i A i x i -B i ≤0⊥ω i ≥0; Apply condition A i x i -B i ≤0⊥ω i ≥0 is replaced with the strong duality condition: Perform a binary expansion on the flexibility unit price to obtain the discrete-time flexibility unit price It is: In the formula, is the minimum unit of the flexibility unit price, and β p,k is a 0-1 variable, and N BE is the number of bits in the binary expansion; Using the big M method, use a very large M as the coefficient for the variables obtained by introducing the binary expansion to convert the non-linear problem into a mixed integer linear problem.

2. The hybrid discrete and continuous-time energy storage rights trading method according to claim 1, characterized in that The discrete-time upper-layer energy storage operation model includes: Objective function: where the subscripts \(i\) and \(t\) represent the renewable energy power station number and the time period number respectively, \(T\) is the time period length, \(R\) ESS is the total revenue of the energy storage power station, is the predicted electricity price in period \(t\), are the charging power and discharging power in period \(t\) of the arbitrage plan respectively; is the flexibility unit price, is the flexibility order subscribed by the \(i\)-th renewable energy power station in period \(t\); Constraint conditions: Among them, is the maximum energy available for the t-th period in the arbitrage plan, and are the maximum energy and minimum energy of the energy storage power station respectively; is the maximum power capacity available for the t-th period in the arbitrage plan, is the maximum power capacity of the energy storage power station.

3. The hybrid discrete and continuous-time energy storage rights trading method according to claim 2, wherein The discrete-time upper-layer energy storage operation model also includes the constraint conditions for restricting the arbitrage process, which are: Among them, and are the energies in the t+1 period and the t period in the arbitrage plan, respectively, and η c and η d are the charging efficiency and the discharging efficiency of the energy storage power station, respectively; and are the charging power and the discharging power in the t+1 period in the arbitrage plan, respectively, is the ramp limit for the energy storage power station to participate in energy arbitrage; I ea,t is the charge-discharge state during the arbitrage process, where 1 represents charging and 0 represents discharging; is the initial energy of the energy storage power station.

4. The hybrid discrete and continuous time energy storage rights trading method according to claim 3, wherein The continuous-time lower-layer new energy power station operation model includes: Objective function: where τ is the time variable and H is the total scheduling period, represents the expected total revenue of the i-th renewable energy power station, is the day-ahead output plan of the i-th renewable energy power station at time τ, P r.s is the probability of the random scenario s, is the penalty coefficient for wind power shortage, is the power supply shortage of the i-th renewable energy power station at time τ; Constraint conditions: Among them, is the real-time output of the i-th renewable energy power station at time τ, and are the charging power and discharging power of the flexibility purchased by the i-th renewable energy power station at time τ, respectively, is the curtailed electricity of the i-th renewable energy power station at time τ.

5. The hybrid discrete and continuous-time energy storage rights trading method according to claim 4, wherein The lower-layer new energy power station operation model also includes the constraint conditions for restricting the flexibility scheduling constraint, which are: Among them, is the energy in the flexibility purchased by the i-th renewable energy power station at time τ under scenario s, η c and η d are the charging efficiency and discharging efficiency of the energy storage power station respectively, is the ramping limit of the flexibility ordered by the i-th renewable energy power station in time period t.

6. The method for trading energy storage rights by mixing discrete and continuous time as claimed in claim 5, wherein The discrete-time lower-layer new energy power station operation model obtained by the enhanced solution space transformation is: Among them, is a vector composed of indicating the k-th interpolation coefficient of (·)(τ) in the t time period, W is a known constant matrix, and J is an enhancement matrix, are the powers P corresponding to the 3rd and 4th control points in the t time period (·)t , are the powers P corresponding to the 1st and 2nd control points in the t time period (·)t .

7. The hybrid discrete and continuous-time energy storage rights trading method according to claim 1, characterized in that Based on the MATLAB platform, solve it through the software GUROBI.

8. A hybrid discrete and continuous-time energy storage rights trading system, characterized in that, Including: The first model construction unit is used to construct a discrete-time upper-layer energy storage operation model for the energy storage power station. The upper-layer energy storage operation model uses the arbitrage plan and the flexibility unit price as decision variables, aims to maximize the sum of the arbitrage income and the flexibility trading income, and sets the constraint conditions for the operation mode in which a part of the power of the energy storage power station is sold as flexibility to the renewable energy power station and the other part is used for the arbitrage plan; The second model construction unit is used to construct a continuous-time lower-layer new energy power station operation model for a renewable energy power station. The lower-layer new energy power station operation model takes the bidding strategy and flexibility order as decision variables, aims to maximize the revenue of the renewable energy power station, and sets constraint conditions according to the operation mode that the renewable energy power station can flexibly control the purchased flexibility to hedge the power supply shortage risk. Among them, the revenue of the renewable energy power station includes predicted output energy revenue, flexibility expenditure, and expected power supply shortage penalty; The discrete-time conversion unit is used to convert the continuous-time lower-layer new energy power station operation model into a discrete-time lower-layer new energy power station operation model by using enhanced solution space transformation; The single-layer optimization model conversion unit is used to merge the discrete-time lower-layer new energy power station operation model into the upper-layer energy storage operation model by using the KKT conditions to obtain a single-layer optimization model; The solving unit is used to solve the decision variables for the single-layer optimization model; Among them, the process of the single-layer optimization model conversion unit obtaining the single-layer optimization model by using the KKT conditions includes: Simplify the discrete-time lower-layer new energy power station operation model to: s.t.A i x i ≤B i :ω i where \(x\) i is a vector composed of decision variables in the lower-layer new energy power station operation model, \(A\) i , \(B\) i and \(F\) i are simplified coefficient matrices, \(\omega\) i is a vector of dual variables, and \(i\) represents the number of the new energy power station; Construct the Lagrange function L i : Obtain the KKT conditions of the lower-layer new energy power station operation model according to the Lagrange function: A i x i -B i ≤0⊥ω i ≥0 Reduce the double-layer optimization to single-layer optimization through the KKT conditions to obtain a single-layer optimization model. The single-layer optimization model is based on the discrete-time upper-layer energy storage operation model and adds constraint conditions: s.t.A i x i ≤B i :ω i A i x i -B i ≤0⊥ω i ≥0; Apply condition A i x i -B i ≤0⊥ω i ≥0 is replaced with the strong duality condition: Perform binary expansion on the flexibility unit price to obtain the discrete-time flexibility unit price It is as follows: In the formula, is the minimum unit of the flexibility unit price, β p,k is a 0, 1 variable, N BE is the number of bits in the binary expansion; Using the big M method, use a very large M as the coefficient for the variables obtained by introducing the binary expansion to convert the non-linear problem into a mixed-integer linear problem.

Images